text
stringlengths
14
5.77M
meta
dict
__index_level_0__
int64
0
9.97k
(function () { "use strict"; function UnpublishController($scope, localizationService, contentEditingHelper) { var vm = this; var autoSelectedVariants = []; vm.id = $scope.content.id; vm.warningText = null; vm.changeSelection = changeSelection; function onInit() { $scope.model.hideSubmitButton = true; vm.variants = $scope.model.variants; vm.unpublishableVariants = vm.variants.filter(publishedVariantFilter) // set dialog title if (!$scope.model.title) { localizationService.localize("content_unpublish").then(value => { $scope.model.title = value; }); } vm.variants.forEach(variant => { variant.isMandatory = isMandatoryFilter(variant); }); // node has variants if (vm.variants.length !== 1) { vm.unpublishableVariants = contentEditingHelper.getSortedVariantsAndSegments(vm.unpublishableVariants); var active = vm.variants.find(v => v.active); if (active && publishedVariantFilter(active) && allowUnpublish(active)) { //ensure that the current one is selected active.save = true; } // autoselect other variants if needed changeSelection(active); } } function allowUnpublish (variant) { return variant.allowedActions.includes("Z"); } function changeSelection(selectedVariant) { // if a mandatory variant is selected we want to select all other variants, we cant have anything published if a mandatory variants gets unpublished. // and disable selection for the others if (selectedVariant.save && selectedVariant.segment == null && selectedVariant.language && selectedVariant.language.isMandatory) { vm.variants.forEach(variant => { if (!variant.save) { // keep track of the variants we automaically select // so we can remove the selection again autoSelectedVariants.push(variant); variant.save = true; } variant.disabled = true; }); // make sure the mandatory isn't disabled so we can deselect again selectedVariant.disabled = false; } // if a mandatory variant is deselected we want to deselet all the variants // that was automatically selected so it goes back to the state before the mandatory language was selected. // We also want to enable all checkboxes again if (!selectedVariant.save && selectedVariant.segment == null && selectedVariant.language && selectedVariant.language.isMandatory) { vm.variants.forEach(variant => { // check if variant was auto selected, then deselect let autoSelected = autoSelectedVariants.find(x => x.culture === variant.culture); if (autoSelected) { variant.save = false; } variant.disabled = false; }); autoSelectedVariants = []; } // disable submit button if nothing is selected var firstSelected = vm.variants.find(v => v.save); $scope.model.disableSubmitButton = !firstSelected; //disable submit button if there is none selected } function isMandatoryFilter(variant) { //determine a variant is 'dirty' (meaning it will show up as publish-able) if it's // * has a mandatory language // * without having a segment, segments cant be mandatory at current state of code. return (variant.language && variant.language.isMandatory === true && variant.segment == null); } function publishedVariantFilter(variant) { //determine a variant is 'published' (meaning it will show up as able unpublish) // * it has been published // * it has been published with pending changes variant.notAllowed = allowUnpublish(variant) === false && variant.active; return (variant.state === "Published" || variant.state === "PublishedPendingChanges") && (allowUnpublish(variant) || variant.active); } //when this dialog is closed, remove all unpublish and disabled flags $scope.$on('$destroy', () => { vm.variants.forEach(variant => { variant.save = variant.disabled = false; variant.notAllowed = false; }); }); vm.checkingReferencesComplete = () => { $scope.model.hideSubmitButton = false; }; vm.onReferencesWarning = () => { $scope.model.submitButtonStyle = "danger"; // check if the unpublishing of items that have references has been disabled if (Umbraco.Sys.ServerVariables.umbracoSettings.disableUnpublishWhenReferenced) { // this will only be disabled if we have a warning, indicating that this item or its descendants have reference $scope.model.disableSubmitButton = true; } localizationService.localize("references_unpublishWarning").then((value) => { vm.warningText = value; }); }; onInit(); } angular.module("umbraco").controller("Umbraco.Overlays.UnpublishController", UnpublishController); })();
{ "redpajama_set_name": "RedPajamaGithub" }
621
Melocactus ernestii ist eine Pflanzenart aus der Gattung Melocactus in der Familie der Kakteengewächse (Cactaceae). Das Artepitheton ehrt den deutschen Botaniker und Entdecker Ernst H. G. Uhle (1854–1915). Beschreibung Melocactus ernestii wächst mit hell gelblichgrünen bis dunkelgrünen, nahezu kugelförmigen bis kurz zylindrischen Körpern, die Wuchshöhen von 9 bis 45 Zentimetern und Durchmesser von 7 bis 35 Zentimetern erreichen. Es sind 9 bis 13 mehr oder weniger gerundete Rippen vorhanden, die eine leicht scharfe Kante aufweisen. Die Dornen sind rot und gelb gebändert oder rötlich oder bräunlich gefärbt. Die 3 bis 8 Mitteldornen, von denen die untersten gebogen oder gerade sind, sind 3,2 bis 9 Zentimeter lang. Die 7 bis 13 geraden oder nach außen gebogenen Randdornen erreichen Längen von 4 bis 15 Zentimetern. Das aus mehr oder weniger rosaroten Borsten gebildete Cephalium, das an der Spitze nur selten mit weißer Wolle bedeckt ist, wird bis 18 Zentimeter hoch und erreicht Durchmesser bis 8 Zentimeter. Die hell bis dunkel rosamagentafarbenen Blüten sind 1,95 bis 2,9 Zentimeter lang und weisen Durchmesser von 9 bis 18 Millimeter auf. Die an der Spitze tiefrosafarbenen bis roten Früchte sind 1,4 bis 4,5 Zentimeter lang. Verbreitung, Systematik und Gefährdung Melocactus ernestii ist im Nordosten Brasiliens verbreitet. Die Erstbeschreibung erfolgte 1920 durch Friedrich Karl Johann Vaupel. Ein nomenklatorisches Synonym ist Melocactus oreas subsp. ernestii (1988). Es werden folgende Unterarten unterschieden: Melocactus ernestii subsp. ernestii Melocactus ernestii subsp. longicarpus In der Roten Liste gefährdeter Arten der IUCN wird die Art als "", d. h. als nicht gefährdet geführt. Nachweise Literatur Einzelnachweise Weiterführende Literatur Frederic Mendes Hughes, Marcelino De La Cruz Rot, Roberto Lisboa Romão, Marina Siqueira de Castro: Dinâmica espaço-temporal de Melocactus ernestii subsp. ernestii (Cactaceae) no Nordeste do Brasil. In: Revista Brasileira de Botânica. Band 34, Nummer 3, 2011, S. 389–402 (PDF; 610 kB). Weblinks Melocactus (Kakteen) Ernestii
{ "redpajama_set_name": "RedPajamaWikipedia" }
7,203
Q: Setting bootstrap dropdown width less than 240px I am using ngx-boostrap (AngularJS for Bootstrap) for drop-downs. I an trying to have the max-width of lets say 120px for my dropdown, but I never gets small than 240px. Is there any reason behind it This photo is the design and this photo is what I am building The html, js codes are as exactly as what is put in ngx-bootstrap, nothing fancy. For css, I am putting max-width=120px on everything inside the dropdown-menu and itself and I can not make that smaller. What is any remedy? A: You're on the right track! If you inspect .dropdown-menu you'll see it has the following value: min-width: 160px Since your max-width is less than that nothings going to happen; min-width is bigger and it's going to win. Easiest solution is going to be to reset that value in your override: .dropdown-menu { min-width: 0px; max-width: 120px; } Just keep in mind that any contents inside of .dropdown-menu that exceed this width will flow out of the container box.
{ "redpajama_set_name": "RedPajamaStackExchange" }
6,377
207th Street may refer to: Fordham Road, or 207th Street, a major street in The Bronx borough of New York City 207th Street Yard, a rail yard of the New York City Subway system New York City Subway 207th Street (IND Eighth Avenue Line); the northern terminal of the train 207th Street (IRT Broadway – Seventh Avenue Line); serving the train
{ "redpajama_set_name": "RedPajamaWikipedia" }
5,034
\section{Introduction} \label{sec:intro} Point cloud is a simple and compact geometric representation of 3D objects, and has been broadly used as the standard output of various sensors. In recent years, the analysis of point clouds has gained much attention due to its wide application in real world problems such as autonomous driving \cite{CVPR_autonomous_driving}, robotics \cite{Robotics}, and navigation \cite{Navigation}. However, it is nontrivial to solve such tasks using traditional deep learning tools, e.g., convolutional neural networks (CNNs). Unlike a 2D image with regularly packed pixels, a point cloud consists of sparse points without a canonical order. Moreover, the spatial distribution of a point cloud is heterogeneous due to factors in data acquisition, e.g., perspective effects and radial density variations. Due to the 3D nature of the problem, various methods have been proposed to convert a point cloud into a 3D volumetric representation, to which 3D CNNs are then applied \cite{3D_ShapeNets,VoxNet}. However, despite their success in analyzing 2D images, CNNs are not satisfactory in this context. The commonly used 3D CNN is extremely memory consuming, and thus can not be trained efficiently. A more serious issue is that converting a point cloud into a volumetric representation introduces quantization artifacts and loses fine-scale geometric details. \begin{figure*}[t!] \centering \includegraphics[width=\linewidth]{RCNet.pdf} \caption{The architecture of RCNet. In the recurrent set encoder, the ambient space of input points is partitioned into parallel beams, where the enclosed points are encoded by a shared RNN. The subregional features from each beam are later processed by a 2D CNN. Depending on the tasks, the aggregated global features are fed forward directly for shape prediction, or tiled and concatenated with the per-point features for semantic segmentation. The feature propagation refers to the operation that propagates the non-local features within each beam to the corresponding component points. The other operations used are: Conv (2D convolution), MP (2D max-pooling), FC (fully connected layer). Batchnorm and ReLU are used in all layers except the last one, and the shared FC is applied per point. Numbers in parentheses represent the size of operation, and the hidden size of RNN is 64 and 128 for classification and segmentation tasks, respectively. The STN block refers to the spatial transformer network \cite{STN,PointNet}. It outputs a transformation matrix and is comprised of a shared MLP(64, 128, 1024), a global max-pooling and another MLP(512, 256, $ d^2 $), where $ d $ is the number of features per input point. } \label{fig:architecture} \end{figure*} Better performance has been achieved by deep networks that avoid the volumetric convolutional architechture and operate directly on point clouds. Representative works include PointNet \cite{PointNet} and PointNet++ \cite{PointNet++}, which process point clouds by combining multi-layer perceptron (MLP) network with symmetric operations (e.g., max-pooling) to learn point features globally or hierarchically. Inspired by PointNet, several recent methods have been proposed to further improve the point feature representation \cite{Kernel-Graph,ShapeContextNet,SO-Net}. This class of networks are invariant to input permutation and have achieved state-of-the-art results. However, due to the reliance on the coarse feature pooling technique, they fail to fully exploit fine-scale geometric details. In this work, we aim to completely bypass the coarse pooling-based technique, and propose a new deep network for point cloud data. At the core of our method is a \textit{recurrent set encoder}, which divides the ambient domain into parallel \textit{beams} and encodes the points within each beam as \emph{subregional} geometric features with an RNN. Our key observation is that when the beam is of moderate size, the RNN is approximately dealing with a sequence of points, as a beam only contains points near a 1D line. Such a sequential input largely benefits the learning of RNN. Meanwhile, noticing that the beams are packed in a regular spatial layout, we use a 2D CNN to further analyze the beam features (called the \emph{convolutional feature aggregator}). Being efficient and powerful at feature learning, the 2D CNN can effectively aggregate the subregional features into a global one, while further benefiting the RNN learning in return. Our method (see Fig.~\ref{fig:architecture}) is surprisingly efficient and effective for point cloud processing. It is invariant to point order permutation, and competes favorably with the state-of-the-arts (SOTAs) in terms of both accuracy and computational efficiency. A few recent works also adopt convolution for point cloud processing. They typically utilize carefully designed domain transformations to map point data into suitable spaces, where convolution could be applied. Examples include SPLATNet \cite{SPLATNet} and PCNN \cite{PCNN_2018}. However, these methods are inefficient as they rely on sophisticated geometric transformations and complex convolutional operations, e.g., continuous volumetric convolution or sparse bilateral convolution. In contrast, our method only employs regular spatial partitioning and sorting, and leverages classic neural network architectures such as RNN and 2D CNN, which are well supported at both software and hardware levels. As a result, our network circumvents much implementation overhead and is significantly more efficient than these SOTAs in computation. It is worth mentioning that, our recurrent set encoder can be seen as a domain mapping function as well. But unlike these SOTAs, it is automatically learned via back-propagation instead of by careful handcrafted design. In this work, we focus on point cloud classification and segmentation tasks, and evaluate the proposed method on several datasets, including ModelNet10/40 \cite{3D_ShapeNets}, ShapeNet part segmentation \cite{ShapePartSeg}, and S3DIS \cite{S3DIS}. Experimental results demonstrate the superior performance of our method to the SOTAs in both accuracy and computational efficiency. In a nutshell, our main contributions are as follows: \begin{itemize} \item We present a new architecture that operates directly on point clouds without relying on symmetric functions (e.g., max-pooling) to achieve permutation invariance. \item We propose a recurrent set encoder for effective subregional feature extraction. To the best of our knowledge, this is the first time an RNN is effectively employed to model point clouds directly. \item We propose to introduce the 2D CNN for aggregating subregional features. This design maximally utilizes the strengths of CNN while further benefiting the RNN encoder. The resulting network is efficient as well as effective at hierarchical and spatially-aware feature learning. \end{itemize} \section{Related Work} \label{sec:related_work} We briefly review the existing deep learning approaches for 3D shape processing, with a focus on point cloud setting. \subsubsection{Volumetric Methods} One classical approach to handling unstructured point clouds or meshes is to first rasterize them into regular voxel grids, and then apply standard 3D CNNs \cite{3D_ShapeNets,VoxNet,3D_CNN_LiDAR,Vol_MVCNN,Orientation_VoxelNet,Segcloud,3DCNN-DQN-RNN}. The major issue with such volumetric representations is that they tend to produce sparsely-occupied grids, which are unnecessarily memory-consuming. Besides, the grid resolutions are limited due to excessive memory and computational cost, causing quantization artifacts and loss of details. To remedy these issues, recent methods propose to adaptively partition the grids and place denser cells near the shape surface \cite{O-CNN,OctNet,Octree}. These methods suffer less from the computational and memory overhead, but still lose geometric details due to sampling and discretization. \subsubsection{View-based Methods} Another strategy is to encode the 3D shapes via a collection of 2D images which are rendered from different views. These rendered images can be fed into traditional 2D CNNs and processed via transfer learning, i.e., fine-tuning networks pre-trained on large-scale image datasets \cite{MVCNN,Vol_MVCNN,PCNN}. However, such view projections would lead to self-occlusions and consequently severe loss of geometric information. Moreover, view-based methods are mostly applied to classification tasks, and are hard to generalize to detail-focused tasks such as shape segmentation and completion. \subsubsection{Non-Euclidean Methods} These approaches build graphs from the input data (e.g., based on the mesh connectivity or k-nearest neighbor relationship), and apply CNNs to the graph spectral domain for shape feature learning \cite{GeoDL,LocalizedSpec_CNN,GraphCNN,FastGraphCNN,SemiSupervised_GraphCNN,AdapGraphCNN}. Graph CNN models are suitable for non-rigid shape analysis due to the isometry invariance. However, it is comparatively difficult to generalize these methods across non-isometric shapes with different structures, largely because the spectral bases are domain-dependent \cite{SyncSpecCnn}. \subsubsection{Point Cloud-based Methods} PointNet \cite{PointNet} pioneers a new type of deep neural networks that act directly on point clouds without data conversions. Its key idea is to learn per-point features independently, and then aggregate them in a permutation-invariant manner via a symmetric function, e.g., max-pooling. While achieving impressive performance, PointNet fails to capture crucial fine-scale structure details. To address this issue, the follow-up work PointNet++ \cite{PointNet++} exploits local geometric information by hierarchically stacking PointNets. This leads to improved performance, but at the cost of computational efficiency. Besides, since PointNet++ still treats points individually at local scale, the relationships among points are not fully captured. In light of the above challenges, a number of recent works have been proposed for better shape modeling \cite{KD_net,SO-Net,Kernel-Graph,SliceNet,ShapeContextNet,PCCN}. These methods overcome the weakness of coarse pooling operation at some degree, and achieve improved performance. Another class of methods have been recently developed without relying on pooling to guarantee permutation invariance. They typically transform the point data into another domain, where convolutions could be readily applied. In SPLATNet \cite{SPLATNet}, the source point samples are mapped into a high-dimensional lattice, where sparse bilateral convolution is employed for shape feature learning. In PCNN \cite{PCNN_2018}, a pair of extension and restriction operators are designed to translate between point clouds and volumetric functions, such that continuous volumetric convolution could be applied. Our method could be considered belonging to this category from the perspective of domain transformation. However, different from existing methods, our domain mapping function is automatically learned rather than by handcrafted design. Moreover, instead of utilizing complex convolutions, we employ the classic 2D convolution for feature aggregation. As a result, our method is more efficient in computation as well as effective at point feature learning. \section{Method} \label{sec:method} In this work, we focus on two tasks: point cloud classification and segmentation, and present two architectures correspondingly, as illustrated in Fig.~\ref{fig:architecture}. The input is a point set $ P = \{p_i \in \mathbb{R}^d, i=1, \cdots, N \} $, where each point $ p_i $ is a vector of coordinates plus additional features, such as normal and color. The output will be a $ 1\times K $ score vector for classification with $ K $ classes, or an $ N \times M $ score matrix for segmentation with $ M $ semantic labels. Our network, termed RCNet, consists of two components: the \textit{recurrent set encoder} and the \textit{convolutional feature aggregator}. The recurrent set encoder aims to extract subregional features from input point cloud, while convolutional feature aggregator is responsible for aggregating these extracted features hierarchically. Below we explain their details. \subsubsection{Recurrent Set Encoder} Given an unordered point set, the recurrent set encoder firstly partitions the ambient space into a set of parallel beams, and then divides the points into subgroups accordingly (see Fig.~\ref{fig:architecture}). The beams are uniformly distributed in a structured manner, spanning a 2D lattice. In particular, suppose the width, height and depth of a beam extends along $ x $, $ y $ and $ z $ axis, respectively. Let $ r $ and $ s $ be the hyper-parameters controlling the number of beams: $ w=(x_{max} - x_{min})/r $ and $ h=(y_{max} - y_{min})/s $, where $ w,h $ are the beam width and height; $ [x_{min}, x_{max}] $ and $ [y_{min}, y_{max}] $ are the maximum spanning ranges of points. Then a point with coordinate $ (x_k, y_k, z_k) $ is assigned to the $ (i,j) $-th beam if $ x_k - x_{min} \in [(i-1)w, iw) $ and $ y_k - y_{min} \in [(j-1)h, jh) $. In our implementation, since the point clouds are normalized to fit within a unit ball, we can simply set $ x_{min} = y_{min}=-1 $ and $ x_{max} = y_{max}=1 $. The subgroups of points are denoted by $ \{S_{ij}\}_{i=1, j=1}^{r,s} $. Note that depending on the tasks, it is also possible to perform non-uniform partition \cite{O-CNN}. In this work we only focus on uniformly partitioned beams. Given points in subgroup $ S_{ij} $, we treat them as a sequential signal and process it with an RNN. In particular, before being fed to RNN, points within each beam are sorted along the beam depth (according to their $ z $ coordinates). The RNN is single-directional, implemented using Gated Recurrent Units (GRU) \cite{GRU} with 2 layers. To the best of our knowledge, our network is the first to \textit{effectively} use an RNN to handle 3D point sets directly. Interestingly, it has been previously observed that an RNN performs poorly on a 3D point cloud due to the lack of a unique and stable ordering \cite{PointNet,OrderMatters}. The key to our success is the beam partition strategy. With the relatively dense partitioning, the points within each beam is of moderate size, and can be approximately considered distributed along a 1D line. In another word, the RNN is approximately handling point signal of moderate length in a 1D space. This facilitates the learning of RNN and makes it behave quite robustly with respect to the input perturbation. The output of recurrent set encoder is a grid of 1D feature vectors, which are taken as a 2D feature map and fed into the subsequent 2D CNN aggregator: \begin{equation} \label{eq:feature_map} I = \begin{bmatrix} \mathcal{R}(S_{11}) & \dots & \mathcal{R}(S_{1s})\\ \vdots & \ddots & \vdots \\ \mathcal{R}(S_{r1}) & \dots & \mathcal{R}(S_{rs}) \end{bmatrix}, \end{equation} where $ \mathcal{R} $ is a shared RNN with hidden size $ \ell $, and $ I \in \mathbb{R}^{r\times s \times \ell} $. Note that, we only utilize RNN to encode nonempty beams, and for those empty ones we pad zero vectors at the corresponding positions of $ I $. \subsubsection{Convolutional Feature Aggregator} We first note that the features encoded by RNN are actually \textit{non-local}, as the points within each beam span a large range along the beam depth. To build a global shape descriptor, we need to connect these non-local features. A natural choice is using 2D convolutional neural network, given the structured output $ I $ in Eq.(\ref{eq:feature_map}). Being efficient and powerful at multi-scale feature learning, a 2D CNN aggregator brings much computational and modeling advantage compared to the sophisticated aggregators in previous methods, as shown in the experiment section. Further, the strength of a 2D CNN alleviates the modeling burden of the recurrent encoder and boosts the overall performance. In this work, we utilize a simple shallow CNN architecture to validate our idea (see Fig.~\ref{fig:architecture}), and leave advanced architectures for future exploration. The aggregated global feature could be used for shape classification directly, or combined with the per-point features for semantic segmentation, as illustrated in Fig.~\ref{fig:architecture}. Note that, for segmentation task we inject additional subregional information into the points via feature propagation, so as to facilitate the discriminative point feature learning. \subsubsection{Remarks} We stress a few key properties of RCNet below. \begin{enumerate} \item It is invariant to point permutation, a result derived from point sorting within beams. \item The amount of context information embedded in the 2D feature maps can be controlled with beam sizes. Smaller beams would preserve richer spatial contexts while larger ones would contain less. In the extreme case, when the ambient space is trivially partitioned, i.e., there is only one beam, RCNet degenerates to the vanilla RNN model for point clouds \cite{PointNet}. The effect of beam size will be investigated in the experiment section in detail. \item RCNet is computationally efficient and converges fast during training, due to the benefits of 2D CNN. Besides, unlike vanilla RNN, our recurrent encoder is parallelizable with each RNN processing a small portion of points. This further facilitates the computational efficiency. \end{enumerate} \subsection{RCNet Ensemble} In RCNet, the beam depth extends along a certain direction, i.e., $ z $ axis. While being effective at extracting subregional features in this direction, the recurrent encoder does not explicitly consider features along other directions. To further facilitate the point feature learning, we propose to capture geometric details in different directions and use an ensemble of RCNets, of which each single model has different beam depth directions. The ensemble unifies a set of ``weak'' RCNets and is able to learn richer geometric features. The resulting model, termed RCNet-E, is flexible and achieves better performance, as shown in our experiments. In practice, we implement an ensemble by independently training three RCNets, whose beam depths extend along $ x $, $ y $ and $ z $ axes respectively. Then we simply average their predictions to produce the final results. Note that, although multiple networks are used, thanks to the high efficiency of RCNet, their ensemble is still quite efficient. Moreover, such ensemble is amenable to parallelization for further speed-up. \section{Experiments} \label{sec:exp} In this section, we evaluate our RCNet on multiple benchmark datasets, including ModelNet10/40 \cite{3D_ShapeNets}, ShapeNet part segmentation \cite{ShapePartSeg}, and S3DIS \cite{S3DIS}. In addition, we analyze the properties of RCNet in details with extensive controlled experiments. Code can be found on the authors' homepage. \subsubsection{Ablation Study and a Baseline Model} To validate the advantages of our recurrent set encoder, we compare it with the widely used pooling-based feature aggregator. In particular, we replace the recurrent encoder in RCNet with an MLP, consisting of two layers whose sizes are the same with that of the corresponding RNN hidden layers. This MLP is shared and applied to each point, followed by a global max-pooling to aggregate the subregional features. Meanwhile, the remaining parts of the model are kept the same with RCNet. We take this modified network as a baseline model. As demonstrated in the following section, our recurrent set encoder is more effective at describing the spatial layout and geometric relationships than pooling-based technique. \subsection{Shape Classification} \subsubsection{Datasets} ModelNet10 and ModelNet40 \cite{3D_ShapeNets} are standard benchmarks for shape classification. ModelNet10 is composed of 3991 train and 908 test CAD models from 10 classes, while ModelNet40 consists of 12311 models from 40 categories, with 9843 models used for training and 2468 for testing. These models are originally organized with triangular meshes, and we follow the same protocol of \cite{PointNet,PointNet++} to convert them into point clouds. In particular, for each model, we uniformly sample 1024 points from the mesh, and then normalize them to fit within a unit ball, centered at the origin. We only use the point positions as input features and discard the normal information. \subsubsection{Training} Following \cite{PointNet,PointNet++,KD_net}, we apply data augmentation during the training procedure by randomly translating and scaling the objects, as well as perturbing the point positions. We set the hyper-parameters $ r=32$ and $ s=32 $. The learning rate is initialized to 0.001 with a decay of 0.1 every 30 epochs. The networks are optimized using Adam \cite{Adam}, and it takes about $ 2 \sim 3 $ hours for the training to converge on a single NVIDIA GTX 1080 Ti GPU. \subsubsection{Results} We compare RCNet with several state-of-the-arts: VoxNet \cite{VoxNet}, volumetric CNN \cite{Vol_MVCNN}, O-CNN \cite{O-CNN}, MVCNN \cite{MVCNN}, ECC \cite{ECC}, DeepSets \cite{Deep_Sets}, vanilla RNN and PointNet \cite{PointNet}, PointNet++ \cite{PointNet++}, KD-Net \cite{KD_net}, Pointwise CNN \cite{PointWise}, SO-Net \cite{SO-Net}, KCNet \cite{Kernel-Graph}, SCN \cite{ShapeContextNet}, and PCNN \cite{PCNN_2018}. The results are demonstrated in Table~\ref{table:classification}. We observe that a single RCNet is able to achieve competitive results against the state-of-the-arts, and with ensemble the performance is further boosted. In particular, RCNet performs better than most existing approaches. While obtaining similar accuracy to PCNN, our network is significantly simpler in design. On the other hand, compared to the baseline model, RCNet outperforms it by a large margin. This validates the effectiveness of recurrent encoder at modeling the relative relationships among points. It is worth noting that, in \cite{SO-Net} the SO-Net also attempted to apply the standard CNN to the generated image-like feature maps, but only led to decreased performance. In contrast, our RCNet is better at incorporating the advantages of CNN into point cloud analysis, thanks to the recurrent set encoder. \begin{table}[t!] \begin{center} \resizebox{1.0\columnwidth}{!}{ \begin{tabular}{ccccc} \hline Method & \# Points & Input & MN10 & MN40 \\ \hline VoxNet & - & Vox & 92.0 & 83.0\\ Vol. CNN & - & Vox & - & 89.9\\ O-CNN & - & Vox & - & 90.6\\ MVCNN & - & Img & - & 90.1\\ \hline ECC & 1000 & PC & 90.8 & 87.4 \\ DeepSets & 5000 & PC & - & 90.0 \\ RNN (vanilla) & 1024 & PC & - & 78.5 \\ PointNet & 1024 & PC & - & 89.2\\ PointNet++ & 1024 & PC & - & 90.7\\ KD-Net & 1024 & PC & 93.3 & 90.6\\ Pointwise CNN & - & PC & - & 86.1 \\ SO-Net & 2048 & PC & 94.1 & 90.9\\ KCNet & 1024 & PC & 94.4 & 91.0\\ SCN & 1024 & PC & - & 90.0 \\ PCNN & 1024 & PC & 94.9 & 92.3\\ \hline Baseline (ours) & 1024& PC & 92.5 & 89.1 \\ Baseline-E (ours) & 1024 & PC & 93.0 & 90.8 \\ RCNet (ours) & 1024 & PC & 94.7 & 91.6 \\ RCNet-E (ours) & 1024 & PC & \textbf{95.6} & \textbf{92.3}\\ \hline \end{tabular} } \caption{Classification accuracies on ModelNet datasets. (``Vox": Voxels; ``Img": Images; ``PC": Point Clouds.)} \label{table:classification} \end{center} \end{table} Finally, our RCNet is computationally efficient. In particular, a single RCNet can be trained in about 3 hours. This is much faster than PointNet++ and PCNN, both of which require about 20 hours for training \cite{PointNet++,PCNN_2018}. Besides, as shown in Table~\ref{table:time}, on average it takes about 0.4 milliseconds for RCNet to forward a shape, while PointNet++ and PCNN require 2.8 and 16.8 milliseconds, respectively\footnote{For PCNN, we run the code released by the authors (https://github.com/matanatz/pcnn), with the default pointconv configuration. For PointNet++, we use the official implementation (https://github.com/charlesq34/pointnet2), and test the MSG model with the default network setting.}. Table~\ref{table:time} also summarizes the number of parameters of different networks. Interestingly, although our model has larger size, it still runs faster than other competitors. This validates that the classic RNN and 2D CNN, which are well supported at both software and hardware levels, contribute largely to the model efficiency. In contrast, since PointNet++ need to perform additional K-nearest neighbor query on the fly on GPU, it is much less efficient in spite of the smaller model size. Similarly, PCNN and SPLATNet$ _\text{3D} $ rely on sophisticated geometric transformations and complex convolutional operations. These operations are much less GPU-friendly and cause a lot of overhead in practice. It is worth mentioning that, since RCNet-E is naturally parallelizable, its inference time is almost the same with that of a single RCNet. \begin{table}[t!] \begin{center} \begin{tabular}{c|cc|cc} \hline \multirow{2}{*}{Method} & \multicolumn{2}{c|}{Infer. Time (ms)} & \multicolumn{2}{c}{\# Param. (M)} \\\cline{2-5} & Class. & Seg. & Class. & Seg.\\ \hline RCNet (ours) & \textbf{0.4} & \textbf{4.5} & 13.3 & 16.7 \\ RCNet-E (ours) & 0.6 & 4.8 & 39.9 & 50.1 \\ PointNet++ & 2.8 & 11.9 & \textbf{1.0} & \textbf{1.7} \\ PCNN & 16.8 & 109.3 & 8.1 & 5.4 \\ SPLATNet$ _{\text{3D}} $ & - & 23.1 & - & 2.7\\ \hline \end{tabular} \caption{Comparison of inference time and model size for different networks. Classification and segmentation are performed on ModelNet40 and ShapeNet part datasets, respectively. Time is measured in milliseconds, which correspond to the cost of forwarding a shape on average. The hardware used is an Intel i7-6850K CPU and a single NVIDIA GTX 1080 Ti GPU. ``M" stands for million.} \label{table:time} \end{center} \end{table} \begin{table*}[t!] \centering \begin{center} \resizebox{2.11\columnwidth}{!}{ \begin{tabular}[t]{c|c| *{16}{c}} \hline & mean & aero & bag & cap & car & chair & ear-p & guitar & knife & lamp & laptop & motor & mug & pistol & rocket & skate & table \\ \hline \# shapes & & 2690 & 76 & 55 & 898& 3758 & 69 & 787 & 392 & 1547 & 451 & 202 & 184 & 283 & 66 & 152 & 5271 \\ \hline PointNet & 83.7 & 83.4 & 78.7 & 82.5 & 74.9 & 89.6 & 73.0 & 91.5 & 85.9 & 80.8 & 95.3 & 65.2 & 93.0 & 81.2 & 57.9 & 72.8 & 80.6\\ PointNet++ & 85.1 & 82.4 & 79.0 & 87.7 & 77.3 & 90.8 & 71.8 & 91.0 & 85.9 & 83.7 & 95.3 & 71.6 & 94.1 & 81.3 & 58.7 & \textbf{76.4} & 82.6\\ Kd-Net & 82.3 & 80.1 & 74.6 & 74.3 & 70.3 & 88.6 & 73.5 & 90.2 & 87.2 & 81.0 & 94.9 & 57.4 & 86.7 & 78.1 & 51.8 & 69.9 & 80.3\\ SPLATNet$ _{\text{3D}} $ & 84.6 & 81.9 & 83.9 & 88.6 & 79.5 & 90.1 & 73.5 & 91.3 & 84.7 & 84.5 & 96.3 & 69.7 & 95.0 & 81.7 & 59.2 & 70.4 & 81.3\\ SO-Net (p.t.) & 84.9 & 82.8 & 77.8 & 88.0 & 77.3 & 90.6 & 73.5 & 90.7 & 83.9 & 82.8 & 94.8 & 69.1 & 94.2 & 80.9 & 53.1 & 72.9 & 83.0\\ RSNet & 84.9 & 82.7 & \textbf{86.4} & 84.1 & 78.2 & 90.4 & 69.3 & 91.4 & 87.0 & 83.5 & 95.4 & 66.0 & 92.6 & 81.8 & 56.1 & 75.8 & 82.2\\ KCNet & 84.7 & 82.8 & 81.5 & 86.4 & 77.6 & 90.3 & 76.8 & 91.0 & 87.2 & 84.5 & 95.5 & 69.2 & 94.4 & 81.6 & 60.1 & 75.2 & 81.3 \\ A-SCN & 84.6 & 83.8 & 80.8 & 83.5 & 79.3 & 90.5 & 69.8 & 91.7 & 86.5 & 82.9 & 96.0 & 69.2 & 93.8 & 82.5 & \textbf{62.9} & 74.4 & 80.8\\ PCNN & 85.1 & 82.4 & 80.1 & 85.5 & 79.5 & 90.8 & 73.2 & 91.3 & 86.0 & \textbf{85.0} & 95.7 & 73.2 & 94.8 & 83.3 & 51.0 & 75.0 & 81.8\\ \hline Baseline (ours) & 84.6 & 83.3 & 76.8 & 87.6 & 78.6 & 90.3 & 73.7 & 90.9 & 86.8 & 82.1 & 95.5 & 69.8 & 94.3 & 82.6 & 58.4 & 76.0 & 81.7\\ Baseline-E (ours) & 85.3 & 84.1 & 77.0 & 87.4 & 79.8 & 90.6 & 73.9 & 91.5 & 87.0 & 83.1 & 95.6 & 70.0 & 94.4 & 83.4 & 58.1 & 75.6 & 82.4 \\ RCNet (ours) & 85.3 & 84.4 & 80.1 & 89.6 & 78.6 & 90.5 & 76.3 & 91.4 & \textbf{87.3} & 82.5 & 96.1 & 73.1 & 94.7 & 84.0 & 61.0 & 76.1 & 82.6\\ RCNet-E (ours) & \textbf{86.0} & \textbf{85.3} & 81.1 & \textbf{90.0} & \textbf{79.9} & \textbf{91.1} & \textbf{77.0} & \textbf{91.8} & \textbf{87.3} & 84.1 & \textbf{96.5} & \textbf{75.1} & \textbf{95.1} & \textbf{84.8} & 61.3 & \textbf{76.4} & \textbf{83.1}\\ \hline \end{tabular} } \caption{Results on ShapeNet part segmentation. mIoU metric is used for evaluation. The instance average mIoU as well as mIoU scores for each shape category are listed. Our RCNet-E outperforms the state-of-the-arts in most categories and achieves the best instance average mIoU.} \label{table:partseg} \end{center} \end{table*} \subsection{Shape Part Segmentation} \subsubsection{Dataset and Configuration} For shape part segmentation, the task is to classify each point of a point cloud into one of the predefined part categories. We evaluate the proposed method on the challenging ShapeNet part dataset \cite{ShapePartSeg}, which contains 16881 shapes from 16 categories. The shapes are consistently aligned and normalized to fit within a unit ball. For each shape, it is annotated with 2-6 part labels, and in total there are 50 different parts. We sample 2048 points for each shape following \cite{PointNet,PointNet++}. As in \cite{PointNet++}, apart from point positions we also use normal information as input features. Following the setting in \cite{ShapePartSeg}, we evaluate our methods assuming that the category of the input 3D shape is already known. The segmentation results are reported with the standard metric mIoU \cite{PointNet}. We use the official train/test split as in \cite{shapenet2015} in our experiment. We follow the same network configuration with the classification task. \subsubsection{Results} Table~\ref{table:partseg} compares RCNet with the following state-of-the-art point cloud-based methods: PointNet \cite{PointNet}, PointNet++ \cite{PointNet++}, Kd-Net \cite{KD_net}, SPLATNet$ _{\text{3D}} $ \cite{SPLATNet}, SO-Net (pre-trained) \cite{SO-Net}, RSNet \cite{SliceNet}, KCNet \cite{Kernel-Graph}, A-SCN \cite{ShapeContextNet}, and PCNN \cite{PCNN_2018}. In Table~\ref{table:partseg}, we report the instance average mIoU as well as the mIoU scores for each category. \begin{figure}[t!] \centering \includegraphics[width=1.0\linewidth]{shape_part_seg.pdf} \caption{Visualization of ShapeNet part segmentation results. From top to bottom: ground truth, baseline, baseline-E, RCNet, RCNet-E. From left to right: airplane, motorbike, lamp, table.} \label{fig:shape_part_seg} \end{figure} As is shown, our method achieves better results than the state-of-the-art works. In particular, a single RCNet is able to achieve average mIoU of 85.3, which is competitive with the performance of PointNet++ and PCNN. With ensemble, the accuracy is further boosted and our method dominates most of the shape categories. Some qualitative segmentation results are illustrated in Fig.~\ref{fig:shape_part_seg}. Specifically, the first two columns show that both RCNet and RCNet-E are able to handle the small details of objects well. The third column indicates that the ensemble helps correct the prediction error of a single model, and is better at capturing the fine-grain semantics than the baseline methods. The last column corresponds to a failure case, which is possibly due to the imperfect model representation ability or caused by shape semantic ambiguity (i.e., the table board in the middle could be interpreted as either table support or tabletop). In Table~\ref{table:time}, we compare the computational efficiency of different networks on part segmentation task. As is shown, our method is more efficient than the state-of-the-arts\footnote{For SPLATNet$ _{\text{3D}} $, we run the code implemented by the authors (https://github.com/NVlabs/splatnet), with the default network configuration. For PointNet++, the MSG model with one hot vector is tested. For PCNN, we use the default pointconv configuration. In the experiment we sample 2048 points for each shape.}. \begin{table}[t!] \begin{center} \begin{tabular}{ccc} \hline Method & Mean IoU & Overall accuracy \\ \hline PointNet & 47.71 & 78.62 \\ A-SCN & 52.72 & 81.59 \\ Pointwise CNN & - & 81.50\\ \hline Baseline (ours) & 50.31 & 81.57 \\ Baseline-E (ours) & 52.38 & 82.98 \\ RCNet (ours) & 51.40 & 82.01 \\ RCNet-E (ours) & \textbf{53.21} & \textbf{83.58} \\ \hline \end{tabular} \caption{Segmentation results on S3DIS dataset. Mean IoU and point-wise accuracy are listed.} \label{table:S3DIS} \end{center} \end{table} \subsection{Semantic Scene Segmentation} \subsubsection{Dataset and Configuration} We evaluate our RCNet on the scene parsing task with Standford 3D indoor scene dataset (S3DIS) \cite{S3DIS}. S3DIS consists of 6 scanned large-scale areas, which in total have 271 rooms. Each point in the scene point cloud is annotated with one of the 13 semantic categories. Following \cite{PointNet}, we pre-process the data by splitting the scene points into rooms, and then subdividing the rooms into small blocks with area 1m by 1m (measured on the floor). As in \cite{PointNet}, we also use k-fold strategy for training and testing. At training time we randomly sample 2048 points for each block, but use all the points during testing. We represent each point using 9 attributes, including XYZ coordinates, RGB values and normalized coordinates as to the room. The same shape segmentation RCNet is used for this task. \begin{figure}[t!] \centering \includegraphics[width=1.0\linewidth]{S3DIS.pdf} \caption{Visualization of S3DIS segmentation results. From top to bottom: input scene, ground truth, baseline, baseline-E, RCNet, RCNet-E.} \label{fig:S3DIS} \end{figure} \subsubsection{Results} We compare our RCNet with PointNet \cite{PointNet}, A-SCN \cite{ShapeContextNet} and Pointwise CNN \cite{PointWise}. The results are reported in Table~\ref{table:S3DIS}. As is shown, our RCNet improves A-SCN by about $ 0.5\% $ in mean IoU and $ 2\% $ in overall accuracy. We visualize a few segmentation results in Fig.~\ref{fig:S3DIS}. It can be observed that RCNet is able to output smooth predictions and segment the small objects well. In contrast, the baseline methods tend to produce large prediction errors. This shows the benefits of our recurrent set encoder and the 2D CNN as feature aggregators. With ensemble, the segmentation accuracy is further boosted and our RCNet-E achieves the best results. \subsection{Architecture Analysis} In this section we show the effects of network hyper-parameters and validate the design choices through a series of controlled experiments. We consider the following two main contributory factors on model performance: (1) the size of beams; (2) the number of points. We use ModelNet40 dataset as the test bed for comparisons of different options. Unless explicitly noted, all the experimental settings are the same with those in the shape classification experiment. \subsubsection{The Size of Beams} The beam size controls how much local context information would be utilized, and is a major contributory factor for the network performance. For RCNet, large beams will lead to a small feature map for the downstream CNN. This would increase the efficiency of CNN but in turn result in the loss of fine-scale geometric details. Moreover, beams with large size would be filled with too many points, and as a result the RNN would perform poorly in feature modeling. On the other hand, if the size of beams is too small, the subregions would contain insufficient amount of points, which is adverse to the feature learning. We conduct several experiments to investigate the influence of beam size on the network performance. In particular, we test RCNet with different specifications of hyper-parameters $ r $ and $ s $. The results are reported in Table~\ref{table:beam_size}. As is shown, both larger and smaller beam sizes would hurt the performance, and $ r\times s =32 \times 32 $ leads to the best results. Note that, although beam size is an important parameter on the performance, our RCNet is still quite robust to this factor. In contrast, the max-pooling based encoder behaves quite sensitively and the performance decreases a lot with large beams. This further validates that pooling is a relatively coarse technique for exploiting geometric details. \begin{table}[t!] \begin{center} \resizebox{1.0\columnwidth}{!}{ \begin{tabular}{c|c|c|c|c|c} \hline $ r \times s $ & $ 8 \times 8 $ & $ 16 \times 16 $ & $ 32 \times 32 $ & $ 64 \times 64 $ & $ 128 \times 128 $ \\ \hline Baseline & 77.2 & 86.3 & 89.1 & 89.3 & 86.7 \\ RCNet & 87.5 & 90.2 & 91.6 & 90.9 & 89.8 \\ \hline \end{tabular} } \caption{The influence of beam size on network performance. The smaller the hyper-parameters $ r $ and $ s $, the larger the beams, and vice versa. The experiments are conducted on ModelNet40, and the metric is classification accuracy.} \label{table:beam_size} \end{center} \end{table} \begin{table}[t!] \begin{center} \resizebox{1.0\columnwidth}{!}{ \begin{tabular}{ccccc} \hline \# Point & Baseline + DP & RCNet + DP & Baseline & RCNet \\ \hline 1024 & 88.9 & 91.1 & 88.2 & 90.2\\ 512 & 88.2 & 90.4 & 68.2 & 76.2\\ 256 & 87.7 & 90.2 & 35.3 & 38.1\\ 128 & 86.4 & 87.8 & 17.8 & 24.9\\ \hline \end{tabular} } \caption{Experiments on robustness to non-uniform and sparse data. DP stands for random point dropout during training. The experiments are conducted on ModelNet40.} \label{table:num_points} \end{center} \end{table} \subsubsection{The Number of Points} Point clouds obtained from sensors in real world usually suffer from data corruptions, which lead to non-uniform data with varying densities \cite{PointNet++}. To validate the robustness of our model to such situations, we randomly dropout the number of points in testing and conduct two different groups of experiments. In the first group, the models are trained on uniform point clouds without random point dropout, while in the second group the models are trained with random dropout as well. In the experiment, we set $ r=s=32 $ as in the shape classification task. The results are shown in Table~\ref{table:num_points}. We observe that models trained with random point dropout (DP) during training are fairly robust to the sampling density variation, with drop of accuracy less than 3.3\% when point number decreases from 1024 to 128. In contrast, those trained only on uniform data fail to generalize well to the cases of non-uniform data. Note that, despite the drop of accuracy, our RCNet still achieves better performance than the baseline model when trained without DP. This validates the superiority of RNN in subregional feature extraction compared to max-pooling. \section{Conclusion and Discussion} \label{sec:conclusion} In this work we present a new deep neural network for 3D point cloud processing. Our network consists of a recurrent set encoder and a 2D CNN. The recurrent set encoder partitions the input point clouds into several parts, which are encoded via a shared RNN. The encoded part features are later assembled in a structured manner and fed into a 2D CNN for global feature learning. Such design leads to an efficient as well as effective network, thanks to the benefits of CNN and RNN. Experiments on four representative datasets show that our method competes favorably with the state-of-the-arts in terms of accuracy and efficiency. We also conduct extensive experiments to further analyze the network properties, and show that our method is quite robust to several key factors affecting the model performance. Finally, we note that the proposed recurrent set encoder can be generalized to other contexts. For example, we can build a KNN graph for the input point cloud and model the local neighborhood for each point with recurrent encoder. In particular, we can sort the $ k $ nearest neighbor points according to their distances to the query point, and then apply RNN to this point sequence for local feature learning. This is different from KCNet \cite{Kernel-Graph} which uses a local point-set kernel, and will be explored in the future. \section{Acknowledgments} This work was partially supported by NSF IIS-1718802, CCF-1733866, and CCF-1733843. \bibliographystyle{aaai} \input{AAAI-WuP.1578.bbl} \end{document}
{ "redpajama_set_name": "RedPajamaArXiv" }
5,288
• Eclectic, BC-bred singer-songwriter Lily Frost brought listeners on an "eerie adventure through the woods" with her last folky set, 2010's Viridian Touch, but her next LP, Do What You Love, is poised to deliver a handful of positive life lessons. • The 10-song set was produced by Me&John (Ryan Kondrat and John LaMagna), and is said to navigate between playful, hummable pop, '60s girl-group melodies, country swing and "trip-pop." The record also closes with a cover of Pink Floyd's San Tropez. • On top of Frost's performance, guests appearing on Do What You Love include Jose Contreras (By Divine Right), Bryden Baird (Zeus, Feist), and Broken Social Scene's Brendan Canning. Album track "Stand" also received a string arrangement courtesy of Todor Kobakov (Small Sins, Emily Haines). • A press release for the album explains that Frost's latest was mostly written in a cabin in Georgian Bay, ON, and was "conceived, written and recorded as a letter to the daughter she was about to have." It allegedly offers insight from Frost's experience in the recording industry, among other topics. • Singer/songwriter Lily Frost draws on influences of '60s pop, fashion, and art to create her engaging brand of not-easily pigeonholed music. Painfully shy, the introverted Frost spent years expressing herself in her diary entries, gradually setting them to music and putting years of classical piano training to use. By 1989, she was studying jazz and French at Concordia University in Montreal and was performing '20s covers in a local group called the Sheiks. To escape from the Montreal winters, Frost took a job singing on a ship in Cairo, Egypt, performing standards and torch songs seven nights a week, for six months, before returning to Vancouver in 1992 to front the Colorifics, who embraced '50s-style swing and attracted a following along the West Coast. The group scored at college radio when they released one of Frost's compositions, "Soda Pop," but Frost, wanting to pursue more contemporary sounds, released the solo effort Lunamarium, which was re-packaged and re-released when she signed with Nettwerk. Promoting the record, which was well-received by critics, Frost had the opportunity to share the stage with acts that included Coldplay, Cowboy Junkies, and the Dandy Warhols. She also continued to work concurrently with the band Mimosa. • As Exclaim! wrote in our review, eclectic, BC-bred singer-songwriter Lily Frost brought listeners on an "eerie adventure through the woods" with her last folky set, 2010's Viridian Touch, but her next LP, Do What You Love, is poised to deliver a handful of positive life lessons. • A press release for the album, which arrives October 2 through Aporia Records, explains that Frost's latest was mostly written in a cabin in Georgian Bay, ON, and was "conceived, written and recorded as a letter to the daughter she was about to have." It allegedly offers insight from Frost's experience in the recording industry, among other topics. • "I wanted to make a fun, upbeat record for my daughter that included lessons I had learned the hard way," Frost said in a statement. "I wanted to leave people feeling inspired with smiles on their faces." • The 10-song set was produced by Me&John (Ryan Kondrat and John LaMagna), and is said to navigate between playful, hummable pop, '60s girl-group melodies, country swing and "trip-pop." The record also closes with a cover of Pink Floyd's "San Tropez." You can catch a sneak peek of Do What You Love via the twinkly piano melodies and breathy "ba-ba-ba" vocal harmonies of "Background Radio" in the stream down below. On top of Frost's performance, guests appearing on Do What You Love include Jose Contreras (By Divine Right), Bryden Baird (Zeus, Feist), and Broken Social Scene's Brendan Canning. Album track "Stand" also received a string arrangement courtesy of Todor Kobakov (Small Sins, Emily Haines). • As an added bonus, the version of the album in Quebec will come with an extra disc featuring songs by Frost sung in French. Plus, Frost has a few Ontario shows on the horizon, and you can see those below. • Singer songwriter Lily Frost is a rare bird from Canada's musical woods. Frost has always danced the line between indie and jazz, thus creating a niche of her own. Frost draws from her real life experiences in her lyrics – in performance, there is an intimacy, a gentle passion that can take you somewhere else and evokes a certain nostalgia – the rainy streets of Paris, a winter picnic or perhaps a beat café in San Francisco. This rarity and distinctiveness can be found on her latest offering, Do What You Love. • Do What You Love puts Lily's distinct sound and exceptional talents together in one concise package. From the lyrics to the arrangements to the striking artwork, it packs a punch and delivers insightful truths about life – embracing femininity with strength, joy and creativity. • Largely written in a cabin in Georgian Bay, Do What You Love was recorded at Pirate Studios in downtown Toronto and was produced by Juno award-winners me&john (Ryan Kondrat and John LaMagna), who also play on the record. The record also features Jose Contreras (By Divine Right) and includes cameos by Brydon Baird, and Brendon Canning and a string arrangement on 'Stand' by Todor Kobokov. • Lily's parents were hobby artists who loved music, taught disco dancing and played records after supper for Lily and her younger brother to dance to. Her mom also encouraged her by sending her to the Royal Conservatory for dance and piano lessons all her childhood life. • Summers were spent up north at the family cottage. It was during these summers that Lily bonded most with her cousin, feminist rapper and singer/songwriter Kinnie Starr. This relationship continues to be inspiring, important and competitive. • By the age of 18, Lily drove too fast and drank too much before flying from the nest of her born-again Christian family – heading to Montreal to study jazz. She soon connected with John Davis (The Sheiks) and found herself performing songs from the 1920's with The Shieks – Bessie Smith, Cab Calloway and Big Mama Thorton. From that first performance, Lily was drawn to the stage and has never looked back. Instead of going back for her third year of university, and with a musical fire in her belly, Lily accepted an offer to sing in Cairo for six months. From Egypt, Lily landed in Vancouver to join Les Minstrels, playing keyboards. They would frequent a local institution called The Railway Club where the late Ray Condo was the star of the jam. • Upon hearing Lily sing Aretha Franklin's "Hold On", Ray asked her to be the featured vocalist in his group, The Swinging Dukes. Ray became a mentor and dear friend to Lily, but after a few months he said to her the words that were to change the course of her life: "You gotta get your own thing going, kid." Lily started busking with Dave Lachance and Jorge Diaz, playing upbeat swing and rockabilly. They recruited local guitarist Bernie Boulanger, called themselves The Colorifics, and began to spearhead the "cocktail nation" (a mid 90's anti-grunge movement) alongside The Squirrel Nut Zippers, The Friends of Dean Martinez and The Pink Martinis. • Five years and three albums later, Lily left The Colorifics in order to finally start performing her own original material. Lily began writing with Chad Horton. This dynamic writing duo led to Lily's first album, Cosmicomic Country (1999) and her follow-up album, Lunamarium (2000), which was released on Nettwerk Records. The style was still retro, drawing from surf, rockabilly and jazz but now the lyrics were more honest, poetic and contemporary. Situation (2004), was Lily's next effort – it was an album that began in Vancouver with producer Steven Drake, and was finished in Toronto for Aporia Records. It was there that she met her fiery soul mate and future husband, José Miguel Contreras. Jose and Lily formed a team that was more prolific than any pairing she had before. They worked together to complete Situation, co-wrote and produced Cine-Magique (2006), Flights of Fancy (2007), Lily Swings (2008) and Viridian Torch (2010). In recent years, Lily has won West Coast Vocalist of the Year (2002, New Music West), a SOCAN award for Situation (2006, NXNE) and also a Gemini Award nominee for co-writing the theme song and score for the CBC hit television show, Being Erica.
{ "redpajama_set_name": "RedPajamaC4" }
2,287
Love to Read: ways of getting back into the habit. For many people, reading is a true pleasure, a way of escaping into a different world, for others the very thought of tackling a book can be so daunting, that they are put off by the experience altogether. Such feelings can come from a number of experiencers, from dyslexia, poor eyesight to just not being confident that they will finish a book, which then feeds into despondency. Thankfully your local library has many options of different reading material to help everyone participate in the pleasure of a good book. Graphic novels: The time when graphic novels were just for teenage boys has long gone, these days the range of graphic novels cover just about any subject you can think of, from the traditional Super Hero to books about Brighton history, Maus, to classic tales. The brilliant thing about graphic novels is there accessibility through pictures and words, making this medium the perfect introduction for the reluctant reader. Brighton and Hove libraries also offer a free download service of hundreds of graphic novels and books that can be borrowed in the same way as physical books. stories. This collection of books are brilliant for people who want to access well known books but are put of by the sheer volume of words. This collection is also a great way for people who wish to learn English as a second language. Many of the Essential skills books are accompanied with a CD allowing the reader to improve their reading skills. Large Print Books are exactly that, a range of fiction an dnon fiction books written in large print to help those with poor vision. Audio Books are another way for everyone to access the joy of story telling. For many people, the last time they were read to was as a child. It is only when you choose to give an audio book a go, do you realise just what a pleasure it is to be read too. Audio books are a great for the visually impaired, to have playing in the car on long road trips, to having a great tale read to you before night-time. Many of Brighton Hove libraries also have a collection of specialist books, from large picture books for children, books in brail and reminiscence books; these books have text and images especially developed for those living with dementia. There is a new project called Love To Read, whose main remit is to get people of all ages from all walks of life back into the habit of reading. There are many reasons why people have stopped reading, from busy lives to just plainly falling out of love with reeading. Have a look on the Love to Read website to find ways of getting back into a habit that once rediscovered can bring so much joy. Ray Bradbury's iconic novel, Fahrenheit 451 first came in to being when he was living in a cramped house with his wife and new born baby daughter, he desperately needed a 'room of his own' to write. it was around this time that he was walking through the University of California, he heard the sound of typing in the basement of the library. it was there he discovered a room filled with twelve typewriters that people could rent for ten cents for half an hour. (Did you know that all fourteen of Brighton and Hove's libraries offer use of computers for registered library uses, the first hour is free (2 hours free for people on certain benefits) and £1 per hour there after. The title of Bradbury's book, Fahrenheit 451, relates to the temperature at which paper burns and it is book burning that is at the centre of this tale. Set in the future, where houses are fireproof, firemen are employed to seek out and burn any books that are found. Bradbury looks back on past history of both governments and religious authorities who have exerted their power and created fear over others through burning books. The authorities in Fahrenheit 451 believe that books are harmful as they would make people question their existence and purpose in the world that is controlled through a banal feed of television programmes in which residence are sent scripts so they can interact with the shows. The main protagonist, Montag, a dedicated firman and destroyer of books, meets a young woman called Clarisse McClellan who unlike the other residents in her neighbourhood notices the natural things in life and finds beauty in them, from a dandelion in the grass to the rain on her face. Montag finds these qualities most strange as he has lived a life of not feeling or thinking for himself for so long. It is through his further encounters with Clarisse that he begins to question why he destroys books. His motives are shaken further when he, along with his team of firemen raid a house (neighbours are encouraged to inform the authorities if they suspect their neighbours are hoarding books). This is a direct echo of the Bradbury's concerns he had of Republican U.S. Senator, Joseph McCarthy and his drive to encourage anyone to expose people deemed to be Communists. On his latest mission, Montag is horrified to find that the person accused of hoarding books has not been arrested and taken to the local asylum. As the books rain down on her and are soaked in kerosene, the woman takes out a box of matches. in the mids of everyones panic, Montag steals one of her books and hides it in his jacket then runs as the woman commits suicide, setting fire to her belovered books as well as herself. From here on in Montag's eyes are forced open as he fights everything he had held as true and seeks out others who find passion beyond the mundane in books. I would highly recommend this book for people who like sci-fi/dystopia novels and for those who have never given this genre a go before. The realism of the settings allow the reader to enter the world of Fahrenheit 451 with ease but will leave them wanting to explore many of the other brilliant novels by Bradbury.
{ "redpajama_set_name": "RedPajamaC4" }
8,336
WEF submits comments on EPAs draft 2017 Construction Stormwater General Permit Washington, D.C. — The Water Environment Federation (WEF) recently provided comments on the Draft National Pollutant Discharge Elimination System (NPDES) General Permit for Stormwater Discharges from Construction Activities. The proposed permit is a renewal of the existing five-year permit that became effective Feb. 16, 2012 and expires on Feb. 16, 2017. The U.S. Environmental Protection Agency made the draft permit available for a 45-day public comment period, which concluded on May 26, 2016. The permit included a number of updates and proposals to which EPA specifically requested public feedback. These were comprised of a proposal for the allowance of group Stormwater Pollution Prevention Plans (SWPPPs); requiring authorized non-stormwater discharges (external structure washdown waters) to not contain hazardous substances, (e.g. PCBs); decreasing timelines for final site stabilization from 14 to 7 days; a request for identification of potential best management practices (BMPs) to reduce the effect of construction associated dewatering activities; increased required site inspection frequencies; a proposed requirement for a specified site inspection frequency for sites with snowmelt runoff; and increased public access to SWPPPs. The proposed permit is immediately applicable only to jurisdictions and tribal groups who do not run their own programs and have designated authority granted by EPA. However, it is reasonable that elements of the updated permit will be mimicked in many state run construction stormwater general permit programs. WEF's comments and recommendations focused on how the permit might affect entities — municipal separate storm sewer system (MS4) communities, localities, tribal groups, etc. — where the permit would be implemented. Highlights of WEF's comments include recognition of the need to make SWPPPs available to the public to be consistent with other federal stormwater permits, such as industrial and MS4s (WEF offered suggestions to EPA on how to address the challenges associated with granting greater public access); general concurrence that addressing potential toxics in source washdown waters is a preferable and more cost-effective method of addressing pollutants than waiting until after they have entered waterbodies and downstream conveyance systems; and recommendations regarding inspection frequencies and final site stabilization timelines. Download a PDF of WEF's complete comments to EPA at http://stormwater.wef.org/wp-content/uploads/2016/06/WEF-Comments-to-EPA-on-General-Permit-for-Stormwater-Discharges-from-Construction-Activities-May-26-2016.pdf.
{ "redpajama_set_name": "RedPajamaCommonCrawl" }
1,746
On Sunday, I got an e-mail that was a long time coming, but even though it was no surprise, the words in the subject line smarted more than I thought they would. The domain is question was the first one I ever registered, leadpencil.net. It was 2001. I was 21, just beginning the spring semester of my senior year. I was, at that point, Over College. I was only taking three classes and was focusing on working more hours for my internship at Boston.com and applying for jobs. Over the previous year and a half, I'd become enamored with online journalism, and I was beginning to put all of my eggs in that basket. One step in that direction was purchasing my own domain. Goodbye, Geocities and Freeservers! Hello, shell account and unlimited potential! I remember agonizing over my domain name. My first choice was littlebluelight.net, inspired by two songs: the Jayhawks' "What Led Me To This Town" (which sports the lyrics "Blue lights are shining over my life") and Miracle Legion's "Little Blue Light." I got talked out of it (too much KMart association) and settled for leadpencil, a name I thought was very poetic at the time but I soon grew tired of. Too late, though. I was branded. And the more your domain and domain-associated e-mail address get out in the world, the more daunting it is to disassociate yourself from them. Having my own domain gave me the opportunity to play with HTML and, eventually, CSS. I built several iterations of my website. The first version actually wasn't so bad, design-wise. I had pages for writing clips, my resume and a bio, but also — ill-advisedly, in retrospect — a link to my Diaryland site (a/k/a Angst Town). Eventually, I hosted a blog, which meant diving into the all-too-fun world of Movable Type installations. At one point, when I was at a crossroads between becoming more of a codehead or continuing to focus on writing, I built a page where I solicited milkshake ratings — for the explicit purpose of learning more about HTML forms. In time, I lessened my emphasis on code, but the HTML playground of those years gave me a basis of understanding that has served me well to this day. The design screencapped above debuted in 2004 and languished for five years. I can understand why. In 2004, I got my current job. I got married. Life began getting a whole lot busier and crazier. There were more pressing things on my agenda than endlessly redesigning my website, as I was wont to do the previous three years. The website remained live, of course, with the resume updated as necessary and a couple of tweaks made now and then. And the e-mail address was still going strong, as well. Beginning in 2008, I realized I needed to transition away from leadpencil.net. It took forever and a day for me to transition my e-mail over to Gmail, including updating my e-mail addresses with every online service from my bank to EddieBauer.com. And, of course, my friends. To tell you the truth, the bank and Eddie Bauer were easier to deal with 🙂 Web-wise, I eventually put in a redirect to a Google Pages site I created. Then, I finally bit the bullet and set up shop at georgycohen.com, my new online hub and portfolio. Every few weeks over the past few months, Dotster would send me increasingly anxious (if only in my mind) e-mail reminders about my pending domain expiration. I thought about extending for another year, but I realized that even though a few stragglers might get an error when trying to e-mail my old address, it was time to cut the cord. So why is this difficult? It's just a domain name, for Christ's sake, right? I guess that the expiration of leadpencil.net makes me think about who I was when I first registered it, my intentions at the time, the professional I wanted to become as I sat in my fourth-floor single and sent my resume to anything and everything web. I'm not sure exactly what I expected to get out of all that effort. So, nine years later, who have I become? I may not be working in the same kind of online journalism that I anticipated as an intern at Boston.com, but I am still working in web communications, a field that has evolved to become something that geeky 21-year-old me would marvel at (though perhaps think "been there, done that"). I think overall that she would be pleased with where I ended up. I suppose that, with the evaporation of leadpencil.net into the domain name ether, this completes my rebranding. But even though I'm setting aside my leadpencil identity, I won't soon forget my humble beginnings and how my little slice of Internet pie (or sip of milkshake, if you will) helped make me the web professional I am today. Conan has been lauded by the press and public alike for his classiness in this situation, especially with regard to his goodbye remarks. But this sentence has particularly resonated with many folks. Part of it is because Conan comes off as incredibly sincere and humble. But I think that another reason is because he mentions something we don't talk about often enough: kindness. You may think that "kind" is not a particularly noteworthy adjective. It means "nice," right? I don't think so. In my mind, being nice is more about being polite. Kindness, in my mind, implies a greater degree of sensitivity and thoughtfulness. Nice is something you might be because you're supposed to be, or it will get you something; kindness is more about the other person's feelings, whether it's a close friend, a colleague or a complete stranger. In the end, I believe, it's more important to be kind than to be nice. Why am I seeing this line quoted everywhere? Why is this hitting home? I think because we all recognize the importance of kindness, and also how unique it is to see someone call it out. The concept has become special, which is great, in a sense. But I wish it didn't seem like such a novelty. Hearing this line, I recalled the National song "Baby, We'll Be Fine," where the insecure protagonist desperately tries to navigate everyday life. "All we've got to do is be brave and be kind," he reassures himself. I'm not sure how Matt Berninger's protagonist turns out, but I hope he meant being kind to himself. That's an important side of kindness, too. We are often too hard on ourselves, too critical and too demanding. Let's forgive ourselves a little bit. Let's give ourselves a break. We can't take anyone for granted, much less ourselves. Anyhow, kudos to Conan. People say he's one of the nice guys in show business, and that may be true. And that's great. But if he's one of the kind guys in show business? Amazing things are definitely in store. I have to admit, when all this "lifestream" business started coming down the pike, I didn't know what to make of it. When people like Steve Rubel started singing the praises of lifestreaming via platforms like Posterous, saying it was just the next iteration of our increasingly real-time online lives, I didn't see how it was different than any other kind of blog. But when I finally sat down and looked at Posterous, I was amazed at how it was just the tool I'd been looking for. I had been noodling a personal publishing conundrum for some time: if I have a piece of media, be it a photograph or video or text or audio file, and I want to publish it to multiple channels (say, Twitter, Facebook and Flickr; or maybe Twitter and Flickr, but not Facebook), how can I do that without posting the piece of content four times, in four separate places, reentering the same data multiple times? I wanted to learn how to COPE (create once, publish everywhere), and selectively at that. When I examined Posterous, I saw that the e-mail-based publishing system they use easily allowed you to do that. After I've connected Posterous to my various social media outposts, all I have to do is e-mail post@posterous.com to send to all of them, twitter@posterous.com to just go to Twitter and Posterous, flickr+twitter@posterous.com to just post to Flickr and Twitter (which is huge, since many a fun photo of mine had been Twitpic'd but never made it to my Flickr archive unless I manually unloaded photos from my phone's SD card and uploaded them), or if I don't want to flood my followers with images, I post it to just my Posterous blog. And so on. If I take the two minutes to add these e-mail addresses into my phone's contact list, I can easily, and selectively, publish on the go. And that's what I've been doing with Georgy To Go, which has become my new go-to personal publishing platform. (I've also added it to the sidebar of this blog.) I've only been using it for mobile photos to date, capturing slices of life from my travels and commutes. But I've been really pleased with my Posterous experience so far, not only for the flexibility and control I have over my publishing, but because it's given me a publishing outlet I did not have previously and allowed me to create new types of content. I love having a venue to showcase the weird, funny or poignant things I see everyday — to the point of this blog, finding the extra in the ordinary. Who Lives in Our Content Village? The other day, as I was walking up to the office, I was thinking about the different roles in the content creation and distribution process. For some reason, in that instant, my brain processed them in a somewhat medieval context. You can look at artisans as content creators, crafting beautiful vessels of meaning. Alternately, that role is also filled by the town crier, broadcasting the news to all who are near enough to hear it. What about the farmers, shepherds and the hunters? Both, to me, represent content aggregation. The farmers and shepherds tend and cultivate a community of content creators, while the hunters go right out and claim content to be, um, repurposed. And of course, there are the craftsmen, the blacksmiths and carpenters who create the structures and mechanisms that store and distribute content — the crop silos and water towers of databases, the plumbing of RSS, the homesteads of websites. Don't forget the soldiers, the security in place to protect the village from outside threats. However, no village can survive in isolation; trade with other villages is essential to growth. The list of metaphors, surely, goes on. The most important thing to keep in mind about our content village, however, is that all of these components are necessary. If one is diminished or removed, the entire system falls apart. And what was it that Hillary Clinton said? "It takes a village." To get the most out of content on the web, ain't that the truth. Talking to Guido at the Buzzup I went to during vacation, I lamented that since I live and work north of the river, I probably wouldn't be able to make it to another one. He had a good suggestion: start my own. So, I am! The first official Buzzup North will be held Wed., Feb. 3 from 8-9AM at Oggi's Gourmet in the heart of Harvard Square (location inside the Holyoke Center Arcade, where the Au Bon Pain is; directions). There's no agenda. It's just an opportunity to start your day by hanging and chatting with some fun folks. RSVP for the event and tweet about it with the hashtag #buzzupnorth.
{ "redpajama_set_name": "RedPajamaC4" }
4,794
Ever since Agriculture Minister Christopher Tufton made his famous speech in which he advocated the return to the cultivation of cassava as a response to the global increases in the price of basic food items, and what is being characterised by many as an impending global food crisis, there is hardly a social gathering in which one does not hear some kind of joke or punchline which centres on cassava cultivation and consumption. It may very well be that people see this as a return to a stage of national development which we have long outgrown and to which there is no desire to return, a hasty and ill-conceived idea which lacks appropriate thought and supporting scientific research before being articulated, or it is one of those uncomfortable realities with which we are confronted and for which laughter becomes the way to cope. What is clear, however, is that the rise in prices is a current reality, a food crisis looms, and any meaningful response on the part of our Government and nation must involve a return to local agricultural production for domestic consumption of which the cassava may be a central element or a symbol of our local staple foodstuff. To that extent, the cultivation of cassava may not be a laughing matter. At the end of April, the House of Bishops and Standing Committee of the Church in the Province of the West Indies (Anglican Church) issued a communiqué to its members and to the people of the Caribbean Community, which sought to address the matter of the increases in food prices and that of petroleum products with its domino effect on almost every aspect of the productive life of the nations of our region. While not advocating the cultivation and consumption of cassava, it sought to join with those regional institutions and individuals who are seeking to find some creative strategies for facing the challenges which the current global situation is creating for us. The communiqué sought to underscore the moral and social imperatives which make it necessary for us to address the current situation with a measure of urgency. It highlighted the reality of poverty under which many of our people live and the dehumanising effects of the same, the difficulties many are facing in providing for the most basic needs of their families, and points to the fact that, if this situation is not addressed, it will only sow the seeds of social unrest as we have already seen in neighbouring Haiti, and even some European cities. This not only places a responsibility on the part of the Government to accelerate the programme of poverty alleviation, it also says that those of us who believe that we are okay because we can make ends meet, and that the problem is to be the concern of some other person or agency, may be in for a rude awakening when the social unrest turns violent in a society already overwhelmed by crime and violence. We cannot ignore the fact that some analyses of the current high level of crime and violence posit the notion that poverty and the lack of means for acquiring the basic demands of daily living constitute a major force in the current scenario. So, even as we treat the call to cassava cultivation as a source of comic relief, it may really have in it the making of a tragedy. In seeking to pursue a creative and responsible path, the Province has committed itself to the pursuit of a number of things, most of which will have to be undertaken at the local diocesan and congregational levels. The proposals include an acceleration of some of the interventions in which we are currently involved, including the operations of feeding programmes which provide a hot meal or care packages for the most needy, and in some instances the provision of cash to assist persons with their needs. Sadly, one of the things which frustrate the attempts of the church to respond to those who are most in need is the spirit of entitlement on the part of those who are not the most needy and who often help themselves to the supplies intended for the needy. This is not unlike the response of civil society to any goods and services which the government provides for the poor and needy. These, however, constitute stopgap interventions. f Collaborating with agencies and organisations involved in the alleviation of poverty. While one can get almost any proposal to look good on paper, the real challenge comes with the implementation of the same. Realism has to inform what the church or any other institution attempts to do by way of poverty alleviation and social outreach. In this regard, it must be acknowledged that in many instances there are certain attitudinal blocks which one must overcome if one is to see social interventions being effective. Granted, it may be true that at a point one is not only being confronted by attitudinal blocks on the part of those one is seeking to help but also the reality of the urgency of certain perceived needs, whether misplaced or otherwise. I recall being at the launch of a goat-rearing project which was intended to empower men and women with hybrid livestock which would have the potential to increase their income, and listening to former minister of agriculture, Roger Clarke, exhorting recipients not to be like those for whom baby chicks and feed were provided, but who did not even leave the precincts of the building in which they were provided with the supplies before they sold them. It has been my experience that in attempting a similar project in which sewing machines were provided, along with a course of training in various techniques in sewing, several participants never returned once they got their machine. Attitudinal issues are also present among a generation of youth who say that they will not work for "slave wages", meaning that they will not work for anything that comes close to the minimum wage. At the same time, many of these individuals are the ones who, when asked what they can do readily respond, "Me can do anything, sir." This is usually another way of saying they have no skill and often are challenged where literacy is concerned. Intervention strategies of a long-term nature which seek to empower persons must involve a commitment on their part to pursue some course of disciplined training, as there are few jobs that do not require some measure of competence, and, even as we talk so glibly about the Government providing jobs for people, we must understand that job readiness requires some commitment on the part of the prospective worker. While the proposal from the church includes empowerment of persons, we recognise that this empowerment is not just about skills and competencies but also education in terms of values. Staff members at the Children's Hospital have spoken at times about seeing children who are malnourished wearing expensive bangles and jewellery on their tiny bodies, a clear case of misplaced values and priorities. The same is true of the many parents who are using their small wages to buy fast food for their lunch each day and keeping their children from school when they cannot find the lunch money for them to do the same thing. There are so many ways in which to prepare a nutritious and inexpensive lunch at home and take it to work and to school. The commitment of the church is therefore to promote ways of budgeting and utilising the resources at hand in the most creative ways to satisfy the needs of families. The church is also clear that the desired intervention at this time cannot come only from Government, but must be a collaborative exercise involving various non-governmental agencies, including the church, along with the mobilisation of the general population to address this major social concern. Whatever the cultivation and consumption of cassava mean to us as Jamaicans, it cannot be just a source of comic relief. It comes out of a recognition of the serious situation we are facing as a people, and that our salvation in the current situation will not come from outside our nation but from a homegrown, domestically driven strategy.
{ "redpajama_set_name": "RedPajamaC4" }
1,253
Q: Downloading the certain folder from GitHub with token method I want to download specific folder from GitHub, and not the whole package. But this one folder is also huge, that's why someone recommended me a token method, but I actually don't know what is it and how it works, and I also did not found the helpful information about this method in Internet. So maybe someone can explain me, how can I download this folder with the token-method? A: First, the token can be the PAT (Personal Access Token), in order to authenticate you, avoiding any anonymous access, and allowing a greater number of queries. Second, "Download a single folder or directory from a GitHub repo" refers to one of the most active tools: refined-github/refined-github. That Chrome extension will add a download button on all folders: And you will enter your token in the options of this extension.
{ "redpajama_set_name": "RedPajamaStackExchange" }
1,660
Q: Spring web project runs with Tomcat in Eclipse fine, but gives 404 when it's on standalone ApacheTomcat server I've already deployed one project to standalone server and it works just fine. I took a copy from it to develop another project with the same dependencies. I changed all that needed to be change(think so), and the problem is that project runs just fine in Eclipse with Tomcat, and gives 404 when I try to run it on standalone server. Is there some specific configuration that need to be done before I export war file? Unfortunately the standalone server is our school's server so I can't provide any logs from it, because i connect to it using ftp so there is to server console. Thank you in advance! A: Check if your web app is deployed and registired at tomcat's manager app gui at http://localhost:8080/manager/html for first access you need to configure it go to tomcat/conf/tomcat-users.xml and add the following lines <role rolename="manager-gui"/> <user username="admin" password="admin" roles="manager-gui"/> this way you can access the manager app with user:admin, passw: admin
{ "redpajama_set_name": "RedPajamaStackExchange" }
3,006
Q: Apply a function using 2 columns grouped by a third column I am having a similar problem to the one described in this post, where I want to apply a function to 2 columns grouped by each level of a third column. I have tried both solutions presented in this post, and neither worked correctly. Here are the first 15 lines of my data frame CMdata: Interval NASC Depth_mean Form 1 0 2.976 0 1 0 4.992 0 1 0 7.008 0 1 0 9.024 0 1 0 11.04 0 1 0 12.96 0 1 0 14.976 0 2 0 2.976 0 2 0 4.992 0 2 0 7.008 0 2 0 9.024 0 2 0 11.04 0 2 0 12.96 0 2 0 14.976 0 3 0 2.976 0 For each Interval, I want to take the sum of Form divided by the sum of NASC (Note: only the beginning of the NASC and Form columns are 0s.) To do this I first created a simple function to do the summing and dividing: CMfunc <- function(arg1,arg2){ sum(arg1)/sum(arg2) } And now I want to apply this to each Interval, 1 through 2104. I tried the solutions presented in the above post, like so: dt = data.table(CMdata) CM <- dt[,CMfunc(CMdata$form,CMdata$NASC),by=CMdata$Interval] But the output is just 23.49 for every row in the new column created. I've also tried various looping functions (tapply, etc.) without success. I feel like this should be very easy but I continue to get errors or incorrect output.
{ "redpajama_set_name": "RedPajamaStackExchange" }
563
On the day of its release, Amazing Spider-Man #569 has sold out at Diamond (though copies will be available at retail level)! The second part of the hotly anticipated "New Ways To Die" storyline marks not only the first appearance of Anti-Venom, but also the second straight sell out for Spidey's summer blockbuster, which IGN.Com deemed a, "well-plotted, intelligent superhero story." Who is Anti-Venom? Why have the Thunderbolts assaulted Peter Parker? And just what does Norman Osborn know about the connection between Peter & Spider-Man? Find out in Amazing Spider-Man #569, in stores now!
{ "redpajama_set_name": "RedPajamaC4" }
789
{"url":"https:\/\/www.khanacademy.org\/science\/in-in-class11th-physics\/in-in-11th-physics-units-and-measurement\/in-in-11th-physics-physical-quantities-and-their-measurement\/e\/angular-measure-2","text":"# Angular Measure 2\n\n### Problem\n\nLet us say that at their closest approach, Mars is about 54.6 million kilometers from the Earth. Mars has a diameter of about 6,779 km. On the same evening let's say our Moon is about 384,400 km away and has diameter of about 3,478 km\nIf we suddenly swapped the position of the Moon and Mars, roughly how many times larger would Mars appear relative to the Moon? (round to one decimal place)\nHere is a picture before the swap occurs:\nDo 4 problems","date":"2017-10-20 09:09:34","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 5, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.3460620045661926, \"perplexity\": 966.8795755622314}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2017-43\/segments\/1508187823997.21\/warc\/CC-MAIN-20171020082720-20171020102720-00393.warc.gz\"}"}
null
null
but I noticed loss of quality in the resulting PDF. Is there some way that I can retain the quality of the PDF i. e. I want the same quality in PDF as in the image. I am not sure if this convert command is using the imagemagick library. Is there any better solution? I have 100s of PDFs, so I have to do it via command line. When looking at the images before converting them into a pdf, you cannot anticipate this error, but LibreOfficeWriter makes it visible and also very easy to correct: just grab the margin of the image, put it into the appropriate form into the page, and it will appear as such in the final pdf. 2) And, when exporting ('File'- 'Export as PDF...') you have several options that let you control the final quality/size of the pdf, which could be very useful if you need to send it by mail and still keep some image quality. I dont know if this is definitely the answer to your question as i am not used to the convert program. PDF is not an image file format, its is a way of formatting documents so that they can be represented consistently not matter what hardware/software you are viewing it from. Therefore when you "convert" the image from JPEG to PDF the JPEG is not actually being converted. The JPEG is embedded in the PDF file itself and the PDF file tells your document viewer how to display the contents on the screen/page. I imagine the loss of quality you are seeing is because of the program with which you are viewing the files. When you view the jpeg with GNOME image viewer the image is displayed exactly as it should be, say and 800x600px image is layout out perfectly on 800x600 pixels on the screen. But when you open your PDF in Document Viewer the viewer lays the 800x600px image out over maybe 25x15cm (or whatever) of the document which when viewed at 100% may not be exactly 800x600 pixels of screen space, it may be more therefore stretching the image and reducing the quality. Based on your comment I tried the links and I see what the problem is. You issue is with the heavy compression that Issuu applies to the pages, and with the Flash interface used to display the page as a lossy image. Your conversion of images to PDF is just fine, the quality loss is occurring when Issuu converts your PDF to their proprietary Flash format and serves it up on the site. Not the answer you're looking for? Browse other questions tagged pdf multimedia convert or ask your own question. How to batch convert an image to a PDF? Why my computer fails when converting many images in Imagemagick with one command? Simple Scan - Changing the default output File Type to PDF and improving Compression? How to programmatically determine DPI of images in PDF file? Why are the images produced by pdfimages different when using the -all flag?
{ "redpajama_set_name": "RedPajamaC4" }
1,684
\section{Introduction} A number of remarkable connections have been observed between gauge theories and integrable systems. They appear to be useful to increase our understanding of both subjects. On the one hand, using powerful integrability techniques one may hope to solve certain gauge theories non-perturbatively. On the other hand, gauge theory can help to formulate and solve integrable system. A spectacular example is given by planar ${\cal N}=4$ super Yang-Mills theory, for which in the last ten years or so tremendous progress has been achieved in solving the theory based on the underlying integrability and on the AdS/CFT correspondence, see \cite{Beisert:2010jr} for a recent review. There exists another amazing connection between gauge theories and integrable models. The gauge theories in this case are ${\cal N}=2$ supersymmetric gauge theories. They do not need to be planar but the connection with integrable models is restricted to a special class of supersymmetric observables. In this paper we focus on an important object in this class, the so-called instanton partition function and its relation with quantum integrable systems. In the groundbreaking work of Seiberg and Witten \cite{Seiberg:1994rs, Seiberg:1994aj}, the exact solution for the low energy effective action of certain ${\cal N}=2$ gauge theories was proposed based on holomorphicity properties and electromagnetic duality. The low energy dynamics are encoded in a single object, called prepotential $\mathcal{F}(\vec{a})$. It is a holomorphic function on the Coulomb moduli space, with coordinates $\vec{a}$, and can be reconstructed from the so-called Seiberg-Witten (SW) curve and SW differential. Shortly after, it was realized that this description provides a direct connection between $\mathcal{N}=2$ gauge theories and classical algebraically integrable systems \cite{Gorsky:1995zq, Donagi:1995cf, {Itoyama:1995uj}}, see e.g.~\cite{D'Hoker:1999ft} for a pedagogical introduction. These are essentially a complex analogue of integrable systems in the sense of Liouville. The challenging program of obtaining the Seiberg-Witten prepotential by a direct gauge theory calculation, developed on \cite{Moore:1997dj,Losev:1997bz, Moore:1998et}, was finalized in \cite{Nekrasov:2002qd}. This problem was solved using powerful localization techniques. Interestingly, this calculation produced a two-parameter, called $\epsilon_1$ and $\epsilon_2$, deformation of the prepotential. The SW prepotential can be obtained as \begin{equation} {\cal F}(\vec{a}; q) \,=\, \lim_{\epsilon_1, \epsilon_2 \rightarrow0} \epsilon_1 \epsilon_2 \log {\cal Z}(\vec a, \epsilon_1, \epsilon_2; q) \,, \end{equation} where we add the explicit dependence on the coupling constant $q$, but suppress dependence on further parameters such as masses. The partition function $\mathcal{Z}$ receives tree level, one loop and instanton contributions. The latter part is usually referred to as Nekrasov instanton partition function. The parameters $\epsilon_1,\epsilon_2$ correspond to a Lorentz rotation $\Omega$, thus the name $\Omega$-deformation, that encodes certain twisted boundary conditions for the four dimensional gauge theory. It was first introduced in \cite{Moore:1997dj,Losev:1997bz} in order to regularize the volume of the instanton moduli space. The $\Omega$-deformation can be understood in a simple way by considering the five dimensional lift of the $\mathcal{N}=2$ theory, further compactified on a circle \cite{Nekrasov:1996cz}. In this set-up it is interpreted as twisting of $\mathbb{R}^4$ by a Lorentz rotation while going around the circle. Introducing a two-parameter generalization of the prepotential triggered a huge progress. An important example is the connection with topological strings. Upon taking $\epsilon_1 = - \epsilon_2 = g_s$, the gauge theory partition function reproduces the topological string partition function with $g_s$ as genus parameter \cite{Iqbal:2003ix,Iqbal:2003zz}. The question of what is the topological string theory analog of the $\epsilon_{1} \neq \epsilon_2$ case leads to the definition of so-called ``refined" topological strings \cite{Hollowood:2003cv, Iqbal:2007ii}. Another spectacular example of progress driven by the calculation of the Nekrasov partition function is given by the Alday-Gaiotto-Tachikawa (AGT) correspondence \cite{Alday:2009aq}. More recently, Nekrasov and Shatashvili \cite{Nekrasov:2009rc} proposed that upon taking the limit $\epsilon_2\rightarrow 0$ and interpreting $\epsilon_1$ as Plank constant, one obtains a correspondence between supersymmetric vacua of a given gauge theory and eigenstates of the corresponding \emph{quantum} integrable model. The relation between the SW prepotential and classical integrable systems is thus quantized. This is usually called Nekrasov-Shatashvili (NS) limit and will be the main focus of this paper. The central role in this correspondence is played by the so-called twisted superpotential ${\cal W}(\vec a, \epsilon_1;q)$ defined as \begin{equation} {\cal W}(\vec a, \epsilon_1; q) \,=\, \lim_{\epsilon_2 \rightarrow0} \epsilon_2 \log {\cal Z}(\vec a, \epsilon_1, \epsilon_2; q) \,. \label{defWINTRO} \end{equation} The Nekrasov-Shatashvili proposal is that, once this function is known, the eigenstates of the quantum integrable system are classified by solutions of the following quantization condition \begin{equation} \exp \Big( {\partial {\cal W}(\vec a) \over \partial a_\lambda } \Big)= 1 \,, \qquad \qquad\lambda\,=\,1,\,\dots\,r\,, \end{equation} where $r$ is the rank of the gauge group. These equations identify the twisted superpotential $\mathcal{W}$ with the so-called Yang-Yang (YY) function \cite{YangYang} of the quantum system. The proposed correspondence provides an efficient general mechanism to define and solve quantum integral models. Remarkably, it can be argued that the instanton part of the prepotential $\mathcal{W}_{\text{inst}}$, defined via \eqref{defWINTRO}, can be characterized as the solution of certain non-linear integral equation of the Thermodynamic Bethe Ansatz (TBA) type \cite{Zamolodchikov:1989cf}. The main goal of this paper is to develope some of the ideas presented in \cite{Nekrasov:2009rc} to give a more explicit derivation, as well as some generalizations, of such TBA equations. The proposal above originates as some sort of extension of the so-called Bethe/gauge correspondence \cite{Nekrasov:2009uh,Nekrasov:2009ui}. The latter is based on the observation that the vacuum equations of two dimensional $\mathcal{N}=4$ gauge theories, broken to $\mathcal{N}=2$ by twisted masses, coincides with Bethe equations for integrable models. The two dimensional twisted superpotential is equal to the YY function that encodes the Bethe equations and Coulomb parameters correspond to the Bethe roots. The generators of chiral ring of the gauge theory \cite{Cecotti:1991me} are mapped to Hamiltonians of the integrable model, while their expectation values mapped to the corresponding eigenvalues. In this way one obtaines a large class of integrable models whose spectrum is characterized by traditional, possibly nested, Bethe equations. Many integrable models do not belong to this class. The simplest example is given by the quantum Toda chain, see e.g.~\cite{Kozlowski:2010tv}, whose classical limit is connected to four dimensional pure $SU(N)$ SYM. From the insight of the Bethe/gauge correspondence it is then natural to propose \cite{Nekrasov:2009rc} the two dimensional twisted superpotential which corresponds to such integrable models. It is the effective low energy action for the four dimensional $\mathcal{N}=2$ gauge theory subject to an $\Omega$-background that preserves two dimensional $\mathcal{N}=2$ super-Poincar\'{e} symmetry, namely $\epsilon_2=0$. This observation provides a stong motivation for the above correspondence. An interpretation of this correspondence was given using brane constructions in \cite{Nekrasov:2010ka}. A further essential step in understanding the nature of the relation between quantum integrable systems and gauge theories has been presented in \cite{Nekrasov:2011bc}. The proposal of Nekrasov and Shatashvili has inspired many other studies. Let us mention a few. In \cite{Mironov:2009uv, Mironov:2009dv} it was shown that, similarly to the prepotential $\mathcal{F}$, the twisted superpotential ${\cal W}(\epsilon_1)$ can be obtained by calculating period integrals of a suitably deformed SW differential. This analysis is also inspired by the AGT correspondence, by which the NS limit corresponds to the semiclassical limit of Liouville CFT \cite{Piatek:2011tp, Mironov:2012uh}. In our work the Coulomb parameters will be assumed to be in generic positions. Extra considerations are needed if they take special values. For example in the conformal $SU(N)$ SYM with $N_f=2 N$, if the Coulomb parameters $a$ are set to be equal to $m_{\text{fund}}-n\,\epsilon_1$, where $n\,\in\,\mathbb{Z}^{N}$, one can quantize the corresponding integrable system to obtain a lenght $N$ spin-chain with infinite dimensional heighest weigth representations of $\mathfrak{sl}_2$ at each site \cite{Dorey:2011pa, Chen:2011sj, Chen:2012we}. Its spectrum is described in terms of traditional Bethe Ansatz equations. Such developments triggered the discovery of a number of new dualities between various integrable models \cite{Bao:2011rc, Mironov:2012ba, Bulycheva:2012ct, Chen:2013jtk, Luo:2013nxa}. The NS proposal has also inspired various studies in (refined topological) string theories \cite{Aganagic:2011mi, Huang:2011qx, Hellerman:2012zf, Huang:2012kn, Antoniadis:2013bja, Hatsuda:2013oxa, Kallen:2013qla,Bao:2013pwa} where the the general $\Omega$-background plays a crucial role. Despite the importance of this correspondence, the precise mechanism by which the instanton part of the twisted superpotential defined in \eqref{defWINTRO} turns out to be characterized as the solution of TBA equations is still to be elucidated. In this paper, we will fill this gap. In order to fully prove the NS's proposal, at least in some example, one should be able to show that the same TBA equation characterizes the spectrum of the corresponding integrable model. In the case of pure $SU(N)$ SYM, which corresponds to the periodic Toda chain with $N$ sites, this was achieved in \cite{Kozlowski:2010tv}. This interesting problem will be studied elsewhere \cite{US}. In the following we briefly outline the main ingredients used in our derivation of the TBA equations for \eqref{defWINTRO}, as well as the structure of the paper. As pointed out in \cite{Nekrasov:2009rc}, it is convenient to start with the contour integral form of the instanton partition function. In this representation the instanton partition function can be interpreted as the partition function of a non-ideal gas of particles. The particular structure of the two-particle interaction potential makes the study of the $\epsilon_2\rightarrow 0$ limit rather subtle. More precisely, this potential is the sum of a short-range (of order $\epsilon_2$) strongly attractive piece and a long-range interaction part. In Sections \ref{sec:simple} and \ref{originDILOG}, we consider the simplified situation in which either the long- or short-range part is set to zero. In order to study these simplified partition functions in the $\epsilon_2\rightarrow 0$ limit, we combine a number of techniques like Mayer expansion \cite{Mayer} (a standard method in statistical mechanics) and the method of expansion by regions \cite{Beneke:1997zp} (a powerful method to compute Feynman integrals in small parameter expansions). For the case with only long-range interactions, the $\epsilon_2\rightarrow 0 $ limit turns the logarithm of the partition function into a sum over certain tree graphs. On the other hand, the free energy corresponding to only short-range interactions gives rise to the dilogarithm function Li$_2$, which can be shown either by direct residue calculation of relevant integrals or via Mayer expansion together with the method of expansion by regions. In order to study the full partition function we find it convenient to use an \emph{iterated} version of Mayer expansion, see \cite{Gopfert:1981zu}. This expansion effectively creates a new partition function whose ``new particles" are \emph{clusters} of the original particles. The interaction within each cluster is governed by the short-range interaction, the one between different clusters by the long-range part. The iterated Mayer expansion thus produces an expression for the twisted superpotential $\mathcal{W}_{\text{inst}}$ as a sum over tree graphs, with vertices given by {\it clusters}. This expansion is carried over in some details in Section \ref{sec:fullexpansion}. The expression can be compared to high order in the instanton number with the expression coming from the solution to the TBA equation as discussed in Section \ref{perturbative_TBA}, providing direct non-trivial check of the equality. There is an elegant way to prove that this equality holds to all orders in the instanton counting parameter $q$. It is based on rewriting the grand canonical partition function of the non-ideal gas in terms of a $(0+1)$-dimensional path integral. The analysis needs some special care as the potential has an unusual feature of depending in a singular way on $\hbar$, which is identified with $\epsilon_2$ in this case. A slight modification of the argument in \cite{Polyakov}, together with the calculation of the contribution from the short range interactions corresponding to the dilogarithm, shows that the instanton partition function in the $\epsilon_2\rightarrow 0 $ limit is obtained by the saddle point evaluation of the path integral. The saddle point equations are nothing but the TBA equations. This is explained in details in Section \ref{sec:nonpert}. In Section \ref{sec:quiver} we present a generalization of the TBA equations corresponding to quiver gauge theories. More precisely, we consider quivers characterized by a Dynkin diagram of $ADE$, or $\widehat{A}\widehat{D}\widehat{E}$ type. The twisted superpotential for such theories is shown to satisfy a set of coupled TBA equations with one equation for each node of the quiver and couplings corresponding to edges in the quiver. The derivation is a simple extension of the one for the single gauge group case. This is so as the short range interaction, responsible for the clustering of particle, is non-vanishing only for particles corresponding to the same gauge group factor in the quiver. In order not to overload the main text, in Appendices we include some review material together with a few technical points concerning the derivation. A review of the contour integral form of the instanton partition function is given in Appendix \ref{app:contour_integral}. Some useful formulas are collected in Appendix \ref{App:UsefulFormula}. A discussion of the method of expansion by regions is given in Appendix \ref{app:region}. In Appendix \ref{App:regionsplusMayer}, we present an alternative derivation of a tree graphs expansion of the instanton partition function. The full partition function is a product of three terms $\mathcal{Z}=\mathcal{Z}_{\text{tree}}\,\mathcal{Z}_{\text{1-loop}}\,\mathcal{Z}_{\text{inst}}$. In this paper we will be only concerned with the study of the instanton part ${\cal Z}_{\rm inst}$. For this reason from now on it will be simply denoted by ${\cal Z}$. \subsection{NS's correspondence} \label{sec:NS} To complete the introduction, we present the integral representation of the instanton partition function for pure $SU(N)$ $\mathcal{N}=2$ super-Yang-Mills and the corresponding TBA equation. The derivation of the TBA starting from the gauge theory expression of the instanton partition function is the main goal of the paper. \subsubsection*{Instanton partition function} The instanton partition function can be written in a contour integral representation as \begin{equation} \label{Zinst} {\cal Z} = \sum_{k=0}^\infty \left({\epsilon \over \epsilon_1 \epsilon_2}\right)^k {q^k \over k!} \int \prod_{I=1}^k {d \phi_I \over 2\pi i} Q(\phi_I) \prod_{ J>I }^k {\cal D}(\phi_{IJ}) \, , \end{equation} where $\phi_{IJ} = \phi_I - \phi_J$, $\epsilon = \epsilon_1 + \epsilon_2$, and \begin{equation} {\cal D}(x) = {x^2 \over (x^2 - \epsilon_2^2)}{ (x^2 - \epsilon^2) \over (x^2 - \epsilon_1^2)} \ , \label{defD} \end{equation} \begin{equation} \label{QP} Q(x) = { 1 \over P(x) P(x+\epsilon) } \ , \qquad P(x) = \prod_{\lambda=1}^N (x-a_\lambda) \ . \end{equation} The origin of this expression is reviewed in Appendix \ref{app:contour_integral}. The parameters entering this integrals, namely $\epsilon_{1,2}$ and $a_{\lambda}$ are taken to be real with a small positive imaginary part $i\, 0$. The domain of integration above should be understood either as a real slice integration or equivalently, upon closing the integration in the upper-half plane, as a multiple contour integral. In Appendix \ref{App:contourRES} we review how the residue evaluation of \eqref{Zinst} reproduces the representation of the instanton partition function as sum over $N$-tuples of Young diagrams. We emphasize that the precise form of $Q(x)$ and ${\cal D}(x)$ does not affect the derivation presented in this paper. This is the main reason why the generalization to quiver gauge theories is rather straightforward. On the other hand the presence of the factor ${x^2 \over x^2 - \epsilon_2^2}$ in ${\cal D}(x)$, which have a particularly singular limit for $\epsilon_2$ small, will play a crucial role and will be responsible for the appearance of the dilogarithm function in the TBA. \subsubsection*{TBA form} The claim of \cite{Nekrasov:2009rc} is that the twisted superpotential, defined as \begin{equation} {\cal W} = {\rm Limit}_{\epsilon_2 \rightarrow 0} \, ( \epsilon_2 \log {\cal Z}) \,, \end{equation} can be written as the critical value of the following integral functional: \begin{equation} {\cal Y}[\rho, \varphi] = {1\over2} \int {d x \over 2\pi i}\,{d y \over 2\pi i} \, \rho(x) G(x-y) \rho(y) + \int {dx \over 2\pi i} \Big[ \rho(x) \varphi(x) + {\rm Li}_2\left( q Q(x) e^{-\varphi(x)}\right) \Big] \ , \label{YYdef} \end{equation} which is \begin{equation} {\cal W} = {\rm Crit}_{\rho, \varphi} \big[ {\cal Y}(\rho, \varphi) \big] = \int {dx \over 2\pi i} \Big[ -{1\over2} \varphi(x) \log\left(1 - q Q(x) e^{-\varphi(x)} \right) + {\rm Li}_2\left( q Q(x) e^{-\varphi(x)}\right) \Big] \, , \label{tbaW} \end{equation} where $\varphi(x)$ is the solution of the following TBA-like equation \begin{equation} \varphi(x) = \int {d y \over 2\pi i} \, G(x-y) \log\left( 1-qQ(y) e^{-\varphi(y)} \right) . \label{tbaphi} \end{equation} $Q(x)$ is defined in (\ref{QP}), and the propagator $G(x)$ is related to ${\cal D}(x)$ as \begin{equation} \label{defG} G(x) = {\rm Limit}_{\epsilon_2 \rightarrow0} {{\cal D}(x) - 1 \over \epsilon_2} = {d \over dx} \log \left( {x+\epsilon_1 \over x- \epsilon_1} \right) \, . \end{equation} \section{Mayer-Cluster expansion} \label{sec:Mayer-Cluster} As mentioned in \cite{Nekrasov:2009rc}, the contour integral form of the instanton partition function (\ref{Zinst}) can be interpreted (for each $k$) as the partition function of a one dimensional non-ideal gas of particles $\phi_1, \ldots, \phi_k$, subject to an external potential and a pair-wise interaction potential respectively given by \begin{equation} \label{potentialU} U^{\rm ext}(x) = - \log\left( Q(x) {\epsilon \over \epsilon_1 \epsilon_2} \right) , \qquad \qquad V^{\rm int}(x) = - \log \left( {\cal D}(x) \right) \, . \end{equation} Upon summing over the number of particles $k$, the instanton partition function takes the form of a grand canonical partition function. The free energy of this gas can be studied by Mayer expansion techniques \cite{Mayer} (see for example \cite{Brydges} for a nice introduction), as pointed out in \cite{Nekrasov:2009rc}. In this section, we will perform such kind of expansion in full details. The limit of $\epsilon_2 \rightarrow 0$ appears to be rather subtle. In order to perform this limit we need to face the problem of studying the leading behavior of some multiple integral where a parameter is small. It turns out that this can be conveniently studied by employing the method of expansion by regions \cite{Beneke:1997zp} discussed in Appendix \ref{app:region}. We introduce this method to provide a unified framework to study certain integrals, but we stress that all the result of this section are also obtained without employing such technique. In order to analyze the behavior of the partition function (\ref{Zinst}) in the limit in which $\epsilon_2$ is small, it is convenient to split the function $ {\cal D}(x)$, see \eqref{defD}, into two parts: \begin{equation} {\cal D}(x)\,=\,\frac{x^2}{x^2-\epsilon_2^2}\, \widetilde{{\cal D}}(x)\,,\qquad \qquad \widetilde{{\cal D}}(x)\,=\,\frac{x^{2}-\epsilon^2}{x^{2}-\epsilon_1^2}\,. \label{splittingPOT} \end{equation} The reason of such decomposition is as follows. The factor $\frac{x^2}{x^2-\epsilon_2^2}$ corresponds to a pair-wise interaction which is strong and attractive at distances of order $\epsilon_2$ and rapidly decreases at large distances. The remaining factor $\widetilde{{\cal D}}(x)$ corresponds to a pair-wise interaction which is different from zero only at distances of order $\epsilon_1$. Thus \eqref{splittingPOT} corresponds to splitting the potential into short- and long-range parts. The natural candidate to study such kind of potentials is the so-called {\it iterated} Mayer expansion \cite{Gopfert:1981zu}. As we will see in Section \ref{sec:fullexpansion}, this effectively creates a new grand canonical partition function whose ``new particles'' correspond to clusters of the original particles. We will start by considering some simplified situation. \subsection{Only long range interactions} \label{sec:simple} Let us first consider a simplified version of (\ref{Zinst}) without the factors\footnote{We also set $\frac{\epsilon}{\epsilon_1}$ to one.} ${\phi_{IJ}^2 \over \phi_{IJ}^2 - \epsilon_2^2}$, i.e \begin{equation} {\cal Z}_{\rm Long} \,:=\, \sum_{k=0}^\infty\, {q^k \over k!}\, \int \prod_{I=1}^k {d \phi_I \over 2\pi i}\,{Q(\phi_I) \over \epsilon_2} \, \prod_{1\,\leq\,I < J\leq\,k}\,\widetilde{{\cal D}}(\phi_{IJ}) \,. \label{Zsimple} \end{equation} The basic idea of Mayer expansion, see e.g.~\cite{Brydges}, is to introduce the function $\widetilde{f}_{IJ}$ as \begin{equation} \label{FG} \widetilde{{\cal D}}(\phi_{IJ}) \,= \,1 + \widetilde{f}_{IJ}\,, \end{equation} and expand the interaction products as \begin{equation} \prod_{1\,\leq\,I < J\leq\,k}\left(1+\widetilde{f}_{IJ}\right)\,=\,1\,+\,\sum_{I<J}\, \widetilde{f}_{IJ}\,+\,\sum_{I<J,I'<J'}\,\widetilde{f}_{IJ}\widetilde{f}_{I'J'}\,+\dots\,. \label{expansionD} \end{equation} Each monomial in the right hand side of this equation can be visualized as a graph (not necessarily connected) on the set $[k]=\{1,2,\dots,k\}$. More precisely, each particle in $\{1,2,\dots,k\}$ corresponds to a vertex, and to each factor $\widetilde{f}_{IJ}$ we associate an edge between particle $I$ and $J$ in the corresponding graph. More explicitly, \begin{equation} \widetilde f_{IJ} \ \ := \begin{tabular}{c}{\includegraphics[height=.8cm]{fIJ}}\end{tabular} \end{equation} As an example for the $k=3$ case, we have the expansion in terms of graphs shown in Figure \ref{fIJ_k=3}. It is clear from the left hand side of \eqref{expansionD} that there are no multiple edges between two vertices, or edges connecting one vertex to itself. \begin{center} \begin{figure}[t] \centerline{\includegraphics[height=2.6cm]{fIJ_k=3} } \caption{\it The graph expansion for the $k=3$ case. The second line contains all connected graphs.} \label{fIJ_k=3} \end{figure} \end{center} This expansion is particularly useful, as the logarithm of the grand canonical partition function can be formally given as a sum over {\it connected} graphs \cite{Mayer} (see for example Appendix A of \cite{Brydges} for a simple derivation): \begin{equation} \label{simple} \log {\cal Z}_{\text{Long}}\,=\, \sum_{k=1}^\infty \,{q^k \over k!}\, \int\, \prod_{I=1}^k\, {d\phi_I \over 2\pi i} {Q(\phi_I) \over \epsilon_2}\, \sum_{g\,\in \,{\cal G}_c^{[k]}}\, \prod _{e(I,J)\,\in\, g} \widetilde{f}_{IJ} \ . \end{equation} Here ${\cal G}_c^{[k]}$ denotes the collection of connected graphs on the set $[k]$ and $e(I,J)$ belongs to the graph $g$ if there is an edge between the vertices $I$ and $J$. Graphs up to four points are shown in Figure \ref{connect_graph}. \begin{center} \begin{figure}[h] \centerline{\includegraphics[height=4.cm]{connect_graph} } \caption{\it Connected graphs in the Mayer expansion. The coefficients correspond to how many different graphs of the given topology belong to ${\cal G}_c^{[k]}$. They are obtained from one another by relabeling the vertices.} \label{connect_graph} \end{figure} \end{center} We stress that \eqref{simple} is an exact relation as a formal power series\footnote{The interesting question of convergence of the Mayer expansion can be addressed in various ways, see e.g. \cite{Brydges}. Here, we will not consider this problem or the analog convergence issue for the solution to the TBA.} in $q$. In the limit of small $\epsilon_2$ one has \begin{equation} \widetilde{f}_{IJ}\, =\, \epsilon_2 \,G(\phi_{IJ}) + {\cal O}(\epsilon_2^2) \, , \label{FG2}\end{equation} where $G$ is defined in (\ref{defG}). As each factor of $\widetilde{f}$ contributes one power of $\epsilon_2$, the leading contribution to the sum in \eqref{simple} is given only by connected {\it tree} graphs $\mathcal{T}^{[k]}_c$. Such graphs have the minimal number of edges ($k-1$) among the connected graphs. The sum of these tree graphs can be shown to be convergent. Collecting the powers of $\epsilon_2$ we conclude that \begin{equation} \label{simpleLIMIT} \log {\cal Z}_{\rm Long}\,=\,\frac{1}{\epsilon_2} \,\sum_{k=1}^\infty \,{q^k \over k!}\, \int\, \prod_{I=1}^k\, {d\phi_I \over 2\pi i} Q(\phi_I)\, \sum_{g\,\in \,{\cal T}_c^{[k]}}\, \prod _{e(I,J)\,\in\, g} \widetilde{f}_{IJ} \,+\, {\cal O}(\epsilon_2^0) \,. \end{equation} Graphs up to five points are shown in Figure \ref{connect_tree_graph}. Notice that the leading behaviour of the free energy \eqref{simpleLIMIT} is at order $\frac{1}{\epsilon_2}$. This fact may be not obvious from the definition \eqref{Zsimple} where for each $k$ the leading contribution is proportional to $\frac{1}{\epsilon_2^k}$. We will see that the similar behaviour applies to the more complicated situations analized in the following. \begin{center} \begin{figure}[h] \centerline{\includegraphics[height=2.7cm]{connect_tree_graph} } \caption{\it Connected tree graphs up to five-point. The coefficients corresponds to how many different graphs of the given topology belong to ${\cal T}_c^{[k]}$, which are related to each other by relabeling the vertices.} \label{connect_tree_graph} \end{figure} \end{center} \subsection{Only short range interactions \label{sec:Li2}} \label{originDILOG} Next we study another simplified version of (\ref{Zinst}). Namely, we set $\widetilde{\mathcal{D}}$, defined in \eqref{splittingPOT}, to zero and consider \begin{equation} {\cal Z}_{\text{Short}}\, :=\, \sum_{k=0}^\infty \,{q^k \over k!} \,\int \, \prod_{I=1}^k {d \phi_I \over 2\pi i}\,{ Q(\phi_I) \over \epsilon_2} \, \prod_{1\,\leq\,I < J\,\leq k} {\phi_{IJ}^2 \over \phi_{IJ}^2 - \epsilon_2^2} \,. \label{ZwithF=0} \end{equation} The main achievement of this subsection is to show that in the limit of small $\epsilon_2$ the logarithm of \eqref{ZwithF=0} is given by \begin{equation} \log{\cal Z}_{\text{Short}}\, = \,\frac{1}{\epsilon_2}\, \int {d\phi_0 \over 2 \pi i} \, {\rm Li}_2[qQ(\phi_0)] \,+\,{\cal O}(\epsilon_2^0) \, . \label{resultLi2}\end{equation} We will see that, as opposed to the case of long range interactions considered in the previous subsection, the right hand side of \eqref{resultLi2} is not given by summing over tree graphs but by a single local term. As will be explained in details in Section \ref{sec:fullexpansion}, this is the underlying mechanism by which the short-range interactions turn a group of particles into a single effective particle. The result \eqref{resultLi2} explains why the dilogarithm function appears in the TBA. It will also be an essential input in the all order proof presented in Section \ref{sec:nonpert}. Considering its importance, we will prove \eqref{resultLi2} in two different ways. \subsubsection{Dilogarithm from a sum over residues} In this subsection we will prove \eqref{resultLi2} by direct evaluating \eqref{ZwithF=0} and \eqref{resultLi2} as a sum over residues. We find it instructive to first consider the partition function for the $U(1)$ gauge theory, i.e.~we evaluate \eqref{ZwithF=0} for \begin{equation} Q(x)\,=\,\frac{1}{(x-a-i\,0)(x-a+\epsilon+i\,0)}\,, \label{QN=1} \end{equation} compare to \eqref{QP}. In \eqref{QN=1} $a$ and $\epsilon$ are real and we wrote explicitely the $i\,0$ prescription. For each $k$ in \eqref{ZwithF=0}, there is only one residue (up to permutation of the integration variables) in the upper half plane, compare to the general discussion in Appendix \ref{App:contourRES}. It is given by $(\phi_1,\dots,\phi_k)=(a,a+\epsilon_2,\dots,a+(k-1)\epsilon_2)$. It follows that \begin{equation} \mathcal{Z}^{U(1)}_{\text{Short}} \,=\,\sum_{k=0}^{\infty}\,\frac{q^k}{k!}\,\frac{1}{\epsilon^k_2}\,\prod_{I=1}^k\, \frac{1}{\epsilon_1+I\,\epsilon_2}\,. \label{ZU1fromRES} \end{equation} One can also directly calculate the one dimensional integral \begin{equation} \int\,\frac{d \phi_0}{2\pi i}\,Q(\phi_0)^\ell\,=\,\frac{(-1)^{\ell+1}}{\epsilon^{2\ell-1}}\,\binom{2\ell-2}{\ell-1}\,, \label{QtothenU1} \end{equation} where $Q(x)$ is given in \eqref{QN=1}. Using the identity \eqref{formulaforres} one concludes the the logarithm of \eqref{ZU1fromRES} in the $\epsilon_2\rightarrow 0$ limit is indeed given by \eqref{resultLi2} with \eqref{QtothenU1}. For the more general $U(N)$ theory, it is easy to classify the poles contributing to \eqref{ZwithF=0} following the same reasoning as in Appendix \ref{App:contourRES}. For fixed $k$, residues are classified, up to permutation of the $k$ particles, by a set of integers $\{ s_1, s_2, \ldots, s_N \}$ such that $\sum_\lambda s_\lambda = k$. The corresponding pole is given by \begin{equation} \phi_{\kappa_{\lambda}+I_{\lambda}}\,=\,a_{\lambda}+(I_{\lambda}-1)\epsilon_2\,, \qquad I_{\lambda}=1,\dots,s_{\lambda}\,,\qquad \kappa_{\lambda}=\sum_{\mu=1}^{\lambda-1}\,s_{\mu}\,, \label{poleF=0} \end{equation} and $\lambda=1,\dots,N\,$. As opposed to the full partition function (\ref{Zinst}), for which residues are classified by $N$-tuple of Young tableaux with a total number of boxes equal to $k$, in the simplified integrals \eqref{ZwithF=0} only Young tableaux with one row, whose length is denoted by $s_{\lambda}$, contribute. It is straigthforward to calculate and collect all residues, see Appendix \ref{App:UsefulFormula}, to obtain \begin{equation} \mathcal{Z}_{\text{Short}}\,=\, \sum_{s_1,\dots,s_{N}\geq 1}\, \prod_{\lambda=1}^N\, \left( \frac{q^{s_{\lambda}}\,\mathfrak{R}^{(\lambda)}_{s_\lambda}}{s_{\lambda}!\,\epsilon_2^{s_{\lambda}}}\right)\, \prod_{\lambda < \lambda'}\prod_{I_\lambda, I_{\lambda'}} {(a_{\lambda \lambda'} + (I_\lambda - I_{\lambda'}) \epsilon_2)^2 \over (a_{\lambda \lambda'} + (I_\lambda - I_{\lambda'}) \epsilon_2)^2 - \epsilon_2^2}\,, \label{ZD=0res} \end{equation} where \begin{equation} \mathfrak{R}^{(\lambda)}_s\,:=\, \prod_{I=1}^s\,\frac{Q^{(\lambda)}(a_{\lambda_1}+(I-1)\epsilon_2)}{\epsilon_1+I\epsilon_2}\,,\qquad Q^{(\lambda)}(x)\,=\,\frac{1}{P^{(\lambda)}(x)P^{(\lambda)}(x+\epsilon)}\,, \label{ZD=0res2} \end{equation} and $P^{(\lambda)}(x)\,=\,\prod_{\mu\neq\lambda}(x-a_{\mu})$. The structure of the result \eqref{ZD=0res}-\eqref{ZD=0res2} represents a simple generalization of the $U(1)$ computation \eqref{ZU1fromRES}. The $\epsilon_2\rightarrow 0$ limit of the logarithm of \eqref{ZD=0res} can be readily obtained using the relation \eqref{usufulresidue2}. One recognizes that \begin{equation} \log\,\mathcal{Z}_{\text{Short}}\,=\,\frac{1}{\epsilon_2}\, \sum_{\lambda=1}^N\,\sum_{\ell=1}^{\infty}\,\frac{q^{\ell}}{\ell^2}\,\text{Res}_{\phi_0=a_{\lambda}}\left[Q^\ell(\phi_0)\right]\,+{\cal O}(\epsilon_2^0)\,\,. \end{equation} This result coincides with the evaluation of \eqref{resultLi2} by residues. This calculation provides a direct proof of \eqref{resultLi2}. \subsubsection{By Mayer expansion and separation of regions} \label{byregions} We now calculate (\ref{ZwithF=0}) for small $\epsilon_2$ by first applying Mayer expansion and then the so-called method of expansion by regions. The first step is to decompose the interaction factor as \begin{equation} {\phi_{IJ}^2 \over \phi_{IJ}^2 - \epsilon_2^2} = 1 + f_{IJ} \,, \qquad f_{IJ} := {\epsilon_2^2 \over \phi_{IJ}^2 - \epsilon_2^2} \, . \label{f_IJ} \end{equation} In a similar way as \eqref{simple}, the logarithm of (\ref{ZwithF=0}) is then given by \begin{equation} \log {\cal Z}_{\text{Short}} = \sum_{k=1}^\infty {q^k \over k! \epsilon_2^k} \int \prod_{I=1}^k {d \phi_I \over 2\pi i} \, {Q(\phi_I) } \sum_{g\,\in\, {\cal G}_c^{[k]}}\, \prod _{e(I,J)\in g} f_{IJ} \, . \label{LogZF=0} \end{equation} To show that this gives \eqref{resultLi2}, we need to evaluate the integrals entering \eqref{LogZF=0} as a Laurant series in $\epsilon_2$. We are actually interested only in the terms with leading negative powers of $\epsilon_2$. We use a powerful method, usually applied to the evaluation of Feynman integrals, called \emph{expansion by regions} introduced in \cite{Beneke:1997zp}. The method goes as follows: $(1)$ divide the integration domain into regions and expand the integrand in a Taylor series in small parameters, $(2)$ extend the integration to the full domain of integration, $(3)$ set to zero scaleless integrals. This method turns out to be particularly efficient to calculate the leading term for the integrals \eqref{LogZF=0}. Concerning these contributions, we do not need to apply the somewhat subtle point $(3)$ above. A more detailed discussion of this method will be given in Appendix \ref{app:region}. To identify the set of relevant regions, we note that the Taylor expansion of the interaction term ${\epsilon_2^2 \over \phi_{IJ}^2 - \epsilon_2^2}$ starts at order $\epsilon_2^2$ except for the \emph{region} in which $\phi_{IJ}$ is of order $\epsilon_2$. It is then natural to expect that regions are classified as follows. Let $\mathcal{B}^{[k]}$ denotes the set of grouping of $k$ labeled particles in clusters, see figure \ref{cluster_3} for the $k=3$ case. For each grouping we define the corresponding region as \begin{equation} \begin{cases} |\phi_{IJ}|\, \sim\,\epsilon_2 & \text{if $I,J$ are in the same cluster,}\\ |\phi_{IJ}|\, \gg\,\epsilon_2 & \text{if $I,J$ are in different clusters.} \end{cases}\, \end{equation} \begin{center} \begin{figure}[t] \centerline{\includegraphics[height=1.1cm]{clusterdecomp} } \caption{\it The decomposition of the contribution for $k=3$ case. Each square represents a cluster labeled by corresponding particles in the cluster.} \label{cluster_3} \end{figure} \end{center} The next step is to Taylor expand the integrand in each region. We denote by $\mathbb{T}^{(\mathsf{b})}\left[\dots\right]$ the operation of Taylor expaning $\dots$ in the region corresponding to the grouping $\mathsf{b}\,\in\,\mathcal{B}^{[k]}$. We have \begin{equation} \mathbb{T}^{(\mathsf{b})}\left[ f_{IJ}\right]\,=\, \begin{cases} {\cal O}(1) & \text{if $I,J$ are in the same cluster,}\\ {\cal O}(\epsilon_2^2) & \text{if $I,J$ are in different clusters,} \end{cases}\, \label{Tbf} \end{equation} and \begin{equation} \mathbb{T}^{(\mathsf{b})}\left[ \prod_{I=1}^k Q(\phi_I)\right]\,=\, \prod_{\mathbf{Y}\,\in\,\mathsf{b}}\,Q(\phi_{\mathbf{Y}})^{|\mathbf{Y}|} \,+\,\dots\,\qquad \phi_{\mathbf{Y}}\,:=\,\frac{1}{|\mathbf{Y}|}\,\sum_{I\,\in\,\mathbf{Y}}\,\phi_I\,, \label{TbQQ} \end{equation} the product on the right hand side runs over the clusters in the grouping $\mathsf{b}$ and $|\mathbf{Y}|$ denotes the number of particles in the cluster $\mathbf{Y}$. From \eqref{Tbf}-\eqref{TbQQ} it is clear that, for each $k$, the leading contribution to \eqref{LogZF=0} comes from the region in which all $\phi_I$ are in the same cluster. Indeed, if there were two or more clusters, the $f_{IJ}$ factor, with $I,J$ in different clusters, would decrease the power of $\epsilon_2$, compare to \eqref{Tbf}. An explicit example of this expansion of $k=3$ is given in appendix \ref{app:region}. As we will shortly see, the leading contribution to \eqref{LogZF=0} starts at order $1/\epsilon_2$ for each $k$. This fact is not obvious from the form \eqref{LogZF=0}. By further separating the integration over the ``center of cluster'' coordinate $\bar{x}$ as \begin{equation} \int\,\prod_{I=1}^sdx_I\,F(x_1,\dots,x_s)\,=\, \int\,d\bar{x}\, \int\,\prod_{I=1}^s dx'_I \,\delta\left(\frac{1}{k}\sum_{I=1}^s\,x'_I\right)\, F(\bar{x}+x'_1,\dots,\bar{x}+x'_s)\,, \label{introduceCenter} \end{equation} where we applied the change of variables \begin{equation} \bar x := {1\over k} \sum_{I=1}^s x_s \,, \qquad x'_I := x_I - \bar x \,, \end{equation} one can rewrite \eqref{LogZF=0} as \begin{equation} \log {\cal Z}_{\text{Short}} \,=\,\frac{1}{\epsilon_2} \sum_{k=1}^\infty {q^k \over k!}\,\int d \phi_0 \,Q^k(\phi_0)\,\mathcal{J}_k\,+\,{\cal O}(\epsilon_2^0)\, \label{logZregionsLi2} \end{equation} where \begin{equation} \mathcal{J}_k\,:=\, \,{1\over \epsilon^{k-1}_2} \int \prod_{I=1}^k {d \phi_I \over 2\pi i} \,\delta\left(\frac{1}{k}\sum_{I=1}^k\,\phi_I\right) \sum_{g\,\in\, {\cal G}_c^{[k]}}\, \prod _{e(I,J)\in g} f_{IJ} \, . \label{defJk} \end{equation} Notice that in \eqref{defJk} all connected graphs, independently on the number of edges, contribute to the leading term. The next observation to be made is that \begin{equation} \text{if $g$ is not connected}\qquad \,{1\over \epsilon^{k-1}_2} \int \prod_{I=1}^k {d \phi_I \over 2\pi i} \,\delta\left(\frac{1}{k}\sum_{I=1}^k\,\phi_I\right) \prod _{e(I,J)\in g} f_{IJ}\,=\,0 \, . \label{zeronotconnected} \end{equation} The validity of this statement can be easily argued as follows. For each connected component in $g$ we can define its center as the average of the $\{\phi_I\}$ in that connected component. The integrand in \eqref{zeronotconnected} does not depend on the distance between the centers. As the integral is calculated by residue, it trivially vanishes in this case as, after integrating trivially the delta function, the $(k-1)$-dimensional residue is absent. The identity \eqref{zeronotconnected} implies that in \eqref{defJk}, we can replace the sum over connected graphs with the sum over all graphs. Finally using the relation \begin{equation} \sum_{g\,\in\, {\cal G}^{[k]}}\, \prod _{e(I,J)\in g} f_{IJ}\,=\,\prod_{1\leq I< J\leq k}(1+f_{IJ})\,, \end{equation} where ${\cal G}^{[k]}$ is the set of all graphs on $[k]$, we conclude that $\mathcal{J}_k=\mathcal{I}_k$, where, using \eqref{f_IJ}, \begin{equation} \mathcal{I}_k\,:=\, \,{1\over \epsilon^{k-1}_2} \int \prod_{I=1}^k {d \phi_I \over 2\pi i} \,\delta\left(\frac{1}{k}\sum_{I=1}^k\,\phi_I\right)\, \prod_{1\leq I< J\leq k} {\phi_{IJ}^2 \over \phi_{IJ}^2 - \epsilon_2^2} \,=\,\frac{1}{2 \pi i }\,\frac{k!}{k^2}\,. \label{resultJk} \end{equation} Notice that the evaluation of this integral is exact, see \cite{Moore:1998et} for a derivation. Using this result we recognize that \eqref{logZregionsLi2} is equal to \begin{equation} \log {\cal Z}_{\text{Short}} \,=\,\frac{1}{\epsilon_2}\, \sum_{k=1}^\infty {q^k \over k^2}\,\int \frac{d \phi_0}{2 \pi i} \,Q^k(\phi_0)\,+\,{\cal O}(\epsilon_2^0)\,\, = \,\frac{1}{\epsilon_2}\, \int {d\phi_0 \over 2 \pi i} \, {\rm Li}_2[qQ(\phi_0)] \,+\,{\cal O}(\epsilon_2^0) \,. \end{equation} \subsection{Expansion of the instanton partition function} \label{sec:fullexpansion} We are now ready to consider the full instanton partition function (\ref{Zinst}). We will combine the considerations from the previous sections. In Section \ref{sec:simple} we learned that the $\epsilon_2\rightarrow 0$ limit singles out certain tree level graphs. In Section \ref{originDILOG} we learned that the factors $\frac{\phi_{IJ}^2}{\phi_{IJ}^2-\epsilon_2^2}$ produce the effect of combining particles into clusters. To exploit the combination of these two mechanisms in the most transparent way we find it convenient to use the so-called iterated Mayer expansion advocated at the beginning of our analysis. We start by reviewing this expansion. Based on this iterative expansion, by combining the discussion from the two previous sections, one obtains a tree graph expansion for the full instanton pationtion as given in Section \ref{sec:ourcase}. In Appendix \ref{App:regionsplusMayer} we show that the same result can be obtained by first applying the method of expansion by regions to the original partition function and then exploiting some combinatorics to conclude that only certain connected graphs contribute to the free energy. \subsubsection{Iterated Mayer expansion} \label{iteratedMayer} We start with a review of the iterated Mayer expansion \cite{Gopfert:1981zu}. Consider the partition function \begin{equation} \mathcal{Z}\,=\,\sum_{k=0}^{\infty}\,\frac{q^k}{k!}\int \,\prod_{I=1}^k d\phi_I A(\phi_I)\prod_{1\leq I<J\leq k}\left(1+\mathcal{F}_{IJ}\right)\,,\qquad 1+\mathcal{F}_{IJ}\,=\,e^{-\left(V^{\text{int}}_S(\phi_{IJ})+V^{\text{int}}_L(\phi_{IJ})\right)}\,, \label{Zforiterated} \end{equation} where, as in \eqref{splittingPOT}, we split the pair-wise interaction potential in a short and long range part. Next we introduce \begin{equation} \mathfrak{g}_{\mathbf{Y}}\,=\, \prod_{\substack{I,J\,\in\,\mathbf{Y}\\I<J}}\,\left(1+\,e^{-V^{\text{int}}_L(\phi_{IJ})}\right)\,,\,\qquad 1+f_{IJ}\,=\,e^{-V^{\text{int}}_S(\phi_{IJ})}\,, \end{equation} \begin{equation} 1+\mathfrak{f}_{\mathbf{Y}_a,\mathbf{Y}_b}\,=\,\prod_{I_a\,\in\,\mathbf{Y}_a}\prod_{I_b\,\in\,\mathbf{Y}_b} \,e^{-V^{\text{int}}_L(\phi_{I_aI_b})}\,,\qquad a\neq b \end{equation} where $\mathbf{Y}$ denotes a set of particles or, in other words, a cluster. Iterated Mayer expansion is the statement that the free enery can be expanded as \begin{equation} \log\, \mathcal{Z}\,=\,\sum_{k=1}^{\infty}\,\frac{q^k}{k!}\, \sum_{\ell\geq 0}\,\sum_{\{\mathbf{Y}_1,\dots,\mathbf{Y}_\ell\}\,\in\,\mathcal{B}^{[k]}_{\ell}}\, \int\,\prod_{I=1}^k d\phi_I \,\, \mathfrak{S}_{\{\mathbf{Y}_1,\dots,\mathbf{Y}_{\ell}\}}\,, \label{Mayer} \end{equation} where \begin{equation} \mathfrak{S}_{\{\mathbf{Y}_1,\dots,\mathbf{Y}_{\ell}\}}\,=\, \left(\prod_{a=1}^{\ell}\,\mathfrak{g}_{\mathbf{Y}_a}\,S_{\mathbf{Y}_a}\right)\, \sum_{g\,\in\,\mathcal{G}^{[\ell]}_c} \,\prod_{e(a,b)\,\in\, g}\,\mathfrak{f}_{\mathbf{Y}_a,\mathbf{Y}_b}\,, \label{Sgothicdef} \end{equation} \begin{equation} S_{\mathbf{Y}}\,=\,\prod_{I=1}^k A(\phi_I)\, \sum_{g\,\in\,\mathcal{G}^{\mathbf{Y}}_c}\, \prod_{e(I,J)\,\in\, g}\,f_{IJ}\,, \label{Sdef} \end{equation} The sum in \eqref{Mayer} is taken over $\mathcal{B}^{[k]}_{\ell}$: groupings of $k$ (labeled) particles into $\ell$ clusters. As the notation may need some time to be digested, in Appendix \ref{App:ExampleIteratedMayer} we spell out the definitions for $k=2,3$. Note that although the cluster here is in a different context, the picture of grouping is similar to that used before in expansion by regions, see for example figure \ref{cluster_3} for $k=3$ case. \subsubsection{Iterated Mayer expansion for the Nekrasov partition function} \label{sec:ourcase} We can apply the iterated Mayer expansion reviewed in Section \ref{iteratedMayer} to the full partition function, compare \eqref{Zforiterated} to \eqref{Zinst}, \eqref{splittingPOT}. Once this is done we need to evaluate the leading contribution for small $\epsilon_2$ to integrals of the type \eqref{Mayer}. The crucial observation is that the Taylor expansion $\tilde{D}(\phi_{IJ})-1\sim \,\epsilon_2\,G(\phi_{IJ})$ is valid everywhere in the domain of integration\footnote{ In the language of the method of expansion by regions we would say that the leading term in the Taylor expansion of $\tilde{D}(\phi_{IJ})-1$ is the same in \emph{any region}.}. This immidiately implies that in the limit of small $\epsilon_2$ one can write, see definition \eqref{Sgothicdef}, \begin{equation} \mathfrak{S}_{\{\mathbf{Y}_1,\dots,\mathbf{Y}_{\ell}\}}\,=\,\epsilon_2^{\ell-1}\, \prod_{a=1}^{\ell}\,S_{\mathbf{Y}_a}\, \sum_{g\,\in\,\mathcal{T}^{[\ell]}_c} \,\prod_{e(a,b)\,\in\, g}\,\prod_{\substack{I_a\,\in\,\mathbf{Y}_a\\I_b\,\in\,\mathbf{Y}_b}} G(\phi_{I_aI_b})+\dots \label{equationTREE} \end{equation} where $\dots$ refers to next to leading contributions in $\epsilon_2$. We point out that, in analogy with \eqref{simpleLIMIT}, only connected \emph{tree graphs} contributes to \eqref{equationTREE}. As opposed to \eqref{simpleLIMIT}, now they are tree graphs on the set $[\ell]$ of $\ell$ clusters rather then the set of $k$ fundamental particles. Notice that we did not expand the $S_{\mathbf{Y}}$ factor. The integration in \eqref{Mayer} is still over $k$ variables. The next step is to explicitly perform the integration over the distances of particles within the same cluster. This turns out to be essentially the same as in Section \ref{byregions}. For each cluster $\mathbf{Y}_a$ we introduce a ``center of cluster'' coordinate $\overline{\phi}_a$ as in \eqref{introduceCenter}. It is clear that the $f_{IJ}$ entering the factors $S_{\mathbf{Y}_a}$, see \eqref{Sdef}, are independent of the ``center of cluster'' coordinates $\overline{\phi}_a$. Using this observation and \eqref{equationTREE} we write \begin{align} \int\,\prod_{I=1}^k d\phi_I \,\, \mathfrak{S}_{\{\mathbf{Y}_1,\dots,\mathbf{Y}_{\ell}\}}\qquad\qquad\qquad\qquad\qquad\qquad\qquad \nonumber \\ =\,\frac{1}{\epsilon_2}\,\int\, \prod_{a=1}^{\ell}\left[d\overline{\phi}_a\,d\mu_{\mathbf{Y}_a}\prod_{I_a\,\in\,\mathbf{Y}_a}Q(\overline{\phi}_a+\phi_{I_a})\right]\, \sum_{g\,\in\,\mathcal{T}^{[\ell]}_c} \,\prod_{e(a,b)\,\in\, g}\, \,\prod_{\substack{I_a\,\in\,\mathbf{Y}_a\\I_b\,\in\,\mathbf{Y}_b}} G(\overline{\phi}_{ab}+\phi_{I_aI_b})+\dots \label{resultMayer} \end{align} where \begin{equation} d\mu_{\mathbf{Y}}\,:=\,\frac{1}{\epsilon^{|\mathbf{Y}|-1}_2}\, \prod_{I\,\in \,\mathbf{Y}_a }\,\frac{d\phi_I}{2 \pi i }\, \delta\left(\frac{1}{|\mathbf{Y}|}\sum_{I\,\in\,\mathbf{Y}}\,\phi_I\right)\, \sum_{g\,\in\,\mathcal{G}^{\mathbf{Y}}_c}\, \prod_{e(I,J)\,\in\, g}\,f_{IJ}\,. \end{equation} As extensively discussed in the previous sections, the leading contribution in $\epsilon_2$ of integrals of the type \eqref{resultMayer}, can be obtained by neglecting the deviation from the center of cluster coordinate in the functions $Q(x)$ and $G(x)$. Doing so, \eqref{resultMayer} becomes \begin{equation} \int\,\prod_{I=1}^k d\phi_I \,\, \mathfrak{S}_{\{\mathbf{Y}_1,\dots,\mathbf{Y}_{\ell}\}}\,=\, \frac{1}{\epsilon_2}\,\int\, \prod_{a=1}^{\ell}d\overline{\phi}_a\,Q^{|\mathbf{Y}_a|}(\overline{\phi}_a)\,\mathcal{J}_{|\mathbf{Y}_a|}\, \sum_{g\,\in\,\mathcal{T}^{[\ell]}_c} \,\prod_{e(a,b)\,\in\, g}\,|\mathbf{Y}_a|\,|\mathbf{Y}_b|\, G(\overline{\phi}_{ab})+\dots \label{resultMayer2} \end{equation} $\mathcal{J}_{k}$ was defined in \eqref{defJk} and computed in \eqref{resultJk}. For convenience we recall it here \begin{equation} \mathcal{J}_{|\mathbf{Y}_a|}\,=\,\int\,d\mu_{\mathbf{Y}_a} \,=\,\frac{1}{2\pi i}\,\frac{|\mathbf{Y}_a|!}{|\mathbf{Y}_a|^2}\,. \end{equation} In the limit of small $\epsilon_2$, the summands in \eqref{Mayer} depend only on the sizes $n_1,\dots,n_{\ell}\geq 1$ of the clusters corresponding to $\mathsf{b}\,\in\,\mathcal{B}^{[k]}$. Converting the sum in \eqref{Mayer} from a sum over groupings of $k$ particles to a sum over the number of clusters $\ell$ and their sizes $n_1,\dots,n_{\ell}$ produces a factor $\frac{k!}{\ell! n_1!\,\dots\,n_{\ell}!}$. Assembling the pieces together we finally arrive at the main result \begin{equation} \lim_{\epsilon_2\to 0}\epsilon_2\,\log\,\mathcal{Z}\,=\,\sum_{\ell=1}^{\infty}\frac{1}{\ell!}\,\sum_{n_1,\dots,n_{\ell}\geq 1}\, \int\,\prod_{a=1}^{\ell}\frac{d\phi_a}{2\pi i}\,\frac{q^{n_a}\,Q^{n_a}(\phi_a)}{n_a^2}\, \sum_{g\,\in\,\mathcal{T}_c^{[\ell]}}\,\prod_{e(a,b)\,\in\,g}\,n_a\,n_b\,G(\phi_{ab})\,, \label{mainresultfromMayer} \end{equation} where we replaced $\overline{\phi}_a$ with $\phi_a$. We emphasize once again that the sum has been rearranged from a sum over particles, weighted by $q$, to a sum over clusters. More precisely, for each $\ell$ there are $\ell$ clusters and $\ell-1$ $G$-factors. For $\ell=1$ one immediatly recovers the dilogarithm \begin{equation} \int\,\frac{d\phi}{2 \pi i }\,\text{Li}_2\left(q\,Q(\phi)\right)) \,. \label{Li2resum} \end{equation} For $\ell=2$ one has \begin{equation} \frac{1}{2}\,\int\,\frac{d\phi_1}{2 \pi i }\frac{d\phi_2}{2 \pi i }\mathcal{Q}_{\text{eff}}(\phi_1)\,\mathcal{Q}_{\text{eff}}(\phi_2)\,G(\phi_1-\phi_2) \,, \qquad \mathcal{Q}_{\text{eff}}(\phi)\,=\,\log(1-q\,Q(\phi))\,. \label{effectiveresum} \end{equation} The result \eqref{mainresultfromMayer} can be visualized diagrammatically in a simple way. For each $\ell$ in the sum one draws all three graphs on the set $\{1,2,\dots,\ell\}$. To each node $a$ of the graph is associated an integer $n_a\geq 1$. Once the diagrams are drawn, the corresponding integrals can be written using the following ``Feynann rules'' \begin{itemize} \item Vertex \begin{equation} \begin{tabular}{c}{\includegraphics[height=.9cm]{rule_vertex}}\end{tabular} \,= \ \,\frac{q^{n_a}}{n_a^2}\int\frac{d\phi_a}{2\pi i }\, Q^{n_a}(\phi_a)\, \end{equation} \item Propagator \begin{equation} \begin{tabular}{c}{\includegraphics[height=.9cm]{rule_prop}}\end{tabular} \,= \ \, n_a\,n_b\,G(\phi_{ab})\,. \end{equation} Here we used a dashed square to indicate that the propagator is associated to the edge only. \end{itemize} The graphs contributing to $\ell=1,2,3$ are given by \begin{eqnarray} \ell=1\,& : & \ \ \ \ \, \sum_{n_1=1}^{\infty}\, \begin{tabular}{c}{\includegraphics[height=.8cm]{rule_example1}}\end{tabular} \\ \ell=2\,& : & \ \ \, \sum_{n_1,n_2=1}^{\infty}\, \begin{tabular}{c}{\includegraphics[height=.8cm]{rule_example2}}\end{tabular} \\ \ell=3\,& : & \ \sum_{n_1,n_2,n_3=1}^{\infty}\, \begin{tabular}{c}{\includegraphics[height=.8cm]{rule_example3}}\end{tabular} \end{eqnarray} The first two graphs correspond to \eqref{Li2resum} and \eqref{effectiveresum} respectively. Notice that each integral obtained by applying the Feynman rules, should be multiplied by an overall $\frac{1}{\ell!}$ factor, which is explicit in \eqref{mainresultfromMayer}. Graphs with more vertices, upon dressing the vertices with positive integers, are as in Figure \ref{connect_tree_graph}. \section{Derivation of the TBA} In this section we show that the logarithm of the instanton partition function in the $\epsilon_2\rightarrow 0 $ limit, whose structure have been studied in last section, satisfies TBA equations. We will first provide a perturbative check to a very high order in the instanton counting parameter $q$, and then present an all order proof based on rewriting the instanton partition function as a $(0+1)$-dimensional path integral. \subsection{Perturbative expansion \label{perturbative_TBA}} Let us recall the expression of the superpotential coming from the TBA \begin{equation} {\cal W} = \int {dx \over 2\pi i} \Big[ -{1\over2} \varphi(x) \log\left(1 - q Q(x) e^{-\varphi(x)} \right) + {\rm Li}_2\left( q Q(x) e^{-\varphi(x)}\right) \Big] \,, \end{equation} where \begin{equation} \varphi(x) = \int {d y \over 2\pi i} \, G(x-y) \log\left( 1-qQ(y) e^{-\varphi(y)} \right)\,, \end{equation} compare to \eqref{tbaW} and \eqref{tbaphi}. The basic observation is that TBA equation has a natural expansion in terms of tree graphs, therefore one can compare it directly with the Mayer expansion of the instanton partition function \eqref{mainresultfromMayer}. One writes $\varphi$ and ${\cal W}$ as \begin{equation} \varphi(x) = \sum_{k=1}^\infty q^k \,\varphi_k(x) \,, \qquad {\cal W} = \sum_{k=1}^\infty q^k \,{\cal W}_k \, . \end{equation} $\varphi$ can be solved recursively via the TBA equation, for example up to the first two orders \begin{eqnarray} \varphi_1(x) &=& - \int {d y \over 2\pi i} G(x-y) Q(y) \, , \\ \varphi_2(x) &=& - {1\over2} \int {d y \over 2\pi i} G(x-y) Q(y)^2 - \int {d y \over 2\pi i} {d z \over 2\pi i} G(x-y) G(y-z) Q(y) Q(z) \,. \end{eqnarray} Substituting the expression for $\varphi$ in ${\cal W}$ and collecting the terms at a given order of $q$, one obtains, here up to order $q^3$ \begin{eqnarray} {\cal W}_1 &=& \int {d x \over 2\pi i} \, Q(x) \, , \\ {\cal W}_2 &=& {1\over 4}\int {d x \over 2\pi i} \, {Q(x)^2} + {1\over2} \int {dx \over 2\pi i} {dy \over 2\pi i} \, {Q(x) G(x-y) Q(y)} , \\ {\cal W}_3 &=& {1\over 9} \int {d x \over 2\pi i} {Q(x)^3} + {1\over2} \int {dx \over 2\pi i} {dy \over 2\pi i} \, {Q(x)^2 G(x-y) Q(y)} \nonumber\\ && + {1\over2} \int {dx \, dy \, dz \over (2\pi i)^3} \, {Q(x) G(x-y) Q(y) G(y-z) Q(z)} \,. \end{eqnarray} One can see that the expansion of ${\cal W}$ has a natural interpretation in terms of connected tree diagrams, with $G(x-y)$ as propagators and $Q(x)^i$ as vertices. This is the same structure as the one obtained starting from the integral representation of the instanton partition function, see \eqref{mainresultfromMayer}. The non-trivial thing to show is that not only the structure is similar but that the coefficients actally match. By staring at their form, it is not obvious that the Mayer expansion \eqref{mainresultfromMayer} and the TBA actually produce the same result. In particular to obtain the contribution of a single graph in \eqref{mainresultfromMayer} one needs to take into account a large number of terms coming from the $\varphi$ expansion. This procedure can be straightforward implemented on a compute. We have checked up to $q^7$ order and found perfect agreement with the Mayer-Cluster expansion of the instanton partition function. It is remarkable that these two different structures are actually equivalent with each other. In the next subsection, we will see the equivalence of the two expansions can be proven to all order in $q$. \subsection{TBA as saddle-point of a $(0+1)$-dimensional path integral \label{sec:nonpert}} As we will shortly review, there is a nice way to rewrite a grand canonical partition function of the form \eqref{Zinst} as a path integral, based on a trick of Polyakov \cite{Polyakov}. Path integrals are particularly well suited to be studied in the classical limit $\hbar\rightarrow 0$, identified with $\epsilon_2$ in this section, where the path integral is obtained by evaluating the classical action on some solution to the equations of motion. In the case of the instanton partition function \eqref{Zinst}, this limit is more subtle as the two particle potential in the original statistical mechanical problem is singular in this limit. Using the considerations from section \ref{originDILOG} we will be able to disentangle such contributions and derive the TBA in a rather transparent way as a saddle point equation of an appropriate action. Let us start by reviewing the steps presented in the appendix of \cite{Polyakov}. Consider the partition function \begin{equation} Z = \sum_{k=0}^\infty {q^k \over k!} \int \prod_{I=1}^k {d \phi_I \over 2\pi i} \prod_{I < J}^k \, e^{-V(\phi_{IJ})} \,. \label{ZforPolyakov}\end{equation} The basic idea is to rewrite the interactions as \begin{equation} \prod_{I<J}^k e^{-V(\phi_{IJ})}\,=\,e^{\Delta({\delta\over\delta \varphi})} \, \prod_{I=1}^k e^{-\varphi(\phi_I)} \Big|_{\varphi\rightarrow0} \,, \label{interPolya} \end{equation} where \begin{equation} \Delta\Big({\delta\over\delta \varphi}\Big) = - {1\over2} \int d x \, dy\, V(x-y) {\delta \over \delta \varphi(x)} {\delta \over \delta \varphi(y)} \,.\end{equation} The operator $e^{\Delta}$ can be pulled out from the sum over the number of particles $k$ to rewrite \begin{equation} Z = e^{\Delta({\delta\over\delta \varphi})} \bigg( \sum_{k=0}^\infty {1 \over k!} \int \prod_{I=1}^k {d \phi_I \over 2\pi i} e^{-\varphi(\phi_I)} \bigg) \bigg|_{\varphi=0} = e^{\Delta({\delta\over\delta \varphi})} \exp\left( \int {d\phi_0 \over 2 \pi i} \, { e^{-\varphi(\phi_0)}}\right) \bigg|_{\varphi\rightarrow 0}\ . \end{equation} All the information about the interaction in the original gas \eqref{ZforPolyakov} is now encoded in the operator $\Delta$. To translate the construction above into path integral language one observes the following equality \begin{equation} e^{\Delta({\delta\over\delta \varphi})}\,\mathsf{F}[\varphi]\Big{|}_{\varphi=0}\,=\, \int [ D \varphi\, D\rho ]\, e^{\frac{1}{\hbar} \left[ \hbar\,\Delta(\frac{i \rho}{2\pi \hbar})+\int\frac{dx}{2\pi i }\rho(x)\,\varphi(x)\right]}\,\mathsf{F}[\varphi]\,. \label{fromDeltatoPATH} \end{equation} This can be formally proved by Fourier transformation\ \footnote{The relation \eqref{fromDeltatoPATH} is a path integral analogue of \begin{equation} e^{\Delta({\partial \over \partial x})} \, \mathbf{f}(x)\big|_{x\rightarrow0} = \int {d p \over 2\pi } \, e^{ i p x + \Delta(i p)} {\hat {\mathbf{f}}}(p)\big|_{x\rightarrow0} = \int {d p dy \over 2\pi} \, e^{-i p y + \Delta(i p)} {\mathbf{f}}(y) \, , \label{FTtrick} \end{equation} where ${\hat {\mathbf{f}}}(p)$ is introduced via Fourier transformation.}. The measure in $[ D \varphi\, D\rho ]$ is defined in such a way that if $\mathsf{F}[\varphi]=1$ then the resulting path integral should also evaluate to $1$. Notice that the relation \eqref{fromDeltatoPATH} holds also when the operator $\Delta$ is not quadratic. If $\Delta$ is linear $\Delta=\int_x \gamma(x){\delta\over\delta \varphi(x)}$, the $\rho$ integration produces a delta function and the resulting path integral gives $\mathsf{F}[\gamma]$ as it should. Indeed in this case $e^{\Delta}$ is just the shift operator. From the discussion in Section \ref{sec:Mayer-Cluster}, it is rather clear that, in the study of the instanton partition function, we should apply the procedure just described in a sightly modified way. In particular we will factor out only the long range part of the potential \eqref{splittingPOT}, corresponding to $\widetilde{\mathcal{D}}$, as we will now describe. The first step is to introduce \begin{eqnarray} \Delta_{\rm L} \Big({\delta\over\delta \varphi}\Big)= - {1\over2} \int d x \, dy\, V_{\rm L}(x-y) {\delta \over \delta \varphi(x)} {\delta \over \delta \varphi(y)}\,, \qquad e^{-V_{\rm L}(x)}\,=\,\widetilde{\mathcal{D}}(x) \,. \end{eqnarray} As above, one uses this operator to rewrite the long range interactions as in \eqref{interPolya}, and factor out their contribution from the sum over the instanton number $k$ as follows \begin{equation} {\cal Z} = e^{\widetilde\Delta({\delta\over\delta \varphi})}\, \mathbf{F}[\varphi] \Big|_{\varphi=0} \,, \qquad \mathbf{F}[\varphi] = \sum_{k=0}^\infty {1 \over k!} \int \prod_{I=1}^k {d \phi_I \over 2\pi i}{q Q(\phi_I) e^{-\varphi(\phi_I)} \over \epsilon_2} \prod_{I < J}^k {\phi_{IJ}^2 \over \phi_{IJ}^2 - \epsilon_2^2}\,.\label{Zgood} \end{equation} Notice that we did not expand the short range interaction ${\phi_{IJ}^2 \over \phi_{IJ}^2 - \epsilon_2^2}$, which still appears in $\mathbf{F}[\varphi]$. The reader may recognize that we already calculated this integral in the $\epsilon_2\rightarrow 0$ limit in Section \ref{originDILOG}. More precisely $\mathbf{F}[\varphi]$ coincides with $\mathcal{Z}_{\text{Short}}$ defined in \eqref{ZwithF=0}, upon replacing $Q(x)$ with $Q(x) e^{-\varphi(x)}$. Using the result \eqref{resultLi2} we can thus rewrite \begin{equation} \mathbf{F}[\varphi] \,=\,\exp\left(\frac{1}{\epsilon_2}\, \int {d\phi_0 \over 2 \pi i} \, {\rm Li}_2[qQ(\phi_0)\, e^{-\varphi(\phi_0)}]+O(\epsilon_2^0)\right)\,. \label{Li2good} \end{equation} The appearance of the dilogarithm is quite remarkable. The derivation presented makes its origin from short range interactions completely transparent. The expression \eqref{Zgood} with \eqref{Li2good}, is well suited for the semiclassical analysis in the $\epsilon_2\rightarrow 0 $ limit. Using \eqref{fromDeltatoPATH} in this case, and recalling that $V_{\rm L}(x)=- \epsilon_2 G(x) + {\cal O}(\epsilon_2^2)$ one immediatly obtains \begin{equation} \mathcal{Z}\,=\,\int [ D \varphi\, D\rho ]\, \exp \left[ { \frac{1}{\epsilon_2}\left(\mathcal{Y}[\rho,\varphi]+O(\epsilon_2)\right)} \right] \,, \label{Zfinal}\end{equation} where $\mathcal{Y}[\rho,\varphi]$ is defined in \eqref{YYdef} and we identified $\hbar$ with $\epsilon_2$. In the limit of small $\epsilon_2$ this path integral is calculated by the sadle point method\,% \footnote{Alternatively, one can also integrate out $\rho$ exactly, as it appears quadratically in the action, before taking the $\epsilon_2\rightarrow 0 $ limit. The resulting path integral, in $\varphi$ in the limit of small $\epsilon_2$, is dominant by the saddle point of $\varphi$, where the saddle point equation is exactly the TBA equation \eqref{tbaphi}.}. The superpotential is thus obtained from the critical valeus of the functional $\mathcal{Y}[\rho,\varphi]$. As explained in Section \ref{sec:NS}, the TBA immidiately follows. This complets the derivation. \subsubsection{Multi-particle interactions} Let us mention that the above discussion may be generalized to the cases containing ``multi-particle" interactions, such as \begin{equation} \mathcal{Z} = \sum_{k=0}^\infty {q^k \over k!} \int \prod_{I=1}^k {d \phi_I \over 2\pi i}{q Q(\phi_I) \over \epsilon_2} \prod_{I < J}^k {\phi_{IJ}^2 \over \phi_{IJ}^2 - \epsilon_2^2} \prod_{m=2}^\infty \prod_{I_1 < \ldots< I_m}^k \, e^{-V_m(\phi_{I_1}, \phi_{I_2}, \ldots, \phi_{I_m})} \,. \end{equation} One can similarly introduce \begin{eqnarray} {\Delta_m} \Big( {\delta\over\delta \varphi} \Big) &=& - {1\over m!} \int d^m x \, V_m(x_1,...,x_m) \prod_{I=1}^m {\delta \over \delta \varphi(x_I)} \, . \end{eqnarray} In analogy with \eqref{fromDeltatoPATH}, this can be brought to the following form \begin{equation} \mathcal{Z} = \int [D \rho D \varphi] \, \exp\left[ - \sum_{m=2}^\infty {(-1)^m\over m!} \int {d^m x \over (2\pi i)^m} \, V_m(x_1,..,x_m) \prod_{I=1}^m {\rho(x_I) \over \epsilon_2} + { \int {d x \over 2\pi i \,\epsilon_2} \rho(x) \varphi(x)} \right] {\rm {\bf F}}[ \varphi] . \label{multi2} \end{equation} Notice that if $m>2$ one can no longer integrate out $\rho$ exactly. The tri-particle interaction is directly related to certain gauge theories. The main example is given by quiver theories containing matter in the trifundamental representation of $U(2)^3$ \cite{Gaiotto:2009we}. The contour integral representation of the instanton partition function for this case can be found in \cite{Hollands:2011zc}. Notice that one has $V_2=-\epsilon_2\,G+{\cal O}(\epsilon_2^2)$, $V_3=\epsilon_2^2\,G_3+{\cal O}(\epsilon_2^3)$. This is the right scaling to produce a $\frac{1}{\epsilon_2}$ factor in the exponent of \eqref{multi2} which is crucial for the semiclassical analysis. \section{Generalization to quiver gauge theories \label{sec:quiver}} Until now the object of study has been the instanton partition function for pure $\mathcal{N}=2$ Super-Yang-Mills with gauge group $SU(N)$, see \eqref{Zinst}. This partition function admits a natural generalization corresponding to other $\mathcal{N}=2$ gauge theories. In order to specify an $\mathcal{N}=2$ supersymmetric gauge theory the first step is to fix the following data: a gauge group $G$ and a set of irreducible representations $R$ of $G$ corresponding to its matter content. Such datum gives rise to a consistent, or UV-complete, theory only if the beta function $\beta_\mathsf{v}$ associated to each simple factor of $G=\prod_{\mathsf{v}}\,G_\mathsf{v}$ satisfies $\beta_\mathsf{v} \leq 0$. This requirement puts strong constraints on $R$ and $G$. A classification of consistent theories can be found in the recent work \cite{Bhardwaj:2013qia}. In particular, with the exception of $G=SU(2)^V$ with $R$ given by trifundamentals introduced in \cite{Gaiotto:2009we}, each irreducible component of $R$ is charged under at most two $G_\mathsf{v}$. In the following we restrict to the case \begin{equation} G\,=\,\prod_{\mathsf{v}=1}^V\,SU(N_\mathsf{v})\,, \qquad R\,=\,\sum_{\mathsf{v},\mathsf{w}} c_{\mathsf{v}\,\mathsf{w}}\, (\square_\mathsf{v}, \overline{\square}_\mathsf{w}) + \sum_\mathsf{v} \,n_\mathsf{v} \square_\mathsf{v}\,. \label{gaugematter} \end{equation} Theories in this family with vanishing beta functions falls into three classes, referred to as type I, II, II* in \cite{Nekrasov:2012xe}. The matrix $c$ in \eqref{gaugematter} is then identified with the adjacency matrix of the Dynkin diagram of type $ADE$ (type I) and $\widehat{A}\widehat{D}\widehat{E}$ (type II and II*). Any other consistent theory in the family \eqref{gaugematter} can be obtained as appropriate limit of theories of type I, II, II*. We fix $G$ and $R$ as in \eqref{gaugematter}. We denote respectively by $\mathsf{V}$ and $\mathsf{E}$ the set of vertices and edges of the graph represented by $c_{\mathsf{v}\,\mathsf{w}}$. We denote the number of vertices by $V=|\mathsf{V}|$. The instanton partition function depends on the following quantities \begin{itemize} \item $\Omega$-deformation parameters $\epsilon_1$ and $\epsilon_2$, we set $\epsilon=\epsilon_1+\epsilon_2$ \item gauge couplings $q_{\mathsf{v}}$, $\mathsf{v} \in \mathsf{V}$ \item Coulomb parameters $a\in$ Cartan subalgebra of $\text{Lie}(G)$, $a_{\mathsf{v},\lambda}$ for $\mathsf{v} \in \mathsf{V}$, $\lambda=1,\dots,N_{\mathsf{v}}$ \item fudametal masses $m_{\mathsf{v},f}$, $\mathsf{v}\in \mathsf{V}$, $f=1,\dots,n_\mathsf{v}$ \item bifundamental masses\footnote{To avoid confusion we recall that in our conventions a bifundamental hypermultiplet associated to the edge $\mathsf{e}$ corresponds to a factor $(\square_{s_{\mathsf{e}}},\, \overline{\square}_{t_{\mathsf{e}}})+(\square_{t_{\mathsf{e}}}, \,\overline{\square}_{s_{\mathsf{e}}})$ in \eqref{gaugematter}.} $m_\mathsf{e}$, $\mathsf{e} \in \mathsf{E}$. \end{itemize} We package masses and Coulomb parameters into polynomials as follows \begin{equation} M_\mathsf{v}(x) = \prod_{f=1}^{n_\mathsf{v}} (x - m_{\mathsf{v},f}) \,, \qquad P_\mathsf{v}(x) = \prod_{\lambda=1}^{N_\mathsf{v}} (x - a_{\mathsf{v},\lambda}) \, . \end{equation} The contour integral representation of the Nekrasov partition function takes the form \begin{equation} {\cal Z} = \sum_{k_1,\, \ldots,\,k_V}^\infty\, \left(\frac{\epsilon}{\epsilon_1\,\epsilon_2}\right)^{k}\, \frac{q^{k_1}_1}{k_1!}\,\dots\,\frac{q^{k_V}_V}{k_V!}\, \int\,\prod_{\mathsf{v}\,\in \,\mathsf{V}}\, \Big{(}[d\phi_\mathsf{\mathsf{v}}] \, z_{\mathsf{v},\,k} ^{\rm gauge}(\phi_\mathsf{v})\, z_{\mathsf{v},\,k}^{\rm fund}(\phi_\mathsf{v})\Big{)}\, \prod_{\mathsf{e}\,\in\,\mathsf{E} }\, z_{\mathsf{e},\,k}^{\rm bifund}(\phi_{s_{\mathsf{e}}},\phi_{t_{\mathsf{e}}})\, \label{Zquiver} \end{equation} where $k=\sum k_\mathsf{v}$ and $[d\phi_\mathsf{v}]=\prod_{I=1}^{k_\mathsf{v}}(2\pi i)^{-1} d\phi_{\mathsf{v},I}$. As above, $\mathsf{V}$ and $\mathsf{E}$ are respectively set of vertices and edges of the quiver. Moreover, given an edge $\mathsf{e}$, its orientation defines two vertices called ${s_{\mathsf{e}}}$ (source) and $t_{\mathsf{e}}$ (target). Using the expressions for $z_\mathsf{v} ^{\rm gauge}$, $z_\mathsf{v}^{\rm fund}$ and $z_{\mathsf{e}}^{\rm bifund}$ given in Appendix \ref{app:quiverblocks} we rewrite the integrand of \eqref{Zquiver} as \begin{equation} \prod_{\mathsf{v}\,\in \,\mathsf{V}} \,\left[ \left(\prod_{I=1}^{k_{\mathsf{v}}} \,Q_{\mathsf{v}}(\phi_{\mathsf{v},I})\right)\, \left( \prod_{1\leq I < J\leq\, k_\mathsf{v}}{\cal D}(\phi_{\mathsf{v},I}-\phi_{\mathsf{v},J}) \right)\right]\, \prod_{\mathsf{e}\,\in \,\mathsf{E}} \,\left[ \prod_{I=1}^{k_{\mathsf{s_{\mathsf{e}}}}}\prod_{J=1}^{k_{t_{\mathsf{e}}}}\,\mathcal{D}_{\mathsf{e}}(\phi_{s_{\mathsf{e}},I}-\phi_{t_{\mathsf{e}},J}) \right]\, \label{ZquiverINT} \end{equation} where \begin{equation} Q_{\mathsf{v}}(x) \,=\, { M_{\mathsf{v}}(x) \over P_{\mathsf{v}}(x)\, P_{\mathsf{v}}(x+ \epsilon)}\, \prod_{\mathsf{v}\stackrel{\mathsf{e}}{\rightarrow}\mathsf{v}'} P_{\mathsf{v}'}(x - m_{\mathsf{e}})\, \prod_{\mathsf{v}\stackrel{\mathsf{e}}{\leftarrow}\mathsf{v}'} P_{\mathsf{v}'}(x + m_{\mathsf{e}}+\epsilon) \, , \label{Qdef} \end{equation} \begin{equation} \mathcal{D}(x)\,=\,\Delta(x)\,\Delta(-x)\,,\qquad \mathcal{D}_{\mathsf{e}}(x)\,=\, [\Delta(m_{\mathsf{e}}+x)\,\Delta(m_{\mathsf{e}}-x)]^{-1}\,, \qquad \Delta(x)\,=\,\frac{x\,(x+\epsilon_1+\epsilon_2)}{(x+\epsilon_1)\,(x+\epsilon_2)}\,, \label{Ddef} \end{equation} Note that $\mathcal{D}$ is the same as for pure SYM, see \eqref{defD}. Moreover, we have the manifest symmetry ${\cal D}(x) = {\cal D}(-x)$ and ${\cal D}_{\mathsf{e}}(x) = {\cal D}_{\mathsf{e}}(-x)$. Explanations regarding the origin of this expression are collected in Appendix \ref{app:contour_integral}. The expression for the contribution to the partition function from bifundamental matter can be found in \cite{Shadchin:2005cc} and \cite{Billo:2012st}. For more general gauge groups and representation see e.g.~\cite{Hollands:2010xa,Hollands:2011zc} and references therein. \subsection{TBA for quiver gauge theories} The partition function \eqref{Zquiver} in the limit $\epsilon_2\rightarrow 0$ can be studied in a similar way as for the simpler case of pure $SU(N)$ SYM, see \eqref{Zinst}. In particular it can be characterized as solution of certain TBA equations given below. The important observation to be made is that as long as the bifundametal masses $m_{\mathsf{e}}$ are large compared to $\epsilon_2$, the pair-wise interaction term in the non-ideal gas interpretation of \eqref{ZquiverINT} splits into two terms with different scales, compare to \eqref{splittingPOT}. The first term ${x^2 \over x^2 -\epsilon^2_2}$, present for each gauge group, gives rise to the dilogarithm function in the TBA. The remaining factor in the pair-wise interaction is responsable for the kernel term in the TBA non-liear integral equation. The resulting TBA is summarized as follows. As in \cite{Nekrasov:2009rc}, the partition function can be written as the critical value of the Yang-Yang functional \begin{equation} {\cal Y}(\rho, \varphi) = {1\over2} \sum_{\mathsf{v},\mathsf{w}} \int \frac{d x}{2 \pi i }\,\frac{d y}{2 \pi i } \, \rho_\mathsf{v}(x) G_{\mathsf{v} \mathsf{w}}(x-y) \rho_\mathsf{w}(y) + \sum_\mathsf{v} \int \frac{d x}{2 \pi i }\Big[ \rho_\mathsf{v}(x) \varphi_\mathsf{v}(x) + {\rm Li}_2\left( q_\mathsf{v} Q_r(x) e^{-\varphi_\mathsf{v}(x)}\right) \Big] , \end{equation} where the variations with respect to $\varphi_\mathsf{v}(x)$ and $\rho_\mathsf{v}(x)$ give \begin{equation} \rho_\mathsf{v}(x) = - \log\left[ 1-q_\mathsf{v} \, Q_\mathsf{v}(x) e^{-\varphi_\mathsf{v}(x)} \right] , \qquad \varphi_\mathsf{v}(x) = - \sum_\mathsf{w} \int \frac{d y}{2 \pi i } \, G_{\mathsf{v} \mathsf{w}}(x-y) \rho_\mathsf{w}(y) \, . \end{equation} The twisted superpotential for the quiver gauge theory is obtained as \begin{equation} {\cal W} = \sum_\mathsf{v} \int \frac{d x}{2 \pi i } \Big[ -{1\over2} \varphi_\mathsf{v}(x) \log\left(1 - q_\mathsf{v}\, Q_\mathsf{v}(x) e^{-\varphi_\mathsf{v}(x)} \right) + {\rm Li}_2\left( q_\mathsf{v} \, Q_\mathsf{v}(x) e^{-\varphi_\mathsf{v}(x)}\right) \Big] , \end{equation} where $\varphi_\mathsf{v}(x)$ satisfy the TBA equation \begin{equation} \varphi_\mathsf{v}(x) = \sum_\mathsf{w} \int \frac{d y}{2 \pi i } \, G_{\mathsf{v} \mathsf{w}}(x-y) \log\left[ 1-q_\mathsf{w} \, Q_\mathsf{w}(y) e^{-\varphi_\mathsf{w}(y)} \right] . \end{equation} The various functions which contain the data of quiver theory are \begin{equation} G_{\mathsf{v}\mathsf{w}}(x) = {d \over d x} \left[ \delta_{\mathsf{v}\mathsf{w}} \log \left( { x+ \epsilon_1 \over x- \epsilon_1} \right) + c_{\mathsf{v}\mathsf{w}} \log \left( { x+ m_{\mathsf{v}\mathsf{w}} \over x + m_{\mathsf{v}\mathsf{w}} + \epsilon_1} \right) + c_{\mathsf{w}\mathsf{v}} \log \left( { x - m_{\mathsf{w}\mathsf{v}}-\epsilon_1 \over x - m_{\mathsf{w}\mathsf{v}}} \right) \right] \, , \label{Rvw} \end{equation} and $Q_{\mathsf{v}}(x)$ defined in \eqref{Qdef}. In \eqref{Rvw} we defined $m_{\mathsf{v}\mathsf{w}}:=m_{\mathsf{e}}$ if the vertices $\mathsf{v}$ and $\mathsf{w}$ are connected by the edge $\mathsf{e}$ and $m_{\mathsf{v}\mathsf{w}}:=0$ otherwise. The expression for the propagator $G_{\mathsf{v}\mathsf{w}}(x)$ given above is obtained as as \begin{equation} G_{\mathsf{v}\mathsf{w} }(x) = \begin{cases} {\rm Limit}_{\epsilon_2 \rightarrow0} {{\cal D}(x) - 1 \over \epsilon_2} & \text{for}\,\,\,\mathsf{v}=\mathsf{w} \\ {\rm Limit}_{\epsilon_2 \rightarrow0} {{\cal D}_{\mathsf{e}}(x) - 1 \over \epsilon_2} & \text{for}\,\,\,\mathsf{v}\stackrel{\mathsf{e}}{\text{\----}}\mathsf{w} \\ 0 & \text{otherwise} \end{cases} \end{equation} Note that in this definition we have $G_{\mathsf{v}\mathsf{w}}(x) = G_{\mathsf{w}\mathsf{v}}(-x)$. \section{Conclusion and Discussion \label{sec-discussion}} In this paper we studied the instanton partition functions of four dimensional ${\cal N}=2$ gauge theories in a special limit of $\Omega$ deformation parameters, namely, taking $\epsilon_2 \rightarrow0$ but keeping $\epsilon_1$ finite. We show explicitly that the instanton part of the twisted superpotential $\mathcal{W}$, which in statistical mechanics language is equal to the free energy, satisfies TBA equations. It is thus naturally identified with the Yang-Yang function of some quantum integrable model \cite{Nekrasov:2009rc}. Based on the proof, we are also able to generalize the correspondence to a general class of ${\cal N}=2$ quiver theories. In this case one obtains a set of coupled TBA equations with one equation for each node of the quiver. The starting point of the derivation is the contour integral representation of the Nekrasov partition function. It can be interpreted as the grand canonical partition function of a non-ideal gas of particles. Due to the singular structure of the two-particles interaction, an interesting effective description of this gas emerges in the $\epsilon_2\rightarrow 0$ limit. In this effective description the ``new particles" are clusters of the original particles and are subject to certain tree-level interaction only. This picture is best obtained by employing the so-called iterated Mayer expansion \cite{Gopfert:1981zu}. This structure is nicely captured by the TBA, in particular the clustering of particles and their effective interaction is nicely reproduced by the presence of the Li$_2$ function\footnote{This makes it drastically different from some other TBA obtained by simply applying saddle point method \cite{Poghossian:2010pn, Fucito:2011pn, Fucito:2012xc}. See also \cite{Bourgine:2013ipa}.}. An all order proof in the instanton counting parameter $q$ of the TBA equations is given applying a $(0+1)$-dimensional path integral trick, dating back to Polyakov \cite{Polyakov}. In studying the NS limit of the instanton partition function, one has to face the problem of evaluating integrals when some parameter, in our case $\epsilon_2$, is small. There is a systematic way to do so, particularly successful for calculating Feynman integrals, called the method of expansion by regions. While our results are obtained also independently of this method, it provides a simple way to single out the leading $\frac{1}{\epsilon_2}$ behavior of the logarithm of the partition function. This structure is somewhat reminiscent of the exponentiation of infrared divergences in gauge theories. It would be interesting to further study the applicability of this method to the type of integral considered here. In this paper we considered quiver gauge theories involving only $U(N)$ gauge group factors and matter in the fundamental/antifundamental or bifundamental representations. It is natural to extend this work by deriving TBA equation in the case of the other classical gauge groups $SO(N)$, $SP(2N)$ and to other matter content, e.g.~symmetric and antisymmetric representations of $U(N)$ or trifundamental of $U(2)^3$. Other interesting generalizations are the deformation of the partition function by chiral ring operators and the five dimensional lift of the partition funtion. The details of these studies will be reported elsewhere. One of the most challenging question for the future is to understand how the proposed TBA equations emerge from the quantum integrable model point of view. In the case of pure $SU(N)$ ${\cal N}=2$ super Yang-Mills, which corresponds to the periodic Toda chain with $N$ sites, this has been achieved in \cite{Kozlowski:2010tv}. In the derivation a crucial role is played by the so-called Baxter Q-operator, which is currently the most powerful and universal tool for determining the spectrum of quantum integrable systems. Despite successful applications in many cases, see \cite{Frenkel:1308} (and also \cite{Nekrasov:2013xda}) for recent progress and relevant references, a complete theory of Q-operators is still to be established. In particular a systematic construction is currently not available in the case for which the representation of the relevant quantum group is not of highest-weight type, see \cite{MT} and references therein for some progress in this direction. Via the NS proposal, quantum integrable systems of this type can be systematically solved by gauge theory methods, offering an entirely new perspective on the structure underlying integrable models. TBA equations of a similar type also appeared in the context of wall crossing effects for the BPS spectrum in ${\cal N}=2$ gauge theories \cite{Gaiotto:2008cd, Alexandrov:2010pp}. It would be interesting to study the connection with our work and find the quantum integrable models corresponding to the TBA appearing there. Another connection is to the study of minimal surfaces in AdS \cite{Alday:2009dv,Alday:2010ku}. In this case, the area of such surfaces is equal, via the AdS/CFT correspondence, to the strong coupling limit of the logarithm of null Wilson-loop expectation values in ${\cal N}=4$ SYM. In \cite{Basso:2013vsa,Basso:2013aha}, it has been proposed that the exact expectation value can be written as a sum of multiple integrals involving some basic building blocks called pentagon transition functions. This form is structurally identical to the contour integral representation of the Nekrasov partition function. Moreover, the strong coupling limit correspond to the NS limit with $\frac{1}{\sqrt{\lambda}}\sim \epsilon_2$. Techniques of the type presented in this paper can be also used to show, directly from this representation of the Wilson loop, that in the strong coupling limit it satisfies the TBA equations derived from the analysis of classical strings in AdS \cite{BassoTalk,BSV}. It would be very interesting to study the possible connection with our work in this respect. \section*{Acknowledgements} We are greatly indebted to J\"{o}rg Teschner for suggesting this problem and valuable discussions at various stages of the project. We would like to thank Yasuyuki Hatsuda and Elli Pomoni for very useful discussions and collaboration on related topics. We would also like to thank Marcos Mari$\tilde{\rm n}$o for his interest in this work and encouragement. An important part of the work was done when G.Yang was a postdoc in the University of Hamburg where he was supported by the German Science Foundation (DFG) within the Collaborative Research Center 676 ``Particles, Strings and the Early Universe". C. Meneghelli is partially supported by a DFG grant in the framework of the SFB 676 ``Particles, Strings, and the Early Universe". G.Yang is supported by a DFG grant in the framework of the SFB 647 ``Space-Time-Matter".
{ "redpajama_set_name": "RedPajamaArXiv" }
7,161
You don't have to be a miracle worker although that would be great. You don't even need previous experience. Join us and you'll benefit from an industry leading training and induction programme as well as one of the sector's most comprehensive teams of in-house specialists. All you need is to share our values and be as passionate as we are about enabling the people we support to achieve the extraordinary each and every day. Working in partnership with the people we support, their family, our specialist support teams and external agencies you'll provide extraordinary support that improves the quality of their everyday life. Working with the manager, deputy manager and other team members you'll implement professionally designed and person centred support plans that maximise choice, independence and the achievement of individual goals, however small or large, each and every day. Your contribution to our mission, to encourage people with disabilities to challenge what they expect from life, from themselves and from wider society, will be invaluable. You'll need to be patient and calm under pressure as well as sociable, caring and supportive but above all you'll want to make a difference. You'll seek new ways of working by thinking outside of the box and being creative with solutions and will take personal responsibility for the quality of care and support delivered to the people we support, striving to always do what is right by them. So apart from transforming lives what's in it for you? You'll benefit from working with a team of passionate and skilled specialists, an outstanding induction and training programme with on-going development and the option to gain formal qualifications as well as the opportunity to progress your career in a sector that genuinely offers careers for life. In addition we also offer a range of benefits including flexible working options, long service and recognition awards, including increasing annual leave and bonus incentives and company pension and life assurance schemes. And we don't expect you to take your work home with you at the end of the day, in fact, it's illegal!
{ "redpajama_set_name": "RedPajamaC4" }
9,969
import _extends from "@babel/runtime/helpers/extends"; import _defineProperty from "@babel/runtime/helpers/defineProperty"; let _ = t => t, _t; function ownKeys(object, enumerableOnly) { var keys = Object.keys(object); if (Object.getOwnPropertySymbols) { var symbols = Object.getOwnPropertySymbols(object); enumerableOnly && (symbols = symbols.filter(function (sym) { return Object.getOwnPropertyDescriptor(object, sym).enumerable; })), keys.push.apply(keys, symbols); } return keys; } function _objectSpread(target) { for (var i = 1; i < arguments.length; i++) { var source = null != arguments[i] ? arguments[i] : {}; i % 2 ? ownKeys(Object(source), !0).forEach(function (key) { _defineProperty(target, key, source[key]); }) : Object.getOwnPropertyDescriptors ? Object.defineProperties(target, Object.getOwnPropertyDescriptors(source)) : ownKeys(Object(source)).forEach(function (key) { Object.defineProperty(target, key, Object.getOwnPropertyDescriptor(source, key)); }); } return target; } import { InputText, InputSearch } from '@looker/components'; import isArray from 'lodash/isArray'; import React, { useMemo } from 'react'; import styled from 'styled-components'; import { inputPlacementStyle, multiInputWidth } from '../../../../../../utils/filter_styles'; import { createOptions } from '../../../../../../utils/option_utils'; import { useOptionFiltering } from '../../../../../../utils/use_option_filtering'; import { usePlaceholder } from '../../../../../../utils/use_placeholder'; export const StringInputLayout = ({ className, onChange, onInputChange, isLoading, item, suggestions, enumerations, validationMessage, id, width: _width = multiInputWidth, height }) => { const options = useMemo(() => { return suggestions ? createOptions(suggestions) : enumerations || []; }, [suggestions, enumerations]); const { filteredOptions, noOptionsLabel, onFilter } = useOptionFiltering({ excludeValues: true, onInputChange, options, value: item.value }); const handleChange = newValue => { onFilter(newValue || ''); onChange === null || onChange === void 0 ? void 0 : onChange(item.id, { value: [newValue] }); }; const comboboxInputRef = React.useRef(null); const handleKeyDown = event => { if (event.key === 'Enter') { var _comboboxInputRef$cur; event.preventDefault(); comboboxInputRef === null || comboboxInputRef === void 0 ? void 0 : (_comboboxInputRef$cur = comboboxInputRef.current) === null || _comboboxInputRef$cur === void 0 ? void 0 : _comboboxInputRef$cur.click(); } }; const placeholderProps = usePlaceholder(item.value, validationMessage); const commonProps = _objectSpread(_objectSpread({}, placeholderProps), {}, { className, height, id, maxHeight: 145, onChange: handleChange, noErrorIcon: true, validationType: validationMessage === null || validationMessage === void 0 ? void 0 : validationMessage.type, width: _width }); const value = isArray(item.value) && item.value.length ? item.value[0] : item.value; return React.createElement(InputSearch, _extends({}, commonProps, { options: filteredOptions, noOptionsLabel: noOptionsLabel, indicator: false, isLoading: isLoading, value: value, onKeyDown: handleKeyDown, isClearable: true, ref: comboboxInputRef, hideSearchIcon: true, openOnClick: true })); }; export const StringInput = styled(StringInputLayout).withConfig({ displayName: "StringInput", componentId: "sc-feqdof-0" })(_t || (_t = _` ${0} ${0} { ${0} } `), inputPlacementStyle, InputText, inputPlacementStyle); //# sourceMappingURL=StringInput.js.map
{ "redpajama_set_name": "RedPajamaGithub" }
6,181
Prisilno polnjeni motor () je batni motor (bencinski ali dizelski) pri katerem se s pomočjo kompresorja (polnilnika) poveča tlak in temperaturo zraka pred vstopom v valje motorja. S tem povečamo gostoto zraka, vbrizgamo več goriva in motor lahko razvija večjo moč. Motorji, ki ne uporabljajo prisilnega polnjenja se imenujejo "atmosfersko polnjeni" ali "neprisilno" polnjeni motorji. Obstajajo trije tipi polnilnikov: Mehansko gnani polnilnik () - prvi polnilniki so bili tega tipa, ročična gred preko jermena poganja kompresor. Najde se jih na letalih in na nekaterih ameriških avtomobilih Turbopolnilnik ali turbinsko gnani polnilnik, ali samo Turbo (): uporabljajo energijo izpušnih plinov, zato imajo večji izkoristek kot mehansko gnani. Na dizelskih motorjih se za prisilno polnjenje večinoma turbopolnilnike Električno gnani polnilnik: alternator poizvaja električno energijo za pogon električnega kompresorja, so redki in imajo najmanjši izkoristek in sposobnosti Z uporabo polnilnikov se zaradi stiskanja poveča tudi temperatura, kar ni zaželeno, zato se v nekaterih primerih uporablja vmesni hladilnik (intercooler), ki ohladi zrak pred vstopom v valje. Nekateri motorji uporabljajo tudi vbrizg vode za isti namen. Polnilniki so še posebej uporabni na letalskih motorjih. Ko se letalo dviga, se zmanjšuje gostota zraka, na vsakih 100 metrov se moč motorja zmanjša za približno en procent. S polnilnikom lahko motor razvija veliko moč tudi na veliki višini. Na veliki višini je zrak redek, kar pomeni manjši zračni upor in letalo lahko leti precej hitreje. Iz turbopolnilnika so se kasneje razvili turbinski letalski motorji kot so turboreaktivni (in kasneje turboventilatorski), turbopropelerski in turbgredni motorji. Glej tudi Atmosfersko polnjeni motor Turbopolnilnik Turbokompaundni motor Sklici Batni motorji Pogonski stroji
{ "redpajama_set_name": "RedPajamaWikipedia" }
4,091
Q: Entity Framework Core self referencing table I have a class called Item which references the next item and the previous item. public class Item { private Item() { } public Item(string itemName) { ItemId = Guid.NewGuid(); ItemName = itemName; } public Guid ItemId { get; set; } public string ItemName { get; set; } public Guid NextItemId { get; set; } public virtual Item NextItem { get; set; } public Guid PreviousItemId { get; set; } public virtual Item PreviousItem { get; set; } public Guid GroupId { get; set; } public virtual Group Group { get; set; } } I have another table called Group it is for grouping it items. public class Group { private Group() { } public Group(string groupName) { GroupId = Guid.NewGuid(); GroupName = groupName; GroupItems = new List<Item>(); } public void AddGroupItem(Item item) { if (Items.Count == 0) { Items.Add(item); } else { item.PreviousItem = Items.Last(); item.PreviousItemId = Items.Last().ItemId; Items.Last().NextItem = item; Items.Last().NextItemId = item.ItemId; Items.Add(item); } } public Guid GroupId { get; set; } public string GroupName { get; set; } public virtual IList<GroupItem> GroupItems { get; set; } } Here's how I create and save items and their group. Group group1 = new Group("first group"); Item item1 = new Item("item 1"); Item item2 = new Item("item 2"); Item item3 = new Item("item 3"); group1.AddItem(item1); group1.AddItem(item2); group1.AddItem(item3); _context.Add(group1); _context.SaveChanges(); How do I write the OnModelCreating to handle the two references to the same table. A: You can do it in next way. First of all you should add two new properties to you model public virtual List<Item> ParentNextItems { get; set; } and public virtual List<Item> ParentPreviousItems { get; set; }. So your model will be something like this public class Item { private Item() { } public Item(string itemName) { ItemId = Guid.NewGuid(); ItemName = itemName; } public Guid ItemId { get; set; } public string ItemName { get; set; } public Guid? NextItemId { get; set; } public virtual Item NextItem { get; set; } public virtual List<Item> ParentNextItems { get; set; } public Guid? PreviousItemId { get; set; } public virtual Item PreviousItem { get; set; } public virtual List<Item> ParentPreviousItems { get; set; } } And than you can configure it in next way protected override void OnModelCreating(ModelBuilder modelBuilder) { modelBuilder.Entity<Item>() .HasKey(x => x.ItemId); modelBuilder.Entity<Item>() .HasOne(x => x.NextItem).WithMany(x => x.ParentNextItems).HasForeignKey(x => x.NextItemId) .Metadata.DeleteBehavior = DeleteBehavior.Restrict; modelBuilder.Entity<Item>() .HasOne(x => x.PreviousItem).WithMany(x => x.ParentPreviousItems).HasForeignKey(x => x.PreviousItemId) .Metadata.DeleteBehavior = DeleteBehavior.Restrict; base.OnModelCreating(modelBuilder); } That's all. But if you want to achieve the same with attribute configuration, than you can skip void OnModelCreating(ModelBuilder modelBuilder) changes and just write next model: public class Item { private Item() { } public Item(string itemName) { ItemId = Guid.NewGuid(); ItemName = itemName; } [Key] public Guid ItemId { get; set; } public string ItemName { get; set; } public Guid? NextItemId { get; set; } [ForeignKey(nameof(NextItemId))] [InverseProperty(nameof(ParentNextItems))] public virtual Item NextItem { get; set; } [ForeignKey(nameof(NextItemId))] public virtual List<Item> ParentNextItems { get; set; } public Guid? PreviousItemId { get; set; } [ForeignKey(nameof(PreviousItemId))] [InverseProperty(nameof(ParentPreviousItems))] public virtual Item PreviousItem { get; set; } [ForeignKey(nameof(PreviousItemId))] public virtual List<Item> ParentPreviousItems { get; set; } } Update Also you should make PreviousItemId, NextItemId optional (Guid?), I did corresponding changes in my answer. And as far as I know, you just can't do it in one .SaveChanges() trip. When you create your second item you should already have first item saved into database. (anyway this is subject for another question) But anyway if you modify you code to something like that Group group1 = new Group("first group"); _context.Add(group1); Item item1 = new Item("item 1"); group1.AddItem(item1); _context.SaveChanges(); Item item2 = new Item("item 2"); group1.AddItem(item2); _context.SaveChanges(); Item item3 = new Item("item 3"); group1.AddItem(item3); _context.SaveChanges(); You can do it in one transaction A: You don't need to have Collection, since as far I understand here you aim to create some type of Two Way LinkedList. A was able to fix it by adding .HasOne() to the main model. modelBuilder.Entity<Item>().HasOne(x => x.NextItem); modelBuilder.Entity<Item>().HasOne(x => x.PreviousItem); A: public int? DeaultNextStateId { get; set; } public State DeaultNextState { get; set; } public virtual ICollection<State> DeaultNextStates { get; set; } public int? OkNextStateId { get; set; } public State OkNextState { get; set; } public virtual ICollection<State> OkNextStates { get; set; } public int? NotOkNextStateId { get; set; } public State NotOkNextState { get; set; } public virtual ICollection<State> NotOkNextStates { get; set; } modelBuilder.Entity<Models.Workflow.State>() .HasKey(x => x.Id); modelBuilder.Entity<Models.Workflow.State>() .HasOne(x => x.DeaultNextState).WithMany(x => x.DeaultNextStates).HasForeignKey(x => x.DeaultNextStateId) .Metadata.DeleteBehavior = DeleteBehavior.Restrict; modelBuilder.Entity<Models.Workflow.State>() .HasOne(x => x.OkNextState).WithMany(x => x.OkNextStates).HasForeignKey(x => x.OkNextStateId) .Metadata.DeleteBehavior = DeleteBehavior.Restrict; modelBuilder.Entity<Models.Workflow.State>() .HasOne(x => x.NotOkNextState).WithMany(x => x.NotOkNextStates).HasForeignKey(x => x.NotOkNextStateId) .Metadata.DeleteBehavior = DeleteBehavior.Restrict;
{ "redpajama_set_name": "RedPajamaStackExchange" }
5,552
\section{Introduction} The frequency of brown dwarfs (BDs) and planetary-mass companions around intermediate-mass main sequence (MS) stars is uncertain. BDs can be formed by several mechanisms \citep[e.g.][]{Padoan2004,Stamatellos2009}, but the expected substellar fractions for B-F type primaries are uncertain. In the case of giant planets formed in the disks of primary stars, some works predict a higher frequency around AB-type stars than in solar-type stars \citep[e.g.][]{Kennedy2008}. However, \citet{Kornet2006} show an opposite result since they conclude that the percentage of stars with giant planets decreases with increasing stellar masses from 0.5 to 4\,M$_{\odot}$. To shed light on this issue, different observational programs have been focused on deriving the frequency of BDs and planetary-mass objects around intermediate-mass stars. As a result, planetary mass companions have been recently detected around three A-type stars through adaptive optics (AO) assisted observations \citep[][]{Marois2008, Kalas2008, Lagrange2010}. Radial velocity (RV) studies, which are sensitive to short period companions, have also reported the presence of substellar objects around several A-F type MS stars \citep{Galland2005,Galland2006,Guenther2009}, with minimum masses ($M. sin i$ ) in the planetary mass regime, that is, they could also be BDs. Transit programs have also detected planetary mass companions around several F-type stars \citep[e.g.][]{Bakos2007,JohnsKrull2008,Joshi2009,Hellier2009}, and one around an A5 star \citep{Christian2006,Cameron2010}. In the case of BDs, direct imaging surveys have allowed to study the fraction of wide substellar companions around intermediate-mass stars. As an example, \citet{Kouwenhoven2005,Kouwenhoven2007} studied the late-B and A-type star population from the Sco OB2 association. Although sensitive to substellar companions, they reported the detection of only two BDs companion candidates. They concluded that the dearth of BD companions to intermediate-mass stars is consistent with the extrapolation of the stellar companion mass distribution into the BD regime (assuming they formed like stars). Recently, \citet{Ehrenreich2010} conducted an AO survey to detect substellar companions in wide orbits around a volume-limited sample of 38 A- and F-type field stars previously observed with RV techniques. They did not report any new BD companion. In fact, up to now there is only one BD companion to an intermediate-mass star, HR7329\,B, confirmed by direct imaging and near-IR spectroscopy \citep[][]{Lowrance2000,Guenther2001}. RV studies have also detected BD companions to A-F type MS stars \citep{Galland2006,Hartmann2010}, and transit observations have reported the presence of a BD around an F-type star, CoRoT-3\,b \citep{Deleuil2008}. As it follows, and despite the efforts, the occurrence of planetary-mass objects and BDs around intermediate-mass stars is still unknown and deserves additional observations. In 2004, we started a project aiming at deriving the binary fraction and properties among a large volume-limited sample of intermediate-mass stars in the field \citep[hereafter, Multi-NETS Project,][]{Ivanov2006}. Thanks to the use of deep AO near infrared imaging with Naos-Conica \citep[NACO][]{Lenzen2003} at the Very Large Telescope (VLT), we have been able to extend our study to substellar companions that are important to understand the formation of low mass ratio binaries. In the course of our survey, we detected a faint, visual companion to the star HR\,6037. In this letter, we report the discovery and the co-moving confirmation of the companion to HR\,6037 based on NACO astrometric observations obtained at two different epochs. We also present ISAAC near-infrared spectroscopic data that confirms, together with the photometry, that this companion is likely to be a new and rare substellar companion to the intermediate-mass star HR\,6037. \section{HR\,6037 stellar properties} \begin{figure} \centering \includegraphics[width=0.8\columnwidth]{HRdiagram2.eps} \caption{HR Diagram with the location of HR\,6037. We have overplotted the \citet{Marigo2008} isochrones for five different ages between 100\,Myr and 500 Myr (from the left to the right). We estimate an age of $\sim$300$\pm$100 Myr for the primary. } \label{hrdiag}% \end{figure} HR\,6037 is a main sequence A-type star classified as 'variable' by \citet{Samus2009} although its type of variability is uncertain. Its proper motion and parallax, according to Hipparcos \citep{Perryman1997}, are $\mu_{\alpha}$=-44.74$\pm$0.56 and $\mu_{\delta}$=-84.65$\pm$0.43\,mas/yr, and 18.13$\pm$0.69\,mas, respectively. The latter value translates into a distance of 55$\pm$2\,pc. We have derived the physical properties of HR\,6037 by analyzing high resolution (R=80.000) optical archival spectroscopy obtained with VLT/UVES (Program ID 266.D-5655(A)). The spectrum was obtained integrating a total of 70 seconds, and was centered at 5800\AA. The primary is an A6V, as derived from spectral synthesis using SYNTHE \citep{Kuruc1993}. Our analysis also yields [Fe/H]=0.00$\pm$0.05, namely solar metallicity, and provides values for the effective temperature and surface gravity (see Table~\ref{properties}). Using the Hipparcos parallax we derive $M_{V}=2.24$, or $log(\frac{L}{L_{\odot}} $)=1.02$^{0.038}_{0.02}$, assuming no extinction. Stromgren H$_{\beta}=2.884$ and b-y$=0.069$ photometry \citep{HauckMermilliod1998} confirms HR\,6037\,A as an A6 dwarf with temperature and gravity in agreement with values derived from the UVES spectrum. We have plotted the object on a Hertzsrpung-Russel diagram (see Fig.~\ref{hrdiag}) and compared it with the Padova evolutionary tracks for solar metallicity \citep{Marigo2008}. We have used isochrones from 100 to 500\,Myr, which is a typical age range for a star of this spectral type. We estimate an age of 300$\pm$100\,Myr and a mass of 1.8$\pm$0.2\,M$_\odot$. Finally, we note that the object was included in a RV survey to look for very close BDs and planetary-mass companions, showing no significant RV variation \citep{Lagrange2009}. \section{Observations and data reduction} \subsection{NACO deep imaging} HR\,6037 was observed in service mode with NACO, the adaptive optics facility at the VLT on June 30, 2004 and June 9, 2006. We used the visible wavefront sensor with the primary as a reference star. We observed in the Ks-band filter with the S27 objective (field of view of $27\arcsec\times27$\arcsec) in 'Autojitter Mode', dithering within a box of 12\arcsec~width. The total on-source exposure time was $\sim$13 minutes. The average coherence time and optical seeing were 1.5\,ms and 1\farcs0, and 1.2\,ms and 1\farcs7, during the first and second epoch, respectively. The data were reduced using {\em Eclipse} \citep{Devillard1997}, and following the standard procedure: dark subtraction, flat-field division, sky subtraction, alignment and stacking. The final image from 2006 is displayed in Fig.~\ref{nacoima}. Apart from the bright primary, we detect a visual companion at a projected separation of $\sim$6\farcs66 and position angle of $\sim$294 degrees. \begin{table}[t] \caption{Physical properties of HR\,6037\,A\&B derived in this work} \label{stellar} \advance\tabcolsep by -2.8pt \begin{tabular}{cclllc} \noalign{\smallskip} \hline \noalign{\smallskip} Name & Sp. Type & $T_{\rm eff}$ & log\,$g_{s}$ & Mass & [Fe/H] \\ & & [K] & [cm/s$^2$] & [$M_{\odot}$]& \\%& [Myr] \\ \noalign{\smallskip} \hline \noalign{\smallskip} HR6037A & A6 & 8120$\pm$100 & 4.2$\pm$0.1 & 1.8$\pm$0.2 & 0.00$\pm$0.05 \\ HR6037B & M9 & 2330$\pm$200 & 5.1$\pm$0.2 & 0.06$\pm$0.02 & -- \\ \hline\hline \end{tabular}\label{properties} \end{table} \begin{figure}[t] \centering \includegraphics[width=0.7\columnwidth]{hr6037b.ps} \caption{NACO/VLT image of HR\,6037. The co-moving companion is encircled.} \label{nacoima}% \end{figure} In order to measure precise separations and position angles, we derived the plate scale and orientation of the detector, CONICA, using archival observations of the astrometric calibrator IDS\,21506-5133 \citep{vanDessel1993} obtained on June 2004 and August 2006. The values for the two campaigns are respectively $27.01\pm0.05$~mas/pix and $27.02\pm0.05$~mas/pix for the plate scale, and $0.0\pm0.2$~deg and $-0.1\pm0.2$~deg for the True North orientation. \subsection{ISAAC near-infrared spectroscopy} Near-infrared spectra of the HR\,6037\,B were obtained in service mode with ISAAC/VLT \citep{Moorwood1998} on 2010-06-06/07 in the $J$ and $H$ atmospheric windows, and on 2010-06-11/12 in the $K$ window, in the ``classical'' nodding-along-the-slit observing strategy. We used the low-resolution mode and the 0.6\arcsec ~wide slit, delivering a spectral resolution of R$\sim$800. We collected six exposures for $J$ and $H$, and twelve for $K$ but one $J$ spectrum was discarded because of a low signal. The total integration times were 1115, 2232, and 4464\,sec, respectively for $J$, $H$, and $K$. The seeing was better than 1\arcsec~during both nights. The sky was clear on 2010-06-06/07, and thin clouds were present on 2010-06-11/12. B-type telluric standards were observed back-to-back with the science targets with the same instrument setup. One of them showed a strong Bracket\,$\gamma$ emission line, which was fitted with a Gaussian and subtracted from the spectrum before applying the telluric correction. The data were reduced using IRAF\footnote{IRAF is distributed by the National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy (AURA) under cooperative agreement with the National Science Foundation.} and following standard steps: flat field division, sky emission removal by subtracting images from corresponding nodding pairs, and extraction and combination of the individual spectra into the final spectrum. The wavelength calibration was performed using arcs. The telluric absorption was removed by divided the target spectra by the telluric standards, and multiplied them by the corresponding spectra from the library of \citet{Pickles1998}. Some of the spectra from this library are featureless models, so artificial emission lines remained in the final product. To remove them, we went back to the telluric spectra and subtracted Gaussian fits to their intrinsic stellar features -- mainly Hydrogen recombination lines. This, together with the Bracket\,$\gamma$ emission mentioned above implies some uncertainty in the spectral regions around the strong Hydrogen lines. \section{Results} \subsection{HR\,6037\,B, a co-moving companion} Fig.~\ref{ppm} shows the difference in right ascension (RA) and declination (DEC) of HR\,6037 and its companion candidate as measured in 2004 and 2006. We have also overplotted the expected difference in RA and DEC of a background object taking into account the proper motion and parallax of the primary. As seen, the companion shows RA and DEC differences consistent with a co-moving object. In fact, the difference in separation and position angle between the two epochs are consistent with a bound companion within the errors (see Table~\ref{binary}). \begin{figure}[t] \centering \includegraphics[width=0.9\columnwidth]{radec_hr6037.ps} \caption{Analysis of the two NACO/VLT epochs of HR\,6037. The axes show the difference in right ascension and declination of the binary members in the two epochs. The solid black and grey circles represent the position of the companion candidate in 2004 and 2006, respectively. The open grey circle represents the expected position of HR\,6037\,B if it were a background object. The data is consistent with HR\,6037\,B being a co-moving companion.} \label{ppm}% \end{figure} \subsection{Spectral characterization of HR\,6037\,B} \begin{figure}[t] \includegraphics[width=0.5\textwidth]{spectra.ps} \caption{Comparison of the $J\&H$ band ISAAC spectra of HR\,6037\,B companion (red) to the template spectra of the young 8\,Myr-old dwarf 2M1207\,A (M8.5) and the field dwarf (M9V). The spectrum of the very late type giant (IO Virginis) is also reported for comparison. All spectra have been normalized to 1.23~$\mu$m and 1.65~$\mu$m and offsetted.} \label{spectro} \end{figure} The ISAAC $JHK$ spectra of HR\,6037\,B were first compared to libraries of template spectra of field dwarfs \citep{Cushing2005,Rayner2009} and moderately young dwarfs from Upper Scorpius, TW Hydrae and $\beta$ Pictoris associations \citep[][]{Allers2009, Rice2010}. The $K$-band spectrum of HR\,6037\,B is much bluer than all M and early-L type dwarfs. This is probably due to a problem of flux loss during the observation. Therefore, only the $J$ and $H$-band spectra were considered for the spectral classification based on the continuum comparison with libraries of field and young dwarfs. The best matches are displayed in Fig.~\ref{spectro}. The $J$ and $H$ continuum of HR\,6037\,B is well reproduced by the spectrum of the young M8.5 dwarf 2M1207\,A from the TW\,Hydrae association (8~Myr) and the M9V field dwarf \citep{Rayner2009}, so we estimate an M9$\pm$1 spectral type. Careful identification of the lines over the $JHK$ spectral range shows the presence of broad molecular absorptions of H$_2$0 (longward 1.33 and 1.6$\mu$m), FeH (at 1.194, 1.222, 1.239, 1.583-1.591 and 1.625~$\mu$m) as well as CO overtones longward 2.29~$\mu$m all typical of late-M dwarfs. There is also the possible presence of VO absorptions from 1.17 to 1.20~$\mu$m. In the $J$-band, the atomic line doublets of Na I and K I at 1.138, 1.169, 1.177, 1.243, 1.253~$\mu$m are well detected.We also detect the K I atomic line at 1.517~$\mu$m. Their strengths are intermediate between spectra of 10 Myr-old dwarfs, and those of field dwarfs with identical spectral types (see Fig.~\ref{spectro}). This finding corroborates the age estimate of HR\,6037\,A and B of a few tens to hundreds Myr (if both component are coeval). \begin{table}[!] \caption{Binary parameters for the two epoch observations.}\label{binary} \begin{tabular}{llll} \hline\hline\noalign{\smallskip} Date & Separation & PA & $\Delta\,K_s$ \\ & (arcsec) & (deg) & (mag) \\ \noalign{\smallskip}\hline\noalign{\smallskip} 2004/06/30 & $6.66 \pm 0.01$ & $293.96\pm0.1$ & $8.2\pm0.1$ \\ 2006/06/09 & $6.67\pm 0.03$ & $293.98\pm0.2$ & $8.7\pm0.2$ \\ \noalign{\smallskip}\hline\noalign{\smallskip} \end{tabular} \end{table} \subsection{HR\,6037\,B physical properties} The difference in $K_s$ magnitude between HR\,6037\,A\&B in the NACO images was derived using standard packages for aperture photometry within IRAF, and is provided in Table~\ref{binary}. Since the 2MASS $K_{s}$ value of the primary is $5.66\pm0.02$ \citep{2MASS}, we estimate an average $K_{s}$=$14.1\pm0.3$\,mag for the secondary, which translates into $M_{K_s}=10.4\pm0.3$\,mag for a distance of $55\pm2$\,pc. We have compared this value with evolutionary tracks by \citet{Baraffe2002}, assuming the age estimate derived from the UVES spectrum of the primary. According to DUSTY evolutionary tracks \citep{Baraffe2002}, a $300\pm100$\,Myr object with $M_{K_s}$ of $10.4\pm0.3$\,mag, corresponds to a $62\pm20$~M$_{Jup}$ BD with $T_{\rm{eff}} = 2330\pm200$\,K, and log\,$g= 5.1\pm0.2$. \subsection{HR\,6037 A\&B: main properties of the binary system} The mass ratio of HR\,6037 A\&B is $q$=0.034. This mass ratio is not common for binaries with intermediate-mass stars as primaries. The projected separation of the binary components, for a distance of 55\,pc, is $\sim$366\,AU. Using Kepler's third law, we derive an orbital period of $\sim$5000\,yr. Even if imaging surveys are sensitive to these long period, small mass ratio binaries, they are uncommon \citep[e.g.][]{Kouwenhoven2007,Ehrenreich2010}. Hence, we can conclude that HR\,6037\,A\&B is a extremely rare binary system. \section{Conclusions} We report the detection of a BD companion to the 300\,Myr old star HR\,6037. Our main results can be summarized as follows: \begin{enumerate} \item HR\,6037 is a binary system with a separation of 6\farcs66 and position angle of $293.9\pm0.1$ degrees. Two epoch observations confirm that HR\,6037\,B is a co-moving companion. \item Near-IR spectroscopy reveals a spectral type of M9$\pm$1 for HR\,6037\,B by comparison of the $J$ and $H$ band continuum to templates. The strength of the gravity-sensitive features are consistent with a dwarf intermediate between a low-gravity young dwarf and high-gravity field dwarf of similar spectral type. This result is consistent with the age derived for the primary, 300\,Myr, i.e. both objects appear to be coeval. \item Evolutionary tracks predict a mass of 62$\pm$20\,M$_{Jup}$, an effective temperature of $T_{\rm{eff}} = 2330\pm200$\,K, and a surface gravity of log\,$g= 5.1\pm0.2$. \end{enumerate} To our knowledge, HR\,6037\,B is the second BD companion confirmed to be bound to an intermediate-mass star by two epoch observations and spectroscopy. Its small mass ratio and long orbital period makes it a rare and uncommon binary system. \begin{acknowledgements} This research has been funded by Spanish grants MEC/ESP2007-65475-C02-02, {\mbox MEC/Consolider-CSD2006-0070, and CAM/PRICIT-1496}. We are grateful to K. Allers and E. Rice for providing their spectra. NH gratefully acknowledges support from the ESO's Visiting Scientist Programme. GCa thanks S. Villanova for helping with the UVES data reduction. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France. \end{acknowledgements}. \bibliographystyle{bibtex/aa}
{ "redpajama_set_name": "RedPajamaArXiv" }
3,021
Az azeri manat Azerbajdzsán jelenlegi hivatalos pénzneme. A manat pénznév az orosz монета (jelentése: érme) szóból származik. Annak idején manat volt a szovjet rubel azeri és türkmén megnevezése is. Története Az Azerbajdzsáni Demokratikus Köztársaság és az Azerbajdzsáni Szovjet Szocialista Köztársaság 1919–1923 között saját pénzt hozott forgalomba. Ezt a pénzt azeriül manatnak (منات), oroszul rubelnek (рубль) hívták, és mindkét nyelven fel volt tüntetve a papírpénzeken. A második manatot 1992-ben vezették be, mely 1 manat = 10 rubel arányban váltotta fel a szovjet rubelt. A valuta az évtized folyamán folyamatosan leértékelődött, de 2002–2005 között már viszonylag szilárd volt a dollárral szembeni árfolyama, sőt, 2005 után még erősödött is – mindez a beáramló olajjövedelmeknek volt köszönhető. Ekkorra azonban a 100 manat alatti bankjegyek – az érmékkel együtt – gyakorlatilag eltűntek a forgalomból, és szükségessé vált a valutareform. 2006. január 1-jén vezették be a jelenleg is használatos manatot 1 új manat = 5000 régi manat átszámítási aránnyal. A bankjegyek meglehetősen hasonlítanak az euróbankjegyekre – tervezőjük az az osztrák Robert Kalina, aki az euróbankjegyeket is tervezte. 2015. december 21-én összeomlott a manat árfolyama az olajválság miatt. Az amerikai dollár ezen a napon 1,55 manat volt, míg januárban csak 0,78 manat. Érmék Bankjegyek 2006-os sorozat 2018. május 24-én bocsátották ki a 200 manatos bankjegyet. 2019-től a bank nevének megváltoztatása miatt (Azərbaycan Mərkəzi Bankı) új felirattal jelennek meg a bankjegyek. 2019. március 18-án az azeri jegybank egy biztonságtechnikai szempontból jelentősen korszerűsített, a 200 manatoshoz hasonlóan hologrammal, SPARK ábrával, irizáló nyomattal ellátott 10 manatost bácsátott ki. A bankjegy külső megjelenése alapvetően nem változott. 2019 áprilisában bejelentették a központi banknál, hogy vizsgálják annak lehetőségét, hogy áttérnek a polimer alapú bankjegyekre. 2021-es sorozat Jegyzetek További információk Ázsiai pénznemek Manat Európai pénznemek
{ "redpajama_set_name": "RedPajamaWikipedia" }
6,922
{"url":"http:\/\/www.absoluteastronomy.com\/topics\/Pierre_de_Fermat","text":"Pierre de Fermat\n\n# Pierre de Fermat\n\nOverview\nPierre de Fermat was a French\nFrench people\nThe French are a nation that share a common French culture and speak the French language as a mother tongue. Historically, the French population are descended from peoples of Celtic, Latin and Germanic origin, and are today a mixture of several ethnic groups...\n\nlawyer at the Parlement\nParlement\nParlements were regional legislative bodies in Ancien R\u00e9gime France.The political institutions of the Parlement in Ancien R\u00e9gime France developed out of the previous council of the king, the Conseil du roi or curia regis, and consequently had ancient and customary rights of consultation and...\n\nof Toulouse\nToulouse\nToulouse is a city in the Haute-Garonne department in southwestern FranceIt lies on the banks of the River Garonne, 590\u00a0km away from Paris and half-way between the Atlantic Ocean and the Mediterranean Sea...\n\n, France\nFrance\nThe French Republic , The French Republic , The French Republic , (commonly known as France , is a unitary semi-presidential republic in Western Europe with several overseas territories and islands located on other continents and in the Indian, Pacific, and Atlantic oceans. Metropolitan France...\n\n, and an amateur mathematician\u00a0who is given credit for early developments that led to infinitesimal calculus\nInfinitesimal calculus\nInfinitesimal calculus is the part of mathematics concerned with finding slope of curves, areas under curves, minima and maxima, and other geometric and analytic problems. It was independently developed by Gottfried Leibniz and Isaac Newton starting in the 1660s...\n\nIn the history of infinitesimal calculus, adequality is a technique developed by Pierre de Fermat. Fermat said he borrowed the term from Diophantus. Adequality was a technique first used to find maxima for functions and then adapted to find tangent lines to curves...\n\n. In particular, he is recognized for his discovery of an original method of finding the greatest and the smallest ordinate\nOrdinate\nIn mathematics, ordinate refers to that element of an ordered pair which is plotted on the vertical axis of a two-dimensional Cartesian coordinate system, as opposed to the abscissa...\n\ns of curved lines, which is analogous to that of the then unknown differential calculus\nDifferential calculus\nIn mathematics, differential calculus is a subfield of calculus concerned with the study of the rates at which quantities change. It is one of the two traditional divisions of calculus, the other being integral calculus....\n\n, and his research into number theory\nNumber theory\nNumber theory is a branch of pure mathematics devoted primarily to the study of the integers. Number theorists study prime numbers as well...\n\n. He made notable contributions to analytic geometry\nAnalytic geometry\nAnalytic geometry, or analytical geometry has two different meanings in mathematics. The modern and advanced meaning refers to the geometry of analytic varieties...\n\n, probability\nProbability\nProbability is ordinarily used to describe an attitude of mind towards some proposition of whose truth we arenot certain. The proposition of interest is usually of the form \"Will a specific event occur?\" The attitude of mind is of the form \"How certain are we that the event will occur?\" The...\n\n, and optics\nOptics\nOptics is the branch of physics which involves the behavior and properties of light, including its interactions with matter and the construction of instruments that use or detect it. Optics usually describes the behavior of visible, ultraviolet, and infrared light...\n\n. He is best known for Fermat's Last Theorem\nFermat's Last Theorem\nIn number theory, Fermat's Last Theorem states that no three positive integers a, b, and c can satisfy the equation an\u00a0+\u00a0bn\u00a0=\u00a0cn for any integer value of n greater than two....\n\n, which he described in a note at the margin of a copy of Diophantus\nDiophantus\nDiophantus of Alexandria , sometimes called \"the father of algebra\", was an Alexandrian Greek mathematician and the author of a series of books called Arithmetica. These texts deal with solving algebraic equations, many of which are now lost...\n\n' Arithmetica\nArithmetica\nArithmetica is an ancient Greek text on mathematics written by the mathematician Diophantus in the 3rd century AD. It is a collection of 130 algebraic problems giving numerical solutions of determinate equations and indeterminate equations.Equations in the book are called Diophantine equations...\n\n.\nDiscussion\n Ask a question about 'Pierre de Fermat' Start a new discussion about 'Pierre de Fermat' Answer questions from other users Full Discussion Forum\n\nRecent Discussions\nQuotations\n\na^p \\equiv a \\pmod\\,\\!\n\nAlso known as Fermat's little theorem|Fermat's little theorem, the formula above is true for any prime p where a\n\n$a^p \\equiv a \\pmod\\,\\!$\n\nAlso known as Fermat's little theorem|Fermat's little theorem, the formula above is true for any prime $p$ where $a$\n\n$a^p \\equiv a \\pmod\\,\\!$\n\nAlso known as Fermat's little theorem|Fermat's little theorem, the formula above is true for any prime $p$ where\nEncyclopedia\nPierre de Fermat was a French\nFrench people\nThe French are a nation that share a common French culture and speak the French language as a mother tongue. Historically, the French population are descended from peoples of Celtic, Latin and Germanic origin, and are today a mixture of several ethnic groups...\n\nlawyer at the Parlement\nParlement\nParlements were regional legislative bodies in Ancien R\u00e9gime France.The political institutions of the Parlement in Ancien R\u00e9gime France developed out of the previous council of the king, the Conseil du roi or curia regis, and consequently had ancient and customary rights of consultation and...\n\nof Toulouse\nToulouse\nToulouse is a city in the Haute-Garonne department in southwestern FranceIt lies on the banks of the River Garonne, 590\u00a0km away from Paris and half-way between the Atlantic Ocean and the Mediterranean Sea...\n\n, France\nFrance\nThe French Republic , The French Republic , The French Republic , (commonly known as France , is a unitary semi-presidential republic in Western Europe with several overseas territories and islands located on other continents and in the Indian, Pacific, and Atlantic oceans. Metropolitan France...\n\n, and an amateur mathematician\u00a0who is given credit for early developments that led to infinitesimal calculus\nInfinitesimal calculus\nInfinitesimal calculus is the part of mathematics concerned with finding slope of curves, areas under curves, minima and maxima, and other geometric and analytic problems. It was independently developed by Gottfried Leibniz and Isaac Newton starting in the 1660s...\n\nIn the history of infinitesimal calculus, adequality is a technique developed by Pierre de Fermat. Fermat said he borrowed the term from Diophantus. Adequality was a technique first used to find maxima for functions and then adapted to find tangent lines to curves...\n\n. In particular, he is recognized for his discovery of an original method of finding the greatest and the smallest ordinate\nOrdinate\nIn mathematics, ordinate refers to that element of an ordered pair which is plotted on the vertical axis of a two-dimensional Cartesian coordinate system, as opposed to the abscissa...\n\ns of curved lines, which is analogous to that of the then unknown differential calculus\nDifferential calculus\nIn mathematics, differential calculus is a subfield of calculus concerned with the study of the rates at which quantities change. It is one of the two traditional divisions of calculus, the other being integral calculus....\n\n, and his research into number theory\nNumber theory\nNumber theory is a branch of pure mathematics devoted primarily to the study of the integers. Number theorists study prime numbers as well...\n\n. He made notable contributions to analytic geometry\nAnalytic geometry\nAnalytic geometry, or analytical geometry has two different meanings in mathematics. The modern and advanced meaning refers to the geometry of analytic varieties...\n\n, probability\nProbability\nProbability is ordinarily used to describe an attitude of mind towards some proposition of whose truth we arenot certain. The proposition of interest is usually of the form \"Will a specific event occur?\" The attitude of mind is of the form \"How certain are we that the event will occur?\" The...\n\n, and optics\nOptics\nOptics is the branch of physics which involves the behavior and properties of light, including its interactions with matter and the construction of instruments that use or detect it. Optics usually describes the behavior of visible, ultraviolet, and infrared light...\n\n. He is best known for Fermat's Last Theorem\nFermat's Last Theorem\nIn number theory, Fermat's Last Theorem states that no three positive integers a, b, and c can satisfy the equation an\u00a0+\u00a0bn\u00a0=\u00a0cn for any integer value of n greater than two....\n\n, which he described in a note at the margin of a copy of Diophantus\nDiophantus\nDiophantus of Alexandria , sometimes called \"the father of algebra\", was an Alexandrian Greek mathematician and the author of a series of books called Arithmetica. These texts deal with solving algebraic equations, many of which are now lost...\n\n' Arithmetica\nArithmetica\nArithmetica is an ancient Greek text on mathematics written by the mathematician Diophantus in the 3rd century AD. It is a collection of 130 algebraic problems giving numerical solutions of determinate equations and indeterminate equations.Equations in the book are called Diophantine equations...\n\n.\n\n## Life and work\n\nFermat was born in Beaumont-de-Lomagne\nBeaumont-de-Lomagne\nBeaumont-de-Lomagne is a commune in the Tarn-et-Garonne department in the Midi-Pyr\u00e9n\u00e9es region in southern France.-History:Beaumont-de-Lomagne, bastide, was founded in 1276 following the act of coregency between the abbey of Grandselve and King Philip III of France - the King was represented by his...\n\n, Tarn-et-Garonne\nTarn-et-Garonne\nTarn-et-Garonne is a French department in the southwest of France. It is traversed by the Rivers Tarn and Garonne, from which it takes its name.-History:...\n\n, France; the late 15th century mansion where Fermat was born is now a museum. He was of Basque\nBasque people\nThe Basques as an ethnic group, primarily inhabit an area traditionally known as the Basque Country , a region that is located around the western end of the Pyrenees on the coast of the Bay of Biscay and straddles parts of north-central Spain and south-western France.The Basques are known in the...\n\norigin. Fermat's father was a wealthy leather merchant and second consul of Beaumont-de-Lomagne. Pierre had a brother and two sisters and was almost certainly brought up in the town of his birth. There is little evidence concerning his school education, but it may have been at the local Franciscan\nFranciscan\nMost Franciscans are members of Roman Catholic religious orders founded by Saint Francis of Assisi. Besides Roman Catholic communities, there are also Old Catholic, Anglican, Lutheran, ecumenical and Non-denominational Franciscan communities....\n\nmonastery.\nHe attended the University of Toulouse\nUniversity of Toulouse\nThe Universit\u00e9 de Toulouse is a consortium of French universities, grandes \u00e9coles and other institutions of higher education and research, named after one of the earliest universities established in Europe in 1229, and including the successor universities to that earlier university...\n\nbefore moving to Bordeaux\nBordeaux\nBordeaux is a port city on the Garonne River in the Gironde department in southwestern France.The Bordeaux-Arcachon-Libourne metropolitan area, has a population of 1,010,000 and constitutes the sixth-largest urban area in France. It is the capital of the Aquitaine region, as well as the prefecture...\n\nin the second half of the 1620s. In Bordeaux he began his first serious mathematical researches and in 1629 he gave a copy of his restoration of Apollonius\nApollonius of Perga\nApollonius of Perga [Pergaeus] was a Greek geometer and astronomer noted for his writings on conic sections. His innovative methodology and terminology, especially in the field of conics, influenced many later scholars including Ptolemy, Francesco Maurolico, Isaac Newton, and Ren\u00e9 Descartes...\n\n's De Locis Planis to one of the mathematicians there. Certainly in Bordeaux he was in contact with Beaugrand and during this time he produced important work on maxima and minima\nMaxima and minima\nIn mathematics, the maximum and minimum of a function, known collectively as extrema , are the largest and smallest value that the function takes at a point either within a given neighborhood or on the function domain in its entirety .More generally, the...\n\nwhich he gave to \u00c9tienne d'Espagnet who clearly shared mathematical interests with Fermat. There he became much influenced by the work of Fran\u00e7ois Vi\u00e8te\nFran\u00e7ois Vi\u00e8te\nFran\u00e7ois Vi\u00e8te , Seigneur de la Bigoti\u00e8re, was a French mathematician whose work on new algebra was an important step towards modern algebra, due to its innovative use of letters as parameters in equations...\n\n.\n\nFrom Bordeaux, Fermat went to Orl\u00e9ans\nOrl\u00e9ans\n-Prehistory and Roman:Cenabum was a Gallic stronghold, one of the principal towns of the Carnutes tribe where the Druids held their annual assembly. It was conquered and destroyed by Julius Caesar in 52 BC, then rebuilt under the Roman Empire...\n\nwhere he studied law at the University. He received a degree in civil law before, in 1631, receiving the title of councillor at the High Court of Judicature in Toulouse, which he held for the rest of his life. Due to the office he now held he became entitled to change his name from Pierre Fermat to Pierre de Fermat. Fluent in Latin, Basque, classical Greek, Italian, and Spanish, Fermat was praised for his written verse in several languages, and his advice was eagerly sought regarding the emendation of Greek texts.\n\nHe communicated most of his work in letters to friends, often with little or no proof of his theorems. This allowed him to preserve his status as an \"amateur\" while gaining the recognition he desired. This naturally led to priority disputes with contemporaries such as Descartes\nRen\u00e9 Descartes\nRen\u00e9 Descartes ; was a French philosopher and writer who spent most of his adult life in the Dutch Republic. He has been dubbed the 'Father of Modern Philosophy', and much subsequent Western philosophy is a response to his writings, which are studied closely to this day...\n\nand Wallis. He developed a close relationship with Blaise Pascal\nBlaise Pascal\nBlaise Pascal , was a French mathematician, physicist, inventor, writer and Catholic philosopher. He was a child prodigy who was educated by his father, a tax collector in Rouen...\n\n.\n\nAnders Hald\nAnders Hald\nAnders Hald was a Danish statistician who made contributions to the history of statistics.He was a professor at the University of Copenhagen from 1960 to 1982.- Bibliography :...\n\nwrites that, \"The basis of Fermat's mathematics was the classical Greek treatises combined with Vieta's new algebra\nNew algebra\nThe new algebra or symbolic analysis is a formalization of algebra promoted by Fran\u00e7ois Vi\u00e8te in 1591 and by his successors...\n\nic methods.\"\n\n### Work\n\nFermat's pioneering work in analytic geometry was circulated in manuscript form in 1636, predating the publication of Descartes' famous La g\u00e9om\u00e9trie. This manuscript was published posthumously in 1679 in \"Varia opera mathematica\", as Ad Locos Planos et Solidos Isagoge, (\"Introduction to Plane and Solid Loci\").\n\nIn Methodus ad disquirendam maximam et minima and in De tangentibus linearum curvarum, Fermat developed a method for determining maxima, minima, and tangents to various curves that was equivalent to differentiation. In these works, Fermat obtained a technique for finding the centers of gravity of various plane and solid figures, which led to his further work in quadrature\nNumerical integration\nIn numerical analysis, numerical integration constitutes a broad family of algorithms for calculating the numerical value of a definite integral, and by extension, the term is also sometimes used to describe the numerical solution of differential equations. This article focuses on calculation of...\n\n.\n\nFermat was the first person known to have evaluated the integral of general power functions. Using an ingenious trick, he was able to reduce this evaluation to the sum of geometric series. The resulting formula was helpful to Newton\nIsaac Newton\nSir Isaac Newton PRS was an English physicist, mathematician, astronomer, natural philosopher, alchemist, and theologian, who has been \"considered by many to be the greatest and most influential scientist who ever lived.\"...\n\n, and then Leibniz, when they independently developed the fundamental theorem of calculus\nFundamental theorem of calculus\nThe first part of the theorem, sometimes called the first fundamental theorem of calculus, shows that an indefinite integration can be reversed by a differentiation...\n\n.\n\nIn number theory, Fermat studied Pell's equation\nPell's equation\nPell's equation is any Diophantine equation of the formx^2-ny^2=1\\,where n is a nonsquare integer. The word Diophantine means that integer values of x and y are sought. Trivially, x\u00a0=\u00a01 and y\u00a0=\u00a00 always solve this equation...\n\n, perfect number\nPerfect number\nIn number theory, a perfect number is a positive integer that is equal to the sum of its proper positive divisors, that is, the sum of its positive divisors excluding the number itself . Equivalently, a perfect number is a number that is half the sum of all of its positive divisors i.e...\n\ns, amicable number\nAmicable number\nAmicable numbers are two different numbers so related that the sum of the proper divisors of each is equal to the other number. A pair of amicable numbers constitutes an aliquot sequence of period 2...\n\ns and what would later become Fermat numbers. It was while researching perfect numbers that he discovered the little theorem\nFermat's little theorem\nFermat's little theorem states that if p is a prime number, then for any integer a, a\u00a0p\u00a0\u2212\u00a0a will be evenly divisible by p...\n\n. He invented a factorization method\u2014Fermat's factorization method\nFermat's factorization method\nFermat's factorization method, named after Pierre de Fermat, is based on the representation of an odd integer as the difference of two squares:N = a^2 - b^2.\\...\n\n\u2014as well as the proof technique of infinite descent\nInfinite descent\nIn mathematics, a proof by infinite descent is a particular kind of proof by contradiction which relies on the fact that the natural numbers are well ordered. One typical application is to show that a given equation has no solutions. Assuming a solution exists, one shows that another exists, that...\n\n, which he used to prove Fermat's Last Theorem for the case n = 4. Fermat developed the two-square theorem, and the polygonal number theorem\nFermat polygonal number theorem\nIn additive number theory, the Fermat polygonal number theorem states that every positive integer is a sum of at most -gonal numbers. That is, every positive number can be written as the sum of three or fewer triangular numbers, and as the sum of four or fewer square numbers, and as the sum of...\n\n, which states that each number is a sum of three triangular number\nTriangular number\nA triangular number or triangle number numbers the objects that can form an equilateral triangle, as in the diagram on the right. The nth triangle number is the number of dots in a triangle with n dots on a side; it is the sum of the n natural numbers from 1 to n...\n\ns, four square numbers\nLagrange's four-square theorem\nLagrange's four-square theorem, also known as Bachet's conjecture, states that any natural number can be represented as the sum of four integer squaresp = a_0^2 + a_1^2 + a_2^2 + a_3^2\\ where the four numbers are integers...\n\n, five pentagonal number\nPentagonal number\nA pentagonal number is a figurate number that extends the concept of triangular and square numbers to the pentagon, but, unlike the first two, the patterns involved in the construction of pentagonal numbers are not rotationally symmetrical...\n\ns, and so on.\n\nAlthough Fermat claimed to have proved all his arithmetic theorems, few records of his proofs have survived. Many mathematicians, including Gauss\nCarl Friedrich Gauss\nJohann Carl Friedrich Gauss was a German mathematician and scientist who contributed significantly to many fields, including number theory, statistics, analysis, differential geometry, geodesy, geophysics, electrostatics, astronomy and optics.Sometimes referred to as the Princeps mathematicorum...\n\n, doubted several of his claims, especially given the difficulty of some of the problems and the limited mathematical tools available to Fermat. His famous Last Theorem\nFermat's Last Theorem\nIn number theory, Fermat's Last Theorem states that no three positive integers a, b, and c can satisfy the equation an\u00a0+\u00a0bn\u00a0=\u00a0cn for any integer value of n greater than two....\n\nwas first discovered by his son in the margin on his father's copy of an edition of Diophantus, and included the statement that the margin was too small to include the proof. He had not bothered to inform even Marin Mersenne\nMarin Mersenne\nMarin Mersenne, Marin Mersennus or le P\u00e8re Mersenne was a French theologian, philosopher, mathematician and music theorist, often referred to as the \"father of acoustics\"...\n\nof it. It was not proved until 1994, using techniques unavailable to Fermat.\n\nAlthough he carefully studied, and drew inspiration from Diophantus, Fermat began a different tradition. Diophantus was content to find a single solution to his equations, even if it were an undesired fractional one. Fermat was interested only in integer solutions to his Diophantine equation\nDiophantine equation\nIn mathematics, a Diophantine equation is an indeterminate polynomial equation that allows the variables to be integers only. Diophantine problems have fewer equations than unknown variables and involve finding integers that work correctly for all equations...\n\ns, and he looked for all possible general solutions. He often proved that certain equations had no solution\nEmpty set\nIn mathematics, and more specifically set theory, the empty set is the unique set having no elements; its size or cardinality is zero. Some axiomatic set theories assure that the empty set exists by including an axiom of empty set; in other theories, its existence can be deduced...\n\n, which usually baffled his contemporaries.\n\nThrough his correspondence with Pascal in 1654, Fermat and Pascal helped lay the fundamental groundwork for the theory of probability. From this brief but productive collaboration on the problem of points\nProblem of points\nThe problem of points, also called the problem of division of the stakes, is a classical problem in probability theory. One of the famous problems that motivated the beginnings of modern probability theory in the 17th century, it led Blaise Pascal to the first explicit reasoning about what today is...\n\n, they are now regarded as joint founders of probability theory\nProbability theory\nProbability theory is the branch of mathematics concerned with analysis of random phenomena. The central objects of probability theory are random variables, stochastic processes, and events: mathematical abstractions of non-deterministic events or measured quantities that may either be single...\n\n. Fermat is credited with carrying out the first ever rigorous probability calculation. In it, he was asked by a professional gambler\u00a0why if he bet on rolling at least one six in four throws of a die he won in the long term, whereas betting on throwing at least one double-six in 24 throws of two dice\nDice\nA die is a small throwable object with multiple resting positions, used for generating random numbers...\n\nresulted in him losing. Fermat subsequently proved why this was the case mathematically.\n\nFermat's principle of least time\u00a0(which he used to derive Snell's law\nSnell's law\nIn optics and physics, Snell's law is a formula used to describe the relationship between the angles of incidence and refraction, when referring to light or other waves passing through a boundary between two different isotropic media, such as water and glass...\n\nin 1657) was the first variational principle\nHistory of variational principles in physics\nA variational principle in physics is an alternative method for determining the state or dynamics of a physical system, by identifying it as an extremum of a function or functional...\n\nenunciated in physics since Hero of Alexandria\nHero of Alexandria\nHero of Alexandria was an ancient Greek mathematician and engineerEnc. Britannica 2007, \"Heron of Alexandria\" who was active in his native city of Alexandria, Roman Egypt...\n\ndescribed a principle of least distance in the first century CE. In this way, Fermat is recognized as a key figure in the historical development of the fundamental principle of least action\nPrinciple of least action\nIn physics, the principle of least action \u2013 or, more accurately, the principle of stationary action \u2013 is a variational principle that, when applied to the action of a mechanical system, can be used to obtain the equations of motion for that system...\n\nin physics. The terms Fermat's principle\nFermat's principle\nIn optics, Fermat's principle or the principle of least time is the principle that the path taken between two points by a ray of light is the path that can be traversed in the least time. This principle is sometimes taken as the definition of a ray of light...\n\nand Fermat functional were named in recognition of this role.\n\n### Death\n\nHe died at Castres\nCastres\nCastres is a commune, and arrondissement capital in the Tarn department and Midi-Pyr\u00e9n\u00e9es region in southern France. It lies in the former French province of Languedoc....\n\n, Tarn. The oldest and most prestigious high school in Toulouse\nToulouse\nToulouse is a city in the Haute-Garonne department in southwestern FranceIt lies on the banks of the River Garonne, 590\u00a0km away from Paris and half-way between the Atlantic Ocean and the Mediterranean Sea...\n\nis named after him: the Lyc\u00e9e Pierre de Fermat. French sculptor Th\u00e9ophile Barrau\nTh\u00e9ophile Barrau\nTh\u00e9ophile Barrau was a French sculptor.Barrau was born in Carcassonne. He was a student of Alexandre Falgui\u00e8re and started at the Salon in 1874. He received awards in 1879, 1880, 1889, and became a Chevalier of the Legion of Honor in 1892...\n\nmade a marble statue named Hommage \u00e0 Pierre Fermat as tribute to Fermat, now at the Capitole of Toulouse.\n\n## Assessment of his work\n\nTogether with Ren\u00e9 Descartes\nRen\u00e9 Descartes\nRen\u00e9 Descartes ; was a French philosopher and writer who spent most of his adult life in the Dutch Republic. He has been dubbed the 'Father of Modern Philosophy', and much subsequent Western philosophy is a response to his writings, which are studied closely to this day...\n\n, Fermat was one of the two leading mathematicians of the first half of the 17th century. According to Peter L. Bernstein, in his book Against the Gods, Fermat \"was a mathematician of rare power. He was an independent inventor of analytic geometry, he contributed to the early development of calculus, he did research on the weight of the earth, and he worked on light refraction and optics. In the course of what turned out to be an extended correspondence with Pascal, he made a significant contribution to the theory of probability. But Fermat's crowning achievement was in the theory of numbers.\"\n\nRegarding Fermat's work in analysis, Isaac Newton\nIsaac Newton\nSir Isaac Newton PRS was an English physicist, mathematician, astronomer, natural philosopher, alchemist, and theologian, who has been \"considered by many to be the greatest and most influential scientist who ever lived.\"...\n\nwrote that his own early ideas about calculus came directly from \"Fermat's way of drawing tangents.\"\n\nOf Fermat's number theoretic work, the great 20th-century mathematician Andr\u00e9 Weil\nAndr\u00e9 Weil\nAndr\u00e9 Weil was an influential mathematician of the 20th century, renowned for the breadth and quality of his research output, its influence on future work, and the elegance of his exposition. He is especially known for his foundational work in number theory and algebraic geometry...\n\nwrote that \"... what we possess of his methods for dealing with curves\nAlgebraic curve\nIn algebraic geometry, an algebraic curve is an algebraic variety of dimension one. The theory of these curves in general was quite fully developed in the nineteenth century, after many particular examples had been considered, starting with circles and other conic sections.- Plane algebraic curves...\n\nof genus 1\nElliptic curve\nIn mathematics, an elliptic curve is a smooth, projective algebraic curve of genus one, on which there is a specified point O. An elliptic curve is in fact an abelian variety \u2014 that is, it has a multiplication defined algebraically with respect to which it is a group \u2014 and O serves as the identity...\n\nis remarkably coherent; it is still the foundation for the modern theory of such curves. It naturally falls into two parts; the first one ... may conveniently be termed a method of ascent, in contrast with the descent\nInfinite descent\nIn mathematics, a proof by infinite descent is a particular kind of proof by contradiction which relies on the fact that the natural numbers are well ordered. One typical application is to show that a given equation has no solutions. Assuming a solution exists, one shows that another exists, that...\n\nwhich is rightly regarded as Fermat's own.\" Regarding Fermat's use of ascent, Weil continued \"The novelty consisted in the vastly extended use which Fermat made of it, giving him at least a partial equivalent of what we would obtain by the systematic use of the group theoretical\nGroup theory\nIn mathematics and abstract algebra, group theory studies the algebraic structures known as groups.The concept of a group is central to abstract algebra: other well-known algebraic structures, such as rings, fields, and vector spaces can all be seen as groups endowed with additional operations and...\n\nproperties of the rational point\nRational point\nIn number theory, a K-rational point is a point on an algebraic variety where each coordinate of the point belongs to the field K. This means that, if the variety is given by a set of equationsthen the K-rational points are solutions \u2208Kn of the equations...\n\ns on a standard cubic.\" With his gift for number relations and his ability to find proofs for many of his theorems, Fermat essentially created the modern theory of numbers.","date":"2013-05-18 16:19:54","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 5, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.6012943983078003, \"perplexity\": 1296.547426118237}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2013-20\/segments\/1368696382560\/warc\/CC-MAIN-20130516092622-00085-ip-10-60-113-184.ec2.internal.warc.gz\"}"}
null
null
package org.nd4j.linalg.util; import org.apache.commons.math3.complex.Complex; import org.apache.commons.math3.util.FastMath; import org.nd4j.linalg.api.complex.IComplexDouble; import org.nd4j.linalg.api.complex.IComplexFloat; import org.nd4j.linalg.api.complex.IComplexNumber; import org.nd4j.linalg.factory.Nd4j; /** * @author Adam Gibson */ public class ComplexUtil { private ComplexUtil() {} /** * Create complex number where the * @param realComponents the real components for the complex * @return the complex numbers based on the given real components */ public static IComplexNumber[][] complexNumbersFor(float[][] realComponents) { IComplexNumber[][] ret = new IComplexNumber[realComponents.length][realComponents[0].length]; for (int i = 0; i < realComponents.length; i++) for (int j = 0; j < realComponents[i].length; j++) ret[i][j] = Nd4j.createComplexNumber(realComponents[i][j], 0); return ret; } /** * Create complex number where the * @param realComponents the real components for the complex * @return the complex numbers based on the given real components */ public static IComplexNumber[][] complexNumbersFor(double[][] realComponents) { IComplexNumber[][] ret = new IComplexNumber[realComponents.length][realComponents[0].length]; for (int i = 0; i < realComponents.length; i++) for (int j = 0; j < realComponents[i].length; j++) ret[i][j] = Nd4j.createComplexNumber(realComponents[i][j], 0); return ret; } /** * Create complex number where the * @param realComponents the real components for the complex * @return the complex numbers based on the given real components */ public static IComplexNumber[] complexNumbersFor(float[] realComponents) { IComplexNumber[] ret = new IComplexNumber[realComponents.length]; for (int i = 0; i < realComponents.length; i++) ret[i] = Nd4j.createComplexNumber(realComponents[i], 0); return ret; } /** * Create complex number where the * @param realComponents the real components for the complex * @return the complex numbers based on the given real components */ public static IComplexNumber[] complexNumbersFor(double[] realComponents) { IComplexNumber[] ret = new IComplexNumber[realComponents.length]; for (int i = 0; i < realComponents.length; i++) ret[i] = Nd4j.createComplexNumber(realComponents[i], 0); return ret; } /** * Return the sin value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber atan(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).atan(); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the sin value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber acos(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).acos(); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the sin value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber asin(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).asin(); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the sin value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber sin(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).sin(); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the ceiling value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber ceil(IComplexNumber num) { Complex c = new Complex(FastMath.ceil(num.realComponent().doubleValue()), FastMath.ceil(num.imaginaryComponent().doubleValue())); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the floor value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber floor(IComplexNumber num) { Complex c = new Complex(FastMath.floor(num.realComponent().doubleValue()), FastMath.floor(num.imaginaryComponent().doubleValue())); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the log value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber neg(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).negate(); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the log value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber log(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).log(); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the absolute value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber sqrt(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).sqrt(); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the absolute value of the given complex number * * @param num the number to getScalar the absolute value for * @return the absolute value of this complex number */ public static IComplexNumber abs(IComplexNumber num) { double c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).abs(); return Nd4j.createDouble(c, 0); } public static IComplexNumber round(IComplexNumber num) { return Nd4j.createDouble(Math.round(num.realComponent().doubleValue()), Math.round(num.imaginaryComponent().doubleValue())); } /** * Raise a complex number to a power * * @param num the number to raise * @param power the power to raise to * @return the number raised to a power */ public static IComplexNumber pow(IComplexNumber num, IComplexNumber power) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).pow( new Complex(power.realComponent().doubleValue(), power.imaginaryComponent().doubleValue())); if (c.isNaN()) c = new Complex(Nd4j.EPS_THRESHOLD, 0.0); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Raise a complex number to a power * * @param num the number to raise * @param power the power to raise to * @return the number raised to a power */ public static IComplexNumber pow(IComplexNumber num, double power) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).pow(power); if (c.isNaN()) c = new Complex(Nd4j.EPS_THRESHOLD, 0.0); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the cos of a complex number * * @param num the tanh of a complex number * @return the tanh of a complex number */ public static IComplexNumber cos(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).cos(); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the tanh of a complex number * * @param num the tanh of a complex number * @return the tanh of a complex number */ public static IComplexNumber hardTanh(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).tanh(); if (c.getReal() < -1.0) c = new Complex(-1.0, c.getImaginary()); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Return the tanh of a complex number * * @param num the tanh of a complex number * @return the tanh of a complex number */ public static IComplexNumber tanh(IComplexNumber num) { Complex c = new Complex(num.realComponent().doubleValue(), num.imaginaryComponent().doubleValue()).tanh(); return Nd4j.createDouble(c.getReal(), c.getImaginary()); } /** * Returns the exp of a complex number: * Let r be the realComponent component and i be the imaginary * Let ret be the complex number returned * ret -> exp(r) * cos(i), exp(r) * sin(i) * where the first number is the realComponent component * and the second number is the imaginary component * * @param d the number to getFromOrigin the exp of * @return the exponential of this complex number */ public static IComplexNumber exp(IComplexNumber d) { if (d instanceof IComplexFloat) return exp((IComplexFloat) d); return exp((IComplexDouble) d); } /** * Returns the exp of a complex number: * Let r be the realComponent component and i be the imaginary * Let ret be the complex number returned * ret -> exp(r) * cos(i), exp(r) * sin(i) * where the first number is the realComponent component * and the second number is the imaginary component * * @param d the number to getFromOrigin the exp of * @return the exponential of this complex number */ public static IComplexDouble exp(IComplexDouble d) { return Nd4j.createDouble(FastMath.exp(d.realComponent()) * FastMath.cos(d.imaginaryComponent()), FastMath.exp(d.realComponent()) * FastMath.sin(d.imaginaryComponent())); } /** * Returns the exp of a complex number: * Let r be the realComponent component and i be the imaginary * Let ret be the complex number returned * ret -> exp(r) * cos(i), exp(r) * sin(i) * where the first number is the realComponent component * and the second number is the imaginary component * * @param d the number to getFromOrigin the exp of * @return the exponential of this complex number */ public static IComplexFloat exp(IComplexFloat d) { return Nd4j.createFloat((float) FastMath.exp(d.realComponent()) * (float) FastMath.cos(d.imaginaryComponent()), (float) FastMath.exp(d.realComponent()) * (float) FastMath.sin(d.imaginaryComponent())); } }
{ "redpajama_set_name": "RedPajamaGithub" }
2,030
Dennis Flinsenberg and I am a 33 year old visual content creator with a long background in UX/UI design. I currently live in Nijmegen (Lent), The Netherlands. In 2017 I moved into the role of visual content creator. My main focus was creating photo and video, but I also worked a lot on graphic design and animations. From 2015 until 2017 I worked as a UX/UI Designer at Iquality. I was the first dedicated designer at the company, making it a very challenging position. I worked on projects for Damen Shipyards, Lease Plan Bank, Athlon and Garantibank. For 5 years I worked as a UX/UI designer at LinkU Internet Media in Nijmegen, occasionally branching out into other fields such as animation or graphic design. I've been involved in the design of projects for clients like N.E.C. Nijmegen, Hunkemöller, Radboud Universiteit, Yournalist and Sophie & James. I was a motion design intern at MediaMonks in the year 2009. Worked on projects for clients asTomTom, McDonalds and Postcode Loterij. My main focus was creating (3d) animations with Cinema4D and After Effects. I worked at Image-in as a 3d visualizer. From February to June as an intern, after that as full-time employee. The internship was focused around creating 3d models and animations with Cinema4D. During the course of my study I focused on user interface design, motion design and 3d modeling.
{ "redpajama_set_name": "RedPajamaC4" }
3,286
CyNetwork is the primary interface for representing a network (graph) data structure in Cytoscape. Specifically, CyNetwork represents a multi-graph as multiple edges may exist between nodes. Edges may be directed, undirected, or both. The name of the default public CyTable that is created by default for CyNetworks, CyNodes, and CyEdges. The name of the default hidden CyTable that is created by default for CyNetworks, CyNodes, and CyEdges. The name of the local public CyTable that is created by default for CyNetworks, CyNodes, and CyEdges. A String column created by default for every CyNetwork that holds the name of the entry. A boolean column created by default for every CyNode or CyEdge that holds the selection state of the entry. This method is used to create and add an edge to this network. This method is used to create and add a node to this network. Determine if this CyNetwork contains a particular edge. Determine if this CyNetwork contains an edge between two nodes. Determine if this CyNetwork contains a particular node. Gets an Iteable of edges that connect to this node. Get the list of edges that connect to this node. Get the list of edges that connect two nodes. A convenience method returns the default attribute table for the edges of this network. A convenience method returns the default attribute table for this network. A convenience method returns the default attribute table for the nodes of this network. Return the CyEdge that has the SUID. Return the number of edges in this network. Return a list of the edges in this network. Get the list of nodes that neighbor this node where the definition of "neighbor" is a node that is connected to this node by the passed edgeType. Return the CyNode that has the SUID. Return the number of nodes in this network. Return a list of the nodes in this network. A convenience method that returns the row in the default table for this object. Returns the row for the specified table name for this object. Returns how (or if) this CyNetwork should be saved. Returns the table with the specified namespace and type from this network. Remove an edge from the network and delete the edge (if it only exists in this network). Remove a node from the network and delete the node (if it only exists in this network). The name of the default public CyTable that is created by default for CyNetworks, CyNodes, and CyEdges. Other CyTables may also be associated with networks -- see CyTableManager for more information. The table should be referenced using this constant: CyNetwork.DEFAULT_ATTRS. The name of the local public CyTable that is created by default for CyNetworks, CyNodes, and CyEdges. This table contains attributes which are specific to only this network. -- see CyTableManager for more information. The table should be referenced using this constant: CyNetwork.DEFAULT_ATTRS. The name of the default hidden CyTable that is created by default for CyNetworks, CyNodes, and CyEdges. Other CyTables may also be associated with networks -- see CyTableManager for more information. Remove a node from the network and delete the node (if it only exists in this network). See CyRootNetwork for information about having the same node in two networks. Remove an edge from the network and delete the edge (if it only exists in this network). See CyRootNetwork for information about having the same edge in two networks. Return a list of the nodes in this network. The list will be empty if there are no nodes in the network. This method should never return null. Modifying this list (if allowed by the implementation) has no effect on the network. Return a list of the edges in this network. The list will be empty if there are no edges in the network. This method should never return null. Modifying this list (if allowed by the implementation) has no effect on the network. Determine if this CyNetwork contains an edge between two nodes. Note that if the edge is directed, the source and targets must match. Return the CyNode that has the SUID. To iterate over all nodes, iterate over getNodeList(). We make no guarantees on what represent valid values for the index. The only valid indices are those accessed from existing nodes. the associated CyNode or null if there is no node with that index in this network. Return the CyEdge that has the SUID. To iterate over all edges, iterate over getEdgeList(). We make no guarantees on what represent valid values for the index. The only valid indices are those accessed from existing edges. the associated CyEdge or null if there is no edge with that index in this network. Get the list of nodes that neighbor this node where the definition of "neighbor" is a node that is connected to this node by the passed edgeType. The CyEdge.Type enum is used to determine whether the list includes undirected, directed, incoming, or outgoing edges. The list will be empty if there are no neighbor nodes found. This method should never return null. Modifying this list (if allowed by the implementation) has no effect on the network. This method only returns immediate neighbors. Get the list of edges that connect to this node. The CyEdge.Type enum is used to determine whether the list includes undirected, directed, incoming, or outgoing edges. The list will be empty if there are no adjacent edges found. This method should never return null. Modifying this list (if allowed by the implementation) has no effect on the network. Gets an Iteable of edges that connect to this node. The CyEdge.Type enum is used to determine whether the list includes undirected, directed, incoming, or outgoing edges. The iterable will be empty if there are no adjacent edges found. This method should never return null. Modifying this list (if allowed by the implementation) has no effect on the network. The Iterable implementation may be more efficient because it doesn't necessarily need to create a list first. However, that depends on implementation details, so you should evaluate the method performance yourself. Get the list of edges that connect two nodes. The CyEdge.Type enum is used to determine whether the list includes undirected, directed, incoming, or outgoing edges. The list will be empty if no connecting edges are found. This method should never return null. This method will NOT find the shortest path between two nodes that are not already immediate neighbors. Modifying this list (if allowed by the implementation) has no effect on the network. the list of edges that include source and target and directed edges. The default attribute table for this network. The default attribute table for the nodes of this network. The default attribute table for the edges of this network. type - Type of CyIdentifiable associated with the table. namespace - the namespace the table should belong to. the table with the specified namespace and type from the network. Returns the row for the specified table name for this object. A null entry or a an entry not found in this network will return null, but otherwise every node or edge in the network is guaranteed to have a row. entry - The entry (node, edge, network) whose row we're looking for. namespace - the namespace of the table from which to extract the row. the row in the table of the specified name for this object. A convenience method that returns the row in the default table for this object. This method is equivalent to calling getCyRow(entry,DEFAULT_ATTRS). A null entry or a an entry not found in this network will return null. the row in the default table for this object. how (or if) this CyNetwork should be saved.
{ "redpajama_set_name": "RedPajamaC4" }
88
{"url":"https:\/\/physics.stackexchange.com\/questions\/228934\/how-can-we-use-conductivity-measurements-of-water-to-find-hydrogen-ion-concentra","text":"# How can we use conductivity measurements of water to find hydrogen ion concentration of water?\n\nHow can we use conductivity measurements of water to find the hydrogen ion ($\\mathrm{H}^+$) concentration in water?\n\n\u2022 For the question are you considering that H+ is the only cation? Or is your solution filled with an assortment of cations and anions. Jan 12, 2016 at 11:26\n\u2022 It is pure water. No other ions are present.\n\u2013\u00a0user102708\nJan 12, 2016 at 13:30\n\u2022 hmmm...are you allowed just to use the pH and figure from there? Jan 12, 2016 at 16:23\n\u2022 No, we want to find H+ ion concentration using the conductivity method. We want to find the pH using that.\n\u2013\u00a0user102708\nJan 13, 2016 at 13:55\n\u2022 Isn't that what a pH meter does? Feb 22, 2019 at 16:41\n\nOkay, this is how I would do it, I would set up cathode and anode to measure the conductivity. I would then compare that to a buffer solution of known ph. If you can't find a buffer solution, see if there is a chart you can find of Hydrogen ion conductivity relationship to pH.(But then you may have to take in temperature of solutions). http:\/\/patentimages.storage.googleapis.com\/EP0229038A2\/imgb0001.png\n\nSuch as this, this will give you a relationship which you should be able to figure out from. You do know what \"Molar\" means? or to figure ppms?\n\n\u2022 If you can't find a buffer solution, see if there is a chart you can find of Hydrogen ion conductivity relationship to pH.(But then you may have to take in temperature of solutions).\n\u2013\u00a0user102708\nJan 13, 2016 at 18:47\n\u2022 Assume you do not have such a chart, you are the first person doing the expt. Buffer conductivity is not only due to H+ ions so you can't compare them directly.\n\u2013\u00a0user102708\nJan 13, 2016 at 18:58\n\u2022 @OddharakTyagi, I added a chart of conductivity and concentrations. But, I guess you are trying to figure it out from \"scratch\". My thought would be to work it back from your conductivity test. Look up the formulas, will deal with Ohms, possibly can figure from there. Jan 13, 2016 at 19:19\n\u2022 Be careful . $H_2$ and $O_2$ will form. Clearly that is outside the scope of Ohm's law. Nov 3, 2019 at 10:10\n\u2022 How does this differ from a standard pH meter ? Nov 3, 2019 at 10:16","date":"2022-06-28 07:13:14","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 1, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.4785514175891876, \"perplexity\": 845.5495106999549}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.3, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2022-27\/segments\/1656103355949.26\/warc\/CC-MAIN-20220628050721-20220628080721-00583.warc.gz\"}"}
null
null
Q: CakePHP redirect after changing language on edit/view action im building an application in cakephp in multilingual, so for example when i am on homepage of application and i change the language it changes ok, but the problems appear when i am on an edit or view part for example: example.com/en/item/edit/1 and i change language from en to it in the edit view it redirects to: example.com/it/item/edit but it doesent keep the passed parameter as it supposed to. what should i do to make it work???? the link that i use to change lanugages are: <?php echo "<ul>"; echo "<li>". $this->Html->link('English', array('language'=>'en')) . "</li>"; echo "<li>". $this->Html->link('Italian', array('language'=>'it')). "</li>"; echo "<li>". $this->Html->link('German', array('language'=>'de')). "</li>"; echo "<li>". $this->Html->link('Franch', array('language'=>'fr')). "</li>"; echo "<li>". $this->Html->link('Spanish', array('language'=>'es')). "</li>"; echo "<li>". $this->Html->link('Russian', array('language'=>'ru')). "</li>"; echo "<li>". $this->Html->link('Chinesse', array('language'=>'zh')). "</li>"; echo "<li>". $this->Html->link('Portuguese', array('language'=>'pt')). "</li>"; echo "<ul>"; ?> The rout that i use are: /*** Languages */ Router::connect('/:language',array('controller' => 'properties', 'action' => 'index','home'),array('language' => '[a-z]{2}')); Router::connect('/:language/:controller',array('controller' => 'pages','action' => 'index'),array('language' => '[a-z]{2}')); Router::connect('/:language/:controller/:action/*', array('controller' => 'pages','action' => null), array('language' => '[a-z]{2}')); And the function on the AppController that i call in beforeFilter is: private function _setLanguage() { if ($this->Cookie->read('lang') && !$this->Session->check('Config.language')) { $this->Session->write('Config.language', $this->Cookie->read('lang')); } else if (isset($this->params['language']) && ($this->params['language'] != $this->Session->read('Config.language'))){ $this->Session->write('Config.language', $this->params['language']); $this->Cookie->write('lang', $this->params['language'], false, '2 days'); } } Also i have overwriten the redirect method as follow on AppController: public function redirect( $url, $status = NULL, $exit = true ) { if (!isset($url['language']) && $this->Session->check('Config.language')) { $url['language'] = $this->Session->read('Config.language'); } parent::redirect($url,$status,$exit); } A: Change all your language switching links to be similar to this: $passedArgs = $this->request->params['named'] + $this->request->params['pass']; $this->Html->link('English', array('language'=>'en') + $passedArgs);
{ "redpajama_set_name": "RedPajamaStackExchange" }
2,088
«"Perché non vieni più?". "Perché il vento cambia, Rocco. Io lo so. Anche tu lo sai"». Dice Antonio Manzini che i suoi romanzi li immagina, non semplicemente come una serie, ma come i «capitoli di un libro più grande» sul vicequestore Rocco Schiavone. Ad ogni episodio, mentre fruga tipicamente svogliato e vigile nel freddo di Aosta, il vicequestore con le sue Clarks entra anche in un pezzo ignoto del suo passato. Di modo che il caso criminale diventa un passaggio esistenziale. Un affondo psicologico dentro di sé avvolto in un'inchiesta di polizia. In Pulvis et umbra due trame si svolgono in parallelo. Ad Aosta si trova il cadavere di una trans. A Roma, in un campo verso la Pontina, due cani pastore annusano il cadavere di un uomo che porta addosso un foglietto scritto. L'indagine sul primo omicidio si smarrisce urtando contro identità nascoste ed esistenze oscurate. Il secondo lascia un cadavere che puzza di storie passate e di vendette. In entrambi Schiavone è messo in mezzo con la sua persona. E proprio quando il fantasma della moglie Marina comincia a ritirarsi, mentre l'agente Caterina Rispoli rivela un passato che chiede tenerezza e un ragazzino solitario risveglia sentimenti paterni inusitati, quando quindi la ruvida scorza con cui si protegge è sfidata da un po' di umanità intorno, le indagini lo sospingono a lottare contro le sue ombre. Tenta di afferrarle e gli sembra che si trasformino in polvere. La polvere che lascia ogni tradimento. Antonio Manzini, scrittore e sceneggiatore. Tra i suoi primi romanzi Sangue marcio e La giostra dei criceti (del 2007), ripubblicato da questa casa editrice nel 2017. La serie con Rocco Schiavone è iniziata con il romanzo Pista nera (Sellerio, 2013) cui sono seguiti La costola di Adamo (2014), Non è stagione (2015), Era di maggio (2015), Cinque indagini romane per Rocco Schiavone (2016) e 7-7-2007 (2016). Nel 2015 ha pubblicato Sull'orlo del precipizio in altra collana di questa casa editrice. La memoria 1073 DELLO STESSO AUTORE in questa collana Pista nera La costola di Adamo Non è stagione Era di maggio Cinque indagini romane per Rocco Shiavone 7-7-2007 La giostra dei criceti nella collana «Il divano» Sull'orlo del precipizio Antonio Manzini # Pulvis et umbra Sellerio editore Palermo 2017 © Sellerio editore via Enzo ed Elvira Sellerio 50 Palermo e-mail: info@sellerio.it www.sellerio.it Quest'opera è protetta dalla Legge sul diritto d'autore. È vietata ogni duplicazione, anche parziale, non autorizzata. EAN 978-88-389-3696-8 # Pulvis et umbra Quando un solo cane si mette ad abbaiare a un'ombra, diecimila cani ne fanno una realtà. CIORAN # Domenica Le luci della sera erano calate da una mezz'ora e l'aria era fresca e piacevole. Qualche ritardatario con passo affrettato rientrava a casa. Lui invece se ne stava lì, fermo, sul marciapiede di via Brean. Non si decideva. Bastava solo attraversare e suonare il citofono, il resto sarebbe venuto da sé. Eppure quel piccolo passo non riusciva a farlo. Le mani nelle tasche, continuava a stropicciare il foglietto di carta con l'indirizzo: via Brean 12, Studio Emme. Cosa lo bloccava? Chi gli aveva inchiodato le scarpe sul marciapiede? «Ciao amigo, vuoi?». Una voce lo fece voltare. Un africano carico di roba incellofanata gli offriva un pacco di calzini filo di Scozia. «Come stai? Dieci euro, amigo...». E allungò la mano libera, Marco come un automa gliela strinse. «Allora vuoi? Dieci euro!». Marco fece no con la testa. «Mi dai qualche spicio? Pe' caffè?». Marco fece sì con la testa ma rimase con le mani in tasca, immobile, una sentinella con una consegna precisa, un palo della luce in mezzo alla strada. Il nero aspettava e lo guardava, poi sorrise coi suoi denti bianchi e scosse la testa un paio di volte. «Amigo, dai spici?» ripeté. Lento Marco tirò fuori il portafogli. Dentro c'erano due banconote da 50 e una da 10. Prese quella da 10 euro e gliel'allungò. Il venditore senza fiatare acchiappò i soldi e in cambio mollò i calzini che Marco afferrò senza guardare. «Ciao amigo...» e con passo dinoccolato se ne andò. Marco tornò a guardare il civico 12. Un palazzo in cortina di due piani con un portone di vetro e ferro battuto, niente portiere, il citofono sul lato destro. Che ore sono? si chiese. Le otto e un quarto. Qual era l'orario? Dalle 15 alle 21 o dalle 15 alle 20? Magari era già andata via. Tirò fuori il cellulare e richiamò il numero che aveva fatto alle 10 del mattino. Attese fin quando rispose la segreteria. «Ciao... me chiamo Sonya... mi trovi a via Brean, che traversa di via Monte Grivola. Vieni... sono bela, calda latina e porto quinta de seno. Sono qui sempre ad aspettare per fare cose che a te ti piacciono... vuoi le coccole? Vuoi fare amore muy lungo? Doppia penetrazione? Ho pure sorpresa per te. Tutto quello che vuoi... ambiente relassado y pulito... vieni oggi domenica dalle quindici a ventuno a via Brean 12 e suona citofono Studio Emme... Emme come Milano... ciao belo te espero!». Era ancora in tempo. Ma lo stomaco continuava a sprofondare e le gambe a restare lì dov'erano. Forse perché la scena se l'era immaginata tante volte. Lei che lo aspettava in guêpière e calze nere fumé. Solo le mutandine, niente reggiseno. I capezzoli scuri sotto la trasparenza della vestaglia, mentre ondeggiava verso di lui sui tacchi vertiginosi che picchiavano sul pavimento. La bocca carnosa, gli occhi socchiusi, i capelli neri e sciolti, un profumo di fiori e pane caldo. Lo invitava a sedersi sul letto, lo baciava, lo spogliava, lo cavalcava per ore sbattendogli sul viso i seni enormi. Ma dentro di sé, in un angolo della coscienza, sapeva benissimo che per scoparsi una come Sophia Loren in Ieri oggi e domani ci voleva ben altro di un annuncio su Gradisca, incontri on line! Chissà cosa ci avrebbe trovato allo Studio Emme di via Brean. C'era una foto sulla pagina, ma era poi veritiera? Mostrava una donna in mutande e reggiseno, col viso nascosto. E quella frase: sorpresa per te, era la cosa che più lo eccitava. Marco non ce la faceva più. A 52 anni suonati, sposato da 25 e con tre figli, non praticava più da due anni. Barbara aveva chiuso il rubinetto, aveva decretato l'embargo da quando caldane e sbalzi di umore avevano preso il posto dei sorrisi e delle carezze. Non era più interessata al sesso, a lui invece andava come ai tempi del liceo. Due anni a digiuno, se si esclude una mezza fellatio che una rappresentante di rubinetterie di Grosseto gli aveva praticato nove mesi prima al convegno delle caldaie a Firenze mentre, ubriaco come una pigna, cantava «We are the champions!» in piazza della Signoria. Neanche si ricordava il nome della tipa, e non fu niente di eccezionale. E comunque prima di comporre il numero di Sonya ci aveva pensato per settimane. Era sempre lì lì per farlo, telefonino in mano, ma poi rinunciava. La notte sognava quell'incontro, e la mattina si svegliava con un'erezione talmente dolorosa che doveva correre in bagno a calmarla prima di colazione. Doveva scopare. Al negozio i due colleghi, Giorgio e Andrea, non facevano che parlare di amanti, mogli insaziabili, signore divorziate sempre disponibili. Lui si limitava a sorridere e a pensare a Barbara che aveva abbandonato da tempo sottovesti e intimo coordinato per pigiamoni con gli orsetti o magliettacce ormai scolorite della sua rivendita di termoidraulica. Via le scarpe coi tacchi per fare posto a ballerine sformate o a infradito casalinghe, il parrucchiere un ricordo lontano. Marco aveva provato ad affrontare la situazione, ma era come discutere con un muro. Inutile anche la puntatina alle terme di Pré-Saint-Didier, sperando che l'acqua calda e i massaggi avrebbero risvegliato in sua moglie, almeno per una notte, un po' di sana voglia. Invece quella alle nove e mezza già dormiva. E a nulla erano valsi i regali che le aveva fatto l'ultimo Natale. Barbara aveva riportato calze, reggicalze in pizzo e sottoveste al negozio e in cambio aveva preferito un bell'accappatoio giallo per Ginevra, la figlia più piccola, e una coppia di asciugamani blu. La frustrazione aumentava con la voglia, Marco non sapeva più dove sbattere la testa. Ecco perché ora si trovava lì, sul marciapiede, a guardare attonito il civico di una che per cento euro gli avrebbe dato una mezz'ora di pelle, profumo e parole sussurrate nelle orecchie. Ne ho il diritto, pensava. Ne ho bisogno. Cazzo, mica sono morto! E allora cosa lo faceva restare lì impalato? La paura. Paura che si portava dietro da quando aveva preso quella decisione. Paura di stringere il corpo nudo di un'estranea, di sentirne l'odore, e soprattutto paura che qualcuno lo vedesse. Aosta non era New York. In quel palazzo non conosceva nessuno, ma lui aveva un negozio, i clienti entravano e uscivano. E se avesse suonato allo Studio Emme e proprio mentre la tipa gracchiava «Vieni dentro belo te aspetavo calda y pronta» una madre coi figlioletti fosse uscita dal palazzo? Una figura di merda colossale. E se un vicino avesse strizzato gli occhi come a dire «ma io questo lo conosco. Che ci fa qui? Non ha un negozio di sanitari?». Le voci girano, si sa. E in meno di tre giorni tutti l'avrebbero saputo. Compresa sua moglie. E quel che è peggio Ginevra. Al liceo l'avrebbero presa in giro per anni cantandole dietro «Tuo padre è un puttaniere, tuo padre è un puttaniere!». Come guardare sua figlia negli occhi? Ci avrebbe più parlato? Già i rapporti con l'adolescente erano difficili, se uno ci aggiungeva quel peso da 30 tonnellate era la fine. Perché non aveva scelto una prostituta in un'altra città? Magari a Torino? Ci aveva pensato. Ma come giustificare alla moglie il viaggio a Torino? Per fare un'installazione? Non era mai successo in tanti anni di onorata carriera di negoziante. Barbara ci avrebbe impiegato venti minuti a scoprire la bugia. C'era da farsi alleati i due soci, chiedere che gli reggessero il gioco, ma allora il tradimento sarebbe diventato di dominio pubblico, o comunque se non pubblico, dominio di Giorgio e Andrea. E non gli piaceva l'idea che i suoi soci sapessero che a casa andava male, tanto che aveva bisogno di un'amante. Conoscevano Barbara da vent'anni. Era una mancanza di rispetto, uno sputare in faccia a Barbara, e questo non lo sopportava. Era una brava moglie, una brava madre, ma lui doveva scopare. Il cervello piccolo, che gli uomini hanno in mezzo alle gambe, non sentiva più ragioni. «Cosce seni chiappa, culo cosce seni labbra!» questi erano i messaggi che mandava, il leitmotiv degli ultimi due anni, e il cervello grande, che i maschi come tutti i mammiferi hanno nel cranio, aveva resistito. Ma la goccia cinese aveva scavato un solco che era diventato un ruscello, poi il letto di un fiume. Ormai non poteva più guardare la televisione, aprire una rivista, osservare il viavai delle donne per il corso che il cervello piccolo urlava il suo bisogno. «Cosce seni chiappa, culo cosce seni labbra!». Basta, si disse. Vado. Un uomo all'angolo con il suo cagnolino al guinzaglio sembrava aspettasse che quello facesse i suoi bisogni. A Marco invece pareva che lo stesse fissando. Sicuro si stava chiedendo: cosa ci fa quello lì in piedi da venti minuti? Chi è? Cosa vuole? Cosa cerca da queste parti? E se l'uomo col cane avesse chiamato il commissariato? «Polizia! C'è un tipo strano in piedi da mezz'ora davanti alla Cassa di Risparmio, venite a dare un'occhiata!». L'avrebbero portato in questura e lì per cavarsela avrebbe dovuto dire la verità: «Commissario, non mi fidavo ad attraversare la strada e citofonare a Sonya muy caliente che mi aspettava». L'ansia gli tolse i chiodi dalle suole delle scarpe, squagliò il silicone dalle giunture e spinse Marco ad attraversare la strada e finalmente arrivò davanti al civico 12. Buttò un occhio nell'androne attraverso i vetri. Nessuno. Nessuno sul marciapiede. L'uomo col cane era sparito. Sul citofono lo Studio Emme era l'interno 3. Dai, ora o mai più, pensò. Allungò il dito. Suonò. La luce a tempo si accese e Marco vide un ragazzo scendere le scale e avvicinarsi al portone. Teneva a tracolla una borsa sportiva. Eccola lì... la figura di merda! si disse. Il tizio s'era fermato a controllare la cassetta della posta. L'aveva aperta senza chiave per prelevare la corrispondenza. Basta! Stava rischiando troppo. Si allontanò di poco e si acquattò dietro una rientranza del palazzo. Il ragazzo uscì dal portone e se ne andò di fretta senza degnarlo di uno sguardo. Marco si infilò veloce come un ratto nell'androne delle scale spingendo il portone che non si era ancora chiuso. Fu una mossa impulsiva. Non avrebbe dovuto farla. Era tramontata la luna, la notte era a metà. Aosta dormiva. Non una luce, non un rumore. Solo il richiamo lontano di qualche uccello notturno. Dalla finestra aperta un profumo dolce di erba e terra bagnata penetrava nella stanza avvolgendo i mobili. E Rocco steso sul letto era sveglio, una mano sotto la testa l'altra poggiata sul fianco di Lupa. Eppure era stanco. Quel silenzio e quegli odori lo tenevano desto, gli ricordavano che l'estate stava arrivando. Chiuse gli occhi e cercò in mezzo a quei profumi qualcosa che lo riportasse indietro. Ce n'era uno pungente di cui non conosceva l'origine. Forse aghi di pino, o forse qualche fiore su un davanzale del dirimpettaio, ma gli arrivarono subito i ricordi di un'estate di mille anni prima. Una pineta, in Calabria, e lui aveva neanche 15 anni. Aveva preso il treno insieme a Brizio con gli zaini carichi e una tenda militare comprata a via Sannio. Un viaggio scomodo lungo una quaresima per arrivare in piena notte in riva al mare, in un campeggio di cui non ricordava il nome. Montarono la tenda sotto le stelle, ridendo e cercando di attutire i rumori del martello sui picchetti avvolgendoci intorno una maglietta. Da una roulotte carica di tedeschi un urlo aveva lacerato la notte: Scheisse! Si erano fermati per infilarsi nella canadese che al buio sembrava una discarica. La mattina dopo aprendo la zip con il sole già alto e bollente videro i campeggiatori che passavano lì davanti e guardando il loro accampamento ridevano a crepapelle. Avevano montato la tenda in discesa. Tutta storta, il telo floscio e gli zaini semiaperti con la roba sparsa sul terreno. Era la prima vacanza che faceva lontano da Roma. E conobbe Beote. Beote era bella. Veniva da un paese della Norvegia, talmente a nord che nei mesi invernali vedevano il sole sì e no a mezzogiorno. Fecero l'amore in riva al mare, e Rocco la ricordava come una delle esperienze più dolorose della sua vita. «Non sai Lupa, una cosa tremenda...» disse al cane che tirò su un orecchio per ascoltare. «No, perché io l'amore non l'avevo mai fatto. Cioè sì, ma roba leggera, vabbè, mi hai capito. E allora non è che fossi proprio pratico. Sapevo più o meno, ecco. Più o meno. Così lei si stende sulla spiaggia e io mi metto sopra. Lupa mia, non capivo più niente, mi girava la testa, il mare la luna la spiaggia. Ecco, la spiaggia. Mo' devi tenere a mente una cosa importante, in quel posto in Calabria la spiaggia era fatta di sassolini. Allora io comincio, la bacio e a un certo punto lei lo prende in mano e lo indirizza. Io mi sono sentito un coglione, Lupa, proprio un coglione. Volevo fare da me. Così le ho spostato la mano, mannaggia alla miseria». Si mise a ridere. «Mamma mia che è successo. Io mi strusciavo, spingevo, premevo, e un dolore...! Un dolore che non puoi capire! E insistevo, pigiavo, strofinavo fino a che non ho più resistito, mi bruciava tutto. Al buio mi sono alzato quasi urlando e sai che ho fatto? Sai che ha fatto 'sto cretino? Mi sono buttato a mare per cercare un po' di frescura, io avevo le fiamme là sotto! E certo, s'era tutto abraso, manco l'avessi scartavetrato. Pieno di tagli e mi vado a buttare in mare, capisci? Beote rideva a crepapelle. Che figura di merda, Lupa mia, che figura di merda... avevo lisciato il bersaglio, praticamente mi sono scopato la sabbia». Una brezza leggera, quasi una carezza lo fece rabbrividire. E sorrise. Sapeva chi la mandava. «Dormo un po'. Buonanotte Mari'...» e chiuse gli occhi. Non aveva preso il loden. A giugno bastava un maglione leggero, niente più pantaloni di velluto. Dell'inverno ormai passato restavano solo le Clarks, sedicesimo paio in dieci mesi. Lupa saltava felice cercando di mordere il guinzaglio. «Piano amore piccolo, piano, mo' scendiamo!». Imboccò le scale. Al terzo gradino sentì una porta sul pianerottolo che si apriva. «Dove va, al lavoro?». Tornò indietro. Gabriele, il vicino adolescente, era sull'uscio. La faccia da sonno, una maglietta nera con un teschio stampato e i calzoncini da basket lunghi al ginocchio. Ai piedi due scarpe da corsa sformate. «Sveglio a quest'ora, Gabrie'?». «Già...». «E non hai messo la musica. Bravo, cominci a imparare. Stammi bene». «Va già al lavoro?». «No. A fare colazione». «A quest'ora il bar è chiuso». «Ettore apre fra poco». «Ma non la fa mai a casa?». «Cosa?». «La colazione?». «No, Gabriele, a casa faccio altro...». «Cioè?». «I cazzi miei. Ora mi dici che vuoi?». Il ragazzo alzò le spalle. «Ho poche speranze, ma ci devo provare lo stesso. Ho un'interrogazione finale in latino. Se vado bene forse con qualche debito passo l'anno, se invece non ce la faccio...». «Cosa che vedo assai probabile...» sottolineò Rocco. «Ecco, se non ce la faccio mi bocciano in tronco». Lupa s'era messa a odorare le scarpe del ragazzo. «E lo so, sono cose che succedono. Andiamo Lupa!». «Lei ha fatto il liceo classico?». Rocco risalì un gradino e guardò Gabriele negli occhi. «Cosa stai cercando di dirmi?». «Niente. Magari mi poteva dare una mano». Il vicequestore sospirò. «Sono passati tanti anni, mi ricordo poco o niente». «Basta che tiene aperto il libro e mi fa le domande!» propose Gabriele con gli occhi sgranati dall'eccitazione. «Quando?». «Ora». «Tua sorella, ora vado al bar». «Ma io poi devo andare a scuola!». Rocco sbuffò alzando gli occhi al cielo. «La prego, venga, tanto mamma non c'è, è a Milano». «A che ora hai l'interrogazione?». «Alle 10». «Ma porc...». Rocco sbatté un piede a terra. «E non potevi venire ieri sera? Vabbè, entri alla seconda. Allora adesso vado al bar. Tu ti prepari, ti lavi, ti vesti da persona normale e non da lobotomizzato e la lezione l'andiamo a fare in questura. Va bene?». «Benissimo!». Seduto a via Croix de Ville fumava e osservava Lupa gironzolare muso a terra fra le vetrine dei negozi chiusi. In strada non c'era nessuno, a parte un camioncino che con due spazzoloni ripuliva il marciapiede. Con un fischio richiamò il cane e si incamminò verso piazza Chanoux. Il dolore sotto il piede sinistro sembrava dargli finalmente tregua. La tallonite dovuta sicuramente all'uso smodato e insistente delle Clarks, per un caso fortuito sembrava sparita. «Possiamo?» gridò entrando. Ettore era dietro il bancone. Davanti a lui una donna sulla quarantina stava facendo colazione. «Prego, prego, buongiorno». Il vicequestore e Lupa entrarono. La cucciola si andò a nascondere sotto il primo tavolino. «Un caffè ristretto e un bel cornetto». «Brioche» lo corresse Ettore. «Cornetto» insistette Rocco. La donna non si era voltata. Continuava a mangiare guardando dritto davanti a sé. Indossava un tailleur pesca e aveva un bel profilo. I capelli pettinati, due orecchini di perle, aveva lasciato l'impronta del rossetto sulla tazza del cappuccino. A Rocco non sfuggì lo sguardo d'intesa che Ettore gli lanciò mentre pressava il caffè nel braccetto. «Che c'è?». «Lei». «Cosa?». «Sapessi chi è...». Il dialogo fra i due maschi era avvenuto nel totale silenzio, solo uno scambio di sguardi e cenni del viso. La donna però aveva seguito quegli ammiccamenti dallo specchio enorme carico di bottiglie alle spalle del barman. «Sì, io la conosco, lei ancora no» disse. «So che lei viene sempre qui per la colazione e volevo farle una sorpresa...». «Se lei sa chi sono, perché non mi dice il suo nome?». La donna si prese tutto il tempo per finire il cappuccino. Poi posò la tazza, si pulì le mani con un tovagliolino e finalmente si voltò verso Rocco. Aveva gli occhi scuri impreziositi da un filo di eyeliner. «Mi chiamo Sandra Buccellato» e allungò la mano. A Rocco tornò su la pizza della sera prima. Sandra Buccellato, la giornalista ex moglie del questore che per mesi gli aveva dato il pilotto dalle pagine del quotidiano. In ogni articolo che scriveva trovava il modo di calunniare lui e la questura per l'omicidio di Adele nel suo vecchio appartamento a rue Piave. Accusava le forze di polizia di voler insabbiare la cosa, con illazioni e maldicenze affliggeva i comportamenti di Rocco. Ora che il nome di Enzo Baiocchi, l'assassino della povera Adele, era uscito alla luce del sole, aveva cambiato tono e gli articoli da colpi di sciabola erano diventati incenso profumato. «Immagino lei abbia letto i miei pezzi...». Ettore poggiò la tazzina e la brioche davanti a Rocco. «Sì» rispose il poliziotto. «Forse sono stata un po' dura?». Rocco addentò il cornetto. «I giornali hanno questo di bello. Li puoi usare per raccogliere la cacca dei canarini nelle gabbiette, d'inverno aiutano ad accendere il camino, se uno ce l'ha. Oppure ci si possono avvolgere gli oggetti fragili in vista di un trasloco». «È così che li ha usati?». «No, il mio è stato un trasloco leggero». E bevve il caffè. «Ciao Ettore» lasciò un po' di monete sul bancone e si avviò verso l'uscita. «E grazie per la bellissima sorpresa». «So che non le sono simpatica, ma faccio il mio mestiere». Rocco fischiò a Lupa che scattò da sotto il tavolino e si infilò nella porta a vetri che il vicequestore teneva aperta. «A me i giornalisti piacciono. Sono gli scassacazzi che detesto con tutto me stesso». «C'è la libertà di parola in questo paese, non lo sapeva?». Rocco ancora sull'uscio sbuffò. «È vero. Anche quella di offendere e calunniare? Allora non me ne vorrà se ne approfitto anche io: vada a fare in culo, signora Buccellato. Con permesso» e finalmente chiuse la porta. Ettore rise sotto i baffi, la giornalista con le labbra strette aprì la borsa. «Quanto le devo?». «Pagato, signora». I corridoi della questura sembravano deserti. Rocco, Lupa e Gabriele avanzavano nel silenzio più assoluto. Il ragazzo s'era messo dei jeans larghissimi, una felpa di qualche gruppo musicale e in testa un cappellino da baseball con scritto «Born to raise hell!» dal quale uscivano i capelli lunghi e bisognosi di shampoo. «È scuro qui, mica tanto bello». «È una questura Gabrie', mica un hotel». Arrivarono davanti alla stanza di Rocco. «Eccoci qui». Il cartello con le rotture di coglioni era ancora appeso lì fuori. Il vicequestore sentì il bisogno di aggiungerne una. Prese la penna attaccata a un chiodo che penzolava con uno spago, idea dell'agente Pierron che l'aveva lasciata lì per ogni evenienza, e all'ottavo livello scrisse: Le sorprese. Qualsiasi sorpresa per Rocco era una rottura di palle immensa. Buona o cattiva che fosse, perché la sorpresa era un deragliamento, un ostacolo improvviso davanti alla regolarità della noia esistenziale, un imprevisto che lo costringeva a reagire, a rispondere, a prendere una decisione. E sapeva anche che una sorpresa non arriva mai da sola. Ci ripensò, infatti, e la cancellò dall'ottavo grado per promuoverla direttamente al nono. «Alziamo il livello, Lupa» disse al cane aprendo la porta dell'ufficio. «Il livello di cosa?» chiese Gabriele. «Di tutto ciò che mi dà fastidio e rende la mia vita un incubo. Si parte dal sesto livello, e si arriva al decimo, dominato da una sola rottura di coglioni: il caso da risolvere». Il ragazzo leggeva e sorrideva. «Radio Maria, le comunioni, i battesimi, i matrimoni, i tabaccai chiusi, la sabbia nelle vongole... lei ha aggiunto le sorprese? Perché?». «La sorpresa è un uovo di pasqua, mio giovane amico, portatrice sana di altre rotture di coglioni. E se prosegui a fare domande metto il tuo nome direttamente al livello numero nove». Finalmente il vicequestore e il cane entrarono nella stanza. Lupa scodinzolò, Rocco invece restò congelato. «Che cazzo...?». La stanza era vuota. Sparita la scrivania, sparito l'armadietto, sparita la poltrona di pelle. «Che succede qui?». «L'hanno licenziata?» chiese Gabriele che stringeva il libro sotto il braccio mentre Lupa si aggirava sperduta nell'ufficio spoglio e senza mobili. «Ma porca... Pierron! Pierron!» urlò, ma a rispondere fu solo l'eco della sua voce. «Shto io!» la voce stridula di D'Intino gli fece l'effetto del trapano di un dentista. La presenza dell'agente chiarì una volta per tutte che le rotture di scatole per Rocco non erano terminate lì. «Che succede?». L'agente apparve nel corridoio. «Mi scusi dotto'» si bloccò a guardare Gabriele. «Chi è questo? L'ha arrestato?». «D'Intino, dove sono i miei mobili? La mia scrivania?». «Ha sentito la novità?». «No, D'Intino, non l'ho sentita». «Sì, del gabinetto?». Rocco poggiò le mani sui fianchi. «Quale gabinetto, D'Intino?». «Aspe', si chiama gabinetto provinciale di una cosa che poi finisce con un'altra parola che mo' non mi ricordo». Il vicequestore alzò gli occhi al cielo. «Cosa stai cercando di dirmi?». «Ci sta il gabinetto provinciale, no? Che allora gli serve una stanza». Schiavone con lo sguardo cercò disperatamente un altro agente, Italo oppure Antonio Scipioni, insomma qualcuno che potesse dargli una mano a tradurre l'agente abruzzese, ma i corridoi restavano deserti. Gabriele lo guardava a bocca aperta. «Non ti capisco, D'Intino, e tempo da perdere non ne ho. Adesso fai una cosa, vattene in un ufficio, ripensa a quello che volevi dirmi, scrivitelo su un foglietto e poi quando ci rivediamo me lo leggi. Sai scrivere, no?». «Freghete se so scrivere. Io alle medie facevo i temi! Comunque dotto', mo' ci cambiano il piano. Cioè no, il piano resta questo, ci cambiano l'ufficio. Cioè, non ci cambiano l'ufficio, lo cambiano solo a lei». Rocco lo scansò con il braccio e cominciò a salire le scale. «Zitto D'Intino, non capisco e non voglio capire». «Perché?». Il vicequestore si fermò a metà scala. «Perché per capirti devo entrare nella tua testa, e non è un posto bello da frequentare. Quindi fai come ti ho detto, mettiti seduto e scrivi il messaggio. Tu Gabriele aspettami qui, pare sia sparito il mio ufficio». Dopo due gradini urlò: «C'è nessuno in questo cazzo di questura?» e salì al piano superiore. D'Intino restò accanto al ragazzo. «Tu chi sei?». «Il vicino». «E perché sei qui?». «Ripassiamo latino». «Lo vuoi ripassa' co' me?». «Escluderei». Alla fine della prima rampa Rocco incrociò l'agente Casella che affannato stava portando una decina di faldoni. «Case', che è successo al mio ufficio?». «Ah, sì. Aspetti» poggiò le cartelle sulla fotocopiatrice e si asciugò la fronte. «No, però questa cosa deve essere chiara» disse a bassa voce, stava parlando a se stesso. «Io non posso fare lavori pesanti, guarda, ho già il fiatone». «Soprattutto non hai un buon odore, Casella. Che cazzo, lavati la mattina». «È la caldaia. S'è rotta da due giorni, non riesco a ripararla». «E ricomprala». «Sto andando in giro, cerco un affare. Ci sarebbero le Ferroli che non sono niente male, anche se una Vaillant o una Beretta...». «Casella!» urlò Rocco che stava lì ad aspettare. «Il mio ufficio!». «Signorsì. Allora arriva il gabinetto provinciale della polizia scientifica». «E?». «Il che significa che ci sarà un nuovo sostituto e un nuovo ufficio che il questore ha deciso doveva essere il suo». «Il mio? E io dove vado?». «Sempre a piano terra, ma alla stanza dall'altra parte del corridoio, dopo la rampetta di scale. Vuole che l'accompagno?». «Chi ha spostato i miei mobili?» gridò. Pensava al suo cassetto chiuso a chiave con dentro il sacco di maria nuovo che gli aveva venduto Brizio a Roma. «Io e Deruta. Ma siamo stati attenti, dotto'... allora l'accompagno?». «No! C'è il questore?». «Il questore a quest'ora in questura non c'è» e si fece una risata da solo. «Si ricorda Totò?». Rocco non gli rispose e ridiscese le scale. Trovò Gabriele e D'Intino ancora sull'attenti davanti alla porta del suo ex ufficio. «D'Intino, ora tu riprendi tutti i mobili miei e li rimetti nella mia stanza». «Ma lu questore ha detto...». «Non me ne frega un cazzo del questore! Fai come ti dico!». «Posso aspettare Deruta?» guardò l'orologio. «Quello mo' torna dal panificio di sua moglie. Io da solo non ce la faccio a sposta' li mobili, dotto'». «Gabriele, con me! Lupa, andiamo!». Si incamminò per il corridoio. «Dove va dottore?». «Me ne vado per i cazzi miei, D'Intino!» e voltò l'angolo del corridoio. «Dottor Schiavone, ma che succede?» gli chiese Gabriele che gli trotterellava accanto per tenere il passo. «M'hanno cambiato ufficio. Io odio i cambiamenti!». «Senta, mi dispiace, ma io alla seconda ora devo entrare... non è che...». «Gabrie', stiamo andando in una stanza a fare lezione, va bene?» e spalancò la porta dell'ufficio degli agenti. C'era solo il giovane napoletano, Miniero, di cui Schiavone non ricordava mai il nome. «Tu, Vomero! Esci dalla stanza, mi serve!». Quello senza aprire bocca si alzò di scatto e si precipitò fuori. Mentre Lupa si sistemava accanto alla finestra, Rocco prese una sedia e la avvicinò al tavolo. «Siediti Gabriele e dammi il libro». Il ragazzo obbedì. «Cominciamo! E vedi di rispondere bene che mi girano a trottola!». «È la musica, la musica ribelle». Gianandrea correva canticchiando la canzone di Finardi che ascoltava negli auricolari. Non era il fiato, il problema erano i muscoli delle gambe. Pasta frolla. Polpacci e quadricipiti già urlavano dopo neanche venti minuti di corsa. Ogni tanto sentiva tirare qualcosa dalle parti dell'inguine. Due mesi di fermo per la cuffia dei rotatori li stava pagando cari. Il campo sportivo di Charvensod alla sua destra doveva ancora aspettare. Gli mancavano i suoi ragazzi, gli mancava il campo, la sfida, il sudore, lo spogliatoio, gli odori delle pomate, delle docce, ma finché non tornava in forma non sarebbe potuto rientrare ad allenarli. Erano finiti terzi in campionato, un ottimo risultato se si pensa che all'inizio avevano esordito con tre sconfitte di fila. E comunque andavano meglio della prima squadra che giocava in promozione e aveva chiuso l'anno in sedicesima posizione. «Che ti dice di uscire e di metterti a lottare...». Era sempre stato così per Gianandrea, da quando era nato 40 anni prima. Mettersi a lottare. Non ricordava un momento della sua vita in cui per ottenere un risultato, anche mediocre, non avesse dovuto combattere con tutte le sue forze. A scuola, sul campo di calcio, con la prima moglie, con la seconda e con i due figli maschi. Gli altri sembravano superare ostacoli con una facilità impressionante, lui invece no. Già far capire che il suo nome si scriveva con una enne e non con due. E il cognome era anche peggio. Appena usciva dalla Valle doveva fare lo spelling, che gli italiani già in Piemonte non lo capivano. Marguerettaz. Diventava Margherittà, o Marchettaz o Margarinaz. Il suo corpo poi, un problema perenne. Aveva subito tre operazioni alla gamba destra, due alla sinistra. S'era rotto clavicola, setto nasale, gomito e due incisivi. Una costellazione di fratture e incidenti che avevano minato la sua carriera di calciatore e lo avevano costretto a passare metà del suo tempo dentro gli ospedali o nei centri fisioterapici. Ma ora non mollava. «Che ti dice di uscire e di metterti a lottare...» suggeriva ancora Finardi, ed eccolo ancora lì a 40 anni suonati a recuperare l'ennesimo ricovero per poter ritornare in campo, dai suoi ragazzi, a lottare. Si voltò verso la Dora che scorreva alla sua destra. I raggi del sole si riflettevano sull'acqua e sulla piccola rapida che schiumava a pochi metri da lui. Sulla riva, c'era un mucchietto di stracci colorati. Rallentò il passo. Non era un cencio, erano un paio di jeans corti, una camicia a fiorellini rossa dalla quale spuntavano la testa immersa sotto l'acqua e le braccia protese in avanti. Nessuno nuota nella Dora. «E le strofe languide di tutti quei cantanti...» si strappò gli auricolari dalle orecchie. «Ma che...?» superò la staccionata e cominciò a scendere verso il greto del fiume. Non sentiva più il sudore, l'inguine non tirava più. Solo il cuore nelle orecchie. Per poco non prese una storta su una buca nel terreno. Poi arrivò sulla riva. L'acqua in mezzo alle pietre cominciò a filtrare nelle scarpe da corsa. Il corpo era lì, a pochi metri da lui, a faccia in giù. «Ero ers ert...». «Ma che stai a di', Gabriele! Indicativo imperfetto di sum! Ma quale ero ers... forza dai. Eram eras erat... poi?». Gabriele deglutì: «Erasmus...». «Erasmus?». «Prima persona plurale, no?». «Eramus, che erasmus!». «Ah! Eramus, erastis». «Eratis, poi?». Gabriele si morse le labbra. «Dai ce la puoi fare... terza persona plurale, eramus, eratis...?». «Errant?» buttò lì il ragazzo. «Ma vaffanculo, Gabriele!». Rocco gettò via il libro di grammatica latina. «Non sai una mazza. Errant! Ma dimmi tu, sei di un'ignoranza allucinante. Erant, che errant! Ma perché hai fatto il classico?». «Mamma dice che apre la testa». «A te per apritte la testa ce vo' un'ascia!». Rocco si alzò scostando rumorosamente la sedia. «Mi dispiace, ti rifai l'anno». Gabriele chinò il capo. «Mamma mi ammazza!». «Ma perché mammina s'aspettava una promozione?». «Sì» e triste tirò fuori una merendina dalla tasca della felpa. «E non s'è mai accorta che a scuola sei una pippa?». «No» con il primo morso strappò metà dolcetto. «Ammazza, basta sentirti una volta e ci si accorge che sei di un'ignoranza crassa». «Già, una volta basterebbe» e con un morso divorò anche l'altra metà. «Gabriele, adesso te ne vai a scuola». «E che ci vado a fare?». «Ti prendi le tue responsabilità». Gabriele alzò le spalle. «Tanto ci vado o no, mi bocciano lo stesso. A questo punto io resto qui». «No, Gabrie', qui no. O vai a scuola o te ne torni a casa. Questa è una questura, mica un kinderheim». Gabriele si arrese. «Va bene. Vado a scuola. A fare l'ennesima brutta figura». Rocco si alzò. «Ora ti faccio accompagnare». «Con la macchina?». «Perché no?». «E posso mettere la sirena?» il ragazzo aveva già cambiato umore. «Se solo ci provi ti faccio passare due notti in guardina». «Ricevuto» recuperò il libro di grammatica latina e si rimise il cappellino da baseball. «Sono pronto». «E di' a mamma che uno di questi giorni mi piacerebbe conoscerla». Il ragazzo sorrise sornione. «Non è male mia madre. Vuole vedere una foto?». «Ma che hai capito? Ci voglio fare una chiacchierata». «Le vuole fare una ramanzina?». «Sei scemo? Tua madre si fa un mazzo così e deve sapere come la ripaghi. Muoversi». Fuori dalla stanza degli agenti incrociò Antonio Scipioni: «Uè Antonio, fammi una cortesia. Ti dispiace portare questo ragazzo in odore di Nobel a scuola?». «E certo, per un premio Nobel però ci vorrebbe la scorta». «Sì, ma è qui ad Aosta per motivi privati. Insomma profilo basso». «Che credete che non ho capito che mi state prendendo in giro? Tanto lo fanno tutti. Vabbè, grazie dottor Schiavone, almeno lei ci ha provato!» e con la testa bassa seguì l'agente Scipioni verso l'uscita della questura. «Erasmus... errant, ma come si fa!» mormorò il vicequestore. «Gabriele!» lo richiamò. Il ragazzo si voltò. «Che c'è?». «Vieni qua!». Sbuffando il ragazzo si avvicinò al vicequestore. «Guarda che poi si cresce e tutto questo finisce, sai?». «Lo spero proprio. Non vedo l'ora. Arrivederci». Rocco lo guardò andare via, col passo lento e la testa incassata nelle spalle. «Capito Lupa? Ora diamoci da fare. Casella!» gridò. Doveva risolvere la questione del suo ufficio. Era sveglio da ore e non aveva ancora fatto un tiro di maria. Sentiva le articolazioni bloccate, il motore grippato, senza lubrificante la macchina non si sarebbe mossa. «Casella!». Invece dal corridoio apparve Italo Pierron pallido in viso. «Dottore?». «È un'ora che ti cerco. Il mio ufficio...». «C'è una cosa urgente. Sulla riva della Dora». «No!». Rocco sgranò gli occhi. «No. Il tempo è bello, il sole splende, una rottura di coglioni di decimo livello no!». Italo allargò le braccia. «Casella e D'Intino sono già sul posto. Abbiamo chiamato Fumagalli». Era il suo giorno libero e Caterina aveva deciso di sacrificare la mattinata alle incombenze domestiche per poi mettersi a lavorare sulla tesi. Le bollette si erano ammucchiate e, per una curiosa compensazione, il frigo svuotato. Scaduta l'assicurazione dell'auto che non usava mai, avrebbe dovuto anche fare un salto all'Aci per il bollo. Stesa sul letto, mezzo vestita, non trovava la forza per affrontare quella giornata. Tutti gli arretrati accatastati potevano far pensare a una ragazza disordinata e distratta, ma bastava dare un'occhiata alla casa per accorgersi dell'esatto contrario. Piccola e accogliente come un rifugio di montagna, niente era fuori posto. Le tende in tono con il divano, tre piccoli quadri discreti e ben realizzati risaltavano sotto delle applique in ottone. La carta da parati a righe, il profumo di lavanda che si spandeva da una boccetta con dei bastoncini di legno infilati dentro, tutto era curato al millimetro. Nell'unica libreria a sette ripiani i libri erano sistemati per case editrici. I cd se ne stavano infilati alla destra del piccolo stereo compatto e i dvd in uno sportello sotto il televisore a 23 pollici. L'angolo cottura del bilocale era uno specchio. Sembrava pronto per una ispezione dei Nas. Il calendario perpetuo con i personaggi di Pinocchio segnava la data esatta, l'orologio dell'Ikea di alluminio spaccava il minuto. Lo spremiagrumi e il frullatore in un angolo, i piatti nello scolapiatti di legno accanto al lavello erano in ordine di grandezza, e i canovacci appesi alla maniglia del forno sembravano appena usciti dalla lavatrice. Ad entrare poi nel piccolo bagno con doccia si aveva la sensazione di essere in un albergo appena sterilizzato per un nuovo cliente. Creme, profumi, trucchi erano schierati come soldati in parata sul lavabo. Dall'altra parte spazzolino dentifricio e filo interdentale. Nei cassetti del mobile del lavabo c'erano dei dischi struccanti, una matita e un pennello per il trucco, qualche medicinale, tutti infilati in graziosi cestini in midollino. Ci teneva all'ordine, il viceispettore Rispoli, e che tutto fosse al proprio posto e che obbedisse a un criterio preciso. Detestava le case approssimative, quelle vuote e coi quadri che pendevano da una parte. Solo due elementi stonavano con la grazia di quell'appartamento: la Beretta di ordinanza infilata nella fondina agganciata a un omino appendiabiti in ferro battuto e il maglione che Italo aveva lasciato appallottolato una settimana prima e non aveva più ripreso. Si alzò di scatto dal letto. Infilò i jeans, le scarpe da ginnastica nere, accese il cellulare che mise in tasca, prese portafogli e chiavi di casa e uscì. «Buongiorno signora Cormet» disse alla donna che stava rientrando dal mercato. «Buongiorno Caterina». «Vado a fare la spesa. Serve qualcosa?». La vecchina aprì la porta. «No tesoro, niente di niente. Grazie». Fece per entrare, poi ci ripensò. «Caterina?». La ragazza si fermò sulle scale e si girò: «Mi dica». «Me la togli una curiosità?». «Certo!». «Perché sorridi?». Caterina impreparata non seppe cosa rispondere. «Non saprei...». «Sei felice? Sei serena?». La ragazza ci pensò su. «Non lo so. Sorrido perché... non c'è una ragione». «Sorridi con la bocca ma gli occhi sono tristi». Il viceispettore chinò appena il capo. «Secondo lei cosa mi manca perché anche gli occhi sorridano?». «Niente. Per questo non capisco». «Meno male, signora Cormet, temevo che anche lei cominciasse con la predica che dovrei trovarmi un fidanzato». «E perché? Perché la gente pensa che noi possiamo realizzarci solo se ci completiamo con un uomo? No, ragazza mia, no. Quello che ti dice la tua vecchia vicina è di sbarazzarti delle cose passate. Sono passate e non devono pesare sull'oggi». «Lei ha una soluzione?». «Io a 82 anni ci sto ancora lavorando. Buona giornata!» e finalmente la donna entrò in casa. Caterina restò pensierosa sulle scale, poi uscì. Stava scegliendo gli spaghetti, poi sarebbe passata ai detersivi, quando la vide davanti al reparto surgelati. Era senza carrello, senza cestino. Si guardava intorno coi suoi occhi a palla e un cespo di capelli spettinati. Una maglietta lunga di un azzurro scolorito, pantaloni della tuta. La pelle bianca, lattiginosa, un paio di occhiali legati a una catenina tempestata di perline le pendevano dal collo. Cosa ci fa ad Aosta? si chiese Caterina mentre guardava distrattamente un pacco di rigatoni. Rapida si nascose dietro la scansia delle paste. Poteva passare attraverso il settore marmellate e, girando dopo le acque minerali, avrebbe guadagnato la cassa e poi l'uscita. Ma doveva lasciare la spesa a metà, abbandonare il suo carrello pieno di prodotti e uscire dal supermercato, e non era un comportamento civile mollarlo lì, in mezzo al corridoio. Si sentiva in colpa per i commessi che avrebbero poi dovuto rimettere tutto a posto. Si affacciò lenta. Sua madre era sparita dal reparto surgelati. E adesso? pensò. Dov'è andata? Si convinse che era una causa di forza maggiore e mollò il carrello, per dirigersi con circospezione verso le marmellate. Non era neanche in quel corridoio. E nemmeno alle sue spalle. Quasi in punta di piedi percorse il reparto vini e liquori. In fondo c'erano solo le casse. Di sua madre neanche l'ombra. Sorrise alla commessa impegnata con una cliente e mostrando le mani, a dire non ho comprato nulla, guadagnò l'uscita. Non era neanche in strada. Pensò allora di essersela immaginata, ma non era possibile. L'aveva vista, in carne e ossa, un cencio umano pallido e ingrassato. Inseguita da quell'ombra fugace si dileguò fra le strade di Aosta svoltando più angoli possibile. Tanto sa dove lavori, se vuole sa dove trovarti, pensò. «Per favore, non c'è niente da guardare». Casella cercava di tenere lontano le persone che già si erano assiepate per godersi lo spettacolo. «Oh, avete sentito? Non ci sta niente da guarda'!». D'Intino aveva fermato un omone senza capelli e col viso butterato che lo sovrastava di almeno mezzo metro. «Sono un giornalista...». «Se è un giornalista può passa', Casella?» chiese l'agente abruzzese al collega. «No, D'Inti'! Uno può essere pure Gesù Cristo a volo radente, di qui non si può passare!». Gianandrea se ne stava seduto sul sedile posteriore dell'auto della polizia. Accanto Miniero gli passò una sigaretta, ma l'allenatore dei ragazzi della Polisportiva Cogne la rifiutò, poi si girò attratto dal rumore degli pneumatici di un'auto con la lucciola sul tetto. Vide scendere un agente e un uomo con la faccia spiegazzata. «Il tipo in calzoncini è quello che ha trovato il cadavere, mi sa» disse Italo indicando con un gesto del mento Gianandrea accanto all'agente napoletano. Rocco non rispose. Non ce n'era bisogno, bastava osservare gli occhi vuoti e spaventati e il pallore del viso. Il furgone della scientifica era fermo accanto alla recinzione del campo sportivo. «Mi ricordi invece come si chiama l'agente del Vomero?». «Miniero» rispose Italo. Due agenti si stavano infilando le tute bianche. «So' già arrivati». Rocco li indicò con il mento. «E certo Rocco, sono i nostri. Hanno messo il gabinetto provinciale, non ci sarà più Farinelli». «Sono quelli che m'hanno fregato la stanza?». «Esatto». «Ecco, già mi stanno sul cazzo». Il vicequestore e Pierron si fecero strada attraverso i curiosi e cominciarono a discendere il piccolo pendio verso il fiume. Rocco si appoggiava a Italo, le suole delle Clarks facevano poca presa sul terreno erboso. «Vai piano, Italo». «Eccolo lì, Rocco». Rocco guardò il cadavere a mezzo metro dalla riva, per metà infilato nell'acqua a testa in giù. Fumagalli era lì vicino con un paio di stivali da pesca alti fino all'inguine. Con lui c'erano due agenti. «Vedi il patologo? Lui si attrezza!». Rocco guardò i piedi di Italo. «Sono anfibi quelli?». «Certo! Quelli in dotazione e...». «Quanto porti?». «44». «Levateli». «Eh?». «Togliti gli scarponcini, è un ordine del tuo superiore». Italo sedette e si sfilò gli stivaletti. Li passò a Rocco che intanto s'era liberato delle Clarks. Infilò gli anfibi di Pierron e si tirò su i pantaloni al ginocchio. «Mi posso mettere le tue scarpe?». Rocco lo guardò freddo. «Le mie Clarks? Sei impazzito? Stattene qui tranquillo, tanto tu vicino al cadavere non ci vieni». L'agente rimase in calzini in attesa vicino ai sassi della riva. Rocco entrò nell'acqua bassa mentre due agenti portavano il corpo all'asciutto. «Bene così ragazzi, bene così» fece Rocco. Lo girarono. «Oh Madonna...» disse quello più giovane. «Va bene ragazzi, potete andare» e i due poliziotti sollevati dallo stare in compagnia di quel cadavere fuggirono quasi di corsa. «E che è successo?» chiese Rocco a Fumagalli una volta soli e coi piedi in un palmo di acqua. Il viso del cadavere era gonfio. Ma la cosa che più colpiva erano le labbra. Sembravano due salsicce. Gli zigomi sporgenti parevano nascondere sottopelle due mandarini, al contrario del naso che era innaturalmente piccolo e schiacciato. Fumagalli sbottonò la camicia a fiorellini. Al terzo bottone apparve la stonatura. Il cadavere aveva due seni grossi e violacei. Fumagalli si chinò a osservare il corpo nudo. Cominciò ad esaminare gli occhi e con l'ausilio di una penna cercò di aprirgli la bocca. Poi abbassò di poco gli shorts. «Abbiamo un M to F». «Un trans?». «Direi di sì». Il petto giallo cera non presentava neanche un pelo, come non c'erano tracce di barba sul viso. «La poverina aveva cominciato il percorso chirurgico. Mammelle e viso li aveva già affrontati. Probabilmente in attesa di togliere il pene» e lo indicò con la penna. Era piccolo e scuro, quasi si vergognasse di stare lì. «Ha delle ecchimosi congiuntivali, e poi guarda il collo. Mi gioco mille euro che è morta strangolata». Era visibile intorno al collo un solco violaceo. «Me la porto in sala, ma non sbaglio». «Ci credo. Secondo te quanti anni ha?». «Una trentina». Rocco si chinò ad osservare la camicia del cadavere. «Deve essersi rivestita di corsa. Ha sbagliato i bottoni, vedi? Ne avanza uno qui in fondo». Si accese una sigaretta. «E non s'è neanche messa le mutande». «È importante?». «Meditiamo su questa cosa... Le scarpe?». «Quelle può averle perse in acqua. Magari aveva delle ciabatte?». «Senza calzini, è chiaro». Sollevò una mano. Le unghie erano dipinte di verde, come quelle dei piedi. «Non ha anelli, bracciali, niente». «Già...». Rocco si rialzò. «Che rottura di coglioni. Io me ne torno in questura. Solito giro con i capi. Stammi bene, Alberto». «Ah, l'hai conosciuta la nuova della scientifica?». «No. Interessante?». «Le novità di questo porto? O piove, o tira vento o sona a morto!» e tornò a esaminare il cadavere. «Che ci dice, vicequestore?». Oltre al giornalista pelato e col viso butterato se n'era aggiunto uno riccetto e scattante. «E che vi dico? Che anche l'anno prossimo la Roma vincerà lo scudetto» e passò oltre. I due desistettero dal fare altre domande, Schiavone non aveva mai dato risposte ai giornalisti se non nelle rare conferenze stampa alle quali il questore lo aveva costretto a partecipare. Un'automobile verde avanzava zigzagando sulla strada accanto al campo sportivo. Mangiata dalla ruggine, il cofano ammaccato, la cinghia di trasmissione che urlava mentre le ruote sgonfie producevano strani tintinnii metallici. Il mezzo salì con due scossoni sul marciapiede e dopo un rantolo arrestò la corsa. Cigolando la portiera si aprì. Scese una donna sulla quarantina. Afferrò una vecchia borsa di pelle e con un calcio chiuse lo sportello. La borsa cadde a terra e ne uscirono piccoli contenitori di plastica multicolore. La donna alzò gli occhi al cielo e cominciò a raccoglierli. Rocco si perse ad osservarla. I capelli lisci e scuri sfioravano le spalle. Aveva il viso di una bimba, e gli occhi neri, grandi e vivaci. Portava una gonna al ginocchio e un paio di anfibi. L'occhio clinico di Rocco riuscì a scorgere il corpo magro e atletico e il seno abbondante nonostante un maglione di tre taglie più grande cancellasse ogni forma. Finalmente la donna aveva radunato le sue cose. Proteggendo la borsa con un braccio neanche fosse un neonato si avvicinò. Si fermò davanti a Rocco. «Salve!». «Salve». «Mi sa che lei è il vicequestore Schiavone, giusto?» aveva l'accento siciliano e sorrideva senza mostrare i denti. Allungò la mano tenendo con l'altra in bilico la borsa: «Michela Gambino, commissario scientifica...» e la borsa cadde nuovamente a terra. Rocco la aiutò a recuperare i contenitori che s'erano sparpagliati per strada un'altra volta. «Madonna, 'sta borsa la devo cambiare...». «Lei è quella nuova?». «Eh sì. Oggi comincio. Grazie» si rialzarono, ma la Gambino aveva lasciato a terra la sacca. Finalmente si strinsero la mano. L'aveva fredda e con le dita affusolate. Niente fede, un solo anello piccolo senza pietra all'anulare. «Cos'abbiamo giù al fiume?». «Un trans morto strangolato». «M'è già capitato, due anni fa a Torre. Quando ero ispettore superiore». «Annunziata?». «Del Greco. Ma io sono di Palermo». «Io di Roma». «Sì, lo so. So tante cose di lei. E senta un po', ci diamo del tu? Mi sa che dobbiamo lavorare insieme». «Per me va bene, Michela». «Ecco, pure per me, Rocco». Poi alzò gli occhi al cielo. Sul volto comparve un sorriso professorale. Anche Rocco sollevò la testa. Il cielo era azzurro. Qualche nuvola. Una striscia chiara, l'orma di un aereo. «Non finirà mai» fece la donna. «Cosa, scusa?». «La vedi quella traccia bianca lasciata da qualche Boeing?» poi tornò con lo sguardo sul vicequestore. «Scie chimiche» disse. «Come?». «Sono scie chimiche» abbassò la voce di un paio di tacche. «Lo fanno dal 1997. Sparano sostanze chimiche nell'aria e tramite agenti psicoattivi prima o poi ci controlleranno tutti». «Non ho capito. Chi le spara?». «I potenti. Sai in quanti comandano il mondo? Sono 300. Il club dei 300. Poi con calma ti dirò anche chi sono». «Sì con calma, anche perché 300 è lunghetta...». «E io sono convinta, Rocco, che questo...» e alzò l'indice al cielo, «è anche un buon metodo per il controllo delle nascite. Ti sei mai chiesto perché George Soros vive su un'isola dove non passa nessuna rotta aerea?». «Guarda, Michela, no, non me lo sono mai chiesto. Ora però lo farò». Il commissario della scientifica scosse la testa. «Hai mai sentito parlare di HAARP?». «No». «E di sandwich elettroconduttivo?». «Mi dispiace Michela, neanche». «Vivi senza sapere niente, vicequestore. Tempo venti anni saranno riusciti a ridurre la popolazione da 7 miliardi a poco più di 500 milioni. Resteranno i ricchi, i potenti, quelli che contano e una manciata di schiavi». Si chinò per recuperare la borsa e superò Rocco. Dopo due passi si voltò. «Come si sta ad Aosta?». «Non saprei» fece Rocco «Ah già, niente sai. E non è che tu e i tuoi uomini avete toccato e lasciato impronte?». «Assolutamente no, Michela, anche perché il morto sta in mezzo all'acqua. Difficile lasciare impronte». «Lo dici tu!». «Senti, mi togli una curiosità? Che marca è la tua auto?». «Una ZAZ Zaporozhets 968. Niente che trovi in Europa. Niente di rintracciabile» e gli fece l'occhiolino. «Amunì» urlò ai due agenti con la tuta bianca e finalmente si avviò verso il fiume. Rocco restò in piedi finché la donna venne inghiottita dal gruppetto di curiosi. Poi allargando le braccia si diresse verso Pierron che lo aspettava accanto all'auto con le Clarks in mano. «Mi ridai gli anfibi?». Rocco cominciò a slacciarli. «Quella è il sostituto della scientifica?». «Già». «E com'è?» chiese Italo. «Se nun so' strani ad Aosta non ce li vogliamo!». Aveva trovato il suo nuovo ufficio, una stanza che somigliava molto a un ex deposito delle scope. Una sola finestra stretta in fondo e la scrivania entrava per miracolo. Per andarsi a sedere bisognava scavalcarla. Lupa l'aspettava lì ai piedi di un armadietto di ferro che doveva fungere da schedario e che invece Rocco sospettava contenesse un mocio un secchio e degli stracci. Si scaraventò ad aprire il suo cassetto personale chiuso a chiave e tirò fuori la canna che si era preparato la sera prima. L'accese. Andò ad affacciarsi alla feritoia. Con orrore si accorse che dava sulla chiostra interna del palazzo della questura e dai finestroni del corridoio centrale tutti potevano sbirciare dentro la sua stanza, meglio, dentro il suo deposito delle scope. Infatti D'Intino stava bussando sul vetro di fronte per attrarre la sua attenzione e lo salutava con la manina. Rocco non ricambiò il saluto. Rientrò nella stanza. «No Lupa, da qui ce ne dobbiamo andare e di corsa». Dal cassetto prese la scatola di biscottini alla carne «Amore!» e ne lanciò uno che la cucciola acchiappò al volo. Cominciò a fissare il telefono poggiato sul piano di legno sotto la lampada verde. «Allora diciamo che entro 10 secondi mi chiama? Uno, due, tre...» il telefono squillò. «Niente, Lupa, l'ho sottovalutato». Alzò la cornetta. «Schiavone, dica, dottore...». Dall'altra parte della linea c'era il questore. «Come faceva a sapere che ero io?». «Sono un poliziotto, ricorda?». «Ogni tanto fa bene a rammentarmelo. Venga a riferirmi i fatti». «Stavo appunto salendo. A proposito, c'è anche una cosetta di cui le vorrei parlare». «Di che si tratta? Deve andare a Roma, immagino». «No. Si tratta della stanza». «Venga su e ne parliamo». Attaccò. «Lupa, stattene buona un altro po', devo salire». Gettò la canna a terra, la schiacciò con il piede, poi la raccolse e la buttò nel cestino. Annusò l'aria. Scosse la testa e andò ad aprire la finestra. «Manco le tende, 'sti purciari». In corridoio si accorse che qualcuno, forse Italo, aveva spostato anche il cartello delle rotture di scatole. Il trasloco aveva l'aria di essere definitivo. Lo punivano? Forse sì. La storia di Roma, di Enzo Baiocchi, l'interrogatorio subito dal questore e dal magistrato avevano avuto la prima conseguenza. L'avrebbero messo a fare le fotocopie, se non fosse che lì dentro era l'unica testa pensante e la macchina fotocopiatrice funzionava a giorni alterni. Magari si stanno dando da fare per trasferirmi, pensava mentre saliva le scale. E allora? La cosa non lo sconvolgeva più di tanto. Non s'era legato a quella città, neanche ai suoi colleghi. Spesso si era domandato se si sarebbe mai più affezionato a qualcuno. Italo Pierron? Un caro ragazzo, gli voleva bene, ma niente di più. Antonio Scipioni, un bravo poliziotto, intelligente, ma restava un collega di lavoro. Gli altri agenti, il questore Costa, Baldi, neanche a parlarne. E nemmeno Fumagalli, troppo diverso, un altro pianeta. C'era Caterina. Caterina meritava un discorso a parte. Gli era entrata dentro piano piano, come fanno certi virus che all'inizio sono poco più di un'influenza, poi ti debilitano a tal punto che non ce la fai neanche ad alzarti dal letto. E dire che a settembre, quando l'aveva conosciuta, era solo una ragazza carina con la quale avrebbe volentieri avuto qualche scambio più intimo di un buongiorno e un buonasera sparati a mezza bocca fuori dall'orario di lavoro. Poi si era accorto che Caterina aveva una marcia in più, e ogni volta che quella aveva il giorno libero o stava poco bene lasciava un vuoto. Per fortuna era fidanzata, proprio con Italo, il che rendeva tutto più semplice. Gliela toglieva dalla testa. Per Rocco la donna di un amico, o di una persona per la quale provava un po' d'affetto, era intoccabile. Era una suora, un prete, un monolite di marmo. Non la vedeva più come una donna da sedurre. C'era stato quel mezzo bacio, è vero, quando le aveva raccontato della morte di Marina. Ma era sull'onda della commozione, dei ricordi, che gli stringevano il petto e la trachea spezzandogli il respiro e togliendogli la capacità di pensare. L'avessero pure spedito in un'altra città. Per lui che si trascinava la vita sulle spalle come una pesante fascina di legna era niente. Il trasloco? Due valigie e un cane. A questo pensava bussando alla porta del questore. «Avanti!». «Eccomi dottore...». «Si accomodi, Schiavone». Costa era seduto alla scrivania intento a leggere il giornale. «Sembra che la mia ex moglie abbia completamente cambiato tono nei suoi confronti. Ha letto?» e gli passò il giornale. C'era un pezzo a firma Sandra Buccellato. Il titolo era «Roma ci ha fatto un bel regalo». Schiavone buttò uno sguardo distratto, era un articolo che elogiava le qualità investigative del vicequestore e insisteva sulla certezza che presto Enzo Baiocchi, l'assassino di Adele Talamonti, la povera fidanzata di Sebastiano Cecchetti, sarebbe stato assicurato alla giustizia. «A proposito» disse Costa riprendendo il giornale che Schiavone gli restituiva, «lei si sente al sicuro? Questo Baiocchi è ancora in giro» e gli fece cenno di accomodarsi. Rocco alzò le spalle. «Io mi sento sicuro. Se vuole tornare, sa dove trovarmi». «Sprezzo del pericolo?». «Noia». Costa lo guardò scuotendo la testa. «E a proposito di noia, posso sapere per quale motivo mi avete sbattuto in quell'ex deposito di scope?». «È una cosa temporanea. Dobbiamo fare spazio al gabinetto della scientifica e...». «Sì, il motivo ufficiale l'ho capito. Si dovrà abituare a una squadra mobile a scartamento ridotto. Cominciamo subito? Allora, Plan Félinaz, vicino al campo sportivo. Abbiamo un cadavere trovato nel letto del fiume, arenato sulla riva. Si tratta di un trans...». Il questore ascoltava con estrema attenzione. «A una prima analisi è sulla trentina. Sulle cause della morte è un po' presto, ipotizziamo strangolamento». Costa si lasciò andare sulla poltrona. «Non è affogato?». «Non credo, ma presto Fumagalli ci deluciderà». «Sappiamo chi è?». «Non lo sappiamo, dottore. E prevedo una serie di rotture di palle per capirlo». Baldi sbuffò. «Dobbiamo andare a cercare in quell'ambiente» fece il magistrato sedendosi alla scrivania. «Quale ambiente, se è lecito?». «Prostituzione. Ha mai sentito di un trans che si guadagna da vivere diversamente?». «Sì» rispose Rocco. «Io no. Comunque cerchiamo di risalire almeno all'identità. Quanti trans ci saranno ad Aosta?». «Perché, lei è sicuro che sia di qui?». La fotografia della moglie di Baldi era di nuovo sparita dalla scrivania. «Non sono sicuro, ma dal momento che lei dice fosse vestito semplicemente, insomma da casa, capace abitasse qui». «O era venuta a trovare qualcuno ed era ospite». Baldi si alzò dalla sedia. Andò alla finestra. «Riferendoci al trans dobbiamo usare il maschile o il femminile?». «Io uso il femminile». «Perché?». «Perché non era un travestito. Era un trans. Aveva il seno, insomma credo ci fossero buone possibilità che si stesse preparando per l'operazione finale. E su questo voglio cominciare a indagare. Se era in attesa della chiamata, qualcosa negli ospedali si dovrebbe sapere». «Mi sembra un'ottima idea. Le dà fastidio se io continuo a usare il maschile riferendomi alla vittima?». «Faccia come meglio crede». «Sa, per me chi nasce maschio resta maschio, non credo molto ai cambiamenti radicali. La sa la storiella dello scorpione e della rana?». «La conosco». «La natura più profonda, quella non la cambi mai!». «Uno può avere una natura femminile incastrata nel corpo di un maschio». Baldi scosse la testa, chiuse la finestra e tornò alla scrivania. «Le piace il suo nuovo ufficio?». Rocco sorrise. «È bellissimo, ma sbaglia a chiamarlo ufficio. Più appropriato sarebbe deposito delle scope. Ma almeno la finestra, meglio la feritoia, dà nel cortile interno, e non vedo le montagne». «Immagino si sarà chiesto perché lei è stato trasferito lì». «Non me lo sono chiesto. Lo so». Baldi puntò gli occhi in quelli del vicequestore. «Credo che lei sia appeso ad un filo». «Ci sono abituato». «Perché non mi dice la verità su Enzo Baiocchi? Perché ce l'ha con lei? Cosa è successo fra voi due? Schiavone, io le sto dando un'ultima possibilità». «Come già ebbi occasione di dirle, lo scopra da solo, dal momento che non mi crede. Ora se non le dispiace avrei da lavorare». Rocco si alzò. «Lei pensa che io la detesti, non è così?». «Lei fa il suo lavoro. È uno che ci crede. Anzi, più di una volta ha chiuso un occhio». «Allora perché non vuole essere collaborativo e evitiamo incidenti futuri?». Il vicequestore non rispose. Aprì la porta, poi si voltò verso il giudice. «La terrò informata sugli sviluppi. Mi saluti sua moglie». Voltò le spalle e il magistrato ne approfittò per sparare l'ultima cartuccia, la più potente: «Com'è morto Luigi Baiocchi, il fratello di Enzo?». «Credo in Sudamerica. Altro non so». I due uomini si guardarono. «Cosa aveva a che fare con lei?». «Ancora? Gliel'ho detto. Era un criminale di mezza tacca. Lo inchiodai per un traffico di cocaina tanti anni fa». «E lei mi rinfreschi la memoria». Rocco alzò gli occhi al cielo e allargò le braccia. «Va bene. Nel 2007 scoprimmo un traffico di cocaina grazie all'omicidio di due ragazzi che s'erano messi a rubare la droga ai trafficanti. Il capo aveva un mobilificio fuori Roma, scoprimmo che si faceva arrivare la coca dal Sudamerica attraverso spedizioni di mobili. La droga era nascosta, meglio, era stata trasformata in finte statuine precolombiane provenienti dall'Honduras. Così in una perquisizione al porto trovammo il carico. Ci fu una sparatoria, un paio di morti, un agente ferito e alla fine inchiodammo la banda. Ma scapparono Luigi Baiocchi e gli altri componenti della banda, e soprattutto non risalimmo mai ai veri capi del traffico, perché sospettavamo, e ne sono tuttora convinto, che dietro c'era qualcuno di molto potente che muoveva tutta la santabarbara». «E quel Luigi Baiocchi, il fratello di Enzo, lei dice che è morto da qualche parte in Sudamerica». «Così venni a sapere». «Non trova curioso che ora spunta fuori il fratello evaso dal carcere per farle la pelle? Andiamo, Schiavone, non offenda la mia intelligenza». «Il giorno che metteremo le mani su Enzo Baiocchi lei avrà modo di chiederglielo e così sapremo la verità. Mi creda, sono curioso anche io». E finalmente Rocco uscì. Tornando verso la questura si era perso a guardare una vetrina. Buttava un occhio sui prodotti esposti senza interesse, non si era neanche accorto che trattasse abbigliamento femminile. Era diventata una situazione insostenibile, lo sapeva. Il cerchio si era stretto, solo questione di tempo e avrebbero scoperto la verità, che era nascosta insieme a Luigi Baiocchi sotto quintali di cemento in un villino a schiera nel quartiere dell'Infernetto a Roma. Gli tornò alla mente quella notte del 2007. La puzza di urina del capannone, la luna e le stelle che facevano capolino fra i vetri rotti. Lo sparo. Luigi Baiocchi steso a terra con un foro rosso in mezzo alla fronte. La bocca della sua pistola che lo guardava per l'ultimo colpo, quello decisivo, quello che avrebbe messo fine a tutto. Poi vide il suo viso riflesso nella vetrina del negozio. Scavato, gli occhi stanchi, i capelli spettinati. Un'ombra gli passò accanto e per un momento si riflesse nel vetro. Una figura bianca, veloce come un cane che fugge, ma che dietro si era lasciata un profumo che Rocco conosceva bene. Si voltò. La strada era deserta. «Che vuoi dirmi Marina?» disse a bassa voce. Tornò a specchiarsi. Curioso, prima non s'era accorto della barba lunga di qualche giorno e di un ciuffo di capelli sulla fronte che si era imbiancato. Riprese a camminare. Incrociò una bambina con le guance piene di lacrime e un gattino in braccio, un uomo dal passo veloce e gli occhi preoccupati e un ragazzo coi capelli lunghi e un giubbottino rappezzato che si mordeva le labbra. E si trovò a pensare che ogni essere umano, dal più giovane al più vecchio, s'è impantanato da qualche parte, e lotta, nella stanza segreta del suo cuore, per guarire un gatto malato, per non perdere un appuntamento, per prendere in mano la propria vita. Alzò gli occhi. Il sole era sempre lì, come il cielo. E mai come in quella mattinata di inizio estate Rocco Schiavone si sentì solo. Neanche Marina veniva più a trovarlo. Passava veloce, come la gente nella piazza, e non si fermava più da lui. Caterina non trovava le chiavi di casa. Niente nella tasca sinistra dei pantaloni, niente nella destra, niente in quella anteriore del giubbotto. Eppure aveva un posto per tutto. Per le bollette, le ricevute, gli scontrini, le medicine. Le chiavi invece erano indomabili. Sembravano vivere di vita propria. Da sempre giocavano a nascondino fra tasche borse e cassetti e non c'era verso di dargli una disciplina. Mi saranno cadute al supermercato? pensò infilando la mano nella tasca sinistra anteriore del bomber, dove aveva già rovistato per ben due volte. Poi le dita toccarono qualcosa di metallico. Erano lì, nascoste dentro la fodera. Il viceispettore con qualche capriola delle dita riuscì a tirarle fuori. «Ciao...» una voce alle sue spalle la freddò. Si voltò. Sul marciapiede c'era sua madre. La donna avanzò di un paio di passi. Caterina doveva mantenere la calma, non cedere, non farsi venire rimorsi e trattarla con distacco, con freddezza, dura e spietata come Agnese era stata con lei per tutti gli anni che aveva vissuto a casa. «Che vuoi!» le disse fissandola negli occhi, le chiavi già infilate nella toppa. «Per favore, Caterina, ti ho vista entrare nel supermercato e poi...». «Che fai, mi segui? Perché sei ad Aosta?». La madre abbassò lo sguardo. «Possiamo cercare di parlare cinque minuti?». Caterina lasciò le chiavi inserite nella serratura e incrociò le braccia. Il cuore le batteva forte. Cercò di respirare ma qualcosa si era bloccato nella trachea. «Poi te ne vai e mi lasci in pace?» ringhiò. «Mi ha chiamato tuo padre». Caterina deglutì, poi stirò un ghigno: «Siete tornati insieme?». «No. Non è per questo. Lui non ha il tuo numero». «Veramente?» chiese sgranando gli occhi. «E come mai?» fece ironica. «Mi ha chiesto di venirti a dire una cosa». «Non mi interessa quello che ha da dirmi» rispose, poi mise la mano sulla chiave e girò per aprire. Ma la madre l'afferrò per un braccio. «Caterina, ti prego. Tuo padre deve fare un'operazione al cuore molto pericolosa». Quel contatto le diede i brividi. Osservò la mano di sua madre, le unghie con lo smalto rosso scorticato, le dita che scoppiavano dentro due anellini di bigiotteria. Odiava quella mano, odiava quella pelle, quell'odore nauseabondo di sedano che emanava. «E allora?». «E allora prima di andare sotto i ferri vorrebbe vederti». «Perché? Rischia di restarci?». Agnese annuì. Allora Caterina si avvicinò e a bassa voce le mormorò sul viso: «È quello che gli auguro!», si liberò dalla stretta, aprì il portone, entrò nel palazzo e con la gamba richiuse l'anta di legno lasciando sua madre in mezzo alla strada. Salì le scale velocemente, senza pensare, con le chiavi in bocca per non perderle un'altra volta. Finalmente entrò nel suo appartamento, accese la luce, appese il giubbotto al gancio di legno vicino alla porta d'ingresso e andò in bagno ad aprire l'acqua della doccia evitando di guardarsi allo specchio. Veloce si spogliò ripiegando i pantaloni sulla sedia, gettò i calzini di cotone e la maglietta nel cesto di vimini per la biancheria sporca, ripose le scarpe nella scarpiera. L'acqua fumava. Entrò nella cabina doccia. Cominciò dai capelli. Due passate di shampoo, poi una dose abbondante di balsamo e il risciacquo. Prese la spugna e ci versò sopra il sapone. Sfregò le ascelle, se la passò fra i glutei, si strofinò le gambe fin quasi a farle diventare rosse, il collo, le braccia, i piedi. Chiuse l'acqua e afferrò l'accappatoio. Si asciugò energicamente. Un asciugamano intorno ai capelli zuppi e poi andò in salone. Accese il televisore e mise il telegiornale a volume altissimo. «... nubifragi stanno martellando le regioni che si affacciano sul bacino centro occidentale del Mediterraneo dove alluvioni lampo...». Riempì il bollitore di acqua, lo mise sul fuoco, preparò la tazza con la bustina di tè. Si sedette sulla poltrona davanti al televisore, prese un respiro e scoppiò a piangere. Seduto alla scrivania, Rocco Schiavone passava in rassegna la sua squadra. L'agente D'Intino dalla provincia di Chieti lo guardava con l'occhio fisso e vitreo di una spigola sui banchi del mercato. Deruta, umido di sudore, respirava a fatica ora che per raggiungere la nuova stanza del capo la strada era diventata troppo lunga, e se a quello si aggiungeva la rampetta di scale, praticamente un percorso di guerra. E poi c'era Casella, con la sua aria svagata di chi è capitato lì per caso, situazione non troppo lontana dalla realtà, e Italo e Antonio, annoiati, accanto alla porta. L'ex deposito di scope era una scatola di sardine. «Amici, colleghi, concittadini, prestatemi orecchio. Come potete vedere, siamo nel nuovo ufficio del vostro capo. Quindi questo è il grado di stima che la questura di Aosta ha per noi, nonostante i recenti successi. Ma, io dico, meglio! Dalle difficoltà si impara. E infatti da oggi si apre una nuova era. E si chiama?» guardò gli agenti che non sapevano cosa rispondere. «Si chiama Schiavone se fa li cazzi sua. E in che modo? Semplice. Sarete voi ad andare in giro a farvi il mazzo mentre io parteciperò alla conferenza stampa del questore e me ne starò qui a pensare e a rimuginare sui fatti della vita fumandomi una quantità industriale di sigarette. Fin qui ci siamo?». «Sì» mormorarono più o meno tutti, tranne D'Intino che si sporse verso Casella. «Je nen so' capite niente». «E mo' lo capisci, D'Intino» disse Rocco. «Allora dividiamo i compiti. D'Intino e Deruta, vi fate un giro alla ricerca di tutti i trans presenti ad Aosta e provincia fino a Biella e mostrate la foto del cadavere rinvenuto». I due agenti sbigottiti si guardavano, sembravano annaspare. «È una parola» fece Deruta. «Sticazzi!» aggiunse Italo. Rocco picchiò un pugno sul tavolo. «Allora, bisogna che qui al nord cominciate a imparare l'uso esatto dei termini e delle locuzioni romane. Sticazzi si usa quando di una cosa non te ne frega niente. Per esempio: Lo sai che Saint-Vincent ha 4.000 abitanti? Sticazzi, puoi dire. Cioè, chissenefrega. Come lo usate voi, Italo, è sbagliato. Devi cercare un ago in un pagliaio? Allora devi dire: mecojoni! Mecojoni indica stupore, lo usi per dire: accidenti! Capisci la differenza Italo? Non puoi usare sticazzi per esprimere meraviglia, sorpresa. Sticazzi lo usi per dire chissenefrega. Ho vinto alla lotteria 40 milioni di euro? Mecojoni, devi dire! Se dici sticazzi significa: non me ne frega niente. Ecco. Ricominciamo. Deruta e D'Intino devono cercare tutti i trans di Aosta e provincia. Tu che devi dire?». «Mei cojoni?». «Mecojoni» lo corresse. «Mecojoni». «Bravo Italo. Invece che a Courmayeur c'è la funivia?». «Sticazzi». «Perfetto. Hai appena imparato l'articolo sette della costituzione romana che recita: uno sticazzi al momento giusto risolve mille problemi. Andiamo avanti. Italo! Dal momento che oggi è il giorno libero di Caterina te ne vai bello come il sole da Fumagalli a chiedere novità sul cadavere». «Io?» chiese Pierron. «Ti chiami Italo? Proprio tu». «Ma dottore, a me l'obitorio...». «Se vuoi fare il poliziotto devi imparare a frequentare l'obitorio, punto e basta. E veniamo ad Antonio Scipioni. Tu e Casella, vi recate dal nostro commissario della scientifica a sentire le novità. È chiaro?». «Sì» risposero tutti, tranne D'Intino che invece si sporse verso Deruta: «Cioè? Je nen so' capite». «Dobbiamo chiedere ai trans se conoscevano il cadavere». «Shine, ma che è un trans?». Rocco tagliò corto. «Andate figlioli e tornate vincitori!». Poco convinti gli agenti si mossero. «Ci vediamo qui fra... diciamo quando vi pare. E mi raccomando, D'Intino e Deruta, tornate vivi. Buon lavoro...». Lasciarono l'ufficetto. Rocco restò solo con Lupa. «E adesso andiamoci a fare la conferenza stampa». La sala era piena per metà. Quando il questore vide arrivare Rocco dal fondo del corridoio, sorrise sollevato. Non poteva crederci. «Schiavone! Affrontiamo insieme i giornalai. Sono molto felice. Posso sapere com'è possibile?». «Molto semplice, dottore. Ufficio nuovo, strategia nuova. Ho mandato sul campo i miei agenti. Io me ne sto qui a fare il lavoro di raccordo». Costa lo afferrò per una manica. «Schiavone, ma che sta dicendo? La sua squadra? Insomma, la conosciamo bene. Ora a parte un paio di elementi...». «Sì, sono un disastro. Ma ora lavoriamo così. Dovranno pur crescere quei ragazzi, no?». Il questore non lo lasciava entrare. «Ma che cosa dice? C'è un omicidio di mezzo, quelli non sono in grado...». «E vabbè, omicidio risolto in più, omicidio risolto in meno. Ora andiamo, non è bene far attendere i giornalisti». Si liberò dalla stretta del superiore e finalmente entrò nella stanza. «Si tratta del ritrovamento di un cadavere sul greto del fiume, in località Plan Félinaz, vicino al campo sportivo. L'omicidio è presumibilmente avvenuto durante la notte, ma attendiamo delucidazioni dal dottor Fumagalli». Costa era in piedi e parlava. Rocco, seduto accanto a lui, osservava i giornalisti. La Buccellato era in seconda fila. I loro occhi si incrociarono più di una volta, mai invece lo sguardo della donna cercò quello dell'ex marito. Dalla prima fila si alzò una mano. Un uomo col nasone chiese la parola. «Prego, Angrisano» fece Costa sbuffando. Era evidente che lo odiasse dal più profondo del cuore. «Ecco, possiamo avere qualche dettaglio in più? Siamo sul generico». «Il delitto è avvenuto qualche ora fa!» gridò il questore. «Ci volete dare il tempo di capire, almeno? Questa non è una conferenza stampa, è un semplice comunicato alla stampa. Le sfugge la differenza?». «Sì, ma a parte questo» insistette Angrisano, «avete qualche sospetto? Qualche pista?». «Angrisano! I dettagli sono segreti, e lei dovrebbe saperlo. È un'indagine, mica una partita a monopoli!». «Dottore» fece una voce di donna. Era la Buccellato. Rocco notò la mano destra di Costa stringersi a pugno. «Qualcosa di più sul morto?» chiese la giornalista. Il questore, saliva azzerata, guardò Rocco cercando aiuto. Rocco prese la parola. «Bene. Il cadavere si trovava in acqua, gonfio, e quando dico gonfio non mi riferisco al corpo, ma al viso. Ha subito diversi interventi chirurgici, naso, labbra, zigomi, seni, però ha anche il membro e non si tratta di un ermafrodita ma di un trans. Escludiamo decisamente l'annegamento come causa della morte». «Perché?» chiese il giornalista riccetto. Rocco neanche gli rispose. «Non aveva documenti addosso quindi buio totale sull'identità». «E sul movente, suppongo» fece la Buccellato. «Bravissima. Dobbiamo percorrere la pista passionale, la pista del ricatto, della vendetta...». «Insomma, tutte le piste?» chiese Angrisano. «Esatto» e il vicequestore tacque. Costa lo guardava ma Rocco non aveva intenzione di dire altro. «Bene» prese la parola il questore. «La squadra mobile ha avviato le indagini e crediamo che al più presto...». «Perché il vicequestore è qui invece di indagare su questo omicidio?» chiese la Buccellato. Fu Rocco a rispondere. «Perché preferisco darvi io direttamente le notizie. Non mi piace quando qualcuno di voi mi mette in bocca le parole e fa della calunnia la sua arma preferita». La Buccellato sorrise. Rocco invece restò serio. Dopo tre secondi scoppiò l'inferno. «Schiavone, mi spiega che razza di atteggiamento è il suo? Viene a parlare con la stampa e mi aizza contro i giornalai? Io con questi ho a che fare ogni giorno, ho pressioni dai loro direttori, dal presidente della Regione, da tutti!». «Dottor Costa, se la considerazione che questa questura ha di me è quella di mettermi in un ex magazzino delle scope, farò in modo di avvalorare questa percezione. E mi comporterò di conseguenza. Visto che questo lavoro, la questura, lei, i giornalisti, Aosta, gli omicidi per me sono delle immense rotture di coglioni e di tutto questo non me ne frega una beneamata, non chiedo di meglio che starmene chiuso negli orari di lavoro nel deposito e lavorare da lì». «Lei sta rinnegando il suo ruolo». «No, siete voi che lo fate. Se ha bisogno di me mi trova fra il detersivo per i pavimenti e la scopa elettrica» e si incamminò per il corridoio. «Ah, a proposito. Credo che andrò a casa. Tanto prima di domattina dalla mia squadra novità non ne arriveranno. La saluto». Il cielo era buio, un alone più chiaro adagiato intorno alle cime dei monti ne seguiva i contorni. I lampioni stradali erano accesi, così come le insegne dei negozi. Rocco tornava verso casa, in mano un pacchetto con il kebab e una vaschetta di patate al forno. Lo vide svoltare in via Croix de Ville diretto verso casa con una busta del minimarket. «Com'è andata, Gabriele?». «Male. Ho preso due». Teneva la testa bassa e i capelli lunghi coprivano metà volto. «Allora quest'anno è andata». «Direi di sì. Alla fine mi ha pure chiesto di tradurre una frase in latino all'impronta!». «A te?» al vicequestore venne quasi da ridere. «Guardi, ce l'ho qui...» posò la busta a terra, si mise la mano nella tasca dei jeans tagliati al ginocchio, tirò fuori un foglietto e lo aprì. «Ibis redibis non morieris in bello». «Ma la sanno tutti! È la frase della Sibilla. A seconda di dove metti la pausa cambia il significato. Se dopo redibis si traduce: andrai, tornerai, non morirai in guerra. Se invece la sposti prima di morieris, allora suona come andrai non tornerai, morirai in guerra. Insomma il soldato prima di andare in battaglia chiedeva alla Sibilla una risposta sul suo destino, in realtà otteneva solo una frase ambigua. Sibillina, appunto». «Chi è la Sibilla?». Rocco allargò le braccia. «Lascia perde, Gabrie'...». Il ragazzo camminava a testa bassa. Rocco si bloccò all'improvviso. Prese Gabriele per mano e lo portò sotto un lampione. «Togli un po' i capelli dalla faccia». «Eh?». «Ho detto: togli i capelli dalla faccia!». Gabriele esitò. Cercò di divincolarsi ma Rocco lo teneva per un braccio. «Gabriele, fammi vedere!». Lento si portò la mano alla testa e con un gesto dolce scostò la ciocca che gli copriva metà volto. Sotto l'occhio aveva un livido giallo-viola. «Come te lo sei fatto?». «A scuola. All'ora di ginnastica. Partita di basket, m'è arrivato il pallone in faccia». Il vicequestore lo guardò negli occhi. «Sai che lavoro faccio io?». «Il commissario?». «Sbagliato. Faccio il vicequestore. Cioè il poliziotto. E quando sento una cazzata, vedi?» tirò su la manica del maglione, «mi viene la pelle d'oca. E ora io ho la pelle d'oca. Questo vuol dire che hai detto una bugia». «Si sbaglia. È la verità». «Gabriele, te lo chiedo una seconda volta. Come te lo sei fatto?». Il ragazzo si grattò la testa. «Niente, è stata una discussione». «Fra te e?». «E uno della scuola». «Mi sta bene. Uno contro uno mi sta bene. L'altro?». «Vuole sapere come si chiama?». «No, voglio sapere, tu porti a casa un livido, l'altro?». Gabriele alzò le spalle. «Eh no, figlio mio, no! Tu devi imparare a picchiare, che cazzo!». «Io i pugni non li so tirare». «Ho detto picchiare, mica boxare. Vabbè, bisogna che ti spieghi un po' di cose». Ripresero la strada verso casa. Fu Gabriele ad aprire il portone. «Mamma non c'è?». «No, stanotte resta a Milano». «E tu dormi solo?». «Sì. Sono abituato» i due entrarono nel palazzo. Rocco dietro, Gabriele davanti, salivano le scale. «E ti devi pure mettere a dieta, hai un culo che pari 'na poltrona. Che ti mangi stasera?». «Ho comprato due hamburger e poi mi sparo un gelato alla vaniglia. Lei?». «Kebab e patate al forno». «Buone!». Arrivarono sul pianerottolo. «Lo sai invece? Ce ne andiamo a cena fuori, io te e Lupa. A festeggiare». «Cosa?». «Che ti bocciano». «E c'è da festeggiare?». «Certo. Non lo sapevi? Se uno festeggia le brutte notizie, sta imparando a vivere. Fatti una doccia e mettiti dei pantaloni lunghi che co' 'sti cosi tagliati al ginocchio mi fai schifo. Si vedono i polpacci, i peli, e pure i fantasmini che escono dalle scarpe da ginnastica. E una maglietta pulita. Ce l'hai?». «Guardi che io la lavatrice la so fare, che crede?». Della casa nuova Rocco apprezzava soprattutto la vicinanza di almeno un paio di ristoranti di tutto rispetto. Trovarono posto al Grottino e ordinò una pasta con le vongole. Mai avrebbe immaginato di trovare ad Aosta una qualità da costiera amalfitana, e quando qualche mese prima l'aveva scoperto aveva eletto quel ristorante il suo preferito. Lì dentro si respirava mare, sole e cielo. Gabriele non amava spigole e orate e si gettò sulla pizza. Rocco si perdeva a guardarlo mangiare. Masticava e deglutiva come se non ci fosse un domani e la prima 4 stagioni finì in pochi secondi. «Una volta ho visto un documentario sui piranha, sai cosa sono i piranha?». «Sta scherzando?» rispose Gabriele con la bocca piena. «Piranha, Piranha paura, Piranha. La morte viene dall'acqua, Mega Piranha... non tocchi argomenti sui quali potrei distruggerla» e si attaccò alla birra. «Sembro uno di quei pesci assassini, vero?». «Esatto. Ne vuoi un'altra?». «Preferirei un calzone. Lo fanno?». «Credo di sì» attirò l'attenzione di un cameriere. «Senta, prepari un calzone qui allo squalo». «Subito» rispose quello e sorridendo sparì. «Lo squalo e Lo squalo 2, Lo squalo 3, e Lo squalo 4 la vendetta. So tutto!» e il ragazzo alzò i pugni al cielo come avesse segnato un goal. «Vuole sapere i registi? Il primo Spielberg, capolavoro assoluto del maestro indiscusso, baciare dove poggia i piedi, poi il secondo l'ha fatto Szwarc, il terzo l'ha girato Alves e l'ultimo diretto dal grande Sargent, che prima aveva realizzato Nightmares-incubi, pellicola da paura, non so se rendo l'idea». Rocco era rimasto col boccone a metà. «Ma di che cazzo stai parlando?». «Dopo andiamo in camera mia. Ho una collezione di film horror da leccarsi i baffi». «Io li odio i film horror» e il vicequestore riprese a mangiare. «Perché ha avuto un'infanzia difficile. Altrimenti li amerebbe». Quell'adolescente ciccione, pieno di brufoli e capelli lunghi e unti, che non apriva un libro, lo aveva lasciato di sasso. Rocco allungò uno spaghetto a Lupa. «Allora tu hai avuto una bella infanzia?». «Io da piccolo ero felice» prese un pezzo di pane. «Le cose sono andate storte dopo, ma fino ai 12 anni me la sono goduta. L'ha mai visto Alien?». «No. È un film, vero?». «Esatto. Ecco, immagini la famiglia come un corpo. Noi dentro avevamo un alien. E non lo sapevamo. Poi è venuto fuori, ma lo abbiamo combattuto e se n'è andato». «Un alien?». «Mio padre». «Dottore, scusi...». Rocco alzò la testa. Davanti a lui c'erano Italo, Casella e Antonio Scipioni. Visi stanchi e immalinconiti. «A proposito di Alien... che c'è? A quest'ora? Non vedete che stiamo mangiando?». «Sì, stavamo venendo a casa sua e l'abbiamo vista dalla strada». Rocco posò il tovagliolo. «La prossima volta, Gabriele, dobbiamo prendere un tavolo più interno. Volete favorire? Abbiamo appena iniziato. L'unico consiglio è non avvicinatevi troppo al pitbull qua...» e indicò Gabriele, «è capace di strapparvi un braccio a mozzichi». «No, veramente, grazie, ce ne andiamo a casa. Solo che...». Italo guardò i colleghi. Prese la parola Casella. «Solo che dotto'. Acqua. Allora la commissaria della scientifica con me e Scipioni non ci ha voluto parlare, dice che non si fidava e che probabilmente eravamo dei servizi...». Rocco alzò gli occhi al cielo. «E io all'obitorio sono svenuto» concluse Italo. Rocco batté le mani. Lupa poggiò le zampe anteriori sulle ginocchia di Rocco. «Ottimo lavoro, ragazzi. Mancano i fratelli De Rege...». «Quelli stanno in giro. Di loro non abbiamo più saputo niente». «Missing in action. Allora a domani. Fate una bella relazioncina, scrivete esattamente i risultati ottenuti e mettetela sul tavolo del questore. Mi raccomando Italo, proprio come l'avete raccontata a me. Ci vediamo domani». I tre poliziotti, coda fra le gambe, lasciarono il ristorante. In quel momento il cameriere portò il calzone a Gabriele. «Vacci piano guaglio', che è l'ultimo!» e ridendo si allontanò dal tavolo. Gabriele afferrò subito le posate. «Ma di che parlavate? Che cercavano?». «Gabrie', lascia perdere. L'horror guardatelo nei film. Com'è il calzone?». «Buonissimo!» posò la forchetta. «È una cosa brutta quella che è successa?». «Parecchio brutta. Dai, mangia che si fredda». «Be', allora buonanotte. Domattina vai a scuola?». Gabriele alzò le spalle. «Hai ragione, che ci vai a fare?». «Vabbè. Grazie per la cena». Rocco aprì la porta. Lupa entrò per prima. Anche Gabriele infilò le chiavi nella toppa. Rocco si voltò. «Senti un po', Gabriele». «Dica» il ragazzo aveva gli occhi grandi e spalancati. «Non è che ti va di dormire da me?». «Magari!» e tolse le chiavi dalla serratura. «Allora dai. Entra!». Gabriele si lanciò in casa. «Grazie». «Oh, io domattina vado a lavorare. Qui la colazione non c'è». «Non si preoccupi, quando mi sveglio torno a casa». Gabriele si guardava intorno mentre Lupa aveva già preso possesso del suo angolo di poltrona. «È bello il suo appartamento. È normale». «Che intendi?». Rocco gettò le chiavi in un piattino sul comò. «Che è una casa bella. La mia sembra un ospedale. Quel quadro?» indicò uno degli sgorbi che Anna, una vecchia fiamma di Rocco, aveva dipinto e lui, per pura carità, aveva acquistato. «Quello? Una crosta». «Be', almeno ha un vantaggio». «Quale? «È biodegradabile, no?». Rocco scoppiò a ridere. «Ti va bene il divano?». «Alla grande». «Lenzuola nell'armadio». Rocco tolse i cuscini, tirò la rete e aprì il divano letto. Lo usava per la prima volta. Chissà se era comodo. «Ce l'ha qualche Marvel?». «Come? Intendi fumetti?». «Sì. Va bene pure Topolino». «Nun t'allarga', ciccio. Intanto c'è il letto, e non mi pare poco». Si misero insieme a sistemare le lenzuola. «Hanno ammazzato qualcuno?» chiese il ragazzo. Rocco annuì. «Lei ne ha visti tanti di morti?». «Tanti...». «Ci ha fatto l'abitudine?». «No». «Infatti» fece Gabriele, «a certe cose secondo me non si fa mai l'abitudine». Rocco rimboccò la coperta di cotone. «Allora facciamo una prima lezione». «Di che?». Gabriele cercava di infilare la federa nel cuscino. «Chi mena per primo mena due volte, lo sapevi?». «Sì. Ma io non ho il coraggio. Non ho mai picchiato nessuno». Rocco si avvicinò. «Ma gli altri sì, quindi ti devi difendere. Dimmi la verità, Gabriele, quanti erano?». Gabriele posò il cuscino. Lo aggiustò, poi senza guardare Rocco rispose: «Tre...». «In tre? Figli di puttana. No, e qui bisogna che impari. Allora guardami. Girati». Il ragazzo obbedì. «Non sai fare a pugni e diciamo che non ci sono oggetti contundenti in giro, che so? Una bottiglia, un sasso, un bastone, un'ascia, niente. Per prima cosa devi stare a mezzo metro dallo stronzo così, vedi? e conti fino a tre. È facile. Uno due e tre, e sai che al tre gli devi mollare un calcio in mezzo ai coglioni». «Devo?». «Sì. Pensa che se non lo fai tua madre muore». «Ah, muore?». «Bravo. Quindi lo devi dare. E lo devi tirare forte. Quello non deve più rialzarsi. Così, guarda» e mimò il calcio. «Senza pensarci due volte. Come se calciassi un rigore. Quello va per terra. Quando sta per terra fai finta che è un pallone e gli molli un altro calcio qui, sotto il mento» toccò il punto preciso sul viso del ragazzo. «È facile che quello a questo punto è svenuto. E tu te ne vai». «Posso urlare goal?». «Puoi urlare quello che ti pare». «E se sono in tre?». «Picchia sempre il capo». «Diego?». «Che cazzo ne so come si chiama? Ma se tu atterri 'sto Diego, gli altri se ne vanno, sta' tranquillo. Hai capito? Calcio alle palle e poi sotto il mento». «Proviamo?». «Sì, ma solo la mossa, non me lo dare il calcio, va bene?». «Bene». Gabriele guardò il punto dove doveva calciare e accennò il movimento. «Errore. Mai guardare dove colpisci, sennò gli dai un vantaggio, quello lo capisce. Lo devi fissare negli occhi, dire una cazzata tipo... che bella giornata oggi, vero? E zac! calcio potente alle palle. Così, guarda». Rocco fece la dimostrazione senza colpirlo. «Vedi? Io ti guardo nelle pupille e poi ho tirato il calcio preciso. Prova». Stavolta lo fece bene. «Bravo. Forte. Seconda ipotesi: quello non cade a terra ma si piega a metà. Invece del calcio sotto il mento sai che devi fare?». «Faccio Thor martello di Odino?». «Non ti seguo...». «Gli do una botta sulla nuca a due mani?». «Sì, ma forte. E grida, se ti può aiutare. Prova. Ecco, io sono chinato...». «Martello di Odino!» urlò Gabriele e mollò una botta a due mani sul collo di Rocco. «Ahhh, ma che sei scemo? Devi fa' per finta!». «Oddio, le ho fatto male?». «Niente, niente» il vicequestore si massaggiò il collo. «Però bene. Ao', sei forte. Basta menare una volta e passa la paura». «Se penso che lei è lì con me, ce la faccio!». «Bravo, pensa che sono lì con te! E che se non tiri il calcio alle palle di Diego tua madre muore. Ora andiamocene a letto». Gabriele si tolse le scarpe da ginnastica. «Quelle mettile fuori dalla finestra che puzzano di topo morto». Il ragazzo le annusò. «Bleah! Che schifo». «Appunto. Altrimenti qui diventa una camera a gas. Andiamo Lupa?» si avviò verso la camera da letto. «Oh, qualsiasi cosa io sto di là». «A domani allora». «A domani Gabriele. Dormi bene» e sorridendo entrò nella stanza. «Rocco?». «Dimmi». «Ma non è che passo da vigliacco se gli do un calcio alle palle e poi basta?». «No. Passi da vigliacco se rientri a casa con un livido come quello». Rocco si svegliò di soprassalto alle tre di notte. Gabriele s'era infilato nel letto e dormiva sereno abbracciato a Lupa. Aveva un sorriso a fior di labbra. Sognava. Anche Lupa sognava, muoveva le zampe come se stesse correndo. Sentì una mano calda avvolgergli il petto. Una scossa piacevole, un massaggio al cervello e al cuore. Appoggiò la testa al palmo della mano e si perse a guardare quel ragazzo e il suo cane sprofondati nel sonno con la testa chissà dove. Era questo che intendeva Marina quando parlava di sicurezza e protezione? Avere un angolo di mondo dove pensieri e paure restano fuori e rimane solo la dolcezza di un sonno tranquillo? Avrebbe voluto chiederglielo, ma non veniva più a trovarlo. «Abbiamo una sola vita, Mari', e ce la siamo giocata male» disse a mezza voce. «Mi correggo. Abbiamo una sola vita e me la sono giocata male. Perché è tutta colpa mia» e questo non l'avrebbe cancellato niente e nessuno. Era la verità evidente e spietata con la quale doveva fare i conti. Poggiò la testa sul cuscino fino a quando vide il soffitto appannarsi e gli occhi cominciarono a bruciare. «Se ti apri alla vita, quella qualcosa in cambio te la dà sempre...» gli risuonò nella testa. Si asciugò gli occhi. Chi aveva parlato? Gabriele dormiva. A Lupa non voleva neanche pensarci, ci mancava un cane con il dono della parola. «Sei qui? Sei qui vero? Sei qui e non ti fai vedere. Eri in strada oggi, ho sentito il tuo profumo. Sei sempre qui, vicino a me? Vuoi sapere chi è questo qui? Non lo so. È un mio vicino. Ha 16 anni, non ha nessuno. Lo vedi come dorme? Io non ho mai dormito così, o forse sì. Ho mai dormito così, Mari'?». «No. Tu russavi». «E questo cos'è?». Costa teneva il rapporto degli agenti davanti al viso di Schiavone fra il pollice e l'indice neanche fosse un calzino sporco. A Rocco tornò in mente la sua maestra delle elementari quando con gli occhi di fuoco e la bocca spalancata ostentava il quaderno alla classe per rivelare la peggiore offesa di un alunno al corpo docente, per denunciare pubblicamente la bestemmia contro la pubblica istruzione, per mostrare alla classe attonita la madre di tutti gli insulti all'insegnamento dai tempi di Giovanni Gentile: il buco nella pagina. Un affronto che si pagava con uno zero spaccato e una mezz'ora dietro la lavagna. «Questo è il frutto di un'intera giornata di indagine da parte della sua squadra? Glielo leggo?». Costa scuoteva il foglio nell'indifferenza totale del vicequestore che se ne stava in piedi accanto alla finestra, a braccia conserte, perso ad osservare la vena del collo del suo superiore che si gonfiava e si sgonfiava, mentre il viso tendeva ad un rosa più acceso. «Vogliamo dire che i suoi nuovi metodi di lavoro fanno schifo?». «Non lo so. Lo vogliamo dire?». «Questo rapporto è indecente!» lo accartocciò e lo gettò nel cestino dei bicchierini di plastica accanto al distributore. «Io adesso le chiedo gentilmente di mettersi a fare il suo lavoro, con serietà...». «Lo sto facendo. Non è colpa mia se ho una squadra di mentecatti. Mi avete dato quegli uomini, e io quello posso fare». «In passato però le cose andavano diversamente, Schiavone!». «In passato sì. Ora le cose stanno così. Ci proverò, ma non assicuro niente». Costa si mise le mani sui fianchi, poi si toccò il mento, infine guardò Schiavone. «Va bene, cosa vuole?». «Lo sa». «Tutto questo per una stanza mefitica?». «Non è la stanza, dottor Costa. È la fiducia. Lei e il dottor Baldi avete mostrato, e senza troppi giri di parole, che di me non vi fidate più». Costa alzò gli occhi al cielo. «Per la storia di Enzo Baiocchi?». «Eh già. Dunque mi dica. Se lei non godesse più della stima e del rispetto dei suoi capi, come si comporterebbe? Se quelli che ti devono dare una mano nelle indagini ti guardano come il primo dei sospetti, che farebbe?». «Il mio lavoro! Ma io e lei siamo differenti. Io ci credo in quello che faccio. Lei potrebbe essere qui oppure su un cargo al largo delle Azzorre, sarebbe lo stesso». «Non sarebbe lo stesso, al largo delle Azzorre starei a poppa a prendere il sole». «Si crede spiritoso, Schiavone?». Nascosti dietro l'angolo del corridoio Italo e Antonio seguivano il dialogo e impedirono a Casella, arrivato in quel momento, di andare verso il distributore del caffè. Sapevano che era una situazione delicata e Rocco stava per esplodere. Era meglio non farsi vedere. «No, Schiavone, lei non è spiritoso. Lei ha dei doveri. Quindi si metta al lavoro e mi porti dei risultati, ammesso che ne sia in grado!». «Vuole dei risultati da me? Da me? Sono qui da settembre e ho fatto più io che qualsiasi altro mentecatto in questa questura di merda. E mi costa parecchio perché a me di sguazzare nelle fogne alla caccia di un pezzo di merda, di nutrirmi di morte e sangue non me ne frega niente, dottore. Niente. Lo faccio perché non ho di meglio e perché davanti all'orrore non so stare fermo. Ma se mi fanno girare i coglioni, e mi creda lei e il dottor Baldi ci state riuscendo alla perfezione, io mollo il colpo. Alzo le mani e non faccio più un cazzo. Spero di essere stato chiaro. Ora mi scriva lettere di ammonimento, mi sbatta in un garage, mi trasferisca in una miniera, la cosa non cambia. Io sul campo mando i miei uomini». Costa ingoiò amaro. Si avvicinò alla macchinetta del caffè. Ci si appoggiò. «Non ho voglia di prendere provvedimenti disciplinari contro di lei». «E io non vorrei mai essere arrivato a questo». Poi il questore lo guardò negli occhi. «Non dipende da me. Io obbedisco a ordini che arrivano dall'alto. Lei non piace, Schiavone. Non piace al tribunale, non piace al ministero, e soprattutto non piace a Roma. Forse io sono il solo che ha sempre cercato di difenderla, perché conosco il suo valore. E allora glielo chiedo col cuore in mano: mi aiuti. Non si metta di traverso e io le garantisco che farò il possibile affinché tutto questo cada nel dimenticatoio». Infilò la chiavetta e premette il pulsante. «Non mi tradisca, o io non avrò più niente in mano per aiutarla. E fino ad oggi, mi creda, ci ho provato. Una cosa deve essere chiara. Io la carriera per lei non me la gioco. Fino a dove posso, conti su di me». Si aggiustò la cravatta. «Lo sa cosa mi disse una volta l'analista? Se una persona è immersa nelle sabbie mobili, per aiutarla, non bisogna gettarsi nel fango, altrimenti si affogherebbe entrambi. Io tendo la mano, sta a lei adesso afferrarla». Prese il caffè ormai pronto. «Ne vuole?». «No grazie. Preferisco vivere». Sorrisero. «Lei torni a fare il suo lavoro. E le prometto che le cose cambieranno. Fa cagare questo caffè!». «Lo so. Io una mezza idea ce l'avrei». «Cioè?». «Compro cialde e macchinetta e quando vuole un caffè, venga da me». «Allora avrà bisogno di una stanza un po' più grande». «Minchia, ve le siete cantate con il questore» disse Antonio che insieme a Italo seguiva Rocco lungo il corridoio. «Che facciamo, Rocco?». «Prendi la macchina, Italo, andiamo a lavorare. Antonio, tu chiama Fumagalli e digli che fra poco siamo da lui. Poi con Rispoli vai sul luogo dell'omicidio. A proposito, Rispoli è tornata?». «Eccomi!». Schiavone si voltò. Finalmente un po' di luce, pensò. Ma non si riferiva al sole che, sebbene un po' velato, faceva di tutto per illuminare la città. Era piuttosto la presenza di Caterina nei corridoi della questura. «Cateri', finalmente!». «Era il mio giorno libero...» aveva il viso grigio e gli occhi spenti. «Che hai il vizio di prenderti quando c'è bisogno di te in questura». Caterina non reagì. «Lei ha bisogno di me?». «Siamo tornati al lei?». «Siamo in questura...». «E certo che ho bisogno di te. M'hai lasciato co' 'st'armata Brancaleone... Sai tutto?». «Sì, Antonio e Italo mi hanno relazionato» e stirò un sorriso forzato e finto. Rocco le si avvicinò abbassando il volume della voce: «Che hai fatto ieri? Non ti sei riposata?». «No. Direi di no». «Sei triste?». «Diciamo che sono fatti miei...». Rocco frenò la camminata. Erano davanti alla sua vecchia stanza. La porta era chiusa a chiave. «Embè?». «È la stanza della Gambino adesso. Chiude sempre a chiave» fece Italo. «E si può?» chiese Antonio. «Boh...» rispose Italo. «Ha mille paranoie». «Facciamo milleuna?» propose Rocco. «Datemi una gomma da masticare». Antonio tirò fuori un pacchetto. Schiavone masticò la cicca per qualche secondo, poi se la sputò sul palmo della mano. La divise in due pezzi. «Scusa Cateri', permetti?» con un gesto rapido strappò un capello al viceispettore. «Ahia! Ma che stai facendo?». Rocco si chinò. Incollò i due pezzi di gomma sulle ante della porta. In mezzo ci incastrò il capello di Caterina. «Avete mai visto i film di spionaggio? Questa è la tecnica per sapere se qualcuno entra nella tua camera d'albergo». «Sì, ma perché alla porta della Gambino? A che serve?». «A niente. Solo a darle una paranoia in più». Ripresero a camminare verso l'uscita. «Dove sono Stanlio e Ollio?». «Ecco, questo non lo sappiamo» rispose Antonio. «Cioè, da ieri non abbiamo più loro notizie». «Meglio, ce li siamo tolti dai coglioni». L'auto correva verso l'obitorio. «Allora, Italo, come va?». «E come deve andare? Male, Rocco, molto male». «Con Caterina?». «Con Caterina credo sia finita. È stato un fuoco di paglia. Non è una donna facile. Anzi, a dirla tutta, è un vero casino». «Mi dispiace». «Non è vero. Sei felice della notizia. Volevi provarci? Ecco, adesso hai campo libero». «Ti sento aggressivo. Io che c'entro?». «Niente. Fa solo un po' male, scusa». «È così triste per colpa tua?». «Non credo. Io non la sento da due giorni. Secondo me c'è altro...». Frenò per far attraversare la strada ad una signora con il carrozzino. «Senti Rocco, ma secondo te non capita un'occasione?». «Ti devi compra' una macchina?». «Voglio dire, una di quelle occasioni per arrotondare...» e ripartì accelerando. Rocco prima lo osservò, poi dalla tasca prese il pacchetto di sigarette. «Cazzo!». Era vuoto. Lo accartocciò e gettò l'involucro a terra. «Dammene una delle tue». «Sicuro? Ho cambiato marca». «Peggio delle Chesterfield non puoi aver preso...». Italo gli allungò un pacchetto di MS Brera. Rocco lo prese in mano, lo guardò, lo soppesò. «Visto che ormai è scientificamente provato che le sigarette uccidono, perché ti suicidi co' 'sto schifo? Se devi morire affogato, Italo, meglio morire affogato nel Pacifico al largo di Bora Bora invece che nel laghetto artificiale per la pesca sportiva a Acquapendente, che ne dici?». «Guarda che non sono male». Rocco restituì le sigarette a Italo. «Non ce la posso fare... se vedi un tabacchi, un bar, un chiosco, un contrabbandiere inchioda. Allora, che volevi sapere?». «Se non ci capita più niente per arrotondare...». «Non lo so, non credo. Da settembre qualche migliaio di euro l'hai presi... finiti?». «Sì...». «Ti vesti di merda, fumi di merda, non vai mai a cena fuori, non hai più la ragazza, non hai la macchina, non hai la moto, dividi l'appartamento con un altro agente e paghi 200 euro al mese, che ci fai coi soldi?». Italo schizzò del sapone sul vetro e azionò i tergicristalli. «Oh, allora? Sto aspettando». «Ma niente, che ne so? Un po' li ho dati a mia zia, un po' avevo dei debiti e poi lo sai Rocco come sono i soldi?». «Vanno e non vengono. Vabbè, dammi una emmesse sennò comincio a leccare il portacenere e non è una cosa elegante». Italo obbedì. «Devo proprio entrare da Fumagalli?». «Facciamo così. Se entri, io ti prometto che vedrò se c'è da arrotondare. Se come al solito resti fuori rimani coi debiti che sicuramente avrai contratto con qualche zoccola di Aosta». «Véi t'fa ampica» ringhiò Italo in mezzo ai denti. Rocco non comprese le parole, il senso generale sì. «Altrettanto, Pierron». Già nel corridoio Italo Pierron presentava un evidente pallore mentre il collo era picchiettato da un firmamento di macchie rosse come tanti morsi di pulce. Gocce di sudore colavano regolari dalle tempie e le palpebre gonfie sbattevano con un ritmo che si intensificava man mano che la porta della morgue si faceva più vicina. Anche il passo s'era fatto incerto. Metteva un piede avanti l'altro, ma non sempre seguendo una linea retta, e il movimento dubbio rendeva il corpo instabile e gelatinoso come un budino. «Ti senti poco bene?» chiese pleonasticamente Rocco. «Sì... ma il cadavere... è proprio combinato male?». «Ma cosa vuoi che ti dica. È morta per strangolamento, pare, quindi è un po' viola, avrà gli occhi fuori dalle orbite, la lingua blu, magari si sarà già decomposta e...». Un tonfo improvviso. Rocco non si fermò. Proseguì lasciando a terra l'agente Italo Pierron che ormai aveva preso una certa familiarità con il pavimento dell'obitorio. Fumagalli era chino sul lettino autoptico. Appena vide entrare Schiavone controllò subito che avesse la plastica intorno alle scarpe. «Non mi mandare più Pierron, fai il favore. Ieri non è riuscito neanche ad entrare nell'edificio». «Oggi invece è entrato». «E dov'è che non lo vedo?». «Giace nel corridoio. Chi è?». «Questo? Un poveraccio che si è schiantato a Pont-Saint-Martin». Coprì il corpo con un gesto rapido. «Vieni con me...». Il cadavere nella sua nudità mostrava tutte le sue contraddizioni. Spalle polsi e caviglie da uomo, il seno abbondante, il viso coi capelli corti ma che voleva somigliare, almeno nella fantasia della proprietaria, a qualche attrice americana, il pene violaceo. «E noi ci si lamenta che abbiamo una vita difficile» fece Fumagalli poggiando il telo ai piedi del lettino. «Ci pensi? Te nasci in un corpo che non vuoi e lo devi cambiare. E mica è un'automobile!». «Albe', saranno cazzi suoi. Raccontami un po' di cose, va'...». «Allora, sul sesso ci siamo chiariti, altezza uno e settantaquattro, peso sessantacinque chili. Piuttosto in forma, pochissimo grasso, muscoli tonici... roba da pilates, zumba, body tone». «Di cosa stai parlando?». «Robe da palestra. E non ti farebbero male neanche a te». «E te che ne sai?». «Io frequento, Rocco. E soprattutto non fumo. Parliamo degli interventi che la poverina ha subito? Allora, blefaroplastica, aumento degli zigomi, chirurgia delle labbra, rinoplastica, lifting facciale, mastoplastica additiva, liposuzione...». «Orca troia, ma quanto tempo ha passato in sala operatoria...». «Già. Allora andiamo con la prima. Ti dico che il tuo patologo preferito ci ha visto giusto. Strangolamento». «Nei polmoni c'era acqua?». «Macché. Quando l'hanno buttata nel fiume la poverina era bella che andata. Sono risalito all'ora della morte. Più o meno verso le otto di sera, aveva cibo non digerito in pancia». «Che ha mangiato?». «Pesce crudo e riso. Tradotto: sushi. E s'era anche bevuta un goccio di vino bianco. Magari può tornare utile». «Come l'hanno strangolata?». «Vieni...» si avvicinarono a un piano sul quale giacevano alcune foto del collo del cadavere in dettaglio. «La corda che hanno usato è alta più o meno un centimetro. Vedi?» e mostrò col dito l'ingrandimento. «Può essere una corda di qualsiasi tipo». «Che palle. Analisi del sangue?». «Le stanno facendo. Per ora escludiamo uso di stupefacenti, almeno negli ultimi cinque giorni. Altra cosa curiosa. Torniamo dalla nostra amica». Si avvicinarono al lettino. «Guarda qui...». Fumagalli indicò un punto una ventina di centimetri sopra la caviglia. «Noti questo livido circolare?». «Sì». «E ne ha un altro identico sull'altra caviglia, vedi?». «Vero. Che cos'è?». «Non ne ho idea. Tu registra. Intanto hai scoperto chi è?». «No». «Cioè, fammi capire, Rocco. Io sto qui a sudare i miei sette camici e tu stai in ufficio a grattarti la pancia? Bellino, se si è un tandem si è un tandem! Sennò sono io che trascino te!». «E chi l'ha mai detto che siamo un tandem? E poi, ma che te voi porta', Alberto...». «Comunque io mi lavoro gli organi, e se va bene ti dico tutto, anche se aveva la dissenteria. Te vedi di dargli un nome». Rocco senza salutare si diresse verso l'uscita. Poi venne fulminato da un pensiero: «Scusa una cosa, Albe'...». «Sono tutt'orecchi». «Segni di lotta? Che so, pelle sotto le unghie, graffi?». «Niente. La poverina è morta senza opporre resistenza. Forse se n'è andata nel sonno?». «Forse. A parte i lividi alle caviglie ne ha altri?». «Poca roba. Uno piccolino al polso della mano destra, un altro vicino al ginocchio, ma se ti fai dare un'occhiata è facile li abbia anche tu. Ah, sopra il labbro superiore. Ma secondo me sono dovuti a una depilazione laser o roba simile». Dopo aver raccattato Italo e aver guidato fino alla questura, decise di riunire le teste pensanti del suo reparto al bar di Ettore. Caterina e Antonio erano stati sul luogo del delitto. La scientifica ci stava lavorando ancora. «Curiosa la Gambino» aveva detto Antonio sedendosi al tavolino sotto il portico. «È convinta che nell'acqua di un fiume ci possano essere delle prove importanti. Secondo me è mezza matta». «Togli il mezza, Antonio...» poi una volta che Ettore ebbe depositato l'ordinazione, Rocco diede inizio alla riunione. «Per capire chi fosse la nostra vittima ci sono pochi e disperati elementi. Abbiamo il sushi, sappiamo che andava in palestra e che ha fatto parecchi interventi chirurgici. In più, se vogliamo dare retta al giudice Baldi, e tutti i torti non li ha, parecchi trans per campare si prostituiscono. Quindi dividiamoci i compiti. Tu Caterina ti fai il giro di tutte le palestre di Aosta». «Ricevuto» rispose con la testa bassa, senza guardare gli altri negli occhi. «Tu Antonio ti occupi dei ristoranti di sushi. A nord del punto di ritrovo». «Perché a nord?». «Supponiamo che l'abbiano uccisa in casa, visto l'abbigliamento spartano. E sempre per l'abbigliamento spartano diciamo che il sushi se l'è mangiato nel suo appartamento. Un take away, e di solito se ne sceglie uno vicino. Ecco perché risalirei dal punto del ritrovamento verso nord. Non è che un cadavere risale la corrente, mica è un salmone. Ovviamente tutte supposizioni che vanno a farsi fottere se non l'hanno uccisa in casa». «Io?» chiese Italo. «Per favore l'obitorio no». «No, stavolta una cosa più piacevole, almeno per un maschio arrapato come te. Guarda tutti gli annunci di escort della Valle, sempre sperando che lavorasse qui e non era in visita a casa di qualcuno. Magari una botta di fortuna». «E tu?» chiese Caterina. Rocco non rispose. Lasciò i soldi della consumazione e si alzò. «Dotto', i chirurghi estetisti sono una mezza dozzina». «Come Case', i chirurghi?». «Estetisti». «Cioè rifanno il rimmel, punture di eyeliner, mastoplastica additiva al fondotinta?». Casella lo guardava senza capire. Teneva le cartelle con gli appunti in mano. «I chirurghi estetici, Case', estetici, quali estetisti! Allora sono una mezza dozzina, e te li fai tu. Vai con la foto e vedi se ne sanno qualcosa. Prima mi porti all'ospedale». «Si sente male?». «Se continuo a stare con te sicuro. Trovata la caldaia?». «Forse...». Il Parini era un ospedale pulito, ordinato, senza avvisi scritti a pennarello e appiccicati con lo scotch sugli stipiti delle porte. A Rocco venne in mente quanto fosse distante da un nosocomio romano, e non solo in termini di chilometri. Salì le scale al primo piano, attraversò un paio di corridoi e arrivò finalmente alla stanza della dottoressa Sara Tombolotti, la psichiatra. Era con lei che voleva parlare per prima. Bussò. Percepì un flebile «avanti» e aprì la porta. Una stanza inondata di luce, un firmamento di pulviscolo, una scrivania al centro, libri ammucchiati nelle librerie che occupavano le pareti, altri impilati a terra, colonne di giornali e riviste sullo scrittoio. Un odore di cannella e menta aleggiava denso come fumo. «Prego» sentì dire, ma della presenza della dottoressa Sara Tombolotti nessuna traccia. Rocco avanzò perplesso fino ad arrivare alla poltroncina davanti al tavolo. «Tolga pure le riviste e si segga, mi scusi per il disordine!». Rocco obbedì. Sentiva rumori di scartocciamenti vari. Si sporse e finalmente la vide. Era una donna minuta, anziana, con un paio di enormi occhiali tondi; pareva un ghiro nella sua tana circondata com'era da scartoffie, libri e periodici. Le sorrise, la dottoressa ricambiò. «Sto cercando di mettere a posto, ma tanto non ci riesco. Aspetti». Scese dalla sedia e fece il giro della scrivania con la mano tesa. «Piacere, Sara Tombolotti». «Rocco Schiavone». Si strinsero la mano. La psichiatra gli arrivava alla pancia. «Allora mi ha detto al telefono...». «Che stiamo cercando di risalire all'identità del cadavere ritrovato sulla riva della Dora». «E si rivolge a me perché è convinto che questa persona fosse in procinto di operarsi, giusto?». Invece di tornare alla scrivania, sedette sul divano e ci sprofondò. I piedi quasi non toccavano il pavimento. «Giusto». «La disforia di genere, mi perdoni, a me parlare di disturbo dell'identità di genere non piace, impone alla persona delle operazioni molto dolorose. Si sottopongono a un vero e proprio calvario. Lei mi diceva che il soggetto ha già avuto degli interventi... ovviamente stiamo parlando di un M to F, no?». «Da maschio a femmina, sì, è così. Ha già fatto interventi al viso, labbra, zigomi, naso, mastoplastica additiva...». «Manca quello più importante. La vaginoplastica». «Sì. Aveva ancora il pene». «S'è mai posto la domanda che magari a lei stava bene così?». «No dottoressa, non me la sono posta». «E ha fatto bene. L'iter è lungo, complesso, serve sempre la presenza dello psichiatra. Mica è un intervento da ridere. Bisogna ricostruire la cavità della vagina e le grandi e piccole labbra con lo scroto. Il clitoride invece con il glande, quindi la zona rimane attiva dal punto di vista della sensibilità, se voleva saperlo». «No, non era nei miei pensieri, ma ora lo so». «Già. Ma guardi che va meglio ai maschi, come sempre. Se una donna vuole diventare uomo non le dico che interventi deve fare...». «Ma io mi fermerei qui, la ringrazio. Lei non ha nessun paziente, diciamo, in cura?». «E diciamolo. No, nessuno. Né so di colleghi qui che affrontino il problema. Che io sappia il Parini non ha mai fatto una simile operazione. Lei dovrebbe andare a cercare a Torino. C'è un ottimo gender team alle Molinette. Io dovessi decidere di cambiare sesso, andrei lì». «Mi può indicare...». «Corrado Berlingeri. Era mio marito. Se è della zona, lui lo sa. Si rivolgono tutti a lui, più o meno». Rocco si alzò. La dottoressa non lo imitò. «Io la ringrazio per il tempo, dottoressa Tombolotti. E...». «Lei come sta?» gli chiese guardandolo dritto negli occhi. «Io?» fece Rocco preso in contropiede. «Bene». «Sciocchezze». «Vero». «Così, dopo una chiacchiera di dieci minuti, solo osservandola, mi permetterei di dire che lei è un uomo sostanzialmente depresso, magari con qualche difficoltà di concentrazione. Non è che soffre di allucinazioni?». «No!» e accompagnò la risposta secca con un gesto del capo. «Bene, soffre di allucinazioni. Come andiamo con l'umore?». «Quale umore?». «Appunto. Lei se ne sta sempre asserragliato in difesa?». «No. Io devo andare a lavorare. E penso anche lei». «Non si arrabbi. Era solo per parlare». «Non mi piace quando la gente mi guarda come se fossi un batterio su un vetrino». «È il mio lavoro. Se avesse bisogno di un aiuto, io sono qui». «Grazie» strinse la mano alla dottoressa Tombolotti e lasciò la stanza. Per la prima volta in vita sua Rocco Schiavone si sentì nudo in mezzo alla strada. Quello gnomo occhialuto l'aveva spogliato in meno di un minuto. Meno di un minuto, e già le voleva bene. Abbassò il telefono e restò seduto a pensare. Corrado Berlingeri era ad Atlanta per un convegno, e prima di qualche giorno non sarebbe tornato. Andare fino a Torino era inutile. Aprì il cassetto e preparò una canna leggera. Sentiva il bisogno di fare un piccolo ripasso, la sostanza psicotropa avrebbe aiutato. Accese, aspirò, trattenne il fumo per qualche secondo, poi espirò fuori dalla finestra. E l'idea gli arrivò. Afferrò il cellulare e chiamò Fumagalli. «Stai lavorando?». «No. Mangiavo un panino. Che vuoi?». «Secondo te era italiana?». Ci fu una pausa. «Perché non ci ho pensato prima ai denti?». «Appunto. Guarda un po', magari abbiamo fortuna». «E mi ci sto recando» fece Fumagalli masticando. «Cioè vai a analizzare un cadavere col panino in mano?». «Embè?». «Come embè?». «Aspetta...». Percepì dei rumori. Alberto stava lavorando sul corpo. «Ecco qui... abbiamo due molari curati con l'oro. Qui non si fa più, e poi devo dire che generalmente non stanno messi bene...» la voce divenne più presente, segno che Fumagalli aveva ripreso il telefono in mano. «Ora si usano delle resine. Direi che al 90 per cento la tipa non è italiana. Bella idea, Rocco. Bravo». «Grazie». «Non ti abituare però... è stato solo culo il tuo. Stammi bene». «Alberto! E se magari sono otturazioni che risalgono a anni fa?». «Ha una trentina di anni. Queste sono state fatte a occhio e croce cinque, massimo sei anni fa a guardare la corona del dente. Comunque ci metto un attimo. Esamino e ti dico» e chiuse la comunicazione. Tempo di spegnere la cicca e bussarono alla porta. «Avanti!» fece Rocco. Entrarono Italo e Caterina. Facce stanche, tese. Evidentemente avevano avuto una discussione. «Strano odore» fece Caterina. «Sono gli impacchi con il rosmarino». «Non hai il raffreddore». «Italo, ma sempre 'sta stessa storia? Non ce l'ho perché mi faccio gli impacchi. Vabbè, tanto lo sapete. Fumo! Contenti? Datemi notizie». «Niente su tutta la linea. Ho visitato 15 palestre di Aosta e provincia, nessuno ha mai visto la nostra amica». Il viceispettore si sedette accanto a Lupa cominciando a carezzarla. «Io ho gli occhi che mi escono fuori dalle orbite» fece Italo, «ho guardato una ventina di siti e la maggior parte delle escort hanno il volto oscurato, altre la mascherina e poi, mi hanno detto, molte foto non sono neanche vere». «Cioè?» chiese il vicequestore. «Cioè mettono la foto di una donna nuda o in déshabillé molto bella per attirare l'attenzione, poi vai lì e invece è tutt'altro». «E chi te l'ha detto?». «Casella. Pare che abbia una certa praticaccia». «E dai la colpa a Casella, va'...» fece Caterina. «E falla finita, Caterina» le disse Italo visibilmente irritato. «Vuoi startene per i fatti tuoi, e comincia a farti i cazzi tuoi». «Oh! Embè?» intervenne Rocco. «Per favore le beghe familiari le tenete fuori dall'ufficio». «Ma quali beghe familiari, Rocco» fece Caterina con gli occhi di fuoco. «Sai quanto me ne frega a me di questo qui!». «Se Dio vuole non ho mai pagato una donna, io» rispose Italo serio. Caterina si alzò di scatto. «Io vado a...». «Tu non vai da nessuna parte, viceispettore Rispoli. Rientriamo nei ruoli!» gridò Rocco. «Questo non è un consultorio, non è un Sert e neanche un circolo del tennis. È una cazzo di questura e qui dentro siete due poliziotti. Fuori di qui fate quello che vi pare, intesi?». Caterina e Italo non risposero. «Ho detto: intesi?». Annuirono. «Buongiorno!». Si voltarono. Antonio Scipioni con la faccia soddisfatta si era affacciato alla porta dell'ufficio. «Un piccolo passo avanti» ed entrò nella stanza. «Xi Huo l'ha riconosciuta». «Chi è Xi Huo?». «Un cinese che ha il take away sushi». «Ma non è una roba giapponese?» chiese Italo. «Lascia perdere 'ste sottigliezze, Italo. Che altro ci dice Xi Huo?». «Che spesso viene a prendere da lui sushi e sashimi ma non lo mangia mai nel ristorante. Se lo porta a casa». Rocco si alzò di scatto. «Bravo Antonio. Lupa, stai a cuccia, ci vediamo dopo...» passando mollò una pacca a Antonio. «E andiamo al ristorante di Xi Huo. Caterina, tu cerca di farti una risata, con questa faccia non sei sopportabile». «E chi ha detto che mi devi sopportare?». Più che un ristorante l'attività di Xi Huo era una vetrina su via Mont Fallère. L'Anello di Giada, così si chiamava, era un negozietto di 20 metri quadrati scarsi. Sul vetro col pennarello bianco c'era scritto il menu take awei. All'interno un bancone dietro il quale spuntava la faccia di un uomo impegnato a riordinare le pietanze e alle sue spalle una porticina che doveva dare sulla cucina. Rocco e Antonio fecero il giro. «Scusa Rocco, perché non entriamo dalla porta principale?». «Perché entrare da dietro ti dà un bel vantaggio. Entri nel loro intimo, si sentono in colpa, magari qualcosa da nascondere ce l'hanno, ce l'hanno tutti, così li hai già messi in uno stato di sudditanza e rispondono alle domande senza rompere i coglioni». «Non si finisce mai di imparare» osservò Antonio. Superarono un cancelletto che dava sul cortile posteriore della palazzina. C'erano parcheggiate due auto e un motorino sgangherato. Dalla puzza di olio fritto dolciastro riconobbero subito la porta del ristorante. «Pensa abitare sopra a questi» fece Antonio, «ti devi fare una lavatrice al giorno». «Prendi casa a piazza Vittorio a Roma e stai più o meno così». Il vicequestore bussò con le nocche alla porta che poco dopo si aprì e apparve il viso tondo di una ragazza. Spaventata osservava i due uomini, particolarmente la divisa di Scipioni. «Ci chiami per favore Xi Huo?». «Sì sì... subito...» e sparì. Provò a richiudere ma Rocco aveva infilato il piede fra anta e stipite. Guardarono dentro. C'era una cucina di una decina di metri quadrati piena fino all'inverosimile di scatole, barattoli e due frigoriferi enormi. Una vecchia era seduta in un angolo e lavorava con le mani a un impasto biancastro. Ma poteva anche essere una matassa di lana. Era tutto pulitissimo, specchiato, soprattutto i due enormi microonde che troneggiavano al centro della stanza. Xi Huo arrivò sorridente. «Sì sì, eccomi qui... buongiorno». «Buongiorno signor Xi Huo». Rocco gli fece il gesto di seguirli fuori. «Fuori?» chiese il proprietario del negozio. Rocco annuendo lasciò la cucina. Appena nel piazzale, a distanza di sicurezza dall'odore dolciastro, si accese una sigaretta. «Lei ha parlato con il mio collega?». «Co suo collega, sì...» annuiva e si asciugava le mani sul grembiule macchiato. «Lei conosceva quella persona?». «Io conoscevo quella persona, sì». «E sa anche dove abita?». «No, no so se abita qui, sì». «Lei sa come si chiama?». Xi Huo guardò il cielo inseguendo un ricordo, Rocco sperò che anche lui non attaccasse con la storia delle scie chimiche. «No, io no ricordo come si chiama... Forse Daiana, Luana, Majana...». «Magliana?» chiese il vicequestore con una smorfia. «Cosa che finisce in ana, sì». «Giuliana?» suggerì Antonio. «Forse, sì». «Eliana?» provò Rocco. «Anche, sì...». «Vabbè, non ne veniamo a capo. Italiana?». «Pure italiana, bel nome». «No, non italiana di nome, italiana di nazionalità, era italiana o straniera?». «Chi?». «Come chi, Xi Huo, di chi stiamo parlando? Della stessa persona che lei ha riconosciuto dalla foto». «Sì conosciuta». «Allora, era italiana?». «Io credo, sì. Non so, sì». «Me sta a rincojoni' questo» disse Rocco all'agente Scipioni che annuì in accordo col superiore. «Era italiana sì o no?» poi si rese conto dell'assurdità della domanda. Xi Huo a malapena capiva quello che gli stava dicendo, come poteva determinare la nazionalità di Giuliana, Eliana o Majana che fosse? «Mezzo uomo mezza femmina, sì». «Va bene Xi Huo, questo lo sappiamo. Allora adesso ascoltami: veniva da sola a prendere il sushi?». «Sì sola, sì». «Bene, e non lo mangiava qui». «No, a ristorante no. A casa». «Perfetto. Tu domenica scorsa l'hai vista?». Xi Huo scrutò di nuovo il cielo. «Chiedi mia figlia. Io no qui l'altra sera...». La figlia, che era affacciata sulla porta della cucina, osservava da lontano la scena. Rocco le fece un gesto perché si avvicinasse, ma quella non si muoveva. «No può venire, lascia ristorante solo?». «Allora datevi il cambio. Va' al ristorante e fa' venire qui tua figlia, sì?». «Lei non parla molto italiano come me». «Antonio, puoi andare al ristorante?». «Io?». «Signor Xi Huo, va il mio collega e dice alla gente di aspettare, va bene?». Il ristoratore scoppiò a ridere mostrando tutti i denti gialli. «Lui guardia del corpo?». Chissà cosa ci trovava di così irresistibile, pensò Rocco, ma non glielo chiese. Antonio si incamminò verso il ristorante e la figlia, come in uno scambio di spie alla porta di Brandeburgo, si fece avanti. «Ecco mia figlia, sì. Lei era ristorante l'altra sera, sì?». La ragazza annuì. Rocco tirò fuori la foto del cadavere e la mostrò. «L'altra sera questa persona era qui?». La figlia del ristoratore sgranò gli occhi spaventata. Poi scoppiò a piangere e si abbracciò al padre. Rocco allargò le braccia e si toccò il mento imbarazzato. «Mi scusi Xi Huo, mi scusi... come si chiama sua figlia?». «Ai». «Scusami Ai...». Ai si calmò un poco. Si soffiò il naso. Poi indicò il foglio che Rocco si era rimesso in tasca. «L'altra sera... preso sushi sashimi... sola. Poi via». «Ottimo. Chiarissimo. Tu sai dove abita?». Il padre tradusse la domanda di Rocco. Ai fece no con la testa. «Ti ricordi com'era vestita? Vestita?» e si toccò il maglione per farsi capire meglio. «Fiori...» Ai toccò la camicia di Rocco, «e jeans» e si portò le mani alle cosce, come a imitare la lunghezza dei pantaloni. «Aveva la camicia a fiori e i jeans corti. Scarpe?». Ai alzò un piede indicando un punto sopra la caviglia. Mormorò qualcosa al padre che sorrise e tradusse: «Dice che aveva scarponcini neri, co lacci». «Bene, grazie». «Io so, perché vuole pure lei scarponcini co lacci viola. Dotto marti». «Dotto marti?». «Marca scarpe. Dotto marti, sì?». «Va bene, non addentriamoci. Ai, ultima domanda. Sai come si chiama?». «Juana». «Juana!». Rocco guardò il ristoratore con disprezzo. «Magliana? Ma dimme te!». «Senti un po', abbiamo telecamere dalle parti di via Mont Fallère vicino al cinese?» chiese Rocco a Caterina che stava appiccicata al computer. «No. Niente» e si stropicciò gli occhi. Italo e Casella davanti a un altro pc osservavano lo schermo. «Ma non è detto che si faccia chiamare Juana» fece Casella. «Di solito usano nomi d'arte». «Casella, ma come mai ne sai così tanto?». «Di escort? Sette anni fa, un caso a Torino e lo abbiamo risolto trovando tutto sulla rete». «Cioè non sei un habitué?» gli chiese Italo. «Pierron, io le donne non le pago. Semmai dovrebbe essere il contrario» e scoppiò a ridere. Fu in quel momento, con il sole ormai tramontato, che Deruta e D'Intino rientrarono in questura. Stanchi, distrutti, pallidi, trascinavano i piedi slacciandosi i bottoni della divisa. D'Intino crollò su una sedia, Deruta sulla scrivania. «Dotto', un massacro» fece Deruta. «Avemo cuntrullate tutti i trams che ci stanno a Aosta». «I trams?». «Voleva dire trans» corresse Deruta. «Shine. Avemo fatto nu giro de chilometri e chilometri. Tutti l'avemo cuntrullate!». «Tutti... in verità solo due» fece Deruta. «Ne abbiamo trovati solo due. Uno sta a Châtillon e fa il parrucchiere. Non ha mai visto questa persona». «N'atre invece dotto', l'avesse visto. Due metri, le spalle tante... ngulo a mammete me sa fruncate addosso...». «Ti è saltato addosso?» chiese Casella quasi ridendo. «No, è che D'Intino cercava di entrare in casa sua e quello non voleva. Allora l'ha preso per la giacca e l'ha buttato per terra». «A te D'Intino?» chiese Pierron. «Eh, a me proprie». «Deruta, vai avanti» disse Rocco. «Allora questo alto due metri ha detto che non ha mai visto la persona in foto, e che comunque coi capelli corti e gli occhi chiusi non lo riconosce. Questo... aspe' come si chiama... Demonia, ma è nome d'arte secondo me». «Veramente, Deruta?». «Sì dottore. Questo Demonia fa la drag queen». «Ma insomma, non l'ha riconosciuto manco Demonia?». «No dottore. Buco nell'acqua». «Da domani riprendete. Ah, sappiate che la nostra amica si chiamava Juana. Datevi da fare. Ora andate pure a letto, per oggi basta così». «Posso fa' una domanda, dotto'?». «Madonna, D'Intino, sei un calcolo renale! E dimmi». «Ma a questi trams... trans... quando ci parliamo li duvemo chiama' al femminile o a lu maschile?». «Secondo me» rispose Rocco calmo, «dobbiamo usare il femminile». D'Intino lo osservava attento. «Perché loro vogliono essere donne e quindi sono donne». Il viso dell'agente si aprì in un sorriso. «So' capite! Grazie dotto'». Schiavone gli mollò una pacca sulla spalla. «Niente D'Inti', va bene così». Mentre D'Intino e Deruta lasciavano la stanza, Rocco si avvicinò alla cartina. «Adesso zitti e torniamo a lavorare. Questo è il negozio di Xi Huo... che facciamo? Ci mettiamo a cercare nel raggio di quattrocento metri?». «Un po' azzardato» fece Caterina. «Siamo a nord del ritrovamento del cadavere. E siamo abbastanza sicuri che a prendere il sushi è andata a piedi...» continuava a guardare la mappa toccandosi il mento. «Dove cazzo sei...Vabbè, è tardi, sono stanco, la giornata è finita e ci devo dormire sopra. Domattina uno di voi va a vedere se ci sono registrati affitti a questo nome. Non è che di Juana ad Aosta ce ne stanno chissà quante. Lo fai tu, Italo?». «Ricevuto». «Anche se è un tentativo disperato. Io non credo che ci sia un affitto registrato». «Mi faccio anche un giro degli alberghi?». «Non serve, Italo. Signori, a casa! Ci vediamo domani a mente fresca». Italo e Casella si alzarono dalla scrivania e si stiracchiarono. Casella sbadigliò. «Tu Caterina vieni con me» disse Rocco. Il viceispettore lo seguì fuori dalla stanza. «Mi dici che hai?» le chiese sulla strada appena illuminata dai lampioni stradali. «Cosa vuoi che ti dica, Rocco. Niente». «Sei assente, sei strana. Hai avuto un giorno libero e ti sei ridotta così. Cos'è successo? Si tratta di Italo?». «Anche» il viso di Caterina era al buio, Rocco riusciva a malapena a distinguere il contorno del volto. La guardò. «Bugia». «No, è la verità. Italo, la mia vita, tante cose che non vanno. Ci sono dei giorni in cui uno fa i conti, addizioni e sottrazioni, e spesso il risultato non è positivo». «Alla tua età fai i conti?». «Non dovrei?». Passò un'auto e per un momento i fari illuminarono il volto della ragazza. Aveva gli occhi lucidi. «Io dico che c'è qualcosa che ti fa stare male. Qualcosa che non mi vuoi dire, e io non voglio sapere. Però ti do un consiglio, perché ti voglio bene. Alla tua età puoi fare quello che ti pare. Puoi diventare quello che vuoi. Qui, in petto, mettiti un segno positivo, come le batterie. Sennò è la fine. Sei un aereo che decolla, Cateri'. E devi decolla'!». Caterina tirò su col naso. «Pigliati 'sta laurea e iscriviti alla scuola di commissari di PS a Roma. Che fai qui ad Aosta? Muoviti! C'è un mondo. Parti! Guarda, te li do io i soldi del biglietto, ovunque ti va di andare». Finalmente Caterina abbozzò un sorriso, il primo della giornata. Alzò il viso che prese luce dal lampione. Puntò lo sguardo dritto su Rocco. «Tu vieni con me?» disse mangiandosi le parole. «Non ho il fiato, Caterina. Mi piacerebbe, ma non ho il fiato». «Io dico che ce l'hai. Ma fai solo finta di non averlo». «Non girare la frittata. Stavamo parlando di te». «E a te, Rocco? Cos'è che ti blocca?». «Poni male la questione. La domanda giusta sarebbe: cos'è che potrebbe darti una spinta, Rocco?». «Io?» rispose Caterina senza staccare gli occhi da quelli del vicequestore. «Tu non sei una spinta, Caterina. Tu rischi di essere una scossa tellurica. E io non sono pronto. Rovinerei la tua vita. Mica ce lo meritiamo». «Sei noioso e bugiardo. Io me ne vado a casa, Schiavone. Ci vediamo domani. Come dicevi poco fa? A mente fresca». Si voltò e a passo spedito si incamminò verso il corso. Rocco la tenne d'occhio fin quando non voltò l'angolo sparendo alla vista. Poi guardò Lupa. «Se ne annamo a casa amore piccolo?» la cagnolina abbaiò e si incamminarono. «Ammazza se perdiamo colpi, cucciola!». Si era tolta la divisa per non dare troppo nell'occhio. Aveva superato la portineria, poi era entrata nel corpo dell'edificio. Sulla legenda scoprì che chirurgia era al secondo piano. Lenta salì le scale con il respiro mozzato in gola. Neanche si ricordava il tragitto da casa sua all'ospedale, lo aveva percorso quasi in trance, senza porsi troppe domande. Doveva calmare tutte quelle vocine che le fischiavano nelle orecchie, che qualcuno chiama coscienza, Caterina invece coercizione educativa. Sin dalle elementari, nella scuola delle suore. Buone quelle. Ricordava suor Ernestina, era austriaca, e la costringeva a mangiare la pastina in brodo che a lei faceva schifo. O le punizioni corporali per aver dimenticato una capitale o una tabellina. Una volta era scoppiata a ridere durante l'Ave Maria, perché una sua compagna aveva fatto la faccia più buffa del mondo. Suor Ernestina l'aveva spedita in direzione e le aveva fatto saltare il pranzo. Tanto a casa non è che le cose andassero meglio. Passava dalla padella alla brace, dall'orrore della sua famiglia all'inferno della scuola. Coercizione educativa. Caterina non aveva mai avuto il coraggio di ribellarsi, alle suore prima, ai professori poi. Sempre quelle stramaledette vocine che le dicevano: il cuore di Gesù sanguina se dici questa cosa, il cuore di Gesù piange se non fai quest'altra, il cuore di Gesù se ne va se non obbedisci. Vergognati. Vergognati. Vergognati. Dio, patria e famiglia. Ridicolo come di tutti questi valori, nel cuore, Caterina non ne riconoscesse uno. Mamma e papà neanche a parlarne, Gesù la Madonna e Dio non ci credeva e dell'Italia erano più le volte che si vergognava che quelle di cui ne andava fiera. Però poi arrivavano le vocine, e la mandavano a messa a Natale, la spedivano a partecipare alla prima comunione dei figli di qualche cugino, le facevano salire le scale verso il reparto di chirurgia a guardare un padre che chiamare tale era un azzardo. «Prima il dovere poi il piacere» dicevano i suoi insegnanti. E ora che ci pensava, salendo i gradini, lei il piacere non l'aveva mai conosciuto. Tutta la sua vita non era stato altro che dovere. I compiti e poi addolcire le lacrime di sua madre, ancora i compiti e di corsa a fare la spesa che sua madre dal letto proprio non riusciva ad alzarsi, studiare e poi in cucina a preparare la cena per entrambe, andare a lavorare e ritirare lo stipendio per mantenere lei e sua madre. Non era mai stata bambina. Non era mai stata ragazza. S'era ritrovata donna senza poterlo neanche raccontare in casa, temeva gli schiaffi della madre e le mani di suo padre. E gli errori li aveva pagati sulla sua pelle. Si era ritrovata a 19 anni arruolata in polizia senza aver mai passato una serata in discoteca, un amorino adolescenziale, niente. Alfredo fu il primo amore. Guidava camion. Fece un solo errore, Alfredo de Joix, quello di alzarle le mani. Almeno essere una poliziotta a qualcosa era servito. Superò la porta del reparto. Lontano nel corridoio un paramedico trasportava un carrello. Sbirciò dentro un paio di stanze. Alla terza incrociò un'infermiera sorridente. «Mi scusi... cercavo Guido Rispoli...» le disse. Quella strizzò gli occhi. «È sicura che sia in chirurgia?». «So che deve operarsi al cuore». «Allora deve salire un altro piano. Terapia intensiva» poi guardò l'orologio. «Se si sbriga la lasciano passare. Lì più di un'ora per le visite non si può». Ancora un piano. A destra sopra una porta chiusa c'era scritto «Terapia intensiva». Caterina suonò il campanello. Attese qualche secondo finché attraverso il vetro smerigliato vide un'ombra avvicinarsi. La porta si aprì. Un infermiere con la barba la guardò con il viso stanco. «Dica?». «Cercavo Guido Rispoli...». Quello annuì. «Ultima stanza a destra. Fra dieci minuti però chiudiamo...». Nel reparto c'erano solo sei camere con le porte a doppia anta e un rettangolo di cristallo per guardare all'interno. Arrivò all'ultima. Si affacciò. Suo padre era steso sul letto. Aveva una flebo infilata nel braccio destro, un monitor sopra la spalliera segnava il battito cardiaco. Teneva gli occhi chiusi e respirava piano. Sembrava più piccolo del suo metro e ottanta. La luce nella stanza era piatta e uniforme come gelatina. Eccolo lì, si disse Caterina restando dietro il vetro, come osservando un acquario con dentro il pescecane più feroce dell'oceano. Teneva le mani bianche poggiate sulle lenzuola. La barba di due giorni, i capelli un po' imbiancati e ricci. In quella stanza il tempo sembrava si fosse fermato e la memoria di Caterina si riavvolse, come un nastro. E tornò indietro, al suo dodicesimo compleanno. Le avevano comprato la torta ma la madre aveva sbagliato, di candeline ne aveva messe solo undici. Regali non ce n'erano, solo quel dolce di cioccolata con la panna sopra e neanche un'amica accanto. Sul tavolo della cucina i resti della cena. «Tanti auguri a Caterina» cantava la madre mentre il padre la inquadrava con una macchinetta fotografica usa e getta. Dopo, seduto sul letto in camera matrimoniale, suo padre le aveva preso le mani e la guardava negli occhi. Aveva l'alito di sigaretta. Sua madre invece, nell'altra stanza, se ne stava seduta sulla poltrona verde davanti al televisore con il volume al massimo. «Sei una donna ormai, Caterina» le aveva detto suo padre, «e devi stare attenta. Le cose più brutte succedono proprio a quelle carine come te...». Le sfiorava le braccia, le carezzava le mani. «Vieni qui, fatti abbracciare, piccola mia...» e se l'era stretta al petto. Non sapeva se doveva stringerlo anche lei, e glielo disse la vocina: abbraccialo, è tuo papà! così lo fece. Non lo strinse. Solo gli poggiò le braccia sulle spalle. Puzzava di sudore, suo padre, e di liquore. Lui la stringeva sempre di più perché, diceva, «Ti voglio tanto bene...». Non l'aveva mai stretta così, solo una volta, tanti anni prima, ma lei lo ricordava appena. «Lo sai che papà ti dice le cose per il tuo bene... e io e te dobbiamo mantenere un segreto...». «Quale?» gli aveva chiesto. «Un segreto di un padre e una figlia». La costrinse a poggiare la testa sulla sua spalla. Poi si accorse che le braccia muscolose non le cingevano più le spalle. «Lo sai che tutti i papà hanno un segreto con le proprie figlie, e allora anche io e te adesso ne avremo uno!». Le mani del padre erano sparite, non le toccavano più la schiena. Piano, come un vento leggero, sentì qualcosa che le sfiorava i glutei. «È per il tuo bene...» mormorava. «Sì papà...». Dai glutei il vento leggero passò sotto la gonna. «E ti devi fidare, perché io ti insegno a fare le cose per bene, a non aver paura delle brutte persone...». Caterina non capiva. Di cosa doveva aver paura? «Sì papà...». Poi le prese una mano. Le alzò il viso per guardarla ancora negli occhi. «Ti fidi di papà?». «Sì...» aveva risposto lei con la voce strozzata. «Brava bambina mia, brava... questo sarà il nostro segreto». Le aveva preso la mano, delicatamente, e se l'era portata in basso. Sempre più in basso. Finché Caterina l'aveva sentito. Guardò, e si accorse che stringeva in mano il cazzo di suo padre. «Ora fai la brava e...». Fuggì dalla camera da letto, mentre il padre rideva alle sue spalle. Attraversò di corsa il salone. Sua madre stava lì, gli occhi incollati al televisore. Scappò in camera e si chiuse dentro, mise anche una sedia per bloccare la porta. Poi pianse per tutta la notte. Non sapeva che quella non sarebbe stata l'ultima, solo la prima notte dei segreti. Ne seguirono altre tre, fino a quando suo padre sparì da casa dopo aver lasciato la madre a terra con un braccio rotto e gli occhi pesti e viola. E intervenne il pezzo grosso, come lo chiamava sua madre. Un amico di suo zio che lavorava in polizia, a Roma, che impedì al padre di riavvicinarsi alla figlia. Fu grazie a lui che Caterina si trovò con una divisa blu e un lavoro che non le era mai passato per il cervello. Ma l'amico di suo zio, l'allora commissario Mastrodomenico, era su un piedistallo agli occhi di Caterina, avrebbe fatto tutto quello che le chiedeva, era la sua ancora di salvezza in un mare molto agitato. Erano passati tanti anni e ora suo padre era lì, steso sul letto della terapia intensiva. Perché c'era andata? Per vederlo ancora una volta? Voleva guardarlo negli occhi sperando fosse l'ultima volta e che Dio, Gesù o qualsiasi cosa ci fosse lassù se lo riprendesse per sempre. In quel momento, dall'altra parte del vetro, il padre voltò il capo e la vide. Eccoli lì, pensò Caterina. Quegli occhi neri e cattivi, che ancora si sognava la notte, che avevano perseguitato la sua vita fino a quel momento. Il padre le sorrise appena e cercò di alzare una mano. Le fece cenno di raggiungerlo. Caterina mosse qualche passo e finalmente entrò nella stanza. C'era puzza di disinfettante misto a brodo vecchio. «Ciao...» le disse. Caterina si fermò ai piedi del letto. Guardò la sovraccoperta bianca e azzurra con il logo dell'ospedale, poi il monitor che segnava pressione sanguigna e battito cardiaco, il dito indice del padre morso da una pinza, la vena del braccio livida con l'ago cannula infilato dentro. «Sono contento che mi sei venuta a trovare. Ti sei fatta grande». «Succede quando si cresce». Guido Rispoli strinse le labbra. «Mi passeresti un po' d'acqua?». «No» fece Caterina senza muoversi. «Devo chiamare l'infermiera?». «Se lei te ne offre un sorso...». «Lo sai che io domani...». Caterina alzò un braccio per interromperlo. «Fermati! E chiariamo subito. Che tu muoia o no non dipende da me. Qualche anno fa ti avrei augurato la morte, oggi sinceramente non me ne frega più niente. All'inizio, tanti anni fa, mi davo anche la colpa di non meritare il tuo amore. Pensa un po'. Sei stato un padre schifoso, e questo ci tenevo a dirtelo. In tutti questi anni ci ho pensato, ho provato a trovare una scusa, un motivo, ma poi ho capito che sei solo un uomo malato. E non mi riferisco al tuo cuore, non ne hai mai avuto uno. Hai reso la mia vita un inferno, mi hai fatto cose che non basta una vita per chiedere scusa e per pentirsi, e non ti è mai venuto in mente di chiedermi scusa. Mai. Ora sei qui, su questo letto, tra la vita e la morte e vuoi chiarire tutto con me? Come fai a chiarire tutto? Spiegamelo». Guido non rispose. Chiuse gli occhi e una lacrima scese giù lenta superando gli zigomi per fermarsi sulle labbra. «Sapessi quante volte avrei avuto bisogno di te. Perché io di te mi fidavo. Come hai potuto? Dimmelo, per favore, come hai potuto?». Ma Guido restò in silenzio. «Allora non abbiamo altro da dirci, mi pare. Sappi che al tuo funerale non ci sarò. Se dovessi campare non venirmi a cercare». «Sei venuta solo per vomitarmi addosso tutto questo schifo?». «Che non è niente rispetto a quello che hai fatto tu. No, sono venuta per un altro motivo...» si avvicinò di due passi. Si chinò fino a guardarlo dritto negli occhi. Li studiò con molta attenzione. «Che fai Caterina?». «I tuoi sono occhi normali, lo sai? Vedo la cornea, la pupilla marrone, qualche pagliuzza gialla in mezzo... e cosa sono questi cosini rossi? capillari? ah no, aspetta, mi pare si chiamino vasi retinici, sì... tutti rossi nella sclera». Si tirò su. «Sono occhi normali tutto sommato. Niente di che. Neanche tanto belli. No, non somigliano ai miei. Niente che vedo su questa faccia somiglia a me. Sono occhi spaventati, lo sai? Fai bene ad avere paura. È una cosa sana, la paura, almeno così dicono. Io lo so, ci ho vissuto insieme per un sacco di tempo. La sai la cosa positiva? Se muori sotto i ferri non te ne accorgi, ti addormentano prima. Goditi l'anestesia, che sia breve o lunga dipende dal chirurgo e da un po' di fortuna. Ma tu quella l'hai sempre avuta, no?». Si voltò e uscì dalla stanza. Il beep del battito cardiaco era aumentato d'intensità. Quando lasciò la terapia intensiva le scappava da ridere. Le venne voglia di fermarsi al pub a prendere una birra. Scura. Una Guinness. Tanto quegli occhi non li avrebbe sognati mai più. Aprì il cartoccio del giorno prima. Il kebab era secco e anche a scaldarlo c'era poco da stare allegri. Puzzava pure di marcio. Rocco fiondò la vaschetta nella pattumiera. Decise che avrebbe cenato con una bottiglia di vino rosso, di uscire non gli andava. Si stravaccò sul divano e accese la televisione. Saltava da un programma all'altro carezzando con la destra Lupa che aveva sbranato i croccantini e sonnecchiava felice. Abbandonò i canali principali e si perse nella selva delle televisioni misteriose. Documentari, film sconosciuti, spose che gareggiavano per il matrimonio più bello, gente che cucinava sul picco di una montagna, una ragazza col seno di fuori che invitava lo spettatore a comporre un numero promettendo chissà quali esperienze erotiche, una partita di calcio di una serie infima. «Porca!» si picchiò con il palmo sulla fronte. «Sono un deficiente! Vieni Lupa» si alzò e uscì dall'appartamento. Bussò alla porta di Gabriele. Il ragazzo aprì dopo qualche secondo. «Buonasera». «Tu hai internet». «Sì». «E hai un computer?». «Pure». «Mi servi» e seguito dal cane il vicequestore entrò nell'appartamento. «C'è mamma?». «No, dorme a Milano anche stanotte. Ha cenato?». «Sì». «Io pure. Mi sono fatto la pasta col pesto». «Sticazzi Gabriele. Forza, andiamo al computer». Gabriele gli fece cenno di seguirlo, poi entrarono nell'ultima stanza di un corridoio con le pareti grigie e marroni. Se l'appartamento era ordinato, minimal, vuoto, quasi giapponese, la camera del ragazzo era l'esatto opposto. Un campo di battaglia. A terra c'erano calzini, scarpe spaiate, giornaletti della Marvel. Sul letto sfatto una chitarra elettrica e custodie di videogiochi. Una piccola libreria ospitava i testi scolastici e qualche romanzo, tutti intonsi, forse mai aperti. Un'altra libreria, molto più grande, accoglieva allineati in un ordine maniacale centinaia di dvd. A dare un'occhiata veloce erano tutti film horror. Lo stereo era poggiato su un mobiletto cinese e intorno c'erano impilate decine di cd. Sulla scrivania un portatile ultima generazione. «Ecco, mi scusi il disordine». Aleggiava un odore di gomma dolciastra mista ad ascella e maglietta umida. Rocco andò ad aprire la finestra. Dava su un cortile interno. «Sente caldo?». «No. C'è una puzza che accora. Forza dai». «Che dobbiamo cercare?» chiese Gabriele sedendosi al tavolo. Rocco prese una sedia e si mise accanto a lui. «Annunci di escort ad Aosta». Gabriele lo guardò stupito. «Non è per me, idiota. Cominciamo». «Sicuro? Va bene. Io sono un treno su Google, sa?». Cominciò a digitare. Aprirono il primo sito. «Ecco qui. Che cerchiamo?» chiese osservando la pagina web. «Escort, Sadomaso, Uomini per uomini, Uomini per donne, Trans?». «Metti escort e trans. E salta donne e maschi». «Cerchiamo un frocione?». «Cerchiamo un trans, imbecille». «Ma chi è?». «Tu fatti i fatti tuoi». «Ma sappiamo il nome?». «Sì, ma non serve perché usano quasi tutti dei nomi d'arte. Quindi datti da fare. Iniziamo dalla lettera A. Lady Angela per esempio. Doppia penetrazione eccetera eccetera. Questo è un trans. C'è un cellulare?». «Mi pare bruttina». «Gabriele, non stiamo cercando compagnia per la serata, lo vuoi capire?». Rocco afferrò il Nokia e compose il numero di Lady Angela. «Pronto?» rispose una voce bassa e cavernosa. Rocco riattaccò. «Bene, non è Lady Angela». «Perché?». «È viva. Avanti». «C'è questa Contessina Barbara. Che poi secondo me non è una contessina». «Sei serio, Gabriele?». Il ragazzo alzò le spalle. Rocco sbuffò. «Vai col cellulare». Chiamarono anche Contessina Barbara. Rispose una segreteria: «Ciao sono Contessina Barbara. Sono speciale, raffinata, capricciosa, ho lo studio attrezzatissimo e sono molto vogliosa. Mi puoi trovare a via Edelweiss la sera fino alle 21... ciao...». Rocco prese l'appunto. «Bene, via Edelweiss, poi lo controlliamo. Prosegui». «Ma vuol dire che sto prendendo parte all'indagine?». «No, vuol dire che io non ho internet in casa e ti sfrutto. Tu sei solo forza bruta, un braccio, poco più di un oggetto. Forza, adesso questa... Brenna... dimmi il numero». Chiamò Brenna. Anche lì rispose una segreteria: «Ciao sono Brenna» l'accento era slavo. Rocco si concentrò nell'ascolto. «Sono pronta a tutto. E ti aspetto fino a venerdì in via Larsinaz». Proseguirono per una buona mezz'ora navigando su diversi siti. Finito di spulciare anche quelli più estremi, Rocco si ritrovò con diciotto indirizzi e l'orecchio che bruciava. Gabriele si abbandonò sulla sedia. «Non credevo che ad Aosta ci fossero tutte queste prostitute. Che cos'è la pioggia dorata?». «Una cosa che fa schifo». «E la doppia penetrazione?». «Pure». «Che significa padding estremo?». «Gabriele, mi deconcentri. Fa' una cosa, chiedilo a tua madre». «Va bene. Abbiamo finito?». «No. Ora apri la mappa di Aosta. Cerca via Mont Fallère per favore. Da lì calcoliamo la distanza di ognuno di questi indirizzi, e vediamo quelli più vicini. Forza, che voglio andare a letto». «Posso dormire da lei pure stanotte?». «Vediamo come ti comporti... cominciamo da Contessina Barbara, via Edelweiss». Vicino al take away L'Anello di Giada c'erano solo tre indirizzi buoni. Quello di Brenna, a via Larsinaz, quello di Sonya, a via Brean, e infine quello di Jesica con una esse a via Capitano Chamonin. «Ottimo lavoro, Gabriele. Sei stato bravo, più bravo di molti dei miei agenti». «Allora dormo da lei?». «Dormi da me. Però sul divano, non fare che a metà notte ti infili nel mio letto. Puzzi, tiri calci e io mi devo riposare». L'alba era spuntata da poco e l'aria era pulita e fresca. Arjan Vlora si appoggiò al pino per schiacciare un'oretta di sonno. Il gregge brucava tranquillo l'erba ancora bagnata di rugiada e i due maremmani, che Arjan aveva ribattezzato Bogdani e Berisha, in onore di due calciatori suoi connazionali, controllavano il pascolo trotterellando e mischiandosi alle pecore. Lontano si sentiva il muggito della Pontina, a quell'ora il traffico dei pendolari era già impazzito, ma ad Arjan non lo riguardava. La città non era più un suo problema. Ormai lavorava da anni all'azienda di Castel di Decima, e quel lavoro non l'avrebbe mollato mai. Portare fuori le pecore, caricare le taniche di latte, badare ai peschi e ai mandorli, controllare i formaggi in stagionatura e riportare le pecore a nanna. Una pacchia. Soprattutto per lui che era arrivato nel 1994 in Italia coi carichi disperati dalla campagna di Fier convinto di trovare chissà che. Ma in fondo ebbe fortuna. Trovò l'unico lavoro che sapesse fare bene, in campagna. E con la paga che gli dava Alessio De Sisti, quasi 800 euro al mese più vitto e alloggio, in neanche due anni sarebbe tornato in Albania per comprare la casa dei Balaj, e avrebbe finalmente messo su famiglia. Ci aveva impiegato un po', ma alla fine l'Italia gli aveva dato un futuro. Chiuse gli occhi e il rumore della strada si allontanò, restarono solo i pochi campanacci delle bestie e l'abbaiare di Berisha. Ma abbaiava troppo. Riaprì gli occhi. Il cane latrava con un tono acuto, d'allarme, non quello basso, cupo di difesa, e Arjan, che coi due maremmani conviveva da sei anni, lo capì immediatamente. Li cercò con lo sguardo senza trovarli. Il latrato proveniva dal limitare del pascolo, vicino ai cespugli di more che Arjan depredava ogni fine estate per fare la marmellata. Si tirò su. Scorse il codone peloso e bianco del pastore maremmano che si agitava, mentre il resto del corpo era infilato nei rovi. Fischiò, ma quello non rispondeva. Bogdani, attratto, corse dal compagno. E come Berisha anche lui infilò mezzo corpo nelle spine agitando la coda. «Ma che... Bery! Bogda! Qui!» fischiò ancora. L'unica a dargli retta fu una delle vecchie pecore che alzò la testa masticando. Lo osservò per un po', poi si rimise a brucare. «Bery! Bogda! Qui!». Niente. Non si muovevano. Forse avevano trovato un serpente. O forse una volpe. Ma le volpi a quell'ora non giravano, al massimo poteva essere una volpe morta sotto una macchina e gettata lì da un autista distratto. E poi i cani non si fissavano mai con le carogne, coi serpenti o coi gatti. Berisha e Bogdani guardavano le pecore. Quello avevano sempre fatto, e quello sapevano fare. Niente li poteva distrarre. Figurarsi un biscione o un corpo di qualche animale del bosco sbudellato la notte prima. Si alzò. Ora la visuale era più chiara. I due cani erano proprio dentro il cespuglio a monte del pascolo, vicino alla rete di recinzione rotta in più punti, che prima o poi doveva decidersi a riparare. Lasciò la sacca e la radiolina che portava sempre con sé e sbuffando si avviò. Un paio di pecore si scansarono per farlo passare. «Bery! Bogda! Ma insomma!». Attraversò quasi tutto il prato e finalmente li raggiunse. «Si può sapere che avete da abbaiare?». Dapprima pensò a dei vestiti abbandonati, ma poi notò che da una manica del giubbotto spuntava una mano. Sentì una scarica che gli fece rizzare tutti i peli del corpo e il cuore mettersi a battere contro le costole. «Che cazzo...». Nel cespuglio di more c'era un uomo. Un taglio netto e preciso come un sorriso gli attraversava il collo da parte a parte. «Dreqi ta haje!». Quando il vicequestore Schiavone e l'agente Pierron bussarono alla porta di Brenna a via Larsinaz erano le sette e mezza del mattino. Dall'appartamento non un suono, non un movimento. Una macchina passò veloce e un cane abbaiò da un balcone. Rocco si guardò intorno. Erano tutte palazzine a due piani, identiche, porte tutte uguali, nessun condomino sembrava dare segni di vita. Suonò di nuovo il campanello. «È presto. A quest'ora starà dormendo» disse Italo. «Sticazzi, si deve svegliare» pigiò e lasciò il dito sul pulsante per almeno dieci secondi. «E se avessimo fatto centro al primo tentativo?». «Allora bisogna che sfondiamo perché di là non ci sarà nessuno». Non ce ne fu bisogno. Qualcuno raspava dall'altra parte della porta. «Chi è?» una voce debole, appena udibile. «Polizia. Apra!». Dopo un paio di secondi sentirono la chiave girare nella toppa. Un paletto di sicurezza rimosso e finalmente la porta si aprì. Apparve Brenna. Aveva i capelli biondo platino attaccati alla fronte. Il viso che aveva subito vari interventi era sudato, pieno di macchie. Indossava una maglietta di Superman e un paio di pantaloni della tuta. Gli occhi pesti e le labbra gonfie, sembrava che qualcuno avesse appena smesso di picchiarla. «Che c'è? Io ho il permesso... lo vado a prendere?». «No. Conosce questa persona?». Rocco allungò il foglio con la foto di Juana presa sul lettino autoptico. Brenna la guardò con gli occhi stanchi. «No, è quella del giornale?». «Sì. Non la conosce? Non l'ha mai vista?». «Vi giuro di no. Sennò venivo da voi a dirlo... è morta strangolata, vero?». Fu Pierron a rispondere: «Sì. Lei ha qualche cliente che potrebbe aver fatto una cosa simile?». Brenna sorrise. Si passò la mano davanti alla faccia. «Volete un caffè? Tanto ormai non mi addormento più...» e fece spazio per farli accomodare. Li aveva portati nel cucinino, piccolo e accogliente. Un gatto tricolore si strusciava sulla gamba di Rocco. «Buono Felix... è molto affettuoso, ma non gli fate le coccole sennò non ve lo staccate più. Volete lo zucchero?». «No grazie, va bene così» rispose Rocco. Brenna gli porse la tazzina, poi si mise seduta e si rassettò i capelli. Aveva il viso teso. «Siete qui per denunciarmi, vero?». «Brenna, non me ne frega niente di come si guadagna da vivere. Non sono della buoncostume. Sto cercando un assassino» disse Rocco e il viso della prostituta sembrò rilassarsi. «Ci penso da quando ho letto la notizia. Poteva succedere a me». «A proposito, lei per caso conosce...» Italo prese il taccuino e lesse «Jesica con una esse sola o Sonya? Jesica abita a...». «Sì, la conosco. Abita a via Capitano Chamonin, qui vicino. L'ho vista ieri. E parlavamo proprio di quella poveretta che avete trovato nel fiume». «Lei non è straniera?». «Sì che lo sono. Sono di Bratislava. Ma al telefono esagero l'accento. Piace di più» e diede un morso a un biscotto al cioccolato. «Io però clienti così mai avuti. Cioè strani, parecchi. C'è quello che si veste da cameriera, quello che vuole lo strap-on...». Italo fece una faccia poco convinta. Rocco lo rasserenò: «Ti spiego dopo». «Poi i soliti manette e fruste e cose così. Ma tutta gente che gioca. Finisce sempre con una risata». «Lei è un'esperta». «Abbastanza» finì il biscotto. «Faccio questo lavoro da quindici anni. Qualsiasi pratica, e costo bei soldi». «Un'idea?». «Dalle cento per una seduta si può arrivare a più di cinquecento per tutta la notte. Soprattutto se devo andare a casa di qualcuno o in albergo, la tariffa sale». «Emette fattura?» chiese Rocco e il viso di Brenna divenne di pietra. «Come?». «Era una battuta» la rassicurò Rocco. «Lei esercita in camera da letto?». «No. Di là. Nell'ex salone. Volete vedere?». «E perché no?» fece Rocco. Si alzarono. Proseguirono nel piccolo corridoio, entrarono nell'ex salone rivestito interamente di moquette rossa. La puzza di deodorante alla rosa stringeva le narici. Appesi alle pareti gli attrezzi del mestiere. Fruste, canne di bambù, manette, catene, museruole. Su un tavolino un'intera collezione di falli di plastica. «Ecco, questo è il mio ufficio». «Bello. Non vedo il fax». Brenna scoppiò a ridere. «Grazie, ci accompagna fuori?». «Ma tu pagheresti cento euro per andare con una così?» chiese Italo tornando alla macchina. «Il caffè lo fa buono». «E adesso?». «Adesso andiamo da Sonya, Jesica con una esse ieri era viva, per ora lascerei perdere... e lascerei perdere pure la macchina, via Brean sono neanche 5 minuti». Non si dissero una parola. Rocco osservava le strade, i palazzi. Italo fissava il marciapiede. L'agente sapeva che quando il vicequestore si azzittiva non andava disturbato. A farci attenzione si poteva sentire il rumore dei suoi pensieri. Arrivarono a via Brean. «Il nostro è al 12?» chiese Rocco leggendo i civici. «Sì, al 12. Ma non lo trovi strano, Rocco?». «Cosa?». «Stiamo cercando l'identità di una persona che non era poi neanche tanto certa della sua». «Perché, sei così sicuro della tua identità?». Pierron si grattò la testa. «Penso di sì...». «Sai tutto di te?». «So parecchio». «Parecchio non è tutto. E a dirtela non è neanche abbastanza». Erano arrivati al 12. Citofonarono allo Studio Emme indicato da Sonya nell'inserzione. Non rispose nessuno. «Starà dormendo pure questa?». «Probabile». Suonarono ancora. «Qui dice che lo Studio Emme è l'interno 3». Rocco indietreggiò di un paio di passi per dare un'occhiata alla palazzina. Due piani. A destra una banca. A sinistra altre palazzine e un ferramenta. «Niente Rocco, questa non risponde». Il vicequestore citofonò ad un interno a caso. Dopo pochi secondi rispose una voce di donna: «Sì?». «Questura di Aosta, apra il portone». Dall'altra parte silenzio. «Signora, mi ha sentito?». «Chi mi assicura che siete proprio della polizia?». Rocco alzò gli occhi al cielo. «Che palle! Si affacci e guardi la divisa del mio agente». La signora Cosma, così recitava la targhetta sul citofono, poggiò l'apparecchio che sbatté rimbalzando sul muro. Stava andando alla finestra. Italo s'era messo in mezzo al marciapiede a braccia larghe per farsi osservare. Intravide una mano scostare le tende e una testolina bianca. Poco dopo la donna tornò all'apparecchio. «Va bene, ma cercate me?». «No. Apra, signora!». Finalmente si convinse e i due poliziotti entrarono nella palazzina. Salirono solo due scalini e arrivarono al primo piano davanti alla porta dell'interno 3. Fu Italo a suonare il campanello. Non accadde nulla. Suonò ancora. Come in alcune case di ringhiera a Milano, la finestra dell'appartamento dava sulle scale. Rocco si affacciò ma il buio impediva la visuale all'interno. Bussò con le nocche sul vetro. E invece si aprì la porta dell'interno 2, dalla quale uscì un ragazzo fresco di doccia. Vide i due poliziotti e trasalì. «Salve... lei conosce la persona che abita qui, Sonya?» chiese il vicequestore. Il ragazzo alternava lo sguardo tra Rocco e Italo, come fosse spaventato che quelli potessero estrarre un'arma e farlo fuori. «Sì... cioè non la conosco» rispose con la voce fragile. «L'ho incontrata qualche volta». Rocco fece segno a Italo che tirò subito fuori la fotografia del cadavere. La passò al vicequestore che la sottopose all'attenzione del ragazzo. «Mi dica un po', è lei?». Il ragazzo impallidì. «Mio Dio, sì. Che succede?». Rocco riprese il foglio. «Pare che l'abbiamo trovata, Italo. Chiama in questura. Voglio Caterina, Casella e voglio anche la Gambino con tutta la squadra». «Ricevuto». «Lei può andare». «Ma che... che succede?» fece il giovane. Finalmente concentrò l'attenzione sul viso del ragazzo. Gli occhi sgranati e cerchiati di nero, chiari e spaventati, la bocca piccola e sfuggente sotto il naso importante e le orecchie a sventola. Rocco lo classificò come un Lemuriformes. Si muoveva a scatti, proprio come quei primati endemici del Madagascar, oscillando fra il terrore e la curiosità. «Non succede niente di buono. Tanti saluti». Il lemure sgusciò via agile. Rocco invece si mise la mano in tasca e tirò fuori il suo coltellino svizzero. Speranze che Sonya o Juana aprisse la porta non ce n'erano più. Arjan aveva chiuso i cani nel recinto e appoggiato all'albero stava raccontando per la terza volta all'ispettore Morganti come era avvenuto il ritrovamento. Le pecore ruminavano tranquille, ogni tanto alzavano la testa ad osservare. A cento metri di distanza, accanto al cespuglio, due uomini della scientifica, incappucciati nonostante il caldo già cominciasse a farsi sentire, ispezionavano foglie e rami. Il corpo dell'uomo era lì a terra, il medico chino sul cadavere, il vicequestore Bonanni in piedi leggeva distratto messaggi sul cellulare. Sulla strada le auto della polizia facevano largo al furgone della mortuaria. «Che le devo dire, Bonanni? Causa della morte un bel taglio sul collo, profondo». «Mmm» rispose quello con gli occhi sul display. Uccio Pichi riprese: «E non è stato ucciso qui». «A quello ci arrivo. Non c'è sangue. Documenti?». «Macché, niente documenti, niente portafogli, niente orologio». Finalmente Bonanni distolse l'attenzione dal telefonino. «Sicuro?». «Al cento per cento». «A me puzza di regolamento di conti». «Ne ha tutta l'aria». Uccio rovistò la tasca interna della giacca. «Qui c'è uno scontrino... vediamo un po'» si avvicinò il foglietto agli occhiali. «Bar Mastrangeli, via Vitellia 31... ha preso un caffè. La data è quella di due giorni fa». «Chissà quando l'hanno seccato» e Bonanni si rimise il telefono in tasca. «Non lo so. Me lo porto in ospedale e magari glielo so dire». «Morganti!» urlò il vicequestore verso l'ispettore. «Accendi la macchina, torniamo in ufficio. Avete finito col pastore?». Il poliziotto annuì e si incamminò. «Appena hai nuove, Uccio, mi chiami?». «Sicuro... sarà un casino capire chi era». «Non sottovalutarci, dottore». La casa di Juana era vuota, solo tre mobili vecchi, neanche un libro, un soprammobile e neppure il televisore. Pavimento nudo come le pareti, in cucina solo un cucchiaino poggiato sul lavello. In camera da letto c'era il materasso nudo e l'armadio aperto senza vestiti. Rocco tornò in salone. Nel mezzo il parquet era scolorito, segno che c'era stato un tappeto. Al centro del soffitto un gancio dove forse era attaccato un lampadario e tre piccoli chiodi alle pareti che dovevano aver sorretto qualche quadro o fotografia. Italo si affacciò sulla porta. «Che succede?». «Niente. Non succede niente. Vuoto, sembra una casa da affittare. Risaliamo al proprietario e facciamogli un po' di domande. Quando arrivano gli altri?». «Sono in strada». Rocco cominciò ad annusare. «Cosa senti?». «Io? Niente. Che sento?». «Fiori. Ma non fiori veri. È roba chimica». «Fa venire i brividi questo posto». «A me fa solo girare i coglioni». In cucina c'era una porta-finestra. Rocco la aprì. Si ritrovò in un minuscolo giardino pavimentato circondato da una recinzione bassa e qualche vaso che ospitava rami secchi. Un rubinetto gocciolava. Si affacciò al di là del piccolo steccato di legno. C'era un parcheggio con poche auto. Ritornò nell'appartamento. Italo era rimasto sull'uscio. «Va bene Italo, aspettiamo giù la compagnia». Scesero a piano terra. Rocco guardò la cassetta delle lettere. «Bene, ci sono 8 interni. Cominciamo a sentire la gente. Vediamo di saperne di più. Partiamo dal secondo piano con la signora Cosma?». In quel momento il portone della palazzina si aprì e riapparve il lemure. «Mi sono dimenticato una cosa» si giustificò e rapido salì i due scalini per tornare nel suo appartamento. Riapparve dopo neanche un minuto mostrando il cellulare ai poliziotti. «Ecco, ormai senza non ci sappiamo più vivere, no?». «Mi scusi, posso sapere come si chiama?» chiese Rocco. «Chi io?». «Se vuole può dirmi anche il nome di sua nonna o di suo padre, ma per ora preferirei il suo». «Certo. Ivano Petrulli». «E mi dica, Ivano. Tre giorni fa, domenica sera, lei era a casa?». Ivano sgranò gli occhi. «Domenica?». «Domenica». «Mi pare di no. No, domenica sera non ero a casa. Ero con la mia fidanzata, a Saint-Vincent. Ho dormito da lei». «E lunedì?». «Sempre sera?». Rocco annuì. «Ero dai miei a Cogne. Abitano lì». «Lei qui non ci dorme mai?». «Ogni tanto. Ieri notte per esempio sì». Aveva fretta di andarsene, invece Rocco aveva tutto il tempo del mondo a disposizione. «Lei sa che lavoro faceva Sonya? O meglio, Juana?». «Credo di sì. Riceveva». «C'è mai andato lei?». «Io? Mai! A me quelli non piacciono». «Quelli chi?». «I cosi, lì... i trans». «Vi siete mai parlati?». «Mah... ci salutavamo quelle rare volte che ci incontravamo. Ma perché mi sta facendo queste domande? Io devo andare». «Perché l'hanno strangolata». Il lemure impallidì. «Oh Madonna...». «Già. Brutto, no? Ammazzano una sua vicina e lei non ne sa niente». Ivano alzava le spalle quasi a scusarsi. «Ma io neanche... le giuro che...». Rocco lo fermò con il gesto della mano. «Lasci i suoi recapiti al mio agente, anche il telefono della sua fidanzata e dei suoi genitori. Sa com'è? Lei dice che non era in casa, ma noi dobbiamo controllare...». «Va bene, va bene» rispose il lemure nervoso sbattendo le palpebre. «Allora» fece Italo pronto con penna e taccuino. «Aspetti, il numero dei miei non lo ricordo». «Lo guardi sul cellulare» suggerì Italo con tutta tranquillità. «Ah già, è vero, che scemo». La signora Cosma era piccola e magra. Sotto ogni soprammobile, fosse un orologio, una foto incorniciata o una statuina di porcellana, c'era un centrino di cotone. I mobili di legno lucido e i tappeti finti iraniani sposavano alla perfezione i colori pastello del tailleur e la collana di perle al collo. Solo i capelli stonavano grazie a una nuance azzurrognola da fare invidia a un vecchio fan dei Sex Pistols. Si erano accomodati in salotto e la padrona di casa si era premurata di offrire un tè al vicequestore che aveva gentilmente rifiutato. Sul tavolino basso davanti al divano c'era la settimana enigmistica, una matita e una gomma per cancellare. «Conosceva Juana, quella dell'interno 3?». Il viso liscio nonostante i 70 anni suonati si trasformò in una smorfia rugosa. «Per carità. Sa quante volte ho litigato con il proprietario? Quello, o qualunque cosa fosse, riceveva clienti nell'appartamento. Creda, il viavai non è cosa piacevole». Su un settimino accanto alle poltrone troneggiava la foto di un uomo in divisa. «Lei abita sola?». «Sono vedova da sei anni. E quello deve ringraziare che mio marito se n'è andato, altrimenti l'avrebbe cacciata a calci nel didietro!». «Se le chiedo di domenica sera, lei ricorda qualcosa?». Giulia Cosma si morse le labbra e guardò in terra. «Domenica... sono tornata dal bridge verso le sette. Poi ho cenato, un po' di televisione e sono andata letto». «Non ha visto o sentito niente di strano?». «Cosa avrei dovuto sentire?». «Juana è stata strangolata» disse Rocco semplice e diretto. Osservò la reazione. Per un attimo negli occhi dell'anziana signora passò una luce ferina. «Mi dispiace» disse toccandosi la collana. «Lei è convinto l'abbiano uccisa in casa?». «Sì, l'hanno uccisa in casa». «E ci sono evidenze, tracce, qualcosa che possa far supporre...». «Signora, le domande le faccio io». «Mi scusi» sorrise con la bocca, mentre gli occhi restarono freddi e distanti. «Mio marito era nell'arma, un po' di deformazione professionale dopo tanti anni di matrimonio ce l'ho anche io». «Nei giorni seguenti lei ha visto qualcosa di curioso? Ha sentito rumori strani?». «Mi faccia un esempio». «Viavai di gente, per esempio, qualcuno che entrava in quell'appartamento?». «E come potrei? No, niente... anzi, ora che ci penso, pace. Sì, pace. Evidentemente fra i clienti s'è sparsa la voce». «Sta facendo dell'ironia?». «Davanti alla morte non mi permetterei mai». «Lei possiede un'auto?». «Sì». «La tiene nel parcheggio sul retro?». «Certo. È un parcheggio condominiale. Ma è sempre vuoto. In questa palazzina solo cinque appartamenti sono affittati e appartengono a un solo proprietario. Ci siamo io, quello al primo piano, poi c'è Ivano, un caro ragazzo, e qui al secondo piano solo altri due condomini. Il signor Fabiani che a quest'ora sarà in ufficio, e accanto c'è il signor Bernardo Valenti, ma è sempre fuori, va a trovare il figlio a Torino». «Ora per esempio Valenti c'è?». «Non lo so». «Sa dov'è l'ufficio del signor Fabiani?». «In centro. A via Aubert. È un commercialista». «Juana aveva l'auto?». «Sì. Una Toyota nuova di zecca. Segno che quel mestieraccio rendeva e come». «Se c'è offerta è perché la domanda è alta». Giulia Cosma guardò Rocco senza capire. «Vabbè, la ringrazio, lei è stata preziosa». «Nei secoli fedele». «Eh già» si disse Rocco. «Obbediamo tutti a una legge morale, chi più chi meno. Lo sa qual è la cosa preoccupante, signora Cosma?». «No». «L'interpretazione di quella legge. Tanti saluti». Qualcosa raspava sulla porta. Riprovò a suonare e il cane si mise ad abbaiare. Un abbaio acuto, sottile come una lama. Poi qualcuno dall'altra parte fece scivolare il paletto. Finalmente aprì. «Dica?». Viso magro, borse sotto gli occhi liquidi che parevano due palle da tennis. La barba incolta e pochi capelli ormai bianchi. Il naso importante che stonava con la magrezza del viso era ricoperto di capillari rotti. Sulle guance scavate chiazze di couperose. Portava un maglioncino di lana pieno di palline. Tra i piedi infilati in un paio di pantofole di lana spuntava il muso di un volpino. «Bernardo Valenti?» chiese Rocco. «Sono io...». «Vicequestore Schiavone». Il viso dell'uomo parve rasserenarsi. «Prego, prego, entri!» e spalancò la porta. «Non si preoccupi, il cane è buono. Ma che sta succedendo?» il volpino s'era messo ad annusare i pantaloni del vicequestore. «Giù, Anubis, giù! Lo scusi...». «Ho un cane anche io». «Ah, ecco perché. Ti piace il cane del signore, eh Anubis? Prego!». Tutta la casa era rivestita di legno di abete chiaro, più che un appartamento pareva un rifugio in quota. «Mi scusi il disordine, ma che cosa accade? Ho sentito un trambusto, mi sono un po' spaventato, per caso vuole un caffè?». «No, signor Valenti, sono a posto così». Il padrone di casa fece cenno a Rocco di accomodarsi su una poltrona di velluto. Il vicequestore sedendosi diede un'occhiata rapida al salone e gli venne da sorridere. Il signor Valenti non ci teneva molto all'arredamento. Come lui non aveva un quadro, un libro, e i mobili anonimi erano sicuramente compresi nel prezzo dell'affitto. Ma Rocco era sicuro che le scansie della cucina e l'armadio di legno fossero pieni di liquori. «Allora signor Valenti, lei conosceva Juana Perez, la donna al primo piano?». «Perché dice conosceva?». La mano destra di Bernardo tremava. Ai lati della bocca s'era formata una leggera bava bianca. «L'abbiamo trovata morta nella Dora». Bernardo si passò la mano tra i capelli. «No! Non è vero». «Invece...» e Rocco allargò le braccia. «Prima o poi doveva succedere... sicuro che non vuole un caffè?». «No, grazie». «Allora un tè? Una tisana? Vuole una birra?». «No grazie, sto a posto così!». «Se permette vado a prendermi un bicchiere d'acqua» e lasciò a passo rapido il salone. Il volpino s'era steso sul suo cuscino liso e osservava Rocco muovendo appena la coda. Passarono dieci secondi e Valenti tornò leccandosi le labbra. «Ecco, ecco, mi scusi, una sete... come è morta? Si è suicidata?». «L'hanno uccisa». Bernardo sbatté le mani, se le strinse al petto e cominciò a scuoterle quasi fosse in preghiera. «Povera Juana, povera Juana». «Perché ha detto che prima o poi doveva succedere?». «Sapesse quante volte ho provato a parlarle, a farle cambiare strada. Lo sa che vita schifosa faceva? Glielo dicevo sempre: Juana, fai un mestiere pericolosissimo! E che si sa?». Rocco accavallò le gambe. «Lei ha sentito niente domenica sera o nelle sere successive?». «Cosa avrei dovuto sentire?». Valenti si sedette sulla poltrona davanti al vicequestore. «Movimenti». Il padrone di casa ci pensò su tastandosi il polso. «No, nessun movimento. Niente. Ma l'hanno uccisa nel suo appartamento?». «Io dico di sì». «È una cosa orribile». «Già. La casa è stata svuotata. Lei è mai entrato in quell'appartamento?». «No. Mai». Il vicequestore si alzò. Il volpino ringhiò. «La saluto. Se ho bisogno di lei la trovo qui?». «Dove altro potrei andare?». «Posso chiederle che mestiere fa?». «Facevo il ragioniere. Ora sono in pensione». «Quant'è che vive ad Aosta?». L'uomo alzò gli occhi verso il soffitto. «Da un po'». «Complimenti!» e si mosse verso la porta di casa. «Per cosa?». «Perché è riuscito a togliersi l'accento. Spesso ci penso. Vede, a Roma, nella mia città, uno può essere di Napoli o di Torino, passano gli anni e l'accento non lo perde. Sa perché? Perché ai romani non interessa da dove vieni. Ti accolgono comunque. Ho notato invece che in molte città del nord anche dopo pochi anni, si tende a prendere l'accento del posto. Chissà, forse per nascondersi?». «Nascondersi?» chiese Valenti. «Mimetizzarsi. Insomma, ci si sforza per essere accettati. Lei dov'è nato?». Il padrone di casa sorrise. «Orecchio fine il suo. Allora le dico che quanno pullecenella pigliaje 'a carrozza, 'o vedettero tutte...». «Napoli?». «Avellino. Per servirla». Rocco era al centro del parcheggio condominiale, ma non c'erano tracce di una Toyota nuova di zecca. Si avvicinò alla recinzione dell'appartamento di Juana, mise un piede sul muretto, la scavalcò per ritrovarsi nel piccolo giardino coi vasi sbreccati e il rubinetto gocciolante. Entrò in casa. Osservò la scena desolata e le tre ombre chiare che foto o quadri avevano lasciato sulla parete. Poi guardò in alto il gancio del lampadario tinto col colore del soffitto. In quel momento entrarono due agenti della scientifica. «Salve» dissero. «Ci rivediamo!». «Eh già». Rocco trascinò l'unica sedia della stanza sotto il gancio, montò sulla seduta e si mise ad osservare l'uncino conficcato nella parete. Nella parte concava del ferro non c'era traccia di vernice. «Cosa controlla, dottore?» disse un agente incappucciato. «Niente». Rocco scese dalla sedia. «Mi chiedevo a cosa servisse un gancio al soffitto se non c'è filo elettrico né una presa per mettere un lampadario». «Mia nonna nella casa di campagna ci appendeva i salami» fece l'altro della scientifica. «Però in cantina». «E vorrà dire che qui fa così freddo che 'sta stanza fungeva anche da frigo» gli rispose il collega. Dalla porta apparve Michela Gambino. Indossava il solito maglione slabbrato dal quale spuntava il collo lungo e magro, ma aveva dato un tocco di vezzosità all'abbigliamento con un paio di anfibi a fiorellini. «Eh no, se mi entrate così mi inquinate la scena!» gridò al vicequestore. «Non lascio tracce io» e mostrò i guanti che aveva appena infilato. «E a proposito di tracce, qui hanno svuotato tutto l'appartamento». «Non c'è niente?». «Niente. Solo 'sti du' mobili». Il sostituto si fregò le mani. «Bello! Bene! Puzza di complotto lontano un chilometro». Rocco la osservò. «Sei stancante, lo sai? Cosa c'è di bello?». «Vivi nell'ignoranza più assoluta. Che fa, ha traslocato dopo morta? Rifletti su cosa è successo negli ultimi tre mesi...». «Meglio di no». «Muore la Thatcher e un mese dopo Andreotti. Non ti dice niente?». «Erano vecchi?». «A parte questo?». «Senti Michela, te lo dico col cuore in mano. Non me ne frega un cazzo che sono morti perché una farfalla in Cina ha sbattuto le ali o un Mahori s'è schiantato con la piroga. Aspetto notizie» la scansò avviandosi verso la porta di casa. «Una sola parola, Schiavone: Datagate! Pensaci». Il vicequestore si batté la fronte con la mano. «Segnato!». «Sappi che in questura ci controllano» gridò. A Rocco venne da sorridere ricordando il chewing-gum che aveva appiccicato alla porta dell'ufficio della Gambino. «Dici?». La Gambino si sfilò il maglione di cotone restando in maglietta e se lo legò ai fianchi. Rocco ci aveva visto giusto. Aveva un corpo magro e scattante. «Ci controllano, ti dico. Ne ho le prove!» fece quella orgogliosa aggiustandosi la chioma. Il vicequestore si voltò per nascondere il sorriso e si allontanò. «A volte basta un capello!». Quando Italo e Rocco bussarono alla porta dello studio Tax Force in via Aubert mezza mattinata se n'era andata. Venne ad aprire una ragazza dall'aspetto dimesso con i capelli lunghi e neri e un paio di occhiali a fondo di bottiglia. «Sì?». «Vicequestore Schiavone. Il dottor Fabiani per favore...». La segretaria li fece entrare senza un sorriso. Nella sala d'attesa c'erano due poltroncine di pelle con lo scheletro in acciaio. Brutti quadri ad olio su delle pareti azzurre contribuivano a dare un tono di squallore all'ambiente. «Il signor Fabiani è al telefono». «E chissenefrega» disse Rocco. «Lo chiami, mica abbiamo tutta la mattina». La segretaria sparì veloce dietro una porta di legno lucido. Rocco andò ad affacciarsi alla finestra. Si mise a tamburellare sul vetro. Un cane senza collare trotterellava in strada. «Buongiorno, come posso esservi utile? Sono Diego Fabiani» l'uomo si fece avanti stringendo la mano di Italo. Rocco restò alla finestra. «Buongiorno a lei, vicequestore Schiavone». Il commercialista era sulla quarantina, portava il pizzetto e una camicia con il colletto talmente alto da sembrare un collarino per la cervicale. Ai polsini splendevano due gemelli. «Domenica sera hanno ucciso Juana Perez, una sua condomina, quella del primo piano. La conosceva?». Diego ingoiò un grumo di saliva. «Uccisa...?». «È quello che ho detto. La conosceva?». «Io? Sì, insomma, l'ho incontrata un paio di volte. Credo fosse da poco che stava da noi». «Lei domenica sera era a casa?». «Domenica? No. Ero qui, in studio». «Di domenica?» fece Italo. «Eh sì. Periodo terribile. Il 16 c'è la prima scadenza delle imposte, questo è un momento molto duro per noi». «Nei giorni successivi alla domenica, ha notato qualcosa?». «Tipo?». «Viavai, rumori...». «Nel palazzo?». «Bravo». «No. Niente di che. Perché me lo chiede?». Rocco guardò Italo. «Perché glielo chiedo, agente Pierron?». «Perché c'è stato una specie di trasloco e nessuno pare se ne sia accorto» si premurò di dire Italo. «Eh già. Evidentemente hanno i muri molto spessi. Il palazzo è antico?». Diego Fabiani alzò le spalle senza sapere cosa dire. «Non lo sa. Pierron, forse c'è bisogno che tutti i condomini di via Brean 12 vadano a fare una visitina da Maico». Pierron sorrise. Diego continuava a non capire. «Per l'udito, dottor Fabiani» disse Italo per chiarirgli le idee. «Vede dottor Fabiani, è una cosa strana. C'è stato un omicidio, molto probabilmente nella casa di Juana Perez, poi quella casa l'hanno svuotata ma né lei, né la signora Cosma, Bernardo Valenti o quel tale, il lemure...». «Il lemure?» chiese Italo. «Sì quello che si scorda il cellulare a casa». «Ah sì, il figlio dell'avvocato... Ivano Petrulli» disse Italo. «Petrulli, appunto, insomma nessuno ha sentito niente. Voi non volete dare una mano alle forze dell'ordine». «Le giuro, commissario...». «Vicequestore». «Sì, scusi, vicequestore. Io non mi sono accorto di niente. Esco la mattina presto, passo tutto il giorno qui allo studio, la sera sono talmente stanco che mi addormento come un ghiro. Mi dispiace...». Sembrava sinceramente seccato di non poter essere di aiuto. «Però mi conforti nel dire che se ricorderà qualcosa, anche una sciocchezza, me la viene a riferire. Dica lo giuro!». «Lo giuro» rispose Diego esitante. «Ottimo. Pierron? Lasciamo il dottor Fabiani alle imposte e torniamo a cercare nella pattumiera qualche brandello di informazione. Tanti saluti, Fabiani». D'Intino e Deruta lo aspettavano sull'attenti davanti alla stanza. Sembrava montassero la guardia. Rocco tentò di svicolare, ma l'agente abruzzese lo aveva visto. «Dottore!» urlò. Schiavone alzò gli occhi al cielo e attese. «Dotto', tutto fatto come ci aveva ordinato. Aveme visto tredici trams». Rocco non ricordava minimamente quale fosse il compito che aveva affidato ai due. «Bravi. E che avete scoperto?». «Che nessuno conosceva il cadavere» rispose Deruta con l'aria triste. In quel momento Rocco era a un bivio. O ammettere che durante l'assenza dei due le indagini erano andate avanti oppure trovare rapidamente qualcos'altro da fargli fare. Ma non gli veniva in mente niente. «Ottimo lavoro. Io ho fatto dei passi avanti e ho scoperto che si chiamava Juana Perez, era argentina. Quindi adesso qual è il vostro compito?». «Boh» risposero all'unisono. «Telefonate alla questura centrale, all'ambasciata, al consolato e cercate di capire se aveva il permesso di soggiorno, da quanto tempo stava in Italia, insomma voglio tutte le informazioni possibili». I due agenti lo guardarono spaventati. «Oh! E allora? Ragazzi, qui siamo di fronte a una cosa grave. Chi ci dice che Juana Perez non fosse una trafficante di droga?». «Maro'» fece D'Intino. «Li 'sti di' pe' da vero?». «Non mi dare del tu, D'Intino!». «Scusi. Ma che li shtite a' di' pe' da vero?». «E che sto a scherza'? Mi raccomando acqua in bocca. Sapete che vuol dire acqua in bocca?». «Sì. Che non dobbiamo parlare con nessuno». «Bravo Deruta! Profilo basso e al lavoro. Siete preziosi». Fecero dietro-front all'unisono e perplessi si allontanarono. Rocco proseguì verso la stanza degli agenti. Trovò Scipioni solo davanti al computer. «Anto', ce l'hai ancora il tuo amico che smanetta coi cellulari?». «Sì, perché?». Prese dalla tasca un foglietto. «Segnati questo numero, è quello di Juana Perez. Vedi di tirare giù le telefonate che ha ricevuto». Antonio sbuffò, poi si guardò intorno. «La magistratura lo sa?». «Ancora no». «Rocco, il mio contatto dice che questa è una cosa illegale» e abbassò la voce «... e che bisognerebbe mettere in mezzo la magistratura per controllare...». Rocco lo interruppe afferrandogli una spalla. «Anto', lo so. Ma mi serve che lui ci faccia un favore. Non è legale, è pericoloso ma è importante». Antonio annuì. «Se si gioca la carriera gli daremo una mano». «Mi spieghi una cosa prima?». «Sì». «Perché di nascosto?». «Ho i miei motivi. La magistratura è allertata e seguirà il suo iter. Ma a me serve adesso. Comprì?». Antonio poco convinto si alzò. «Vedo che riesco a fare. Ma fuori di qui però...». Lupa guardava minacciosamente il tappeto del magistrato. Aveva un conto aperto con quel tessuto di lana e seta, sbavava dal desiderio di aggredirlo, ma sapeva anche che il suo padrone su questo era stato categorico: «No!». E a quel suono secco e perentorio Lupa obbediva sempre. Dopo pappa e Lupa, No era la terza parola che aveva imparato. La quarta era amore, che significava dover scodinzolare e leccare il padrone possibilmente sul viso per ricordargli che era un cucciolo indifeso. Baldi se ne stava sorridente seduto sulla sua poltrona a giocherellare con una penna. La fotografia della moglie era riapparsa sulla scrivania. «Sicuro che il cane non mi attacca il tappeto?». «Tranquillo, dottor Baldi. Se le dico di no Lupa si ferma. Ha imparato tre parole: Lupa, pappa e no» omise la quarta, la più importante. «Be', conosco un sacco di persone che non ne hanno imparata neanche una. Allora che idea s'è fatto?». «L'idea è che è stata uccisa in casa. Era nuda e prima di buttarla al fiume l'hanno rivestita. L'hanno gettata lì per depistare, o quantomeno confondere le acque, mi passi il termine. Ora quello che resta piuttosto oscuro è perché abbiano svuotato l'abitazione di Juana Perez». Il magistrato si alzò. «Ottimo. Ha tratto delle conclusioni?». «Qualcosa, ma è ancora presto». Andò alla finestra. «E non vuole condividerla con me?». «Certo. Io con lei condivido tutto». «Bum!». Rocco cercò di calmare l'accelerazione improvvisa del battito cardiaco respirando, poi parlò. «È evidente che togliendo di mezzo vestiti e oggetti si cerca di nascondere qualcosa. Non l'identità, magari, ma insabbiare sì. Gli oggetti parlano, lo sappiamo, e parlano anche pavimenti e pareti se uno sa leggerli. Così ci lasciano un deserto da decodificare. Questo mi fa pensare che Juana era legata a qualcuno che deve restare nell'ombra. Fin qui siamo nell'ovvio. Quello che non è chiaro è chi sia questa persona». Baldi si voltò di scatto: «Lei pensa a un pezzo grosso?». «Forse. Oppure semplicemente è un assassino molto meticoloso che cancellando ogni traccia cerca di ripulire il suo passaggio. Ma, e questa è solo una supposizione, anche se Juana abitava in quella casa da poco, è probabile che roba da portare via ne avesse accumulata. Senza un aiuto? Questo non mi convince. Nella palazzina nessuno ha sentito niente. Per traslocare oggetti e vestiti ci sarà stato un minimo di viavai. Diciamo che è avvenuto dalla porta sul retro e che l'assassino è uno molto organizzato. Nel bel mezzo della notte, si mette al lavoro. Motore spento eccetera eccetera. Se lo vede? Io no. In una città dove tutti si fanno i cazzi degli altri, contrariamente a quanto credevo mesi fa, può passare inosservata una cosa simile? La strada, via Brean, è piena di palazzine. E questo si mette a fare un mezzo trasloco e nessuno si accorge di nulla?». «Già. E se fa sparire tutto» intervenne Baldi, «significa che fra gli oggetti di Juana magari c'era qualcosa di compromettente. Un regalo? Un indumento? Una lettera? Un messaggio?». «Se era un cliente abituale o aveva una relazione con lei è una cosa che non scarterei. Lupa no!» aveva bloccato il cane con le fauci spalancate verso la frangia del tappeto. «Ora però dobbiamo controllare il cellulare di Juana Perez». La punta della penna si spezzò fra le mani di Baldi che divennero nere d'inchiostro. «Ma porca... ma guarda te...». Rocco gli passò un fazzoletto di carta ma quello preferì asciugarsi con un paio di fogli che aveva sulla scrivania. «Una mia relazione» disse a Rocco. «Per quello che serve. Ha il numero della vittima?». Tirò fuori il foglietto con il numero e lo dettò al magistrato. «Va bene Schiavone, ora mi do da fare. Fra poco avrò l'autorizzazione». «Grazie». Schiavone si alzò. «Resta il mistero della sparizione degli oggetti della vittima. È arrivato a una conclusione?». «Ancora nulla. Vado a parlare con il proprietario dell'immobile, le saprò dire di più». «Non è vero. Lei comincia già a farsi un'idea, ma non la vuole condividere, come al solito». «Meglio di no. Potrei fare la fine della Gambino». «La Gambino? Che c'entra il sostituto della scientifica?». «C'entra, c'entra... annamo Lupa!». Carlo Berthod abitava a Chenoz ma aveva l'attività a Courmayeur. Caterina lasciò l'auto in piena zona pedonale obbedendo agli ordini di Rocco. Lupa rimase dentro accucciata sui sedili posteriori. Appena scesi dall'auto un vigile si lanciò sui trasgressori. «Scusate, qui non si può...». «Facciamo presto, collega» disse Rocco mostrando il tesserino. «A proposito, io darei un'occhiata a quell'Alfa blu che sta entrando in piazza». Il vigile e Caterina si voltarono all'unisono. «Perché?» chiese l'agente della municipale. «Così, hai visto mai...» e si incamminò seguito dal viceispettore verso il negozio. Il vigile scuotendo la testa si allontanò. «Perché gli hai detto di controllare quella macchina?». «Secondo me ci stava seguendo» disse Rocco tranquillo. «Veramente?». Caterina sbarrò gli occhi. «Può essere» e finalmente entrarono nel negozio di articoli sportivi Mont Blanc. Il prezzo di una giacca a vento rosa in vetrina si mangiava metà stipendio di un operaio. Il negozio era tutto rivestito di legno, anche lì c'era profumo di cannella. Subito si fece avanti una ragazza. «Buongiorno, posso essere utile?» chiese col miglior sorriso della casa. «Cerchiamo Carlo Berthod, è il proprietario, no?». «Posso chiedere chi...?». «Vicequestore Schiavone, questura di Aosta». Due pomelli rossi si accesero sulle guance della donna. «Sì subito» balbettò, poi si voltò e accelerando il passo sparì dietro un séparé di stoffa e legno scuro. «Perché ogni volta che diciamo che siamo della polizia la gente ha questa reazione?». «Siamo gente scomoda, Caterina». Il proprietario seguito dalla commessa arrivò con passo deciso. Energico, si tolse gli occhiali da vista e allungò la mano per stringere quella di Rocco. Aveva un bel maglione rosso acceso e pantaloni blu. «Salve, Berthod». «Schiavone. Lei è il viceispettore Rispoli. Come va?». «Bene, benissimo, qual buon vento?». «Buono non lo so. Veniamo per una sua affittuaria. Si chiama Juana». Tutta la sicumera sparì come una nuvola sotto la tramontana. Si voltò verso la ragazza e la liquidò con uno sguardo. «Juana... Juana...». «Il trans che ha affittato l'interno 3 di via Brean, palazzina di sua proprietà». «Ah, capito. Perez, mi pare si chiamasse. Sì. Argentino, no? Che ha combinato?». «Perché usa il passato?». «Cosa?». «Perché dice: mi pare si chiamasse? Dovrebbe dire: mi pare si chiami». «Vabbè, non sottilizziamo sull'italiano». «A volte serve». «Comunque Juana Perez. Che gli è successo?». «S'è fatta ammazzare». Berthod si portò le mani alla bocca. «Eh, brutte storie. Lo sapevo». «Che sapeva?». «Quella gente prima o poi finisce così... prostitute, omosessuali, sempre storie di omicidi...». «Abbiamo un criminologo qui, Caterina. È un criminologo lei?». «No. Sono un uomo che a 62 anni un po' la vita la conosce». «Allora se conosce la vita, mi mostra qualche documento dell'affitto?». «Eh?». «Dico, ha il contratto con le generalità?». «Mica è della finanza lei» si oppose deciso l'uomo. «Senta Berthod, a me che affitta in nero non me ne sbatte un cazzo. Voglio solo sapere quando l'ha affittata e le generalità complete dell'affittuaria. Può farlo o devo tornare con un mandato del magistrato?». «E certo, abbia pazienza, io...». «Non ce l'ho». «Cosa dottore?». «La pazienza io non ce l'ho». «Vado a prendere un quaderno dove segno tutto... aspetti» e a gambe strette corse verso la cassa. Entrarono due clienti, un uomo e una donna griffati dalla testa ai piedi, e si misero ad osservare la merce. La commessa li raggiunse. Rocco diede un colpetto di gomito al viceispettore. «Li vedi quei due con i loghi stampati sulle magliette?». «Certo». «Me la spieghi una cosa? Perché uno porta in giro la marca di quello che indossa e lo mostra gratis? Voglio dire, se io mi devo mettere le iniziali del sarto scritto con gli Swarovski pretenderei dei soldi. O no?». «È per far vedere che te li puoi permettere, credo». Con il suo passo deciso, Berthod tornò da Rocco e Caterina. «Ecco qui... allora allora allora» si leccava le dita e sfogliava il quaderno. «Non lo faccia, per favore» quasi gridò Rocco. «Cosa?». «Non si lecchi le dita per girare i fogli. È una cosa che mi fa vomitare». «Mi scusi. Allora... ecco qui. Juan Pedro Perez, detta Juana» e strizzò l'occhio a Rocco. «Che è?» chiese il poliziotto. «Juan Pedro... era il suo nome da uomo». «Ci arrivo, signor Berthod». «E poi l'ha cambiato in Juana. Il nome. Il cognome invece l'ha lasciato» e ridacchiò. Rocco guardò Caterina. Il viceispettore cominciò a temere per la salute del negoziante. «Berthod, amico mio caro, le devo ricordare che è morta una donna per farle passare 'st'aria divertita? Allora mi dica». Gli occhi dell'uomo tornarono sul quaderno: «Sì, bene. Allora Juana o Juan Pedro era nato a Rosario, Argentina, il dieci maggio del 1982. L'anno che abbiamo vinto i mondiali!». Rocco sospirò. La voglia di prenderlo a schiaffi era diventata una necessità. Berthod proseguì: «Io gli ho affittato casa ad aprile». «Due mesi fa... bene. È sempre stata puntuale nei pagamenti, anche se in nero?». «Sempre...». «Lei è mai andato in casa da quando Juana l'aveva presa?». «Per carità!» e alzò le mani schifato. «Eh già. Caterina, ricorda: pecunia non olet! Andiamo avanti, quindi lei non è a conoscenza di chi frequentava, se avesse amici, insomma della sua vita non sa niente». «Dottore, credo battesse. Ci andavano i clienti». «Ne era consapevole?». «Se glielo sto dicendo?». «Capisce che lei a cuor leggero ha commesso una fracca di reati?». «Io?». «E no, mi' nonno! Potremmo aprire una bella inchiesta, che ne dice viceispettore!». «Accidenti se potremmo». «E chissà, magari troviamo nell'appartamento un po' di droga e aggiungiamo pure spaccio». «Io affitto casa, quello che ci fa l'inquilino sono fatti suoi!». «Lei affitta casa in nero a un'inquilina per farne un bordello e non a sua insaputa. Prendeva la stecca? La percentuale?». «Come si permette?». «Giù la cresta Berthod, lei non è in grado di minacciare, tantomeno di innervosirsi. Lei deve solo sperare che questi due bravi poliziotti, cioè io e la signorina, chiudano un occhio. E diciamo che per far chiudere un occhio serve uno sforzo!». Carlo Berthod annuì. «Ho capito. Potevate dirlo prima e la facevamo breve. Quanto volete?». Rocco guardò il negoziante dritto negli occhi. «E mettiamoci pure corruzione a un pubblico ufficiale. Berthod, ma lei ci tiene tanto ad andare sotto processo?». «Io non ci sto capendo più niente! A voi chi vi manda?». Rocco sgranò gli occhi. «Chi ci manda? Ma lei è rincoglionito o gioca su un altro tavolo?». In quel momento si palesò la commessa. «Signor Berthod, qual è il prezzo della giacca a vento rosa?». «580 euro». La ragazza sparì e il negoziante tornò a guardare Rocco. «Insomma, mi state trattando come se fossi un assassino. Be', e allora è giunto il momento di dirvi un paio di cose». «Siamo al punto del lei non sa chi sono io» disse Rocco sottovoce a Caterina che sorrise mentre i due clienti griffati osservavano da lontano il dialogo con la faccia spaventata. «Io sono stato assessore in Regione, sono un onesto cittadino che paga le tasse». «Insomma». Ma Berthod continuò ignorando il commento del vicequestore. «Sono membro del Rotary e faccio beneficenza. La mia famiglia vive a Courmayeur da sempre e vanto amicizie molto in alto. Intesi? Molto in alto!». Caterina non riuscì a trattenersi: «Tipo sul Monte Bianco?». «Tiri qualcosa fuori dalla memoria. Qualche cliente o amico della defunta signora Juana Perez... ci pensi bene». «Io non ho nient'altro da dirle. Per ogni eventuale incontro futuro la prego di rivolgersi al mio avvocato». «Va bene, signor Berthod. A questo può rispondermi senza la presenza dell'avvocato. Lei prima di affittare la casa a Juana Perez, l'ha ritinteggiata?». «Certo! È compito del proprietario consegnare la casa pulita e in ordine». «Bravo. Mi piace la gente che rispetta le regole» esclamò Rocco. Poi mormorò qualcosa nell'orecchio di Caterina, infine allungò la mano. «Lasci il nome dell'avvocato in questura. Signor Berthod, glielo dico da uomo a uomo, lei s'è cacciato in qualcosa più grande di lei. Tanti saluti e si tenga a disposizione. Andiamo Caterina, o vuoi prima vedere una giacca a vento rosa?». «Non ricordavo che fosse così divertente andare in giro con te». «Hai la faccia appesa da stamattina, sorridere non fa mai male». Salirono in auto. «E poi a venire con me impari un sacco di cose, o no?». La strada si era riempita di turisti. Eleganti, benestanti, i vecchi tendevano ad apparire dei giovani coi capelli bianchi, le vecchie erano tirate come archi. «Per esempio?». «Bisogna far incazzare quel tipo di persone, perché è gente che ha sempre vissuto nella bambagia e basta poco e commettono un passo falso». «Berthod l'ha fatto?». «Secondo me sì. E adesso ti faccio un'altra breve lezioncina». «Sono tutta orecchi». «No, devi essere tutt'occhi. Guarda la gente che passeggia qui in centro. Lo sai da dove si vede che sono ricchi?». «Da cosa?». «Addosso non hanno niente di eclatante. Ma le scarpe sono fatte a mano e i pullover sono di cashmere. Sono i finti ricchi che si mettono i marchi degli stilisti sul petto, griffati fino alle mutande. E che comprano giacche a vento rosa». «Ho capito cosa mi stai dicendo, ma non ho capito cosa mi stai dicendo». Rocco scoppiò a ridere. «I due che sono entrati dopo di noi. Li hai notati?». «Sì. Che c'era che non andava?». «Stonavano, non erano del posto. Invece Berthod? Cos'hai capito di lui?». «Che è un pallone gonfiato?». «Non solo Cateri'. Ora voglio sapere se Antonio ha qualche novità del cellulare. Torniamo all'ovile. Non prima di fermarci alla Dolce Voglia e spararci un bignè con la crema». «Come fai a conoscerla?». «Mi informo». Aveva bisogno di una doccia. La giornata era stata lunga e pesante. Lupa aveva sbranato i croccantini e ne voleva altri. Rocco la accontentò con un pezzo di pane secco che quella ingoiò senza neanche masticare. Cominciò a spogliarsi. Aprì l'acqua. Era fredda. Attese guardandosi allo specchio il torso nudo. Masticato, ecco cosa sembrava. Giorno dopo giorno, settimana dopo settimana, anno dopo anno il tempo aveva triturato pelle e muscoli. Si mise di profilo. Non andava bene. Intorno ai fianchi si stava formando una ciambellina adiposa e pure la pancia tendeva a gonfiarsi e a superare la cintola. Ripensò al suo corpo quando aveva vent'anni. Magro e senza grasso anche se mangiava come una betoniera. Intorno agli occhi poi s'era formata una ragnatela di rughe e pure dalle tempie molti capelli se n'erano andati chissà dove. Valeva la pena star lì a guardare il proprio corpo ammucchiarsi come uno straccio, un pezzo di carne che si sostiene coi medicinali? Aveva sempre denigrato le donne che si ritoccavano pur di dilatare e stirare il tempo, come facevano con rughe e pelle avvizzita. Ora cominciava a capirle. «Tanto quello passa lo stesso, mica lo fermi». Sta lì Marina, si intravede in piedi vicino alla finestra, ma so che c'è. «Finalmente. Allora per vederti c'è bisogno di uno specchio...». «Lo specchio non mente mai. Dovresti ricordartelo, lo raccontavano già nelle favole». «Perché non vieni più?». «Perché il vento cambia, Rocco. Io lo so. Anche tu lo sai». «Guarda come mi sono ridotto Mari'... guarda 'sta pancia, pare 'na borraccia, eh? Meno male che non mi hai visto, sennò capace che te ne andavi!». Ride Marina. «Mia nonna diceva che il tempo mozzica. E noi siamo meno di un abbaio di un cane. Anche se è strana 'sta storia del tempo, no? Lo sai come la vedo? È come stare su una scala mobile ma nessuno sa quanto è lunga, poi quella finisce e pure tu hai chiuso». Ma lei non c'è più. È stato un attimo, un'ombra appena, un'onda del mare. «La tua è stata troppo breve, amore mio». Non c'è più. E l'acqua è ancora fredda. Segno evidente che la caldaia era rotta. Masticando una bestemmia fra i denti andò in cucina. Uscì a petto nudo e sentì subito il fresco della sera dargli una pacca poco amichevole sulla cervicale. Il display era rosso e lampeggiava una lettera che pareva una S. «Che cazzo vuol dire?». Che la caldaia era in blocco. Premette il tasto reset. Non succedeva nulla. «Ma porca puttana!» mollò un cazzotto allo scafandro di lamiera. Contemporaneamente suonarono alla porta. «Chi è che scassa il cazzo?» rientrò in casa. Andò ad aprire. C'era Gabriele con la faccia scura. «Che vuoi?». «È successa una cosa» e gli allungò un foglietto. «Che è 'sta roba?». «La legga per favore». «Le sai aggiustare le caldaie?». «La mia si blocca sempre. Se vuole le do un'occhiata». «Fuori al balcone della cucina» disse Rocco e cominciò a leggere il foglio mentre Gabriele scattò. Era una nota della scuola. «Famme capi' Gabriele, che hai combinato?». «È un richiamo del dirigente» urlò quello dal balconcino. «Domani mi devo presentare con un genitore o chi ne fa le veci!». «Ripeto: che hai combinato?». «Ecco, ora funziona!» era rientrato in casa con il sorriso stampato sulla faccia. «Si deve premere reset per almeno cinque secondi». «Che hai combinato? E tre!». «Ho fatto come mi ha detto. Sono andato da Diego. Prima gli ho tirato un calcio alle palle e poi un martello di Odino sul collo e quello l'hanno portato all'ospedale!». «Ma porca... ma l'hai fatto dentro la scuola?». «E sì, in aula». «Gabriele sei un coglione! 'Ste cose si fanno fuori dalle mura scolastiche, per strada». «Questo lei non me l'ha detto!». Rocco accartocciò la nota. «Ma che sei un robot che obbedisci a occhi chiusi? Dovevi capirlo da solo! Sei dotato di un cervello o in testa hai solo un pacco di merendine?». Gabriele abbassò il capo. I capelli gli coprirono il volto. «Non ci ho pensato». «E sei un cretino!» e se ne andò in bagno. Gabriele lo seguì. Rocco si sedette sulla tazza e cominciò a spogliarsi. «Mo' mi devo fare la doccia, per favore stai fuori!». Gabriele sparì dietro lo stipite. «Si può sapere come ti è venuto in mente?». «Dovevo agire, no? Me l'ha ordinato lei. E allora mi sono detto: la scuola sta per finire, o lo faccio adesso oppure sono punto e accapo. Perché anche Diego lo bocciano e me lo ritrovo sicuramente l'anno prossimo, magari in classe». Rocco entrò nel getto caldo e cominciò a lavarsi. «Ma dimmi tu». Gabriele era tornato nel bagno. Poteva vedere la sua figura massiccia attraverso la plastica smerigliata del box. «Ma non vuole sapere cosa è successo dopo la rissa?» lo sentì dire. «Che t'hanno mandato dal preside». «Si chiama dirigente scolastico» lo corresse Gabriele. «Vabbè, come se chiama se chiama... allora?». «No, prima del dirigente. Insomma è andata proprio come aveva previsto lei, i due amichetti di Diego sono scappati. Uno mi ha mandato a dire che eravamo amici e che lui era sempre stato con me, non con Diego». «Bravo». «E Serenella mi ha spedito un sms con la foto della sua bocca che mi manda un bacio». «Chi è Serenella?». «Una della prima B. È bionda e le piacciono i Metallica». «Annamo bene». Chiuse l'acqua. «Passame un po' l'accappatoio?» e tirò fuori una mano. Se lo infilò e grondante uscì dalla doccia. Gabriele si era seduto al lavandino. «Scendi sennò me lo sbrachi» e quello obbedì. «Insomma ti sei fatto valere. Vabbè, hai sbagliato il luogo e i tempi, ma bravo. Sono orgoglioso di te. Mo' dimmi perché mi porti la nota». «Domani devo andare con un genitore dal dirigente». «Ti sei accorto che hai una madre e da qualche parte pure un padre?». «No». «No che?». «Mamma sta ancora a Milano. E papà è all'estero». «E tu di' a mammina di chiamare il preside e che appena torna ad Aosta andrà a parlarci». Gabriele non rispose. Osservava Rocco che si sfregava i capelli con un asciugamano. «Embè?». «Mamma non lo deve sapere». «Cosa?». «Che ho fatto a botte. Sennò mi manda dai preti, e io dai preti non ci voglio andare». Rocco lo guardò negli occhi. «Fammi capire, ci devo venire io?». «Lei può inventarsi che è l'uomo di mia madre e che ha la patria potestà». «Ma che stai a di'? La patria potestà. Non sai manco che significa!». Gabriele prese fiato. «La potestà in diritto è la situazione giuridica soggettiva che consiste nel dare un potere a una persona per tutelare un interesse altrui. L'ho letto sulla Treccani on line». «Oltre a impararlo a memoria, sai cosa vuol dire?». «No». «Vuol dire che i tuoi genitori sono separati ma vivi e vegeti, stanno bene, non sono in galera e hanno loro l'affidamento condiviso o congiunto finché sei un minorenne. Io non sono tuo padre, io non sono tua madre, io sono il vicino e non mi scassare il cazzo con queste storie» e uscì dal bagno. Gabriele lo seguì. «Quindi non viene?». «No». «Per favore». «Ho detto no». «Mi buttano fuori». «Peggio di così non può andare. L'anno te lo sei giocato, no?». «Può andare peggio, invece. Se mi buttano fuori mi spediscono dai preti». «E sai che ti dico? Che mica ti farebbe male il liceo dai preti». Gabriele si gettò sul divano e cominciò a carezzare Lupa. «Io non credo in Dio». «E allora?». «Se uno mi chiede della trinità o del mistero dell'immacolata concezione, che gli rispondo? E se pure gli rispondo, mi buttano fuori di sicuro e mamma mi manda a fare il benzinaio o il ragazzo spazzola dal barbiere. Guardi che il mio dirigente è una persona buona». «Stai diventando una rottura di coglioni. Anzi, domani devo aggiungere all'ottavo livello delle rotture di palle la caldaia rotta quando uno è già mezzo nudo e pronto a farsi la doccia. Se poi succede d'inverno sale di diritto al nono». «Addirittura al nono?». «Non parlavo con te». «Facciamo così... lei domani viene dal dirigente e io per un giorno faccio lo schiavo». «Tu sei deficiente, secondo me». «Lo so. Lo dice anche mamma». «Ah, allora ne è cosciente». «Per favore, non mi mandi dai preti. Mi tagliano i capelli, mi devo mettere la cravatta e le scarpe di cuoio nere. Non voglio! Per favore, per favore, per favore!». «E sei una piattola! Vabbè. Vengo io. Ma ti costa caro». «Sono pronto» fece Gabriele. «Cosa devo fare?» chiese scattando in piedi. «Ancora non lo so. Ci devo pensare, ma sarà qualcosa che non ti piacerà sicuramente. Ora siccome ho fame e a casa non c'è una mazza io e te ci andiamo a fare una pasta con le vongole al Grottino. Va bene?». «Ci sto. Pago io?». «Mi pare il minimo». Uccio Pichi era a pezzi. Aveva gli occhi rossi e il dolore dietro le scapole era diventato insostenibile. Prima o poi doveva decidersi a tornare dal suo amico fisioterapista. Sapeva che lo schiacciamento delle vertebre non si risolve con due massaggi, ma almeno per un paio di giorni alleviava il dolore. Però quello abitava a Roma sud, lui a Roma nord e raggiungerlo era un'impresa. Aveva saltato il pranzo e anche la cena sembrava un miraggio. Il cadavere con il taglio sotto la gola giaceva sul lettino autoptico. Cominciò a svestirlo mentre il suo assistente seduto alla piccola scrivania davanti al computer attendeva con gli occhi semichiusi per la stanchezza. «Allora, segna che il corpo viene portato in sala autoptica vestito con una giacchetta impermeabile marca Old Eastern di colore marrone chiaro...». Il giovane assistente cominciò a digitare. Il patologo la infilò in una busta trasparente. Poi cominciò a tagliare la camicia. «Una camicia bianca con evidenti macchie di sangue marca sconosciuta» e anche il secondo capo finì in una busta per le prove. «Passiamo alle scarpe. Marca Alexander, suola sporca di fango... calzini di cotone neri, anzi no, credo sia filo di Scozia». Messi anche quelli al sicuro, osservò i piedi. «Evidente micosi sull'unghia dell'alluce destro, questo non lo scrivere ma tieni l'appunto per l'esame. Passiamo ai pantaloni. Cinta di cuoio colore marrone con marchio indecifrabile sul retro. Pantaloni color crema marca... boh... rock qualcosa... scrivi rock puntini puntini... Nelle tasche» infilò la mano nella tasca destra. Vuota. Poi frugò la tasca sinistra. Cambiò espressione. «Alt. C'è qualcosa...» era un foglietto di carta a quadretti con uno scarabocchio. «Bene, che cos'è?». Un numero di telefono. «Bene, molto bene...» lasciò il tavolo di acciaio e andò al telefono. Compose un numero. «Sono Uccio Pichi, mi passi il vicequestore Bonanni per cortesia». Attese guardando l'assistente. «Questa è una cosa interessante Ugo, molto interessante». Quello annuì. Non poteva fare altro dal momento che era muto dalla nascita. La voce di Uccio era l'unica che da due anni risuonava in quella sala della morgue, e questo cominciava a pesare sulla sua psiche. Per vincere quel silenzio pesante e carico di morte Uccio parlava sempre. E continuava a farlo anche a casa. Ma non viveva da solo, Spartaco Pichi detto Uccio conviveva con una donna da tre anni e dare voce ai propri pensieri, anche quelli più segreti e intimi, poteva farlo incorrere in situazioni sgradevoli. Ma non riusciva a frenarsi e sapeva che quella convivenza aveva i giorni contati. «Dimmi Uccio» la voce del vicequestore dall'altra parte del telefono era annoiata e stanca. «Allora, il corpo senza identità aveva un bigliettino in tasca. C'è segnato un numero di cellulare». «Ottimo!». Bonanni sembrò risvegliarsi. «È un punto di partenza. Dammelo!». Aspettavano il dolce. Gabriele s'era alzato per andare al bagno. Rocco giocherellava con uno stuzzicadenti. Lo teneva per le punte fra pollice e indice e cercava di raccogliere nella testa tutti i dati che si erano accumulati durante il giorno. Premeva il legno provando a spezzarlo. Le punte acuminate creavano un avvallamento nei polpastrelli. Abbandonò le riflessioni sulla rottura di coglioni del decimo livello, ossia il caso Juana Perez, e si concentrò su quella stupida operazione. Ormai stava diventando una lotta fra lui e il legnetto. Aumentò la pressione, resistette al dolore e finalmente quello si spezzò lasciandogli due segni rossi sulle dita. Rocco le osservò. Chiuse gli occhi. Così lo trovò Gabriele tornato dalla toilette. Pensò si fosse addormentato. Si sedette e si limitò a guardarlo. Rocco riaprì gli occhi e prese il cellulare senza degnare il ragazzo di uno sguardo. Compose un numero. «Albe'?». L'anatomopatologo rispose con la bocca piena: «No, assolutamente, non mi rompi i coglioni». «Ascolta una cosa. I piccoli lividi sopra le caviglie...». «Stiamo parlando di Juana?». «Esatto. Di forma circolare. Può essere stato un bastone?». Per qualche secondo sentì solo un mugugnare indistinto. «Scusa, stavo ingoiando. Sì, un bastone, più che altro una sbarra. Sono segni perfettamente circolari. Motivo?». «Grazie. Piano piano il velo si sta alzando». «Fa' una cosa bellino, segnati tutte le idee che ti vengono e comunicamele lontano dai pasti o dalle ore di riposo. Così io e te si continua ad avere un rapporto di reciproco rispetto». «Che mangi?». «Mi sono preparato spinaci con la mozzarella squagliata sopra. Tu?». «Spaghetti alle vongole». «Cioè tu mi vuoi dire che sei in grado di fare la pasta con le vongole?». «Sono al ristorante». «Bella la vita! Sei con una donna?». Rocco guardò Gabriele. «No, sono con un adolescente grasso, pieno di brufoli e problemi, che ascolta l'heavy metal e che verrà bocciato, e in più lo vogliono espellere perché ha picchiato un cafone di nome Diego». «Vai a ruba la sera, eh? Portalo un giorno da me, chissà che non gli faccia bene». «Amico, questo ragazzo ha più problemi che capelli, eviterei di aggiungere altri traumi». «Bello sapere che il mio lavoro lo consideri un probabile trauma». «Perché, cos'è?» e chiuse la comunicazione. Gabriele lo guardava. «Ero io il ciccione coi brufoli che sente l'heavy metal?». «Ne vedi altri?». Il ragazzo si guardò intorno. «No. Allora è deciso, domani viene a scuola a parlare col dirigente?». «Deciso. E tu dovrai darmi qualcosa in cambio. Prima o poi mi verrà in mente. Ma tua madre è sicuro che sta a Milano oppure è fuggita con un amante?». «L'ho sentita oggi pomeriggio. È a Milano». Rocco si versò dell'acqua. «Ti arrabbi mai con lei?». «No. Dovrei?». «Te l'ho detto. Tua madre lavora come una bestia, tu dovresti fare il tuo dovere, cioè studiare, e renderle la vita più facile». In quel momento l'inno alla gioia, la suoneria personalizzata del vicequestore, risuonò. Rocco rispose senza guardare il display: «Che c'è Albe'?». «Con chi parlo?» fece una voce seria dall'altra parte. «No, con chi parlo io!» rispose Schiavone. «Vicequestore Bonanni, questura di Spinaceto, Roma, ora mi dice con chi parlo?». «Vicequestore Schiavone, questura di Aosta. Ci conosciamo?». Ci fu un silenzio dall'altra parte del filo. «Pronto?» fece Rocco. «Sei un collega?». «Perché mi chiami?». «Abbiamo trovato un cadavere in campagna, a Castel di Decima. Niente documenti e un solo foglietto in tasca. E sopra c'era scritto il tuo numero». Rocco poggiò il volto sul palmo della mano. Gabriele lo osservava in silenzio. Il cameriere servì i dolci e se ne andò senza dire niente. Il ragazzo, più per riflesso condizionato che per appetito, prese un cucchiaino di cioccolato e lento lo portò alla bocca. Gli occhi erano due punti interrogativi. Grigio in volto Rocco digrignò i denti. «Bonanni, quant'è alto 'sto cadavere?». «Aspetta... Un metro e settantatré». Rocco tirò un sospiro di sollievo. Non poteva essere nessuno dei suoi amici. Aveva lasciato Gabriele sul divano letto con un giornaletto e Lupa a fargli compagnia. Si era chiuso in stanza. Osservava il cellulare e pensava. Doveva tornare a Roma, almeno fino a quando la storia di Enzo Baiocchi non si fosse risolta. Perché quel cadavere puzzava di Enzo Baiocchi. E se fosse stato proprio quel corpo? Se lo augurava. Sebastiano non rispondeva, teneva il cellulare spento. Compose il numero di Brizio. Rispose al terzo squillo. «Brizio, sono io». «Che succede, Rocco?» la voce calda e rilassata dell'amico lo distese. «Ascolta, ma Seba dov'è?». «Non lo vedo da un paio di giorni. Che è successo?». «Hanno trovato un cadavere a Castel di Decima. Niente documenti. A parte un foglietto in tasca, con sopra il mio cellulare». Brizio ascoltava. «Gola tagliata». Sentì Brizio accendersi una sigaretta. «Pensavi a Enzo Baiocchi?». «Esatto. Mandano la foto in questura domattina, ma ci spero poco». «Me pare strano. Diciamo che quello è Baiocchi. La cosa che stona è la gola tagliata. Seba j'avrebbe tirato du' schioppettate in petto». «È quello che pensavo io. E poi Seba non gli avrebbe messo il mio numero in tasca. Per farmelo sapere m'avrebbe chiamato di persona. Anzi, sarebbe salito a offrirmi una cena». Rocco si lasciò andare sul letto. «Ma te sei sicuro che in qualche modo c'entra Baiocchi?». «Me lo sento. C'entra lui». «Cioè tu dici che ha ammazzato questo e gli ha messo il tuo numero in tasca?». «Non lo so. In effetti, Brizio, non ha senso». «Appunto. Magari è uno che doveva sentirti per chissà quale motivo...». Rocco puntava gli occhi sul soffitto. «Magari...». «Hai chiamato Furio?». «No». «Mo' lo chiamo io. Che fai, scendi?». «E non lo so». «Dico ufficiosamente». «Brizio, qualunque cosa faccia qui la vengono a sapere. Non mi chiedere perché, ma è così. In caso ci vediamo a Roma». «Io intanto metto la voce in giro. Certo è una bella rottura di coglioni questa». «Enorme. Supera il decimo livello, amico mio...». Chiuse la telefonata. La porta si aprì. Palpebre serrate, come un morto vivente, Gabriele entrò in stanza e si infilò sotto le lenzuola. Poco dopo anche Lupa saltò sul letto. A Rocco vennero gli occhi lucidi al pensare che qualcuno lo considerasse un'ancora di salvezza, un muro di protezione. Carezzò i capelli di Gabriele. Non si era accorto che quello era sveglio. Al tocco della mano sorrise e si sistemò meglio sul materasso. «Finalmente la conosco!» disse il dirigente scolastico alzandosi dalla poltrona e andando incontro a Rocco con la mano tesa. «Sono tre anni che Gabriele frequenta l'istituto e non l'ho mai vista». Rocco gli strinse la mano. «Sì, ma io non sono il papà di Gabriele». Il sorriso sulle labbra del preside sparì. «Ah». Rocco gettò uno sguardo truce al ragazzo. «Sono la persona... insomma il convivente della madre» poi si morse le labbra. Segno inequivocabile che aveva appena mentito, ma il preside non se ne accorse. Si limitò a ripetere: «Ah!». «E insomma, siccome sua madre...». Gabriele pronto gli suggerì: «Laura...». «Laura...» fece Rocco, «allora, siccome Laura sta a Milano per lavoro, insomma eccomi qua!» e allargò le braccia. «La prego, si accomodi che non...». «Mi scusi ma non posso, fra dieci minuti devo essere in questura. Sono un poliziotto, vicequestore Schiavone». Il sorriso tornò sul viso del dirigente «Ah. Bene. Molto bene. E allora andando subito al sodo. Gabriele ieri è stato protagonista di una rissa. Lei si renderà conto che io, in qualità di dirigente scolastico, devo prendere i provvedimenti del caso». Rocco annuì. «Ora si tratterebbe di una sospensione di qualche giorno, siamo a fine anno e diciamo che Gabriele terminerebbe qui la sua frequentazione». Rocco assentì. «Per altro so che i voti del ragazzo sono insufficienti, e quindi l'anno l'ha già perso. Non sarà certo un sette in condotta a rovinarlo». Rocco continuava a tacere. «Ecco, io credo però che sta alle famiglie il compito di sforzarsi affinché certi episodi non accadano più». «Capito, Gabrie'? Perdi l'anno. Grazie, signor dirigente scolastico, le sue parole sono state chiare e limpide» gli strinse la mano e si voltò per andarsene. «Io spero che non accada mai più». «Non si preoccupi, il ragazzo ha imparato la lezione, vero Gabriele?». «E spero anche che cominciate, lei e la signora, a stare dietro a Gabriele, controllare i suoi compiti, insegnargli l'educazione di cui ha bisogno». Rocco si fermò sulla porta. «Gabriele, esci un momento e portati Lupa, devo parlare con il dirigente scolastico». «Ma...». «Se ti dico esci tu che devi fare?». «Uscire» afferrò il guinzaglio di Lupa e sparì. «Caro preside». «Dirigente scolastico». «Dirigente scolastico» si corresse Rocco, «sicuramente lo seguiremo di più nei suoi studi perché è un caprone, ma di quelli d'alta montagna». Il preside annuì soddisfatto. «Ora però mi permetta due parole. Gabriele è un ragazzo estremamente educato, gentile, premuroso, sensibile. È ignorante come la porchetta, ma è onesto e sincero. Quel Diego che ha picchiato sono anni che insieme ad altri due massacra questo ragazzo fuori e dentro la scuola e Gabriele non s'è mai difeso, ha provato a parlare con un paio di professori ma diciamo che la situazione non è mai cambiata. Allora se lei mi ricorda giustamente i doveri di padre, io le rammento quelli di preside, pardon, dirigente scolastico: la sua scuola ha un problema con il bullismo, dottore, e lei non ha fatto niente per evitarlo. E quello che mi manda in bestia è che quando un ragazzo bravo e obbediente anche se ignorante come Gabriele si ribella, passa dalla parte del torto. Ecco, io la esorto a fare più attenzione a quello che succede nel suo istituto anche fuori dagli orari di lezione. Altrimenti sa cosa si rischia? Di perdere i ragazzi migliori per colpa di quattro deficienti. Lei a questo Diego ha mai fatto una ramanzina?». «Lo sa commissario...». «Vicequestore...». «Pardon, vicequestore... Diego è uno studente complicato. Ha già perso tre anni». «Ed era a conoscenza del fatto che picchiava con cadenza bisettimanale Gabriele?». «Io? No, so che è un tipo violento e...». «Vede, dottore, a me succede spesso di incontrare fuori dal carcere i peggio figli di puttana e osservo invece la legge accanirsi contro quelli più indifesi. E lo sa? È una cosa che mi ha stancato. Sembra un discorso da bar, ma è la realtà. E qui, con le dovute proporzioni, succede la stessa cosa. Abbiamo un occhio di riguardo per Diego che è uno studente difficile e picchiamo duro su Gabriele. Due pesi e due misure mi fa girare i coglioni. Gabriele si merita l'espulsione, fuori dall'istituto e bocciato senza riserve. Però mi aspetto un trattamento simile anche per l'altro». «L'altro l'hanno portato in ospedale!» disse calmo il preside. «E non se la sente di dire che se lo meritava? Non mi pare che Gabriele abbia aggredito un povero studente con gli occhiali, mingherlino e indifeso, no?». Poggiò la mano sulla maniglia e aprì la porta. «Ah, e per la cronaca io ripasserò le lezioni con il ragazzo, gli insegnerò a studiare, ma gli ho insegnato anche a picchiare. Solo che, essendo un caprone di alta montagna, il cretino l'ha fatto fra le mura scolastiche». «Questo è l'insegnamento che dà a suo figlio?». «Se fosse mio figlio avrebbe tutti nove e Diego lo avrebbe massacrato il primo giorno di scuola, all'uscita. Ma ci sto lavorando sopra e ne farò uno studente modello. Con permesso». Gabriele era in corridoio che l'aspettava. «Allora?» chiese restituendo il guinzaglio a Rocco. «L'hai sentito, no? Perdi l'anno». «Ma davvero lei pensa quelle cose di me?». Rocco si fermò. «Cosa?». «Che sono un ragazzo estremamente educato, gentile, premuroso, sensibile?». «Hai origliato». «Che significa?». «Hai spiato?». «Sì». «Penso solo che tu sia ignorante...». «Come una porchetta». «Esatto!» e riprese a camminare. Gabriele lo seguì euforico. «Cos'è la porchetta, Rocco?». Un uomo sui trent'anni con un baffetto che incorniciava le labbra superiori e la mosca sul mento. Capelli neri, viso magro. Rocco era abituato a guardare le facce dei cadaveri che poco somigliano a quando erano vivi, e quel viso non gli diceva niente. Restituì la foto appena stampata al questore. «Mai visto». «Un'idea se l'è fatta?» chiese Costa aprendo la porta dell'ufficio. Rocco e Lupa lo seguirono. «Assolutamente no». «Nessun documento, niente di niente tranne il suo numero scritto a penna su un foglietto di carta che teneva in tasca». Cominciarono a scendere le scale. Incrociarono Pierron che salutò i due portandosi militarmente la mano alla fronte. «Perché un uomo a lei sconosciuto aveva il suo numero? E poi oggi i numeri si mettono direttamente sul cellulare!». «Ma se uno non vuole lasciare traccia di chiamate lo fa da un telefono anonimo e non si segna sul proprio il numero chiamato». «È vero, Schiavone. Peraltro pare che il telefonino della vittima non sia stato trovato». Attraversarono il corridoio. Lupa controllava ogni angolo. «Gola tagliata. Sembra un agguato, no, Schiavone?». «Sicuramente l'assassino l'ha preso di sorpresa, questo sì». Uscirono dalla questura e si ritrovarono nel parcheggio. «Lei lo capisce, la prima cosa che mi è venuta in mente è Enzo Baiocchi. E credo di non aver mancato di molto il bersaglio». Rocco allargò le braccia. «Forse, o forse no. Venti anni di polizia a Roma creano parecchi nemici. Di questo almeno se ne sarà accorto». «Tutto dipende da come uno se li fa, i nemici». Costa sorrise stirando le labbra. «Ma veniamo a noi. So che lei è tornato sul campo». «Esatto, dottore». «Bene. Benissimo. Abbiamo novità dal cellulare di Juana Perez?». «Ancora no. E aspetto notizie dalla scientifica». Come se l'avesse chiamata, si materializzò la Zaz della Gambino che parcheggiò con rumore di ferraglie. Il sostituto scese dall'auto. Stavolta aveva optato per un cardigan di cotone liso sopra una polo di un centro fitness. La borsa di pelle le scivolò a terra e lei si chinò a raccoglierla. «Secondo lei è brava?». «Credo di sì, dottor Costa. Certo, strana è strana». «E chi non lo è?». Finalmente la donna li raggiunse. «Buongiorno questore, salve Schiavone!». «Gambino, qualcosa di interessante per noi?». «Per caso sappiamo se Juana Perez andasse a cavallo?». La domanda lasciò i due poliziotti in silenzio. «Mi', andiamo bene. Abbiamo trovato...» e dalla borsa tirò fuori una decina di bustine. «Ecco, innanzitutto questa... è un frammento di una scatola, sicuramente medicinali. Sopra si legge solo Ox. Oxi qualcosa...». Rocco e Costa si passarono l'evidenza. «Invece queste altre tracce, vedete? Peli. Che possono essere un crine, oppure un animale domestico. Escluderei topo e anche volpe». «Volpe?» fece Costa. «Volpe, sì. Una volta in un omicidio vicino Benevento in una casa in campagna trovammo diverse tracce di volpe». «Dottoressa Gambino» disse Costa, «Juana Perez abitava in città. Mi sembra curioso pensare a una volpe. A meno che...». Costa si voltò verso Rocco. «Dottor Schiavone, è andato nell'appartamento insieme a lei?» e indicò Lupa che se ne stava accucciata. «Al suo Saint-Rhémy-en-Ardennes?». «No...». «Quel cane è tuo?» la Gambino strabuzzò gli occhi. «Sì. È vietato?». «E magari vive in casa con te?». «E dove deve stare?». «Ma porca miseria! Allora sai quanti peli di quella bestia hai addosso?». «Non è una bestia, si chiama Lupa» precisò Rocco. Lupa infilò il muso in mezzo alle zampe. L'istinto di sopravvivenza la stava avvertendo che l'argomento dell'alterco era lei, e sapeva che in quei casi la cosa migliore da fare era mostrarsi contrita e molto, molto pentita. «Sulle scarpe, sui pantaloni, sul maglione, peli ovunque! Una pelliccia! E li vai seminando in giro! E dire che ti avevo anche avvertito di non inquinare la scena del crimine!». Rocco si morse le labbra. «E mi dispiace...» fece, «starò più attento». «Tu e il tuo cane!». «Aho! Mo' basta! Vuoi continuare a smembrarmi l'apparato riproduttivo o possiamo tornare a lavorare?». Costa si intromise per sedare la discussione che stava prendendo una brutta piega. «Grazie, dottoressa Gambino, lei è davvero preziosa». «Lo so dottor Costa, lo so, e mi faranno fuori» era diventata improvvisamente calma. Lo sbalzo di umore colse i due uomini di sorpresa e restarono lì a guardarla stralunati. «Quelli bravi danno fastidio, non lo sapete? Spesso la verità è scomoda. Se qui dentro c'è una mano potente che ha guidato l'omicidio, be', mi creda, io, lei, la questura non ne verremo a capo». «Potente? E perché?» fece Costa. La donna si tolse gli occhiali. «Perché i fili li muovono sempre poche persone». Costa e il vicequestore si scambiarono uno sguardo. Rocco cercava di comunicare al suo superiore di soprassedere a quella discussione e di tagliare corto, ma Costa non afferrò al volo il suggerimento. La Gambino prima si guardò alle spalle, neanche fosse seguita da un commando del Mossad, poi con voce decisamente più bassa attaccò: «Non è un mistero che a comandare il mondo sia il club dei 300. Volete sapere chi sono?». «Immagino che la spigolatrice di Sapri non c'entri nulla» commentò Rocco. «Non siate ingenui» rispose Gambino con un sorrisetto saccente. «I 300... volete saperlo o no?». Costa allargò le braccia. «Se pensa sia una cosa utile, ma facciamo una cosa di giorno». «Sì, prima che ci sorprenda il santo Natale» aggiunse Rocco. «Allora vi dico chi è a capo dei 300?». «E ce lo dica» sbottò Costa. «La regina Elisabetta II d'Inghilterra». «Li comanda lei?» chiese Costa guardando il parcheggio. Sperava arrivasse l'agente con l'auto, un funzionario, qualunque cosa pur di togliersi di lì. «Già, quell'adorabile vecchina. In mezzo ci sono Rothschild, Rockefeller, Gustavo di Svezia, Gorbaciov». «Pure Gorbaciov?». «Guardi dottore, lei potrebbe restare di sasso se le rivelassi alcuni segreti...». Rocco provò ad approfittare della pausa a effetto lasciata dalla Gambino per cercare di tagliare corto. Aprì la bocca, ma la donna ormai lanciata lo anticipò. «Tutto si spiega attraverso il comando dei 300. La guerra in Iraq, le torri gemelle, il complotto dell'euro, Putin e gli oligarchi, Kim Jong-un, tutto!» e li guardò con passione. Rocco sorrise. «Bene. È tutto chiaro». «Chiarissimo» fece Costa. L'auto guidata da Casella frenò a pochi metri dal gruppetto. «Ma è arrivata la mia macchina. Grazie Gambino, sono felice che lei si sia unita a noi, avevamo proprio bisogno di una persona efficiente e capace come lei» e sgattaiolò lasciando Rocco ad affrontare da solo il sostituto. «Ora, Rocco, sai una parte di verità. Guarda il mondo attraverso quello che hai imparato oggi e non ti sembrerà più lo stesso. Non dirmi poi che non ti avevo avvertito». «No, Gambino, lo ricorderò» e si voltò per entrare in ufficio. «Schiavone!». «Che c'è?». «La prossima volta sulla scena del crimine fai entrare prima noi. È meglio!». Aveva provato a chiamare Sebastiano altre tre volte ma il telefono era sempre staccato. Da Brizio nessuna novità. Il cervello si era inchiodato al cadavere trovato a Castel di Decima, non riusciva a pensare ad altro. Si doveva muovere, stare lì ad Aosta in attesa degli esami della scientifica o del cellulare di Juana era una perdita di tempo. Caterina capitò nel momento giusto. Lupa appena la vide saltò dalla poltrona e le andò incontro per farle le feste. «Caterina! Novità?». «Ancora niente da Scipioni. Intanto ho qui tutte le schede degli altri affittuari di Berthod nella palazzina di via Brean». Poggiò un plico sulla scrivania già invasa di scartoffie. «Fantastico. Ora io ho bisogno di te. Mi devi tenere Lupa». La ragazza sbuffò alzando gli occhi al soffitto. «Ancora? Che succede?». «Una brutta storia a Roma. La saprai tra poco, non è un segreto». «Di nuovo Baiocchi?». «Credo di sì». «Ci vai ufficialmente o ufficiosamente? Insomma, Costa lo sa?». «Non lo sa». Caterina scosse la testa. «Chiaro che non vuoi che ti venga a dare una mano». «Chiaro». «E se chiedono di te?». «Mi sono preso la febbre». «Febbre? L'abbiamo già usata». «Un parente in fin di vita?». «Non ne hai, Rocco, e lo sanno tutti». «Un incidente domestico? Tipo la perdita di una tubatura a casa mia che richiede la mia presenza?». «No. Te lo dico io. Sfrutta lo psichiatra di Torino». «Io non ho uno psichiatra a Torino». «Parlo di quel Berlingeri, il marito della Tombolotti. Di' che sei andato da lui a parlare della vittima che forse si era rivolta a lui. Anzi ci inventiamo che l'aveva contattato e ti serviva un quadro generale della situazione psicofisica di Juana Perez per proseguire le indagini». «Berlingeri è negli States, credo» disse Rocco pensieroso. «E da quando ti fai tutti questi problemi? Io dirò così, e tu vedi di non stare via troppo». Rocco la guardò negli occhi: «Grazie Caterina, come sempre. Italo non sa cosa si è perso». Caterina poggiò le mani sulla scrivania. «No scusa, che c'entra Italo?». «C'entra eccome!». «Non la pianterai mai di ficcanasare nella mia vita privata?». «Vale per tutti i miei collaboratori. Per esempio, sapevi che D'Intino benché nato in Abruzzo ha una madre molisana?». «E questo che c'entra?». «Un cazzo, ma serve a depistare!» si alzò e raggiunse Lupa. «Ci risiamo amore. Fa' la brava e da Roma ti porto un regalo». «Rocco?». «Che c'è?». «Stai attento. Quello è ancora vivo». Scese le scale con un mezzo sorriso sul volto. Sapeva il motivo di quella strana serenità che si era andata a mischiare al sangue nelle vene, e un po' se ne vergognava. Ma non poteva negarlo, le ultime parole di Caterina erano state una carezza. «Sei peggio di un liceale» si disse, e pensando all'adolescenza si ricordò che doveva lasciare un bigliettino a Gabriele per avvertirlo che quella sera non poteva restare a dormire da lui. Il telo si sollevò. Il corpo sembrava di cera. I tagli sulla carne erano stati ricuciti. Non aveva più niente di umano. Era un oggetto e Rocco faticava a pensare che quella materia dalla forma vagamente antropomorfa fino a tre giorni prima camminava, rideva, e un tempo era stato un neonato, tondo e paffuto, e la madre l'aveva tenuto fra le braccia. Gli succedeva sempre più spesso, anche nelle macellerie, guardando i quarti di bue aperti con le ossa al vento, non riusciva più a immaginare l'animale al pascolo in mezzo a un prato, a strappare fili d'erba, a scacciare le mosche con la coda. «Cominci a invecchiare» gli disse Uccio Pichi. Rocco stava lì a osservare il corpo martoriato di quell'uomo senza nome. «Pensavo che con l'età ci si facesse l'abitudine a certe cose. E invece...». «Sai cos'ho scoperto? Può dipendere dall'umore». Uccio si allontanò dal lettino autoptico e andò a prendere una bottiglia di acqua poggiata su una mensola. «Se mi sento giù, lavorare mi risolleva lo spirito. Se invece sono felice rischio di sprofondare nel baratro». Rocco distolse gli occhi dal cadavere: «Cioè fammi capire. Se sei triste aprire un corpo ti aiuta?». «È così. E penso che è un caso che su quel lettino non ci sia io». «Sarà come dici te. Io credo invece che dentro di me non c'entra più neanche uno spillo. Comunque» e fece cenno a Uccio di ricoprire il morto «... mi fai vedere i vestiti?». «Vieni...» si incamminarono verso lo studio attiguo. Nella stanza correva una libreria carica di scatole e di libri di medicina. Un odore pungente di muschio e detersivo si infilava nelle narici e come fumo di un camino si avviluppava agli abiti. Su un tavolo enorme al centro c'erano disposti gli indumenti della vittima. Rocco li osservò attentamente. «Aveva qualcosa sotto le unghie? Epidermide, stoffa...». «Niente. Segno che è stato colto alla sprovvista. Magari alle spalle. Insomma non ha lottato». «Denti?». «Ordinaria amministrazione. Sani». Quei vestiti anonimi non gli dicevano niente. «Tutta roba che si può trovare ovunque». «Già» fece Uccio. Rocco prese le scarpe, le osservò. «Queste sono ottime scarpe, ma le venderanno in decine di negozi». Le poggiò sul tavolo. Prese la cinta. «Qui dietro c'è un logo» e lo indicò a Uccio. «Sì, l'avevo notato. L'ho fotografato e mandato in questura. Ma ho poche speranze». Rocco si mise le mani sui fianchi scuotendo la testa. «Chi sei?». «E perché aveva il tuo numero in tasca?». «Fammi vedere il bigliettino». «I bigliettini» disse Uccio andando a prelevare una piccola scatola dalla libreria. «Perché i bigliettini?». «Nei pantaloni aveva il tuo numero, nella giacca uno scontrino». Rocco tirò fuori le due bustine. Il suo cellulare era scritto su un foglietto a quadretti. Afferrò la busta con lo scontrino. «Bar Mastrangeli, via Vitellia 31...». «Già». «Questa strada è a Monteverde» fece Rocco. «Ah sì?». «È vicino casa mia». I due si guardarono. Neanche dieci minuti dopo era sotto casa di Sebastiano a Trastevere. Citofonò ma nutriva poche speranze che quello rispondesse. Si affacciò sora Letizia, la vecchia vicina del primo piano. «Chi è?». «Sora Letizia, sono Rocco». «E com'è a 'st'ora?». «Sto a cerca' Sebastiano». «È da un po' che nun lo vedo. Voi entra' fijo?». «Me apra, sì». La vecchina sparì in casa. Poco dopo aprì il portone. Rocco salì al primo piano. La donna era già sulla porta. «Me fai preoccupa'» disse a Rocco. «Ce stavo a pensa' pure io, non lo vedo, non lo sento. Da quando la povera Adele se n'è annata non lo riconosco più...». «Lo chiamo dall'altro giorno e non mi risponde». «Vieni, entra» e si scansò. Rocco puntò verso il balconcino della cucina. In casa c'era odore di fritto. «Come sta Sabatino?». «Domani se va a fa' le transaminacee. Mo' sta già a fa' una pennica. Prego, vieni». «Sora Leti', se deve fa' le transaminasi e je prepara i fritti?». «Boni però» fece la donna con un sorriso birichino. «Mele, cavolo e pure er baccalà». Aprì la finestra. Rocco uscì sul balconcino che dava sul retro del palazzo. Poggiò un piede sulla balaustra, afferrò il divisorio e ci montò sopra. «Sta' attento Rocco, nun te fa' male». Il vicequestore si aggrappò al parapetto del piano di sopra, si tirò su, lo scavalcò e atterrò sul terrazzino di Sebastiano. La porta-finestra era chiusa, dentro buio. Prese il coltellino svizzero e dopo neanche un minuto l'aveva aperta. Entrò. Accese la luce. L'appartamento era in ordine. Il letto intonso, la cucina pulita. Sebastiano aveva riempito casa con decine di foto di Adele. Ce n'era una presa al mare, dove la ragazza coi capelli scompigliati dal vento sorrideva felice. Era piazzata sul mobile del salone dietro un vasetto con dei fiori di creta viola. Nel cassetto all'ingresso trovò le copie della V-Strom 650, la Suzuki di Sebastiano, ma non le originali. Non c'era bisogno di andare a controllare al garage di vicolo del Cinque. L'amico era partito. E teneva il cellulare spento. Questo lasciava due alternative: o non voleva farsi trovare oppure Sebastiano non era più di questo mondo. Percorreva i viali del cimitero mentre il sole ormai quasi spento sbavava di arancione i tetti delle cappelle. Superò i pini. Le lapidi erano davanti a lui, in file lunghe si perdevano fra i pendii del prato. Un uomo e una donna, lontani, stavano in piedi con la testa china. Superò la terza fila e arrivò alla tomba di sua moglie. I fiori erano freschi, Laura e Camillo, i genitori di lei, dovevano essere passati da poco. Ai piedi della lastra di marmo con il nome e le date c'era ancora il quaderno che le aveva lasciato. Quello sul quale Marina segnava le parole difficili. Si era gonfiato e scolorito, ma la piccola tettoia in alto lo aveva riparato dalla pioggia. La matita invece non c'era più. «Ciao Mari'... eccomi qua». Prese le margherite e le infilò accanto ai fiori lasciati da Laura e Camillo. Sfogliò le pagine. «Hai aggiunto qualcosa?». La grafia minuta e precisa di Marina gli strinse il cuore. Qua e là la grafite si era cancellata, soprattutto vicino agli angoli delle pagine, il resto era ancora leggibile. Lo sguardo gli cadde sulla parola eonismo. Travestitismo, aveva scritto Marina, deriva dal cavaliere d'Eon Charles de Beaumont, che amava indossare abiti femminili. «Eh sì, proprio così» disse Rocco. «Ma allora lo sai. Sai quello che mi succede...». La parola che seguiva era ancora più criptica. Epatta. Numero che rappresenta la differenza in giorni fra l'anno lunare e quello solare. «Questa non l'ho capita Marina, proprio no. Me la devi spiegare. Che stai dicendo?». Lo stava prendendo in giro? Che c'entravano ora la luna e il sole? Eppure Marina non parlava mai a caso. Rimise a posto il quaderno. «Qui succedono un sacco di cose brutte, amore mio. Sono in mezzo alle ombre, sono dappertutto, e non si può acchiappare un'ombra. Però devi tornare a trovarmi, magari una volta ogni tanto, mica sempre. È da un pezzo che non vieni». Poi alzò lo sguardo. Sul cielo che si stava imbrunendo notò che la luna era già salita. Era una scheggia di luna, bianca e polverosa. Solo allora capì e sorrise. «Devo aspettare? E vabbè, io aspetto». Le mandò un bacio e si alzò. Dopo pochi passi arrivò alla lapide di Adele. Come promesso, l'avevano sepolta vicino a Marina. Sebastiano aveva scelto una foto dove rideva. Quelle due lapidi segnavano la stessa cicatrice nel cuore di Rocco, e non avrebbe smesso di sanguinare. Marina e Adele, morte al posto suo. Per due volte era stato graziato ed era stato peggio che morire. Sulla tomba di Adele mancavano i fiori. In un vaso di cristallo rosso ce n'era un mazzo ormai rinsecchito. Non era un bel segno, Sebastiano non veniva a trovarla da un pezzo. Sostituì i rami secchi con un mazzo di azalee. Ritornò sui suoi passi, mandò ancora un bacio a Marina. Non c'era più nessuno. Doveva affrettarsi. Cominciava l'ora in cui i vivi non erano più desiderati in un cimitero, quando la luna e la notte scacciavano le ombre, sbiadivano i colori e scandivano il tempo senza meridiana. Anche se tardi, a Roma Sparita per loro c'era sempre posto. Aspettando la pasta Rocco tirò fuori la foto del cadavere e la mostrò a Brizio e Furio. Che scossero la testa. «No, mai visto» e Furio restituì il foglio. «Manco io». Rocco se la rimise in tasca. «Da quando non vedete Seba?». «Da un po'. Un tre giorni?» chiese Brizio a Furio, che annuì. «Non è manco andato a trovare Adele al cimitero». Rocco li fissò. «Che sta succedendo?». «È da ieri che ci penso» fece Brizio, poi bevve un sorso di vino. «Le cose stanno così. Se c'è di mezzo Enzo Baiocchi come dici te, perché ha ammazzato questo? Che c'entra con Enzo, con te?». «Infatti non quadra» disse Rocco. «In più m'hai detto che la polizia dice che non è stato ucciso nel campo a Castel di Decima. Quindi ce l'ha portato». «E buttato in un cespuglio». «Dunque l'assassino ha una macchina, Baiocchi?». «Magari ha usato quella della vittima. Ma so' tutte supposizioni» concluse Rocco. In quel momento arrivò il cameriere. «Eccole qua, tre cacio e pepe da fa' resuscita' i morti» disse. «Magari! Così quello ce dice chi è» esclamò Furio. «Quello chi?» chiese il cameriere. «Niente Aldo, lascia perdere» e Rocco si versò un bicchiere di vino. «Come volete voi. Buon appetito!». Aspettarono che si allontanasse. «Sentite un po'. Enzo Baiocchi è da solo, lo cerchiamo noi e lo cerca la polizia da quando è evaso dal carcere di Velletri. Fino a qui ci siamo?». «Mi pare proprio di sì, Rocco» rispose Brizio con la bocca piena. «Bene. Cosa fai da solo? Mica puoi stare nascosto. Enzo Baiocchi non è un latitante della 'ndrangheta. È un cane sciolto, e senza appoggi». I due amici masticavano e riflettevano. La sera di giugno stava regalando una brezza leggera che muoveva l'edera sul muro e giocava coi capelli. «Allora rubi una macchina e cosa fai? Minimo cerchi di passare la frontiera». «E devi pure rimedia' i documenti». «Ma per rimediare i documenti hai bisogno di soldi. In cosa era forte Enzo Baiocchi?». I due amici non risposero. «Ve lo dico io. Se faceva i benzinai». «I benzinai?». «Esatto. Mo' al di là di questo cadavere che non sappiamo chi è, dove ve ne andreste voi se foste Enzo?». «Slovenia, una volta che ho i documenti in Slovenia. Da lì in poi chi me trova più». Rocco annuì. «E la Svizzera?». «No. Pericolosa. Slovenia tutta la vita» aggiunse Furio. «Facciamo che Seba l'ha capito, non so come, e s'è messo dietro a Enzo. È lì che puntereste?». Brizio scosse il capo. «Troppo vago, Rocco. Qualche traccia ce la dobbiamo avere. Troppo vago». «La targa della moto di Seba ve la ricordate?». «Ma che cazzo ne so!» fece Brizio. «E poi perché?» chiese Furio. «Perché se sta in giro magari una multa, un'autostrada, una cazzata e ci può dare una mano, no?». «Troppo vago» ripeté Brizio. «Hai rotto il cazzo! Io da qualcosa devo comincia'!» sbottò Rocco. «I soldi» fece Brizio. «Cioè?». «Per andare in giro hai bisogno di soldi. Io e Seba abbiamo il conto alla stessa banca. Il direttore è amico mio». «E che voi fa'?» chiese Furio. «Domani ce vado a parla'. Se ha usato il bancomat a quelli gli risulta. Hai visto mai?». «È già una cosa» fece Rocco. «Io mi muovo con la targa». Rientrò in casa e l'assalì un odore di polvere e di chiuso che gli tappò le narici. Andò subito ad aprire la finestra del balcone. Il limone se ne stava nel solito angolo e aveva messo i fiori. Negli altri vasi c'era il disastro. Rocco non aveva il pollice verde, quello era compito di Marina. Gli sterpi infilati nel terriccio ormai arido, le foglie accartocciate e mai raccolte sembravano dire la stessa cosa. Che il tempo passa e le stagioni si avvicendano e toccava a lui andargli dietro altrimenti restavano rami rinsecchiti buoni solo per accendere il fuoco. Una volta quel terrazzo era un trionfo di colori, e quando il sole ci batteva sopra esplodeva in tutti i suoi profumi. C'erano le bouganvillee, i rincospermi, le petunie, il glicine rampicante, e fiori che Rocco non aveva mai imparato a nominare. Il viola si sposava con il rosa e generava il cremisi, il bianco si intrufolava nell'azzurro e nel giallo. Sembravano gareggiare in potenza, come un fuoco d'artificio che impazziva ogni maggio e si andava a riparare coi primi freddi in attesa della primavera che sarebbe tornata. Adesso non c'era più niente. Solo il limone, isolato fra gli altri vasi, sembrava guardarlo senza riconoscerlo. Si sedette sulla panchina di ferro dove una volta cadevano i rami del glicine. Si rollò una canna e l'accese guardando le luci e il cielo di Roma attraversato da due aerei che sembravano in rotta di collisione e che invece chissà a quale distanza stavano viaggiando. La luna dondolava pigra in mezzo a due puntini di stelle. Gli pesava addosso quel cielo, gli pressava i polmoni. Anche la marijuana aveva un brutto sapore. All'interno la plastica che rivestiva i mobili crepitava al vento leggero. Lontano tuonava il traffico. Prese il telefono. Lo accese. La luce azzurrognola del display gli illuminò il viso in ombra. «Cateri'? Dormivi?». «No Rocco». «Tutto bene? Lupa come sta?». «È qui accanto a me. Non mi hai mai chiamato da quando sei a Roma. Che succede?». «Niente» aspirò una boccata di fumo. «Ci sono novità?». «Ancora nulla... Tu quando rientri?». «Al più tardi domani in serata». Per qualche secondo non parlarono. «Che c'è Rocco? Che hai?». «Niente. Avevo bisogno di sentirti». Nel buio la brace della canna gli illuminava appena il naso. «Rocco, io...». «Buonanotte Caterina. Ci vediamo domani» e chiuse la comunicazione. Spense la cicca in un vaso, si alzò e rientrò in casa senza accendere la luce. Preferiva guardare le ombre dei mobili e il riverbero tenue che entrava dalle finestre. Ma sonno non ne aveva e la massa scura del tavolo, del divano, del televisore pareva minacciarlo. Gli sembrò che volessero parlare, dirgli qualcosa che lui non aveva voglia di stare a sentire. Se ne andò in camera da letto. Lì era ancora peggio. Il materasso respirava, come l'armadio vuoto che chissà perché aveva spalancato le fauci delle ante. Si sentì un ospite indesiderato, di quelli che si invitano per dovere ma di cui non vediamo l'ora di sbarazzarci. «Questa non è casa mia» mormorò. Fino a quel momento si era nascosto in una piega del tempo, lo capì solo quella sera. Aveva finto di guardare, di pensare, di vivere. S'era trascinato fin lì, in quell'appartamento che non aveva più una voce familiare e un colore, un odore che lo riportassero indietro. Tutto era evaporato, si era dissolto, lasciando solo lo scheletro delle forme che una volta conosceva. Come da un fuoco che ardeva nel camino era rimasta cenere grigia, fredda e inospitale. Tornò in salone, si rimise la giacca e uscì di casa, alla ricerca di un albergo. L'alba era spuntata da poco. Caterina infreddolita si chiuse il bomber e guardò il cielo. Lupa pigra annusava i marciapiedi. Il sole coperto dalle nuvole aveva fatto capolino dalle cime delle montagne. La città sembrava ancora dormire, solo un'auto lontana correva verso l'autostrada. Poi la vide, ferma all'angolo davanti al negozio di casalinghi ancora chiuso. Portava un giubbino viola con riflessi argentati e la solita tuta con le scarpe da ginnastica vecchie e sporche. La stava guardando fissa senza decidersi a muovere un passo verso di lei. Fu Caterina ad attraversare la strada per andare incontro a sua madre. Man mano che si avvicinava notava gli occhi rossi e gonfi di chi ha passato la notte in bianco. Teneva le mani nella tasca del giubbetto e spostava lentamente il peso da un piede all'altro. «Cosa vuoi?» le chiese. Sua madre chiuse gli occhi. Una lacrima le scese sulla guancia. «Non ce l'ha fatta» disse con voce roca. Caterina annuì. Poi tornò verso Lupa che aveva finalmente scelto l'alberello giusto. «Non mi saluti neanche?» le gridò dietro sua madre, ma Caterina non si voltò. Per anni si era chiesta come avrebbe reagito quando le avessero portato la notizia della morte di suo padre. Immaginava che si sarebbe sentita liberata, invece si stupì della mano di ferro che le strizzava lo stomaco e ancora di più nel percepire gli occhi inumidirsi. Non era colpa dell'aria fresca del mattino, ma stava accadendo. Non immaginava di ritrovarsi così, divisa perfettamente a metà, con il cervello che le diceva di cominciare la giornata di lavoro in questura e il corpo che non riusciva a compiere le azioni più elementari. Due lacrime dagli occhi, i polmoni spremuti, la pelle rinsecchita. E dentro una voce che le ordinava di andarsi a fare una doccia, vestirsi, dare la pappa al cane. Non voleva piangere la morte di suo padre, ma da quel momento Caterina non aveva più alibi, era arrivata l'ora di guardarsi davvero allo specchio. «Caterina?» la chiamò la madre. Allora si voltò. La vide piccola e sconfitta dagli anni, con le braccia aperte. «Ti prego, vieni qui da me». «Mi dispiace. Non ho tempo e tu sei in ritardo» e finì di attraversare la strada. Lupa l'aspettava davanti al portone. Infilò le chiavi ed entrò senza voltarsi indietro. Solo quando l'anta si chiuse rimbombando nella tromba delle scale, si appoggiò alla ringhiera per scivolare sul primo gradino. Voleva dare sfogo al pianto, ma non le riuscì. Si era bloccato da qualche parte in mezzo alla gola. Papà è morto, pensò. «Poteva farlo prima» si rispose. «Dai Lupa, andiamo a fare la pappa» e si rialzò. Salì le scale. Dopo la doccia e la colazione le venne in mente di chiamare il vecchio amico, quello che l'aveva aiutata, a cui doveva tutto, compresa la divisa che in quel momento riposava nell'armadio. «Dottor Mastrodomenico? Sono io, Caterina». «Ciao Caterina, hai novità?». «Non ce l'ha fatta». Il dirigente non parlava. «Se n'è andato sotto i ferri». «È una bella morte. Tu come stai?». «Mi scusi se l'ho disturbata, volevo solo che lo sapesse». «Hai fatto bene, Caterina. Ci sentiamo...». «Ci sentiamo». Scese alla stazione di Torino che non era ancora mezzogiorno. Aveva contattato un suo compagno di liceo alla motorizzazione per avere notizie sulla targa di Sebastiano. Poi aveva sentito De Silvestri, il suo vecchio amico, agente al commissariato dell'EUR, per tracciare gli eventuali spostamenti di quella moto. In più gli aveva chiesto di verificare negli ultimi dieci giorni se ci fossero stati furti presso benzinai di Roma e provincia. Dalla banca Brizio non aveva ricevuto ancora notizie, restare in città era inutile. L'auto con Casella a bordo lo attendeva a via Nizza. Appena montò percepì un odore di menta chimica nauseante. «Casella, devi cambiare profumo, pari er pavimento di un ospedale» fu la prima frase che gli uscì dalla bocca mentre si accomodava sul sedile mezzo sfondato della Lancia. «Non sono io. Io non uso profumo. Anzi, finalmente mo' ho la doccia che funziona. Ho comprato una caldaia che...». «E allora 'sta puzza di menta mista a uovo sodo marcio che è?». «Deve essere l'arbre magique, D'Intino li mette in tutte le macchine della questura». Rocco cercò con lo sguardo la sorgente di quella puzza che impestava l'abitacolo. «Non lo vedo 'st'arbre magique». «Li sbriciola nel portacenere». «Ma perché non lo trasferiscono in Abruzzo 'sto mentecatto?». «Faccia una lettera e così non ci pensa più!». «Sì, ma deve andare in un posto tremendo, dove ripensare ogni giorno alla bella vita che poteva fare qui se non fosse l'idiota che è». Casella sorrise: «Io un'idea ce l'avrei». «Spara!». «Faccia una cosa, non lo mandi in Abruzzo che da quelle parti hanno già i problemi loro, non gliene aggiungerei un altro. Lo mandi a San Severo, vicino Foggia». «Tosta?». «Vuole che non lo sappia? Ci sono nato». «San Severo, provincia di Foggia. Segnato». Rocco si accese una sigaretta. Casella fece una smorfia, Rocco lo prevenne: «Fino ad Aosta capace che me ne fumo un'altra, quindi guida e non mi rompere il cazzo». «Si figuri dotto'... tutto a posto. Com'è andata con lo psichiatra?». «Eh?». Rocco aveva dimenticato l'alibi da due soldi concordato con Caterina. «Doveva incontrate 'sto Berliguer, Berlinger...». «Ah, sì, bene. Arrivato niente dalla magistratura?». «Boh...». Le strade erano libere, e l'auto già attraversava i quartieri periferici. «Di' un po' Casella, ma tu una moglie? Una compagna?». «Ce l'avevo. Tanti anni fa. Non eravamo sposati. Poi mi ha lasciato per un rappresentante di mulinelli da pesca». «Hai sofferto?». «No. Non l'amavo più». «E da allora come ti arrangi?». Casella si grattò la testa. «E che le devo dire? Ogni tanto un'avventuretta qua, un'avventuretta là...». «Paghi?». «Sia chiaro. Io non ho mai cacciato una lira per fare l'amore, gliel'ho già detto!». Rocco spense la sigaretta nel portacenere. «Che ne pensi di questa storia?». «Del trans? Boh, a chi piace...». «Secondo te perché hanno svuotato l'appartamento?». «Dotto', ma che ne so? Io a certe cose mica ci arrivo». «Buttala lì». Casella si mise a pensare. Tacque per alcuni lunghissimi minuti. Diminuì la velocità, si asciugò la fronte. «Secondo me...» disse a bassa voce, «qualcuno che si vergogna forte, ecco. Qualcuno che magari è sposato con una donna ricca che se scopre la tresca lo caccia da casa». «E svuota l'appartamento? Non faceva prima a scappare? Invece toglie vestiti e ammennicoli vari e getta il cadavere nel fiume. Un po' strano, no? Poteva scappare e basta». «Già. E le impronte digitali?». «Allora tu sospetti un pregiudicato». «Be', perché non può essere? E poi se magari le aveva regalato, che ne so, dei vestiti della moglie? Oppure delle cose che aveva in casa sua?». «Ma secondo te, perché l'ha strangolata?». Casella ingranò la marcia e accelerò verso l'autostrada. «Perché magari è uno che non sta bene con la testa, un mezzo matto, ha presente quelli che si menano per godere? Ecco, si sono menati e gli è saltato il piripicchio e l'ha strozzata. Comunque sia, è un figlio di puttana e noi gli dobbiamo mettere le mani addosso!». «Su questo ti do ragione, Case'. E secondo me con il fatto delle impronte non ci sei andato lontano». Casella sparò un bel sorriso: «Veramente?». «Veramente». Quando superarono Pont-Saint-Martin il cielo era nuvoloso e il sole era sparito. Italo gli andò subito incontro insieme a Caterina. Avevano il viso stanco, grigio, se avesse potuto li avrebbe mandati una settimana al mare. «Dottore... ci sono novità». «Innanzitutto dov'è Lupa?». «Dorme nella sua stanza» rispose Caterina con un filo di voce. «Ditemi...». «Allora abbiamo controllato l'alibi di Ivano Petrulli, ricorda? Il ragazzo del primo piano». «Il lemure, sì...». «Il lemure?» chiese Caterina. «Dunque?». «Aveva detto che domenica, il giorno dell'omicidio, aveva dormito dalla fidanzata a Saint-Vincent. Ora...» Italo controllò i fogli, «Raimonda Ferretti, questo è il nome, domenica non era a Saint-Vincent». «Ah no?». «No. Era a Milano. Il giorno dopo ha avuto un colloquio di lavoro e ha dormito lì». «Che coglione...» fece Rocco. «Non tu Italo» si premurò a specificare notando l'espressione dell'agente. «Come ha fatto questo Petrulli a prendere una toppa simile?». «Di solito la Ferretti dormiva a casa di amici. Stavolta ha scelto un albergo. Se avesse mentito lo avremmo saputo. Evidentemente Ivano non era a conoscenza della cosa e l'alibi è andato a farsi benedire». Rocco si toccò il mento. «E il giorno dopo? Era dai suoi a Cogne come ci aveva riferito?». Intervenne Italo. «Ecco. Qui sorge un problema. Suo padre, l'avvocato Petrulli, ha detto che risponde solo se convocato in questura». «Per dirci dov'era suo figlio vuole essere convocato in questura?». «Già» disse Caterina. Faticava a parlare e guardava fisso in terra. Ingoiò e riprese. «Che facciamo?». «E noi lo convochiamo in questura. Così viene con il papà e prendiamo due piccioni con una fava». «Senti Rocco» Italo abbassò la voce e si avvicinò al vicequestore. «Scipioni ti aspetta al bar di fronte. Non so cos'avete combinato, ma è molto preoccupato». «Preoccupato?». «Già...». «Vado» fece per muoversi, poi si bloccò e guardò il viceispettore. «Caterina, cos'hai?». «Niente». Rocco la guardò. «Sicura? Vuoi andare a casa?». «Ma perché? No, sono qui, lavoro...». Piano piano gli occhi le si inumidirono. Rocco allungò una mano per toccarle la spalla ma il viceispettore si morse le labbra e piangendo scappò via lasciando Rocco e Italo in mezzo al corridoio. «Che succede?». «Ti giuro, Rocco, non lo so...». L'agente Antonio Scipioni era in piedi accanto al bancone. Nervoso si guardava intorno. Quando vide entrare il vicequestore scattò e gli andò incontro. «Vieni...» disse. Rocco lo seguì. Le macchine correvano sul corso Battaglione, i due girarono l'angolo di una palazzina, in una piccola via deserta, e finalmente Scipioni si arrestò. «Mi dici che succede?». «Una cosa molto strana, Rocco. Il mio amico alla compagnia telefonica». «Sì?». «Ha controllato i numeri del cellulare di Juana Perez. E non gli è mai capitata una cosa simile». Rocco si accese una sigaretta. «Dimmi un po'?». «Non c'è niente». «Come non c'è niente?». «Nessun numero composto o telefonate ricevute. Vuoto. Meglio, c'è solo il tuo di numero, quando l'hai chiamata per cercare l'indirizzo». Rocco aspirò il fumo. «Cazzo...». «Tu hai un'idea di cosa significhi?». «Pallida, come il sole oggi. Questa cosa, Antonio, non la deve sapere nessuno. Solo io e te». L'agente annuì. «Ma che succede?». «Succede che la cosa è molto più grave di quanto pensiamo. Anche se tutto sembrava volercelo dire. Allora dobbiamo spremerci le meningi e farci venire un'idea». Andò da solo a via Brean. Il cielo coperto si stava tingendo di grigio. Anche l'aria s'era raffreddata e un vento leggero strisciava in mezzo alle strade e alle case. La via era silenziosa e deserta, poche macchine parcheggiate. Rocco guardava il palazzo, un edificio come tanti, a due piani, alcune imposte chiuse e impolverate, segno che quegli appartamenti erano disabitati da tempo. Verso l'incrocio con via Monte Grivola c'era un bar, più in là una farmacia e un negozio di abbigliamento. Dall'altra parte una cartoleria e a una decina di metri una banca. Rocco aguzzò la vista e si avvicinò. Sopra l'ingresso c'era una telecamera. Non sembrava inquadrare la strada, ma era comunque una chance. Entrò. Un paio di persone in fila alla cassa e due impiegati che lavoravano dietro dei box di vetro. Rocco si affacciò nel primo. «Salve. Vicequestore Schiavone. Il direttore, per favore». L'impiegato lo guardò, poi si alzò dalla sedia. «Se ha un momento di pazienza lo chiamo subito» e lasciò il suo posto. Schiavone tornò all'entrata, alle due vetrine che affacciavano sulla via. Riusciva appena a scorgere il marciapiede di fronte al civico di Juana Perez. Neanche un minuto e apparve il direttore, un uomo sulla sessantina con un paio di baffi da messicano, in testa un cespuglio di capelli neri e una nevicata di forfora sulle spalle della giacca di cotone blu scuro. «Dottore! Cosa posso fare per lei?». «La telecamera» disse Rocco. «Quale? Quella in strada?». «Esatto. Mi dica che ha ancora le registrazioni di domenica sera». «Non dipende da me. C'è un istituto di sorveglianza che tiene i dati. Vuole il numero?». «Grazie». Il direttore rientrò nel suo ufficio. Rocco guardò un impiegato e quello, che lo stava osservando, distolse gli occhi e si riconcentrò sulle sue carte. Sgambettando il direttore tornò con un foglietto in mano. «Ecco. Questo è il numero. Vigilpol, si chiama». Il vicequestore prese il foglietto. «Grazie». «Posso fare altro?». «No, lei è stato gentilissimo» e accennando a un sorriso si avviò all'uscita. «Immagino sia per l'omicidio dell'altra sera». «Immagina bene». Il direttore scosse la testa. «Che tristezza». «Vero» rispose Rocco, «e la cosa peggiore è sguazzarci dentro, mi creda». «Cosa le devo dire, Schiavone?». Baldi aveva l'aria sommessa, quasi spaventata. In mano teneva un foglio che non si decideva a mostrare al vicequestore seduto davanti alla scrivania. «Questi sono i tabulati di quel cellulare che mi ha dato. Non c'è un numero». Rocco finse sorpresa. «Mi sta dicendo che quel cellulare non ha ricevuto o fatto chiamate?». «Così sembra. Anzi a essere più precisi ci sono solo due numeri. Il suo e quello della questura». «Sì, io l'ho chiamata martedì mentre cercavo di rintracciarla...». Sospirò. «Dottor Baldi, la cosa è assurda, non trova?». «Porca miseria se è assurda. Qui puzza di bruciato, Schiavone». «Chi può fare una cosa simile secondo lei?». Il magistrato non rispose. Si alzò dalla sedia. A passi lenti se ne andò alla finestra. Scostò la tenda. La luce gli illuminò il viso. Strizzò un poco gli occhi. La foto della moglie era sul tavolo, al solito posto. «Schiavone, qui l'affare si sporca». Aveva abbassato la voce di tre tacche. «Mi dice che ha anche la sensazione che qualcuno l'abbia seguita». «È così». Rimise a posto la tenda e tornò al tavolo. Era pallido. «Che succede, dottor Baldi?». «Facciamo due conti. Svuotano l'appartamento, e nessuno sente niente, cancellano i tabulati telefonici di Juan Perez, la seguono...». «Bisogna sapere qualcosa di più su Berthod, il proprietario di casa, e su Giulia Cosma, Ivano Petrulli, Bernardo Valenti e Diego Fabiani». «Sarebbero i vicini?» chiese Baldi prendendo appunti. «Esatto. Io intanto ho convocato in questura Petrulli, ha sparato un alibi fasullo. La notte dell'omicidio non era dalla fidanzata». «Agli altri ci penso io». «Che succede, Rocco?» chiese Italo. «Succede che qui fra poco scoppia un bel casino». Bussarono alla porta. «Avanti!» urlò Schiavone. Apparve il viso pallido di Michela Gambino. «Mi hai fatto chiamare? Accidenti, questa non è una stanza, è un deposito di scope!». «Vieni Michela, siediti per favore e chiudi la porta». Il tono serio e il viso grigio del vicequestore spensero il sorriso del sostituto della scientifica. «Posso?» chiese a Lupa che saltò giù dalla sedia e andò ad accucciarsi sotto la scrivania. «Che succede?» disse una volta seduta. «Ti chiedo il massimo della riservatezza per quello che sto per dirti». Michela annuì, poi guardò Italo che con un cenno della testa negò di sapere qualcosa. «Chi può cancellare i tabulati telefonici di un'utenza?». «Che stai dicendo, Rocco?». «Ti prego Michela, rispondi alla domanda». Il sostituto si morse le labbra. «Non è una cosa facile. Bisogna infiltrarsi nei computer della compagnia e soprattutto fare in modo che la compagnia non se ne accorga o, se se ne accorge, taccia». «Qualcuno molto in alto?». «Molto in alto» disse Michela. Poi il sorriso tornò sul suo viso. «Ma è successo davvero?». «Sì. Qualcuno non vuole che si sappia quali numeri di telefono ha ricevuto Juana Perez sul suo cellulare». «Oh porca!» e Gambino sbatté le mani eccitata. «Lo vedi Rocco? Quando ti dico che ci controllano? Non vedo l'ora di ridarti la stanza. Io voglio andare nel seminterrato!». «Nel seminterrato?» chiese Rocco. «Esatto. Lì almeno i cellulari non prendono». Italo intanto era diventato bianco come un cencio. «Scusate. Scusi dottor Schiavone» disse. «Sta dicendo che c'è qualcuno interessato a insabbiare l'inchiesta?». «È così». «E noi che facciamo?» chiese Italo. «Noi andiamo avanti. Michela, tu hai qualche novità?». «Ho esaminato i primi tre peli trovati nell'appartamento. Sono di cane, e lo sai? Hanno pigmenti rossastri. Non come quello lì» e severa indicò Lupa che scodinzolò felice. «Ma c'è di più. I miei continuano a lavorare in casa e di peli ne hanno trovati ancora. Solo per esaminarli ci passerò le ferie». «Possono essere tutti miei?» disse Rocco sorridendo. «A meno che non ti sia fatto un maglione con il manto di Lupa... si chiama così, no?». Si avvicinò e infilò le mani nel pelo folto del cane che scambiò quel controllo per una coccola e si gettò a pancia all'aria. «Che stai facendo al mio cane?» le chiese Rocco. Ma Michela non rispose. «No, anche quelli sottopancia sono bianchi e qualcuno nero. Rossastri non ne ha». Si alzò. «Bene Rocco, senza fare prove di dna che costano ai contribuenti e ci vuole una quaresima posso affermare che i peli del cane non erano i tuoi. Cioè, i suoi, voglio dire, non li hai portati tu». «Forse Juana aveva un cane?». Michela alzò le spalle. «Se l'aveva ora è nel fiume da qualche parte... adesso però devo controllare anche gli altri peli. Hanno un colore più scuro, ma credo di non sbagliarmi di molto se dico che risulteranno canini anche quelli lì». «Ma che aveva un canile in casa?» chiese Italo. «Magari è lo stesso cane ed è un bicolore» osservò il vicequestore. «Hai preso in considerazione il gancio sul soffitto?». «È strano. Non c'è presa né filo della corrente, quindi niente lampadario, però è stato usato dopo la riverniciatura dell'appartamento». «Brava Michela, bel lavoro». Italo buttò fuori l'aria che aveva trattenuto. «È proprio un gran casino» disse. «Già. Una cosa la sappiamo. Chi ha cancellato i numeri l'ha fatto prima di martedì sera, quando io stesso ho chiamato Juana Perez. Dunque la cosa è avvenuta intorno al giorno dell'omicidio, la domenica». «E questo che vuol dire?» chiese Michela. «Che chiunque sia stato s'è mosso in maniera rapida e precisa». «Sono stata alla Vigilpol, come mi hai ordinato. Non hanno fatto storie. Questi sono i filmati della telecamera esterna della banca di domenica sera» disse Caterina scuotendo un dischetto argentato. La squadra intera si riunì di fronte al computer mentre il viceispettore aveva già infilato il dvd. «Speriamo...». Si vedeva la strada. In alto sullo schermo la data della registrazione e l'orario. Le diciotto e venti. Qualche passante, una macchina che parcheggia dalla quale scende una donna con due bambini. Rocco, Antonio, Italo, Caterina e Casella stavano con gli occhi puntati sul monitor. «Vai avanti veloce, puoi?» chiese il vicequestore e Caterina obbedì. La luce del sole si affievolisce sulla scena col passare dei minuti. Sfilano corpi rapidi come in una vecchia comica di Ridolini, le ombre si allungano veloci, sul display corrono i minuti. Poi un uomo, fermo sul marciapiede. «Stop!» disse Rocco. Ore 20:00. Un tizio se ne sta impalato a guardare qualcosa di fronte a lui. «Che fa?» chiese Casella, ma nessuno gli rispose. Un africano gli si avvicina. Parlano. Cerca di vendergli qualcosa, ma quello sembra non ascoltare. Prende il portafogli e gli allunga una banconota. L'africano gli consegna un pacchetto, sembrano calzini, e se ne va. L'uomo resta lì, inchiodato a osservare il palazzo di fronte. Poi prende il cellulare. Fa una telefonata. «Strano è strano» disse Antonio. Poi come preso da un raptus l'uomo attraversa la strada e infine sparisce dall'inquadratura. Caterina bloccò il filmato e guardò Rocco. «Sarebbe bello ingrandire l'immagine» disse il vicequestore. «Ci penso io». Antonio prese il posto di Caterina. Tornò indietro, al momento dove l'uomo si era voltato a favore della telecamera per dare i soldi al venditore ambulante. Bloccò l'immagine. L'avvicinò. La luce era poca. A ogni zoomata bisognava aspettare che i pixel si riassestassero. Arrivò al massimo dell'ingrandimento. Il viso era per metà in ombra ma la parte in luce era abbastanza dettagliata da rendere riconoscibile il volto. «Chi sei?» disse Italo. «Possiamo stamparlo?». Antonio neanche rispose. Un paio di clic sul mouse e si alzò. Dalla stampante uscì il viso dell'uomo. «Ecco qui» fece mostrandolo alla squadra. «E non c'è bisogno...» fece Casella, «io questo lo conosco...» si avvicinò a Scipioni e gli strappò il foglio dalle mani. «Dove l'ho visto?». Tutti erano concentrati sull'agente che si era portato una mano al mento, come se quel contatto aiutasse la memoria. «Dove l'ho visto...». «Forza Casella...» lo incitò Schiavone. Poi il poliziotto si mollò una manata in fronte e sorrise. «Sì!» urlò. «Ci ho comprato la caldaia!». «La caldaia?» chiese Italo. «Sì, quella nuova, 800 euro, una Ferroli che mi dicono buonissima». «Perché non hai preso una Vaillant come ti avevo detto?» chiese Italo. «Oh! Ma sticazzi delle caldaie» gridò Rocco. «Dove sta? Chi è?». «Ha il negozio di termoidraulica a via Festaz» rispose l'agente. «Sono le sette. Facciamo ancora in tempo. Daje Italo» e il vicequestore seguito dall'agente Pierron si precipitò fuori dalla stanza. «Questo lo deve decidere lei. Io le consiglio i sanitari del bagno in sospensione, e sa perché? Per la pulizia, diventa tutto più facile». La signora De Belli doveva cambiare solo tazza e bidet a uno dei quattro bagni del suo appartamento di 200 metri quadrati e da ore non riusciva a decidersi. Era il quarto catalogo che Marco le stava mostrando ed era stanco. Ormai si avvicinava l'ora di chiusura. «Un cesso è un cesso» pensava, «prendilo bianco blu o rosso sempre la stessa cosa ci devi fare». Giorgio, il socio, lo guardava dalla scrivania con il sorriso negli occhi. Ci godeva che la De Belli fosse toccata a lui. Era un patto che avevano sancito fin dall'inizio: la De Belli una volta a testa. Due settimane prima se l'era presa Andrea, il terzo socio, che in quel momento se ne stava in palestra a fingere di allenarsi. Poi era stato il turno di Giorgio, che se l'era cavata con una mezz'ora per una semplice manopola, e quella sera toccava a lui. La De Belli, soprannominata al negozio «la signora non saprei», era lì, con gli occhi inchiodati sulla pagina della linea «Gaeta» e storceva la bocca insoddisfatta. «Non saprei... vede, questa forma moderna un po' circolare non mi convince». «Ma signora mia, prima aveva detto che voleva una linea più curva, che non le piacevano gli spigoli, che doveva somigliare vagamente al lavandino di marmo!» e incrociò le dita, fosse mai che la signora decidesse di cambiare anche quello, si entrava nell'orrore puro. I modelli erano migliaia, c'era da arrivarci a ferragosto. «Sì, lo so» disse la cliente, «ma non mi convince. E poi il prezzo?». Marco stirò un sorriso. «La differenza fra questi che le ho mostrato e gli altri si aggira sui 20 euro, più o meno». «Eh, venti euro, li butti via!». Sarebbe stato un pensiero valido per chi si affannava ad arrivare a fine mese, non per la signora De Belli, moglie del notaio, gente che viaggiava a 300.000 euro all'anno. «Allora faccia una cosa signora...» dopo due ore di fotografie, cataloghi e disquisizioni sull'importanza della forma ovale invece di quella quadrata, più avveniristica ma sicuramente scomoda soprattutto nel caso del bidet, Marco aveva preso la decisione di piantarla lì. «Lei ci pensi, ci rifletta, modelli ne ha visti tanti, poi con comodo torna, fa l'ordine e io vedrò di accontentarla». «Non mi ha ancora parlato dei tempi di consegna». «A costo di andare a prenderli di persona in Umbria e caricarmeli sulle spalle, il meno possibile!» le rispose. «Mi creda, si tratta di qualche giorno». «E per le rubinetterie?». Marco sentì un vuoto allo stomaco. Se solo avessero cominciato a guardare i rubinetti per quel bidet ci avrebbe fatto notte. «No signora, io i rubinetti non glieli mostro». «Perché?» fece quella aggiustandosi gli occhiali. «Perché mi sono rotto i coglioni e ho voglia di andare al bar» avrebbe voluto rispondere. Invece le disse con un sorriso: «E come fa a scegliere il rubinetto se non conosciamo ancora il bidet sul quale lo monteremo?». La cosa convinse la signora De Belli. «Ha ragione, certo che ha ragione». Un brivido di soddisfazione passò sul cuore di Marco. Ma durò un attimo. Fuori dalla vetrina, sulla strada, vide un'auto della polizia parcheggiare di fronte al negozio. L'ansia per la signora De Belli era una sorella minore di quella che lo colse alla vista della macchina. Cominciò a sudare, a impallidire. «Si sente bene?» chiese la signora. «Eh? Sì, tutto bene. Tutto a posto» poi guardò il collega ancora indaffarato alla scrivania. «Giorgio, puoi venire?». Quello da lontano fece no col ditino. «Per favore!». Dall'auto scese un poliziotto giovane in divisa accompagnato da uno più vecchio coi capelli spettinati che gettò per terra una sigaretta e si stiracchiò guardando le vetrine del negozio. Erano lì per lui. Piantò la signora De Belli e corse nel magazzino. Il negozio era spazioso e ben illuminato. Una parete a sinistra dedicata all'esposizione di rivestimenti, ceramiche, rubinetterie, un bancone in acciaio e cristallo lungo tutto il muro di fondo, decorato con mattoncini bianchi a vista. A destra due scrivanie, sempre di acciaio e cristallo, attrezzate con computer e circondate da librerie colme di cataloghi. Ai lati dell'entrata splendide stufe di ghisa riempivano le vetrine. Una donna anziana con degli occhiali da vista con la montatura verde era concentrata a guardare delle fotografie sparpagliate sul ripiano del bancone. Un uomo sui 50 con il pizzetto e senza capelli era seduto alla scrivania e li osservava. Si alzò per andare incontro al vicequestore non appena varcarono la soglia. «Prego?» disse fregandosi le mani. «Lei è Marco Rosset?». «No... Marco Rosset è il mio socio, io sono Giorgio Calegari. Posso esservi utile?». «No, se lei non è chi cerchiamo» rispose Rocco. «Dov'è il suo collega?». «Era qui un attimo fa» fece il negoziante voltandosi verso la signora ancora intenta a osservare i dépliant. «Signora De Belli, ha visto Marco?». La donna si tolse gli occhiali e guardò Rocco e Italo strizzando gli occhi. «Era qui proprio adesso. Guardavamo i sanitari, poi è sparito nel retro...» e indicò una porta di legno bianca che si apriva proprio alle spalle del bancone che dominava tutto il negozio. «Ce lo va a chiamare?» disse Rocco. Giorgio Calegari scattò e infilò la porta. «È scappato?» sussurrò Italo a Rocco. «Può essere. Vai a dare un'occhiata fuori». Italo obbedì e tornò in strada. «Anche lei cerca sanitari?» fece la donna con un sorriso invitante. «No signora, polizia». «Questo lo vedo. Ma un poliziotto non usa i sanitari?». «Se è per questo ci sguazzo dalla mattina alla sera. Sono qui per lavoro, signora». Giorgio Calegari si affacciò alla porta «È in bagno... non apre e non risponde...» fece allargando le braccia a chiedere scusa. Rocco lo raggiunse ed entrò nella stanza sul retro. Se il negozio era pulito moderno e ordinato, il magazzino era buio e polveroso. Un locale enorme nel quale correvano scaffalature di metallo assiepate da tubi, raccordi, pompe elettriche, bulloni, rubinetti. In fondo una piccola scala portava alla toilette. Rocco bussò alla porta: «Marco Rosset?». Dall'altra parte non ci fu risposta. «Ma è successo qualcosa?» chiese Giorgio che cominciava a preoccuparsi. «Non lo so, me lo dica lei» gli disse Rocco e bussò ancora. «Rosset, per favore apra!». Si sentì lo sciacquone, poi la porta si aprì. Apparve Marco sorridente e fresco come una rosa. «Mi scusi... è me che cerca?» chiese infilandosi il cellulare in tasca. «No, ero senza fare un cazzo e mi dicevo: perché non passare il tempo a bussare nei cessi di questo negozio?». Marco guardò tranquillo Giorgio. «Vai pure Giorgio... qui ci penso io. Chiudi tu con la signora De Belli, per piacere?». Il socio con lo sguardo teso e poco convinto si allontanò. «Allora, che cosa succede dottor...?». «Vicequestore Schiavone». «Bene, era me che cercava?». «Proprio lei. Parliamo qui?». Marco si guardò intorno. «Nel magazzino no, commissario». «Vicequestore!». «Sì, mi scusi, vicequestore, c'è un bellissimo bar qui di fronte, un caffè è quello di cui ho bisogno, dopo un'ora di dépliant con la signora De Belli penso di meritarmelo» e sorridente fece strada. Il bar aveva carta da parati di finto broccato alle pareti, i tavoli bianchi e le sedie imbottite lo facevano somigliare a un salone da tè che aspirava a Piccadilly ma si fermava dalle parti di Lugano. Profumava di crema e la ragazza che portò i caffè ristretti indossava una divisa nera allietata da un grazioso grembiule bianco. Poggiò il vassoio accanto al cellulare di Rosset e veloce sparì dalla saletta. «Lei sa perché siamo qui?» chiese Rocco. «Immagino per la denuncia del furto che ho subito in casa l'anno scorso?». Rocco guardò Italo che girava il cucchiaino nella tazzina. «Pierron, eri a conoscenza del fatto?». «No» rispose quello. «No, signor Rosset, non è per quello». Gli sfuggiva la somiglianza. C'era qualcosa nella faccia del commerciante che gli ricordava un uccello dal becco fino e lungo, però gli occhi assenti e la bocca piccola e sfuggente lo spiazzavano. E anche la testa, magra, sembrava trovare equilibrio solo in quel naso lungo e appuntito. Fu all'improvviso che capì, guardandogli le orecchie prive di lobo, piccole, quasi nascoste. Il naso non era un becco e Marco Rosset non era un uccello. Era un Myrmecophaga tridactyla anche detto formichiere gigante, il bradipo che vive in America Centrale e del Sud. Si aspettava che succhiasse il caffè infilando una lingua lunga e sottile nella tazzina di porcellana, invece lo sorseggiò come la maggior parte degli esseri umani. «Ah, non è per la denuncia? E allora per cosa?». «Una prostituta che lavorava al 12 di via Brean. Lei domenica sera era lì, sul marciapiede davanti al civico, c'è stato per parecchi minuti, poi è andato a bussare». «Io? Io non sono andato da nessuna prostituta. Figuriamoci...». «C'è il filmato di una banca, signor Rosset». Rocco bevve il caffè. «È corretto?». «Che, il caffè?» chiese il commerciante. «No, dico, è corretto quello che dico?». «No, commissario...». «La prego, sono vicequestore». «Mi scusi, no, vicequestore. Le giuro che non capisco di cosa sta parlando». Rocco si assestò sulla sedia. «Senta, signor Rosset. A me non piace, ma è il mio lavoro. C'è di mezzo un morto». «Oh porca...». «Già. Lei è un cittadino onesto, paga le tasse. Le paga, vero?». «Accidenti se le pago!». «Ecco, e io non ne voglio approfittare». «Le giuro commissario che...». «Vicequestore, e tre» lo corresse Rocco. «Scusi. Le giuro vicequestore che io non so di cosa stiamo parlando». «Si offende se le faccio qualche domanda in più?». «Ma si figuri» e Rosset bevve il caffè. «Può dirmi, sempre se lo ritiene opportuno, cosa ci faceva a via Brean domenica sera?». Il tono suadente, gentile di Rocco stonava alle orecchie di Italo. «Aspettavo un amico che abita da quelle parti». «Che poi non è venuto?». «Già. Allora ho preso e me ne sono andato». Rocco guardò Italo con l'espressione triste. «Accidenti, agente Pierron, non ne caviamo un ragno dal buco». «Cosa ci rimane da fare?». «Niente. Lei, signor Rosset, proprio non ha notato nulla di strano? Qualcuno che è uscito dal portone, che so? Qualcosa che l'ha colpita». Rosset si morse le labbra, poi scosse la testa. «Proprio no, mi dispiace». «Magari ora non lo ricorda, però mi promette che ci pensa?» gli chiese Rocco. «Sa, a volte a noi della polizia basta un dettaglio insignificante, anche un'ombra, e arriviamo a qualche conclusione» poi abbassò la voce. «Non le nascondo che stiamo brancolando nel buio». Rosset aveva la faccia dispiaciuta. Sembrava in colpa per non poter aiutare le forze dell'ordine. «Sì, certo, ci penso...». «Mi lasci il suo cellulare. Per qualsiasi evenienza, la posso chiamare?». «E certo!». Rocco prese una penna e trascrisse il numero che il negoziante gli dettava. «Bene, grazie, lei è stato di una gentilezza squisita!». Rocco si alzò dal tavolino. «Le chiedo di scusarci per l'invasione, e spero di non doverla infastidire più». «Ma si figuri, dovere. Posso sapere come si chiamava?». «Chi? La prostituta?». «Già». «Juana Perez». «Spagnola?». «No, argentina. Tanti saluti, Rosset» e i poliziotti uscirono dal bar. Marco Rosset dentro esultava per la facilità con cui se l'era cavata. Gli parve molto a buon mercato. Anche troppo. «Sembra un poveraccio» fece Italo. «Magari dell'omicidio sapeva tutto, ma se l'è fatta sotto». «Ti sbagli Italo, ci ha mentito». Rocco salì in macchina. «Come, ci ha mentito?» chiese Italo. «Te lo dico io». «Infatti neanche gli hai chiesto il nome dell'amico con cui aveva appuntamento». «Tanto non c'era. Sparava una cazzata, o forse si era già organizzato con qualcuno. Quello sa parecchie cose». Italo guardava dritto davanti a sé. Aspettava che Rocco proseguisse il ragionamento. «Per quale motivo al nostro arrivo s'è andato a chiudere in bagno?». «Un'urgenza?». «Col cellulare? Non credo proprio. Bisogna fare una cosa sporca, Italo, che al giudice Baldi non piacerà, ma qui si sta mettendo male, molto male...». Prese il telefonino e chiamò. «Antonio? Il tuo vicequestore preferito». «Novità?». «Sì. Ora ti do un numero di cellulare. Devi ricontattare il tuo uomo». Rocco tirò fuori il tovagliolino con il numero di Marco Rosset. «Oddio, ancora?». «Sì. Servono i tabulati di questa utenza. Ma mi bastano le chiamate di stasera se non può fare di più». «Va bene, Rocco. Quello vorrà qualcosa in cambio, però». «Tu digli che gli leviamo tutte le multe che becca da qui a Natale». «E lo possiamo fare?». «No, ma tanto la gente non lo sa». «Vabbè, Rocco, capito... senti, ti sta cercando la Gambino. Dice che ha una cosa molto urgente da dirti». «Cioè? Che nei dentifrici che usiamo ci sono particelle infinitesimali radianti che indicano ai sovietici la nostra posizione 24 ore su 24?». Antonio rise. «No, dice che riguarda Juana Perez. Ti aspetta qui in ufficio». Parcheggiarono davanti alla questura. Ormai era scesa la notte. Michela Gambino li attendeva davanti all'entrata a braccia conserte. Guardandosi intorno, come fosse inseguita da qualche nemico misterioso, si avvicinò ai due poliziotti. «Ehilà Rocco, stanco?». «Da morire. Tu che hai da dirmi di tanto urgente?». La donna portò una ciocca di capelli dietro l'orecchio. «I peli». Rocco aveva rimosso. «Ancora?». «Gli altri tre che per ora ho esaminato, non sono peli di cane». «Bensì?». «Juta». Rocco aggrottò le sopracciglia. «Juta?». «Corda di juta. Ecco, magari serve». Il vicequestore annuiva in silenzio. «Ora io mi chiedo» continuò la Gambino, «perché usare una corda di juta in casa? Per metterci un'altalena?». Ma Rocco non rispose e tirò dritto verso l'ufficio. Sulle scale incrociò D'Intino e Deruta. Avevano la faccia stanca e lo sguardo sommesso. «Non ho tempo» gli disse. «Dotto', noi semo cercato in tutte le ambasciate, come ci ha detto lei». «E di questa Juana Perez nessuna notizia» aggiunse Deruta. «Mi sa che il traffico di droga non c'entra niente». Non aveva tempo di mettersi a pensare a un nuovo compito da affidare ai due. Li scansò e seguitò a salire le scale. «Che facciamo mo'?» chiese D'Intino. «Andatevene a casa. Non ho bisogno di voi» gridò sparendo alla vista dei poliziotti. D'Intino guardò Deruta. «Che cascetta!» disse. «Che vuol dire?». «Uno lavora, fatija, e po'? Manco nu grazie». «Vabbè, D'Intino, noi il lavoro lo abbiamo fatto. Io me ne vado a dormire, così domattina all'alba do una mano a mia moglie alla panetteria». «Oh, Deruta, ma io e te serviamo veramente a qualcosa qui dentro?». «D'Inti', se ti devo dire la verità non lo so più...». Appena entrato in questura gli squillò il telefono. Era a pezzi, ma non poteva fermarsi. «Chi scassa?» disse. «Bonanni» rispose una voce, stanca anche quella. «Bonanni?» si passò la mano fra i capelli. «Vice? Spinaceto? Roma?». «Ah, Bona', scusami, momento complicato. Che succede?». «Magari vuoi sapere cosa abbiamo scoperto sul cadavere anonimo di Castel di Decima. Vedi se può aiutare». «E dimmi». «Forse ricordi che aveva una cinta di pelle. Con un marchio strano. Siamo risaliti al marchio». «Mbè?». «Il fabbricante è Mario's Pura Piel, una fabbrica importante che sta in Honduras». «Honduras?». «Già. Abbiamo chiamato. Loro le cinte non le producono. Solo su richiesta. E anni fa hanno fornito alcune ambasciate, per esempio quella italiana. Insomma, un cadeau, e visto che stai ad Aosta il francese mi pare d'obbligo». «Hounduras...». «Honduras, Schiavo'... che c'è?». «Niente, Bonanni. Ci devo riflettere. Se mi viene in mente qualcosa ti richiamo». «Buon lavoro» e attaccò. Honduras. Il cervello di Rocco si mise a correre su più binari. Juana Perez era argentina. Ma poi cosa c'entrava col cadavere misterioso che portava il suo numero di cellulare in tasca? Honduras. Si fermò sulle scale. Non ricambiò il saluto di un agente che scendeva veloce. Honduras. Scosse la testa. Prima o poi gli sarebbe venuto in mente. Era arrivato alla porta dell'ufficio. L'aprì. Il lemure, al secolo Ivano Petrulli, era venuto senza l'avvocato minacciato da suo padre. Osservava l'ufficio di Rocco, stupito che qualcuno potesse lavorarci, figurarsi un vicequestore. Rocco lo guardava in silenzio. Fuori era già buio, e la luce sparata della stanza ingrandiva ancora di più gli occhi del ragazzo. «Hai mentito» fu la prima cosa che disse Schiavone. Ivano teneva le dita intrecciate poggiate sulle ginocchia. «Dov'eri la sera dell'omicidio se non stavi dalla tua fidanzata?». Il ragazzo tirò su il mento, sospirò. «Come vede sono senza l'avvocato, la questione è delicata». «Più di un omicidio?». Rocco aprì la finestra direttamente dalla poltrona. Poi si accese una sigaretta. «Dispiace?». «No. Faccia con comodo, è casa sua» rispose il lemure. «Allora?». «Non ho una ragazza» disse. Rocco aspettò che continuasse. «Ho una storia, ma non con una ragazza». «Capito. Allora eri con la persona con cui hai una storia?». Ivano sorrise. «Mio padre è un uomo diciamo molto attaccato alla famiglia, all'onore, al lavoro eccetera eccetera. Io a casa non resistevo più, ho affittato quel piccolo appartamento a via Brean perché costa poco e posso viverci la mia vita, lontano dai miei, dai loro amici e dalle loro cene. Non è facile...». «Immagino». «No, non può immaginarlo. Raimonda, la mia amica, mi protegge, è la mia scusa da almeno tre anni. Mi dispiace se l'ho messa nei guai con questa storia, ma mi ero spaventato. Me ne dà una?» e indicò la sigaretta. Rocco allungò il pacchetto insieme all'accendino. Ivano l'accese. «Tanto prima o poi sarebbe venuto fuori. Era anche arrivato il momento di dirlo. Comincio ad avere un'età. Non è tanto per mio padre, di lui e dei suoi progetti sul mio futuro non me ne frega niente. Penso a mia madre, che fa quella che capisce tutto, sempre pronta con una buona parola. Non appena lo saprà...». «Senti Ivano, non voglio passare per un insensibile o uno dal cuore di pietra, ma a me delle tue abitudini sessuali non me ne frega una mazza. Io voglio solo sapere se domenica sera eri a casa tua con la persona che ami oppure no». «No. Non ero a casa mia». «Oh, bene. E questo fa uno. Ora dimmi» spense la cicca nel portacenere, «c'è modo di sapere dove ti trovavi?». «Se lo dico però metto un'altra persona nei casini». «Se non me lo dici nei casini ci stai tu». «E va bene. Ero a via Aubert». «Che è una strada non lunghissima ma credo ci vivano almeno un trecento persone. Puoi essere più preciso?». Ivano si grattò la testa. Si portò la sigaretta alla bocca. Tremava. «Ero da Diego». «Diego Fabiani? Il commercialista?». «Esatto». «Il tuo vicino di casa?». «Esatto». «Perché eravate nello studio e non in casa dal momento che abitate uno sotto all'altro?». «Mi aveva chiamato, verso le otto, dicendomi di passare a trovarlo lì. Io tornavo dal tennis e ci sono andato». «Alle otto?». «Sono arrivato verso le otto e mezza. Poi sa com'è?». «No, com'è?». «Una frase, un gesto... e insomma siamo rimasti lì e poi a cena fuori». Rocco incrociò le braccia davanti al petto. «Visto che siamo in tema di sincerità, anche io voglio essere sincero con te. Nel palazzo c'è stato un trasloco, tu non ti sei accorto di niente?». «Giuro di no». «I tuoi vicini, a parte Diego ovviamente, li conosci bene?». «Poco. La signora è affabile, gentile, mi saluta. L'altro, quel...». «Bernardo Valenti». «Sì, ogni tanto lo vedo mentre porta a spasso il cane, oppure rientra dalla spesa. Ci ho parlato qualche volta. Mi pare facesse il ragioniere, ma è in pensione». «A proposito di cane, un piccolo dettaglio che mi sovviene proprio ora. Sai se Juana Perez ne avesse uno?». «No, nel modo più assoluto». «Ne sei sicuro?». «Sicuro. I cani abbaiano, vanno portati fuori, lei poi amava i gatti. Lo sa? Quando andavo a prendere la macchina sul retro trovavo del cibo in piccole scodelline sulla recinzione del giardinetto. E ce n'era sempre qualcuno che gironzolava lì intorno». «Grazie. Torniamo ai tuoi condomini. Del tuo padrone di casa che mi dici?». «Chi, Berthod? Uno stronzo» e schiacciò la sigaretta nel posacenere. «Me l'ha presentato Diego. Lui mi ha trovato la casa. Non possiamo andare a vivere insieme, e allora abbiamo preferito diventare almeno vicini». «Quant'è che conosci Diego?». «È arrivato ad Aosta due anni fa e ha aperto quello studio. L'ho conosciuto in palestra». «Cos'altro sai di lui?». «Che viene dal Monferrato. La sua famiglia è lì». «Lavora molto?». «Sì». «Ci vai spesso nel suo studio?». «La verità?». «Ti prego». «Quella sera sarà stata la terza volta in due anni. C'è una segretaria, farmi vedere troppo non è bene. Si vive un po' nascosti, almeno in una città piccola come questa». «Ti sei mai sentito preso per il culo?». «Non capisco...». «Da Diego, intendo. Ti ha mai presentato ad altri amici, o ai familiari...». «No, mai. Frequenta i miei amici, lui non ne ha molti, credo. Clienti, quelli sì». «Ivano, da me non uscirà una parola. Ti prego di fare lo stesso». Il ragazzo annuì. «Intendo, noi questo colloquio non l'abbiamo mai avuto. Tu non mi hai mai parlato della tua relazione, e l'alibi di Raimonda per noi della polizia tiene». «Perché?». «Perché sei in mezzo a una cosa più grande di te. E più grande di me. Quindi meno parli e meglio è». Si avvicinò guardandolo negli occhi. «Ivano, sto parlando della tua incolumità». Attraversava i corridoi deserti della morgue. Le luci al neon esaltavano lo squallore dei muri dipinti di bianco e verdino. Le porte di legno a doppio battente erano color crema, bisognose di una ritinteggiatura. Finalmente raggiunse quella dello studio di Fumagalli. La aprì senza bussare. Il patologo era seduto ad una scrivania, sfogliava un libro alla luce di una lampada da tavolo. La stanza era poco più di tre metri quadrati, affogata nei libri e nelle carte. Alberto alzò lo sguardo e si tolse gli occhiali. «A bussare ti si rovinano le nocchine?». «Ci hai pensato?» disse Rocco. «Ci ho pensato e mi sa che hai ragione». Allungò a Rocco il libro che stava leggendo. «Questa può essere... i due segni alle caviglie, la juta, lo strangolamento. Non hai tutti i torti». Rocco prese il libro. C'era una fotografia. Una donna legata come un salame. «Shibari o kinbaku» attaccò Alberto. «È una forma d'arte giapponese. In pratica si lega una persona con dei nodi e degli intrecci di corde molto complessi, diciamo così, che a detta dei praticanti hanno un intenso significato estetico. Se a uno piace, va da sé». Rocco scorreva le immagini. Donne legate alle caviglie, ai polsi, corde che correvano lungo tutto il corpo con una quantità di nodi inverosimile. «Io continuo a pensare che le Stanze di Raffaello o le teste di Brancusi siano meglio, ma sono uno all'antica Rocco, lo sai. Dunque, qualcuno la considera una performance, ma per molti è soprattutto una pratica sadomasochista. Io credo che Juana Perez ci stava in mezzo». «Cioè si è fatta legare...». «Per volontà del cliente, è chiaro. Solo che, guarda l'immagine a pagina 43» suggerì Alberto. Rocco sfogliò il libro. Una ragazza appesa al soffitto. La corda le teneva i polsi stretti dietro alla schiena, poi proseguiva in mille giri intorno al corpo passando anche sotto la gola. Fra le caviglie una sbarra per divaricare le gambe. «Ecco, vedi? Una cosa del genere...» fece Alberto. «E si usano corde di juta?». «Anche. Almeno così dice questo libro. E così ho visto su internet. Sai Rocco, io questa roba non la pratico. L'unico legame che ho è alla tradizione: io nudo, lei nuda, e via si tromba. Invece funziona che il maestro di nodi, si chiama nawashi, lega la vittima. Nel caso di Juana Perez a scopi sessuali, altrimenti resta inspiegabile il motivo della divaricazione delle gambe. Serve a una eventuale penetrazione una volta che il corpo immobile e inoffensivo penzola alla tua mercé dal soffitto. Ci sono un'infinità di nodi. È una tecnica che risale ai tempi dei samurai, quando non avevano le celle per tenerci dentro un prigioniero e allora lo impacchettavano in quel modo». «E quella c'è rimasta strangolata. È così?». «Credo proprio di sì. Non è la prima volta che lo sento». «E si spiega l'uso del gancio al soffitto. Un incidente?». Alberto annuì e riprese il libro che Rocco gli stava restituendo. «E lo sai? Dalle analisi del sangue, mi è venuto fuori un antistaminico, l'ha dovuto prendere non molto tempo prima di morire». Rocco chiuse gli occhi, come a richiamare alla memoria un'immagine, una parola. Li riaprì soddisfatto. «Se ti dico Oxi, che ti viene in mente?». «Oxi... e poi?». «Era un frammento di una scatola di medicinali che la Gambino ha trovato in casa di Juana». «Oxi... oxi...» poi sorrise. «Oximix 3. Antistaminico che si usa... si usa... aspetta» spostò la sedia e prese un prontuario. Cominciò a sfogliarlo umettandosi l'indice. «Ti prego!». «Cosa?». «Non ti leccare le dita mentre sfogli le pagine, è una cosa che mi fa vomitare». «Cioè, siamo dentro un obitorio e a te viene da vomitare per questa cosa?». «È una vecchia storia. Per favore non farlo». «Ecco qui. Trovato! Oximix 3, si usa per le allergie». «Anche al pelo dei cani?». «Anche, ma non solo. Perché, aveva un cane?». «Lei no». Fumagalli posò il prontuario. «Ora però non capisco perché l'assassino abbia gettato il cadavere nel fiume se poi lui o qualcun altro ha ripulito l'appartamento». «Panico, sapeva che la cosa non doveva uscire fuori. Buttandolo nella Dora avrebbe preso tempo, aveva bisogno di tempo, Alberto. Ha degli alleati, il pezzo di merda». Alberto si alzò. «Che stai dicendo?». «Sto dicendo che dopo lo hanno aiutato. Hanno liberato la casa della poveraccia, fatto sparire la sua auto, cercato di insabbiare tutto cancellando pure i tabulati del suo telefono. Credo che abbiano anche parlato coi vicini». «Diobono...» mormorò Alberto. «Non mi stupirei se a quest'ora l'assassino fosse già molto lontano da Aosta. Abbiamo fatto tardi, cazzo!». Le tre di notte. Tutto era immobile, non tirava un alito di vento ma l'aria era fredda. Solo qualche lampione illuminava via Aubert. Negozi serrati, finestre spente. La luna, se c'era, se ne stava dietro le nuvole grigie. L'ombra di Rocco svoltò un angolo di strada avvicinandosi a passi veloci a un palazzo. Raggiunse un'altra ombra nascosta nel buio, Italo. «Vieni con me». Girarono in via Torre del Lebbroso e si trovarono alle spalle del palazzo di via Aubert. In quella strada invece tirava un vento leggero che penetrava i vestiti e faceva dimenticare di essere ormai in pieno giugno. Rocco guardò in alto. «Fammi la scaletta» disse, Italo obbedì. Gli montò sopra, coi piedi sulle spalle fino ad aggrapparsi a un poggiolo del primo piano. Con uno sforzo si tirò su. Poi salì coi piedi sulla ringhiera, afferrò il dente del balconcino più in alto e finalmente lo raggiunse. Una volta al sicuro si affacciò dal secondo piano. Coi gesti indicò a Pierron di rifare il giro e tornare sulla via principale. Quello obbedì. Si chinò. In mano aveva il coltellino svizzero mille usi. Scelse una seghetta metallica e cominciò a lavorare sulla finestra. Lontano il rombo di un camion ruppe il silenzio della notte. Ci mise meno di un minuto e la aprì. Una volta dentro l'appartamento raggiunse a tentoni la porta d'ingresso, afferrò il citofono e aprì il portone, lo stesso fece con la porta di casa girando la manopola della blindata. Prese il cellulare e accese la torcia. L'ufficio era ordinato, profumava ancora di vernice. Pochi secondi dopo l'agente apparve sull'uscio. Entrò e veloce richiuse. «Ci avranno visto?» la luce della pila inquadrò il viso spaventato di Italo. «Io me la faccio addosso». Rocco gli fece cenno di tacere. Passarono nello studio principale di Diego Fabiani. Una scrivania e una grande libreria piena di faldoni. Qualche attestato alla parete, una fotocopiatrice, il monitor di un computer spento. «Almeno mi dici che cerchiamo?». «Chiudi gli scuri» gli ordinò. Italo eseguì. Finalmente poté accendere la luce. «Prima o poi ci beccheranno» disse l'agente. Rocco aveva afferrato il primo faldone a portata di mano. «Speriamo di no». Lo aprì. Sorrise. «Che hai trovato?». Mostrò il contenuto della cartella. Quotidiani, riviste, buste di plastica. Lo stesso negli altri faldoni. «Carta straccia, Italo. Niente di niente. Ci siamo». «Ci siamo cosa?». «Diego Fabiani non è un commercialista, sempre che Diego sia il suo vero nome». «E chi è?». «E che cazzo ne so?» e indicò il computer. Italo si avvicinò e lo accese. Attesero. «C'è una password» disse Italo. «Prova con Ivano». Italo eseguì, ma non successe niente. «Diego?». Nessun risultato. «Come facciamo? Qui ci vorrebbe uno smanettone». «Stacca la spina e spegnilo. Poi riattaccala, cerchiamo di non lasciare tracce. Ce ne andiamo. Tu dalla porta principale, io faccio la stessa strada». «E così lascerai la finestra aperta!». «Meglio che la porta blindata senza chiusura, credimi. Ci vediamo giù...». Non era riuscito a fermare il cervello per quello che restava della notte e solo la luce dell'alba riportò i pensieri alle loro giuste proporzioni. Si alzò dal letto, tanto ormai non c'era più niente da fare. Diede la pappa a Lupa, poi insieme uscirono di casa. Gli squillò il telefono. Tutto si sarebbe aspettato, tranne di ricevere una telefonata a quell'ora da Antonio Scipioni. «Sono davanti alla questura, Rocco. Bisogna che ci vediamo». «Vienimi incontro» gli disse camminando. «Che succede?». «Succede che la situazione è brutta assai. Ma non ne voglio parlare al telefono». Affrettò il passo. Lo stomaco gorgogliava, desiderava il caffè. Il cervello invece sembrava bloccato, quello richiedeva la preghiera mattutina. La voce di Antonio gli aveva messo i brividi. Lo vide arrivare quasi di corsa, in borghese, si guardava intorno come se avesse paura di essere spiato. Raggiunse Rocco col fiato che colorava l'aria. «Rocco, vieni» e lo prese sottobraccio portandolo verso un baretto. Dentro l'aria era più calda. Al bancone due uomini in tuta da lavoro facevano colazione. Ma Antonio non si avvicinò per ordinare. «Allora qui le cose sono brutte, parecchio brutte. Il mio amico è risalito ai contatti del cellulare di Marco Rosset. È andato parecchio indietro». «E?». «Prima di tutto ti dico che lui alle otto di domenica sera ha telefonato a Juana. Poi la cosa più strana è la chiamata fatta all'ora che mi hai detto, quando era chiuso nel cesso del magazzino per capirci... era un cellulare. E indovina a chi appartiene?». «A chi?». «A Diego Fabiani». Rocco se ne andò al bancone. Ci appoggiò i gomiti. Guardò il barista. «Due caffè, per favore». Antonio si avvicinò. «E non è finita. Nei giorni precedenti ha ricevuto per due volte una telefonata dallo stesso numero. Un ufficio». «Che ufficio, Anto'?». «Prefettura di Biella...». Rocco chiuse gli occhi e quello che vide metteva paura. «La prefettura...» disse a bassa voce. «... hai capito, Antonio?». Il barista depose i due caffè. «Qualcosa, e quello che ho capito mi fa paura. Ma chi è 'sto Diego Fabiani? Perché quello chiama un commercialista mentre è chiuso nel bagno?». «Fabiani non è un commercialista». Scolò il caffè in un solo sorso. «Fabiani gioca su un altro tavolo. E quello che è peggio, Antonio, è che a quel tavolo noi non siamo invitati». «Fabiani è...?». «La butto là?» fece Rocco. Poi lo guardò serio. «Servizi» mormorò. Poi afferrò il cellulare e componendo un numero uscì dal bar. Antonio e Lupa lo seguirono. «Dottore? La sveglio?». «Ma che mi sveglia Schiavone, non ho chiuso occhio tutta la notte» rispose Baldi. «Sto facendo il caffè con la moka. L'avrei chiamata subito». «Dobbiamo vederci, dottor Baldi». «Lo penso anche io» disse il magistrato. «Non in tribunale». «Mi dica lei». «Ci vediamo fra due minuti...» sentì un rumore di stoviglie. «Porca... bruciato... vabbè, allora ci vediamo fra cinque minuti davanti alla cattedrale, al criptoportico». «Sono già sulla strada». «Senta, è vicino a un bar?». «Ne sono uscito adesso». «Mi prenda un caffè, per favore...». «D'accordo... senta, sono con un mio uomo». Ci fu un breve silenzio. «È fidato, Schiavone?». «Ci metto la mano sul fuoco». «Non parliamo di fuoco, Schiavone, ho appena incendiato la macchinetta del caffè. A fra poco». Attendevano Baldi sul sagrato della chiesa. Una luce diafana soffocata dalle nuvole toglieva corpo agli oggetti, come al crepuscolo dopo che il sole è tramontato. Invece non erano neanche le sette del mattino. Lo videro arrivare trotterellando. Anche Baldi si guardava le spalle, come sospettasse di essere seguito. Appena li raggiunse, Rocco gli allungò il caffè che il magistrato finì in un sorso. Poi accartocciò il bicchierino e se lo mise in tasca. «Buongiorno a tutti». «Agente Antonio Scipioni, dottore». «Perché qui?» domandò Schiavone al magistrato. «Lo spazio è grande e non ci sente nessuno». Un brivido di freddo corse nelle ossa dell'agente Antonio Scipioni che aprì bocca ma non disse nulla. Si limitò a guardarsi intorno con le mani in tasca. «Comincio io, dottore. Allora, abbiamo una relazione intima fra Ivano Petrulli e Diego Fabiani, entrambi inquilini del palazzo. Diego Fabiani non è un commercialista come asserisce». «E che cos'è?». «Adesso ci arrivo. La cosa più interessante, dottore, è uno dei clienti di Juana Perez, si chiama Marco Rosset. L'abbiamo interrogato ma nega di sapere qualcosa. Siamo risaliti alle telefonate del tipo e abbiamo appurato che ha chiamato proprio Diego Fabiani, esattamente quando io e l'agente Pierron lo siamo andati a cercare, e in più ha ricevuto due telefonate da un ufficio e qui la cosa si fa pesante, dottor Baldi. Questo ufficio è la prefettura di Biella». «Oh cazzo» uscì al magistrato. Rocco proseguì: «È ovvio che l'assassino è un uomo protetto, uno che deve restare fuori dalle indagini». «Come ha fatto a venire a conoscenza dei tabulati telefonici di questo Rosset?». Rocco non rispose. Il magistrato scuotendo il capo proseguì: «Va bene, non lo voglio sapere, Schiavone, non ho sentito niente. Questo assassino secondo lei era un cliente abituale?». «Secondo me sì. Perché sono convinto che si intrattenessero con dei giochetti complessi. Shibari. Sa cos'è?». Antonio alzò la mano come a scuola. «Quella cosa giapponese che si legano con la corda?». «Bravo. Quella cosa giapponese. Non è una pratica molto comune, il nostro cliente doveva fidarsi di Juana, sicuro l'avevano già fatto». Baldi si portò la mano al mento. Stava pensando. «Ora tocca a me, Schiavone. Lo sa perché non ho dormito tutta la notte?». «No». «È semplice. Ci sono tre cose che mi spaventano e non mi tornano. La prima, ho fatto un po' di indagini sul proprietario della palazzina, quel Berthod». «Sì, quello che minacciava di avere amicizie molto in alto». «Ce le ha, dottor Schiavone. Suo fratello è il generale Berthod, l'ha mai sentito?». «Macché». «Io sì. Lavora a Roma, a via Giovanni Lanza. Lei sa cosa c'è a via Giovanni Lanza?». «L'Aisi» fece Rocco con un fiato. «Cos'è?» chiese Antonio. «Agenzia informazioni e sicurezza interna» rispose Baldi. Antonio deglutì. «E mica finisce qui, Schiavone. Secondo punto. La vedova Cosma, quella del secondo piano, suo marito era un carabiniere. Antonio Cosma, si chiamava. Anche lui lavorava per il Sisde... è morto nel 2006, prima che il Sisde fosse sostituito dall'Aisi». «Quella è una casa protetta...» concluse Rocco. «Esatto. Come diavolo ci sia capitata Juana Perez non lo sappiamo, ma credo che...». «Magari a qualcuno faceva comodo» azzardò Rocco. «Cosa intende?». «Intendo che se in una casa protetta come quella ci abita qualcuno di cui non dobbiamo sapere niente, forse quel qualcuno aveva delle necessità e Berthod non si è opposto». «È una cosa azzardata. Quant'è che abitava lì?». «Due mesi, mi pare di ricordare». «Allora lei...». «Mi viene in mente un tipo con un cane, che parla con un buon accento di qui, e che però viene da Avellino». «Bernardo Valenti?» fece Antonio. «Bernardo Valenti» confermò il vicequestore. «E questa è la terza cosa che non torna» intervenne Baldi. «Lui è un'ombra» disse preoccupato. «Lavorava alla Regione, così dicono le carte, solo che alla Regione non c'è traccia di Bernardo Valenti». «Chi cazzo è allora?». «Non lo so. Cosa vuole fare?». «Mi faccia avere un mandato per questo Valenti, e vediamo di alzare un bel casino. Intanto io lo vado a trovare. Ammesso che sia ancora ad Aosta». I tre uomini si guardarono. «Corriamo un bel rischio...» obiettò Baldi. «Lei si tira indietro?» gli chiese il vicequestore. «Io? Ma che dice, Schiavone». In una macchina Rocco e Caterina, nell'altra Italo e Antonio, frenarono davanti al civico 12 di via Brean. Scesero solo l'ispettore e il vicequestore. Due passanti si fermarono a guardare la scena. Schiavone citofonò a Valenti. Nessuno rispose, allora premette il pulsante della signora Cosma, quella invece era in casa. «Polizia signora, apra!». «Chi mi assicura che...». «Non rompa i coglioni e apra 'sto cazzo di portone!» urlò Rocco. Forse spaventata, forse aveva riconosciuto la voce, la signora aprì. Rocco e Caterina arrivarono di fronte all'interno del signor Valenti. Bussarono. Il suono del campanello echeggiava nell'appartamento. «Valenti apra!» gridò Rocco bussando alla porta. «Apra!». «Questo non c'è» sbuffò Caterina. «Dottore? Valenti è andato a Torino, dal figlio» la signora Cosma si era affacciata sul pianerottolo. «Sto cazzo» rispose Rocco che continuava a tempestare la porta di pugni. «Tre giorni fa...» aggiunse la vicina. Ma Schiavone neanche si voltò a guardarla. Sbuffò e prese il suo coltellino. Si mise a lavorare alla serratura. «Che sta facendo?». «Signora, torni dentro» scattò Caterina, «non la riguarda» e si parò col corpo fra la vecchia e il suo superiore che armeggiava accanto alla porta. «Perché non mi credete?». «Signora, per favore!» esplose Caterina. Quella rientrò in casa come gli occhi della chiocciola al tocco di un bambino curioso. Finalmente la porta si aprì. Rocco fu il primo ad entrare. I pochi mobili erano ancora lì. Ma erano stati svuotati. Non c'era più una fotografia, non c'era più il cuscino di Anubis e neanche il tappeto liso al centro del saloncino. In cucina le credenze erano vuote, il frigorifero aperto. Solo stoviglie, piatti e bicchieri. In camera da letto l'armadio era stato liberato, come i comodini e il settimino sotto la finestra. Rocco si affacciò. In strada i curiosi erano diventati una decina. Italo e Antonio erano accanto alla macchina. «Italo!» gridò Rocco. «Dica?» fece quello alzando la testa. «Chiama la nostra esperta di peli». «Non ho capito». «La Gambino. Fai venire la squadra, subito!». «Porca miseria...» imprecò Caterina guardandosi intorno. «E uno quando va a trovare il figlio svuota tutto l'appartamento?». «Siamo arrivati tardi, io lo sapevo! Lo sapevo, cazzo!» mollò un calcio alla poltrona dove qualche giorno prima s'era seduto a chiacchierare con Valenti. «Dovevamo essere più veloci. Invece niente. Ecco il risultato. Il figlio di puttana è sparito». «Non lo possiamo rintracciare?». «E chi? Bernardo Valenti che non è neppure il suo nome?». «Perché era così protetto?». «Non lo so, Caterina, non lo so... Ora vieni con me!» e come una furia uscì dall'appartamento. In strada era già arrivato il furgone della scientifica. Italo e Antonio lo osservavano in silenzio. «E che so già qui?» chiese Rocco. Italo alzò le spalle, non sapeva cosa rispondere. Antonio allargò le braccia. «Infatti, l'abbiamo chiamati neanche tre minuti fa...». Dal furgone scesero due agenti che Rocco non aveva mai visto. Alle sue spalle da una Lancia Delta appena parcheggiata spuntò Farinelli, il sostituto di Torino. «Uelà, Schiavone, ci si rivede!». «Com'è che stai qui?». «Ispezione alla casa di Juana Perez. Evidentemente non siete ancora in grado di cavarvela da soli». Rocco lo guardava storto. «La Gambino lo sa?». «E chi è la Gambino? Ti vedo furioso, che succede?». I due agenti intanto avevano indossato le tute bianche e afferrato un paio di valigette. I curiosi erano diventati una ventina. «Farinelli, spiegami perché sei qui» disse Rocco con la voce dura. «È semplice. Dobbiamo dare un'occhiata all'appartamento e cercare nuove tracce. Evidentemente il tuo questore ne vuole sapere di più e ha chiamato i migliori. Ora se mi permetti, lasciaci lavorare!». «Farine', io lo so già chi è l'assassino!». «Meglio, così ti do la certezza scientifica. Con permesso...» superò Rocco ed entrò nel palazzo. Rocco guardò i suoi agenti con gli occhi spalancati. «Questa me la devono spiegare» ma lo aveva detto a se stesso. Era un momento di calma, clienti non ce n'erano. Giorgio era andato a fare un sopralluogo in un cantiere a Châtillon, Andrea invece aveva preso la giornata libera. Marco Rosset leggeva il giornale quando all'improvviso alle sue spalle si aprì la porta che dava sul magazzino. «Ma che...». Apparve Rocco Schiavone seguito da una poliziotta. «Le presento il viceispettore Rispoli. Allora, come sta?». «Bene... è entrato dal magazzino?». «Così stavolta non si va a chiudere nel cesso a chiamare gli amici, Rosset». Il commerciante si guardava intorno come una preda braccata. «Rosset, io so tutto. So che lei è andato da Juana Perez e so che ha visto qualcosa e io ora voglio sapere cosa». «Niente!». «Mi vuole seguire in questura?». «Con il mio avvocato!». «Lo chiamerà da lì. Metta il cartello chiuso per ferie, perché la questione è lunga, e ci segua». «Le conviene» aggiunse Caterina. «Io non c'entro niente, porca miseria!» urlò il negoziante. «E allora se non c'entra niente, parli». «Non ho niente da dirle». «Caterina, mentre io porto Rosset in commissariato, fammi una cortesia. Va' a casa sua, avverti la moglie che suo marito è invischiato nell'omicidio di un trans e che presto però glielo rimandiamo all'ovile». Caterina abbozzò un sorriso e si avviò verso l'ingresso del negozio. «Dove va?» fece Marco. «A casa sua. Aspetti... l'indirizzo ce l'hai, Cateri'?». «Frazione Gioannet». «Aspetti, aspetti, dove sta andando?» gridò Marco Rosset. «Gliel'ho appena detto, va da sua moglie. Preferisce che la chiamiamo al telefono? Sa, di persona è più tranquillizzante, una telefonata è sempre un po' fredda. Invece il viceispettore queste cose le sa gestire bene, giusto?». «Giusto» rispose Caterina. «Ho molto tatto». «Gliel'ho detto, non so niente!» la fronte del negoziante si era imperlata di sudore. «Se non sa niente di che si preoccupa?». «Se lei va da mia moglie io...». «Fa una figura di merda, lo so. Ma tanto o prima o dopo che abbiamo parlato col suo avvocato, la figura di merda è inevitabile». Rosset scoppiò a piangere. «È un incubo, è un incubo» diceva fra i singhiozzi. «Io non ho fatto niente». «Rosset, le sto offrendo una via di uscita. Gli altri, chiunque siano, non saranno così gentili». «E avevo pure cambiato idea. Perché sono entrato, perché?». Rocco scambiò uno sguardo d'intesa con Caterina che si avvicinò. «Forza Rosset, non è niente di grave. È una cosa che può capitare. Insomma, cercare un po' di compagnia è umano». «Io non volevo. Proprio non volevo. Perché? Perché?» e con il fazzoletto di carta che il viceispettore gli aveva allungato si asciugò il naso. «Ero appena entrato. Al primo piano. Sono arrivato alla porta. Era semichiusa. Non volevo entrare. E allora ho guardato dalla finestra che dà sulle scale. Era lì, legata, penzolava dal soffitto come... come un prosciutto. Bianca come un lenzuolo. Le cadeva la bava dalla bocca... e gli occhi... gli occhi spenti... Dio, non lo scorderò finché campo». «Poi cos'ha visto?». «Dalla porta è uscito un cane. Piccolo, un volpino, e s'è messo ad abbaiare. Io sono scappato. Ho camminato per strada, avevo paura, non sapevo che fare. Volevo andare alla polizia, giuro che ci ho pensato». «Se l'avesse fatto... vada avanti». «Sono rientrato a casa e mi sono messo a letto. Ma non riuscivo a dormire. La telefonata è arrivata alle sei». «La telefonata di chi?». «Non l'ha detto subito. Ha detto che ci dovevamo incontrare. Che per me sarebbe stato meglio. Io ho avuto paura. Non sapevo chi fosse, ma quello mi ha detto che era delle forze dell'ordine e non avevo nulla da temere. E così il giorno dopo sono andato». «Dove?». «A Biella. Alla prefettura. Fuori dalla prefettura. L'uomo mi aspettava lì». «L'uomo chi?». «Quello che mi aveva chiamato. Non si è presentato. Mi ha dato solo un numero di cellulare. E mi ha detto che se avessi avuto problemi con...». «Con?». «Con la questura di qui, avrei dovuto comporre quel numero. È stato gentile, ma mi ha minacciato che se sentiva una sola parola le cose potevano prendere una brutta piega. E adesso io ho paura. Lo capisce? Che gente è quella?». Rocco si allontanò dal bancone lasciando Rosset nella sua angoscia. Sembrava misurare a passi il negozio. «Tutto per una stronzata» mormorò, ma nessuno lo capì. «Lei mi ha detto la verità, Rosset?». «Lo giuro. Non so come uscirne, non sapevo a chi chiedere aiuto!». «Avrebbe dovuto pensarci quella domenica sera, invece di farsela sotto per la paura che sua moglie avrebbe scoperto 'sta cazzata, doveva venire da noi e tutto sarebbe stato più semplice. Invece è stato zitto, per proteggere le sue piccole sicurezze. E guardi che bordello ha combinato. Se ne vada a casa, si butti malato. Non so che altro consiglio darle, perché anch'io a questo punto non so più cosa fare. Le useremo la cortesia di uscire dal retro, anche se ormai sapranno già che sono venuto a cercarla. Andiamo Caterina, qui non c'è altro per noi». E si avviarono verso la porta del magazzino. «Mi lasciate così?». «Si sentirebbe più al sicuro in questura?». «Sì». «Mica si può trasferire lì». «Adesso ho troppa paura. La prego, non mi lasci solo». Rocco sbuffò. «Cateri', portalo in ufficio». «E lei dove va?» chiese il viceispettore. «Una visitina a Fabiani. Lo porto dentro per i capelli, il pezzo di merda» e uscì. Svoltò in via Trottechien e gettò un'occhiata al teatro. Le locandine annunciavano lo spettacolo di un comico che Rocco non aveva mai sentito nominare. Sui vetri dell'edificio vide una figura riflessa. Un uomo fermo sul marciapiede, si stava accendendo una sigaretta e lo guardava. Rocco si bloccò in mezzo alla strada e si voltò. Aveva una cinquantina d'anni, i capelli bianchi, un vestito di cotone blu. Gettò il fiammifero a terra e gli sorrise. Poi si mosse per andargli incontro. Il viso abbronzato e il naso schiacciato. «Salve Schiavone, come va?». «Va. A lei?». «A me benissimo. Ha cinque minuti?». Portava la fede all'anulare. All'indice un anello di argento intrecciato. Sotto la giacca un leggero rigonfiamento, proprio sopra il cuore. «Dipende» rispose. «Solo due parole, potrà trovare cinque minuti della sua giornata da dedicarmi». Non riusciva a capire l'inflessione dialettale. Sembrava uno speaker della radio dalla dizione perfetta e la voce impostata. «Di solito non parlo con gli sconosciuti. Ce l'ha un nome?». «Serve a qualcosa?». «Non mi piace che lei conosca il mio e io non so chi cazzo sia». «Paolo? Carlo? Come preferisce». «Mi stia bene...» e prese a camminare. «Come va col cadavere che aveva il suo foglietto in tasca?» fece l'uomo. Rocco si fermò. Si girò a guardarlo. «Ne sa di cose, allora». «Abbastanza. Però non le sarò di aiuto. Non so chi sia quel tipo di Castel di Decima. Diciamo che non mi occupo di quell'affare». Schiavone prese un respiro. «E mi dica un po', Paolo e Carlo, dove li vuole trascorrere 'sti cinque minuti?». «Ci facciamo due passi?». «E famose due passi». Si incamminarono verso via Aubert. «Passiamo al tu? È meglio, che ne dice?» propose l'uomo. «Passiamo al voi, Paolo e Carlo. Che cazzo volete?». L'uomo scoppiò a ridere. Una risata falsa, impostata come la voce. «Sei forte Rocco. Ma lo sapevo. Sei uno che conosce poco le buone maniere. Dimmi un po', come vanno le cose a Roma?». «Non lo so, vivo ad Aosta». «E ti piace questa città?». «Vai al punto che mi sono già rotto i coglioni». «Va bene, se preferisci». Si fermarono dove la strada diventava poco più di un vicolo. L'uomo gettò la sigaretta. «Tu non lo capisci quando è il momento di fermarti...». «No, se non me lo spiegano gentilmente». «Allora te lo spiego gentilmente. Sei stato bravo, hai fatto il tuo dovere, hai giustificato lo stipendio, ora però basta così». «Chi è Bernardo Valenti?». «Allora non capisci o non vuoi capire?». «La seconda». Il labbro dell'uomo ebbe un leggero fremito. «È uno che ha avuto un piccolo incidente, ma ora è tutto a posto». «Mica tanto, se la vedi dalla prospettiva di Juana». «E ci vogliamo andare a preoccupare per un mezzo uomo?». «Aridaje co' 'sta storia dell'uomo e del mezzo uomo. Mettiti in testa che avete ammazzato una donna. Ce la puoi fare a capire?». Una signora con il passeggino passò accanto ai due fermi in mezzo alla strada. Rocco aspettò che si allontanasse per riprendere il dialogo. «Sì, io mi preoccupo. Vedi Paolo e Carlo, Juana Perez è diventata un cadavere per colpa del tuo amico Valenti, e questa è una cosa che mi fa preoccupare». «È stato un incidente, lo hai capito. Facevano dei giochini e...». «Allora spieghiamolo a un giudice e tutto si aggiusta. Certo, c'è occultamento di cadavere, ma con un buon avvocato se la cava con un paio di anni. A meno che il sedicente Valenti non sia un recidivo. È forse un recidivo?». L'uomo si accese un'altra sigaretta. «Vedi Rocco, io ho la possibilità di renderti la vita un inferno. L'avrai capito. E mi dispiace, perché io e te in fondo giochiamo con la stessa squadra. Solo che il mister in questo momento ha deciso di toglierti dal campo e sostituirti, e tu devi andare a sederti in panchina; la partita per te è finita. Chiaro il concetto?». «Chiaro. Mi piacerebbe che fosse il mister a dirmelo». L'uomo scoppiò a ridere. «Mi sa che non hai capito chi è il mister. Pensi che abbia tempo per venire a parlare con te? Con un vicequestore della polizia di Aosta? Rocco, ti facevo più intelligente. Te lo dico da persona che in fondo ti stima, e te lo dico nella tua lingua, così lo capisci: levati dar cazzo». «Patetico, non sei neppure di Roma. Arroti male la erre e poi si dice: levate dar cazzo. Stai a fa' a chi ce l'ha più lungo, e la questione è un'altra. Non è fra me e voi, Paolo e Carlo, la questione è che io voglio sapere chi è Bernardo Valenti, altrimenti non mi fermo». «Va bene, solo perché in fondo sei simpatico. Valenti è uno che ha la lingua lunga, e in questo momento, per certi affari, diciamo che la sua lingua per noi è molto preziosa. Fattelo bastare». «È un pentito?». «Fattelo bastare». «Sotto protezione? È questo Bernardo Valenti?». «Hai scassato la minchia». «Ecco, sei siciliano. Catania?». «Fottiti. E se rivedi la mia faccia è una notizia brutta». «Dai, posa per terra il ferro che porti sotto la giacca e chiudiamo 'sta storia adesso, in mezzo alla strada». L'uomo scosse la testa. «Sei rimasto un ragazzino di Trastevere, Rocco. La storia, se deve finire, finisce come diciamo noi e quando diciamo noi. Ah, se stavi andando da Fabiani ti evito l'incombenza. Dubito che ce lo ritroverai. Goditi Aosta, Schiavone, e pensa solo a mantenerti in vita, che è già una cosa piuttosto complicata. Puoi dire al negoziante di tornare a casa tranquillo, non ha niente da temere. Mica siamo assassini, noi...». «No?». «No. E per la cronaca, l'assassino di Juana Perez è tale Paul De Vries, un pluriomicida che abbiamo fermato alla frontiera due giorni fa. Ci tenevo a dirtelo, per evitarti figure di merda, anche se ci sei abituato». «Paul De Vries?». «Belga. Un bastardo». «E se quello vi sputtana?». «Difficile. È morto in uno scontro a fuoco». «E non sareste degli assassini?». L'uomo non rispose. Girò le spalle e si incamminò. «Un errore l'avete fatto però, Paolo e Carlo». Quello si fermò, poi si girò lentamente. «Vero. Non dovevamo fare entrare Juan Pedro Perez in quella casa. Ci è sfuggito di mano. Ma non si può prevedere tutto, no?». Rocco non lo perse di vista fin quando l'uomo non imboccò via Aubert sparendo alla vista. Il vicequestore stringeva i pugni con la voglia di spaccare una vetrina, prendere a calci un'auto, dare fuoco a un intero isolato. Tornò sui suoi passi e fumando come la ciminiera dell'impianto siderurgico si incamminò verso la questura. Appena aprì la porta del suo sgabuzzino restò di sasso. Era vuoto. Non un mobile, neanche la scrivania. Il foglio con le rotture di coglioni era sparito. E non c'era traccia di Lupa. «Italo!» chiamò. «Rispoli? Scipioni? Oh, ma c'è nessuno?». Apparve la faccia di Casella. «Dotto'... ha visto?». «Cosa? Dove è adesso il mio ufficio? Direttamente in strada?». «Macché... sta di nuovo dove stava... quello solito». «Cioè mi hanno ridato la stanza?». «Sì, Deruta e D'Intino hanno spostato tutto. Pure Lupa, che però ha schiaffato un mozzico alla coscia di D'Intino, chissà come mai». «Perché è intelligente e difende il territorio». «Venga, l'accompagno. Ah, questa è per lei» e gli allungò una busta. Rocco l'aprì. Dentro c'era un biglietto scritto dal questore di proprio pugno. Gli ordinava di presentarsi alla conferenza stampa, la cosa era della massima urgenza. Peccato, pensò Rocco, gli sarebbe piaciuto farci prima due chiacchiere. «E dov'è il commerciante, Rosset?». «È nella sua stanza che l'aspetta». Scesero le scale. «E la Gambino?». «S'è fatta mettere nel seminterrato, vicino al garage. Dice che ci sta benissimo. Novità?». «Lascia perdere, Casella, lascia perdere...». «Fra mezz'ora ci sta la conferenza stampa sul caso di Juana Perez. Che fa, ci va?». «Stavolta sì, Casella, ci vado eccome!». Erano arrivati al vecchio ufficio di Rocco. Il cartello delle rotture di coglioni era tornato al suo posto. Rocco ne aveva almeno una decina da aggiungere, ma rimandò l'incombenza. Appena aprì la porta sentì: «Scopa!». Alla sua scrivania D'Intino giocava a carte con Rosset. Appena lo videro, scattarono in piedi. Deruta s'era addormentato sul divano accanto a Lupa. «Deruta!» gridò Rocco. Quello si svegliò di soprassalto. «D'Intino! Ma che è? Avete preso il mio ufficio per un bar?». «Stavamo a passa' lu tempo, dotto'. Je so' fatto una scopa proprio mo' che...». «Fuori, fuori di qui! Tu pure Deruta, via!» e li scacciò. «Sciò, sciò, via, aria, aria!». Rosset lo guardava pallido. «Pure lei, se ne vada, non la voglio più vedere!». «Ma...». «Vada via! Torni da sua moglie e non mi scassi più il cazzo!» urlò. Rosset a passetti rapidi e brevi riguadagnò la porta. «Ma se io...». «Cosa le ho detto? Vada via!». Poi sbatté la porta e crollò sulla poltrona. Lupa lo guardava con le orecchie in su, allarmata. Rocco prese la chiave e aprì il cassetto. Poi andò alla finestra e si accese la canna. Il panorama era tornato quello abituale, i monti erano sempre lì e anche il cielo, ancora coperto, che non lasciava vedere il sole. «Vita di merda...» disse fra i denti. La porta si aprì, fece appena in tempo a gettare la canna fuori dalla finestra. Si voltò. Era il questore. Costa era teso, pallido, sembrava avesse addosso lo stesso vestito da una settimana. La cravatta, solitamente annodata alla perfezione, era slacciata come il primo bottone del colletto. «Ha ricevuto il mio biglietto?» aveva la voce incrinata. «Sì, l'ho ricevuto. Mi dica però come ci sono arrivati Farinelli e la sua squadra a via Brean». «Avevamo bisogno di una prova definitiva». «Non capisco». «Lo capirà in conferenza stampa. Una cosa sola. Si attenga ai fatti che descriverò». «Dove sono finiti i suoi doveri morali, dottore?». «Non so di cosa stia parlando. Già gliel'ho spiegato una volta, io servo lo Stato, Schiavone, e quello dovrebbe fare anche lei». «Lei sa quanto me che...». «Che cosa? Che ci sono cose più grandi, Schiavone, molto più di me e di lei. E non è una testa dura come la sua che può cambiare le cose. C'è una bilancia, e stavolta il peso sull'altro piatto è una tonnellata, quindi io le dico di venire immediatamente alla conferenza stampa e ascoltare con molta attenzione le informazioni che io, capo di questa questura, darò ai giornalisti». È fuori di sé, pensò Rocco. Non li chiamava più giornalai. «Cercherò di stare attento. Ma a differenza di lei non riesco a chiamare le cose con un altro nome». «Sarebbe?». «Lei, dottore, si trincera dietro una barricata di cazzate e finto senso del dovere. Lo sa cosa siamo noi adesso?». «Me lo dica». «Siamo peggio di Bernardo Valenti. Perché almeno lui una colpa ce l'ha e con la sua fuga l'ammette. Noi invece ce ne laviamo le mani e convochiamo una conferenza stampa». «Crede che la cosa mi renda felice? Io sono pronto a sacrificare Juana Perez solo perché la partita è molto più grande e di mezzo ci va la vita di un sacco di gente». «Ha ragione. Juana Perez è come una pedina degli scacchi. Sacrificabile, no?». «Ha detto la cosa giusta. E la prego, accolga i miei suggerimenti. Io di ripescare gente nella Dora ne ho le palle piene!» aveva gli occhi rossi e umidi. La mano, ancora poggiata sulla maniglia della porta, tremava. «Nonostante tutto, mi creda, io ho stima di lei e le voglio bene, Schiavone. La prego, se ne voglia anche lei...». I due uomini si guardarono. E per la prima volta Rocco vide il suo superiore sotto una luce diversa. Anche a lui toccava una vita che avrebbe volentieri scambiato con un'altra. «Si sa almeno a che serve 'sto Valenti?». «All'incolumità di un paio di magistrati, Rocco. Ma io non te l'ho detto». Era passato al tu, forse per distrazione, forse perché era davvero stanco di menzogne e sotterfugi. «Non me l'hai detto» sospirò Schiavone. «Questa giornata di merda non la dimenticheremo tanto facilmente, Andrea». «Non la dimenticheremo, Rocco». Poi guardò la stanza e strinse gli occhi con un sorriso tirato. «Ha visto? Promissio boni viri est obligatio. Ora tocca a lei». «In che senso?». «Mettere la macchinetta del caffè con le cialde. Ci vediamo giù in sala fra dieci minuti, dottor Schiavone!». Andrea se n'era andato ed era tornato il dottor Costa, il capo della questura di Aosta. Sulle sedie un mucchio di giornalisti. Dietro il tavolo il questore, accanto a lui Farinelli, serio e con una cartellina di plastica davanti. Baldi era in piedi, in un angolo, teneva gli occhi bassi e la giacca poggiata sulle braccia. In maniche di camicia, si notava il sudore sotto le ascelle. Da quando Rocco era entrato e s'era messo in fondo alla sala, il magistrato gli aveva rivolto solo un'occhiata. Si vergognava di essere lì, si vergognava di quello che stava per sentire, si vergognava dell'intera sua vita. Benvenuto nel club, pensò Rocco. «Il caso di Juana Perez, nota come Sonya nell'ambiente della prostituzione, è arrivato alla sua conclusione» esordì Costa, mentre i giornalisti prendevano appunti. «La mano misteriosa dell'omicida, tanto misteriosa non è più. Abbiamo arrestato alla frontiera Paul De Vries, noto criminale italo-belga che in uno scontro a fuoco ha perso la vita. A confermare i nostri sospetti il sostituto Farinelli ha trovato nell'appartamento di Juana Perez impronte digitali del pluriomicida...». Farinelli prese la parola. Guardava la sua cartellina, non l'uditorio. «È stato piuttosto semplice» esordì. «Ce n'erano un po' dappertutto, come io e i miei uomini abbiamo rilevato. I riscontri con De Vries, già noto alle forze dell'ordine e che fino a qualche giorno fa alloggiava in un appartamento a Châtillon, hanno dato esito positivo». «In più» continuò Costa, «dopo un tentativo di fuga e la sparatoria che l'ha visto cadere sotto i colpi degli agenti di confine, sul cadavere di Paul De Vries sono stati ritrovati alcuni effetti personali che appartenevano alla vittima». Si alzò la mano di Angrisano. «Il movente?». «Un gioco erotico finito male. Paul De Vries era uso a questo tipo di attività, se avete voglia di dare un'occhiata al suo curriculum, vi accorgerete che era già stato fermato nel '98 e nel 2005 per violenza carnale». Qualche giornalista si mise a guardare il plico che il questore aveva distribuito. «Ora tengo a dire che le indagini, tenute in sinergia con la Criminalpol e la Dia, hanno dato i loro frutti. Vorrei ringraziare...». Rocco non ascoltava più. Osservava i giornalisti prendere appunti, le telecamere con l'occhio rosso, Baldi con la testa poggiata sulla parete bianca. Aveva lo stomaco contratto. Si chiese se era il caso di alzare la mano e dire la sua. Stava per farlo, poi ci ripensò. Cercò lo sguardo di Farinelli, ma quello continuava a tenere gli occhi bassi, concentrati sui suoi appunti. Sei una merda, pensò, e decise di uscire dalla stanza. Quello che gli pesava di più era parlare coi suoi uomini che sapevano come lui la verità, ma non era la verità giusta. Di nuovo si sentì sporco e stanco, spiegazzato e infangato. La frustrazione del fallimento si mischiava alla marcia sensazione di aver sguazzato per l'ennesima volta in una palude a cercare serpenti velenosi. Ma non era il momento dell'autocommiserazione. Ora doveva sforzarsi di preservare l'incolumità dei suoi uomini e cercare di proteggerli, come Costa con lui. Ragione di Stato, erano le parole impresse nel cervello. Sì, avrebbe parlato chiaro, senza peli sulla lingua. Erano giovani, si sarebbero arrabbiati, forse Antonio Scipioni avrebbe urlato, ma avevano ancora parecchia strada da fare. Era giunto anche per loro il momento di ingoiare una bella badilata di merda. «Cosa?». Caterina con gli occhi di fuori non credeva a quanto aveva appena sentito. «È così, viceispettore. E non si cambia una virgola». Antonio scuoteva la testa, Italo invece fissava il pavimento. «Ma se... insomma Rocco, noi lo sappiamo. Noi quattro la verità la conosciamo. Perché non...». «Perché è così, perché non contiamo un cazzo! Perché fai 'sto mestiere per quattro soldi e ti tieni il rospo senza rompere i coglioni. Pensi che a qualcuno interessi un tuo rimorso di coscienza, Caterina? O la tua vergogna, Italo? Oppure la rabbia di Antonio? O la mia? La partita si chiude così, il cattivo l'hanno preso i colleghi e noi siamo felici così». «Ma insomma almeno si sa chi era questo Valenti?» chiese Italo con un filo di voce. «Un testimone sotto protezione. Di cosa non lo so, ma diciamo che preferiscono conservarselo». «A pensarci bene, se questo mette bocca sulla camorra, o sulla mafia, insomma sulla criminalità organizzata e ci aiuta...». «Ma che cazzo dici, Italo?» Antonio esplose. «Che cazzo dici? Siccome è un testimone di chissà che cosa se la passa liscia dopo aver ammazzato una persona?». «Tecnicamente non l'ha ammazzata. È morta per un gioco che stavano facendo assieme». «L'ha buttata nel fiume!» gridò Antonio. «Questo una persona normale non lo fa!». «Una persona normale no. Un camorrista del cazzo evidentemente sì». «Rocco, ci fermiamo qui?». Antonio aveva il respiro mozzato. «Dimmi solo se ci fermiamo qui». Schiavone abbassò la testa. «Finisce qui». E calò il silenzio. «Io non posso mettere a rischio la vostra vita, Antonio. Non me la sento e non voglio». «E io non so se voglio ancora lavorare con te» gli rispose Antonio Scipioni. «Liberissimo di andartene. Solo che questa cofana di merda è solo la prima della vostra carriera. Il consiglio che vi do è di farci l'abitudine». «Insomma abbiamo perso». «Antonio, questo non è un gioco. Non si perde e non si vince. Questa è la vita» fece Caterina che aveva dato voce a un suo pensiero senza quasi volerlo. «Secondo me abbiamo pareggiato» disse Italo. Ma nessuno sorrise. «Ti ricordi, Antonio, quando sei venuto a comunicarmi i numeri di telefono di quel Rosset?». «Certo che mi ricordo». «Posso sapere a cosa pensavi?». «A tante cose. Che il gioco era duro, che eravamo tutti in pericolo e che...». «Appunto. Basta così. Te la facevi sotto, Antonio. E io come te. È normale. Ma ti dirò la verità, e ti prego di credermi. Se fosse per me, per me solo, io Bernardo Valenti lo andrei a prendere pure nel culo dell'inferno. Ma solo non sono. Voi ne sapete quanto e più di me. Ora spero ti sia chiaro perché ci fermiamo qui». «Non è per questo che sono un poliziotto». «Tu devi servire lo Stato. E lo Stato mi pare sia stato abbastanza chiaro su quello che vuole da te. Ti piaccia o meno è così! Non è una tua partita privata». «Non ti riconosco più» gli disse il giovane agente. «Facci l'abitudine. Io non sono un supereroe, e manco un eroe. Sono una persona normale Anto', normale. E le persone normali devono capire quando hanno perso. Vuoi aprire un'indagine su questo Valenti, cercarlo per anni senza trovarlo mai, passare il resto dei tuoi giorni a farti rodere il fegato, lasciarci la pelle per la cirrosi epatica, o per un tumore? Vuoi sapere la verità? Quando Bernardo Valenti non servirà più, ce lo daranno su un piatto d'argento, lo Stato spesso fa così. Tradisce più di me o di te». «Ma allora perché facciamo i poliziotti?». «Perché non abbiamo di meglio, perché fin dove possiamo diamo il nostro contributo. Ora il contributo che ti chiedono è di stare buono e zitto. E tu attacchi l'asino dove ti dice il padrone». «Non mi piace. Non mi piace proprio...». «Neanche a me, Antonio. Siamo finiti in un gioco più grande di noi. E per dirtela tutta, a noi le carte neanche ce le danno». «Vedila così» fece Caterina. «Siamo in trincea sotto il fuoco nemico. Ci stanno massacrando. Se però uno di noi si sacrifica, gli altri se la caveranno. Lo faresti?». «Se fosse così. Ma io non mi fido» rispose Antonio. «Non mi fido perché se sto in trincea so chi è il nemico, qui sono solo ombre. Vorrei solo essere un generale, così...». «Co 'ste premesse Anto' tu generale non ci diventerai mai!» disse Rocco alzandosi dalla scrivania. Guardò Lupa che capì e scodinzolando si avvicinò al padrone. «Ragazzi, io vi voglio bene, e questo giorno non lo scorderò mai. State protetti, state nascosti, e cercate di sopravvivere. Io di morti per colpa mia ne ho le palle piene. Ora vado al bar da Ettore a prendere una cosa pesante, molto pesante. Se volete unirvi a me, mi fa piacere». Appena uscito dalla stanza vide la Gambino ferma davanti al distributore che sorseggiava un caffè. Si fissarono per un istante. Quella fece sì con la testa. Non c'era bisogno di altro, e Rocco cominciò a guardarla sotto un'altra luce. Camminava da solo verso il centro storico, osservava la gente. Lupa incontrò un boxer e ci giocò per qualche minuto. Bella la vita. Dai una sniffata per capire con chi hai a che fare, qualche zampata per saggiare la pericolosità o la giocosità del nuovo amico, tiri un abbaio e poi vai dritto per la tua strada in attesa della nuova scodella, della nuova pappa, della pisciatina serale. Voleva passarsi un panno sul cervello e cancellare tutto quello che s'era accumulato in quei giorni affilati. Due ragazzi si rincorrevano e ridevano. Quattro persone con la barba bianca vestite da montanari parlavano a bassa voce sedute a un tavolino con un bicchiere di bianco davanti. Un uomo passando lo salutò con un sorriso, Rocco ricambiò senza avere la minima idea di chi fosse. Poi all'angolo con via Croix de Ville vide la Buccellato ferma davanti a una vetrina che sorridendo gli andò incontro. «Salve Schiavone, allora un altro successo?». «Così pare. Non mio stavolta». «Ce l'ha ancora con me?». «Le cose passano». «Ma le persone restano». «No, anche quelle cambiano». La giornalista lo guardò seria. «È una storia assurda quella di via Brean». «Mica tanto. Un omicidio, un colpevole, la risoluzione del caso». «Tocca a me scrivere l'articolo. Vuole aggiungere qualcosa che il mio ex marito non ha detto?». «No. È stato piuttosto esaustivo». «A me questa storia non convince» obiettò la giornalista. «Ah no? E se le dicessi che in mezzo ci sono i servizi? Che l'assassino l'hanno fatto sparire e che se viene fuori ci va di mezzo la sicurezza di un sacco di gente?». «Secondo lei me la bevo?». «No, non se la beve. E ora che mi viene in mente ho sete. La saluto, signora Buccellato» e tirò dritto verso piazza Chanoux. Arrivato al bar di Ettore fu felice di vedere Italo e Antonio che lo aspettavano. Prese una sedia e si unì ai suoi uomini. «Oh, però paghi tu» disse Antonio. «Pago io. Che vi va?». «Rhum?» fece Antonio. «Con la cioccolata». «Bella idea. Pure io» rispose Italo. «Vada per il rhum e crepi l'avarizia. Prendiamo uno Zacapa Royal di trent'anni!». «Addirittura!». «Bisogna festeggiare, Antonio». «Cosa?». «La sconfitta. Solo così si riesce a vivere, secondo me». Si alzò dal tavolo per ordinare. Dopo due passi si fermò. «E Caterina?». «Non è voluta venire. Ha ricevuto una telefonata ed è scappata» replicò Italo. «Che ha 'sta ragazza?» chiese Rocco. «Non lo so. Non ci capisco niente e detto fra noi mi sono anche stufato». Solo, nel deposito della morgue, guardava la bara che conteneva il corpo di Juana Perez in attesa che la venissero a prendere per rimandarla dai genitori a Rosario, in Argentina. «Ciao...» disse a bassa voce toccando il legno chiaro. «Ora ti rimandiamo a casa, scusami Juana se non ce l'ho fatta, fa' buon viaggio». Alberto lo aspettava nel corridoio. «Era come dicevi tu?». «Come dicevo io. Abbiamo fatto tardi». «Chi c'è dietro?». «Noi» fece Rocco. «Io, te, la questura, la procura, tutti». «Puoi essere un tantino più specifico?». «Servizi. Va meglio ora?». «No, non va meglio, troiaccia della miseria. Proprio no. Tu che fai?». «Che faccio? Niente, che vuoi che faccia, Alberto? Quel figlio di puttana chissà dove l'hanno nascosto ora. Sotto quale nome. Sempre se sta ancora in Italia». «Ti va di andare a bere da qualche parte?». «Ho bevuto coi miei uomini e ho la testa che gira. Voglio andare a casa. A dormire, se ci riesco». Fumagalli gli diede una pacca sulla spalla. «Stammi bene...». «Anche tu, mi raccomando» e si incamminò per il lungo corridoio. «Oh, Rocco». «Che c'è?». «Però siamo stati bravini anche stavolta, no?». «Ariconsolamose co' l'ajetto» gli rispose. «Potresti parlare italiano? Lingua che, ti sembrerà strano, parliamo ancora in tutta la penisola e usiamo per poterci comprendere nonostante le enormi differenze antropologiche e sociali?». «Riconsoliamoci con l'aglio. È un modo di dire. Non servono altre spiegazioni, credo. Grazie Alberto». «E di che?». Sciolse il guinzaglio di Lupa annodato a una panca e si incamminò verso l'uscita. Stava fuori che l'aspettava accanto alla sua Lancia Delta. Rocco non aveva voglia di parlarci, aveva bisogno di buttarsi sul letto e chiudere gli occhi ma Farinelli gli andò incontro deciso. Rocco alzò le mani. «Non mi interessa, qualsiasi cosa tu abbia da dirmi, non mi interessa». «Ascoltami! Non è colpa mia. Ho dovuto farlo». «Appunto, lo so. Che vuoi da me? L'assoluzione? Cercati un prete». «Ma devi saperlo. Mi hanno chiamato, non avevo scelta, lo capisci? Non pensare che l'abbia fatto a cuor leggero». «La Gambino s'è rifiutata, mi pare». «E ti pare male. La Gambino non è neanche stata interpellata». «Tu non hai trovato impronte in casa, vero?». «Sì, ne ho trovate, corrispondono per una decina di punti a quelle di De Vries». «Non ci credo. E poi dieci punti non sono niente». «Però posso dormire tranquillo, come te, come i tuoi uomini». «Su questo ti do ragione. L'unica cosa, mi piacerebbe sapere il vero nome di Bernardo Valenti». Farinelli sorrise. «Basterà leggere i giornali, credo. Ora torno a Torino, hai un sostituto tutto per te qui ad Aosta, quindi non credo ci rivedremo. Sappi che lavorare con te è stato un incubo, ma sei comunque in gamba» e gli allungò la mano. Rocco non gliela strinse. «Posso sapere chi ti ha dato l'ordine?». «Perché lo vuoi sapere?». «Perché se ti hanno contattato vuol dire che si potevano fidare. E se si possono fidare io e te non abbiamo più un cazzo da dirci. E le stronzate che l'hai fatto per la tua famiglia, per tua moglie, per dormire sonni tranquilli, valle a raccontare a chi ci crede. Farine', lavorare con te è stato un incubo, ma non pensavo fossi un pezzo di merda fino a questo punto». Si incamminò verso la strada superando il cortile pieno di alberi. Sentì il rumore dell'auto che poi gli passò a pochi centimetri. La guardò finché i fanalini posteriori furono inghiottiti dalla notte. Il sapore del rhum era svanito. Alzò gli occhi al cielo. Il vento era calato e aveva lasciato nuvole grigie e minacciose ammassate sulla città. La prima goccia lo centrò in pieno viso. «Daje a ride...» mormorò. Superò il parcheggio e si avviò verso casa. I fari delle auto e le insegne dei negozi ormai chiusi illuminavano la pioggia. Aumentò il passo. Lupa trotterellava al suo fianco immersa nei suoi odori. Mollò una pisciata accanto agli pneumatici di un'Alfa parcheggiata. Attraversarono insieme la piazza e si trovarono in via Aubert. Gettò un'occhiata all'ufficio di Fabiani. Aveva le luci spente. Proprio di fronte alla vecchia chiesa vide Caterina parlare con una donna anziana. Rocco si fermò ad osservare. Il viceispettore Rispoli se ne stava in silenzio con gli occhi sgranati a guardare la donna, il corpo teso, aggressivo, pareva la corda di un violino. L'altra invece sembrava una tartaruga che cercava di richiudersi nel guscio, ritraeva il collo e parlando allargava le braccia, come stesse chiedendo scusa. Il vicequestore capì che stava violando qualcosa di molto privato, mise il guinzaglio a Lupa, semmai la cucciola avesse riconosciuto Caterina e le fosse andata incontro per salutarla, e scantonò nel vicolo. Avrebbe allungato un po' per arrivare a Croix de Ville, ma era meglio così. La stradina era buia e deserta, superò un albergo e puntò verso la traversa che lo avrebbe riportato sulla via principale. Caterina era un mistero. Non parlava mai delle sue cose, della sua vita, ma non ci voleva un fine psicologo per capire che qualcosa la stava tormentando. Da un po' non era più lei, per la precisione da quando era tornata dal suo giorno libero. Avrebbe voluto parlarle, cercare di capire se poteva aiutarla, ma poi si rese conto che se l'aiuto non era richiesto lui la mano doveva tenersela in tasca, non tenderla in soccorso. Svoltò nel vicolo e dopo pochi metri tornò sulla via principale. Si era lasciato la vecchia chiesa alle spalle, non c'era più pericolo di disturbare il viceispettore. E invece se la ritrovò davanti, proprio in mezzo alla strada. «Caterina!» disse sorpreso. «Buonasera» rispose quella. Lupa le saltò subito addosso. «Com'è che sbuchi da quel vicolo?» chiese carezzando la cucciola. «Così, allungavo prima di tornare a casa». Caterina guardò la strada, proprio verso la vecchia chiesa. «Dalla questura è una strada un po' tortuosa, no?». «Il lavoro ti sta distruggendo, amica mia. Ormai sospetti di tutto». «C'è qualcosa di male? No, è che mi viene automatico pensare che questa è la strada per tornare a casa tua, invece hai preso questo vicolo prima della chiesa vecchia. Proprio dove ero io neanche un minuto fa». «Va bene. Ho notato che parlavi con una donna e non mi sembrava il caso di farmi vedere». Caterina si concentrò sul cane. «Bella Lupa, andiamo a mangiare la pappa?». Rocco tolse di nuovo il guinzaglio alla cucciola che fu attratta da un cartello di divieto di sosta. «Io me ne torno a casa» disse. «Se ti va di accompagnarmi facciamo un pezzo insieme». Caterina annuì e si incamminarono verso via Croix de Ville. La pioggia s'era calmata. Poche gocce che cadevano isolate. «Mi ha detto Antonio che il rhum era buonissimo». «Potevi venire anche tu». «Purtroppo... un piccolo problema da risolvere». Camminarono ancora in silenzio. «Ti ringrazio Rocco» gli disse passandosi la mano nei capelli per asciugarli. «E di cosa?». «Di non farmi domande». «È la tua vita, Caterina. Io non c'entro niente» e si accese una sigaretta. «È che a volte le scelte sono difficili. Ma ci si rialza, no?». «Questo è quello che pensi tu. Io la vedo diversamente». Un ragazzo che usciva da un pub salutò il viceispettore Rispoli con un cenno del capo. «Rocco, quello che è successo oggi... credo che ognuno di noi abbia bisogno di tempo per metabolizzarlo». «Sì, anche se non lo digeriremo mai. Poi dipende da come uno vede la vita». «E tu come la vedi?». «Da quando ero piccolo ho sempre avuto la sensazione di stare nella camera della morte, hai presente? Quel percorso che fanno fare ai tonni nelle mattanze? Per quanto sia tortuoso, pieno di angoli e svolte, finiscono tutti nella trappola per essere trasformati in scatolette. Ecco, per me è la stessa cosa. Qualsiasi decisione tu prenda nella vita arrivi sempre nello stesso posto, nella scatoletta. Ci illudiamo di fare delle scelte, ma la strada è già segnata e questo non me lo toglie nessuno dalla testa. Pensa a quello che è successo oggi». Caterina lo guardò. «È un modo un po' duro di vederla». «Quello che cambia è una sosta, un bacio, innamorarsi come dei coglioni, ma per il resto siamo sempre lì perché la strada è una sola». «Mi sarebbe piaciuto conoscerti quando eri un bambino. Sai come ti immagino? Sempre zitto, con pochi amici, non molto obbediente...» e per la prima volta da giorni sorrise veramente. «Ti sbagli. Ero un bambino aperto. Non ci crederai ma avevo un sacco di amici. Ed ero obbediente. Pensa, andavo pure bene a scuola». «E poi cos'è successo?». «Ho capito che quello che provavo era vero. E che la vita, pure se è bella, è bella da morire». «Su questo non hai torto. Vuoi sapere chi era quella donna con cui parlavo?». «Se vuoi, altrimenti lascia perdere». «Era mia madre. Non ci parliamo da anni, a parte stasera e poco altro. Te l'ho detto, la mia non è proprio una famiglia modello. Non lo è stata, insomma. Ora però siccome non ne posso più, ti invito a prendere una cosa da bere. Ricordi? Avanzo un rhum! A proposito, ce la fai?». Le diede un pizzicotto sulla guancia. «Mi sottovaluti, cara...». Avevano perso il conto dei gin-tonic. Rocco sentiva caldo e gli sembrava di avere qualche diottria in meno, ma era una sensazione piacevole. Aveva ingollato più alcol in quel giorno che negli ultimi tre mesi. Camminavano incerti fra le strade del centro storico, il lastricato bagnato rifletteva le luci dei lampioni. Ridevano per qualsiasi sciocchezza dicessero. Come due sommozzatori in una grotta sottomarina, avevano trovato una bolla d'aria e la stavano respirando a pieni polmoni. «Quella cosa della vita che hai detto prima...» Caterina si aggrappò al braccio di Rocco, «non è proprio così secondo me. Io la vedo diversamente. Ci è dato un tempo, e lo sappiamo, e durante quel tempo dobbiamo fare il possibile per onorarlo». «Sono discorsi troppo complessi dopo quella caterva di gin-tonic che ho bevuto». «No, no, dico sul serio. Lo sappiamo tutti dove andiamo a finire, ma fino ad allora, dico io, viviamola questa vita, no?». «Certo. È ovvio. Anche se tu non mi sembra ci stia lavorando». Caterina si fermò e lo guardò. «Perché dici così?». «Perché non hai ancora trent'anni e potresti fare qualsiasi cosa. Invece guardati! Lavori come una bestia, hai un passato alle spalle che non sai affrontare e che ti pesa come un macigno, non ti liberi dei rami secchi, o almeno ancora non ci riesci. Volevi frequentare la scuola di PS e invece stai ancora qui ad aspettare di dare quella tesi per laurearti. In più sprechi le serate insieme a un vecchio elefante diretto a spargere le ossa accanto a quelle degli avi». «Hai detto un sacco di cose giuste, vicequestore Schiavone, ne hai sbagliata solo una». «Sarebbe?». «Non sei un vecchio elefante, e io non sto sprecando le mie serate». Lenta avvicinò il viso a quello di Rocco e lo baciò. Sentì un sapore di alcol misto a menta. Era tentato di ritirarsi, sapeva quanto fosse sbagliato, ma la lingua di Caterina era calda, e il suo corpo sembrava appiccicarsi al suo. Non devi Rocco, non è giusto, pensava, potrebbe essere tua figlia. Si staccò dalle labbra della ragazza e la guardò negli occhi grandi e luminosi, più dei lampioni stradali. La pelle liscia e i denti bianchi, i capelli tirati all'indietro, la fronte serena, le labbra umide e scure. «Ma sticazzi» disse, e abbracciandola con tutta la forza che aveva seguitò a baciarla. Si svegliò di soprassalto. Si stropicciò gli occhi. Accanto a lui Caterina dormiva, una gamba fuori dal letto. Ricordava perfettamente quello che era successo, ma la luce era poca, i fumi dell'alcol tanti e non si era goduto quanto avrebbe voluto il corpo di Caterina Rispoli. Ora lo poteva guardare a lungo, accarezzato dalla luce lattiginosa del mattino che penetrava dalla finestra. Era bellissimo. Con il viso schiacciato sul cuscino Caterina dormiva profondamente. Avvicinò la mano, non resisteva, voleva toccare la pelle di seta, sentire sotto il palmo il calore delle gambe, dei glutei. Ma a pochi centimetri si fermò. L'avrebbe svegliata, invece voleva continuare a osservarla ancora per un po'. Ciocche bionde sparse sul guanciale coprivano parte della bocca e degli occhi, ma sembrava che la ragazza stesse sorridendo. Si soffermò sul collo, dove i capelli diventavano leggerissimi, fili di seta. Quello era il punto che aveva sempre preferito in una donna. Il collo e dietro le orecchie. Da ragazzo si vergognava a dirlo agli amici, loro parlavano solo di tette, culo, cosce. Lui non osava dire «A me fa impazzire il collo dietro le orecchie!». Che figura ci faceva? Avvicinò le labbra e ci depositò sopra un bacio sfiorando appena la pelle. Profumava di fiori, ma probabilmente era il deodorante che usava. Invece dalle sue ascelle proveniva un afrore acre, acido. Gli tornò la voglia di fare l'amore ma la luce fuori dalla finestra parlava chiaro; dovevano essere le otto passate. Lupa era restata in salone a dormire sul divano. Appena lo vide si stiracchiò e lo seguì all'angolo cottura. Rocco riempì la ciotola di croccantini e ci aggiunse il tonno dalla scatoletta. Mentre la cucciola sbranava il pasto se ne andò a fare la doccia. La caldaia funzionava. Indugiò a lungo sotto il getto caldo e accogliente riordinando le idee e lasciandosi andare ai ricordi della notte appena passata. Ebbe un'erezione ma la ignorò. Non era il caso. Tornò in salone asciugandosi i capelli proprio nel momento in cui la suoneria del cellulare si mise a cantare l'inno alla gioia di Beethoven. Si precipitò a rispondere e sbatté lo stinco su un mobiletto basso dell'ingresso. «Porca troia...» disse fra i denti. «Rocco? Tutto bene?». «Chi cazzo è?». «So' Brizio!». Crollò sul divano. «Ho dato una botta allucinante... Dio che dolore». «Mi dispiace, Rocco». «Pure a me. Aspe' che metto il vivavoce e almeno mi massaggio la gamba... dimme Brizio...». «Allora» la sirena di un'ambulanza coprì la voce dell'amico e rimbombò nel salone. «Non capisco un cazzo, non te sento!». «Ora?». «Sì, dove ti trovi?». «A Testaccio. C'è una manifestazione. Allora, sono stato alla banca. Il mio amico m'ha dato una mano». Rocco si osservò la gamba. Il mobiletto aveva lasciato un graffio profondo sulla pelle. «Seba ha prelevato 600 euro, poi ha usato il bancomat». «Dove?». «A Foligno, in Umbria». «In Umbria? Ci conosce qualcuno?». «Che io sappia no». «Ce l'hai una cartina?». «E che so, l'Aci?». «Vabbè, grazie, grande notizia. Stamme bene». «Che fai vieni?». «Devo controllare una cosa. Tu resta alleprato». Un clacson potentissimo risuonò nel microfono del vivavoce. «Che è?». «E sto alleprato sì. Per poco non m'arrotavano. Fatte vivo!» e chiuse la comunicazione. Aprì il cassetto del settimino e prese una canna già preparata. Doveva togliere le ragnatele dell'alcol dal cervello, mettere in moto gli ingranaggi cognitivi. L'accese guardando fuori dalla finestra. La nebbia nascondeva la piazza e lasciava intravedere il palazzo di fronte e il bar aperto. Due bambini che si rincorrevano furono inghiottiti dal biancore. Al terzo tiro afferrò il cellulare e chiamò Alfredo De Silvestri. Rispose al primo squillo. «Alfredo? Sono Schiavone!». «Dottore, aspetti» si sentiva lontano il rumore di una stampante che poi si affievolì. Segno che Alfredo De Silvestri aveva chiuso la porta. «Allora mi dica, dalla voce sembra cosa urgente». «Lo è, Alfredo. Ma sei al lavoro pure oggi?». «Mi tocca». «Allora ascolta, mi dovresti fare una ricerca rapida. Ho bisogno di sapere se ci sono denunce per furti a benzinai. Ma non guardare i benzinai grossi tipo autostrada o raccordo anulare, quelli piccoli, sulle statali per capirci. E isolati». «Zona?». «La butto lì, diciamo Roma nord, Salaria, Flaminia». «Quanto tempo mi dà?». «Zero!». «Come sempre. Ricevuto. La richiamo appena possibile». Spense la cicca con l'acqua del lavello e la gettò nella pattumiera. Poi cominciò a rivestirsi. La gamba gli doleva ma non voleva pensarci. Riprovò ancora a chiamare Sebastiano, ma quello continuava a tenere il cellulare spento. Se ci aveva visto giusto, se l'amico era sulla coda di Enzo Baiocchi, lo addolorava che lo stesse facendo senza dirgli niente. Perché in fondo, e ormai Rocco l'aveva capito, Sebastiano non si fidava più di lui. Altrimenti lo avrebbe chiamato, lo avrebbe coinvolto. E non alleviava il suo disappunto sapere che l'amico non si fosse fidato neanche di Brizio e Furio. Sebastiano aveva scelto di giocarsi la partita da solo. E immaginarlo in giro, probabilmente armato in un territorio non suo, accrebbe la sua ansia. Uscì di casa lasciando la porta aperta e bussò al vicino. Attese mezzo minuto, poi apparve la faccia sonnacchiosa di Gabriele. «Ah, Rocco, come va? Ha visto che...». «Mi devi un favore, ricordi?». «Eh? Sì certo, che...». «Tu hai il wi-fi?». «Sicuro» rispose quello. «E secondo te a casa mia prende?». «Certo che sì. È potentissimo». «Allora prestami il portatile». Gabriele lo guardò senza muoversi. «Quale parte non ti è chiara?». «Se deve fare una ricerca le posso dare una mano». «No, Gabriele, meglio se te ne stai a casa». Il ragazzo sparì per tornare poco dopo con il computer. «Ecco qui, l'ho già acceso. Per andare su internet può richiederle la password». «E qual è?». «Hey teacher leave the kids alone, tutto attaccato». «Bravo! Cominci a imparare». «Dal momento che è sveglio, posso mettere i Saxon?». «Ti taglio la giugulare». «Ma perché? È sveglio!». «Io sì...». Un sorrisetto furbo apparve sulla bocca del ragazzo. «Non è solo». «Gabriele, tu proprio non vuoi imparare, vero?». «A farmi i cazzi miei? No, lo so. Allora metto gli Iron Maiden, conosce?». «Fanno schifo». Per risposta Gabriele mostrò la sua maglietta nera con il logo del gruppo. «Metto un pezzo dolcissimo, gliel'assicuro». «Se schitarrano e vomitano rutti busso di nuovo e sei un adolescente morto» si girò e tornò nel suo appartamento. «Ah, forse le farà piacere sapere che ormai è ufficiale. Con la sospensione e tutto ho perso l'anno». «Mi dispiace, ma era la cronaca di una morte annunciata. L'hai detto a tua madre? Ah no, sarà a Torino o a Milano o a Bergamo per lavoro». «No. È di là che fa colazione. Sì, gliel'ho riferito. Sa cosa ha detto?». «S'è incazzata?». «No. Ha scosso la testa e ha pronunciato le seguenti parole: Gabriele, Gabriele, cosa ho fatto per meritarmi questo?». «Ha tutta la mia comprensione» poi entrò in casa. Si sistemò con il computer sul tavolo del salone e inserì la password. Aprì una cartina. Guardò la posizione di Foligno, in Umbria. Fra Spello e Spoleto. Esattamente sulla statale 3. Che ci faceva Sebastiano da quelle parti? Evidenziò il corso della strada che passava da Terni per arrivare fino a Rimini e proseguire verso nord. Come promesso Gabriele aveva acceso lo stereo. Una canzone terrificante riusciva ad attraversare i muri. Murders in the Rue Morgue. Le solite schitarrate violente e orribili che impestavano le orecchie dell'adolescente, la batteria assordante, Rocco storse la bocca ma lasciò correre anche perché il cellulare si mise a suonare. Rispose con il vivavoce per restare impegnato sulla cartina. «Sì». «Dottore, sono De Silvestri». «Un fulmine! Che hai da dirmi?». «Allora, non so se può servire. Ci sono due denunce. Una di una pompa sulla Flaminia, altezza Sacrofano. L'altra invece sulla Salaria, a Monterotondo». «Si sa anche quello che è successo?». «A Sacrofano hanno divelto la colonnina del self-service. A Monterotondo invece... aspetti che mi hanno mandato copia della denuncia». Sentì scartabellare. Alfredo si schiarì la voce. «Dunque, io sottoscritto Ermanno Pioli eccetera eccetera. Sì, ecco, un uomo ha fatto il pieno alla macchina poi pare abbia stordito il povero Pioli che si è risvegliato con un trauma cranico e il portafogli vuoto». «Quindi questo Ermanno Pioli l'ha visto in faccia?». «Così pare». «La stazione di rifornimento non ha telecamere?». «No, qui non dice niente». «E chiaramente non ha preso la targa». «Descrive l'auto dicendo che è una Fiat Cinquecento bianca». «Una Fiat Cinquecento bianca... capirai. Potrei chiederti un'altra ricerca ma è allucinante». «Vuole sapere quante denunce di furto di Cinquecento bianche ci sono state a Roma nelle ultime tre settimane?». «Ma perché, lo sai?». «Dottore, è l'auto più rubata in assoluto a Roma e provincia. Nove solo qui all'EUR. In neanche un mese». «È in gamba, se ci ho visto giusto» fece Rocco. «Di chi parla?». «Del ladro. Se avesse preso un'auto che viene rubata raramente sarebbe stato più facile rintracciarla, no?». «Sì, non ci avevo pensato. Insomma si mimetizza dice?». «Si mimetizza. E questo depone a favore del sospetto che ho. Va bene, Alfredo, dammi l'indirizzo preciso del benzinaio». «Salaria Monterotondo chilometro 32 e 300». Rocco lo segnò su un foglio. «Bene Alfredo... grazie!». «E di che? Qualsiasi altra novità la chiamo». «Ci conto». Chiuse la comunicazione e guardò la mappa sul monitor. La Salaria portava a Rieti, da lì proseguendo sulla statale 3 si saliva verso nord. Ci si poteva arrivare in Slovenia. Cliccò sul confine e la strada divenne un serpente blu da Roma al Friuli. Poteva essere la prima traccia. Sentì un rumore e si voltò. Forse Caterina si era svegliata. Entrò nella stanza in punta di piedi e la vide ancora lì, sul letto. Si era coperta col lenzuolo ma dormiva profondamente. Decise di prendere Lupa e lasciarle un messaggio. Arrivò alle cinque del pomeriggio. Lupa aveva sonnecchiato per tutto il viaggio. Caterina l'aveva chiamato a mezzogiorno protestando per la sua scomparsa e più che altro per cose pratiche, se c'era da chiudere qualche finestra, dove fosse Lupa e soprattutto dove stesse andando lui. Le aveva mentito rispondendole Milano, a trovare un suo caro amico che non vedeva da tempo. Il viceispettore sembrava l'avesse bevuta e si erano salutati con un bacio. Per tutto il viaggio pensò alla notte trascorsa con lei. Si era fatto un sacco di domande su cosa sarebbe diventato quel rapporto, se avrebbe dovuto considerarlo un'avventura breve e piacevole o qualcosa di più. Rivedersi al lavoro e far finta di niente oppure parlarne e magari col tempo rendere la cosa di dominio pubblico? Cosa avrebbe voluto lui, sinceramente? Doveva dare una risposta alla domanda principale, la stessa che poneva ai suoi amici intrappolati in storie d'amore difficili e contorte: puoi pensare di stare lontano da lei? Poteva pensare di stare lontano da Caterina? Non lo sapeva. Avrebbe fatto meglio ad evitare di mettersi in quella situazione e cercare di rendersi la vita più semplice, per quel poco di semplice che c'era nella sua vita. Ma erano mesi che la desiderava, che la osservava, la punzecchiava. Caterina ormai era diventata parte integrante della sua esistenza, e ripensò alle schermaglie un po' ingenue che facevano in questura o durante le indagini. Erano divertenti, alludevano, procrastinavano qualcosa che forse, un giorno, sarebbe accaduto. E adesso quel qualcosa era successo. Ne sarebbe nato un rapporto o stava correndo troppo? Non ce la poteva fare. Non era per lui. All'altezza di Roncobilaccio pensò a Marina al suo sorriso e alla sua malinconia. S'era chiuso a chiave in quel ricordo. Ma s'era scordato una finestra aperta, e Caterina s'era affacciata. «Lupa, che cazzo devo fare?». Quella non aveva neanche tirato su le orecchie. Era una pompa di benzina con solo due colonnine per il rifornimento e senza il self-service. Pompe bianche, così si chiamano quelle che non appartengono ai grandi gruppi petroliferi e che applicano prezzi più vantaggiosi. La Range Rover di Brizio era parcheggiata vicino al gabbiotto. Rocco scese insieme a Lupa che ne approfittò per fare i bisogni. L'amico spuntò da un cespuglio abbottonandosi la patta. «Ciao Rocco» e si avvicinò. «Mica me vorrai abbraccia'!» gli disse guardandogli le mani. «Me so' pulito su una foglia». «Lascia perdere. Te voglio bene uguale. Sta qui?». «No, c'è la moglie. Lui sta a casa. Dice che gli gira ancora la testa». «E dove abita?». «Lì» e indicò una casetta proprio dietro il distributore. I due amici arrivarono al cancelletto di ferro dipinto di rosso sul quale, con del fil di ferro arrugginito, era assicurata una vecchia cassetta delle poste. Lo varcarono. Subito tre cagnolini microscopici cominciarono ad abbaiare come indemoniati. Lupa si spaventò e decise di non entrare. «Dai Lupa, non ti fanno niente» fece Rocco, ma la cucciola temeva il piccolo branco che intanto si era avvicinato all'intrusa. Cominciarono ad annusarla, mentre la lupetta teneva la coda in mezzo alle zampe. Poi fu un attimo e partirono tutti e quattro di corsa nel giardino a giocare. «Si stanno simpatici» commentò Brizio. La casa era senza intonaco. Davanti alla porta d'ingresso c'era un patio coperto con un bandone ondulato verde trasparente. Dietro la costruzione si intravedeva una vecchia recinzione bassa e malandata e un chiocciare sommesso denunciava la presenza di qualche gallina. La porta era screpolata, sulla destra un pulsante di plastica riportava il cognome del proprietario. Suonarono e poco dopo il portoncino si aprì. «Chi siete?». «Ermanno Pioli?». «Sono io». Superava di poco il metro e sessanta e aveva i capelli bianchi. La barba ispida, era in pantaloni di pigiama e canottiera di cotone. Portava delle vecchie ciabatte di cuoio screpolato. Un enorme porro sulla guancia destra. «Voi chi siete?». «Vicequestore Schiavone, polizia di Stato». L'uomo annuì e li fece entrare. «È per la rapina, vero?». «Esatto» fece Brizio entrando dietro a Rocco. La casa puzzava di umidità. Il piccolo ingresso aveva la carta da parati rossa staccata in più punti. «Venite, venite, accomodatevi» e fece strada. Entrarono nel salone. Era pulito come un bisturi. Sui divani a fiori i cuscini erano ancora protetti dal cellophane, sul tavolo con le zampe di legno il piano di marmo era lucido e specchiato. Il trumò, stessa pietra del tavolo, sopportava un'enorme specchiera con la cornice dorata. Al centro un vaso di peltro con delle rose di plastica. «Prego, sedetevi pure» e il padrone di casa accennò alle sedie, non ai divani, che evidentemente stavano lì solo per bellezza. «Vi posso offri' qualcosa?». «Dell'acqua» fece Rocco. «A me niente» rispose Brizio. «Ecco... torno subito» e l'omino sparì dietro una porta. Il muro era dipinto di un azzurro spento. Alle pareti coperchi di scatole di cioccolatini incorniciati con delle vedute del Bel Paese. C'era il golfo di Napoli, le tre cime di Lavaredo, il Canal Grande. Una Madonna postmoderna sbalzata in argento troneggiava sul muro sopra il televisore. Ermanno tornò con una bottiglia di vetro decorata con delle strisce arancioni e un bicchiere. «Ecco, mannaggia quella minerale l'ho finita». «Fa niente, signor Pioli» disse Rocco. «Mo' è passato qualche giorno. Era martedì scorso. Quello è sceso, ha fatto il pieno, io mi sono girato per rimettere a posto la pompa e poi un dolore alla testa. Sono cascato per terra e ho visto che teneva una pistola in mano. M'ha dato un'altra zaccagnata e poi... tutto nero. Quando mi sono svegliato stavo in ospedale». «Senta, se io le mostro una fotografia, lei è in grado di riconoscere il bandito?». «Teneva degli occhiali da sole e un cappello di lana in testa, tanto che io m'ero detto: ma che ci fa questo con il cappello di lana? Siamo a giugno!». Rocco si mise la mano in tasca. Prese il portafogli e lo aprì. «Te lo porti tipo santino?» fece Brizio. «Già» confermò Rocco sorridendo e tirò fuori la fotografia di Enzo Baiocchi. «È di qualche anno fa» disse a Ermanno passandogliela. Quello la prese e si mise a guardarla. «È difficile» fece, «gli occhi non glieli ho visti...». «Aspetti...». Rocco prese una penna dal taschino. Poi disegnò due lenti sulla foto che restituì a Ermanno. L'uomo sorrise. «Mo' va meglio» disse divertito mostrando qualche dente mancante nella chiostra. «Io direi... sì, potrebbe essere lui. Potrebbe, sì. Però... no, no, potrebbe essere lui». «Sua moglie ha visto niente?». «No, mia moglie stava in casa». «Quanti soldi ha rubato?». «Io direi qualche duecento euro...». «E 'ndo ce va?» fece Brizio. «Già. Quindi ha ripetuto l'impresa, Brizio, te lo dico io». L'aveva ripetuta. Stavolta Alfredo De Silvestri ci aveva impiegato meno di mezz'ora cercando tra le denunce di furti a distributori di benzina sulla strada statale 3. Era avvenuto a Scheggia, poco distante dal confine marchigiano. «Sta lasciando le mollichelle» fece Brizio appoggiato al suo Range Rover, «tipo quello delle favole, come si chiama...». «Pollicino» rispose Rocco. «Che ore so'?». «Le sei» fece Brizio. «Che facciamo?». «Tu tornatene a Roma e avverti Furio». «E tu dove vai?». «A fare un controllo. A Velletri». «A Velletri?». Un'altra ora di macchina. Lottava con le palpebre che volevano chiudersi mentre la schiena urlava di dolore. Il sedere cominciava a prendere la forma della poltrona della Volvo. Quando arrivò in vista del carcere il sole stava tramontando. Parcheggiò, fece scendere Lupa per i bisogni, poi la lasciò in auto con i finestrini un po' aperti per l'aria. «Torno subito, fai la brava» le disse. E si avvicinò alla casa circondariale. «Chi si rivede, come va, Rocco?». Francesco Selva, il direttore del carcere, nonostante l'ora era ancora al lavoro ma sembrava fosse appena arrivato. Fresco come una rosa, abbronzato e con i capelli in perfetto ordine profumava di dopobarba. Si strinsero la mano. «Mi dicevi, allora, ancora Enzo Baiocchi?». «Sempre lui». «Quel figlio di buona donna». Si aggiustò il ciuffo con un rapido gesto della mano. «E cosa posso fare per te?». «Mi serve sapere se ricordi qualcuno che lui frequentasse assiduamente». «Qui dentro? Se c'è qualcuno che lo sa è un secondino, Luciano Domizi. È quasi in pensione, ma dei nostri ospiti conosce vita morte e miracoli». «Dove lo posso trovare?». «Vieni con me. È giù in ufficio. Senti, è un po' burbero, ma è una brava persona. Trent'anni qui dentro, non gliene puoi fare una colpa». Gli uffici erano deserti. In una stanza piccola e con le pareti screpolate sedeva un uomo con un paio di baffi a manubrio e i capelli tagliati a spazzola. Stava rimettendo a posto una pila di documenti. Quando Selva e Rocco entrarono non alzò gli occhi. «Lucia', c'è il vicequestore Schiavone, vorrebbe parlare con te». «Ecco, un attimo, io due mani ho!». Sbatté la pila di carte sul tavolo e le infilò in una cartellina. Poi finalmente sollevò lo sguardo. «Te sei Schiavone, quello che quel figlio di puttana di Baiocchi voleva ammazzare?». «In persona». Luciano guardò l'orologio appeso alla parete. «Io me ne devo andare a casa. Facciamo una cosa di giorno». «Rapidissimo. Senta Luciano...». «Diamoci del tu. Non ci riesco con il lei. Mica sono un signorino come voi che ha fatto le scuole. Io faccio lavorare queste» e alzò le mani tozze e scure. Selva sorrise, abituato alle reazioni poco diplomatiche del suo uomo. «Va bene, Lucia'. Dammi una mano. Qui dentro, quando ospitavate Baiocchi, c'era qualcuno con cui lui legava?». Luciano si alzò dalla sedia. Era parecchio in sovrappeso e quel semplice movimento gli causò un po' di fiatone. «Ce n'erano un paio. Uno sta ancora dentro, Damiano Mezzi, ti ricordi diretto'?». Selva annuì. «Sì, deve fare parecchi anni. È un recidivo. Rapine, furti e omicidio colposo. Buona condotta, dà una mano in carcere». Rocco si passò la mano sul mento. «L'altro?». «L'altro è uscito da un sei mesi. Mino Coppetti. Ha girato le carceri di mezza Italia. Dice che a Velletri si trovava bene» e scoppiò a ridere. «Io non mi ci abituo a questa cosa. I carcerati ci hanno scambiato per un albergo. Danno i voti al vitto, alle stanze, ai cessi. Io gli sparerei in bocca a tutti, 'sti fiji de 'na mignotta». «Coppetti. Che specialità?». «Falsario. Ma mettici pure un paio di rapine, tanto per restare allenato». «E con questo Coppetti, erano molto amici?». «Stavano sempre culo e camicia. Io dico che qualche mignottata fuori l'hanno pure fatta insieme». «Ed è uscito». «E te l'ho detto, no? Che sei sordo? Sei mesi fa. Ha scontato». «Ti ricordi di dov'è 'sto Coppetti?» chiese Rocco a Selva, parlare con il secondino cominciava a dargli sui nervi. «No, io no». «E chiedi a me, no? Che è, te faccio schifo?» intervenne Luciano. «Cividale del Friuli. Oh, quando parlava al telefono con la madre non si capiva niente. Parlano strano da quelle parti». «Parlano friulano, Luciano» lo informò il direttore. «Si può avere l'indirizzo di questo tipo?». «Te lo rimedio io, Rocco» fece Selva. «Grazie, Luciano, sei stato molto utile». «Prego, direttore. Ah, Schiavone, se rimetti le mani su Enzo, vedi di non portarlo qui». «Fosse pe me lo lascerei 'ndo lo trovo» rispose Rocco. «Mi piaci, vicequestore» e finalmente sorrise. È facile perdere le cattive abitudini, e Rocco dopo mesi di Aosta aveva perso quella del traffico congestionato. Tornava verso Roma stretto in mezzo ad autovetture e luci posteriori rosse. Era dentro un enorme blocco di acciaio, plastica e vetro che si muoveva di mezzo metro ogni dieci minuti. Gli era tornato alla mente più di una volta Bernardo Valenti, e la faccia con il naso schiacciato di Carlo e Paolo. Gli scottava ancora sulla pelle, ma il fatto che lui all'omicida ci fosse arrivato nonostante i depistaggi era una vittoria, anche se tutta l'indagine aveva lasciato un livido che ci avrebbe messo del tempo a riassorbirsi. Fermo in mezzo all'imbottigliamento, mentre la radio mandava El día que me quieras, un tango argentino, come dalle acque della Dora riaffiorò nella mente il cadavere di Juana. Accanto spuntò quello di Castel di Decima, e l'ultima telefonata fatta con il vice Bonanni. Honduras. Perché gli era familiare? L'immagine arrivò come uno schiaffo. Le statuine precolombiane! Quando anni prima aveva fermato il traffico di droga nel quale c'era in mezzo il bastardo di Luigi Baiocchi, al porto di Civitavecchia, la cocaina era stata trasformata in finte statuine precolombiane provenienti proprio dall'Honduras. Una coincidenza? Poteva essere. All'epoca arrestarono sì il braccio romano di quel traffico, ma non riuscirono mai a risalire ai veri fornitori di droga. C'entrava quel nuovo cadavere ritrovato a Castel di Decima? Perché? E se nell'omicidio c'era di mezzo Enzo Baiocchi, che ancora latitante stava sicuramente cercando la fuga, c'era un anello che legava l'omicidio a quel vecchio traffico di cocaina? Fermo sulla provinciale all'altezza di Ciampino stava pensando seriamente di fermarsi al primo albergo che gli capitava, tanto di andare a Monteverde non era aria. Suonò il cellulare. Conosceva quel numero. Inserì il vivavoce dell'auto. «Ciao Caterina...». «Ciao Rocco. Dove sei?». «Sto andando a Varese». «Ti cercava il questore. Gli ho detto che ti eri preso una pausa. Non ha fatto una piega». «Meglio così, meno lo vedo e meglio sto». «Avrei voglia di parlarti...». «Sei a casa?». «Sì». «Però al telefono non mi piace». «Neanche a me. Baldi sta cercando una piccola vendetta, lo sai?». «E cioè?». «S'è messo a indagare sui conti di quel Berthod, il padrone di casa col negozio». «Brucia anche a lui». «Vedrai, ti chiamerà». «L'ha già fatto, ma non mi va di sentirlo». «Torni presto?». «Non lo so. Credo di sì». «Che c'è di bello a Varese?». «Boh, mai stato». «Ti raggiungo?». «Lascia perdere Caterina...». Ci fu un silenzio. «Allora ci vediamo appena torni. Lupa sta bene?». «Poveraccia. Dorme sul sedile di dietro». «Ti abbraccio Rocco» e chiuse la telefonata. All'altezza del raccordo anulare il traffico s'era sciolto, e preferì proseguire per Trastevere e trovare un albergo lì. Dopo la doccia e la pappa a Lupa presa da un droghiere che aveva i prezzi di un negozio d'alta moda, andò all'appuntamento con gli amici. Lo aspettavano seduti a piazza Sant'Egidio. Forse perché la scottatura era ancora fresca, raccontò tutta la storia di Juana Perez. Impiegarono due Ceres a testa prima che finisse. L'unico commento lo fece Furio: «Fiji de 'na mignotta...», poi ordinarono la terza consumazione. «Da dove cominciamo?» chiese Brizio sorseggiando la birra. «C'è un tipo che frequentava Enzo. Tale Coppetti, uno di Cividale del Friuli». «Perché da lui?». «Perché se ci abbiamo visto giusto è laggiù che sta andando. Da lì la Slovenia è a un passo, e quell'amico magari può dargli una mano coi documenti». «E se abbiamo sbagliato tutto il ragionamento?». «Non lo pigliamo più» fece Rocco. «Del cadavere, quello senza nome, s'è saputo qualcosa?». «Addosso aveva una cinta fatta in Honduras». I due amici lo guardarono interrogativi. «E io ci ho pensato a lungo. E mi sono ricordato del 2007, la storia della coca e di quel pezzo di merda di Luigi Baiocchi...». «Je possa brucia' il culo ovunque si trovi» disse Brizio alzando il bicchiere. «È così. Be', c'erano quelle statuine fatte di droga, vi ricordate? Venivano dall'Honduras». «Me pare un po' pochino» obiettò Furio. «Che vuoi fare?». «Un salto all'ambasciata tanto per capire. Poi ci diamo una mossa e puntiamo su Cividale». Furio sbuffò, Brizio invece si morse le labbra. «Ce l'abbiamo un indirizzo?». «Al carcere me ne hanno dato uno vecchio, della madre, che però risulta morta da tempo. Voi sentite in giro. 'Sto Coppetti era un mezzo falsario. Secondo me in città lo conoscono pure. E comunque vedo con gli amici al commissariato se mi possono dare una mano». «Vabbè, annamo a magna' che ci ho lo stomaco che parla» propose Furio alzandosi. Rocco lasciò i soldi sul tavolino, insieme si incamminarono verso il loro ristorante. «E comunque Seba continua a tenere il telefono spento» fece Brizio. «Già. Ormai ho perso le speranze di sentirlo...». «E a proposito di cellulari» disse Furio accendendosi una sigaretta, «a me ancora non torna il fatto che il morto aveva il tuo numero in tasca...». La camera era piccola, poco pulita e piena di brutti quadri. Rocco li rigirò uno per uno, altrimenti non avrebbe preso sonno. Chiuse i battenti salutando Roma e Trastevere, poi se ne andò in bagno a lavarsi i denti. Fu nello specchio che la vide. Era lontana, il viso stanco e appannato. Non lo guardava, se ne stava di profilo come se nell'altra stanza ci fosse qualcosa di più interessante. «Non ci torno Marina. Non ci sto bene. Preferisco starmene in albergo. Stavo pensando seriamente di venderla. Vale ancora un sacco di soldi, semmai mi trovo qualcos'altro proprio qui a Trastevere. A me basta un buco». Il viso scomparve. Rocco sputò il dentifricio nel lavandino, poi si sciacquò la bocca. L'acqua di Roma sapeva di calcio. Tornò in camera da letto. Eppure era sicuro di aver chiuso le imposte. S'erano riaperte, e si era aperta anche la finestra. Si commosse quando nel cielo vide solo una stella farsi spazio fra due nuvole sporche, forte e luminosa. La luna era un filo d'argento. «Buonanotte pure a te amore mio». L'addetto stampa arrivò trafelato. Un vestito di lino blu impeccabile, fazzoletto da taschino in tinta con la cravatta, sembrava pronto per un matrimonio. «Buongiorno, sono Alejandro Giménez» e gli strinse la mano. «Che cosa posso fare per lei?». Il suo italiano era perfetto. «Chiedere al Comune qualche parcheggio per i visitatori» fece Rocco. «Qui è un vero incubo». Giménez sorrise. «La fortuna di lavorare in un paese straniero, l'unica mi creda, è che almeno abbiamo il posto riservato. Gradisce un caffè, un tè?» e con il braccio destro indicò una scalinata che saliva al piano superiore. Una segretaria con un vestito rosso e i capelli ricci, seduta a un desk proprio sotto le scale, sorrise. «Sono a posto, grazie». «Allora...» l'addetto si strofinò le mani, «rimaniamo qui nell'ingresso o preferisce venire nel mio ufficio?». «Stiamo pure qui». «Almeno usufruiamo di quelle comodissime Frau» e fece strada in un angolo del salone d'ingresso dove, accanto a una finestra, due poltroncine rosse e un tavolino basso prendevano un po' di luce. «Al telefono mi diceva della cinta...». «Esatto. Può aiutarmi? So che quei fabbricanti di pellami...». «Mario's Pura Piel. I migliori del nostro paese». «Esatto, possono aver prodotto un regalo per voi. Una cinta». «Guardi, mi sono andato a informare. Effettivamente due anni fa l'ambasciata ne ha ordinate un certo numero. E le ha usate per dei regali». «È troppo chiedere a chi?». Alejandro allargò le braccia. «Non è troppo ma è molto complesso. Si tratta di un regalo minore. Credo, ma ne sono abbastanza sicuro, che siano state usate per persone che collaborano con l'ambasciata, magari qualche giornalista, qualche fornitore di piante e fiori, oppure qualcuno del catering». Il vicequestore si mise la mano in tasca. Tirò fuori la foto del cadavere senza nome trovato a Castel di Decima. «Lei quest'uomo non l'ha mai visto?». «Me l'hanno già mostrata, dottor Schiavone. No, non so chi sia, e non l'ho mai visto. Era lui ad avere la cinta, vero?». «Esatto». «Ma sa cosa? Non è detto che sia una persona del mio paese. Sa come succede coi regali? Molti li riciclano». «È vero. Dalle parti di Pisa sotto Natale si tirano fuori i regali peggiori e si mettono in palio. La chiamano la tombola dei troiai». Alejandro scoppiò a ridere. «Ma non è il caso della cinta. Mario's Pura Piel fa cose eccellenti». «Non lo metto in dubbio. Però, certo, qualcuno potrebbe aver utilizzato il regalo per sdebitarsi, insomma come diceva lei, riciclarlo. Ecco, mi viene in mente una cosa. Secondo lei, quanto può costare una cinta come quella?». «Guardi, sono fatte a mano una per una. Anche se simili, sono tutti pezzi unici. Una quarantina di euro?». «Me lo chiede?». «No, non glielo chiedo. Direi una quarantina di euro». «Allora mi viene in mente che, diciamo una persona non molto ricca, difficilmente riciclerebbe un regalo simile. Può essere?». «A questo non avevo pensato. Sì, può essere». «Mi segua nel ragionamento. Io ricevo dall'ambasciata questo regalo, non so che farmene perché di soldi ne ho parecchi, allora lo mollo a una persona che può apprezzare. Nel nostro caso il cadavere. Perché, ammettiamo che sia onduregno, non può averla comprata al negozio. Lei mi ha detto che Mario's Pura Piel non produce cinte». «No, esatto, il ragionamento è plausibile». «E allora, sempre ragionando per assurdo, lei mi può trovare qualcuno che ha ricevuto quella cinta e che magari può averla in qualche modo riciclata?». «Aspetti» si alzò di scatto. Non era felice di quell'incombenza, ma lo nascondeva benissimo. «Carmen!» chiamò. «La nostra memoria storica!». La segretaria col vestito rosso si avvicinò. L'addetto stampa si mise a parlottare con lei. Quella non perse il sorriso e si inerpicò sulla scala. Alejandro tornò da Rocco. «Se abbiamo qualcosa, Carmen la trova di sicuro». «Lei nel 2007 era qui?» chiese Rocco. «No, ero al consolato a Torino. Perché?». «Niente. Non ricordavo di averla conosciuta. Io venni per un brutto affare di droga, al porto di Civitavecchia, contattai l'ambasciata. Ma non risolvemmo nulla». «Un brutto affare di droga?». «Già. Fermammo un carico proveniente dal suo paese, avevano trasformato la coca in alcune statuine finte precolombiane». Il viso di Alejandro Giménez si oscurò. «Ho detto qualcosa che non va?». «No... però si renderà conto che non è piacevole sentire una cosa simile. Un traffico di cocaina dal Sudamerica o dall'Honduras?». «Dal suo paese». «E come si possono controllare tutti i cittadini dell'Honduras, o tutti gli italiani che abbiano a che fare col mio paese. Solo a Roma sono migliaia». «Lo so. Infatti ci fermammo. All'epoca parlammo con un suo collega. Meglio, era uno che stava nel gabinetto del viceconsole, mi pare di ricordare. Aspetti, il nome... Juan Gonzalez qualcosa...». Alejandro schioccò la lingua. «Juan Gonzalez Barrio. Non era nel gabinetto del viceconsole. Teneva i rapporti con le piccole e medie imprese». «È ancora qui?». «Per carità!» disse alzando le mani. «No, non è più qui. Non c'è più da anni». «Chissà dove è finito». «Non lo so. Terminò il suo rapporto con noi e, credo, tornò in Honduras. Ricordo che aveva una moglie molto bella, ma poco altro. E a proposito di belle donne» e si voltò. Carmen scendeva dalle scale. In mano aveva un plico non molto spesso. Lo consegnò a Giménez che le era andato incontro. «Ecco qua. Vede ad avere una grande segretaria? Tutto al suo posto. Allora, l'ordine era del... del 2010, tre anni fa, mi correggo. Ce n'era uno precedente, di cinte intendo, ma risale al 2001. Troppo lontano, no? Sì, la fabbrica di pellami ci rifornì di 35 cinte, 150 portafogli e 150 portadocumenti. Solo che io non ho i destinatari di quei regali». Abbassò il foglio. «Mi dispiace». Rocco guardò la segretaria che se n'era tornata alla sua scrivania sotto la scalinata. I loro occhi si incrociarono e quella non li abbassò. «Già. Era un tentativo». «Le assicuro che se dovessi trovare qualche traccia...». «Me lo farà sapere. O può comunicare col mio collega, vicequestore Bonanni. Anche se qualcosa mi dice che io e lei non ci sentiremo più». «Non sia così pessimista». «Sono realista, dottor Giménez». Il vicequestore si alzò allungando una mano. «È stato un vero piacere». «Piacere mio, dottor Schiavone. E ci pensi a visitare il nostro paese. È meraviglioso!». «Non lo metto in dubbio. Soprattutto se le donne sono tutte come Carmen!». La donna sorrise arrossendo. Non è che Rocco Schiavone fosse un profondo conoscitore dell'animo femminile, ma di quello umano sì. Sapeva leggere gli occhi e i movimenti del corpo. Da come una persona camminava, da uno sguardo, perfino dal modo in cui si aggiustava i capelli a volte traeva qualche conclusione utile. Quindi le diede dieci minuti, il tempo sufficiente per prendere una decisione e agire di conseguenza. Si era fermato al bar davanti all'ambasciata, seduto a un tavolino di ferro soffocato dalle macchine parcheggiate. Aveva preso un caffè e teneva d'occhio la finestra del primo piano. Carmen si era affacciata dopo tre minuti, e dopo sette la vide uscire dal portone. Mento sollevato, capelli al vento, forse si era anche data un'aggiustatina al trucco. Attraversò la strada e puntò verso di lui. Entrò nel bar salutandolo con un piccolo gesto del capo e sorridendo. Era la sua arma preferita, il sorriso, la pelle del viso si stendeva e gli occhi, grazie al contrasto con la dentatura perfetta, risultavano più grandi e luminosi. Rocco si alzò e la seguì. Passò alla cassa a pagare l'ordinazione. Poi si appoggiò al bancone, un piede intrecciato all'altro, in attesa. «Buongiorno Carmen» disse Rocco. «Buongiorno dottor Schiavone» rispose puntandogli addosso i suoi occhi. «È una giornata magnifica, non trova?». Carmen si mise a ridere. Fuori nuvole acide ingrigivano la città. «Ne ho viste di migliori». «Lei è la memoria storica dell'ambasciata, così ho sentito». La donna non rispose. Mise le mani nella pochette nera che portava sotto il braccio e tirò fuori un foglio. «Queste sono le persone che hanno ricevuto la cinta in regalo. Magari le può tornare utile». Rocco prese la carta senza guardarla. «Se il signor Alejandro dovesse accorgersene...». «Non è cattivo. Ma lui per poter dare le informazioni deve passare attraverso tante di quelle persone che ci può fare notte». «Lei non è italiana. Ma parla benissimo». «Sono qui da venti anni» rispose Carmen. «Un tempo abbastanza lungo per imparare una lingua». «Perché mi sta aiutando?». Con un sorso finì il succo e poi poggiò il bicchiere. «Non se ne abbia a male se non rispondo. Diciamo che ci sono brutte persone con cui ho lavorato, che mi hanno anche fatto del male, e non vedo l'ora che ne paghino le conseguenze». «Può essere più precisa?». «No. Anche perché io non le ho dato niente e niente le ho detto». «Perché si fida di me?». «Ha gli occhi giusti. Hasta luego, dottor Schiavone» girò i tacchi e uscì dal bar. Rocco prese il foglio, lo aprì. C'era una lista con una decina di nomi. Alzò gli occhi al cielo. Gli toccava cercare il solito ago nel pagliaio. Erano tutti uomini dell'ambasciata, il vecchio addetto stampa, un fotografo, due segretari di gabinetto. Ma il decimo nome gli fece rizzare i peli sul collo. Era quello di Juan Gonzalez Barrio. S'era perso e ritrovato a piazza del Popolo con il foglio in mano, come un rabdomante che cercasse acqua. E dire che la fontana di Della Porta acqua ne aveva da affogarci. Perché Barrio aveva richiesto quella cinta? Era poi quella che aveva indosso il cadavere di Castel di Decima? Non ci arrivava, non riusciva a fermare i pensieri, a trovare un solo binario dove convogliarli. I ricordi si confondevano con ricostruzioni azzardate. Cercò di richiamare alla memoria le indagini di quella partita di coca che aveva intercettato a Civitavecchia, del mobilificio, dei nigeriani, ma non ne usciva. I visi si sovrapponevano, parole, rumori. Chiuse gli occhi e cercò di isolarsi dalla città, dal frastuono delle auto e della gente che parlava. Dov'era la chiave? E dov'era Sebastiano? Poi si ricordò che aveva chiuso Lupa in macchina. Era ora di farle prendere aria. Con il cane al guinzaglio, risaliva per Villa Borghese. Il cielo era sempre grigio e i taxi scendevano come bolidi. Il Muro Torto invece era un blocco di auto ferme. Mentre Lupa assaggiava l'erba lui si perse a guardare le macchine incolonnate. Ce n'era una che attirò la sua attenzione. Una vecchia auto verdina che scatarrava fumo dal tubo di scappamento. Gli ricordò quella di Michela Gambino, il sostituto con la testa piena di complotti. Ci pensò qualche secondo e gli venne l'idea. «Dica pure...» era stato venti minuti in fila al negozio e finalmente era il suo turno. «Mi serve un cellulare. Semplice. Con una carta e un nuovo numero». «Ha preferenze?». «Quello che costa di meno» rispose Rocco. Il commesso aprì una piccola vetrina. Tirò fuori una specie di citofono nero. «Ecco qui. È un'ottima marca, niente di che, però...». «Lo prendo. Ci metta un numero nuovo». «Vuole fare un contratto o una ricaricabile?». «Ricaricabile, grazie». Il commesso aprì un cassetto e si mise a trafficare con bustine e batterie. Lupa se ne stava stesa a terra a godersi la frescura del marmo. Non sapeva dire se la paranoia di Michela Gambino l'avesse contagiato, ma lasciare il suo vecchio cellulare in albergo e muoversi con quello nuovo che nessuno conosceva gli sembrò una mossa intelligente. «Ecco, ci sono 10 euro di ricarica e...». «Me ne metta 100!». Il ragazzo lo guardò stranito. «Spende più di ricarica che di cellulare?». Rocco annuì. «Già... ci posso vedere le mail?». «Certamente. Glielo imposto?». Brizio alla guida superava allegramente i 130 chilometri orari, limite di velocità consentito sulle autostrade. Rocco seduto accanto guardava fuori dal finestrino. Furio, sul sedile posteriore carezzava il cane. «Non è male 'sta Volvo, Rocco» fece osservando il cruscotto. Avevano preferito la macchina di Schiavone per via dei peli che Lupa avrebbe lasciato sui sedili nuovi della Rover di Brizio. «Certo la Range è un'altra storia, ma va benone». Nessuno gli rispose. «Dateme chiacchiera sennò m'addormo!». «Che sei Nino Manfredi?» fece Furio. «Quanto ce vorrà fino in Friuli?». «Boh... Rocco, che è 'sta faccia lunga?» gli domandò Brizio. «A che stai a pensa'?». «Devo dirvi una cosa della lista che m'hanno dato all'ambasciata». «La storia della cinta? Mbè?». «Una delle persone che ha avuto quel regalo è Barrio». Brizio guardò nello specchietto cercando lo sguardo di Furio che teneva gli occhi puntati sul poggiatesta di Rocco. «Chi è Barrio?» chiese. «Lavorava all'ambasciata dell'Honduras, addetto ai rapporti con le piccole e medie imprese». «Ancora non ci arrivo» fece Brizio. «Io lo contattai tanti anni fa, nel 2007, e mi sembrò una persona a posto. Un po' arrogante, ma a posto. Però la faccia di Alejandro Giménez quando ho pronunciato il suo nome è diventata grigia come 'st'autostrada». «Vabbè, magari je stava sur cazzo» concluse Brizio. «Sarà... però 'sto tipo è stato allontanato dall'ambasciata». «Avrà preso bustarelle, oppure scroccava». «Oppure» disse Rocco «acchittava cose sporche. All'ambasciata in qualche modo è arrivata una voce e l'hanno cacciato». «Boh» fece Furio che aveva afferrato le orecchie di Lupa per massaggiarle. «Me pare che stai a anda' troppo veloce». «Ora, dico io, uno che ha a che fare con le piccole e medie imprese italiane ha un sacco di soldi che gli passano davanti, questo è sicuro. Che ci fa con una cinta?». «Se doveva sdebita' con qualche sfigato» concluse Furio. «Pure secondo me». Poi Rocco afferrò il cellulare nuovo. Brizio lo guardò appena e si sentì in dovere di commentarlo. «Ammazza, te sei fatto un capolavoro... quello te dura sì e no due giorni». «E quello deve durare... come si fa a nascondere il numero?». «Cancelletto poi 31 poi cancelletto. Ma perché?» disse Furio. «Meno se sa il numero e meglio è». Rocco eseguì, compose un numero e poi attese. «Che è tutto 'sto mistero?» chiese Brizio. «Ci hai paura che te spiano?». «Poi te lo dico Furio... Pronto? Uccio? Sono Rocco!». «Uelà, m'hai letto nel pensiero. Chiami dall'ufficio che non vedo il numero?». «Già. Dimmi se ci sono novità». «E te lo dico sì. Sono ore che ti chiamo al cellulare ma ce l'hai sempre staccato. Ho una bella notizia, amico mio. Riguarda il cadavere ritrovato a Castel di Decima. Ho avuto l'intuizione, ci ho provato e ho fatto tombola». «E dimmi un po'?». Furio e Brizio cercavano di sentire. Rocco premiò i loro sforzi mettendo in vivavoce. «Allora ho mandato le impronte digitali alla polizia di Tegucigalpa». «E 'ndo cazzo sta?» chiese Brizio sottovoce. «Honduras» sussurrò Rocco. «Vabbè Uccio, e allora?». «Allora, ci hanno messo qualche giorno ma c'è stato il riscontro. Diciassette punti in comune con Ruben Montoro. Ti mando la foto con un sms così lo vedi. È lui, Rocco!». «Ruben Montoro? E chi è?». «Padre siciliano madre onduregna. Nel suo paese era ricercato per un sacco di reati. Credo fossero felici di saperlo cadavere qui da noi». «Ruben Montoro... e da quanto era in Italia?». «Stanno cercando, ma non ci sono tracce. Né quando è entrato e neppure se e quando è uscito. Capace avesse la doppia cittadinanza, doppio passaporto, documenti falsi, io che ne so? Però se vuoi ti faccio mandare la sua bella cartellina per mail. Così ti leggi i precedenti. Sempre che tu abbia una mezzoretta a disposizione». «Va bene, mandami pure una mail con la foto di Ruben Montoro». «Qualsiasi novità mi faccio vivo! Ad maiora!» e attaccò. «Mi spieghi 'sta storia del cellulare?» chiese Brizio. «È per evitare che mi rintraccino». «Chi?». «Te l'ho detto: qualcuno mi ha messo gli occhi addosso. E meno sa meglio è...». «Daccelo pure a noi 'sto numero» fece Brizio, «se dovesse servi'. Oppure non te fidi?» e sorrise. La suoneria annunciò che era arrivata una mail. Uccio era stato di parola. La foto segnaletica di Ruben Montoro non lasciava alcun dubbio. «Cazzo, è lui! Il cadavere di Castel di Decima...» fece il vicequestore. «Cosa ha a che fare con Enzo Baiocchi?». «Questa è la domanda principale, e io una risposta comincio a vederla. Mo' fermati al prossimo benzinaio sennò me la faccio sotto!». «Figata!». Brizio per gli autogrill aveva un'attrazione irrefrenabile. Mentre Brizio svaligiava gli scaffali e Furio appoggiato al banco era in attesa di un caffè, Rocco si era messo accanto ai bagni dove il cellulare prendeva. «Carmen, sono Schiavone, si ricorda?». «Come potrei dimenticare. Posso esserle utile?». «Parecchio. Ruben Montoro, lo ha mai sentito nominare?». Ci fu un breve silenzio. «Sì, il nome non mi è nuovo. Lei è a Roma?». «No, sono in viaggio». «Mi chiami appena rientra, dobbiamo fare due chiacchiere ma non al telefono». «Sicuro. Mi dica solo se la strada è quella giusta». «Non al telefono» ripeté Carmen e chiuse la comunicazione. Rocco restò con il cellulare in mano. Brizio, carico di cioccolata, un pallone di peluche e altri giocattoli, si avvicinò. «Sono pronto, 'nnamo?». «Spiegami che ci devi fare col pallone di peluche. Guarda qui, pure un trenino di plastica». «Per il nipote di Stella. Ogni volta che mi vede mi dice: hai portato un regalo a Tommy? Ecco, gli ho preso pure il trenino. Bello, no?». «Saresti un bel padre» gli disse sorridendo. «Sei scemo? Io? Ma m'hai visto bene? No Rocco, certe cose è meglio che si chiudano». «Che vuoi dire?». «La storia del dna, me l'ha spiegata Stella. Io ci ho quello de papà, ti ricordi papà, no?». «Poco». «E certo, stava più a Regina che a casa. E pure nonno frequentava le stesse celle. E suo padre prima de lui. Insomma, 'na schiatta di gente marcia. Mo' con me basta, dico io. È ora di fermare la catena». «Sei pessimista. Potrebbe prendere da Stella». «Non credo. La mela marcia vince sempre su quelle sane, lo sai. Prova a metterle nello stesso cesto. Dopo du' giorni so' tutte fraciche. Col dna funziona uguale». «Te l'ha detto Stella?». «No, ce so' arrivato da solo». Arrivarono a Cividale che erano le otto di sera. Il paese sembrava deserto. Riuscirono a trovare posto in un albergo che accettava i cani. Era pulito e accogliente, sul bancone accanto alle chiavi c'era una quantità di dépliant che illustravano le bellezze della cittadina ma a nessuno dei tre venne in mente di prenderne uno. Rocco non si fermò. Lasciò gli amici a farsi una doccia e uscì dall'albergo. Da via San Lazzaro a piazza Armando Diaz erano poco più di cinque minuti, come il solerte uomo della reception gli spiegò dandogli in mano una cartina colorata della città. Passò sul ponte del Diavolo e si perse a guardare le acque del Natisone. Le case che spuntavano dalle rocce a picco erano un bello spettacolo. Peccato non ci fosse il sole, pensò. Ma di fare il turista non era il momento. Entrò nel commissariato. Un agente lo fermò. «Vicequestore Schiavone» disse Rocco. «Chi cerca?». «Vicequestore Cascone...». «È fortunato, è ancora in ufficio». Prese il telefono. «Dottore? C'è il vicequestore Schiavone... questura di...?» e guardò Rocco. «Aosta». «Aosta, la sta cercando... va bene, riferisco» abbassò la cornetta. «Scende subito». «Grazie mille». Neanche un minuto dopo vide apparire un modello di Giorgio Armani. Sotto i quarant'anni, alto, biondo, spalle larghe, un cardigan che teneva aperto per mostrare una maglietta di cotone a V con il ritratto di Jimi Hendrix. Sul petto ciondolavano delle collane fatte con qualche seme essiccato. Il passo agile, portava delle vecchie scarpe nere da ginnastica. Aveva la faccia stanca e gli occhi azzurri erano cerchiati di nero. L'unica nota stonata che lo teneva lontano dalle passerelle di Milano moda era il naso, che qualche cazzotto gli aveva spostato. «Augusto, piacere» e gli strinse la mano. «Rocco Schiavone...». «Com'è da queste parti?» gli chiese appoggiandogli la mano sulla spalla mentre si incamminavano verso l'uscita. «Dove andiamo?» gli chiese Rocco. «Non so te, ma io tengo una certa sete e sono stanco morto. A quest'ora una birra non me la toglie nessuno». Usciti in strada si diressero verso un bar che si apriva sulla piazza. «Senti Augusto, è una cosa facile. Sto cercando un tale Mino Coppetti, che è di queste parti». L'uomo strizzò gli occhi. «Aiutami un po'?». «Allora l'ultima residenza era la casa circondariale di Velletri. Precedenti di furto, falsificazioni, due tentati omicidi... anni 59...». Augusto si grattò la testa. Poi aprì la porta del bar e fece entrare Rocco. C'erano parecchi tavolini inchiodati alle pareti rivestite di legno e due ragazze sui venti anni che servivano al bancone. L'età media degli avventori non superava i trenta. Appena entrato Augusto salutò con il suo vocione: «Grazia e Stefania, Augusto è arrivato!». «E finalmente» rispose una delle due. «Birra per me e il mio amico» disse. Una musica da ascensore finto orientale ottundeva i timpani. «Vieni Rocco, andiamoci a sedere lì» e presero posto ad un tavolo in fondo al locale facendosi spazio fra ragazzi che bevevano liquidi colorati e parlavano ad alta voce. «Allora... Mino Coppetti. A me questo nome non mi dice niente. Hai una fotografia, qualcosa?». «No». «Allora ho l'asso nella manica» e gli fece l'occhiolino. Prese il cellulare e chiamò. «Uè, Dario, sono il vicequestore. Fammi un favore, vedimi in archivio se abbiamo qualcosa su Mino Coppetti... anni 59... sì ma ti devi sbrigare!». Si rimise il telefono in tasca. «È bello avere lo schiavo pronto. L'agente Dario Esposito. Sta qui da venti anni, è un'enciclopedia, tiene un quoziente di intelligenza bassino, ma a cercare è il numero uno. Pensa che se mi perdo qualcosa a casa lo chiamo e lui me la trova. Con uno così, che mi piglio moglie a fare? E dimmi un po', come si sta ad Aosta?». «Ce l'hai un'altra domanda?». «Qui sto in grazia di Dio. Mo' a luglio sono due anni e ti posso dire la verità? Non lo cambierei con nessun posto...». Una delle ragazze si avvicinò con le due birre. «Ecco qui. Ci volete le patatine?». «No Stefa', quali patatine. Magari porta Grazia e fateci un po' di compagnia!» e le sorrise. «E al banco chi resta? Ci vai tu?» si girò di scatto e tornò al lavoro. Sapeva che Augusto e Rocco le avrebbero guardato il sedere, cosa che fecero all'unisono. «Maronna...» bofonchiò Augusto. «Mo' capisci perché qui mi piace?». Bevve un sorso. «Ahhh, sulla birra li devi lasciare perdere. Scusa, per caso ti ritrovi un preservativo?». Rocco lo guardò serio. «Perché dovrei?». «Sennò mi tocca arrivare alla farmacia, e non ho moneta». «No Augusto, non ce l'ho un preservativo». «Che magari di aids non ne parlano più, ma quello ci sta sempre. Io dico, facciamola finita di dare retta solo ai giornalisti. Sono come le mode, no? Quando non va più un argomento, ecco che non lo vedi sui giornali. Ma l'argomento c'è ancora, dico bene? Il preservativo serve sempre. E non solo per le malattie, tipo l'epatite, ma se ti inguai? Tu figli ne hai?». «Non ne ho, Augusto». «E sei sposato?». «Augu', non so' venuto fino in culonia per farmi due chiacchiere con te. Io sto cercando una persona». «Già. Come si chiama?». «Mino Coppetti» disse Rocco spazientito. Il tempo di concentrazione del vicequestore Cascone era inferiore a quello di un bimbo di sette anni. «E tu sei sicuro che sia da queste parti?». «No». «Ma posso sapere perché lo stai cercando?». «Nemmeno». «È bello parlare con te» e alzò il bicchiere in un brindisi ideale. Rocco sospirò, poi si alzò mettendo venti euro sul tavolo. «Stammi bene, Augusto». «Be'? E mo' te ne vai?». «Non ho tempo, figlio mio. Salutami Grazia e Stefania e digli che la birra è un po' troppo amara» e lasciò il collega al tavolino con la faccia spaurita. Non tornò in albergo, ma in commissariato. Di nuovo dall'agente all'entrata. «Mi chiami l'agente Dario Esposito». Quello eseguì. Poi gli passò la cornetta. «Dica?». «Esposito, sono il vicequestore Schiavone, questura di Aosta. È a me che servono notizie». «Venga, dottor Schiavone, faccia il piacere. Sono qui appiccicato al computer e...». «Arrivo» restituì la cornetta all'agente. «Piano?». «Quale piano, dotto', sempre dritto...». Esposito lo aspettava davanti alla porta dell'ufficio. Aveva superato da un pezzo i cinquant'anni ma s'era tinto i capelli di un nero talmente esagerato che qualche ciocca dava sul blu cobalto. Le sopracciglia enormi e aggrottate, sulle guance un'ombra di grigio. «Dario Esposito, dottore» e inchinò appena la testa. «Il vicequestore mi ha detto che cerca Mino Coppetti, ma io l'unica traccia che ho di lui è la compagna. Cioè, c'è un vecchio indirizzo della madre che però è morta da anni. La casa se la sono ripresa i proprietari». Entrarono nella stanza. Era piccola con un solo tavolo e un computer. Sulla parete di destra correva un enorme schedario di ferro. L'angelo della giustizia incorniciato e due disegni di qualche nipotino. «La compagna invece, Mariella Carimini, abita ancora qui. A Bottenicco. Via dei Gelsi. Lei deve superare Villa dei Claricini e prendere la prima... o la seconda a destra». «Abita lì da sempre?». «La residenza ce l'ha dal '78». Rocco si mise una sigaretta in bocca senza accenderla. «E ci sono altri parenti? Suoi o della compagna?». «La compagna ha solo una sorella, Renata, ma non si parlano da anni. Quella è sposata con un ex dirigente delle poste, e poco hanno da spartire con Mariella e con quel malacarne del marito». «Mi dice qualcosa di Mino Coppetti?». «Certo. È un figlio di buona donna. Soprattutto era amico di due altri personaggi, Renato e Carmelo Sabatini. Due fratelli che trafficavano oltre confine». «Slovenia?». «Già». «Che fine hanno fatto 'sti due fratelli?». «Carmelo l'ho arrestato io nel 2010 e mi sa che sta a Poggioreale. Renato è morto due anni fa». «Esposito, lei è stato di grande aiuto». «Si figuri, dottore, dovere. Me la toglie una curiosità?». «Se lei ne toglie una a me». «E certo. Allora, io volevo solo sapere che ci fa un vicequestore di Aosta da queste parti». «Inseguo un criminale evaso di prigione tempo fa». «Coppetti?». «No, un suo amico. Ora tocca a lei. Perché il vicequestore dice che lei ha un quoziente intellettivo basso?». Il poliziotto sorrise. «Perché per parlare con lui mi devo abbassare. E siccome sempre a proposito di altezze, lui è più alto di me, sia in centimetri che di ruolo, la cosa forse gli dà fastidio, ma chissene fotte». «Lei mi farebbe comodo ad Aosta». «Dotto', io fra qualche anno torno ad Acciaroli. Ho pure la barca e arrivederci e grazie». Chiamò Alfredo De Silvestri mentre era a tavola con Furio e Brizio a fare colazione. «Se interessa, abbiamo un terzo furto presso un benzinaio. A Mortegliano, vicino Udine». «E questa è una bella notizia, Alfre'. Sempre stessa faccenda?». «Esatto. Il benzinaio è stato aggredito e stordito da un tizio. Dice che l'ha colpito con il calcio di una pistola. Dottore?». «Dimmi, Alfredo». «Ma lei chi sta cercando?». Rocco ingoiò un pezzo di cornetto. «Lascia perdere Alfredo, è una storia lunga». «Sempre a disposizione». Appoggiò il telefono sul tavolo e guardò i due amici. «Mortegliano». «Lontano da qui?» chiese Furio mentre Brizio s'era perso dietro i seni della cameriera. «No, e io dico che siamo sulla strada giusta». «E lo dico pure io. Annamo da 'sta Mariella, va'... Senti un po', Rocco, ci si può fidare delle guardie qui?». «Mi pare di sì». «E come la vuoi fare?». «Se 'sta tizia sa qualcosa del marito, sicuro che non ce la dice. Brizio!» richiamò l'attenzione dell'amico che subito distolse lo sguardo dalla ragazza. «E 'nnamo, Brizio. Devo chiamare Stella?». «No no, ascolto». «Come no» fece Furio. «Allora questa, se sa, non ci dirà niente. Quindi facciamo così. Voi imboccate dentro, io la faccio cacare sotto e voi vedete come reagisce». «Chiaro. Beviti 'sto caffè Brizio, che sono già le otto». C'erano tutti e sette i nani e pure Biancaneve, la fontana con un putto che pisciava acqua direttamente dal pisellino e due fenicotteri scoloriti accanto a un bambi senza una zampa. La villetta a due piani era rosa maiale, due colonne di finto marmo facevano la guardia alla porta di ingresso, protetta da una tettoia di vetro e ferro. L'alluminio anodizzato color ottone splendeva al sole. I fiori erano tutti morti. «Sì?» rispose una voce di donna al citofono. «Mariella Carimini?». «Sì?». «Vicequestore Schiavone. Apra!». Rocco passò il piccolo cancello di ferro nero e entrò nel giardino. Mentre superava i sette nani alla porta si affacciò la donna. Teneva un cucchiaio di legno in mano e lo stringeva come volesse spaccarlo. Aveva superato la sessantina. Portava una maglietta viola con degli strass che disegnavano la frase Best girl in town, i capelli corti e lisci biondo tinto e un paio di occhiali da vista argentati. Magra come un ramo, non sorrideva. Aveva l'aria preoccupata. «Dica pure, che succede?». «Succede che mi sono rotto i coglioni di cercare suo marito. Dov'è?». Gli occhi dietro le lenti diventarono ancora più grandi. «Io non lo so... e poi non è mio marito, non siamo sposati». «Ecchisenefrega» fece Rocco. «E allora non abita qui?». «No. Io e Mino ci siamo lasciati». «E io non ci credo. 'Ndo cazzo sta, signo'? Non ho fatto centinaia di chilometri per famme imbotti' de cazzate. Quand'è l'ultima volta che l'ha sentito?» gridò. «Non lo so... saranno mesi...». «E dov'era?». «A Napoli». Troppo pronta, troppo precipitosa, mentiva. «A Napoli? E che faceva a Napoli?». «Non lo so. Gliel'ho detto, io non lo vedo più». «Che palle» fece Rocco. «Allora suo marito s'è messo nei guai seri, lo sa?». Quella fece no con la testa. «Ora ne è a conoscenza. In più, signora Carimini, se lei sa dove si trova e me lo nasconde, finisce dentro e buttiamo pure la chiave». «Ma perché? Che ha fatto?». «Che gliene frega? Se le dico che voglio sapere dov'è, lei me lo dice senza se e senza ma. Sennò io comincio a sospettare che lei ne sappia di più, che non parla, che lo nasconde e allora andiamo dritti dritti nel penale. Mo' sono stato chiaro?». «Io le giuro che non lo so. Quel disgraziato...» unì le mani davanti al volto. «Senta, le prometto che appena ho qualche notizia la chiamo». «E chi ci crede?». «Lei, no?». «Appunto, io non ci credo. Torno con l'ordine del tribunale, le entro in casa, la metto sottosopra finché non trovo qualcosa. E allora la porto in catene fino in carcere. Glielo giuro. Ha fatto male a non collaborare» si girò e si avviò verso l'uscita. Poi davanti al bambi si fermò. «E butti 'sto coso de plastica. Non lo vede che ha una zampa di meno?» sbatté il cancello e sparì alla vista di Mariella Carimini che si precipitò in casa. Superò l'ingresso e andò in cucina. Dal ripiano di marmo nero prese il cellulare. «Pronto? Pronto Mino? Ti stanno cercando...» disse. «Polizia. Non ti muovere. Vengo io, sta' tranquillo... arrivo subito, dammi il tempo, santa pace!» chiuse la telefonata. Se ne andò in salone a infilarsi un golf di cotone. Non si accorse che la finestra sul retro era aperta. Uscì dal giardino e montò su una piccola auto giapponese. Rocco aspettava in macchina a duecento metri dal cancello. Brizio s'era acceso una sigaretta. Lupa scodinzolava. «Eccola!» fece Furio. Aspettarono che svoltasse l'angolo, poi il vicequestore mise in moto e la seguì. Attraversarono Cividale, poi la Mitsubishi puntò verso Ponte San Quirino. «Pure troppo facile» fece Furio. «Almeno accelerasse» disse Rocco tenendosi sempre a distanza con una macchina in mezzo per non dare troppo nell'occhio. «Se sta nascosto un motivo c'è» osservò Brizio. «Mo' basta sape' qual è». «Voi avete provato a chiamare Sebastiano?». «Ieri sera, sempre staccato. Tu Furio?». «Uguale» rispose giocherellando con una sveglia di ottone. «E quella?». «Carina no?». Rocco alzò gli occhi al cielo. «E vabbè, prima che la signora se ne accorge sai quanto ci mette». Superarono anche Ponte San Quirino. Brizio controllava la mappa sul cellulare. Arrivarono alla frazione di Clenia dove Mariella parcheggiò ed entrò in un minimarket. Rocco, a una cinquantina di metri, fermò l'auto. La videro uscire con una busta. Poi, a piedi, raggiungere il bancomat lì vicino. Neanche un minuto per tornare all'auto. Salì, rimise in moto, lo stesso fece Rocco. Lasciò il piccolo paese e prese una strada che si inerpicava in mezzo ai boschi. Era deserta, Rocco non poteva starle troppo attaccato. «Siamo su via del Klančič» fece Brizio. Furio guardava fuori dal finestrino. Erano in mezzo a un bosco, le fronde tagliavano la luce del sole che cercava di farsi largo fra rami e foglie. La strada attraversava la macchia con la sinuosità di un serpente. «Bello qui» fece, «vero Lupa? Te piacerebbe fa' una corsetta?». «Ma 'ndo cazzo stamo!» fu invece il commento di Brizio. Rocco continuava a tenere d'occhio la carreggiata. Dopo una curva c'era un rettilineo, ma la Mitsubishi della signora Carimini era sparita. «'Nd'è annata?». Rocco rallentò. Superarono una carraia in terra battuta sulla sinistra. Il vicequestore fermò l'auto. «Ha girato qui...» fece. «E noi? Seguiamo?». «No. Quante case vuoi che ci siano da 'ste parti? Dammi un po' la mappa, Brizio...». L'amico gli consegnò il cellulare. La strada partiva dal paesello, attraversava la collina per finire in un altro paese, San Pietro al Natisone. «L'ha scelto bene il rifugio, con due vie d'uscita, ma noi siamo in tre. Sapete che facciamo? Uno si va a mettere dalla parte di San Pietro, gli altri due invece a Clenia. Così qualunque strada prendano, lo sappiamo». «Ma scusa, non ci conviene muoverci ora?». «No, aspettiamo e vediamo se ci sono movimenti» fece Rocco. «Se qualcuno cerca di andarsene lo seguiamo». «Vabbè, portace a 'sto paese, io e Furio ci mettiamo lì». Rocco fece retromarcia e veloce accompagnò i due amici a Clenia. Trovarono un bar e si piazzarono a guardia della strada. «Ao', cellulari sempre accesi!» si raccomandò Rocco. Poi sgommando riprese via del Klančič. Riaffrontò le curve in mezzo al bosco, superò la stradina sterrata dove la macchina di Mariella era sparita e ridiscese fino a San Pietro al Natisone. Fermò l'auto davanti a un bel ristorante, proprio di fronte a via del Klančič. Scese con Lupa al guinzaglio e la vide. Era nascosta dietro un enorme masso ricoperto di muschio. La moto di Sebastiano. Si guardò intorno. Un negozio che riparava computer, più in là una paninoteca. All'angolo l'insegna di un bar tabacchi. Il vicequestore si avvicinò, aprì la porta del locale e lo vide seduto ad un tavolino all'angolo, guardava fuori dalla vetrina e masticava lento una brioche. Sulla barba era nevicato lo zucchero della glassa. La faccia stanca, gli erano cresciute le rughe. Capelli sporchi, portava un giubbotto di pelle da moto rosso e blu. Quando alzò gli occhi e vide il suo vecchio amico sulla soglia non cambiò espressione. Rimase lì, a masticare lento. Non era sorpreso e neanche sembrava felice di rivederlo. Rocco si avvicinò al tavolo e si sedette senza dire una parola. «Te sto a cerca'». «E m'hai trovato». «È lassù?» gli chiese. «Credo di sì» e bevve un sorso di caffellatte. «Come ci sei arrivato?». «Ho fatto dei giri». Non lo guardava negli occhi, e questo feriva Rocco. Sebastiano sembrava una bomba ad orologeria pronta ad esplodere. Come un artificiere esperto, Rocco doveva provare a disinnescare l'ordigno. Facile nei film, un po' più complesso nella vita. «Mi vuoi guardare per piacere?». Finalmente Sebastiano voltò la testa. Aveva gli occhi rossi e stanchi. «Io sto con te, Seba». «Ma io voglio stare da solo». «Brizio e Furio sono a Clenia, semmai lo stronzo provasse a scappare dall'altra parte». «So' due giorni che sto qui e non sono mai usciti». «Ma io ho dato un'agitata alle acque». Sebastiano si passò la mano tra i capelli. «Ero dentro al bosco. Dormo lì». «Per questo puzzi che accori?». «Vuole qualcosa?» la voce estranea del barista fece sobbalzare Rocco. «Niente, grazie» quello sorrise e se ne andò. C'erano due slot machine e un vecchio flipper che sparavano le lucine a intermittenza. Un calcio balilla impolverato riposava in un angolo. «Ti sei avvicinato alla casa?». «Sì. È una catapecchia di legno in mezzo agli alberi». «Perché stanno nascosti lì secondo te?». «L'infame ha bisogno di documenti. E qui li fa solo 'na mezza stronza nana che però torna domani». «Una stronza nana?». «Sì, la chiamano Mélie Copete, ci ho messo du' ore a capì che era Amelia Coppetti. E ha pure un soprannome. La chiamano 'na cosa tipo falcon, che vuol dire coltellaccio, boh...». «Te posso chiede come ci sei arrivato quaggiù?». «Ti ricordi Flavio Buglioni? Quello che gli ha venduto la pistola?». «Embè?». «Ho fatto le poste e l'ho beccato. Ha parlato facile. Enzo era stato da lui il giorno prima, voleva dei documenti ma Flavio non glieli poteva procurare». «E allora?». «Hanno parlato della Slovenia. Voleva uscire da lì. E io a quello ci avevo già pensato. Ma per andarsene in giro e prendere un aereo per chissà dove, c'è bisogno di un passaporto». «E poi?». «Me so' informato. A fa' documenti falsi da 'ste parti so' boni in due. Uno sta a Udine, l'altra è la nana. Coppetti. La cugina de Mino». Rocco scuoteva la testa. «Ma chi ti dice che veniva proprio da 'ste parti, come dici te. Se poteva pure ferma' ad Ancona, no?». «Tu sei andato a Velletri, vero?». «Come lo sai?». «Perché pure io ho preso le informazioni da lì. Tu puoi parla' con il direttore, io ci ho radio carcere... e come vedi ce so' arrivato prima de te. E lo sai perché? Perché ci ho messo il cuore, Rocco. Io all'infame lo devo massacrare» e tornò a guardare fuori dalla finestra. «Tu non mi hai cercato perché vuoi risolverla da solo o perché non ti fidi più di me?» gli chiese il vicequestore all'improvviso. Sebastiano indurì le mascelle. «Io e te siamo amici da sempre. E solo il fatto di essere arrivati a questa domanda è una coltellata. Io di te mi fido, lo sai, ma lo voi capi' che è una questione personale?». «Quando c'era di mezzo Luigi Baiocchi tu stavi accanto a me». «E per fortuna. Sennò te sparavi in faccia». «Mo' perché io non dovrei stare con te?». «Sei cambiato...» gli disse. «Io?». «Sì. Forse sei più vecchio, forse a tante cose ci hai ripensato». «Mi dispiace che la vedi così». «Rocco, io Adele non me la scordo più. Lo sai che ci parlo?» gli occhi del bestione cominciarono a inumidirsi. «Mo' per esempio, prima che arrivassi te, stava qui fuori a prendere il sole su quella panchina. Ci guardo da un pezzo, ma se n'è andata. Sei passato a Prima Porta?». «Sì». «Hai messo i fiori pure a lei?». «Certo che li ho messi, Seba...». La lacrima trattenuta scappò via. Fu solo un attimo, perché Sebastiano se l'asciugò con la manica del giubbotto. «Io sto qui, Seba. E sto con te». Con la manona l'amico prese il bicchiere e finì il caffellatte. «Noi stasera lo andiamo a prendere» disse Rocco. «Noi?». «Noi». Suonò il cellulare di Rocco. «Dimme, Brizio». «La signora è appena passata con la macchinetta. Stava con un uomo. Uno basso e pelato. Sarà quel Coppetti, mi sa. Novità?». «Grossa quanto un orso» rispose Rocco. «Nun me di'?». «Sì, sta qui...». Tornarono alla sterrata su via del Klančič. Lasciarono l'auto ben nascosta dietro la curva. Poco dopo arrivò Sebastiano sulla moto tenendo bassi i giri del motore. Sparì nella boscaglia. Ne riuscì a piedi e si avvicinò agli amici aprendosi la zip del giubbotto. «Ciao...» disse. «Ciao stocazzo» fece Brizio e partì rapido colpendolo al mento con un pungo secco e preciso. La testa di Sebastiano si girò e lui barcollò all'indietro. «Che sei scemo?» urlò con la mano sul mento. «So' scemo io? 'Cci tua, so' giorni che te stamo a cerca'. Pensavo che eri morto, brutto coglione. Non mi vuoi parlare, mandame un cazzo di messaggio, no? Sta a fa' er misterioso 'sto rincojonito!». «Ha ragione» fece Furio che non s'era staccato dal tronco dell'albero. Sebastiano sputò per terra. «M'hai fatto usci' er sangue, Brizio». «Pure poco» gli disse. «Lo sai? M'andrebbe de pijatte a calci la moto pe' quanto sto incazzato». «Ecco, vi siete stretti la mano. Ci incamminiamo?» fece Rocco. I tre amici lo seguirono sulla sterrata. Lupa sembrava rinata in mezzo alla boscaglia. «Fa' una cosa, Rocco, levaje il collare. Così se la vedono la scambiano per un randagio» suggerì Brizio. Rocco eseguì. Dopo un centinaio di metri la carraia spariva sotto un manto di erba sul quale era evidente il passaggio di un'auto. Poi gli alberi si aprirono per lasciare spazio a un terreno incolto, un prato piuttosto ampio che una volta doveva essere un recinto per animali. C'erano rimasti solo vecchi pali di legno conficcati nel terreno e qualche metro di filo spinato. Alla fine del pratone una piccola casa, una baracca di pietre a un piano col tetto di legno vecchio. Due finestre buie affacciavano sul bosco. I quattro amici si fermarono dietro gli ultimi tronchi a osservare. «La casa è quella, doveva essere una stalla» disse Sebastiano. «Non ha alberi vicino. Se ci avviciniamo possono vederci». «Infatti noi non ci avviciniamo» disse Rocco. «Bisogna aspettare che cali il sole. Diamo un'occhiata sul retro, sempre dentro il bosco, mi raccomando, a distanza...». Misurarono tutto il perimetro. Altre case vicine non ce n'erano. Un fosso d'acqua correva a valle, a una ventina di metri. Il rumore di un'auto attrasse la loro attenzione. Si accucciarono tenendo stretta Lupa. Sbucò la Mitsubishi di Mariella, sobbalzando sulla sterrata raggiunse la casa. Scese solo l'uomo, un tipo bassino senza capelli. In mano teneva una busta. L'ometto aprì la porta e sparì all'interno. «Quel piccoletto è Coppetti?» chiese Furio. «Già». Lo videro uscire poco dopo a mani vuote, accompagnato da Enzo Baiocchi. «Eccolo!» fece Furio. Coppetti scambiò due parole con il pregiudicato, salì sull'auto della compagna che ingranò la retromarcia e ripartì. L'auto percorse la sterrata e sparì nel bosco mentre Baiocchi rientrò in casa. Il rumore del motore si affievolì fino a sparire. Rimasero solo gli uccelli a cantare e il respiro affannato di Lupa. «Enzo Baiocchi è solo?». «Ora sì. Per aspettare il buio possiamo pure tornare sulla strada principale» disse Rocco. Furio e Brizio s'erano offerti di andare al paese a prendere birra e panini. Sebastiano e Rocco erano di nuovo soli, seduti su due massi che spuntavano dal terreno. Fumavano e guardavano il cielo, mancava un po' al tramonto. Sebastiano aveva riacceso il cellulare che dopo qualche secondo si scatenò con una serie di suonerie. «Mo' ci metto una vita a leggere tutti questi messaggi». «I vaffanculo sono i miei» disse Rocco. Sebastiano si concentrò nel lavoro. «La maggior parte so' rotture di palle» disse con gli occhi puntati sul display. Rocco spense la sigaretta, si alzò dal masso e andò ad appoggiarsi a un albero. «Sebastia'... hai mai sentito nominare Juan Gonzalez Barrio?». «Chi?». «Juan Gonzalez Barrio». «Mai sentito, chi è?». «È uno che una volta lavorava all'ambasciata dell'Honduras e che io contattai per il traffico di coca del 2007, ti ricordi?». «E come faccio a scordarmelo, Rocco. Certo che me lo ricordo». «Ruben Montoro, invece?». «Oh, ma che me stai a fa' il terzo grado? No, mai sentito». Gli lanciò il cellulare. «Apri le foto, l'unica che trovi è quella segnaletica di Montoro. Dimmi se lo conosci». Sebastiano eseguì. «Mai visto» e gli restituì il telefono. «Chi è?». «Un cadavere che hanno trovato vicino a Castel di Decima. Gola tagliata. E aveva il mio numero in tasca». «Perché lo dovrei conoscere?». «Curiosità». Sebastiano lo guardò. «Sei strano Rocco. Sei molto strano». «Non sono strano, mi mancano un po' di dettagli e il quadro non mi è ancora chiaro, tutto qui». Un rumore di foglie calpestate li interruppe. Sebastiano scattò mettendosi la mano nella cintola. Ma erano Furio e Brizio con due buste che scansavano rami e avanzavano nella boscaglia. «Arriva la pappa» fece Furio. Brizio guardò i due amici. «Abbiamo preso da bere. Ma non so se me la sento di fare un brindisi» e lanciò una birra a Sebastiano. Finalmente il sole stava tramontando. Radunarono carte e lattine e lasciarono la piccola radura. Era scesa la temperatura, o forse era solo aumentata l'umidità, e Rocco ebbe un brivido mentre metteva Lupa in macchina a dormire. Si incamminarono sulla sterrata che portava alla baracca. Non dissero una parola. Avanzavano silenziosi. Il giorno si stava spegnendo come una candela a fine corsa insieme al canto degli uccelli. Per fortuna la luna alta nel cielo era poco più di un'unghia e non avrebbe disturbato la notte. Raggiunsero l'ultima curva prima della casa. Sebastiano li fermò con un gesto. «Da qui vado avanti da solo» disse. «Che sei matto?» fece Brizio. «Noi veniamo con te». «No Furio, è una questione mia, e mia soltanto. Grazie che ve siete preoccupati, grazie Rocco che te sei dato da fa' a cerca' quel bastardo, mo' è una questione personale. Io devo resta' da solo con quell'infame. E non vi voglio vicino». «Quando Rocco anni fa...» provò a dire Furio, ma Sebastiano lo interruppe. «Quando era il momento di Rocco io ci dovevo stare sennò non ce la faceva. Ma io non corro 'sto rischio. Io lo devo ammazza' con le mie mani, e voi ne state fuori. Giuratelo!». I tre amici lo guardavano senza dire niente. «Giuratelo!». Il primo ad annuire fu Brizio. «È giusto, Rocco, ha ragione. Questa è una storia sua. Noi fino a qui ci siamo arrivati, ora tocca a lui». «Pure secondo me». Rocco strinse le labbra. Era arrivato al punto, al bivio e ormai doveva scegliere quale strada prendere, una decisione che aveva rinviato fino a quel momento. Lasciare che l'amico avesse la sua vendetta e diventare complice dell'omicidio, oppure seguirlo e tentare di impedirglielo? Qualche anno prima non si sarebbe mai posto il dubbio, consegnargli Baiocchi senza battere ciglio sarebbe stato un gesto naturale. Forse Sebastiano aveva ragione, stava invecchiando. «Me l'hai giurato» disse l'amico che lesse l'indecisione negli occhi del vicequestore, «me l'hai giurato davanti al corpo di Adele. E le parole so' piombo, lo sai». «Sì, lo so. E hai ragione, Seba. Vai e fa' quello che devi fare. Noi ti aspettiamo qui». L'amico si incamminò da solo nella boscaglia. «Oh» lo richiamò. «Che vuoi?». «La moglie di Coppetti li ha avvertiti che la polizia sta cercando il falsario. Quindi estrema cautela, capace che sta sul chi va là, Enzo è armato». «Pure io» disse Sebastiano sfilando la pistola dalla cintola. «Quel ferro lavora?» gli chiese Rocco. «La ragazzina lavora, e lavora bene. A fra poco» e sparì dietro i tronchi. «È giusto così» fece Furio. Si sedette su un sasso e si accese una sigaretta. «Siamo venuti fino a qui, abbiamo scoperto dove sta nascosto l'infame, abbiamo visto che Seba sta bene, missione conclusa, no?» disse Brizio. «Hai ragione, è giusto così» fece Rocco. Sebastiano era arrivato al pratone davanti alla casa. Una luce fioca illuminava di giallo sporco la finestra accanto alla porta d'ingresso. Accucciato e silenzioso, avanzò correndo sul prato, allo scoperto. Era il momento più delicato. Un rumore, un solo sospetto e quello l'avrebbe visto. Giunto a metà strada, a una trentina di metri dalla baracca, la porta si spalancò. Sebastiano si gettò a terra fra l'erba bassa. La pistola nella cintola lo colpì sull'inguine, trattenne un'imprecazione e ringraziò Dio di aver messo la sicura. Sarebbe stata una fine poco nobile. Dalla casa vide uscire Enzo Baiocchi che fece un paio di metri, si aprì la patta dei pantaloni e pisciò guardando la notte stellata. Finita l'operazione se ne tornò dentro sbattendo il portoncino di legno. Sebastiano serrava i denti così forte che rischiava di scheggiarli. Si rialzò e veloce riprese la corsa. Finalmente si appoggiò alla baracca. «Rega', comunque stiamo con gli occhi aperti. Semmai dovesse tornare Coppetti con la moglie lo blocchiamo» disse Brizio sgranchendosi il collo. «Vabbè, e che problema c'è? Quello appena ce vede se la fa sotto. Anzi magari, così Rocco lo porta in questura e ci ha pure l'alibi». «Questo è vero» fece Brizio. I rami smossi dal vento parevano mani enormi che afferravano l'aria. «Che silenzio» disse Brizio. «Non ci siamo abituati. Aosta com'è, silenziosa?». «Abbastanza» fece Rocco. «I primi giorni non riuscivo a prendere sonno». «Poi?». «Poi uno si abitua. Pare che ci si abitui a tutto». Sebastiano si affacciò lentamente alla finestra illuminata. E lo vide, seduto alla luce di una lanterna da campeggio, che controllava delle carte. Sul tavolaccio c'erano due bicchieri, un piatto con avanzi del pasto. Teneva la pistola, l'arma che aveva tolto la vita a Adele, poggiata accanto alla bottiglia di vino. La casa era spoglia e piena di ragnatele. Lontana si sentiva una canzone, forse da una radio. Sebastiano tornò alla porta. Si concentrò. Appena entrato doveva sparare mirando al petto, non al volto. Voleva avere il tempo di guardarlo morire e voleva che Baiocchi avesse i secondi necessari per realizzare che a ucciderlo era stato lui. Godersi ogni secondo dell'agonia fino a che avesse sputato l'anima dai denti, anche se l'anima Baiocchi non ce l'aveva e di questo Sebastiano ne era certo. Aveva aspettato tanto, e ora ce l'aveva, proprio dietro quella porta di legno, a pochi metri. Bastava aprire e sparare. Tolse la sicura alla pistola. Prese fiato per calmare i battiti del cuore. «Uno... due...». Tre soli si accesero togliendogli la vista. Il bosco, la casa, tutto s'era illuminato a giorno sotto la potenza di fari che sparavano migliaia di watt squarciando la notte. Una moltitudine di insetti sorpresi nel volo notturno sfarfallavano nei fasci di luce. Sebastiano con la mano davanti agli occhi, accecati e indeboliti, cercava di capire cosa l'avesse colpito con quella violenza. Scorgeva tre ombre armate che avanzavano in silenzio. «Gettate le armi e incrociate le mani sulla testa» si udì riecheggiare nel bosco. Sebastiano obbedì lasciando a terra la pistola. Poi finalmente le ombre divennero uomini armati di MP5, scuri come il buio che li circondava, con il giubbotto anti-proiettile e il viso coperto dal passamontagna. Dalla casa uscì Enzo Baiocchi, già ammanettato, stretto fra due colossi con l'elmetto e le pistole in mano. Passò davanti a Sebastiano e gli sorrise. «Hai fatto tardi...» gli sibilò. Sebastiano scattò e gli si avventò al collo, ma l'intervento di due agenti impedì che gli strappasse la pelle dalla faccia. Ci volle l'aiuto di un altro uomo per bloccarlo. «Figlio di puttana, lasciame! Lasciame!» urlava. Ma Enzo Baiocchi, scortato dai due poliziotti, si perse nell'oscurità al limitare del bosco. Impotente, con gli occhi iniettati di sangue, stretto da dieci braccia, sputava bava sulla barba e sulla divisa di uno dei poliziotti mentre il respiro a fatica pompava aria nei polmoni. Brizio, Furio e Rocco erano rimasti impietriti. Dagli alberi avevano visto i potenti fari illuminare la notte. «Che cazzo succede?» aveva detto Furio con la voce smozzicata. Rocco fece per avvicinarsi alla fonte luminosa quando l'ordine «Gettate le armi e incrociate le mani sulla testa» riecheggiò per tutto il bosco. Furio lo bloccò per un braccio. «'Ndo vai? Fermo qui...». «Le guardie?» chiese Brizio spaventato. Rocco non capiva. «Rocco, queste so' guardie. Le hai chiamate tu?». «Io? Ma come te viene in mente, Furio...». «A rega', poche chiacchiere. Via per ora, via!» veloci si persero fra i rami verso la strada. Saltavano radici, evitavano i tronchi. Un ramo schiaffeggiò Rocco sul volto, Furio inciampò ma si rialzò subito. Si precipitarono verso la strada. Finalmente la raggiunsero. «Alla macchina!» gridò Furio. «No!» fece Rocco. «La macchina sta infrattata, non l'hanno vista. Aspettiamo qui». Si accucciarono fra i cespugli e attesero. Dalla carraia videro spuntare una prima auto che a velocità folle puntò dritta verso San Pietro. Era grigia con il lampeggiante spento e i vetri oscurati. Passarono due fuoristrada con la luce blu che invece a intermittenza illuminava le fronde degli alberi. «Ma quanti cazzo so'?» chiese Brizio. Infine un furgone. Poi quando anche quello sparì dietro la curva, tornarono il silenzio e i rumori del bosco. Lenti i tre uomini si alzarono in piedi. «Rocco, ma che è successo?». «Hanno arrestato Baiocchi. E pure Sebastiano». «Come facevano a sapere che...» poi Furio si interruppe. Brizio concluse la domanda «... che stavano qui, eh, Rocco?». «Non lo so, proprio non lo so». «Hai parlato con qualcuno alla questura di Cividale, no? Ci avranno seguito». «Ma quando mai» e ripensò all'agente Esposito e al vicequestore biondo che cercava un preservativo. «Quelli erano reparti speciali. Altro che questura di Cividale». «Guarda, io me voglio fida'» e Brizio si accese una sigaretta. Nel buio le braci illuminarono di rosso i suoi occhi stanchi. «Se non hai parlato con nessuno, qualcuno però ha parlato di te». «Su questo non c'è dubbio». «Hai gli occhi addosso, allora. Di chi? Ma se sapevano che stavamo qui, che cercavamo Baiocchi, significa che ogni tua mossa, ogni tuo spostamento qualcuno lo conosce». «Io devo andare da Seba. Non posso lasciarlo da solo». «Rocco, ma che dici? Se vai in questura alzi un vespaio che non finisce più» obiettò Brizio. «Nessuno sa che siamo qui. Annamosene. A Seba ci pensiamo dopo». «Concordo» si unì Furio. «Nessuno sa che siamo qui? Ma che dici, Brizio. Mi hanno seguito, lo sanno che sono qui. Quindi giochiamo scoperti». «Secondo me fai una pazzia». «Non lo posso lascia' da solo» insistette Rocco. «Ci devo parlare, qualcosa mi invento». Davanti alla questura di Cividale c'erano due auto parcheggiate che tenevano i lampeggianti accesi. La luce blu spettrale rimbalzava sulle facciate e sulle finestre dei palazzi intorno. Rocco salutò un paio di poliziotti ed entrò. Esposito stava scendendo le scale. Gli sorrise. «Dottore, ha sentito l'arresto a San Pietro?». «No, mi chiedevo infatti di questo casino...». «Se vuole su c'è il vicequestore. È arrabbiato nero». «Immagino. Je tocca lavora'». Ridendo Esposito lo superò e uscì sulla piazza. Rocco salì al primo piano. Agenti entravano ed uscivano dagli uffici. Si sentiva il ticchettio dei computer, voci al telefono. L'ultima porta era quella del vicequestore Augusto Cascone, il modello Armani che aveva sbagliato carriera. Era aperta. «Permesso?». Cascone era al telefono. Gli fece cenno di accomodarsi. «Sì... sì, per ora ho mandato gli uomini a controllare... sì, Coppetti. Sì, dottore, certo. Certo, non si preoccupi. Appena ho notizie la chiamo». Attaccò il telefono, si stiracchiò stropicciandosi gli occhi. «Schiavo', mannaggia alla miseria m'hai portato un casino, stavamo tanto tranquilli!». «Io?». «Eh sì... hanno appena arrestato un pregiudicato molto pericoloso, Baiocchi... che stava proprio da quel Coppetti che cercavi te. L'abbiamo beccato con le mani nel sacco. Dodici passaporti, due carte di credito. Quel Baiocchi era pronto a volare via e adesso se ne sta a Udine bello carcerato». Si tirò su le maniche del cardigan. «E noi che tenevamo d'occhio la cugina...». «La Amelia Coppetti, detta il falcon?». Augusto sgranò gli occhi cerchiati di rosso. «La conosci?». «Te l'ho detto. Indagavo su quel traffico». Cascone fischiò. «Però, sei un grande. Dammi il cinque» e si sporse sulla scrivania offrendo il palmo della mano. Rocco con imbarazzo alzò il braccio proprio mentre un agente scelto si affacciò alla porta. «Dottore, quando vuole Coppetti è pronto». «Grazie Siniscalchi» e l'agente scelto sparì. «Senti Augu', a parte che mi dispiace che la serata è andata a farsi fottere». «Lascia stare, è pure rotta la macchina dei preservativi». «Quando se dice la sfiga. Dicevo, oltre a Coppetti e a 'sto Baiocchi che non me ne frega un cazzo, hanno arrestato qualcun altro?». Augusto dondolò la testa. «Sì, un bestione di Roma. Perché?». «Sebastiano Cecchetti, vero?». «Ma te sei una sorpresa continua, Schiavo'! Maronna del Carmine, un segugio. Sta di là. Aspettiamo il magistrato». «Da collega a collega, mi ci faresti parlare un minuto?». «E non lo so... prima dell'arrivo del magistrato...». «Per favore. Mi risparmi una trafila. Un minuto!». «Ma lui che c'entra?». Rocco abbassò la voce. «È un mio informatore...». Augusto aggrottò le sopracciglia sui begli occhi azzurri. «Un informatore?». «Sì, ma non lo sputtanare... sennò mi rovini». Il vicequestore si portò la mano al cuore. «E che sei matto? Solo che resta una cosa strana. Stava alla baracca armato. Che faceva?». «È pazzo» gli disse Rocco. «S'era messo in testa di aiutarmi, ma ha esagerato. Fammici parlare, vedrai che mi conosce». «Da collega a collega, ti do cinque minuti» poi chiamò: «Siniscalchi!». Dopo pochi secondi si affacciò di nuovo l'agente scelto. «Dica?». «Accompagnami il collega dal bestione romano». «Subito. Venga dottore». Rocco si alzò. «Grazie Augusto, ti debbo un favore». «Ma figurati, se non ci si aiuta fra noi. Al magistrato la dico 'sta cosa dell'informatore?». «E digliela» e seguì l'agente. L'agente scelto gli aprì l'inferriata che proteggeva la porta, poi prima di afferrare la maniglia si raccomandò: «Sto qui fuori, dottore. Se dovesse servire». «Stia tranquillo, è un amico» gli disse, poi entrò. Sebastiano era seduto davanti a un tavolo. In alto una finestrella con le sbarre. L'arredamento era tutto lì. Aveva gli occhi arrossati. «Bravo Rocco, sei felice?». «Io non ne so un cazzo!». «Ah no? Non ne sai un cazzo? E chi ce l'ha portati quelli? Mi' nonno?». «Se ti dico...». Ma l'orso non ascoltava. « Quando era Luigi Baiocchi io ero con te! Con te! Tu invece m'hai accoltellato alle spalle, Rocco Schiavone. Hai ucciso Adele un'altra volta. Quel faccia di merda se n'è andato, vivo! E io sto qui a strilla' in cella! Era questo che volevi? Eravamo amici io e te, eravamo fratelli!». «Seba, io non ne so niente, come te lo devo dire? Io non ho manco capito che è successo!». «Eppure è facile. Quelli sono reparti speciali, e io non l'ho chiamati. E chi sarà stato? Sei una guardia, Rocco, e guardia resterai per sempre!». Rocco si avvicinò ma Sebastiano scattò in piedi come se fosse in presenza di un serpente. «Nun t'avvicina', Rocco, che te metto le mani addosso». Ma Rocco non ascoltò il consiglio. «T'ho detto nun t'avvicina'!». Arrivato a mezzo metro Sebastiano lo colpì con un cazzotto in pieno volto che fece cadere il vicequestore all'indietro. Lo guardava dall'alto, coi pugni serrati, il volto una maschera di odio. «Te l'avevo detto, giuda traditore. Ma come hai fatto? Come cazzo hai fatto?». Rocco si rialzò. Si mise in piedi davanti all'amico asciugandosi il sangue dal labbro. «Non sono stato io, Seba. Tu mi devi credere». Ma Sebastiano lo caricò come un toro con la testa sullo stomaco. Caddero a terra. Rocco cercava solo di scansare i colpi che Sebastiano menava senza logica, con rabbia. Il pavimento era freddo, l'amico grugniva, gli arrivò un colpo sul naso che gli accese un milione di stelle negli occhi. La porta della cella si aprì e entrarono l'agente scelto con un collega per fermare la rissa. «Fermi! Fermi!» urlò Rocco, ma i due avevano già afferrato Sebastiano per le spalle tentando di scioglierlo dal vicequestore. «Vi ho detto fermi!» urlò Rocco. I poliziotti riuscirono a disincastrare i due corpi e a portare Sebastiano a distanza di sicurezza. Rocco si alzò in piedi. «Lasciatelo. Vi ho detto lasciatelo!» urlava guardando l'amico negli occhi. «Dotto', questo la massacra» gridò uno dei due. «Meno male che era un amico» disse l'agente scelto. «Fate come vi ho detto!». I due agenti col fiatone piano piano mollarono il collo e le braccia di Sebastiano Cecchetti che respirava a fatica e non si divincolava più. Rocco con il sangue dal naso si rimise la camicia strappata nei pantaloni. «Uscite e lasciateci soli». I due poliziotti si guardarono senza capire, si rassettarono la divisa e finalmente lasciarono la stanza. Sebastiano puntò lo sguardo sul pavimento. Rocco si avvicinò all'amico. Lo abbracciò forte e finalmente Sebastiano scoppiò a piangere. «Perché m'hai fatto questo... perché?». «Te lo provo che non c'entro niente!». Sebastiano si liberò dall'abbraccio e si rimise seduto. Si prese la testa fra le mani. Poi con una voce bassa, cavernosa, disse: «Vattene Rocco, io e te non abbiamo più niente da dirci». Schiavone aspettava fuori dalla porta che finisse la riunione. Era passata l'una di notte e ogni tanto la stanchezza aveva la meglio e un colpo di sonno gli faceva cadere il mento sul petto. Baldi era arrivato da mezz'ora e stava chiuso nella stanza col magistrato De Biase e il vicequestore Cascone. Avevano anche riso, ed era riuscito a sentire qualche brandello di conversazione. Poca roba, solo che Baldi e De Biase erano stati colleghi al tribunale di Bari qualche anno prima. Fuori il cielo era ancora scuro, ma questione di poche ore e l'alba sarebbe spuntata. Aveva la bocca secca. Era almeno riuscito a cancellare il sapore del sangue bevendo acqua e il pessimo caffè del distributore. Il viavai s'era calmato. Ora il commissariato era più tranquillo e s'erano zittiti i rumori delle stampanti e gli squilli dei telefoni. La serenità s'era interrotta solo quando due agenti avevano portato dentro un ragazzo magro e imbottito di roba tanto da non reggersi in piedi. Finalmente la porta si aprì, e uscì Baldi. Il viso era teso, la pelle tirata come un tamburo, gli occhi rossi e i capelli spettinati. Si infilò la giacca, prese Rocco sotto braccio. «Andiamo fuori, facciamo due chiacchiere». Si incamminarono fra le strade del centro mentre la città dormiva. «Io le ho retto il gioco, Dio solo sa perché, ma adesso lei mi racconta un po' di cose». Le ombre dei due si allungavano sul selciato di via Pellico. «Come ho sempre sospettato, lei gioca su troppi tavoli. Mi corregga se sbaglio». «Sbaglia, dottore, la verità è un'altra». «E mi dica...». «Io ero dietro Sebastiano Cecchetti, che lei ricorderà essere...». «Il fidanzato di Adele Talamonti, sì, lo so, vada avanti». «L'ho cercato e seguito finché ho capito che era diretto qui, a cercare Enzo Baiocchi in fuga». «So dove vuole arrivare e la fermo subito. Lei voleva evitare che quello gli sparasse». «In qualche modo sì». Baldi scoppiò a ridere. «E io me la bevo?». «Faccia come vuole. Ma è la verità». «Cioè lei mi sta dicendo che non era alla caccia di Baiocchi bensì alla ricerca del suo amico». «No, temevo per Sebastiano». «Mi prende per scemo?». «Allora risponda a questo. Se, come dice, ero dietro a Enzo Baiocchi, com'è che al momento dell'arresto di Sebastiano Cecchetti io non ero lì?». La domanda prese in contropiede il magistrato. «Non c'ero, e sa perché? Perché non ho fatto in tempo. Se avessi cercato Baiocchi a quest'ora io mi trovavo in cella insieme a Sebastiano Cecchetti. Invece sono qui e parlo con lei!». «Quindi ringrazia i reparti speciali per averle tolto le castagne dal fuoco?». «Dio benedica i reparti speciali». Erano arrivati al ponte del Diavolo. Il fiume scorreva ancora avvolto nell'oscurità. «Lei sembra un grillo, di quelli che saltano di stelo in stelo e non li acchiappi mai». Si appoggiò al parapetto a guardare il paesaggio. «Già, un grillo. È bello qui». «Non lo so, non ho avuto il tempo di girare la città». «Vede laggiù? Là dietro c'è la stretta Cornelio Gallo. Io abitavo lì». «Lei è di qui?». «No, mio padre era direttore di banca. Ho vissuto un po' in giro. Qui a Cividale quattro anni. Ma ero piccolo, ricordo casa mia, il Natisone e la scuola elementare. Già, migriamo tutti, chi per lavoro, chi per affetto, chi per punizione» e lanciò un'occhiata sarcastica al vicequestore. «A proposito, lo sa che io avrei la soluzione per l'immigrazione africana nel nostro paese?». Baldi era tornato ai discorsi sui grandi temi socio-economici. Questo mise Rocco di buon umore, era il segnale che il magistrato si stava rilassando. «La vuole sapere?». «Certo». «È molto semplice. Si fa una domanda all'immigrato appena sbarcato. A seconda di quale lingua parla, lo si manda nel paese d'appartenenza». «Si spieghi meglio» fece Rocco. «Parla francese? In Francia. Parla inglese? In Inghilterra. Portoghese? In Portogallo. Italiano? In Italia. Perché francese, inglese, spagnolo, non sono lingue native del Senegal, dell'Angola, della Nigeria, del Corno d'Africa. No. Quelle sono le niger-kordofaniane, le lingue fur, le semitiche. Se parlano quelle europee è perché un paese europeo per centinaia di anni ha occupato e sfruttato la terra dei loro nonni. Quindi il minimo che ora quel paese colonizzatore possa fare è ospitarli. Punto e basta. Si chiama indennizzo». Rocco ci rifletté sopra. «Non è del tutto sbagliato, dottore». «Ne conviene?». Poi guardò Rocco in viso. «Perché non s'è fatto aiutare?». «Con il mio amico? Perché mi muovo più veloce da solo». «Ora Baiocchi è agli arresti a Udine. Io non so perché si sono mossi con i reparti speciali, ma deve essere uno che sa un sacco di cose». «Anch'io sono arrivato alla stessa conclusione, dottor Baldi». «E soprattutto forse ora ci dirà cosa lo spingeva a farla fuori». «Già, sono curioso anche io». «Io mi sarei fatto un'altra idea. La vuol sentire?». «Tempo ne ho». «Lei conosce il motivo per cui Baiocchi la vuole morto, e chiudergli la bocca era una priorità. Lo cercava per azzittirlo una volta per tutte. La vuole morto perché nel passato lei gli ha fatto qualcosa. Ma se così fosse lei sarebbe una specie di giustiziere, un vendicatore. Diciamo pure un assassino. E io non credo di essermi sbagliato così tanto nei suoi confronti. Lei è un poco di buono, arronza, fa le cose di testa sua, ma non è un assassino. No, voglio credere alla storia che mi ha appena raccontato. Sarà che questa città mi riporta all'infanzia, sarà l'aria fresca di questa notte, sarà che ancora non digerisco Bernardo Valenti, ma le voglio credere. E a proposito di Bernardo Valenti...» alzò gli occhi al cielo. «No, sono troppo stanco. Ne parliamo ad Aosta, io sono andato avanti. Aspetto un paio di riscontri. Schiavone...?». «Dica». «È l'ultima volta che le salvo il culo, lo tenga bene a mente». Brizio guidava. Furio, seduto accanto, ascoltava la radio a bassissimo volume. Le luci di un autogrill passarono veloci sui loro volti mentre il cielo si colorava di rosso. Superarono un camion che trasportava vitelli. Si vedevano i musi delle povere bestie che cercavano di rubare ancora qualche boccone di aria e di vita. Rocco steso sul sedile di dietro accanto a Lupa guardava fuori dal finestrino. Si sentiva sporco e freddo. Uno straccio usato e lasciato sul pavimento di una stanza. Dove aveva sbagliato? Chi continuava a superarlo senza neppure mettere la freccia? S'era giocato anche l'amicizia con Sebastiano, uno dei tre esseri umani su cui poteva sempre contare. Qualche risposta doveva trovarla. E subito. Quel Barrio, quell'ex dipendente dell'ambasciata allontanato e quasi dimenticato dai suoi vecchi datori di lavoro era il nodo principale da sciogliere. E il commando dei Nocs, come era arrivato fino a lui? Guardò le nuche dei suoi vecchi amici. Loro non potevano essere, non c'era neanche da pensarci. Passò ad Alfredo De Silvestri, il vecchio agente dell'EUR. Impossibile. Neanche era a conoscenza che stessero dando la caccia a Baiocchi. Il direttore del carcere di Velletri? Richiamando alla memoria il viso di Francesco Selva percepì una nota stonata. Con lui aveva parlato. La strada per Udine e dintorni l'aveva imboccata grazie alle informazioni avute dal secondino. Era lui la spia? Oppure l'agente di Cividale, quell'Esposito che gli aveva fornito l'indirizzo della compagna di Coppetti? Difficile. In commissariato erano stupiti tanto quanto lui di quell'intervento dei reparti speciali. Ombre. Come la storia di Bernardo Valenti. C'era sempre l'ombra di qualcuno più veloce e preparato di lui che gli pesava sul collo. Ma il dettaglio che stonava di più era sempre lo stesso. E gli diede voce Brizio che per conto suo seguiva i suoi pensieri: «Sebastiano se la caverà?». «Non lo so. Dipende da De Biase, il procuratore. Baldi mi ha detto che c'è il problema del tentato omicidio». «Ma Seba non ha sparato. Cioè, mica si possono processare le intenzioni». «Era lì con un'arma, Furio. Che stava a fa'? Una gita?». «L'avvocato lo tira fuori secondo me» fece Brizio mettendo la freccia per superare un'auto. «Dirà che era lì per fermarlo ed era sua intenzione consegnarlo alla giustizia». «E chi ce crede?». «A Furio, ce devi crede tanto quanto al fatto che je voleva spara'!». «Invece, Rocco, io vorrei sapere perché scomodano un commando dei Nocs per un pezzo di merda come Baiocchi?». «Gli serve». «Per cosa?» fece Furio. «Non lo so. Ma deve conoscere cose interessanti». Furio scosse la testa. «Non capisco, non capisco proprio. Baiocchi stava in carcere, avevano tutto il tempo di interrogarlo. Perché ora dopo la sua fuga?». «Perché avranno indagato, avranno scoperto qualcosa su di lui, io che ne so?». «È così» affermò Furio. «S'è mosso a Roma, pensava di essere invisibile invece gli tenevano gli occhi addosso». «Ma secondo te ci siamo arrivati insieme al nascondiglio?». «Può darsi di sì, Brizio. Oppure come dicevi te, qualcuno sapeva i miei spostamenti, e ha aspettato il momento giusto per intervenire. Così hanno inchiodato Baiocchi e pure me. Perché adesso mi tengono per le palle». «Tu dici che tutto 'sto fango viene da Roma?». «Il fango viene sempre da Roma, Furio...». Tornò il silenzio accompagnato dal rumore del motore. Lupa sbadigliò. Rocco avrebbe voluto stare in un letto. A considerare come, in pochi secondi, la sua vita si era capovolta un'altra volta. Aveva pure la sensazione che i colpi dell'avversa fortuna non avessero ancora finito di martellarlo. Poi gli tornò alla mente Gabriele e i suoi problemi con la madre assente e la scuola che lo intrappolava. E Caterina e il suo corpo, dove avrebbe voluto nascondersi ancora per un po'. Brizio parcheggiò vicino al ministero. Scesero, stanchi del viaggio. Trastevere era una bolgia. Furio si incamminò per recuperare l'auto e tornarsene a casa. Lupa andò a controllare gli pneumatici da innaffiare. «Be', ci vediamo» fece Rocco. Per tutta risposta Brizio si accese una sigaretta. «Che c'è?» gli chiese Rocco. «Sta succedendo una cosa brutta» e Brizio tirò una boccata, «e solo il fatto che ci abbia pensato mi fa stare male». «A che ti riferisci?». «Vedi Rocco? Te vojo bene da quarant'anni, però quello che è successo m'ha stranito» si appoggiò all'auto che aveva ancora il cofano caldo per il viaggio. «Sto parlando di Sebastiano». «Dove vuoi arrivare?». «Io lo so che non c'entri, lo so che tu non hai giocato sporco, di te mi fido, mi fido da sempre. Ma...» fece un tiro alla sigaretta, sputò il fumo chiaro dalla bocca «... è successo». Guardò Rocco negli occhi. «E la domanda è: sarebbe successo lo stesso se tu non fossi stato con noi?». Rocco incrociò le braccia. Scosse appena la testa, la sua era una domanda silenziosa. Brizio continuò: «Io dico di no, che non sarebbe successo. Perché tu sei un poliziotto, Rocco, noi no». «E allora?». «E allora che t'abbiano seguito o meno, sei tu la causa dell'arresto di Sebastiano. Solo per il fatto di essere un poliziotto. Basta quello. E prima o poi a questo punto ci dovevamo arrivare». Gettò la sigaretta a terra, la brace rimbalzò sull'asfalto. «Stiamo su due sponde diverse. E pure se ci vogliamo bene, se rispettamo, tu sei tu e io so' io. Mi capisci?». «Quindi anche tu non ti fidi di me?». «No, te l'ho detto. Io di Rocco mi fido e lo farò sempre. Ma sei pure il vicequestore Schiavone, e questo non lo puoi cancellare». Rocco aggrottò le sopracciglia. «Fino ad oggi però 'sto ragionamento non l'hai mai fatto». «Fino ad oggi è andato tutto pulito. Ora Seba è dietro alle sbarre. E sai che ti dico? Che forse quel bestione aveva ragione a volerci andare da solo a prende quel disgraziato di Baiocchi. Perché da solo, a quest'ora, la partita l'aveva già chiusa da un pezzo». «Io di te ho bisogno» disse Rocco guardando per terra. «Di te, di Furio e di Sebastiano. Siete un pezzo della mia vita, siete come le braccia e le gambe, non mi fare questa cosa, Brizio...». Brizio si allontanò dall'auto. Abbracciò il suo amico. Stettero in silenzio per qualche secondo. Poi si staccò. Aveva gli occhi umidi. «E il giorno che faccio una cosa brutta e tu lo vieni a sapere, quel giorno, che farai? Chiudi un occhio? Sei un poliziotto, Rocco, e non puoi». «Questo lascialo decidere a me. Ti chiedo solo di continuare a essere come prima. Per quanto possa valere, ti dimostrerò che avevo ragione, che qualcuno mi ha tradito». «Non c'è bisogno, io lo so già. Se sei un'aquila voli, amico mio, se sei un cavallo trotti. E né io né te possiamo farci niente». L'auto con Furio a bordo si fermò lì accanto. Brizio mollò una pacca a Rocco. «Ciao...». «Fatte 'na dormita, Rocco» lo salutò Furio. Poi insieme partirono e lo lasciarono lì, in mezzo alla strada. Si svegliò di soprassalto al terzo squillo del vecchio cellulare che aveva riacceso prima di andare a dormire. Impiegò una decina di secondi per ricordarsi dove fosse: Roma, Trastevere, albergo Santa Maria, camera 62, primo piano. Allungò il braccio. «Chi è?». «Dove sei?» la voce di Caterina. «Ma che ore sono?». «Mezzogiorno, Rocco. È dall'altro ieri che ti chiamo ed era sempre staccato. Ce l'hai la tv?». Guardò la parete di fronte. Sì, ce l'aveva. «Embè?». «La stai guardando?». «Stavo a dormi', che stavo guardando...». «Sei di buonumore, vedo. A mezzogiorno dormi ancora? Accendi sul telegiornale. Ora, subito!». Rocco trovò il telecomando sul comodino. Accese. «Sul primo!» fece Caterina e chiuse la comunicazione. Apparve il viso di un giornalista in giacca e cravatta. «Arrestato in un blitz nella notte il pregiudicato Enzo Baiocchi» e alle spalle dello speaker una foto che ritraeva due agenti col passamontagna che tenevano stretto il figlio di puttana. «... catturato in provincia di Udine. Enzo Baiocchi, evaso dal carcere di Velletri lo scorso maggio, era ricercato per numerosi delitti, non ultimo l'omicidio di Adele Talamonti avvenuto ad Aosta. Gli inquirenti lo ritengono persona a conoscenza di fatti fondamentali riguardo traffici di droga nella capitale. Al momento è a disposizione dei magistrati nella casa circondariale di Udine dove ha passato la prima notte di interrogatori. Le sue rivelazioni sono ancora protette dal segreto istruttorio...». Rocco spense la televisione. Si alzò dal letto. Si accorse che s'era addormentato vestito. Non aveva tempo da perdere. Una doccia rapida e uscì dopo essersi raccomandato con la cucciola di restare buona, cosa che lei fece mettendosi ai piedi del letto. Appese il cartello «Non disturbare» fuori dalla porta e prima di uscire si raccomandò all'uomo della reception di portare da mangiare al cane. Non era venuta da sola. Accanto a lei c'era Alejandro Giménez, l'addetto stampa dell'ambasciata. Lo aspettavano ai tavolini di un bar in piazza San Cosimato. Appena lo videro arrivare si alzarono in piedi. «Ben trovato, dottor Schiavone». Si strinsero la mano. Giménez era sempre impeccabile, stavolta in un completo di cotone color crema, Carmen aveva optato per un tailleur pantalone blu. «Allora, Carmen mi ha detto tutto. E siamo felici di poterla aiutare...» fece un gesto. La segretaria prese una cartellina di pelle e tirò fuori un foglio. «Ruben Montoro lo conosciamo, dottor Schiavone. Un bandito, uno che la polizia del nostro paese teneva d'occhio da tanti anni, ma è sempre sfuggito. Sapevamo che aveva a che fare con Barrio, forse addirittura lavorava per lui». «Certo che lavorava per lui» affermò Carmen col viso serio. «Gliel'assicuro, dottor Giménez, faceva l'autista per Barrio. Solo che non riuscimmo mai a provarlo» disse rivolta a Rocco. «Perché Barrio è stato allontanato?». «Non aveva un buon odore» rispose Giménez. «Trafficava con gente come Montoro, e noi dell'ambasciata non ci fidavamo più di lui. C'era anche stato un mezzo scandalo con delle fatture poco chiare, così lo rimuovemmo». «Capisco. Io questo Barrio lo devo vedere». Per tutta risposta l'addetto stampa gli allungò il foglio che teneva in mano. «Questo è il suo indirizzo». Rocco lo lesse. «Abita vicino casa mia... ora si spiega il bar di via Vitellia». «Non la seguo». «Montoro aveva uno scontrino in tasca del bar Mastrangeli a via Vitellia. Che è vicino casa di Barrio». «Bene» fece Giménez alzandosi dalla sedia, subito seguito da Carmen. «Siamo felici che sia giunto a queste conclusioni. E speriamo riesca a mettere le mani su quella banda di farabutti. L'ambasciata sarà sempre a sua disposizione». Anche Rocco si alzò. «Perché ora sì?». Alejandro Giménez guardò Carmen che finalmente sorrise. «Sono le regole, dottor Schiavone. Non si sputa su un parente, a meno che qualcuno non ci dimostri che su quel parente faremmo meglio a sputarci». «Anche se voi già sapevate che era un bastardo?». «Anche se noi lo sapevamo già». «Chiaro e tortuoso, dottor Giménez». «Lo so. Un consiglio? Non intraprenda la carriera diplomatica». «È un rischio che non corro, mi creda». Aveva portato Lupa a fare la pipì per poi rimetterla in stanza a sonnecchiare. Risaliva via Garibaldi sul fianco del Gianicolo puntando verso Monteverde Vecchio. Roma era davanti a lui, rossa e gialla sotto un cielo azzurro solcato da piccole nuvole bianche di ovatta dove in mezzo ci correvano i gabbiani. La città ruttava con un rumore sordo e continuo, il mugghio di un branco di bisonti feroci. Avrebbe potuto prendere un taxi, invece decise che la giornata era bella e si meritava una passeggiata. Superò Porta San Pancrazio e proseguì su via Vitellia finché la strada ne incrociò un'altra, che scendeva ripida verso l'Aurelia Antica. Lì finiva il traffico e si sentivano solo uccelli. Sudato, Rocco proseguiva quella passeggiata. Era più lunga di quanto ricordasse, e ginnastica e moto erano solo un ricordo sbiadito. A destra e a sinistra solo recinzioni, nessun civico. In fondo sembrava ci fosse un cancello, e doveva essere di una villa perché ovunque girasse lo sguardo vedeva solo fronde di alberi e neanche una palazzina. Seguitò per altri trecento metri e finalmente arrivò al civico giusto. Un cancello di ferro. Spalancato. Il giardino enorme era curatissimo, seguì un viale che si perdeva fra alberi e cespugli per un pezzo, poi superati due pini arrivò alla casa. Sembrava un grande cottage, con le finestre bianche a quadretti. Davanti c'era un'auto con gli sportelli aperti. Non un cenno di vita. Rocco arrivò al patio, coperto da un tetto di tegole e legno. A terra dei cocci di un bicchiere. Le tre grandi porte-finestre erano spalancate. Delle tende bianche si muovevano pigre al vento. Entrò in casa. Il salone era gigantesco, non riuscì a contare la quantità di divani bianchi sparsi per la stanza. Un camino di ferro pendeva dal soffitto. «Lei chi è?». Stesa su uno dei divani, un bicchiere in mano, una donna lo osservava. Aveva passato la quarantina. I capelli lunghi e neri cadevano ricci sulle spalle. Le labbra carnose e la pelle ambrata, il corpo risultato di molto allenamento in palestra si mostrava in tutta la sua perfezione grazie a un vestito giallo che lo fasciava come una seconda pelle. Alzò il viso. «Ho chiesto lei chi è?» disse con una voce roca da fumatrice. «Vicequestore Schiavone». «Lo avete appena portato via. Che altro volete?». Rocco avanzò un paio di passi. «Ruben Montoro... lavorava per voi?». «Chi?». «Montoro. Faceva l'autista per suo marito?». La donna alzò appena la spalla. «... autista...» borbottò scettica. Poi raggiunse un tavolino basso di legno nero intarsiato, aprì un portasigarette d'argento. Se ne infilò una in bocca e l'accese con un accendino pieno di brillantini che lanciò sul tavolo. Soffiò il fumo e guardò Rocco. «Perché lo vuole sapere?». «Quello che mi piacerebbe sapere è il motivo per cui è morto col mio numero di cellulare in tasca, dal momento che non lo conoscevo». «Affari di mio marito» e con un gesto di sufficienza si voltò incamminandosi verso una porta col passo incerto di chi ha già bevuto troppo. «Vuole un caffè?». «Volentieri». «Anita!» gridò. «Ah, già, dimenticavo. Niente servitù. Sono spariti pure loro. Se vuole prenda pure da bere» e indicò un bancone con decine di bottiglie di liquori poggiati sopra. «No grazie. Lei non mi sembra molto triste...». «Dovrei?» e fece un altro tiro dalla sigaretta. «Ma la prego, si accomodi. Tanto oggi la palestra la salto. Diciamo che non è giornata...». «Suo marito stava partendo?» chiese Rocco indicando la valigia ancora poggiata sulle piastrelle in cotto del patio. «Scappando è il termine più appropriato» rispose la donna. «E non mi sono ancora presentata. Penelope». Rocco sorrise appena. «Scappava?». «Certo. Era davanti alla televisione, parlavano di quell'arresto di ieri notte, è impallidito e s'è messo a preparare i bagagli. Ma non ha fatto in tempo». Rocco si sedette su una poltrona di pelle. «Lei sa cosa faceva suo marito?». «Juan? Tante cose. Alcune pulite, altre molto sporche. E credo che lo abbiano cacciato per questo dall'ambasciata». «E lei?». «Lei cosa?». «Ci viveva tranquilla?». Scoppiò a ridere. Spense la sigaretta direttamente sul tavolino di fronte al divano. Poi scosse la testa. «Signor vicequestore, che Juan se ne andasse di casa era questione di giorni. Io e lui condividevamo questo umile tetto, niente di più» e con uno sguardo abbracciò il soffitto a cassettoni. «Sarà dura dirlo a mio padre, l'ha sempre odiato. Se l'immagina quanti te l'avevo detto! mi dovrò sorbire? Mio padre era un generale, sa? Tutto d'un pezzo! Come mio nonno e il padre di mio nonno, siamo una famiglia antica e un po' noiosa... e forse aveva ragione quando mi diceva di stare alla larga da quel Barrio... ma ero giovane, e da giovani si fanno sciocchezze». Con uno scatto della mano si portò i capelli all'indietro. Poi si alzò dal divano e riuscì ad arrivare al mobile dei liquori. «Juan e io facevamo due vite diverse» svuotò il restante whisky nel bicchiere e lasciò cadere la bottiglia sul tavolo di legno. «Mio marito si farà qualche anno di carcere, io pretenderò il divorzio, tornerò a casa mia, nel mio paese, e ricomincerò a vivere» e si lasciò cadere nuovamente sul divano felice di poter bere un sorso di liquore. «Cocaina» fece Rocco e si sedette su una poltrona. «Cosa?». «Suo marito trafficava in cocaina». «Già. Per i potenti...» e gli fece un occhiolino. «Se si mette a parlare ne sentirà di belle. Cocaina, sì. E io lo sapevo. Lei pensa che meriti il carcere anche io? Tutto sommato sì» e buttò giù un altro sorso. «Fino ad oggi guardi come ho vissuto, nel lusso. Sapevo da dove arrivavano quei soldi? Non era difficile intuirlo vista la gente che frequentava mio marito». «Tipo Montoro?». «Montoro... Ruben gli guardava le spalle, lo seguiva come un cagnolino fedele. Una volta era un militare, ma lo cacciarono dall'esercito. Un tipo violento, orribile. È morto?». «Lei non ne era a conoscenza?». «No. Erano giorni che non si vedeva in giro. Posso sapere come?». «Sgozzato». Penelope fece una smorfia che somigliava a una risata silenziosa. «Brutta morte». «Forse lei mi può aiutare, visto che è così gentile». «Non sono gentile, è che al momento non ho niente da fare e sono anche parecchio ubriaca». «Mi piacerebbe andare a dare un'occhiata allo studio di suo marito. C'è uno studio, immagino». «Faccia pure. Ma ci sono già passati i suoi colleghi. Sono suoi colleghi, vero?». Si guardarono in silenzio. Rocco si alzò dalla poltrona. La donna allungò una mano indicando una porta. «Di là c'è un corridoio, in fondo trova lo studio. Faccia con comodo. Io non l'accompagno». Tanto la casa era ampia, ricca, piena di quadri e mobili, tanto lo studio era microscopico, spoglio, essenziale. Una stanza monastica dove potersi concentrare senza essere distratti da orpelli e fantasie. C'era un piccolo tavolo sul quale giacevano cavi di un computer che era stato portato via. La libreria conteneva delle scatole di cartone aperte e svuotate, anche lì il contenuto era stato prelevato dai colleghi. A Rocco non rimase che aprire il piccolo cassetto dello scrittoio. Un caricabatterie, un vecchio portafogli pieno di biglietti da visita, penne, matite, qualche centesimo, un blocchetto di post-it ancora incartato, un temperamatite, un vecchio sigaro sbriciolato, un anello di poco valore, un contenitore di inchiostro per stampante vuoto, un dépliant di un centro sportivo. Prese il portafogli e cominciò a guardare i biglietti da visita. Autofficina Rosario, il pin e il puk di un cellulare, Marco Ierosi fabbro, Eusebio Valor giornalista dell'«Independiente», Motorcar, un carrozziere, un albergo Bellavista a una stella, Michele Servaggelli trasporti. Li controllò uno per uno. C'era una stonatura, l'aveva già percepita mentre li guardava con sufficienza, ma non aveva capito quale. Albergo Bellavista, una stella. Via Pontina km 35. Era una zona di Roma che Rocco conosceva bene. Quel posto non era lontano da Castel di Decima, la Pontina ci passava proprio accanto. Cosa doveva farci il ricco Gonzalez Barrio con un misero e pidocchioso hotel a una stella sulla Pontina? Penelope ancora sul divano parlava al cellulare. «Sì... sì, io ti aspetto. Vieni quando ti pare». Appena vide Rocco, chiuse la comunicazione. «Ha trovato qualcosa di interessante?». «Sì, grazie. Arrivederci, signora Barrio». «Gradirei che mi chiamasse con il mio nome vero, quello della mia famiglia. Barrio non lo voglio più sentire». «Se me lo comunica, l'accontento subito». «Penelope de la Vega y Fernández de Tejera». «Non credo di poterlo ricordare. Ma c'entra nel passaporto?». La donna scoppiò a ridere mostrando i bellissimi denti. «Allora mi chiami solo con il mio nome. Basta e avanza». «Arrivederci Penelope. Buona fortuna». «Sì, quella fino a oggi mi è mancata. Dottor Schiavone, lei mi ha fatto un sacco di domande, ora vorrei farne una a lei». «Prego». «Chi è veramente? Perché ci credo poco che sia un poliziotto». «Sono un vicequestore della polizia, anni fa ho indagato su suo marito senza successo, e soprattutto uno che, negli ultimi tempi, pare arrivi sempre un po' in ritardo». «Le va tutto storto, mi dispiace. Ha il viso così stanco, come mai?». «Avere a che fare con gente come voi, mi creda, è molto stancante. E il patto era una sola domanda. Tanti saluti...». Risaliva la strada con le idee più chiare nonostante avesse dormito poco o niente. Barrio arrestato per le rivelazioni di Enzo Baiocchi voleva dire che ci aveva visto giusto. I due si conoscevano, infatti stava emergendo il vecchio traffico di cocaina di Barrio e della banda di Luigi Baiocchi. Barrio aveva mandato un sicario ad eliminare Enzo, per zittirlo. E il numero di telefono che Ruben Montoro aveva in tasca? Forse Barrio, che era al corrente della storiaccia fra lui e i Baiocchi, l'avrebbe fatto contattare per chiedergli qualcosa in cambio. Tornò con la mente a Sebastiano. Il magistrato De Biase sembrava aver creduto alla storia dell'informatore, ma Baldi poteva metterci lo zampino. Tentato omicidio. Sebastiano rischiava il processo e il carcere. E la cosa peggiore era che l'amico lo considerasse il responsabile di tutto questo. Il discorso di Brizio aveva dato il colpo di grazia. «Un'aquila vola, un cavallo trotta». E un crotalo non è responsabile del suo veleno, pensò, e io ne sono portatore sano. Recuperò Lupa e la macchina e partì verso Roma sud. Percorreva la Pontina nel solito traffico. Superò Spinaceto, Castel di Decima e finalmente lo vide. Un edificio a due piani, un parallelepipedo rosso con un balcone proprio sopra un distributore di benzina. L'hotel Bellavista non era un posto per turisti, neanche per commessi viaggiatori, al massimo poteva funzionare come scortico di bassissimo livello. Parcheggiò nel distributore, lasciò Lupa in macchina ed entrò. Se la doveva giocare sporca. Odore di muffa. Sulle pareti a metà altezza correva una boiserie di plastica marrone. La reception era un bancone di finto marmo, una grata con una decina di chiavi appese sovrastata da una brutta riproduzione di due putti del Tiepolo che si abbracciavano. In un vaso di vetro colorato c'erano delle spighe di grano rinsecchite e impolverate annodate in un nastrino di raso. Le pareti erano ciclamino e a terra una guida putrida e piena di macchie attutiva il rumore dei passi. Lo specchio sulla destra era coperto da decine di cartoline incastrate alla cornice. Da una porticina apparve un uomo. Portava i capelli lunghi ai lati, sulla sommità del cranio invece ne era privo. Un paio di baffetti bianchi accentuavano ancora di più le guance risucchiate. Aveva una maglietta azzurra che mostrava le braccia piene di tatuaggi. Si asciugava le mani su uno straccio sporco di grasso di motore. «Dica...» fece serio. «Mi manda Juan» rispose Rocco. Quello lo squadrò dalla testa ai piedi. «Juan chi?». «Non fare il coglione. Lo sai». «Non sei uno sbirro?». «Si te do 'na crocca in mezzo all'occhi cambi idea o mi cachi ancora il cazzo?». «Te manna Juan?». «Sei sordo?». «Che voi?». «Non ho più notizie di Ruben» disse Rocco. L'uomo annuì. «...Vie' de qua...» gli disse facendogli cenno di seguirlo. Passarono dalla porticina dalla quale era entrato l'uomo, e attraverso un piccolo corridoio scrostato sbucarono in una stanza che voleva essere una cucina. Ma della cucina aveva solo la macchina del gas e due scansie di formica marrone. Uscirono da una porta-finestra per ritrovarsi in un cortile sul retro dell'hotel. C'erano due macchine impolverate, dei pallet accatastati e un vecchio carrello della Conad. Su un lato la bocca di un garage nero di fuliggine vomitava una musica gracchiata da una radiolina. «Che voi di' che non hai più notizie?». «Che voglio di'? Che non s'è più visto né sentito». «Ao', stamme a senti'. Diglielo a Juan, io le chiavi gliel'ho date. Quello il lavoro l'ha fatto. Poi che cazzo ha combinato io non lo so». «E dove l'ha fatto?». «In mansarda». «Portamece». «Perché?». «Fatte li cazzi tua e portamece!». L'uomo alzò gli occhi al cielo. «Annamo» disse. Rientrarono. Si fermò un attimo al bancone e prese delle chiavi da un cassetto. «Vie'» fece e cominciarono a salire le scale di marmo. Al primo piano si affacciavano una decina di stanze. Ma l'uomo proseguì. «Sopra» disse. Rocco lo seguiva in silenzio. La puzza di muffa s'era tramutata in benzina. Il rumore del traffico sulla statale era assordante. Le scale terminavano in una porticina. L'uomo la aprì e passarono in un piccolo disimpegno dove si affacciava una sola stanza senza numero. «Ecco...». La camera era un buco con un letto singolo, niente bagno. Era in ordine e la puzza di varechina torceva le narici. «Stava qua?» chiese Rocco. «Sì». «L'hai pulita?». «Eh?». «T'ho chiesto se l'hai pulita!». «Ao', cominci a cacarmi il caz...». Una gomitata secca lo colpì sul setto nasale. L'uomo cadde a terra mugolando. Rocco gli fu addosso e gli mollò un cazzotto in piena bocca. «T'ho chiesto se l'hai pulita!». Era una maschera di sangue. Cercava di tamponarselo, ma un altro cazzotto di Rocco sul naso lo fece quasi svenire. «Ultima volta, poi te apro! L'hai pulita?». «S... sì...». «Che è successo?». «Che cazzo ne so? C'era sangue dappertutto, mannaggia alla...». Rocco gli afferrò un baffo, tirò forte e strappò un ciuffo di peli. L'uomo gridò, poi cominciò a tossire per il sangue che gli andava di traverso. «Lo devi sapere che è successo!». «Quello m'ammazza». «Chi t'ammazza, coglione?». Gli stringeva la maglietta urlandogli a un centimetro dalla faccia ridotta una poltiglia. «Chi!». «Quello che stava... qui...». «Baiocchi?». L'uomo non rispose. Rocco non ne poteva più. «Ripeto: Baiocchi?». Non disse niente. Fece solo sì con la testa. Rocco lo lasciò andare sul pavimento. Poi andò al letto, strappò via la federa e gliela gettò. «Tiè, asciugate, pezzo de merda!». Subito il cotone bianco divenne rosso di sangue. «Perché non gliel'hai detto a Juan?». «Se parlavo tornava e m'ammazzava». «Non piangere, cazzo, e parla che te capisco. Baiocchi ti ammazzava?». «Sì... j'ha preso la macchina, l'ha caricato e se n'è annato...». «E tu zitto...». «Io ho imparato a famme i cazzi miei». «E hai fatto male. Come te chiami?». «Io?». «No, 'sta ceppa. Come te chiami!». «Demetrio». «Che nome è?». «Mi nonno se chiamava così». «Demetrio e poi?». «Demetrio Seganti. Ma perché, me conosci?». «Ma chi t'ha mai visto. Come esci da 'sta situazione Demetrio?» si sedette sul letto e si accese una sigaretta. «Come esco perché?». «Juan vuole Ruben. Tu l'hai visto cadavere e non hai parlato. Ma t'è annata male. Perché a Baiocchi se lo so' bevuto, invece noi semo qui. Hai tradito!». Demetrio deglutì. «Io non ho tradito, diglielo a Juan». «Ma non hai tradito una volta sola. Hai già cantato con qualcun altro, vero Demetrio?». «No». «Io dico di sì. Non so' il primo che viene a cerca' Ruben e Enzo, è così?». «Sei er primo invece». «Chi altro è venuto, Demetrio?». Ma l'uomo dell'hotel Bellavista non rispose. Rocco gettò la sigaretta a terra e si alzò di scatto. Demetrio si riparò la faccia. «Te giuro, io non te dico bugie. Non è venuto nessuno!». «Dimmi se è venuto uno grosso, capelli lunghi, barba che pareva un orso». «No». «È venuto sì o no?». Demetrio tremava. Poi disse un semplice «No» che cadde dritto nello stomaco del vicequestore strappandogli un sorriso. Guardò la faccia tumefatta di Demetrio, si alzò e prese il cellulare. «Pronto? Bonanni sei tu?». «Schiavone caro, dimmi tutto!». «Hotel Bellavista, al chilometro...» si rivolse a Demetrio: «Che chilometro è questo?». «35» rispose con un filo di voce. «Chilometro 35, c'è la stanza dove hanno ammazzato Ruben Montoro. Il testimone oculare è Demetrio Seganti...» lo guardò di nuovo «sei il proprietario tu?». «Sì...». «Che è il proprietario di questo grand'hotel». «Grazie Schiavone. Chi l'ha ammazzato?». «Baiocchi Enzo, arrestato stanotte vicino Cividale del Friuli!». «Ma... puoi provarlo?». «Quando vuoi. Ma ti basta il testimone qui...». «Schiavone, come ci sei arrivato?». «Storia lunga. Stammi bene, Bona'...» rimise in tasca il cellulare. «Ciao Demetrio...». «A te nun te manna Juan!» fece quello continuando a tenere la federa ormai rossa premuta sul naso. «Hai un certo intuito...» e prima di uscire dalla stanza si raccomandò: «Mo' vengono i colleghi. Tu digli le stesse cose che hai detto a me. E soprattutto che non hai parlato perché hai avuto paura di quei banditi. Ne esci pulito, Demetrio, non aver paura perché Barrio è dentro e pure Enzo Baiocchi. Anzi, fai la figura del bravo cittadino». Stava per uscire dalla stanza, ma si fermò ancora. «Dimenticavo. Avverti tutte le zoccole che ospiti al momento di sparire di corsa sennò te fai qualche mese dentro». «E la faccia? Che je dico? Che so' cascato?». «No, digli pure che sono stato io. Sai quanto me ne frega? Stammi bene e curati» e finalmente se ne andò. «Grazie» disse Demetrio e si sentì un deficiente, come nonno buonanima non si risparmiava mai di ricordargli. La schiena spezzata, gli occhi pesanti, in bocca un sapore dolciastro e zuccheroso, le palpebre di piombo lasciò la macchina sul lungotevere dove avrebbe preso sicuramente la multa. Ma di pensare non era più tempo, solo dormire. E lo fece cadendo in un pozzo senza fondo. Ne riuscì solo il giorno dopo. Parcheggiò l'auto davanti all'ingresso della questura di Aosta che il sole stava tramontando. Barcollò fino alla sua stanza e chiuse a chiave la porta, si gettò sulla poltrona aprendo la finestra e finalmente si accese una canna. Non si era mai sentito così nella vita professionale. Forse solo quando era in attesa del suo trasferimento. Non erano le umiliazioni subite sul lavoro, la sconfitta di non aver chiuso il cerchio su Barrio all'epoca dei fatti di Luigi Baiocchi, quelle gli pesavano poco o niente. Dopo quel viaggio massacrante si portava sulle spalle tre sconfitte, amare e definitive. Aveva perso la fiducia di Sebastiano, e quella era la prima sconfitta. Gli occhi dell'amico, carichi di odio, il risentimento, le parole terribili che gli aveva sparato in faccia, assai più dolorose di pugni e gomitate. La seconda era il discorso di Brizio del giorno prima, che aveva tutta l'aria di un commiato, comunque un pezzo di vita che non c'era forse più. La terza sconfitta era quella di sentirsi controllato ad ogni passo. Diventare un batterio su un vetrino era una sensazione che a Rocco non piaceva per niente. A quello pensava fumando la canna, senza che l'erba riuscisse ad alleviare le ferite. Qualcuno abbassò la maniglia della porta tentando di entrare. Rocco non gli diede peso. Ma quello insisteva, prese a bussare istericamente. Si alzò, gettò il mozzicone dalla finestra e andò ad aprire. C'era Italo, pallido e con i capelli spettinati. «Che c'è?». «Dove sei stato fino a oggi?». «In giro. Che vuoi?». «Hanno arrestato Baiocchi» e gli mollò il giornale. Al centro delle pagine della cronaca, un bell'articolo della Buccellato che Rocco non volle leggere. Restituì il quotidiano all'agente. «Lo so. L'ho visto in televisione». «C'è Baldi». «Dove?». «Giù. Davanti ai caffè. Ti aspetta e dice che tu lo sai». Era la prima volta che il magistrato si scomodava per andare da lui. «Scendo» e lasciò la stanza. Anche Baldi era pallido e spettinato, aveva le occhiaie e la barba incolta. Appena lo vide gli andò incontro. «Andiamoci a fare un giro» propose senza aspettare risposta. A Rocco non restò che seguirlo. Appena usciti dalla questura Baldi allungò una mano: «Mi dia una sigaretta» gli disse. «Lei non fuma...». «Lo so». Se l'accese e non tossì. Dal modo in cui aspirava si vedeva che il giudice un tempo era stato un buon fumatore. «Perché ricomincia?» gli chiese Rocco. «Io non ho intenzione di mollare» esordì. «Su Valenti, voglio dire, non con le sigarette. Non ho intenzione di mollare. Ma devo andarci coi piedi di piombo. Ho riflettuto sul nostro incontro a Cividale, e va bene così. Ma da oggi le chiedo di essere pulito e trasparente con me. È un patto che si sente di fare?». «Dottore, io sono sempre stato pulito e trasparente». «Bene, è un patto che non ha intenzione di suggellare». «Cosa le devo dire per dimostrarle il contrario?». «Per esempio perché Enzo Baiocchi ce l'ha con lei. Cosa le ha fatto?». Rocco alzò gli occhi al cielo. «E non è possibile, è la sedicesima volta che me lo chiede! Gliel'ho detto, ho incastrato suo fratello anni fa, quello è scappato e ci ha lasciato la pelle in Sudamerica. Lui mi ritiene responsabile. Tutto qui». «Vabbè» fece Baldi sputando fuori il fumo. «Buona!» e gettò via mezza Camel. «Stanno facendo parecchi arresti, grazie alle rivelazioni di quel bandito». «Così pare». «Quindi le gole profonde servono sempre, non crede?». «Spie e collaboratori aiutano, dottor Baldi. Lo sappiamo entrambi. Guardi cosa è successo qualche giorno fa con Valenti». «Esatto. Parliamo di lui. Come le ho detto, aspettavo un po' di conferme. Ho acquisito informazioni». Attraversarono la strada. «È molto protetto, questo ormai lo abbiamo capito. Quello che non sapevamo era il suo vero nome. Pasquale Iovine, nato in provincia di Avellino nel luglio del '49. Fino al 2011 legato al clan Barbieri, collaboratore di giustizia da quella data. Almeno adesso abbiamo un quadro più preciso». «Che non ci serve a un cazzo» rispose Rocco. «Sputa!». «Come?». «Scusi, dicevo al cane. Sputa!». Lupa depositò a terra un pezzo di ossobuco scovato chissà dove. Il vicequestore lo raccolse e lo gettò in un tombino. «Dove l'hanno portato a questo Iovine?». «Non ci sono arrivato. Da qualche parte sotto altro nome. E se lo terranno stretto finché inizia il processo a quel clan che controlla un bel pezzo del casertano». «Quindi che cosa vuole fare?». «Io? Aspetto che non serva più. Poi colpisco duro». «Dottor Baldi». Rocco si fermò in mezzo alla strada. Le macchine sfrecciavano lì accanto. «Quando questo accadrà, lei sarà in pensione, io chissà dove e del delitto di Juana Perez non se ne avrà più memoria. È stato un incidente e a noi è toccato andarci di mezzo. Lei mi ha sempre parlato del dovere, del rispetto morale verso le istituzioni. Ecco, applichi le sue prediche». «Siamo di fronte a un criminale che deve rispondere di un delitto!». «Ma i suoi superiori hanno già deciso per lei, e hanno decretato che un trans non vale un processo di camorra. Se ne faccia una ragione. Io ci sto provando». «Quindi lei chiude gli occhi?». «Io non chiudo gli occhi, Baldi. Non l'ho mai fatto. Ma a differenza di lei capisco quando la partita è persa». «Sono un magistrato, Schiavone! E conosco l'amaro calice della sconfitta!». «Questa partita era persa in partenza!» anche Rocco alzò un poco la voce. «Cosa pretende da me? Che raccolga ancora prove contro Bernardo Valenti o Pasquale Iovine o come si chiama quel vecchio alcolizzato con il suo volpino del cazzo? L'ho già fatto, ce n'è da riempire un camion con le prove. Ma le mie capacità si fermano qui. In mezzo a quel mare io non so navigare. Sono solo un vicequestore, niente di più. E lei è un magistrato del tribunale di Aosta». Poi i due uomini si guardarono per qualche secondo. «'Ndo cazzo annamo, dottor Baldi?». Il magistrato strinse le mascelle e abbassò lo sguardo. Si guardò le scarpe. «Almeno facciamola pagare a chi l'ha protetto!». «L'ha protetto lo Stato, i servizi, il ministero degli interni, quelli che stanno molto sopra di noi. Cosa crede, che a me faccia piacere prenderla in quel posto? L'avevamo sotto il naso, l'assassino, e ce lo siamo fatti scappare. A quest'ora chissà dov'è, come si chiama, dove abita e che vita conduce. Una cosa sola le garantisco, per quel poco che vale. Dovesse ricadere nello stesso errore non ci saranno servizi o ministeri a proteggerlo, ovunque sia lo inchiodo». «Magari costerà un'altra vita umana». «Ha ragione, ma non sono io che ho fatto le regole di questo gioco di merda». Baldi cercò di pettinarsi i capelli passandoci una mano. «Le va di farla pagare almeno a Berthod, il proprietario dell'appartamento?». «Certo. Le prometto che domattina vengo da lei con una bella idea e gli facciamo passare un brutto quarto d'ora». «Sono d'accordo. È un po' poco... ma è meglio che niente». Si lasciarono senza salutarsi, tornando alla loro vita di ogni giorno, con la rabbia in corpo e la frustrazione sulle spalle come fossero mostrine di un generale. Spense il cellulare, non se la sentiva di parlare con nessuno. Voleva rimandare tutto al giorno dopo. Il questore, la squadra, le bugie, tutto. Era troppo stanco, era ora di tornare a casa a stendersi sul letto e farsi sorprendere dal sonno. Risaliva lento le scale cercando dal mazzo la chiave giusta. Arrivò sul pianerottolo ma prima che riuscisse ad entrare la porta del vicino si aprì. «Dottore?». Era Gabriele. Stavolta aveva optato per una maglietta di qualche squadra di football americano e calzoncini da tennis. Ai piedi sempre le stesse scarpe sporche e slabbrate. «Che vuoi?». «Lo so che è stanco ma...». «Niente latino, niente dialogo col preside, niente professori e niente mamma, per favore». Gabriele sorrise. «No, è che... lei è andato via e non mi ha ridato il computer». «Ah no?». «No...». «Scusami. Dai, vieni dentro che te lo riprendi» e carezzando la testa di Lupa entrò in casa. Il computer era ancora lì, chiuso. Mentre Rocco si toglieva la giacca il ragazzo lo afferrò. «Ah, l'ha lasciato in carica, meglio!» disse. Rocco andò a controllare la camera da letto. Caterina aveva rimesso tutto in ordine. S'era portata via il biglietto che le aveva lasciato. «Ha già cenato?» sentì dall'altra stanza. «No Gabriele. Non mangio. Voglio solo andare a letto» e tornò in salone. «Tua madre è a casa?». «Sì. È appena tornata. Ci vuole parlare?». «Non ce la posso fare». Si avventurò in cucina per dare la pappa a Lupa. «Lo sa? Diego l'hanno dimesso dall'ospedale. Non aveva niente». «Chi è Diego?» chiese Rocco bevendo acqua direttamente dalla bottiglia. «Quello che ho picchiato. Ora mi sta lontano, lo sa? L'ho incontrato ieri alla sala giochi, neanche s'è avvicinato. Io ero pronto comunque, con il calcio alle palle e il martello di Odino!». «Bravo. Ora se non ti dispiace vorrei andare a dormire. Sono distrutto». «Vado, vado...» finì di avvolgere il filo del caricabatterie del computer intorno alla mano. «Se domani non ha da fare...». «Ecco, sì, magari. Ci vediamo domani, Gabriele. E scusa per il computer. Me lo sono dimenticato». «Non si preoccupi. Ho risparmiato qualche diottria». «Ascolta, l'hai detto a tua madre?». «Cosa?». «Della rissa?». «No. Muto come un pesce» e facendo l'occhiolino uscì dall'appartamento. Si tolse le scarpe, i pantaloni, spense la luce e andò a gettarsi sul letto. Il riverbero di un lampione stradale dall'esterno colpiva il soffitto producendo strani disegni. Si perse a osservarli. Un cavallo che rampava, o forse era un uomo in guardia davanti a un avversario coi pugni chiusi e pronti alla scazzottata. Chi mi ha tradito? «Mi sa che ho perso il ritmo, Marina. Pare che nessuno è quello che è. E io mi sono stancato di inseguire le ombre. Non faccio altro da giorni. L'unica che vorrei prendere sei tu, ma non ti fai più vedere. L'hai detto, te ne sei andata, e non torni più». Vorrei fare una doccia, ma non ce la faccio manco ad alzare un braccio. Se la casa andasse a fuoco rimarrei sul letto, come un vecchio guerriero vichingo, a bruciare con tutta la barca. Sono rimasto io e un cane. Bella prospettiva, no? Mi prendesse il sonno, almeno, me ne andrei per un po'. Ha ragione Baldi, altroché. Ci hanno preso per il culo e noi ce ne stiamo zitti ad abbozzare. Ed è pericoloso, perché neanche riesco ad arrabbiarmi su 'sta cosa. Invece dovrei. Un coglione che si mette a giocare con le corde, a fare il giapponese, e quella che ci lascia la pelle. Ma sono stanco. Sono troppo stanco. Sul cuscino accanto c'era ancora il profumo di fiori che aveva lasciato Caterina qualche giorno prima. E non gli diede una sensazione piacevole. «I still don't know what I was waiting for / And my time was running wild / A million dead-end streets...». Eccolo Gabriele, rimette il disco, sa che lo sto ascoltando. Chissà, magari un giorno potremo parlare così. In faccia ci diciamo poco o niente, poi lui torna nella cameretta coi suoi film horror e mi manda un messaggio. Va bene Gabriele, bisogna cambiare, io questo l'ho capito, pure se ti ritrovi in mezzo a incroci con milioni di binari morti, ci devi provare lo stesso. Hai 16 anni tu. Io un po' di più. C'era una poesia, com'era? Del poeta e del bambino che aveva paura a chiedergli l'età. T'avrei risposto ottanta, come tu mi hai detto tre. Era più o meno così. «Ch-ch-Changes / Turn and face the stranger, / Ch-ch-Changes / Just gonna have to be a different man». I conti si faranno alla fine. A destra la colonna col segno più, a sinistra quella col segno meno e in mezzo quella enorme e vuota delle intenzioni e dei rimorsi. Abbozzò un sorriso e si addormentò. Quando si svegliò il sole era già alto. Ancora un sonno buio e senza interruzioni che non gli aveva dato riposo. Anzi, se possibile, si sentiva più stanco del giorno prima. Anche se aveva saltato la cena fame non ne aveva. Lupa sì. Già scodinzolava pronta. Si alzò e le dette la pappa. Gli toccava andare in ufficio, per quale motivo non lo sapeva. Per parlare con il questore? O per ascoltare Michela Gambino affermare che Jim Morrison non era mai morto e si era trasformato in Barry Manilow? Doveva venirgli un'ideuzza per farla pagare a Berthod, come d'accordo con Baldi. Avrebbe cominciato con gli affitti in nero per poi inchiodarlo con qualcos'altro. Prese il cellulare. Appena entrò nella sua vecchia stanza ci fu un ululato che lo fece sobbalzare. Erano tutti lì. Italo, Antonio, Caterina, Casella e anche i fratelli De Rege, al secolo Deruta e D'Intino. Tutti a battere le mani sorridenti, pronti con lo spumante che Antonio sturò facendo volare il tappo addosso a D'Intino. «Mi sposo!» gridò quello mentre batteva le mani. «Che è?» chiese Rocco. «Per Baiocchi! L'hanno preso, dotto'!» urlò Casella. «Brindiamo!» gridò Italo mentre Antonio versava lo spumante nei bicchieri di plastica. «Che possa marcire dentro, quel figlio di buona donna» andava dicendo l'agente siculo marchigiano. «A me poco» fece Deruta. «No, perché ho passato la notte al panificio e...» ma nessuno gli diede retta. Riempiti i bicchieri Caterina invitò tutti a un brindisi. «Alla squadra, al futuro!». «Grazie, grazie» fece Rocco. «A tutti noi!». «La Buccellato la cerca per un'intervista, dotto'!» fece Antonio che davanti agli altri tornava al lei. «E falla cerca'... non ho niente da dirle». «Anche Costa vuole vederla» aggiunse Italo. «Si vuole complimentare». «E di che?» obiettò Rocco. «Io non ho fatto niente!» e guardò Caterina che gli sorrideva con gli occhi felici, luminosi. «Però è un pericolo scampato, no?». «Pura fortuna. L'hanno preso quelli vestiti di nero, quelli che sanno sempre tutto e agiscono per la nostra sicurezza e incolumità». «Noto una certa ironia, dotto'». «Dici, Antonio? Parrà a te. Be', adesso al lavoro, compagni. D'Intino e Deruta prendete!» andò alla scrivania, afferrò tre evidenziatori e li consegnò nelle mani dei due agenti. «Questi sono evidenziatori. Io ve li consegno per le pratiche di oggi. Mi raccomando, fatene buon uso!». Deruta e D'Intino guardarono i grossi pennarelli colorati, li afferrarono e con un rapido gesto del capo salutarono per uscire dalla stanza. «Perché fanno così con questi evidenziatori?» chiese Antonio. «Non si sa, Scipioni» rispose Rocco, «nessuno può indagare le pieghe della mente dei nostri fratelli De Rege. C'è qualcosa di ancestrale in questo attaccamento agli evidenziatori. Ma nessuno saprà mai cosa. Casella! Tu oggi sei di riposo». «Dice veramente?». «E che scherzo? Vattene all'ufficio passaporti, al bar, fatti un giro per negozi. Tu oggi ti godi la giornata». «Grazie, veramente, ne avevo bisogno». Prese il berretto e salutò: «Con permesso...». Rocco era con la sua squadra. «Noi invece ci divertiamo un po'. Dobbiamo far vedere i sorci verdi a Carlo Berthod. Una piccola vendetta ci vuole, no?». «Cos'hai in mente?». «Una perquisizione nel negozio di Courmayeur...». «Una perquisizione?». «Già. Una cosa facile facile. Quello ha affitti in nero, schifezze varie e chissà quante ne combina col negozio. Io vado dal magistrato a prendere l'occorrente. Voi mi aspettate davanti al tribunale. Due macchine. Io con Italo e tu Caterina viaggi con Antonio». «Ricevuto» fecero i poliziotti quasi in coro. «Niente di personale, vero?» aggiunse Caterina. «Personalissimo, invece!». Prima di uscire dalla stanza Caterina si avvicinò: «Poi stiamo un po' insieme?». Diretto al tribunale non si aspettava di incontrare Costa davanti al portone. Sembrava fresco e rilassato, la bella copia del questore dell'ultimo incontro. Sorridente si avvicinò per stringergli la mano che Rocco accettò con una certa riluttanza. «Non l'ho vista. Pausa di riflessione?». «Diciamo di sì». Un agente passò svicolando e salutò sottovoce i superiori. «Pare che anche quel Baiocchi sia molto interessante per i magistrati. Questo non fa che convincermi che alla fine con Valenti abbiamo agito nel migliore dei modi, secondo coscienza». «Dottor Costa, io non andrei a scomodare la coscienza. Diciamo che abbiamo agito come ci è stato chiesto di fare» rispose Rocco guardandolo negli occhi. «Sia come sia, a me resta un sapore di muffa in bocca, e non è piacevole. Detto questo l'argomento l'abbiamo già sviscerato in lungo e in largo». «Già...» il questore guardò lontano, come se cercasse l'ispirazione. «Baiocchi ne avrà da dire anche su di lei?». «Può dire tutto quello che gli pare. Ma vede, anche quando questa gente di merda si veste da santarellino e comincia a collaborare con la giustizia, ci sono le pinze...». «Le pinze?». «Sì, quelle con cui si dovrebbero prendere le parole che escono da quelle fogne. Va tutto controllato, provato. Baiocchi di me potrà dire il peggio. Ma c'è un fatto incontrovertibile, dottor Costa, lui si è introdotto in casa mia, lui ha sparato ad Adele Talamonti, non io». «Già. Ma, come dice lei, sempre attenendoci ai fatti, io vorrei conoscere il movente di quell'omicidio». «Ci gira intorno da tempo. E non si vuole arrendere» aprì l'anta di vetro per uscire dalla questura. Il rumore delle auto aumentò di intensità. «Ne ha la facoltà» gli disse. «Ora se non le dispiace, vorrei tornare al lavoro». «Prego...» e con un gesto del braccio lo invitò ad uscire. «A proposito» lo richiamò, «la presenza del suo amico Sebastiano Cecchetti sul luogo della cattura, come la vede?». «Ha fatto prima di me. E sa perché? Lui ci ha messo il cuore, io solo il raziocinio». «Lei l'ha spacciato per un suo informatore». «E lo è. Lo è stato e spero continuerà ad esserlo. Gli informatori, come i pentiti, servono. Oppure facciamo due pesi e due misure?». Costa allargò le braccia, poi si strinse il nodo della cravatta. «Schiavone, sa cosa vedo quando la guardo?». «Me lo dica...». «Un'auto che sbanda sul ghiaccio e che, miracolosamente, riesce a tenere la strada. Ma non sempre va così». «Ce l'ha messo lei il ghiaccio?». «No, Schiavone, è sulla strada. È un fatto naturale». «Allora se non è stato lei io dormo tranquillo. A presto». Rocco uscì dalla questura. Baldi era già pronto con l'autorizzazione. Felice di dare del filo da torcere a Berthod, un toccasana per la sua autostima che al momento viaggiava sotto il livello stradale. Impiegarono meno di mezz'ora per raggiungere Courmayeur. Parcheggiarono proprio davanti al negozio Mont Blanc in pieno divieto di sosta. Il vigile stavolta neanche si avvicinò. Entrarono nel negozio. Appena la commessa vide i quattro poliziotti impallidì e corse a chiamare il principale. Carlo Berthod, sguardo fiero e pomposo, si avvicinò con un sorriso falso stampato sul viso. «Ci rivediamo, dottor Schiavone!». «Ci rivediamo. Come va?». «Io bene. A cosa debbo l'onore?». «A questa!» e mostrò il mandato del magistrato. «Dovremmo dare un'occhiata in giro, se non le dispiace». «Mi dispiace sì...» fece il commerciante guardando il mandato. «E sticazzi!» fece Rocco. «Giusto, Italo?». «Sì, l'hai usato correttamente». «E posso sapere il motivo?». «Si ricorda quando le ho detto che non mi importava del fatto che lei affittasse in nero?». «Embè?». «Bugia. Mi interessa invece. Forza ragazzi, prima ci diamo una mossa prima torniamo a casa». Antonio e Italo si diressero verso la cassa. Caterina restò accanto alla porta. Rocco invece superando Berthod gli chiese: «Ha anche un ufficio in questa splendida boutique? Posso dare un'occhiata alle fatture?». «Mica siete della finanza!». «Non si opponga, Berthod. Se un vicequestore le chiede una fattura, lei la fattura la tira fuori. E non si preoccupi: detto fra noi, ora che i ragazzi non mi sentono, è routine...». «Cosa vuole da me, Schiavone?» gli sibilò tra i denti. «La storia di Juana Perez è chiusa, mi sembra». «Appunto, le sembra. Forza, tempo non ne abbiamo!» e si diresse verso un angolo del negozio che immetteva in una stanza piccola e piena di cartelline, libri e faldoni. «Allora eccoci qua... Intanto signor Berthod, dia al mio viceispettore il quaderno con gli affitti che ci ha mostrato l'altra volta. Quello nero». «Non ce l'ho!». «Dobbiamo trovarlo noi?». Berthod scosse la testa. «Non finisce qui, stronzetto» mugugnò, e se ne andò alla cassa. Afferrò il quaderno e lo consegnò ai poliziotti. «Ecco qua! Contenti?». Antonio lo prese e cominciò a sfogliarlo. «Che spera di trovarci, agente? La data della sua fine carriera?». «Oh porca...» sentirono gridare Rocco dall'ufficetto. «Che succede?». Italo e Antonio accorsero, anche Caterina e Berthod. Chiudeva la fila la giovane commessa ancora sconvolta. Rocco era in piedi, di spalle all'entrata. «Caro Berthod, si ricorda le tanto millantate amicizie molto in alto che lei possiede?» poi si voltò. In mano aveva una busta gonfia di marijuana. «E mi sa che qualche telefonata lassù la deve fare!». Alla vista della busta, Antonio e Italo sorrisero contenti. Caterina sgranò gli occhi. La commessa si mise le mani davanti alla bocca. Berthod era bianco come la vetta del monte di cui il negozio portava il nome. «Ma... ma... non è roba mia!». «E com'è che stava qui, fra due faldoni di fatture dell'anno in corso, peraltro vuoti?». «Io non uso quella robaccia!». «Va bene. Lo spiegherà al giudice. Antonio, Italo, prendete questo galantuomo e portatelo in questura». «Mi sentirà, cazzo se mi sentirà. Ce l'ha messa lei!». Rocco divenne serio. «È sicuro di quello che ha appena detto? Ci sono quattro testimoni. Lei mi sta accusando di aver messo questa busta nel suo ufficio? Ci pensi bene, Berthod, fa il paio con il tentativo di corruzione dell'altra volta...». «Io la farò spedire in Barbagia!». «Dice? Io la Sardegna la amo... forza, muoviamoci, mica abbiamo tempo da perdere. Signorina!» salutò la commessa mentre Italo e Antonio scortavano il proprietario del Mont Blanc fuori dal negozio. «Ma che ti salta in mente?» gli disse Caterina viaggiando verso Aosta. «Niente. Un piccolo scherzetto che gli costerà in avvocati, ci rimetterà la faccia, e gli tolgo un po' di ore di sonno. Mica male, no? Mi pesa solo dover rinunciare a quasi un etto di quella roba che, credimi, è meravigliosa!». Caterina scoppiò a ridere. «Quanto la vuoi portare avanti?». «Lo facciamo solo spaventare un po'. Fondamentale è spargere la voce per Courmayeur che il nostro amico è sospettato di possesso illecito di sostanze stupefacenti. Basterà a dargli un sacco di problemi nella sua piccola e laboriosa comunità. Vedi Caterina, se avesse messo le mani su qualche appalto, se avesse giocato sporco con forniture e simili, se avesse rubato, parlo politicamente, nessuno dei suoi amici col cashmere avrebbe avuto niente da ridire. Così fan tutti, no? Invece la maria... che cafonata! Che scivolone per un membro del Rotary...». «Certo se lo merita... Io però non so se me la sento» fece Caterina. «E che? Adesso ti fai venire i rimorsi di coscienza?». Rocco si accese una sigaretta. «Non è il momento. Ricordati come ce l'hanno messa in quel posto qualche giorno fa! Non è stata una sensazione piacevole». I due poliziotti non parlarono più fino all'arrivo alla questura di Aosta. Camminavano per corso Battaglione Aosta, Rocco serio e pensoso, Italo che lo guardava e non capiva perché l'avesse voluto accanto mentre si dirigeva all'appuntamento con la Buccellato. Un camion rumoroso passò, veloce e assordante. Trasportava tronchi d'albero. A Italo parve che Rocco avesse aperto la bocca, ma non aveva sentito quello che gli aveva detto. «Italo» gli disse a voce più alta. «Ma da quando sono arrivato ad Aosta, tu non hai notato niente di strano?». «Ma di che parli?». «Ascoltami bene, io di te mi voglio fidare. Mi conosci, e allora rispondi a questa domanda: chi è che mi controlla?». La domanda scese pesante nello stomaco dell'agente Italo Pierron. «Ti controllano?». «Ne sono certo. Vuoi sapere cos'è successo l'altro giorno vicino Cividale del Friuli?» e gli raccontò l'arresto di Enzo Baiocchi. Italo non ci credeva. Scuoteva la testa. «Com'è possibile?». «È possibile solo se i miei movimenti sono stati controllati da qualcuno. Ogni volta che vado a Roma per gli affari miei, qui lo vengono a sapere. A grandi linee sanno perché vado e cosa combino. Lo sanno in questura e in tribunale, lo sanno Costa e Baldi. E allora ti chiedo: hai notato niente?». «Mamma mia, Rocco, non è che la vicinanza della Gambino sta avendo una brutta influenza?». Rocco si fermò in mezzo alla strada. «Italo, non sto scherzando». Il volto e il tono di voce severo scossero Pierron. «Io non ho mai notato niente. Adesso ci penso, cerco di capire, ma non ho mai notato niente. Tu qualche sospetto ce l'hai?». «No. Io non ce l'ho. Non mi sono mai guardato le spalle qui ad Aosta. E ho fatto male». Ripresero a camminare. «Dopo quello che è successo con il caso Perez non mi stupisco più di niente» disse Italo. «Voglio dire, quella gente che ci ha insabbiato l'indagine viene da Roma, tu vieni da Roma, è possibile che ti abbiano messo gli occhi addosso dal primo giorno che sei sbarcato qui». «Ormai ne sono certo». «E allora se sei controllato, se questo ordine viene dalla questura centrale o peggio dagli interni, io guarderei in alto» suggerì Italo. «Stai pensando al questore?». «Anche. E a Baldi. Lui ha sempre chiuso un occhio, dava la sensazione di sapere sempre più di quanto fosse lecito. Non ci hai mai pensato? Ricordi la storia dei cingalesi che portammo a Torino?». «Sì...». «Ricordi anche la prova di Enzo Baudo?». «Anche». «Ecco, io ho sempre pensato che Baldi ti lasciasse fare, ma sapesse molto di più». «Io sono sicuro che c'entra Baldi. Ma... c'è un ma... deve avere qualcuno dentro la questura che mi sta accanto e che mi guarda da vicino». «E chi?». «Tu?». Stavolta fu Italo a bloccarsi in mezzo alla strada. «Ma sei matto?». «Tu di me sai tutto. Potresti essere tu la spia». «Rocco, ma non dire cazzate, per favore!» e riprese a camminare. «Ti sei offeso?». «Vaffanculo». «Ti sei offeso». Fecero un po' di passi in silenzio. «Deficiente, scherzavo» gli disse Rocco e gli mise un braccio intorno alla spalla. Il primo gesto di vero affetto dopo tanti mesi. Italo da principio lo scansò ma poi gli tornò il sorriso. Sandra Buccellato lo aspettava al tavolino del bar. «Che vi porto?» chiese Ettore un po' stupito nel vedere Rocco accanto alla donna che diceva di disprezzare. «Una spremuta» disse la Buccellato. «Io il solito caffè, Ettore». Quando il proprietario del bar andò via, Rocco guardò la giornalista negli occhi. «Ho qualcosa da raccontarle che mi farebbe piacere vedere nelle pagine del giornale». «Sono tutta orecchi». «Abbiamo fermato Carlo Berthod, lo conosce?». «Certo che lo conosco. E posso sapere perché?». «Abbiamo trovato un bel po' di marijuana nascosta nell'ufficio del suo negozio. Eravamo andati lì per una perquisizione, sempre sul caso Juana Perez, per capire alcuni dettagli, e per caso ci siamo imbattuti in quel sacchetto». «Oh oh» sorrise la Buccellato e gli occhi le luccicarono. «Bene, benissimo. Non le nascondo, Schiavone, che mi fa piacere. Era ora che qualcuno bucasse quel pallone gonfiato... e, se non è indiscreto, conosce il nome dell'avvocato che lo difende?». «Petrulli, ovviamente». «Perché ovviamente?». «Suo figlio, affittuario di Berthod, è stato indagato all'inizio, sempre per il caso Juana Perez». «Benissimo. Avrà l'articolo, però...» la frase della donna fu interrotta dall'arrivo di Ettore che poggiò le ordinazioni sul tavolo, lasciò lo scontrino e filò via senza neanche uno sguardo. Rocco prese la sua tazzina. La Buccellato non sfiorò la spremuta. «Però voglio una cosa in cambio». «Quale?». «Il caso era chiuso e lei è andato da Berthod solo per sputtanarlo. Lo capisco e l'appoggerò. Mi dia la sua versione su Juana Perez». «Gliel'ho già detta». «No, lei mi ha riferito pari pari quella del questore. Io voglio la sua, personale». «Paul De Vries abitava a Châtillon da due mesi. Questo è stato appurato. Era un ricercato. Gli piaceva andare coi travestiti, ha tirato fuori la pistola nel momento sbagliato, è morto, era il colpevole». «Quali prove a suo carico?». «Impronte digitali in casa della Perez». «Strano...» fece la giornalista e finalmente afferrò il bicchiere. «La sua collega, Michela Gambino, mi ha detto tutt'altro». «Tipo?». Abbassò il bicchiere. Aveva due baffi arancioni che asciugò con un tovagliolino di carta. «La casa era stata svuotata e non c'erano tracce di impronte digitali neanche a pagarle». «È stata chiamata una seconda squadra da Torino, evidentemente più affidabile della Gambino. Lo chieda a Farinelli, è il capo. Lui ha trovato le evidenze». Sandra Buccellato finì la spremuta, posò il bicchiere. «Sciocchezze» disse. «Può darsi che la verità assuma dei tratti incredibili e a volte fantasiosi, ma è andata proprio così». «Le dico come la vedo io?». «Sicuro». «Voi avete scoperto il vero assassino di Juana Perez, ma lo tenete nascosto per motivi a me ancora poco chiari. Avete trovato un colpevole qualsiasi e gli avete affibbiato l'omicidio. Mi dica se sbaglio». Rocco finì il caffè. «Sbaglia. Ma è un paese libero, così dicono, e lei queste cose, se vuole, può scriverle sul suo giornale». «E mi becco qualche querela?». «Non da me» l'assicurò Rocco. «C'è altro?». «Enzo Baiocchi. L'hanno preso». «Un problema in meno». «Ma non l'ha preso lei». «No». «Peccato» fece la Buccellato. «Già. L'importante è che l'abbiano messo dentro. Sta parlando, stanno facendo arresti, insomma si è rivelato un tesoro, no? Sono importanti i pentiti e i chiacchieroni, per la giustizia». «Cosa sta cercando di dirmi?». «Semplicemente quello che le ho detto» poi il vicequestore prese il portafogli, lasciò una banconota sul tavolino e si alzò dalla sedia. «Buon lavoro, signora Buccellato» e si diresse verso il margine della piazza dove lo aspettava Italo col giornale in mano. I due si incamminarono verso la questura. Si era fatto la doccia, una camicia pulita, si era sbarbato e spruzzato il profumo, ma la faccia era sempre quella. Almeno aveva l'aria pulita. «Vabbè Lupa, sono così, che ci vuoi fare?». Aveva dato appuntamento a Caterina al Grottino solo per passare una serata tranquilla, per farsi due chiacchiere e alzare un po' il gomito. Poi chissà. Se erano rose sarebbero fiorite. Non che ne avesse voglia, ma una divagazione quella sì, di quella c'era bisogno. Chiuse la porta. Qualcuno saliva le scale. Era Gabriele, con la solita aria malinconica, la stessa maglietta della sera prima, le stesse scarpe sfondate e i pantaloni al ginocchio. «Le è piaciuto ieri sera? Le ho messo David Bowie». «Apprezzato, grazie». «Non si ricordava più la password, eh?». Rocco lo guardò senza capire. «La password del wi-fi. Hey teacher leave the kids alone!». «Me la ricordavo eccome. Che stai dicendo?». «Perché ho aperto il portatile e ho visto che è andato a cercare un'altra connessione. È arrivato fino alle poste!». Rocco fece una smorfia. «Non ti seguo, Gabriele». «Io le ho prestato il computer per andare su internet, lei non c'è riuscito da casa ed è andato giù in strada, s'è agganciato al wi-fi delle poste». «Io non sono andato in strada» fece Rocco. «Non mi sono mosso da casa mia». «Strano» fece Gabriele. «Le connessioni del computer parlano chiaro». «Aspetta Gabriele, fammi capire questa cosa. Qualcuno ha usato il tuo computer ed è andato fino alle poste per cercare qualcosa su internet?». «No, non ha cercato nulla. Era sempre sullo stesso sito. Cioè lei all'inizio la password la sapeva, ed è andato a vedere la strada che da Roma porta in Friuli, mi scusi se mi sono messo a ficcare il naso. Poi è tornato di nuovo on line, stavolta dalle poste, e ha aperto lo stesso sito, la stessa strada. Ecco perché non capivo». Rocco sentì un brivido, nonostante la serata fosse calda. «Fammi ricordare». Tornò indietro con la memoria. «Quella mattina mi aveva chiamato Brizio, ci ho parlato, mi ha detto del bancomat usato in Umbria, poi tu mi hai dato il computer, io ho cercato una strada, ho parlato con De Silvestri per sapere delle rapine ai distributori... aspetta!» si voltò e riaprì la porta di casa. Gabriele lo osservava senza capire, scostandosi il ciuffo di capelli davanti agli occhi con un soffio. «Che fa?». Rocco si era accucciato a controllare la serratura. Ci passò sopra un dito. Poi guardò il legno della porta. «No, niente...». Si precipitò alla porta-finestra del balconcino. Era chiusa. Si mise a controllare anche quella. «Sta cercando di capire se qualcuno è entrato in casa sua?». «Qualcuno deve essere entrato, Gabrie'!». Abbandonò il balconcino per ispezionare le altre finestre dell'appartamento. «Qualcuno che ha preso di nascosto il computer». Finì l'indagine sugli infissi e chiuse la finestra. «Qui non è entrato nessuno» e guardò Gabriele. Poi sussurrò: «Niente...». «Come niente?». «Niente, Gabriele». Gli occhi di Rocco avevano perso la luce e il viso era diventato un cencio sporco. «Niente...». Caterina lo aspettava fuori dal ristorante. Sorrideva con un filo di trucco sugli occhi. Rocco teneva la testa bassa. Lupa trotterellava al suo fianco. Appena la raggiunse il sorriso svanì dal viso della ragazza. «Rocco... che hai?». «Entriamo...» e aprì la porta del ristorante. Non c'era molta gente, il suo tavolo era libero e si andarono a sedere. Subito il cameriere gli lasciò i menu. «Acqua?». «Liscia» rispose Rocco. «Che ti succede?». Il vicequestore alzò lo sguardo. «Sto male, Caterina. Ho un dolore allo stomaco, un pessimo sapore in bocca, e mi succede spesso di sentirmi così. E poi mi sento sporco. Sporco e marcio». «Perché?». «Non mi piace sguazzare nel fango. A cercare un assassino, un traditore. È una sensazione alla quale non mi abituerò mai». «Lo so, l'hai sempre detto». «Ma la cosa peggiore è quando il fango ce l'hai in casa». «Non ti seguo...» fece Caterina preoccupata. «A chi ti riferisci? A Baiocchi?». «Anche. A Baiocchi, a Valenti... a chi ti colpisce dove fa più male, dove non hai difese. E il dolore è doppio. A te è mai successo?». «Tante volte, Rocco. L'ultima qualche giorno fa». «A me stasera». Il cameriere portò l'acqua. «Avete scelto?». «No». Se ne andò a servire un altro tavolo. «Rocco, mi stai facendo preoccupare. Cos'è successo stasera?». «La cosa peggiore che potesse accadere. E fa male». La guardò negli occhi. «Non l'avevo capito, sai? Sei stata brava, Caterina, veramente brava. Te l'ho sempre detto che sei una donna in gamba». Caterina impallidì. «Io non capisco, proprio non ti capisco». «A letto con me ci sei venuta per dovere o per piacere?». Il sangue con una vampata improvvisa tornò a irrorarle le gote. «Ma che cazzo dici?». «Almeno questo puoi dirmelo. Però, per una volta, dimmi la verità. Allora?». «Ci sono venuta perché mi andava, Rocco, perché mi piaci, perché piano piano...». «Perché piano piano mi sono fidato di te. Giorno dopo giorno. Ogni volta che andavo a Roma ti lasciavo Lupa, perché mi fidavo. E mi succede di rado. Sai tutto di me. Com'è morta mia moglie, chi sono i miei amici, e soprattutto sapevi chi cercavo e dove. O mi sbaglio?». Caterina scuoteva lenta la testa. «Deve essere stata brutta quando mi chiamavi e trovavi il cellulare sempre spento. L'avevo lasciato a Trastevere, in albergo. Andavo in giro con questo» tirò fuori il telefonino nuovo e lo sbatté sul tavolo. «Un numero che non conoscevi. Sarebbe stato tutto più semplice, no? Per trovare dove se n'era andato quel vecchio pazzo di Schiavone bastava controllare le celle alle quali il numero s'era agganciato. Vabbè, t'ho dato un po' di lavoro in più da fare...». Caterina era a bocca aperta, non riusciva a parlare. Due fili di saliva le partivano dagli incisivi fino all'attaccatura delle labbra. «Dimmi almeno se eri sola o ti aiutava qualcuno». Gli occhi di Caterina si inumidirono. «Puoi dirmi chi ti ha ordinato di starmi addosso?». Il viceispettore scoppiò a piangere. «Non...» si asciugò una lacrima con la manica del maglione. «Rocco, non è così». «E allora com'è? Dimmelo, sono qui che ascolto. E non vedo l'ora di crederti». Si mise a braccia conserte. La ragazza poggiò i gomiti sul tavolo. Guardò il soffitto, le lacrime le impedivano di parlare. «Lo sai». «Tu non ce la fai a parlare, io non ce la faccio ad ascoltare, Caterina. Dimmi solo se lo stai facendo dal primo giorno che sono arrivato». «No... è successo dopo. Io devo molto a un uomo che sta a Roma, che mi ha salvato quando ero una ragazza, una bambina. Lui me l'ha chiesto e non potevo...». «Sei stata tu a mandare a puttane il caso Perez...». «No! Non c'entro niente. In quella storia ero con te!». «Aveva ragione Brizio. Tempo fa mi aveva detto che un uomo della mia età non deve frequentare donne più giovani. Abbiamo poco da spartire, meglio, io ho poco da spartire, tu di cose da me ne volevi e tante. Vedi che avevo ragione? La vita è come la mattanza dei tonni». «Per favore, dammi una possibilità di...». «Vaffanculo Caterina Rispoli. Esci dalla mia vita, di' al tuo contatto di Roma, all'uomo che ti ha salvato da piccola, che non puoi avere più informazioni da questo stronzo e che la tua permanenza ad Aosta è inutile. Peccato, perché scopi bene». Si alzò di scatto dal tavolo e si avvicinò al cameriere. «Tutto quello che mangia la signorina lo metti sul mio conto». Diede un'ultima occhiata al tavolo. «Offre la casa! È l'ultimo regalo che ti faccio» e uscì dal ristorante con la morte nel cuore. Se ne andò in giro per la città, senza una meta. Gli sembrava che i piedi affondassero nell'asfalto. Faticava a tenere il passo, ma non riusciva a fermarsi. Le orecchie piene di ovatta e i denti stretti mordevano l'aria e si maciullavano. Il dolore appuntito sotto al tallone tornò, ma non ci fece caso. Se c'era da zoppicare avrebbe zoppicato, l'importante era continuare a camminare. Non c'era spazio per piangersi addosso, neanche per andare a cercare qualche ricordo dolce. Solo camminare finché il fiato reggeva, finché le scarpe resistevano e il tallone non si sarebbe messo a urlare. Odiava quell'odore, cos'era? Qualche fiore, o ancora erba appena tagliata. Si accese una sigaretta per annullarlo, per scacciarlo. Ma la gettò alla prima boccata, era amara, stopposa. Si ritrovò in mezzo alle rovine, seduto sulla pietra scura che una volta doveva essere un'arcata che sosteneva le gradinate della platea del teatro romano. Dove spesso anche Marina veniva a trovarlo, ma ormai non ci sperava più. Non gli restò che guardare il cielo. Lupa gli saltò in braccio. Voleva essere abbracciata, carezzata sotto quel manto di stelle. La luna non c'era. Era sparita. Epatta, gli tornò in mente quella parola. E così la vide. Marina, fresca e riposata, se ne stava seduta due pietre più in là a guardare il cielo. «Hai imparato a contare?» mi dice. «Sì». «Le notti senza luna sono le più strane. Ci pensi ai marinai tanti anni fa?». «Avevano le stelle». «Sì, ma la luna è un faro. Le stelle mentono. Lo sai che un sacco di queste sono già morte?». Mi viene da ridere. Anche io ci penso spesso a questa storia delle stelle già morte. «Vieni accanto a me?». «Accontentati che sono venuta» poi si gira e mi guarda. «Povero Rocco, a forza di inseguire le ombre stai diventando un'ombra pure tu. Devi stare attento». «È vero. A forza di rincorrere le ombre, si rischia di perdere la sostanza». «Mi dispiace che resti sempre solo». «Non sono solo» le dico. «Ho Lupa. E ho te, quando vieni». Sorride ancora. «Ce la fai?». «Mi tocca. Mi sa che vendo la casa di Roma». «E il limone?». «Me lo porto qui». «Non hai il terrazzo. E qui fa troppo freddo. Dallo ai miei». «Ma se manco mi parlano». «Lasciaglielo fuori la porta. Capiranno». «E se te lo portassi?». «Me lo pianti vicino? È una bella idea. Ma l'inverno chi lo copre?». «Già». Mi viene da dire che morirebbe, e quel limone non deve morire. «È bella questa notte». «Una volta su un gozzo a Positano mi dicesti che sapevi i nomi delle stelle». «Li so ancora». «E dimmeli un po'?». Alzo un dito e indico la prima che mi capita. «Aldebaran, Sirio, Bentegodi». Marina comincia a ridere. «Navicella, Stupor mundi, Aileselassié, Rodolfo!». E ci mettiamo a sghignazzare come due scemi. «Ciao Lupa. Ciao Rocco. Alla prossima luna che se ne va, amore mio». «Alla prossima luna». Hai visto, Lupa? È tornata a trovarci. Eppure la vita deve essere bella, lo sai? Se pure un vitello che ha fatto una vita schifosa chiuso in una gabbia piange quando lo portano al mattatoio, allora sì, allora deve essere proprio bella. È una lezione che dovrei ripetermi ogni giorno. Ma io oggi non riesco neanche a respirare. Tu non lo sai cucciola mia, ma un sacco di animali vanno in letargo quando arriva il gelo dell'inverno. Si accucciano in una buca sotto terra, chiudono gli occhi e muoiono per un po' di mesi. Quando tornano al sole sono nati un'altra volta e ricominciano a sorridere, a saltare, perché è vita nuova, piena di colori e di odori. Noi no. Noi a dormire non ci andiamo mai sul serio, e così invecchiamo e la pelle si raggrinzisce, come il sangue. Tutto si stanca, Lupa, si consuma e non torna più come prima. Mi guardi, con la lingua di fuori, e siamo soli io e te, un'altra volta, e sei tu che mi devi dare coraggio, amica mia, perché io non ce l'ho più. Stai qui, attaccata a me. Chiudi gli occhi. Dormi, Lupa. Sogna gli ossi e i prati dove correre. Vola pure. Io da qui ti guardo e aspetto di capire come si fa. Ti giuro, appena ci riesco ti seguo. # Indice Pulvis et umbra Epigrafe Domenica
{ "redpajama_set_name": "RedPajamaBook" }
8,168
<?php namespace Parse; use Exception; use Parse\Internal\ParseRelationOperation; /** * ParseRelation - A class that is used to access all of the children of a many-to-many relationship. Each instance * of ParseRelation is associated with a particular parent object and key. * * @author Mohamed Madbouli <mohamedmadbouli@fb.com> */ class ParseRelation { /** * The parent of this relation. * * @var ParseObject */ private $parent; /** * The key of the relation in the parent object. * * @var string */ private $key; /** * The className of the target objects. * * @var string */ private $targetClassName; /** * Creates a new Relation for the given parent object, key and class name of target objects. * * @param ParseObject $parent The parent of this relation. * @param string $key The key of the relation in the parent object. * @param string $targetClassName The className of the target objects. */ public function __construct($parent, $key, $targetClassName = null) { $this->parent = $parent; $this->key = $key; $this->targetClassName = $targetClassName; } /** * Makes sure that this relation has the right parent and key. * * @param $parent * @param $key * * @throws \Exception */ private function ensureParentAndKey($parent, $key) { if (!$this->parent) { $this->parent = $parent; } if (!$this->key) { $this->key = $key; } if ($this->parent !== $parent) { throw new Exception('Internal Error. Relation retrieved from two different Objects.'); } if ($this->key !== $key) { throw new Exception('Internal Error. Relation retrieved from two different keys.'); } } /** * Adds a ParseObject or an array of ParseObjects to the relation. * * @param mixed $objects The item or items to add. */ public function add($objects) { if (!is_array($objects)) { $objects = [$objects]; } $operation = new ParseRelationOperation($objects, null); $this->targetClassName = $operation->_getTargetClass(); $this->parent->_performOperation($this->key, $operation); } /** * Removes a ParseObject or an array of ParseObjects from this relation. * * @param mixed $objects The item or items to remove. */ public function remove($objects) { if (!is_array($objects)) { $objects = [$objects]; } $operation = new ParseRelationOperation(null, $objects); $this->targetClassName = $operation->_getTargetClass(); $this->parent->_performOperation($this->key, $operation); } /** * Returns the target classname for the relation. * * @return string */ public function getTargetClass() { return $this->targetClassName; } /** * Set the target classname for the relation. * * @param $className */ public function setTargetClass($className) { $this->targetClassName = $className; } /** * Set the parent object for the relation. * * @param $parent */ public function setParent($parent) { $this->parent = $parent; } /** * Gets a query that can be used to query the objects in this relation. * * @return ParseQuery That restricts the results to objects in this relations. */ public function getQuery() { $query = new ParseQuery($this->targetClassName); $query->relatedTo('object', $this->parent->_toPointer()); $query->relatedTo('key', $this->key); return $query; } }
{ "redpajama_set_name": "RedPajamaGithub" }
6,562
Thoughts From The Music(al) Man (2020) on… An American In Paris (1951) Neil "The Musical Man" Powell Uncategorized February 9, 2020 May 30, 2021 4 Minutes "It's very clear our love is here to stay." In case you haven't guessed already, the next movie I want to talk about is that classic 1951 Gene Kelly musical, An American In Paris, also starring Leslie Caron, Oscar Levant and Georges Guetary. Coming Up Shorts! with… Service With A Guile (1946) Popeye and Bluto help Olive repair an admirals car. Another fun outing as Popeye and Bluto try to one-up each other in fixing the car, resulting in even more trouble. And a fun ending I didn't quite see coming after Popeye eats his spinach and repairs the car. While still voiced by Harry Welch instead of regular Jack Mercer, I didn't notice it as strongly this time, which made things better. While some of the gags may not be new, they worked well enough I had a good time watching this one! Certainly another fun short that continues to make this set (and seeing some of these old Popeye shorts) well worth it! Ex-G.I. Jerry Mulligan (Gene Kelly) is living in Paris as a painter, alongside his pianist buddy Adam Cook (Oscar Levant). One morning when displaying his paintings, Jerry ends up selling two of them to Milo Roberts (Nina Foch), who decides to help support him as an artist. They go out to a club that night, where Jerry runs into Lise Bouvier (Leslie Caron). He is instantly infatuated with her, although she is less than thrilled with his attentions (and the same could be said for Milo as well). The next day, Jerry tries to ask Lise out again, and with a little persistence, she says yes. What Jerry doesn't know is that she is engaged to Adam's friend Henri "Hank" Baurel (Georges Guetary), who had raised her after her parents were killed in the war. Hank is given an offer to go to America, and he hopes that he and Lise can get married before they have to leave. Meanwhile, Milo is doing all she can to help Jerry towards giving an exhibition of his paintings, by providing a new place for him to work from and helping make contacts. When Jerry is given advice to tell Lise that he loves her, he does, only to find out she is engaged to another man (and to Hank, who had given him that advice without knowing who Jerry was in love with). In frustration, Jerry takes Milo to a party, where they run into Hank and Lise before they prepare to leave. The idea for the movie famously came to producer Arthur Freed after he attended a concert for George Gershwin's song An American In Paris. He liked the title for a movie, and went about getting the rights to the song (along with a number of other George Gershwin tunes). With Gene Kelly quickly cast, they ended up giving the role of his romantic interest to newcomer Leslie Caron, after Gene saw her performing in a French ballet and lobbied for her to get the part. Of course, the final ballet, set to the title tune, would prove a controversial addition, as previous attempts at lengthy ballets (especially in the 1945 Fred Astaire musical Yolanda And The Thief, also directed by Vincente Minelli) had failed to connect with audiences. But Arthur Freed and company stuck to their guns, and it became a high point of the movie (and the beginning of a trend whereby many musicals in the fifties would make use of dream ballets). I can't deny that a lot of the fun here is indeed the music and dancing! Gene Kelly gets a lot of the fun, especially with the likes of his tap solo to "I Got Rhythm," where he gets to work with a bunch of French children as he teaches them a little English. Of course, he also has his romantic duet with Leslie Caron to "Our Love Is Here To Stay," which is a thing of beauty in and of itself. But that ballet to the title tune is definitely a highlight, especially since some of the music should solidly be stuck in your head by that point, after having been used as background music for most of the movie. But the variety in dance styles and sets during that ballet is just so much fun to watch! And this movie works so well as a comedy, too! From the character introductions for Jerry Mulligan, Adam Cook and Hank Baurel, we get the camera "mistakenly" showing somebody else before showing us the actual characters (especially a hoot with Oscar Levant's Adam, if you know how much of a sourpuss Oscar's characters tend to be and then we are shown a guy that is "too happy" before moving on to Adam)! And then the comic interactions between them on songs like "By Strauss" and "Tra-La-La (This Time It's Really Love)." Of course, it's hard not to laugh at Adam after Jerry tells him he is in love with Lise (especially since Hank had told Adam about Lise at the beginning of the movie) and then, when Hank comes in, Adam is nervously re-lighting his cigarette and drinking all the tea while he waits for Jerry or Hank to say just the wrong thing. Just priceless to watch! Honestly, this movie sells itself, it is just so wonderful, I can't even begin to recommend it enough! "'S Wonderful!" This movie is available on Blu-ray and DVD from Warner Home Video. *ranked #1 in Top 10 Movies Watched In 2020 Summer Stock (1950) – Gene Kelly – Singin' In The Rain (1952) Leslie Caron – Father Goose (1964) Oscar Levant – The Band Wagon (1953) an american in paris georges guetary service with a guile thoughts from the musical man Published February 9, 2020 May 30, 2021 Previous Post What's Old Is A New Release Again (2019) with… Steamboat Bill, Jr. (1928) Next Post What's Old Is A New Release Again (2019) with… The Bells Of St. Mary's (1945)
{ "redpajama_set_name": "RedPajamaCommonCrawl" }
2,523
Wanarejan Utara is een bestuurslaag in het regentschap Pemalang van de provincie Midden-Java, Indonesië. Wanarejan Utara telt 9704 inwoners (volkstelling 2010). Plaats in Midden-Java
{ "redpajama_set_name": "RedPajamaWikipedia" }
2,213
Q: IntelliJ - Show errors/warning on the left of code I use IntelliJ after moving from Eclipse (With Java) I like that errors/warnings are displayed on the code map on the right but I miss the indications of errors and warning that appear next to the problem on the left (gutter) Is there a way to turn it on? Thanks! A: No it's not possible but you can hit F2 to cycle through the errors quickly.
{ "redpajama_set_name": "RedPajamaStackExchange" }
7,161
\section{Introduction} In this paper, we consider the stationary flow of a viscous incompressible fluid, in the case where the viscosity of the fluid depends on the temperature. Such a fluid-thermal coupling appears in many physical phenomena and it is involved in wide range industrial applications, e.g. cooling processes in steel industries, industrial furnaces, boilers, heat exchangers and nuclear power plants. Due to its importance in many practical applications, the numerical approximation of coupled {N}avier-{S}tokes and heat equations have been widely studied. In \cite{oyarzua:2014} a generalized Boussinesq problem modelling the motion of a nonisothermal incompressible fluid subject to a heat source has been discretized by means of mixed finite element methods. A spectral discretization of the {N}avier-{S}tokes equation coupled with the heat equation has been proposed in \cite{Bernardi:2015}, in the stationary case, and in \cite{Bernardi:2015-b} in the unsteady one. Finite Element approximation of the time dependent Boussinesq model with nonlinear viscosity depending on the temperature has been studied in \cite{Aldbaissy:2018}. Recently, the finite element approximation of the heat equation coupled with Stokes equations with nonlinear slip boundary conditions has been analyzed in \cite{Djoko:2020}. Finite element methods for {D}arcy's problem coupled with the heat equation has been studied in \cite{Bernardi:2018}. In this paper we are interested in the Virtual Element discretization of the coupled thermo-fluid dynamic problem. The Virtual Element method (VEM) is a generalization of the Finite Element Method that takes inspiration from modern Mimetic Finite Difference schemes \cite{volley,autostoppisti,projectors}. It belongs to the family of polytopal element methods, i.e. finite element methods that can support polygonal/polyhedral meshes. Polytopal methods such as polygonal FEM, polygonal and Hybridizable Discontinuous Galerkin, Hybrid High Order methods, see e.g.\cite{N:2020, BdPD:2018, dPK:2018, AdP:2018, CdP:2020, LVY:2014, CFQ:2017,Antonietti-FSI-1,Antonietti-FSI-2} have received substantial attention in the last years for fluid mechanic problems, thanks to their flexibly in handling complicated geometries and their ability preserving the underlying physical models' fundamental properties. The VEM has received substantial attention in the last ten years for a wide range problems stemming from of Engineering applications. We refer to the book \cite{Antonietti_BeiraodaVeiga_Manzini:2022} where the current state of the art on the Virtual Element Method is reported, collecting contributions from many of the most active researchers in this field. VEMs for Stokes and Navier-Stokes equations have been proposed and analyzed in \cite{ABMV:2014,BLV:2017,vacca:2018,BLV:2018,BDV:2020}; we refer, e.g., to \cite{berrone-supg,BdV-supg,BdV-oseen}, for the development of VEM for convection dominated problems. The remaining part of the manuscript is organized as follows. Section \ref{sec:model_problem} describes the coupled thermo-fluid dynamic problem together with the theoretical results. Section \ref{sub:VEM} is devoted to the Virtual Elements discretization. The existence of discrete solutions is established in \ref{sec:exi} whereas the \textit{a priori} convergence analysis is presented in Section \ref{sec:conv}. In Section \ref{sec:num} we present some numerical experiments to test the practical performance of the proposed VEM scheme. Finally, in Section\ref{sec:conclusions} we draw some conclusions. We close this introduction with some standard notation. Let $\Omega \subset \numberset{R}^2$ be the computational domain, we denote with $\boldsymbol{x} = (x_1, \, x_2)$ the independent variable. With a usual notation the symbols $\nabla$ and $\Delta$ denote the gradient and the Laplacian for scalar functions, while $\boldsymbol{\Delta} $, $\boldsymbol{\nabla}$, $\boldsymbol{\epsilon}$ and ${\rm div}$ denote the vector Laplacian, the gradient, the symmetric gradient operator and the divergence operator, whereas $\boldsymbol{{\rm div}}$ denotes the vector valued divergence operator for tensor fields. Throughout the paper, we will follow the usual notation for Sobolev spaces and norms \cite{Adams:1975}. Hence, for an open bounded domain $\omega$, the norms in the spaces $W^s_p(\omega)$ and $L^p(\omega)$ are denoted by $\|{\cdot}\|_{W^s_p(\omega)}$ and $\|{\cdot}\|_{L^p(\omega)}$ respectively. Norm and seminorm in $H^{s}(\omega)$ are denoted respectively by $\|{\cdot}\|_{s,\omega}$ and $|{\cdot}|_{s,\omega}$, while $(\cdot,\cdot)_{\omega}$ and $\|\cdot\|_{\omega}$ denote the $L^2$-inner product and the $L^2$-norm (the subscript $\omega$ may be omitted when $\omega$ is the whole computational domain $\Omega$). \section{Continuous problem}\label{sec:model_problem} Let $\Omega \subseteq \numberset{R}^2$ be a polygonal bounded domain with a Lipschitz-continuous boundary $\partial \Omega$. We consider the following coupled thermo-fluid dynamic problem \begin{equation} \label{eq:pb primale} \left\{ \begin{aligned} & \begin{aligned} - \boldsymbol{{\rm div}} (\nu(\theta) \,\boldsymbol{\epsilon} (\boldsymbol{u})) + (\boldsymbol{\nabla} \boldsymbol{u} ) \,\boldsymbol{u} - \nabla p &= \boldsymbol{f}\qquad & &\text{in $\Omega$,} \\ {\rm div} \, \boldsymbol{u} &= 0 \qquad & &\text{in $\Omega$,} \\ - {\rm div}(\kappa \, \nabla \theta) + \boldsymbol{u} \cdot \nabla \theta &= g \qquad & &\text{in $\Omega$,} \\ \boldsymbol{u} &= 0 \qquad & &\text{on $\partial \Omega$,}\\ \theta &= \theta_{\rm D} \qquad & &\text{on $\partial \Omega$,} \end{aligned} \end{aligned} \right. \end{equation} where the unknowns $\boldsymbol{u}$, $p$ and $\theta$ represent the velocity, the pressure and the temperature of the fluid respectively; the function $\nu$ is the temperature depending viscosity of the fluid, $\kappa$ is the thermal conductivity, $\boldsymbol{f}$ and $g$ stand for the the external volume source terms and $\theta_{\rm D}$ is the Dirichlet datum for the temperature. Equation \eqref{eq:pb primale} models the stationary flow of a viscous incompressible fluid (governed by a Navier-Stokes equation), where the viscosity of the fluid depends on the temperature (governed by an elliptic equation). For sake of simplicity we here consider Dirichlet homogeneous boundary conditions for the velocity field, different boundary conditions can be treated as well. Moreover, the analysis of the three dimensional case could be developed with similar arguments. From now on, we assume that the model data satisfy the following assumptions. \smallskip\noindent \textbf{(A0) Data assumptions:} \begin{itemize} \item the conductivity $\kappa$ is positive and bounded, i.e. there exist $\kappa_*$, $\kappa^* > 0$ s.t. \[ 0 < \kappa_* \leq \kappa(\boldsymbol{x}) \leq \kappa^* \quad \text{for all $\boldsymbol{x} \in \Omega$;} \] \item the viscosity $\nu$ is positive, bounded and Lipschitz continuous, i.e. there exist $\nu_*$, $\nu^*$, $\nu_{\rm lip} > 0$ s.t. \[ \begin{gathered} 0 < \nu_* \leq \nu(\theta) \leq \nu^* \qquad \text{for all $\theta \in \mathbb{R}$,} \\ |\nu(\theta_1) - \nu(\theta_2)| \leq \nu_{\rm lip} |\theta_1 - \theta_2| \qquad \text{for all $\theta_1, \theta_2 \in \mathbb{R}$;} \end{gathered} \] \item the external loads satisfy $\boldsymbol{f} \in [L^2(\Omega)]^2$, $g \in L^2(\Omega)$; \item the Dirichlet datum satisfy $\theta_{\rm D} \in C^0(\partial \Omega)$. \end{itemize} Notice that the regularity assumption on the boundary datum is required in order to define its nodal interpolant in the discrete VEM setting (see \eqref{eq:thd}). Let us define the continuous spaces \[ \begin{gathered} \boldsymbol{V} := \left[ H_0^1(\Omega) \right]^2, \quad Q := L^2_0(\Omega) = \left\{ q \in L^2(\Omega) \,\,\, \text{s.t.} \,\,\, \int_{\Omega} q \,{\rm d}\Omega = 0 \right\}, \\ \Sigma := H^1(\Omega), \quad \Sigma_0 := H^1_0(\Omega), \quad \Sigma_{\rm D} :=\{\sigma \in H^1(\Omega) \,\,\, \text{s.t.} \,\,\, \sigma_{|\partial \Omega} = \theta_{\rm D} \}, \end{gathered} \] endowed with natural norms, and the forms \begin{align} \label{eq:forma av} {a_V}(\cdot; \, \cdot, \cdot) &\colon \Sigma \times \boldsymbol{V} \times \boldsymbol{V} \to \numberset{R}, &\quad {a_V}(\theta; \boldsymbol{u}, \boldsymbol{v}) &:= \int_{\Omega} \nu(\theta) \boldsymbol{\epsilon}(\boldsymbol{u}) : \boldsymbol{\epsilon} (\boldsymbol{v}) \, {\rm d}\Omega \,, \\ \label{eq:forma as} {a_T}(\cdot, \cdot) &\colon \Sigma \times \Sigma \to \numberset{R}, &\quad {a_T}(\theta, \sigma) &:= \int_{\Omega} \kappa \, \nabla \sigma \cdot \nabla \theta \, {\rm d}\Omega \,, \\ \label{eq:forma cv} {c_V}(\cdot; \, \cdot, \cdot) &\colon \boldsymbol{V} \times \boldsymbol{V} \times \boldsymbol{V} \to \numberset{R}, &\quad {c_V}(\boldsymbol{w}; \, \boldsymbol{u}, \boldsymbol{v}) &:= \int_{\Omega} ( \boldsymbol{\nabla} \boldsymbol{u} ) \, \boldsymbol{w} \cdot \boldsymbol{v} \,{\rm d}\Omega \,, \\ \label{eq:forma cs} {c_T}(\cdot; \, \cdot, \cdot) &\colon \boldsymbol{V} \times \Sigma \times \Sigma \to \numberset{R}, &\quad {c_T}(\boldsymbol{u}; \, \theta, \sigma) &:= \int_{\Omega} \boldsymbol{u} \cdot \nabla \theta \, \sigma \,{\rm d}\Omega \,, \\ \label{eq:forma b} b(\cdot, \cdot) &\colon \boldsymbol{V} \times Q \to \numberset{R}, &\quad b(\boldsymbol{v}, q) &:= \int_{\Omega}q\, {\rm div} \boldsymbol{v} \,{\rm d}\Omega \,. \end{align} Notice that under the assumptions \textbf{(A0)} the forms are well defined, furthermore the following hold \smallskip\noindent \textbf{(P0) Stability properties of the continuous forms:} \\ $\bullet$ \,\, for any $\theta \in \Sigma$ the bilinear form ${a_V}(\theta; \, \cdot, \cdot)$ is coercive and continuous, i.e. \[ {a_V}(\theta; \, \boldsymbol{v}, \boldsymbol{v}) \geq \nu_* \vert \boldsymbol{v} \vert_{1, \Omega}^2 \,, \quad {a_V}(\theta; \, \boldsymbol{u}, \boldsymbol{v}) \leq \nu^* \vert \boldsymbol{u} \vert_{1, \Omega} \vert \boldsymbol{v} \vert_{1, \Omega}\,, \qquad \text{for all $\boldsymbol{u}$, $\boldsymbol{v} \in \boldsymbol{V}$;} \] $\bullet$ \,\, the bilinear form ${a_T}(\cdot, \cdot)$ is coercive and continuous, i.e. \[ {a_T}(\sigma, \sigma) \geq \kappa_* \vert \sigma \vert_{1, \Omega}^2 \,, \quad {a_T}(\theta, \sigma) \leq \kappa^* \vert \theta \vert_{1, \Omega} \vert \sigma \vert_{1, \Omega}\,, \qquad \text{for all $\theta$, $\sigma \in \Sigma$;} \] $\bullet$ \,\, the convective trilinear forms ${c_V}(\cdot; \, \cdot, \cdot)$ and ${c_T}(\cdot; \, \cdot, \cdot)$ are continuous with continuity constant $C_{\rm conv}$, i.e. \[ \begin{aligned} {c_V}(\boldsymbol{w}; \, \boldsymbol{u}, \boldsymbol{v}) &\leq C_{\rm conv} \vert \boldsymbol{w} \vert_{1, \Omega} \vert \boldsymbol{u} \vert_{1, \Omega} \vert \boldsymbol{v} \vert_{1, \Omega} \,, \qquad & \text{for all $\boldsymbol{w}$, $\boldsymbol{u}$, $\boldsymbol{v} \in \boldsymbol{V}$,} \\ {c_T}(\boldsymbol{u}; \, \theta, \sigma) &\leq C_{\rm conv} \vert \boldsymbol{u} \vert_{1, \Omega} \Vert \theta \Vert_{1, \Omega} \Vert \sigma \Vert_{1, \Omega} \,, \qquad & \text{for all $\boldsymbol{u} \in \boldsymbol{V}$, $\theta$, $\sigma \in \Sigma$;} \end{aligned} \] \\ $\bullet$ \,\, the bilinear form $b(\cdot, \cdot)$ is continuous and realizes the inf-sup condition with inf–sup constant $\beta >0$, i.e. \[ \begin{aligned} &b(\boldsymbol{v}, q) \leq |\boldsymbol{v}|_{1, \Omega} \Vert q \Vert_{0,\Omega} &\qquad &\text{for all $\boldsymbol{v} \in \boldsymbol{V}$ and $q \in Q$,} \\ &\sup_{\boldsymbol{v} \in \boldsymbol{V}} \frac{b(\boldsymbol{v}, q)}{|\boldsymbol{v}|_{1, \Omega}} \geq \beta \Vert q \Vert_{0,\Omega} &\qquad &\text{for all $q \in Q$.} \end{aligned} \] \smallskip\noindent The variational formulation of Problem~\eqref{eq:pb primale} reads as follows: \begin{equation} \label{eq:pb variazionale} \left\{ \begin{aligned} & \text{find $(\boldsymbol{u}, p, \theta) \in \boldsymbol{V} \times Q \times \Sigma_{\rm D}$, such that} \\ &\begin{aligned} {a_V}(\theta, \, \boldsymbol{u}, \boldsymbol{v}) + {c_V}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}) + b(\boldsymbol{v}, p) &= (\boldsymbol{f}, \boldsymbol{v}) \qquad & &\text{for all $\boldsymbol{v} \in \boldsymbol{V}$,} \\ b(\boldsymbol{u}, q) &= 0 \qquad & & \text{for all $q \in Q$,} \\ {a_T}(\theta, \sigma) + {c_T}(\boldsymbol{u}; \theta, \sigma) &= (g, \sigma) \qquad & &\text{for all $\sigma \in \Sigma_0$.} \end{aligned} \end{aligned} \right. \end{equation} Let us introduce the kernel of the bilinear form $b(\cdot,\cdot)$ that corresponds to the functions in $\boldsymbol{V}$ with vanishing divergence, i.e. \begin{equation*} \label{eq:Z} \boldsymbol{Z} := \{ \boldsymbol{v} \in \boldsymbol{V} \quad \text{s.t.} \quad {\rm div} \, \boldsymbol{v} = 0 \}\,. \end{equation*} Then, Problem~\eqref{eq:pb variazionale} can be formulated in the equivalent kernel form: \begin{equation} \label{eq:pb variazionale ker} \left\{ \begin{aligned} & \text{find $(\boldsymbol{u}, \theta) \in \boldsymbol{Z} \times \Sigma_{\rm D}$, such that} \\ &\begin{aligned} {a_V}(\theta, \, \boldsymbol{u}, \boldsymbol{v}) + {c_V}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}) &= (\boldsymbol{f}, \boldsymbol{v}) \qquad & &\text{for all $\boldsymbol{v} \in \boldsymbol{Z}$,} \\ {a_T}(\theta, \sigma) + {c_T}(\boldsymbol{u}; \theta, \sigma) &= (g, \sigma) \qquad & &\text{for all $\sigma \in \Sigma_0$.} \end{aligned} \end{aligned} \right. \end{equation} We can observe by a direct computation, that, for a fixed $\boldsymbol{w} \in \boldsymbol{Z}$, the bilinear forms ${c_V}(\boldsymbol{w}; \, \cdot , \cdot)$ and ${c_T}(\boldsymbol{w}; \, \cdot , \cdot)$ are skew symmetric, i.e. \begin{equation} \label{eq:skew} \begin{aligned} &{c_V}(\boldsymbol{w}; \boldsymbol{u}, \boldsymbol{v}) = -{c_V}(\boldsymbol{w}; \boldsymbol{v}, \boldsymbol{u}) \qquad &\text{for all $\boldsymbol{u}$, $\boldsymbol{v} \in \boldsymbol{V}$,} \\ &{c_T}(\boldsymbol{w}; \theta, \sigma) = -{c_T}(\boldsymbol{w}; \sigma, \theta) \qquad &\text{for all $\theta$, $\sigma \in \Sigma$.} \end{aligned} \end{equation} Therefore, for $\boldsymbol{w} \in \boldsymbol{Z}$, the forms ${c_V}(\boldsymbol{w}; \, \cdot, \cdot)$ and ${c_T}(\boldsymbol{w}; \, \cdot, \cdot)$ can be equivalently defined as follows \begin{align} \label{eq:forma cvskew} &c_V^{\rm skew}(\boldsymbol{w}; \, \boldsymbol{u}, \boldsymbol{v}) := \frac{1}{2} \bigl({c_V}(\boldsymbol{w}; \, \boldsymbol{u}, \boldsymbol{v}) - {c_V}(\boldsymbol{w}; \, \boldsymbol{v}, \boldsymbol{u}) \bigr) \qquad &\text{for all $\boldsymbol{u}$, $\boldsymbol{v} \in \boldsymbol{V}$,} \\ \label{eq:forma csskew} &c_T^{\rm skew}(\boldsymbol{w}; \, \theta, \sigma) := \frac{1}{2} \bigl ({c_T}(\boldsymbol{w}; \, \theta, \sigma) - {c_T}(\boldsymbol{w}; \, \sigma, \theta) \bigr) \qquad &\text{for all $\theta$, $\sigma \in \Sigma$.} \end{align} However, at the discrete level ${c_V}(\boldsymbol{w}; \, \cdot , \cdot)$ and $c_V^{\rm skew}(\boldsymbol{w}; \, \cdot , \cdot)$ (resp. ${c_T}(\boldsymbol{w}; \, \cdot , \cdot)$ and $c_T^{\rm skew}(\boldsymbol{w}; \, \cdot , \cdot)$) will lead to different bilinear forms, in general. In the context of the analysis of incompressible flows, it is useful to introduce the concept of Helmholtz--Hodge projector (see for instance \cite[Lemma~2.6]{john-linke-merdon-neilan-rebholz:2017} and \cite[Theorem~3.3]{GLS:2019}). For every $\boldsymbol{w} \in [L^2(\Omega)]^2$ there exist $\boldsymbol{w}_0 \in H({\rm div}; \, \Omega)$ and $\zeta \in H^1(\Omega) / \numberset{R}$ such that \begin{equation} \label{eq:helmholtz-hodge} \boldsymbol{w} = \boldsymbol{w}_0 + \nabla \, \zeta, \end{equation} where $\boldsymbol{w}_0$ is $L^2$-orthogonal to the gradients, that is $(\boldsymbol{w}_0 , \, \nabla \phi) = 0$ for all $\phi \in H^1(\Omega)$ (which implies, in particular, that $\boldsymbol{w}_0$ is solenoidal, i.e. ${\rm div} \, \boldsymbol{w}_0 = 0$). The orthogonal decomposition \eqref{eq:helmholtz-hodge} is unique and is called Helmholtz--Hodge decomposition, and $\mathcal{P}(\boldsymbol{w}) :=\boldsymbol{w}_0$ is the Helmholtz--Hodge projector of $\boldsymbol{w}$. Combining the argument in \cite[Theorem~2.2]{Bernardi:2015} and the definition of Helmholtz--Hodge projector, the following existence result holds. \begin{theorem} \label{thm:exi} Under the data assumptions \cfan{(A0)}, Problem \eqref{eq:pb variazionale} admits at least a solution $(\boldsymbol{u}, p, \theta) \in \boldsymbol{V} \times Q \times \Sigma_{\rm D}$. Moreover the solution satisfies the bound \begin{equation} \label{eq:bound cont} \vert \boldsymbol{u} \vert_{1, \Omega}^2 + \Vert \theta \Vert_{1, \Omega}^2 \leq {C}_{\rm est}^2 \left( \vert \mathcal{P}(\boldsymbol{f}) \vert_{-1, \Omega}^2 + \vert g \vert_{-1, \Omega}^2 + \Vert \theta_{\rm D} \Vert_{1/2, \partial \Omega}^2 \right) \end{equation} where the constant ${C}_{\rm est}$ depends on the domain $\Omega$ and on the constants $\kappa_*$, $\kappa^*$ and $\nu_*$ in the data assumptions \cfan{(A0)}. \end{theorem} \noindent Assuming suitable bounds on the data of the problem and on the velocity solution it is possible to establish the following uniqueness result \cite[Proposition 2.3]{Bernardi:2015} \begin{theorem} \label{thm:uni} Under the data assumptions \cfan{(A0)}, assume moreover that there exist two positive constants $C_{\rm data}$ and $C_{\rm sol}$ depending on the domain $\Omega$ and on the constants $\kappa_*$, $\kappa^*$ and $\nu_*$ in the data assumptions \cfan{(A0)} s.t. \begin{itemize} \item [i)] the data of the problem satisfies \begin{equation} \label{eq:cdata} C_{\rm data}^2 \left( \vert \mathcal{P}(\boldsymbol{f}) \vert_{-1, \Omega}^2 + \vert g \vert_{-1, \Omega}^2 + \Vert \theta_{\rm D} \Vert_{1/2, \partial \Omega}^2 \right) < 1 \,, \end{equation} \item [ii)] Problem \eqref{eq:pb variazionale} admits a solution $(\boldsymbol{u}, p, \theta)$ with $\boldsymbol{u} \in W^{1}_q(\Omega)$ where $q > 2$ and the following bound holds \quad \begin{equation} \label{eq:csol} C_{\rm sol} \, \nu_{\rm lip} \, \vert \boldsymbol{u} \vert_{[W^{1}_q(\Omega)]^2} < 1 \,, \end{equation} \end{itemize} then this solution is unique. \end{theorem} \section{Virtual Elements discretization} \label{sub:VEM} \subsection{Notation and preliminaries} \label{sec:notations} We now introduce some basic tools and notations useful in the construction and the theoretical analysis of Virtual Element Methods. Let $\{\Omega_h\}_h$ be a sequence of decompositions of the domain $\Omega \subset \numberset{R}^2$ into general polytopal elements $E$ where $h := \sup_{E \in \Omega_h} h_{E}$. We suppose that $\{\Omega_h\}_h$ fulfils the following assumption. \noindent \textbf{(A1) Mesh assumption:} \\ there exists a positive constant $\rho$ such that for any $E \in \{\Omega_h\}_h$ \begin{itemize} \item $E$ is star-shaped with respect to a ball $B_E$ of radius $ \geq\, \rho \, h_E$; \item any edge $e$ of $E$ has length $ \geq\, \rho \, h_E$. \end{itemize} We remark that the hypotheses above, though not too restrictive in many practical cases, could possibly be further relaxed, combining the present analysis with the studies in~\cite{BLR:2017,brenner-guan-sung:2017,brenner-sung:2018}. Using standard VEM notations, for $n \in \numberset{N}$ and $s \in \numberset{R}^+$ and $p=1, \dots, +\infty$ let us introduce the spaces: \begin{itemize} \item $\numberset{P}_n(\omega)$: the set of polynomials on $\omega \subset \Omega$ of degree $\leq n$ (with $\numberset{P}_{-1}(\omega)=\{ 0 \}$), \item $\numberset{P}_n(\Omega_h) := \{q \in L^2(\Omega) \quad \text{s.t} \quad q|_{E} \in \numberset{P}_n(E) \quad \text{for all $E \in \Omega_h$}\}$, \item $W^s_p(\Omega_h) := \{v \in L^2(\Omega) \quad \text{s.t} \quad v|_{E} \in W^s_p(E) \quad \text{for all $E \in \Omega_h$}\}$ \end{itemize} equipped with the broken norm and seminorm \[ \begin{aligned} \|v\|^p_{W^s_p(\Omega_h)} &:= \sum_{E \in \Omega_h} \|v\|^p_{W^s_p(E)}\,, &\quad |v|^p_{W^s_p(\Omega_h)} &:= \sum_{E \in \Omega_h} |v|^p_{W^s_p(E)}\,, &\qquad \text{if $1 \leq p < \infty$,} \\ \|v\|^p_{W^s_p(\Omega_h)} &:= \max_{E \in \Omega_h} \|v\|^p_{W^p(E)}\,, &\quad |v|^p_{W^s_p(\Omega_h)} &:= \max_{E \in \Omega_h} |v|^p_{W^s_p(E)}\,, &\qquad \text{if $p= \infty$.} \end{aligned} \] Let $E \in \Omega_h$, we denote with $h_{E}$ the diameter, with $|E|$ the area, with $\boldsymbol{x}_{E} = (x_{E, 1}, x_{E, 2})$ the centroid. A natural basis associated with the space $\numberset{P}_n(E)$ is the set of normalized monomials \begin{equation*} \label{eq:M_n} \numberset{M}_n(E) := \left\{ m_{\boldsymbol{\alpha}}, \,\,\, \text{with} \,\,\, |\boldsymbol{\alpha}| \leq n \right\} \end{equation*} where, for any multi-index $\boldsymbol{\alpha} = (\alpha_1, \alpha_2) \in \numberset{N}^2$ \[ m_{\boldsymbol{\alpha}} := \prod_{i=1}^2 \left(\frac{x_i - x_{E, i}}{h_{E}} \right)^{\alpha_i} \qquad \text{and} \qquad |\boldsymbol{\alpha}|:= \sum_{i=1}^2 \alpha_i \,. \] Moreover for any $m \leq n$ we denote with \[ \widehat{\numberset{P}}_{n \setminus m}(E) = {\rm span} \left\{ m_{\boldsymbol{\alpha}}, \,\,\, \text{with} \,\,\, m +1 \leq |\boldsymbol{\alpha}| \leq n \right\} \,. \] Furthermore, we introduce the following notations: let $\{\mathcal{X}^E\}_{E \in \Omega_h}$ be a family of forms $\mathcal{X}^E \colon \prod_{j=1}^{\ell} W_{p_j}^{s_j}(E) \to \numberset{R}$, then we define \begin{equation} \label{eq:XG} \mathcal{X}\colon \prod_{j=1}^{\ell} W_{p_j}^{s_j}(\Omega_h)\to \numberset{R} \,, \quad \mathcal{X}(u_1, \dots, u_{\ell}) := \sum_{E \in \Omega_h}\mathcal{X}^E(u_1, \dots, u_{\ell}) \,, \end{equation} for any $u_j \in W^{s_j}_{p_j}(\Omega_h)$ and $j=1, \dots, \ell$. For any $E$, let us define the following polynomial projections: \begin{itemize} \item the $\boldsymbol{L^2}$\textbf{-projection} $\Pi_n^{0, E} \colon L^2(E) \to \numberset{P}_n(E)$, given by \begin{equation} \label{eq:P0_k^E} \int_{E} q_n (v - \, {\Pi}_{n}^{0, E} v) \, {\rm d} E = 0 \qquad \text{for all $v \in L^2(E)$ and $q_n \in \numberset{P}_n(E)$,} \end{equation} with obvious extension for vector functions $\Pi^{0, E}_{n} \colon [L^2(E)]^2 \to [\numberset{P}_n(E)]^2$ and tensor functions $\boldsymbol{\Pi}^{0, E}_{n} \colon [L^2(E)]^{2 \times 2} \to [\numberset{P}_n(E)]^{2\times 2}$; \item the $\boldsymbol{H^1}$\textbf{-seminorm projection} ${\Pi}_{n}^{\nabla,E} \colon H^1(E) \to \numberset{P}_n(E)$, defined by \begin{equation*} \label{eq:Pn_k^E} \left\{ \begin{aligned} & \int_{E} \nabla \,q_n \cdot \nabla ( v - \, {\Pi}_{n}^{\nabla,E} v)\, {\rm d} E = 0 \quad \text{for all $v \in H^1(E)$ and $q_n \in \numberset{P}_n(E)$,} \\ & \int_{\partial E}(v - \, {\Pi}_{n}^{\nabla, E} v) \, {\rm d}s= 0 \, , \end{aligned} \right. \end{equation*} with extension for vector fields $\Pi^{\nabla, E}_{n} \colon [H^1(E)]^2 \to [\numberset{P}_n(E)]^2$. \end{itemize} In the following the symbol $\lesssim$ will denote a bound up to a generic positive constant, independent of the mesh size $h$, but which may depend on $\Omega$, on the ``polynomial'' order of the method $k$ and on the regularity constant appearing in the mesh assumption \textbf{(A1)}. We finally recall the following well know useful results: \begin{itemize} \item Poincar\'e inequality \cite[Theorem 1.3.3]{quarteroni-valli:book} \begin{equation} \label{eq:poincare} \Vert \phi \Vert_{1, \Omega} \lesssim \vert \phi \vert_{1, \Omega} \qquad \text{for any $\phi \in H^1_0(\Omega)$;} \end{equation} \item Sobolev embedding $H^1(\Omega) \subset L^p(\Omega)$ \cite[Theorem 1.3.4]{quarteroni-valli:book}: let $2 \leq p <\infty$, then \begin{equation} \label{eq:sobemb} \Vert \phi \Vert_{L^p(\Omega)} \lesssim \Vert \phi \Vert_{1, \Omega} \qquad \text{for any $\phi \in H^1(\Omega)$;} \end{equation} \item Polynomial inverse estimate \cite[Theorem 4.5.11]{brenner-scott:book}: let $1 \leq q,p \leq \infty$, then for any $E \in \Omega_h$ \begin{equation} \label{eq:inverse} \Vert p_n \Vert_{L^q(E)} \lesssim h_E^{2/q - 2/p} \Vert p_n \Vert_{L^p(E)} \quad \text{for any $p_n \in \numberset{P}_n(E)$;} \end{equation} \item Bramble-Hilbert Lemma \cite[Lemma 4.3.8]{brenner-scott:book}: let $0\leq t \leq s \leq n+1$, and $1 \leq q,p \leq \infty$ such that $s - 2/p > t - 2/q$, then for any $E \in \Omega_h$ \begin{equation} \label{eq:bramble} \vert \phi - \Pi^{0,E}_n \phi \vert_{W^t_q(E)} \lesssim h_E^{s-t + 2/q - 2/p} \vert \phi \vert_{W^s_p(E)} \quad \text{for any $\phi \in W^s_p(E)$.} \end{equation} \end{itemize} The present section is devoted to the construction of the proposed virtual elements scheme. In Subsections \ref{sub:v-p spaces} and \ref{sub:t-space} we present the inf-sup stable divergence--free velocities-pressures pair of spaces and the temperatures spaces, respectively. In Subsection \ref{sub:forms} we define the discrete computable forms. Finally in Subsection \ref{sub:vem problem} we show the virtual elements discretization of Problems \eqref{eq:pb variazionale} and \eqref{eq:pb variazionale ker}. Let $k \geq 2$ be the ``polynomial'' order of the method. The lowest order case $k=1$ can be treated as well using a slightly different approach \cite{ABMV:2014}. We recall that, in standard finite element fashion, the VEM spaces are first defined elementwise and then assembled globally. In the following we will denote by $E$ a general polygon having $\ell_e$ edges $e$, while $\boldsymbol{n}_E$ will denote the unit vector that is normal to $\partial E$ and outward with respect to $E$. \subsection{Virtual Elements velocities space and pressure space} \label{sub:v-p spaces} In the present section we outline an overview of the divergence-free Virtual Elements spaces for the Navier-Stokes equation \cite{BLV:2017,vacca:2018,BLV:2018}. We consider on each polygonal element $E \in \Omega_h$ the ``enhanced'' virtual space \begin{equation} \label{eq:v-loc} \begin{aligned} \VV_h(E) := \biggl\{ \boldsymbol{v}_h \in [C^0(\overline{E})]^2 \,\,\, \text{s.t.} \,\,\, (i)& \,\, \boldsymbol{\Delta} \boldsymbol{v}_h + \nabla s \in \xx^{\perp} \numberset{P}_{k-1}(E), \,\,\text{ for some $s \in L_0^2(E)$,} \\ (ii)& \,\, {\rm div} \, \boldsymbol{v}_h \in \numberset{P}_{k-1}(E) \,, \\ (iii) & \,\, {\boldsymbol{v}_h}_{|e} \in [\numberset{P}_k(e)]^2 \,\,\, \forall e \in \partial E, \\ (iv) & \,\, (\boldsymbol{v}_h - \Pi^{\nabla,E}_k \boldsymbol{v}_h, \, \xx^{\perp} \, \widehat{p}_{k-1} )_E = 0 \,\,\, \text{$\forall \widehat{p}_{k-1} \in \widehat{\numberset{P}}_{k-1 \setminus k-3}(E)$} \,\,\biggr\} \,, \end{aligned} \end{equation} where $\xx^{\perp} = (x_2, -x_1)$. We here summarize the main properties of the space $\VV_h(E)$ (we refer to \cite{BLV:2018} for a deeper analysis). \begin{itemize} \item [\textbf{(P1)}] \textbf{Polynomial inclusion:} $[\numberset{P}_k(E)]^2 \subseteq \VV_h(E)$; \item [\textbf{(P2)}] \textbf{Degrees of freedom:} the following linear operators $\mathbf{D_{\boldsymbol{V}}}$ constitute a set of DoFs for $\VV_h(E)$: \begin{itemize} \item[$\mathbf{D_{\boldsymbol{V}}1}$] the values of $\boldsymbol{v}_h$ at the vertexes of the polygon $E$, \item[$\mathbf{D_{\boldsymbol{V}}2}$] the values of $\boldsymbol{v}_h$ at $k-1$ distinct points of every edge $e \in \partial E$, \item[$\mathbf{D_{\boldsymbol{V}}3}$] the moments of $\boldsymbol{v}_h$ $$ \frac{1}{|E|}\int_E \boldsymbol{v}_h \cdot \boldsymbol{m}^{\perp} m_{\boldsymbol{\alpha}} \, {\rm d}E \qquad \text{for any $m_{\boldsymbol{\alpha}} \in \numberset{M}_{k-3}(E)$,} $$ where $\boldsymbol{m}^{\perp} := \frac{1}{h_E}(x_2 - x_{2,E}, -x_1 + x_{1,E})$, \item[$\mathbf{D_{\boldsymbol{V}}4}$] the moments of ${\rm div} \boldsymbol{v}_h$ $$ \frac{h_E}{|E|}\int_E {\rm div} \boldsymbol{v}_h \, m_{\boldsymbol{\alpha}} \, {\rm d}E \qquad \text{for any $m_{\boldsymbol{\alpha}} \in \numberset{M}_{k-1}(E)$ with $|\boldsymbol{\alpha}| > 0$;} $$ \end{itemize} \item [\textbf{(P3)}] \textbf{Polynomial projections:} the DoFs $\mathbf{D_{\boldsymbol V}}$ allow us to compute the following linear operators: \[ {\Pi^{0, E}_k} \colon \VV_h(E) \to [\numberset{P}_{k}(E)]^2, \qquad {\boldsymbol{\Pi}^{0, E}_{k-1}} \colon \boldsymbol{\nabla} \VV_h(E) \to [\numberset{P}_{k-1}(E)]^{2 \times 2} \,. \] \end{itemize} The global velocity space $\VV_h$ is defined by gluing the local spaces with the obvious associated sets of global DoFs: \begin{equation} \label{eq:v-glo} \VV_h := \{\boldsymbol{v}_h \in \boldsymbol{V} \quad \text{s.t.} \quad {\boldsymbol{v}_h}_{|E} \in \VV_h(E) \quad \text{for all $E \in \Omega_h$} \}\,. \end{equation} The discrete pressure space $Q_h$ is given by the piecewise polynomial functions of degree $k-1$, i.e. \begin{equation} \label{eq:q-glo} Q_h := \{q_h \in Q \quad \text{s.t.} \quad {q_h}_{|E} \in \numberset{P}_{k-1}(E) \quad \text{for all $E \in \Omega_h$} \}\,. \end{equation} The couple of spaces $(\VV_h, \, Q_h)$ is inf-sup stable \cite{BLV:2017}, and we denote with $\widehat{\beta} >0$ the inf-sup stability constant, i.e. $\widehat{\beta}$ is such that \begin{equation} \label{eq:infsup} \sup_{\boldsymbol{v}_h \in \VV_h} \frac{b(\boldsymbol{v}_h, q_h)}{\vert \boldsymbol{v}_h\vert_{1, \Omega}} \geq \widehat{\beta} \|q_h\|_{0,E} \qquad \text{for any $q_h \in Q_h$.} \end{equation} Let us introduce the discrete kernel \begin{equation} \label{eq:z-glo} \ZZ_h := \{ \boldsymbol{v}_h \in \VV_h \quad \text{s.t.} \quad b(\boldsymbol{v}_h, q_h) = 0 \quad \text{for all $q_h \in Q_h$}\} \end{equation} then recalling $(ii)$ in \eqref{eq:v-loc} and \eqref{eq:q-glo}, the following kernel inclusion holds \begin{equation} \label{eq:kernel} \ZZ_h \subseteq \ZZ \,, \end{equation} i.e. the functions in the discrete kernel are exactly divergence-free. \subsection{Virtual Elements space for the temperatures} \label{sub:t-space} In this section, we briefly introduce the $H^1$-conforming virtual space for the temperatures, that consists on the so-called nodal ``enhanced'' virtual space \cite{projectors}. We thus consider on each element $E \in \Omega_h$ the space \begin{equation} \label{eq:w-loc} \begin{aligned} \W_h(E) := \biggl\{ \sigma_h \in C^0(\overline{E}) \,\,\, \text{s.t.} \,\,\, (i)& \, \, \Delta \sigma_h \in \numberset{P}_{k}(E), \\ (ii) & \, \, {\sigma_h}_{|e} \in \numberset{P}_k(e) \,\,\, \forall e \in \partial E, \\ (iii) & \, \, (\sigma_h - \Pi^{\nabla,E}_k \sigma_h, \, \widehat{p}_{k} )_E = 0 \,\,\, \text{$\forall \widehat{p}_{k} \in \widehat{\numberset{P}}_{k \setminus k-2}(E)$} \,\,\biggr\} \,. \end{aligned} \end{equation} We here summarize the main properties of the space $\W_h(E)$ (see \cite{projectors}). \begin{itemize} \item [\textbf{(P4)}] \textbf{Polynomial inclusion:} $\numberset{P}_k(E) \subseteq \W_h(E)$; \item [\textbf{(P5)}] \textbf{Degrees of freedom:} the following linear operators $\mathbf{D_{\boldsymbol \Sigma}}$ constitute a set of DoFs for $\W_h(E)$: \begin{itemize} \item[$\mathbf{D_{\boldsymbol \Sigma}1}$] the values of $\sigma_h$ at the vertexes of the polygon $E$, \item[$\mathbf{D_{\boldsymbol \Sigma}2}$] the values of $\sigma_h$ at $k-1$ distinct points of every edge $e \in \partial E$, \item[$\mathbf{D_{\boldsymbol \Sigma}3}$] the moments of $\sigma_h$ $$ \frac{1}{|E|}\int_E \sigma_h \, m_{\boldsymbol{\alpha}} \, {\rm d}E \qquad \text{for any $m_{\boldsymbol{\alpha}} \in \numberset{M}_{k-2}(E)$;} $$ \end{itemize} \item [\textbf{(P6)}] \textbf{Polynomial projections:} the DoFs $\mathbf{D_{\boldsymbol \Sigma}}$ allow us to compute the following linear operators: \[ {\Pi^{0, E}_k} \colon \W_h(E) \to \numberset{P}_{k}(E), \qquad {\Pi^{0, E}_{k-1}} \colon \nabla \W_h(E) \to [\numberset{P}_{k-1}(E)]^2 \,. \] \end{itemize} The global velocity space $\W_h$ is defined by gluing the local spaces with the obvious associated sets of global DoFs: \begin{equation*} \label{eq:w-glo-g} \W_h := \{\sigma_h \in \Sigma \,\,\, \text{s.t.} \,\,\, {\sigma_h}_{|E} \in \W_h(E) \,\,\, \text{for all $E \in \Omega_h$} \}\,, \end{equation*} with its homogeneous boundary counterpart \begin{equation} \label{eq:w-glo-z} \W_{0,h} := \{\sigma_h \in \W_h \,\,\, \text{s.t.} \,\,\, {\sigma_h}_{|\partial \Omega} = 0 \}\,. \end{equation} In order to treat non-homogeneous Dirichlet data, accordingly to the data assumptions \textbf{(A0)}, let us consider the ``nodal'' interpolant $\theta_{\rm D, h}$ of $\theta_{\rm D}$ defined by: \begin{equation} \label{eq:thd} \begin{gathered} \theta_{\rm D, h} \in C^0(\partial \Omega) \quad \text{s.t.} \quad {\theta_{\rm D, h}}_{|e} \in \numberset{P}_{k}(e) \quad \text{for any edge $e \subset \partial \Omega$,} \\ \theta_{\rm D, h}(\boldsymbol{x}_i) = \theta_{\rm D}(\boldsymbol{x}_i) \qquad \text{for any node $\boldsymbol{x}_i \in e$, for any edge $e \subset \partial \Omega$.} \end{gathered} \end{equation} Then the discrete counterpart of the space $\Sigma_{\rm D}$ is defined as follows: \begin{equation} \label{eq:w-glo-d} \W_{{\rm D}, h} := \{\sigma_h \in \W_h \,\,\, \text{s.t.} \,\,\, {\sigma_h}_{|\partial \Omega} = \theta_{\rm D, h} \}\,. \end{equation} \subsection{Virtual Element forms} \label{sub:forms} The next step in the construction of our method is the definition of discrete versions of the continuous second order forms in \eqref{eq:forma av} and \eqref{eq:forma as}, of the convective forms in \eqref{eq:forma cv} and \eqref{eq:forma cs} and the approximation of the right-hand side terms. It is clear that for an arbitrary function in $\VV_h$ or $\W_h$ the forms are not computable since the discrete functions are not known in closed form. Therefore, following the usual procedure in the VEM setting, we need to construct discrete forms that are computable by {means of the only knowledge of} the DoFs. In the light of properties \textbf{(P3)} and \textbf{(P6)} we define the computable discrete forms. \noindent $\bullet$ \, Discrete second order forms (cf. \eqref{eq:forma av} and \eqref{eq:forma as}) \begin{align} \label{eq:forma avhl} {a_{V,h}^E}(\theta_h; \boldsymbol{u}_h, \boldsymbol{v}_h) &:= \int_{E} \nu\bigl({\Pi^{0, E}_k} \theta_h \bigr) \, {\boldsymbol{\Pi}^{0, E}_{k-1}} \boldsymbol{\epsilon}(\boldsymbol{u}_h) : {\boldsymbol{\Pi}^{0, E}_{k-1}}\boldsymbol{\epsilon} (\boldsymbol{v}_h) \, {\rm d}E +{s_{V}^E}(\theta_h; \boldsymbol{u}_h, \, \boldsymbol{v}_h) , \\ \label{eq:forma ashl} {a_{T,h}^E}(\theta_h, \sigma_h) &:= \int_{E} \kappa \, {\Pi^{0, E}_{k-1}}\nabla \theta_h \cdot {\Pi^{0, E}_{k-1}} \nabla \sigma_h \, {\rm d}E + {s_{T}^E}(\theta_h, \, \sigma_h) , \end{align} where the VEM stabilizing terms are given by \begin{align} \label{eq:svstab} {s_{V}^E}(\theta_h; \boldsymbol{u}_h, \, \boldsymbol{v}_h) &:= \nu(\Pi^{0,E}_0 \theta_h) \, S_V^E \bigl( (I - {\Pi^{0, E}_k}) \boldsymbol{u}_h, \, (I - {\Pi^{0, E}_k} ) \boldsymbol{v}_h \bigr) \,, \\ \label{eq:ssstab} {s_{T}^E}(\theta_h, \, \sigma_h) &:= \bigl(\Pi^{0,E}_0 \kappa \bigr) \, S_T^E \bigl( (I - {\Pi^{0, E}_k}) \theta_h, \, (I - {\Pi^{0, E}_k} ) \sigma_h \bigr) \,, \end{align} and where $S_V^E(\cdot, \cdot) \colon \VV_h(E) \times \VV_h(E) \to \numberset{R}$ and $S_E^E(\cdot, \cdot) \colon \W_h(E) \times \W_h(E) \to \numberset{R}$ are computable symmetric discrete forms satisfying \begin{equation} \label{eq:stab} \begin{gathered} \vert \boldsymbol{v}_h \vert^2_{1,E} \lesssim S_V^E(\boldsymbol{v}_h, \boldsymbol{v}_h) \lesssim \vert \boldsymbol{v}_h \vert^2_{1,E} \quad \text{for all $\boldsymbol{v} \in \VV_h(E) \cap \ker {\Pi^{0, E}_k}$,} \\ \vert \sigma_h \vert^2_{1,E} \lesssim S_T^E(\sigma_h, \sigma_h) \lesssim \vert \sigma_h \vert^2_{1,E} \quad \text{for all $\sigma_h \in \W_h(E) \cap \ker {\Pi^{0, E}_k}$.} \end{gathered} \end{equation} Many examples can be found in the VEM literature \cite{volley,BDR:2017}. In the present paper we consider the so-called \texttt{dofi-dofi} stabilization defined as follows: let $\vec{\boldsymbol{u}}_h$, $\vec{\boldsymbol{v}}_h$ and $\vec{\theta}_h$, $\vec{\sigma}_h$ denote the real valued vectors containing the values of the local degrees of freedom associated to $\boldsymbol{u}_h$, $\boldsymbol{v}_h$ in the space $\VV_h$ and to $\theta_h$ and $\sigma_h$ in the space $\W_h$ respectively, then \begin{equation*} \label{eq:dofi} S_V^E(\boldsymbol{u}_h, \boldsymbol{v}_h) :=\vec{\boldsymbol{u}}_h \cdot \vec{\boldsymbol{v}}_h \,, \qquad S_T^E(\theta_h, \sigma_h) :=\vec{\theta}_h \cdot \vec{\sigma}_h \,. \end{equation*} \noindent $\bullet$ \, Discrete convective forms (cf. \eqref{eq:forma cv} and \eqref{eq:forma cs}) \begin{align} \label{eq:forma cvhl} c_{V,h}^E(\boldsymbol{w}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) &:= \int_{E} \bigl[{\boldsymbol{\Pi}^{0, E}_{k-1}}( \boldsymbol{\nabla} \boldsymbol{u}_h ) \, {\Pi^{0, E}_k} \boldsymbol{w}_h \bigr] \cdot {\Pi^{0, E}_k} \boldsymbol{v}_h \,{\rm d}E, \\ \label{eq:forma cshl} c_{T,h}^E(\boldsymbol{u}_h; \, \theta_h, \sigma_h) &:= \int_{E} \bigl({\Pi^{0, E}_k} \boldsymbol{u}_h \cdot {\Pi^{0, E}_{k-1}}\nabla \theta_h \bigr) \, {\Pi^{0, E}_k} \sigma_h \,{\rm d}E \,, \end{align} and their skew-symmetric formulation (cf. \eqref{eq:forma cvskew} and \eqref{eq:forma csskew}) \begin{align} \label{eq:forma cvhlskew} &{c_{V,h}^{{\rm skew}, E}}(\boldsymbol{w}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) := \frac{1}{2} \bigl(c_{V,h}^E(\boldsymbol{w}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - c_{V,h}^E(\boldsymbol{w}_h; \, \boldsymbol{v}_h, \boldsymbol{u}_h) \bigr) \,, \\ \label{eq:forma cshlskew} &{c_{T,h}^{{\rm skew}, E}}(\boldsymbol{w}_h; \, \theta_h, \sigma_h) := \frac{1}{2} \bigl (c_{T,h}^E(\boldsymbol{w}_h; \, \theta_h, \sigma_h) - c_{T,h}^E(\boldsymbol{w}_h; \, \sigma_h, \theta_h) \bigr) \,. \end{align} \noindent $\bullet$ \, \, Discrete external forces \begin{equation} \label{eq:forma fh gh} {\boldsymbol{f}_h}_{|E} := {\Pi^{0, E}_k} \boldsymbol{f} \,, \quad {g_h}_{|E} := {\Pi^{0, E}_k} g \,, \qquad \text{for all $E \in \Omega_h$.} \end{equation} \noindent The global forms, defined accordingly to the notation \eqref{eq:XG}, satisfy the following. \smallskip\noindent \textbf{(P7) Stability properties of the discrete forms:} \\ $\bullet$ \,\, for any $\theta \in \Sigma$ the bilinear form ${a_{V,h}}(\theta; \, \cdot, \cdot)$ is coercive and continuous, i.e. \[ {a_{V,h}}(\theta; \, \boldsymbol{v}_h, \boldsymbol{v}_h) \geq \alpha_* \nu_* \vert \boldsymbol{v}_h \vert_{1, \Omega}^2 \,, \quad {a_{V,h}}(\theta; \, \boldsymbol{u}_h, \boldsymbol{v}_h) \leq \alpha^* \nu^* \vert \boldsymbol{u}_h \vert_{1, \Omega} \vert \boldsymbol{v}_h \vert_{1, \Omega}\,, \] for all $\boldsymbol{u}_h$, $\boldsymbol{v}_h \in \VV_h$; \noindent $\bullet$ \,\, the bilinear form ${a_{T,h}}(\cdot, \cdot)$ is coercive and continuous, i.e. \[ {a_{T,h}}(\sigma_h, \sigma_h) \geq \beta_* \kappa_* \vert \sigma_h \vert_{1, \Omega}^2 \,, \quad {a_T}(\theta_h, \sigma_h) \leq \beta^* \kappa^* \vert \theta_h \vert_{1, \Omega} \vert \sigma_h \vert_{1, \Omega}\,, \] for all $\theta_h$, $\sigma_h \in \W_h$; \noindent $\bullet$ \,\, the convective trilinear forms ${c_{V,h}^{\rm skew}}(\cdot; \, \cdot, \cdot)$ and ${c_{T,h}^{\rm skew}}(\cdot; \, \cdot, \cdot)$ are continuous, i.e. \[ \begin{aligned} {c_{V,h}^{\rm skew}}(\boldsymbol{w}; \, \boldsymbol{u}, \boldsymbol{v}) &\leq \widehat{C}_{\rm conv} \vert \boldsymbol{w} \vert_{1, \Omega} \vert \boldsymbol{u} \vert_{1, \Omega} \vert \boldsymbol{v} \vert_{1, \Omega} \,, \qquad & \text{for all $\boldsymbol{w}$, $\boldsymbol{u}$, $\boldsymbol{v} \in \boldsymbol{V}$,} \\ {c_{T,h}^{\rm skew}}(\boldsymbol{u}; \, \theta, \sigma) &\leq \widehat{C}_{\rm conv} \vert \boldsymbol{u} \vert_{1, \Omega} \Vert \theta \Vert_{1, \Omega} \Vert \sigma \Vert_{1, \Omega} \,, \qquad & \text{for all $\boldsymbol{u} \in \boldsymbol{V}$, $\theta$, $\sigma \in \Sigma$,} \end{aligned} \] where $\alpha_*$, $\alpha^*$, $\beta_*$, $\beta^*$, $\widehat{C}_{\rm conv}$ are positive constants depending on the domain $\Omega$, on the ``polynomial'' order of the method $k$ and on the regularity constant appearing in the mesh assumption \textbf{(A1)}. \subsection{Virtual Element problem} \label{sub:vem problem} { Having in mind} the spaces \eqref{eq:v-glo}, \eqref{eq:q-glo}, \eqref{eq:w-glo-d} and \eqref{eq:w-glo-z}, the discrete forms \eqref{eq:forma avhl}, \eqref{eq:forma ashl}, \eqref{eq:forma cvhlskew}, \eqref{eq:forma cshlskew}, the form \eqref{eq:forma b}, the discrete loading terms \eqref{eq:forma fh gh}, the virtual element discretization of Problem \eqref{eq:pb variazionale} is given by \begin{equation} \label{eq:pb vem} \left\{ \begin{aligned} & \text{find $(\boldsymbol{u}_h, p_h, \theta_h) \in \VV_h \times Q_h \times \W_{{\rm D}, h}$, such that} \\ &\begin{aligned} {a_{V,h}}(\theta_h, \, \boldsymbol{u}_h, \boldsymbol{v}_h) + {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) + b(\boldsymbol{v}_h, p_h) &= (\boldsymbol{f}_h, \boldsymbol{v}_h) \quad & &\text{$\forall \boldsymbol{v}_h \in \VV_h$,} \\ b(\boldsymbol{u}_h, q_h) &= 0 \quad & & \text{$\forall q_h \in Q_h$,} \\ {a_{T,h}}(\theta_h, \sigma_h) + {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \theta_h, \sigma_h) &= (g_h, \sigma_h) \qquad & &\text{$\forall \sigma_h \in \W_{0,h}$.} \end{aligned} \end{aligned} \right. \end{equation} Recalling the definition of { the}discrete kernel $\ZZ_h$ in \eqref{eq:z-glo}, the previous problem can be also written in the kernel formulation, i.e. \begin{equation} \label{eq:pb vem ker} \left\{ \begin{aligned} & \text{find $(\boldsymbol{u}_h, \theta_h) \in \boldsymbol{Z}_h \times \W_{{\rm D}, h}$, such that} \\ &\begin{aligned} {a_{V,h}}(\theta_h, \, \boldsymbol{u}_h, \boldsymbol{v}_h) + {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) &= (\boldsymbol{f}_h, \boldsymbol{v}_h) \quad & &\text{$\forall \boldsymbol{v}_h \in \boldsymbol{Z}_h$,} \\ {a_{T,h}}(\theta_h, \sigma_h) + {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \theta_h, \sigma_h) &= (g_h, \sigma_h) \quad & &\text{$\forall \sigma_h \in \W_{0,h}$.} \end{aligned} \end{aligned} \right. \end{equation} \section{Existence of discrete solutions} \label{sec:exi} In the present section we establish the existence of discrete solutions. \begin{proposition} \label{prp:existence} Under the data assumptions \cfan{(A0)}, assume moreover that there exists $C_{\rm exi}$ depending on the domain $\Omega$ and the shape regularity constant in assumption \cfan{(A1)} s.t. \begin{equation}\label{aux:Prop:5:1} \Vert \theta_{\rm D, h}\Vert_{1/2, \partial \Omega} \leq C_{\rm exi} \min\{\nu_*, \kappa_* \} \end{equation} then Problem \eqref{eq:pb vem} admits at least a solution $(\boldsymbol{u}_h, p_h, \theta_h) \in \ZZ_h \times Q_h \times \W_{{\rm D}, h}$. Moreover, the following bound holds \begin{equation} \label{eq:bound disc} \vert \boldsymbol{u}_h \vert_{1, \Omega}^2 + \Vert \theta_h \Vert_{1, \Omega}^2 \leq \widehat{C}_{\rm est}^2 \left( \Vert \boldsymbol{f}_h \Vert_{\ZZ_h^*}^2 + \vert g \vert_{-1, \Omega}^2 + \Vert \theta_{\rm D} \Vert_{1/2, \partial \Omega}^2 \right) \end{equation} where the constant $\widehat{C}_{\rm est}$ depends on the domain $\Omega$ and on the constants $\kappa_*$, $\kappa^*$ and $\nu_*$ in the data assumptions \cfan{(A0)} and the shape regularity constant in assumption \cfan{(A1)}. \end{proposition} \begin{proof} Let $\theta^{\partial}_h \in \W_{{\rm D}, h}$ be the solution of the following auxiliary problem \begin{equation*} \label{eq:tparh} \left\{ \begin{aligned} & \text{find $\theta^{\partial}_h \in \W_{{\rm D}, h}$, such that} \\ &\begin{aligned} \int_{\Omega} \nabla \theta^{\partial}_h \cdot \nabla \sigma_h \, {\rm d}\Omega &= 0 \quad & &\text{$\forall \sigma_h \in \W_{0,h}$.} \end{aligned} \end{aligned} \right. \end{equation*} Then, by Lax-Milgram Lemma, $\theta^{\partial}_h$ satisfies the bound \begin{equation} \label{eq:tpar-bouh} \Vert \theta^{\partial}_h\Vert_{1, \Omega} \leq C_\Omega \, \Vert\theta_{\rm D, h}\Vert_{1/2, \partial \Omega} \,. \end{equation} Notice that Problem \eqref{eq:pb vem ker} can be equivalently formulated as follows \begin{equation} \label{eq:pb vem kerZ} \left\{ \begin{aligned} & \text{find $(\boldsymbol{u}_h, \theta^o_h) \in \ZZ_h \times \W_{0,h}$, such that} \\ &\begin{aligned} {a_{V,h}}(\theta^o_h + \theta^{\partial}_h, \, \boldsymbol{u}_h, \boldsymbol{v}_h) + {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) &= (\boldsymbol{f}_h, \boldsymbol{v}_h) & &\text{$\forall \boldsymbol{v}_h \in \ZZ_h$,} \\ {a_{T,h}}(\theta^o_h + \theta^{\partial}_h, \sigma_h) + {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \theta^o_h + \theta^{\partial}_h, \sigma_h) &= (g_h, \sigma_h) & &\text{$\forall \sigma_h \in \W_{0,h}$.} \end{aligned} \end{aligned} \right. \end{equation} We now prove that Problem \eqref{eq:pb vem kerZ} admits solutions. Let $\beth := \ZZ_h \times \W_{0,h}$ and, for any $(\boldsymbol{u}_h, \theta^o_h) \in \beth$ let $\mathcal{A}^{(\boldsymbol{u}_h, \theta^o_h)}\colon \beth \to \numberset{R}$ given by \begin{equation} \label{eq:prpex1} \begin{aligned} \mathcal{A}^{(\boldsymbol{u}_h, \theta^o_h)}(\boldsymbol{v}_h, \sigma_h) &:= {a_{V,h}}(\theta^o_h + \theta^{\partial}_h, \, \boldsymbol{u}_h, \boldsymbol{v}_h) + {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - (\boldsymbol{f}_h, \boldsymbol{v}_h) \\ & \quad + {a_{T,h}}(\theta^o_h + \theta^{\partial}_h, \sigma_h) + {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \theta^o_h + \theta^{\partial}_h, \sigma_h) - (g_h, \sigma_h) \end{aligned} \end{equation} for all $(\boldsymbol{v}_h, \sigma_h) \in \beth$. Then employing property \textbf{(P7)} and the skew-symmetry of the convective forms ${c_{V,h}^{\rm skew}}$ and ${c_{T,h}^{\rm skew}}$ (cf. \eqref{eq:forma cvhlskew} and \eqref{eq:forma cshlskew}) we obtain \begin{equation} \label{eq:prpex2} \begin{aligned} \mathcal{A}^{(\boldsymbol{u}_h, \theta^o_h)}(\boldsymbol{u}_h, \theta^o_h) & \geq \alpha_* \nu_* \vert \boldsymbol{u}_h\vert_{1, \Omega}^2 + \beta_* \kappa_* \vert \theta^o_h\vert_{1, \Omega}^2 - \Vert \boldsymbol{f}_h \Vert_{\ZZ_h^*} \vert \boldsymbol{u}_h\vert_{1, \Omega} \\ &\qquad - \vert g_h\vert_{-1, \Omega} \vert \theta^o_h \vert_{-1, \Omega} + {a_{T,h}}(\theta^{\partial}_h, \theta^o_h) + {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \theta^{\partial}_h, \theta^o_h) \,. \end{aligned} \end{equation} From property \textbf{(P7)} and \eqref{eq:tpar-bouh} it holds that \begin{equation} \label{eq:prpex3} \begin{aligned} {a_{T,h}}(\theta^{\partial}_h, \theta^o_h) &\geq -{a_{T,h}}(\theta^{\partial}_h, \theta^{\partial}_h)^{1/2} \, {a_{T,h}}(\theta^o_h, \theta^o_h)^{1/2} \geq -\beta^* \,\kappa^* \, C_\Omega \, \Vert \theta_{\rm D, h} \Vert_{1/2, \Omega} \, \vert \theta^o_h\vert_{1, \Omega} \,, \\ {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \theta^{\partial}_h, \theta^o_h) &\geq - \widehat{C}_{\rm conv} \Vert \theta^{\partial}_h \Vert_{1, \Omega} \, \vert \boldsymbol{u}_h\vert_{1, \Omega} \vert \theta^o_h\vert_{1, \Omega} \geq - \frac{\widehat{C}_{\rm conv}}{2} C_\Omega \, \Vert \theta_{\rm D, h} \Vert_{1/2, \Omega} (\vert \boldsymbol{u}_h\vert_{1, \Omega}^2 + \vert \theta^o_h\vert_{1, \Omega}^2) \,. \end{aligned} \end{equation} Therefore inserting \eqref{eq:prpex3} in \eqref{eq:prpex2}, we infer \begin{equation} \label{eq:prpex4} \begin{aligned} \mathcal{A}^{(\boldsymbol{u}_h, \theta^o_h)}(\boldsymbol{u}_h, \theta^o_h)& \geq \left( \min\{\alpha_*, \beta_* \}\min\{\nu_*, \kappa_* \} - \frac{1}{2}\widehat{C}_{\rm conv} C_\Omega \,\Vert \theta_{\rm D, h} \Vert_{1/2, \Omega} \right) (\vert \boldsymbol{u}_h\vert_{1, \Omega}^2 + \vert \theta^o_h\vert_{1, \Omega}^2) + \\ & \quad - \left( (\Vert \boldsymbol{f}_h \Vert_{\ZZ_h^*}^2 + \vert g_h\vert_{-1, \Omega}^2)^{1/2} + \beta^*\, \kappa^*C_\Omega \, \Vert \theta_{\rm D, h} \Vert_{1/2, \Omega} \right) (\vert \boldsymbol{u}_h\vert_{1, \Omega}^2 + \vert \theta^o_h\vert_{1, \Omega}^2)^{1/2} \,. \end{aligned} \end{equation} Let us set \[ C_{\rm exi} \leq \frac{\min\{\alpha_*, \beta_* \}}{\widehat{C}_{\rm conv} C_\Omega } \,, \] then from { \eqref{aux:Prop:5:1}} and \eqref{eq:prpex4} we get \begin{equation} \label{eq:prpex5} \begin{aligned} \mathcal{A}^{(\boldsymbol{u}_h, \theta^o_h)}(\boldsymbol{u}_h, \theta^o_h)& \geq \frac{1}{2} \min\{\alpha_*, \beta_* \}\min\{\nu_*, \kappa_* \} (\vert \boldsymbol{u}_h\vert_{1, \Omega}^2 + \vert \theta^o_h\vert_{1, \Omega}^2) + \\ & \quad - \left( (\Vert \boldsymbol{f}_h \Vert_{\ZZ_h^*}^2 + \vert g_h\vert_{-1, \Omega}^2)^{1/2} + \beta^* \,\kappa^* C_\Omega \,\Vert \theta_{\rm D, h} \Vert_{1/2, \Omega} \right) (\vert \boldsymbol{u}_h\vert_{1, \Omega}^2 + \vert \theta^o_h\vert_{1, \Omega}^2)^{1/2} \,. \end{aligned} \end{equation} Let \begin{equation} \label{eq:rho} \rho := 2\frac{(\Vert \boldsymbol{f}_h \Vert_{\ZZ_h^*}^2 + \vert g_h\vert_{-1, \Omega}^2)^{1/2} + \beta^* \,\kappa^* C_\Omega \, \Vert \theta_{\rm D, h} \Vert_{1/2, \Omega}} {\min\{\alpha_*, \beta_* \}\min\{\nu_*, \kappa_* \}} \end{equation} and $ \mathcal{S} := \{ (\boldsymbol{v}_h, \sigma_h) \in \beth \quad \text{s.t.} \quad \vert \boldsymbol{u}_h\vert_{1, \Omega}^2 + \vert \theta^o_h\vert_{1, \Omega}^2 \leq \rho^2\} $ then \eqref{eq:prpex5} implies \[ \mathcal{A}^{(\boldsymbol{u}_h, \theta^o_h)}(\boldsymbol{u}_h, \theta^o_h) \geq 0 \qquad \text{for any $(\boldsymbol{u}_h, \theta^o_h) \in \partial \mathcal{S}$.} \] Employing the fixed-point Theorem \cite[Chap. IV, Corollary 1.1]{girault-raviart:book2} there exists $(\boldsymbol{u}_h, \theta^o_h) \in \mathcal{S}$ s.t. $\mathcal{A}^{(\boldsymbol{u}_h, \theta^o_h)}(\boldsymbol{v}_h, \sigma_h) = 0$ for all $(\boldsymbol{v}_h, \sigma_h) \in \beth$, i.e. Problem \eqref{eq:pb vem kerZ} admits solution. Let $\theta_h = \theta^o_h + \theta^{\partial}_h$, then $(\boldsymbol{u}_h, \theta_h) \in \ZZ_h \times \W_{0,h}$ is a solution of Problem \eqref{eq:pb vem ker}. Bound \eqref{eq:bound disc} follows from \eqref{eq:rho} and \eqref{eq:tpar-bouh} and the definition of $ \mathcal{S}$. Finally, the existence of solutions $(\boldsymbol{u}_h, p_h, \theta^o_h) \in \VV_h \times Q_h \times \W_{0,h}$ of the equivalent Problem \eqref{eq:pb vem} is a direct consequence of \eqref{eq:infsup} (see for instance \cite{boffi-brezzi-fortin:book}). \end{proof} \begin{remark} \label{rm:existence} Notice that, differently from the continuous problem, in order to establish the existence result for the discrete problem \eqref{aux:Prop:5:1} we need that the conductivity $\kappa$ and the viscosity $\nu$ are sufficiently large with respect to the Dirichlet datum $\theta_D$ (cf. bound \eqref{eq:pb vem}). A careful inspection of the proof of Theorem \ref{thm:exi} (see \cite[Theorem~2.2]{Bernardi:2015}) shows that the gap between the continuous and the discrete case is the different construction of the (skew-symmetric) convective forms $c_T(\boldsymbol{u}; \, \cdot, \cdot)$ and $c_T^{\rm skew}(\boldsymbol{u}_h; \, \cdot, \cdot)$. \end{remark} We now investigate whether Problem \eqref{eq:pb vem ker} admits a unique solution. We introduce the analysis with the following lemma. \begin{lemma} \label{lm:lipshitz} Let $\sigma_{h,1}$, $\sigma_{h,2} \in \W_{{\rm D}, h}$, let $\boldsymbol{v}_h \in \VV_h$ such that $\boldsymbol{v}_h \in [W_1^p(\Omega_h)]^2$ with $p>2$. There exists a positive constant $\widehat{C}_{\rm lip}$ only depending on $\Omega$, $k$, and the shape regularity constant in \cfan{(A1)} such that for any $\boldsymbol{w}_h\in \VV_h$ \begin{equation} \label{eq:cclip} \begin{aligned} \vert {a_{V,h}}(\sigma_{h,1}; \boldsymbol{v}_h, \boldsymbol{w}_h) - {a_{V,h}}(\sigma_{h,2}; \boldsymbol{v}_h, \boldsymbol{w}_h) \vert \leq \widehat{C}_{\rm lip} \nu_{\rm lip} \vert \sigma_{h,1}-\sigma_{h,2}\vert_{1, \Omega} \, \vert \boldsymbol{v}_h \vert_{[W^1_q(\Omega)]^2} \, \vert \boldsymbol{w}_h \vert_{1, \Omega} \,. \end{aligned} \end{equation} \end{lemma} \begin{proof} From definition of ${a_{V,h}}(\cdot, \cdot)$ (cf. \eqref{eq:forma avhl}) we have \begin{equation} \label{eq:lmlip1} \begin{aligned} \vert {a_{V,h}}(\sigma_{h,1}; \boldsymbol{v}_h, \boldsymbol{w}_h) - {a_{V,h}}(\sigma_{h,2}; \boldsymbol{v}_h, \boldsymbol{w}_h) \vert \leq \sum_{E \in \Omega_h} \left \vert {s_{V}^E}(\sigma_{h,1}; \boldsymbol{v}_h, \boldsymbol{w}_h) - {s_{V}^E}(\sigma_{h,2}; \boldsymbol{v}_h, \boldsymbol{w}_h) \right \vert + \\ + \sum_{E \in \Omega_h} \int_{E} \bigl \vert \nu\bigl({\Pi^{0, E}_k} \sigma_{h,1} \bigr)- \nu\bigl({\Pi^{0, E}_k} \sigma_{h,2} \bigr) \bigr \vert \, \bigl\vert {\boldsymbol{\Pi}^{0, E}_{k-1}}\boldsymbol{\epsilon} (\boldsymbol{v}_h) \bigr \vert \bigl\vert {\boldsymbol{\Pi}^{0, E}_{k-1}}\boldsymbol{\epsilon} (\boldsymbol{w}_h) \bigr \vert \, {\rm d}E := \eta_1 + \eta_2 \,. \end{aligned} \end{equation} Concerning the term $\eta_1$, from \eqref{eq:svstab} and \eqref{eq:stab}, and recalling property \textbf{(P0)}, we infer \[ \begin{aligned} \eta_1 &\leq \sum_{E \in \Omega_h} \bigl \vert \nu(\Pi^{0,E}_0 \sigma_{h,1}) - \nu(\Pi^{0,E}_0 \sigma_{h,2}) \bigr \vert \, \bigl \vert S_V^E \bigl( (I - {\Pi^{0, E}_k}) \boldsymbol{v}_h, \, (I - {\Pi^{0, E}_k} ) \boldsymbol{w}_h \bigr) \bigr \vert \\ &\lesssim \nu_{\rm lip} \sum_{E \in \Omega_h} \vert \Pi^{0,E}_0 (\sigma_{h,1} - \sigma_{h,2}) \vert \, \vert (I - {\Pi^{0, E}_k}) \boldsymbol{v}_h \vert_{1,E} \, \vert(I - {\Pi^{0, E}_k} ) \boldsymbol{w}_h \vert_{1,E} \,. \end{aligned} \] Then employing the inverse estimate \eqref{eq:inverse} and the Bramble-Hilbert Lemma \eqref{eq:bramble} \[ \begin{aligned} \eta_1 &\lesssim \nu_{\rm lip} \sum_{E \in \Omega_h} h_{ E}^{-2/p} \Vert \Pi^{0,E}_0 (\sigma_{h,1} - \sigma_{h,2}) \Vert_{L^p(E)} \, h_E^{1 -2/q}\vert \boldsymbol{v}_h \vert_{[W^{1}_q(\Omega_h)]^2} \, \vert \boldsymbol{w}_h \vert_{1,E} \end{aligned} \] where $p$ is such that $1/p + 1/q = 1/2$. Therefore, since $\Pi^{0,E}_0$ is continuous with respect to the $L^p$-norm, from the H\"older inequality for sequences we infer \[ \begin{aligned} \eta_1 &\lesssim \nu_{\rm lip} \sum_{E \in \Omega_h} \Vert \sigma_{h,1} - \sigma_{h,2} \Vert_{L^p(E)} \, \vert \boldsymbol{v}_h \vert_{[W^1_q(E)]^2} \, \vert \boldsymbol{w}_h \vert_{1,E} \\ &\lesssim \nu_{\rm lip} \biggl( \sum_{E \in \Omega_h} \Vert \sigma_{h,1} - \sigma_{h,2} \Vert_{L^p(E)}^p \biggr)^{1/p} \biggl( \sum_{E \in \Omega_h} \vert \boldsymbol{v}_h \vert_{[W^{1}_q(\Omega_h)]^2}^q \biggr)^{1/q} \biggl( \sum_{E \in \Omega_h} \vert \boldsymbol{w}_h \vert_{1,E}^2 \biggr)^{1/2} \\ &\lesssim \nu_{\rm lip} \Vert \sigma_{h,1} - \sigma_{h,2} \Vert_{L^p(\Omega)} \vert \boldsymbol{v}_h \vert_{[W^{1}_q(\Omega_h)]^2} \vert \boldsymbol{w}_h \vert_{1,\Omega} \,. \end{aligned} \] Then, since $p \geq 2$, from the Sobolev embedding \eqref{eq:sobemb} we obtain \begin{equation} \label{eq:lmlip2} \eta_1 \lesssim \nu_{\rm lip} \vert \sigma_{h,1} - \sigma_{h,2} \vert_{1,\Omega} \vert \boldsymbol{v}_h \vert_{[W^{1}_q(\Omega_h)]^2} \vert \boldsymbol{w}_h \vert_{1,\Omega} \,. \end{equation} For what concerns the term $\eta_2$ in \eqref{eq:lmlip1}, using similar argument to that used for $\eta_1$, we have \begin{equation} \label{eq:lmlip3} \begin{aligned} \eta_2 &\leq \sum_{E \in \Omega_h} \int_{E} \bigl \vert \nu\bigl({\Pi^{0, E}_k} \sigma_{h,1} \bigr)- \nu\bigl({\Pi^{0, E}_k} \sigma_{h,2} \bigr) \bigr \vert \, \bigl\vert {\boldsymbol{\Pi}^{0, E}_{k-1}}\boldsymbol{\epsilon} (\boldsymbol{v}_h) \bigr \vert \bigl\vert {\boldsymbol{\Pi}^{0, E}_{k-1}}\boldsymbol{\epsilon} (\boldsymbol{w}_h) \bigr \vert \, {\rm d}E \\ &\lesssim \nu_{\rm lip} \sum_{E \in \Omega_h} \Vert {\Pi^{0, E}_k} (\sigma_{h,1} - \sigma_{h,2}) \Vert_{L^p(E)} \, \vert \boldsymbol{v}_h \vert_{[W^{1}_q(\Omega_h)]^2} \, \vert \boldsymbol{w}_h \vert_{1,E} \\ &\lesssim \nu_{\rm lip} \vert \sigma_{h,1} - \sigma_{h,2} \vert_{1,\Omega} \vert \boldsymbol{v}_h \vert_{[W^{1}_q(\Omega_h)]^2} \vert \boldsymbol{w}_h \vert_{1,\Omega} \,. \end{aligned} \end{equation} The thesis follows combining \eqref{eq:lmlip2} and \eqref{eq:lmlip3} in \eqref{eq:lmlip1}. \end{proof} We now state the following result concerning the uniqueness of the discrete solution. \begin{proposition} \label{prp:uni} Under the data assumptions \cfan{(A0)}, assume moreover that there exist two positive constants $\widehat{C}_{\rm data}$ and $\widehat{C}_{\rm sol}$ depending on the domain $\Omega$, on $k$, on the constants $\kappa_*$, $\kappa^*$ and $\nu_*$ in the data assumptions \cfan{(A0)} and on the shape regularity constant in assumption \cfan{(A1)} s.t. \begin{itemize} \item [i)] the data of the problem satisfies \begin{equation} \label{eq:cdatah} \widehat{C}_{\rm data}^2 \left( \vert \boldsymbol{f}_h \vert_{\ZZ_h*}^2 + \vert g_h \vert_{-1, \Omega}^2 + \Vert \theta_{\rm D, h} \Vert_{1/2, \partial \Omega}^2 \right) < 1 \,, \end{equation} \item [ii)] Problem \eqref{eq:pb vem} admits a solution $(\boldsymbol{u}_h, p_h, \theta^o_h)$ with $\boldsymbol{u}_h \in [W^{1}_q(\Omega_h)]^2$ where $q > 2$ and the following bound holds \begin{equation} \label{eq:csolh} \widehat{C}_{\rm sol} \, \nu_{\rm lip} \, \vert \boldsymbol{u}_h \vert_{[W^{1}_q(\Omega_h)]^2} < 1 \,. \end{equation} \end{itemize} Then, referring to \eqref{eq:bound disc} and \eqref{eq:cclip}, if \begin{equation} \label{eq:uni} K := \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\alpha_* \nu_* \widehat{C}_{\rm data}} \left( 1 + \frac{\widehat{C}_{\rm lip}}{\beta_* \kappa_* \widehat{C}_{\rm sol}} \right) < 1 \end{equation} this solution is unique. \end{proposition} \begin{proof} We limit to sketch the proof since it follows the guidelines of the proof of \cite[Proposition 2.3]{Bernardi:2015}. Let $(\boldsymbol{u}_{h,i}, \theta_{h,i}) \in \ZZ_h \times \W_{{\rm D}, h}$ with $i=1,2$ two solutions of Problem \eqref{eq:pb vem ker} and let \[ \boldsymbol{\delta}_h := \boldsymbol{u}_{h,1} - \boldsymbol{u}_{h,2} \in \ZZ_h \,, \qquad \zeta_h := \theta_{h,1} - \theta_{h,2} \in \W_{0,h} \,. \] Then from \eqref{eq:pb vem ker}, employing the skew-symmetry of the convective forms ${c_{V,h}^{\rm skew}}$ and ${c_{T,h}^{\rm skew}}$, we infer \begin{equation} \label{eq:prpuni1} {a_{T,h}}(\zeta_h, \zeta_h) = - {c_{T,h}^{\rm skew}}(\boldsymbol{\delta}_h; \theta_{h,1}, \zeta_h) \,, \end{equation} and \begin{equation} \label{eq:prpuni2} {a_{V,h}}(\theta_{h,2} , \, \boldsymbol{\delta}_h, \boldsymbol{\delta}_h) = - {c_{V,h}^{\rm skew}}(\boldsymbol{\delta}_h; \, \boldsymbol{u}_{h,1}, \boldsymbol{\delta}_h) -\left({a_{V,h}}(\theta_{h,1}, \, \boldsymbol{u}_{h,1}, \boldsymbol{\delta}_h) - {a_{V,h}}(\theta_{h,2}, \, \boldsymbol{u}_{h,1}, \boldsymbol{\delta}_h) \right) \,. \end{equation} Recalling properties \textbf{(P7)}, from \eqref{eq:prpuni1} we derive \[ \beta_* \kappa_* \vert \zeta_h \vert_{1, \Omega}^2 \leq \widehat{C}_{\rm conv} \vert \boldsymbol{\delta}_h \vert_{1, \Omega} \Vert \theta_{h,1} \Vert_{1, \Omega} \vert \zeta_h \vert_{1, \Omega} \,. \] Thus recalling that $(\boldsymbol{u}_{h,1}, \theta_{h,1})$ satisfies bound \eqref{eq:bound disc}, from \eqref{eq:cdatah} \begin{equation} \label{eq:prpuni3} \vert \zeta_h \vert_{1, \Omega} \leq \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data} \beta_* \kappa_*} \vert \boldsymbol{\delta}_h \vert_{1, \Omega} \,. \end{equation} In a similar way, employing properties \textbf{(P7)} and Lemma \ref{lm:lipshitz}, from \eqref{eq:prpuni2} we infer \[ \alpha_* \nu_* \vert \boldsymbol{\delta}_h \vert_{1, \Omega}^2 \leq \widehat{C}_{\rm conv} \vert \boldsymbol{u}_{h,1} \vert_{1, \Omega} \vert \boldsymbol{\delta}_h \vert_{1, \Omega}^2 + \widehat{C}_{\rm lip} \nu_{\rm lip} \vert \boldsymbol{u}_{h,1} \vert_{[W^1_q(\Omega_h)]^2} \vert \zeta_h \vert_{1, \Omega} \vert \boldsymbol{\delta}_h \vert_{1, \Omega}\,. \] Using again \eqref{eq:bound disc} and \eqref{eq:cdatah}, from \eqref{eq:csolh} and \eqref{eq:prpuni3} we derive \[ \alpha_* \nu_* \vert \boldsymbol{\delta}_h \vert_{1, \Omega}^2 \leq \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data} } \vert \boldsymbol{\delta}_h \vert_{1, \Omega}^2 + \frac{\widehat{C}_{\rm lip}}{\widehat{C}_{\rm sol} } \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data} \beta_* \kappa_*} \vert \boldsymbol{\delta}_h \vert_{1, \Omega} ^2 \,, \] that is \[ \vert \boldsymbol{\delta}_h \vert_{1, \Omega}^2 \leq K \, \vert \boldsymbol{\delta}_h \vert_{1, \Omega}^2 \,. \] Now the proof follows from \eqref{eq:uni}. \end{proof} \section{Convergence results: \textit{a priori} analysis} \label{sec:conv} In present section we derive the rate of convergence for the proposed VEM scheme \eqref{eq:pb vem}. We preliminary recall some optimal approximation properties for the discrete temperatures spaces $\W_{0,h}$, $\W_{{\rm D}, h}$ and the velocities space $\VV_h$ that can be found in \cite{Sutton:2018,BV:2022} and \cite{BLV:2018} respectively. \begin{lemma}[Approximation property of $\W_{0,h}$, $\W_{{\rm D}, h}$] \label{lm:int WDG} Under the assumption \cfan{(A1)} for any $\tau \in \Sigma_0 \cap H^{s+1}(\Omega_h)$ (resp. $\Sigma_{\rm D} \cap H^{s+1}(\Omega_h)$) there exists $\tau_{I} \in \W_{0,h}$ (resp. $\W_{{\rm D}, h}$) such that for all $E \in \Omega_h$ it holds \[ \|\tau - \tau_{I}\|_{0,E} + h_E |\tau- \tau_{I}|_{1, E} \lesssim h_E^{s+1} |\tau|_{s+1,E} \] where $0 < s \leq k$. \end{lemma} \begin{lemma}[Approximation property of $\VV_h$] \label{lm:int VDG} Under the assumption \cfan{(A1)} for any $\boldsymbol{v} \in \VV \cap [H^{s+1}(\Omega_h)]^2$ there exists $\vb_{\cal I} \in \VV_h$ such that for all $E \in \Omega_h$ it holds \[ \|\boldsymbol{v} - \vb_{\cal I}\|_{0,E} + h_E |\boldsymbol{v}- \vb_{\cal I}|_{1, E} \lesssim h_E^{s+1} |\boldsymbol{v}|_{s+1,E} \] where $0 < s \leq k$. Furthermore if $\boldsymbol{v} \in \ZZ$ it holds $\vb_{\cal I} \in \ZZ_h$. \end{lemma} In order to derive the error estimates, we state the following assumption on the solutions and on the data of Problem \eqref{eq:pb primale}. \smallskip\noindent \textbf{(A2) Regularity assumptions:} \begin{itemize} \item the solution $(\boldsymbol{u}, p, \theta)$ satisfy $\boldsymbol{u} \in [H^{s+1}(\Omega_h)]^2$, $p \in [H^{s}(\Omega_h)]$, $\theta \in [H^{s+1}(\Omega_h)]$, \item the viscosity $\nu$, the conductivity $\kappa$ satisfy $\nu(\theta) \in W^{s}_\infty(\Omega_h)$, $\kappa \in W^{s}_\infty(\Omega_h)$, \item the external loads $\boldsymbol{f}$ and $g$ satisfy $\boldsymbol{f} \in [H^{s+1}(\Omega_h)]^2$, $g \in [H^{s+1}(\Omega_h)]$. \end{itemize} for some $0 < s \leq k$. In the following $\mathcal{R}(\cdot)$ will denote a generic constant depending on the norms of $\boldsymbol{u}$, $p$, $\theta$, $\nu$, $\kappa$, $\boldsymbol{f}$ and $g$ in the aforementioned spaces. \begin{proposition} \label{prp:conv} Under the assumptions \cfan{(A0)}, \cfan{(A1)} and under the assumptions of Theorem \ref{thm:uni} and Proposition \ref{prp:uni}, let $(\boldsymbol{u}, p, \theta)$ be the solution of \eqref{eq:pb variazionale} and $(\boldsymbol{u}_h, p_h, \theta_h)$ be the solution of \eqref{eq:pb vem}. Assuming moreover \cfan{(A2)} the following error estimates hold: \[ \begin{gathered} \vert \boldsymbol{u}_h - \boldsymbol{u}\vert_{1, \Omega} + \Vert \theta - \theta_h \Vert_{1, \Omega} \leq \mathcal{R}(\boldsymbol{u}, \theta, \nu, \kappa) \, h^s \,, \\ \Vert p - p_h\Vert_{0, \Omega} \leq \mathcal{R}(\boldsymbol{u}, p, \theta, \nu, \kappa) \, h^s \,. \end{gathered} \] \end{proposition} \begin{proof} We preliminary observe that under the assumptions of Theorem \ref{thm:uni} (resp. Proposition \ref{prp:uni}) and recalling \eqref{eq:bound cont} (resp. \eqref{eq:bound disc}) the following hold \begin{equation} \label{eq:bounds} \vert \boldsymbol{u} \vert_{1, \Omega}^2 + \Vert \theta \Vert_{1, \Omega}^2 \leq \frac{{C}_{\rm est}^2}{{C}_{\rm data}^2}\,, \qquad \qquad \vert \boldsymbol{u}_h \vert_{1, \Omega}^2 + \Vert \theta_h \Vert_{1, \Omega}^2 \leq \frac{\widehat{C}_{\rm est}^2}{\widehat{C}_{\rm data}^2} \,. \end{equation} The proof follows the following steps. \noindent \textbf{Step 1. Interpolation error estimates.} \noindent Let us introduce the following error quantities \[ \begin{aligned} \ee_{\cal I} &:= \boldsymbol{u} - \uu_{\cal I}\,, \qquad & \sigma_{\cal I} &:= \theta - \theta_{I}\,, \qquad & \rho_{I} &:= p - p_{I}\,, \\ \boldsymbol{e}_h &:= \uu_{\cal I} - \boldsymbol{u}_h\,, \qquad & \sigma_h &:= \theta_{I} - \theta_h\,, \qquad & \rho_h &:= p_{I} - p_h\,, \end{aligned} \] where $\theta_{I}$ and $\uu_{\cal I}$ are the interpolant function of $\theta$ and $\boldsymbol{u}$ of Lemma \ref{lm:int WDG} and Lemma \ref{lm:int VDG} respectively and $p_{I} \in Q_h$ is defined elementwise by $p_{I}|_E = \Pi^{0,E}_{k-1} p$ for any $E \in \Omega_h$ (cf. \eqref{eq:P0_k^E}). Therefore from Lemma \ref{lm:int VDG}, Lemma \ref{lm:int WDG} and Bramble-Hilbert Lemma \eqref{eq:bramble}, we have \begin{equation} \label{eq:int-errors} \vert \ee_{\cal I}\vert_{1 ,\Omega} \leq \mathcal{R}(\boldsymbol{u}) \, h^s \,, \qquad \Vert \sigma_{\cal I}\Vert_{1 ,\Omega} \leq \mathcal{R}(\theta) \, h^s \,, \qquad \Vert \rho_{I}\Vert_{1 ,\Omega} \leq \mathcal{R}(p) \, h^s \,. \end{equation} Notice now that since $\boldsymbol{u} \in \ZZ$, from Lemma \ref{lm:int VDG} it holds that $\boldsymbol{e}_h \in \ZZ_h$. Whereas from Lemma \ref{lm:int WDG}, $\sigma_h \in \W_{0,h}$. \noindent \textbf{Step 2. Error equation for the temperature.} \noindent Employing Property \textbf{(P7)} and manipulating the second equation in \eqref{eq:pb variazionale ker} and \eqref{eq:pb vem ker} we obtain \begin{equation} \label{eq:conv1} \begin{aligned} \beta_* \kappa_* \vert \sigma_h \vert_{1, \Omega}^2 &\leq {a_{T,h}}(\sigma_h, \sigma_h) = {a_{T,h}}(\theta_{I}, \sigma_h) - {a_{T,h}}(\theta_h, \sigma_h) \\ & = \left( {a_{T,h}}(\theta_{I}, \sigma_h) - {a_T}(\theta, \sigma_h) \right) + \left( {a_T}(\theta, \sigma_h) - {a_{T,h}}(\theta_h, \sigma_h) \right) \\ & = \left( {a_{T,h}}(\theta_{I}, \sigma_h) - {a_T}(\theta, \sigma_h) \right) + \left( {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \, \theta_h, \sigma_h) - c_T^{\rm skew}(\boldsymbol{u}; \, \theta, \sigma_h) \right) + (g - g_h, \sigma_h) \\ & =: \eta_{a, T} + \eta_{c, T} + \eta_g \,. \end{aligned} \end{equation} We now estimate the three terms above. \noindent $\bullet$ estimate of $\eta_{a, T}$: we split $\eta_{a, T}$ into local contributions $\eta_{a, T}^E$. Recalling definitions \eqref{eq:forma as} and \eqref{eq:forma ashl} and the property of the $L^2$-projection we have \[ \begin{aligned} \eta_{a, T}^E &:= \int_{E} \kappa \, {\Pi^{0, E}_{k-1}}\nabla \theta_{I} \cdot {\Pi^{0, E}_{k-1}} \nabla \sigma_h \, {\rm d}E + {s_{T}^E}(\theta_{I} , \, \sigma_h) - \int_{E} \kappa \, \nabla \theta \cdot \nabla \sigma_h \, {\rm d}E \\ & = \int_{E} \kappa \, ({\Pi^{0, E}_{k-1}}\nabla \theta_{I} - \nabla \theta)\cdot {\Pi^{0, E}_{k-1}} \nabla \sigma_h \, {\rm d}E - \int_{E} \kappa \, \nabla \theta \cdot (I - {\Pi^{0, E}_{k-1}})\nabla \sigma_h \, {\rm d}E + {s_{T}^E}(\theta_{I} , \, \sigma_h) \\ & = \int_{E} \kappa \, ({\Pi^{0, E}_{k-1}}\nabla \theta_{I} - \nabla \theta)\cdot {\Pi^{0, E}_{k-1}} \nabla \sigma_h \, {\rm d}E - \int_{E} (I - {\Pi^{0, E}_{k-1}}) (\kappa \, \nabla \theta) \cdot \nabla \sigma_h \, {\rm d}E + {s_{T}^E}(\theta_{I} , \, \sigma_h) \,. \end{aligned} \] Then from Property \textbf{(P0)}, \eqref{eq:ssstab} and \eqref{eq:stab}, employing Bramble-Hilbert Lemma \eqref{eq:bramble}, we obtain \[ \begin{aligned} \eta_{a, T}^E & \leq \left( \kappa^* \Vert {\Pi^{0, E}_{k-1}}\nabla \theta_{I} - \nabla \theta\Vert_{0,E} + \Vert (I - {\Pi^{0, E}_{k-1}}) (\kappa \, \nabla \theta) \Vert_{0,E} + \kappa^* \vert (I - {\Pi^{0, E}_k}) \theta_{I}\vert_{1, E} \right) \vert \sigma_h \vert_{1,E} \\ & \leq \kappa^* \left( \vert \sigma_{\cal I}\vert_{1,E} + \vert (I - {\Pi^{0, E}_k})\theta\vert_{1,E} + \frac{1}{\kappa^*} \Vert (I - {\Pi^{0, E}_{k-1}}) (\kappa \, \nabla \theta) \Vert_{0,E} + \right) \vert \sigma_h \vert_{1,E} \\ & \lesssim \kappa^* h_E^{s} \left( \vert \sigma_{\cal I}\vert_{1,E} + \vert \theta\vert_{s+1,E} + \frac{\Vert \kappa\Vert_{W^s_\infty}(E)}{\kappa^*} \Vert \theta\Vert_{s+1,E} \right) \vert \sigma_h \vert_{1,E} \,. \end{aligned} \] Therefore summing the local contributions and employing \eqref{eq:int-errors} we obtain \begin{equation} \label{eq:conv2} \eta_{a,T} \leq \mathcal{R}(\theta, \kappa) \, h^{s} \, \vert \sigma_h \vert_{1,\Omega} \,. \end{equation} \noindent $\bullet$ estimate of $\eta_{c, T}$: employing the skew-symmetry of ${c_{T,h}^{\rm skew}}(\cdot; \cdot, \cdot)$ (cf. \eqref{eq:forma cshlskew}) we have \begin{equation} \label{eq:conv3a} \begin{aligned} \eta_{c,T} &= {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \, \theta_h, \sigma_h) - {c_{T,h}^{\rm skew}}(\boldsymbol{u}; \, \theta, \sigma_h) + {c_{T,h}^{\rm skew}}(\boldsymbol{u}; \, \theta, \sigma_h) - c_T^{\rm skew}(\boldsymbol{u}; \, \theta, \sigma_h) \\ &= - {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \, \sigma_{\cal I}, \sigma_h) - {c_{T,h}^{\rm skew}}(\boldsymbol{u} - \boldsymbol{u}_h; \, \theta, \sigma_h) + {c_{T,h}^{\rm skew}}(\boldsymbol{u}; \, \theta, \sigma_h) - c_T^{\rm skew}(\boldsymbol{u}; \, \theta, \sigma_h) \,. \end{aligned} \end{equation} From Property \textbf{(P7)}, bounds \eqref{eq:bounds} and \eqref{eq:int-errors} we have \begin{equation} \label{eq:conv3b} \begin{aligned} {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h; \, \sigma_{\cal I}, \sigma_h) &+ {c_{T,h}^{\rm skew}}(\boldsymbol{u} - \boldsymbol{u}_h; \, \theta, \sigma_h) \leq \widehat{C}_{\rm conv} \left( \vert \boldsymbol{u}_h\vert_{1, \Omega} \Vert \sigma_{\cal I}\Vert_{1, \Omega} + \vert \boldsymbol{u} - \boldsymbol{u}_h\vert_{1, \Omega} \Vert \theta\Vert_{1, \Omega} \right) \vert \sigma_h\vert_{1, \Omega} \\ & \leq \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data}} \Vert \sigma_{\cal I}\Vert_{1, \Omega} \vert \sigma_h\vert_{1, \Omega} + \frac{\widehat{C}_{\rm conv} {C}_{\rm est}}{{C}_{\rm data}} \left(\vert \ee_{\cal I}\vert_{1, \Omega} + \vert \boldsymbol{e}_h\vert_{1, \Omega} \right) \vert \sigma_h\vert_{1, \Omega} \\ & \leq \frac{\widehat{C}_{\rm conv} {C}_{\rm est}}{{C}_{\rm data}} \vert\boldsymbol{e}_h\vert_{1, \Omega} \vert \sigma_h\vert_{1, \Omega} + \mathcal{R}(\boldsymbol{u}, \theta)\, h^s \, \vert \sigma_h\vert_{1, \Omega} \,. \end{aligned} \end{equation} Furthermore employing \cite[Lemma 4.3]{BLV:2018} \begin{equation} \label{eq:conv3c} {c_{T,h}^{\rm skew}}(\boldsymbol{u}; \, \theta, \sigma_h) - c_T^{\rm skew}(\boldsymbol{u}; \, \theta, \sigma_h) \leq \mathcal{R}(\boldsymbol{u}, \theta) \, h^{s} \, \vert \sigma_h \vert_{1,\Omega} \,. \end{equation} Therefore collecting in \eqref{eq:conv3a}, bounds \eqref{eq:conv3b} and \eqref{eq:conv3c} we obtain \begin{equation} \label{eq:conv3} \eta_{c,T} \leq \frac{\widehat{C}_{\rm conv} {C}_{\rm est}}{{C}_{\rm data}} \vert\boldsymbol{e}_h\vert_{1, \Omega} \vert \sigma_h\vert_{1, \Omega} + \mathcal{R}(\boldsymbol{u}, \theta) \, h^{s} \, \vert \sigma_h \vert_{1,\Omega} \,. \end{equation} \noindent $\bullet$ estimate of $\eta_{g}$: employing the definition of $L^2$-projection \eqref{eq:P0_k^E} and Bramble-Hilbert Lemma \eqref{eq:bramble} we infer \begin{equation} \label{eq:conv4} \eta_g \leq \mathcal{R}(g) \, h^{s +2} \, \vert \sigma_h \vert_{1,\Omega} \,. \end{equation} Finally combining \eqref{eq:conv2}, \eqref{eq:conv3} and \eqref{eq:conv4} in \eqref{eq:conv1} we get \begin{equation} \label{eq:convtheta} \vert {\sigma_h} \vert_{1, \Omega} \leq \frac{\widehat{C}_{\rm conv} {C}_{\rm est}}{{C}_{\rm data} \beta_* \kappa_*} \vert\boldsymbol{e}_h\vert_{1, \Omega} + \mathcal{R}(\boldsymbol{u}, \theta, \kappa) \, h^{s} + \mathcal{R}(g) \, h^{s +2} \,. \end{equation} \noindent \textbf{Step 3. Error equation for the velocity.} \noindent We now analyse the velocity error equation. Combining the first equation in \eqref{eq:pb variazionale ker} and \eqref{eq:pb vem ker} together with Property \textbf{(P7)}, we have \begin{equation} \label{eq:conv5} \begin{aligned} &\alpha_* \nu_* \vert \boldsymbol{e}_h \vert_{1, \Omega}^2 \leq {a_{V,h}}(\theta_h; \, \boldsymbol{e}_h, \boldsymbol{e}_h) = {a_{V,h}}(\theta_h; \, \uu_{\cal I}, \boldsymbol{e}_h) - {a_{V,h}}(\theta_h; \, \boldsymbol{u}_h, \boldsymbol{e}_h) \\ & = \left( {a_{V,h}}(\theta_h; \, \uu_{\cal I}, \boldsymbol{e}_h) - {a_V}(\theta; \, \boldsymbol{u}, \boldsymbol{e}_h) \right) + \left( {a_V}(\theta; \, \boldsymbol{u}, \boldsymbol{e}_h) - {a_{V,h}}(\theta_h; \, \boldsymbol{u}_h, \boldsymbol{e}_h) \right) \\ & = \left( {a_{V,h}}(\theta_h; \, \uu_{\cal I}, \boldsymbol{e}_h) - {a_V}(\theta; \, \boldsymbol{u}, \boldsymbol{e}_h) \right) + \left( {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h; \, \boldsymbol{u}_h, \boldsymbol{e}_h) - c_V^{\rm skew}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{e}_h) \right) + (\boldsymbol{f} - \boldsymbol{f}_h, \boldsymbol{e}_h) \\ & =: \eta_{a, V} + \eta_{c, V} + \eta_{\boldsymbol{f}} \,. \end{aligned} \end{equation} We now estimate each term above. \noindent $\bullet$ estimate of $\eta_{a, V}$: we split the term into local contributions $\eta_{a, V}^E $. Recalling definitions \eqref{eq:forma av} and \eqref{eq:forma avhl} we have \begin{equation} \label{eq:conv6a} \begin{aligned} \eta_{a, V}^E &:= \int_{E} \nu({\Pi^{0, E}_k} \theta_h) \, {\boldsymbol{\Pi}^{0, E}_{k-1}}\nabla \uu_{\cal I} \cdot {\boldsymbol{\Pi}^{0, E}_{k-1}} \nabla \boldsymbol{e}_h \, {\rm d}E + {s_{V}^E}(\theta_h; \, \uu_{\cal I} , \, \boldsymbol{e}_h) - \int_{E} \nu(\theta) \, \nabla \boldsymbol{u} \cdot \nabla \boldsymbol{e}_h \, {\rm d}E \\ & = \int_{E} \nu({\Pi^{0, E}_k} \theta_h) \, ({\boldsymbol{\Pi}^{0, E}_{k-1}}\nabla \uu_{\cal I} - \nabla \boldsymbol{u}) \cdot {\boldsymbol{\Pi}^{0, E}_{k-1}} \nabla \boldsymbol{e}_h \, {\rm d}E + {s_{V}^E}(\theta_h; \, \uu_{\cal I} , \, \boldsymbol{e}_h) + \\ & \quad - \int_{E} \nu(\theta) \, \nabla \boldsymbol{u} \cdot (I - {\boldsymbol{\Pi}^{0, E}_{k-1}})\nabla \boldsymbol{e}_h \, {\rm d}E + \int_{E} (\nu({\Pi^{0, E}_k} \theta_h) - \nu(\theta)) \, \nabla \boldsymbol{u} \cdot {\boldsymbol{\Pi}^{0, E}_{k-1}} \nabla \boldsymbol{e}_h \, {\rm d}E \\ & =: \zeta_1^E + \zeta_2^E + \zeta_3^E + \zeta_4^E \,. \end{aligned} \end{equation} Recalling Property \textbf{(P0)}, the stability bounds \eqref{eq:svstab} and \eqref{eq:stab}, and the definition of $L^2$-projection \eqref{eq:P0_k^E}, the first three sub-terms can be bounded as follows \[ \begin{aligned} \sum_{i=1}^3\zeta_i^E &\leq \bigl( \nu^* \Vert {\boldsymbol{\Pi}^{0, E}_{k-1}}\nabla \uu_{\cal I} - \nabla \boldsymbol{u}\Vert_{0,E} + \Vert (I - {\boldsymbol{\Pi}^{0, E}_{k-1}})(\nu(\theta)\nabla \boldsymbol{u})\Vert_{0,E} + \nu^* \vert ( I - {\Pi^{0, E}_k}) \uu_{\cal I}\vert_{1,E} \bigr) \vert \boldsymbol{e}_h \vert_{1,E} \\ & \leq \nu^* \left( \vert \ee_{\cal I}\vert_{1,E} + {\vert (I - {\Pi^{0, E}_k}) \boldsymbol{u}\vert_{1,E}} + \frac{1}{\nu^*} \Vert (I - {\Pi^{0, E}_k}) (\nu(\theta) \, \nabla \boldsymbol{u}) \Vert_{0,E} \right) \vert \boldsymbol{e}_h \vert_{1,E} \\ & \lesssim \nu^* h_E^{s} \left( \vert \ee_{\cal I}\vert_{1,E} + \vert \boldsymbol{u}\vert_{s+1,E} + \frac{\Vert \nu(\theta)\Vert_{W^s_\infty}(E)}{\nu^*} \Vert \boldsymbol{u}\Vert_{s+1,E} \right) \vert \boldsymbol{e}_h \vert_{1,E} \,. \end{aligned} \] Therefore \begin{equation} \label{eq:conv6b} \begin{aligned} \sum_{E \in \Omega_h}\sum_{i=1}^3\zeta_i^E & \leq \mathcal{R}(\boldsymbol{u}, \nu) \, h^s \, \vert \boldsymbol{e}_h \vert_{1, \Omega} \,. \end{aligned} \end{equation} For the term $\zeta_4$, employing similar computations to that in \eqref{eq:lmlip3} (recalling that $p$ is such that $1/p + 1/q =1/2)$ we infer \[ \begin{aligned} \zeta_4^E &\leq \nu_{\rm lip} \Vert {\Pi^{0, E}_k} \theta_h - \theta\Vert_{L^p(E)} \vert \boldsymbol{u} \vert_{[W^1_q(E)]^2} \vert\boldsymbol{e}_h \vert_{1,E} \\ & \leq \nu_{\rm lip} \left( \Vert {\Pi^{0, E}_k} (\theta_h - \theta)\Vert_{L^p(E)} + \Vert (I - {\Pi^{0, E}_k}) \theta\Vert_{L^p(E)} \right) \vert \boldsymbol{u} \vert_{[W^1_q(E)]^2} \vert\boldsymbol{e}_h \vert_{1,E} \\ & \lesssim \nu_{\rm lip} \left( \Vert \theta_h - \theta\Vert_{L^p(E)} + \Vert \theta\Vert_{W^s_p(E)} \, h_E^s \right) \vert \boldsymbol{u} \vert_{[W^1_q(E)]^2} \vert\boldsymbol{e}_h \vert_{1,E} \\ & \lesssim \nu_{\rm lip} \left( \Vert \sigma_h\Vert_{L^p(E)} +\Vert \sigma_{\cal I}\Vert_{L^p(E)} + \Vert \theta\Vert_{W^s_p(E)} \, h_E^s \right) \vert \boldsymbol{u} \vert_{[W^1_q(E)]^2} \vert\boldsymbol{e}_h \vert_{1,E} \end{aligned} \] where we also used the continuity of the $L^2$-projection w.r.t. the $L^p$-norm and Bramble-Hilbert \eqref{eq:bramble}. From the H\"older inequality for sequences and the Sobolev embedding \eqref{eq:sobemb}, we infer \[ \sum_{E \in \Omega_h} \zeta_4^E \lesssim \nu_{\rm lip} \left( \vert \sigma_h\vert_{1, \Omega} +\Vert \sigma_{\cal I}\Vert_{1,\Omega} + \mathcal{R}(\theta) h^s \right) \vert \boldsymbol{u} \vert_{[W^1_q(\Omega_h)]^2} \vert\boldsymbol{e}_h \vert_{1,\Omega} \,. \] Therefore from \eqref{eq:int-errors} and \eqref{eq:csol} \begin{equation} \label{eq:conv6c} \begin{aligned} \sum_{E \in \Omega_h} \zeta_4^E & \leq \frac{C_{\rm lip}}{{C}_{\rm sol}} \vert \sigma_h\vert_{1, \Omega} \vert\boldsymbol{e}_h \vert_{1,\Omega} + \mathcal{R}(\theta) \, h^s \, \vert\boldsymbol{e}_h \vert_{1,\Omega} \end{aligned} \end{equation} for a suitable positive constant $C_{\rm lip}$. Collecting \eqref{eq:conv6a} and \eqref{eq:conv6b} in \eqref{eq:conv6c} we obtain \begin{equation} \label{eq:conv6} \begin{aligned} \eta_{a,V} & \leq \frac{C_{\rm lip}}{{C}_{\rm sol}} \vert \sigma_h \vert_{1, \Omega} \vert \boldsymbol{e}_h\vert_{1, \Omega} + \mathcal{R}(\boldsymbol{u}, \theta, \nu) \, h^s \, \vert \boldsymbol{e}_h\vert_{1, \Omega} \,. \end{aligned} \end{equation} \noindent $\bullet$ estimate of $\eta_{c,V}$: direct application of \cite[Lemma 4.3 and Lemma 4.4]{BLV:2018} yields \begin{equation} \label{eq:conv7} \eta_{c,V} \leq \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data}} \vert\boldsymbol{e}_h\vert_{1, \Omega}^2 + \mathcal{R}(\boldsymbol{u}) \, h^{s} \, \vert \boldsymbol{e}_h \vert_{1,\Omega} \,. \end{equation} \noindent $\bullet$ estimate of $\eta_{\boldsymbol{f}}$: using the same argument used above for $\eta_g$ we obtain \begin{equation} \label{eq:conv8} \eta_{\boldsymbol{f}} \leq \mathcal{R}(\boldsymbol{f}) \, h^{s +2} \, \vert \boldsymbol{e}_h \vert_{1,\Omega} \,. \end{equation} Collecting \eqref{eq:conv6}, \eqref{eq:conv7} and \eqref{eq:conv8} in the error equation \eqref{eq:conv5} we obtain \begin{equation} \label{eq:convu1} \alpha_* \nu_* \vert \ee_{\cal I}\vert_{1, \Omega} \leq \mathcal{R}(\boldsymbol{u}, \theta, \nu) h^s + \mathcal{R}(\boldsymbol{f}) h^{s+2} + \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data}} \vert\boldsymbol{e}_h\vert_{1, \Omega} + \frac{C_{\rm lip}}{{C}_{\rm sol}} \vert \sigma_h\vert_{1, \Omega} \,. \end{equation} \noindent \textbf{Step 4. Error estimates for the velocity and the temperature} \noindent Now employing the estimate \eqref{eq:convtheta} in \eqref{eq:convu1} we infer \[ \left(\alpha_* \nu_* - \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data}} - \frac{\widehat{C}_{\rm conv} {C}_{\rm est}}{{C}_{\rm data}} \frac{C_{\rm lip}}{{C}_{\rm sol} \beta_* \kappa_*} \right) \vert \ee_{\cal I}\vert_{1, \Omega} \leq \mathcal{R}(\boldsymbol{u}, \theta, \nu, \kappa) h^s + \mathcal{R}(\boldsymbol{f}, g) h^{s+2} \,. \] Therefore assuming that $\nu_*$ and $\kappa_*$ large enough \begin{equation} \label{eq:convu2} \vert \ee_{\cal I}\vert_{1, \Omega} \leq \mathcal{R}(\boldsymbol{u}, \theta, \nu, \kappa) h^s + \mathcal{R}(\boldsymbol{f}, g) h^{s+2} \,. \end{equation} We finally collect the estimates \eqref{eq:convu2}, \eqref{eq:convtheta} and the error bounds \eqref{eq:int-errors} obtaining \begin{equation} \label{eq:conv-theta-u} \vert \boldsymbol{u} - \boldsymbol{u}_h \vert_{1, \Omega} + \Vert \theta - \theta_h \Vert_{1, \Omega} \leq \mathcal{F}(\boldsymbol{u}, \theta, \nu, \kappa) h^s + \mathcal{F}(\boldsymbol{f}, g) h^{s+2} \,. \end{equation} \noindent \textbf{Step 5. Error estimate for the pressure.} \noindent In the last step we briefly study the convergence error for the pressures. From the first equation in \eqref{eq:pb variazionale} and \eqref{eq:pb vem}, simple computations yield \begin{equation} \label{eq:conv10} \begin{aligned} b(\boldsymbol{v}_h, \rho_h) &= b(\boldsymbol{v}_h, p_{I}) - b(\boldsymbol{v}_h, p_h) = b(\boldsymbol{v}_h, p) - b(\boldsymbol{v}_h, p_h) - b(\boldsymbol{v}_h, \rho_{I}) \\ & = \left( {a_{V,h}}(\theta_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - {a_V}(\theta; \, \boldsymbol{u}, \boldsymbol{v}_h) \right) + \left( {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - c_V^{\rm skew}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}_h) \right) + \\ & \qquad + (\boldsymbol{f}- \boldsymbol{f}_h, \boldsymbol{v}_h) - b(\boldsymbol{v}_h, \rho_{I}) \\ & := \gamma_{a, V} + \gamma_{c, V} + \gamma_{\boldsymbol{f}} + \gamma_b \,. \end{aligned} \end{equation} \noindent $\bullet$ estimate of $\gamma_{a, V}$: the estimate of $\gamma_{a, V}$ follows the same steps for the bound of $\eta_{a, V}$ \[ \begin{aligned} \gamma_{a, V} &\leq \nu^* \vert \boldsymbol{u} - \boldsymbol{u}_h\vert_{1, \Omega} \vert \boldsymbol{v}_h\vert_{1, \Omega} + \frac{C_{\rm lip}}{{C}_{\rm sol}} \Vert \theta - \theta_h \Vert_{1, \Omega} \vert \boldsymbol{v}_h\vert_{1, \Omega} + \mathcal{R}(\boldsymbol{u}, \theta, \nu) \, h^s \, \vert \boldsymbol{v}_h\vert_{1, \Omega} \end{aligned} \] therefore from \eqref{eq:conv-theta-u} we have \begin{equation} \label{eq:conv11} \begin{aligned} \gamma_{a, V} &\leq \left(\mathcal{R}(\boldsymbol{u}, \theta, \nu, \kappa) \, h^s + \mathcal{R}(\boldsymbol{f}, g) \, h^{s+2}\right) \vert \boldsymbol{v}_h\vert_{1, \Omega} \,. \end{aligned} \end{equation} \noindent $\bullet$ estimate of $\gamma_{c, V}$: simple computations yield \begin{equation} \label{eq:conv12a} \begin{aligned} \gamma_{c, V} &= \left({c_{V,h}^{\rm skew}}(\boldsymbol{u}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - {c_{V,h}^{\rm skew}}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}_h) \right) + \left({c_{V,h}^{\rm skew}}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}_h) - c_V^{\rm skew}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}_h)\right) \\ &= \left({c_{V,h}^{\rm skew}}(\boldsymbol{u}_h - \boldsymbol{u}; \, \boldsymbol{u}_h, \boldsymbol{v}_h) + {c_{V,h}^{\rm skew}}(\boldsymbol{u}; \, \boldsymbol{u}_h - \boldsymbol{u}, \boldsymbol{v}_h) \right) + \left({c_{V,h}^{\rm skew}}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}_h) - c_V^{\rm skew}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}_h)\right) \,. \end{aligned} \end{equation} Employing Property \textbf{(P7)} and bounds \eqref{eq:bounds} and estimate \eqref{eq:conv-theta-u}, we have \begin{equation} \label{eq:conv12b} \begin{aligned} {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h - \boldsymbol{u}; \, \boldsymbol{u}_h, \boldsymbol{v}_h) + {c_{V,h}^{\rm skew}}(\boldsymbol{u}; \, \boldsymbol{u}_h - \boldsymbol{u}, \boldsymbol{v}_h) & \leq \widehat{C}_{\rm conv} \left( \vert \boldsymbol{u} \vert_{1, \Omega} + \vert \boldsymbol{u}_h \vert_{1, \Omega} \right) \vert \boldsymbol{u} - \boldsymbol{u}_h \vert_{1, \Omega} \, \vert \boldsymbol{v}_h\vert_{1, \Omega} \\ & \leq \widehat{C}_{\rm conv} \left( \frac{{C}_{\rm est}}{{C}_{\rm data}} + \frac{\widehat{C}_{\rm est}}{\widehat{C}_{\rm data}} \right) \vert \boldsymbol{u} - \boldsymbol{u}_h \vert_{1, \Omega} \, \vert \boldsymbol{v}_h\vert_{1, \Omega} \\ & \leq \left(\mathcal{R}(\boldsymbol{u}, \theta, \nu, \kappa) \, h^s + \mathcal{R}(\boldsymbol{f}, g) \, h^{s+2}\right) \vert \boldsymbol{v}_h\vert_{1, \Omega} \,. \end{aligned} \end{equation} Whereas applying \cite[Lemma 4.3]{BLV:2018} we have \begin{equation} \label{eq:conv12c} \begin{aligned} {c_{V,h}^{\rm skew}}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}_h) - c_V^{\rm skew}(\boldsymbol{u}; \, \boldsymbol{u}, \boldsymbol{v}_h) & \leq \mathcal{R}(\boldsymbol{u}) \, h^s \, \vert \boldsymbol{v}_h\vert_{1, \Omega} \,. \end{aligned} \end{equation} Therefore \eqref{eq:conv12a}, \eqref{eq:conv12b} and \eqref{eq:conv12c} implies \begin{equation} \label{eq:conv12} \begin{aligned} \gamma_{c, V} \leq \left(\mathcal{R}(\boldsymbol{u}, \theta, \nu, \kappa) \, h^s + \mathcal{R}(\boldsymbol{f}, g) \, h^{s+2}\right) \vert \boldsymbol{v}_h\vert_{1, \Omega} \,. \end{aligned} \end{equation} \noindent $\bullet$ estimate of $\gamma_{\boldsymbol{f}} + \gamma_b$: the term $\gamma_{\boldsymbol{f}}$ can be bounded as $\eta_{\boldsymbol{f}}$. The term $\gamma_b$ can be bounded, recalling Property \textbf{(P0)}, by Bramble-Hilbert Lemma \ref{eq:bramble}. Then we have \begin{equation} \label{eq:conv13} \gamma_{\boldsymbol{f}} + \gamma_b \leq \left(\mathcal{R}(\boldsymbol{f}) \, h^{s+2} + \mathcal{R}(p) \, h^s \right) \vert \boldsymbol{v}_h\vert_{1, \Omega} \,. \end{equation} Employing the inf-sup stability \eqref{eq:infsup}, from \eqref{eq:conv11}--\eqref{eq:conv13}, we obtain \[ \widehat{\beta} \|\rho_h\|_{0,E} \leq \sup_{\boldsymbol{v}_h \in \VV_h} \frac{b(\boldsymbol{v}_h, \rho_h)}{\vert \boldsymbol{v}_h\vert_{1, \Omega}} \leq \mathcal{F}(\boldsymbol{u}, \theta, p, \nu, \kappa) \, h^s + \mathcal{F}(\boldsymbol{f}, g) \, h^{s+2} \,. \] The pressure error estimates follows from the previous bound and \eqref{eq:int-errors}: \[ \|p - p_h\|_{0,E} \leq \mathcal{F}(\boldsymbol{u}, \theta, p, \nu, \kappa) \, h^s + \mathcal{F}(\boldsymbol{f}, g) \, h^{s+2} \,. \] \end{proof} \begin{remark} \label{rmk:conv} Note that the proposed VEM scheme \eqref{eq:pb vem ker} has the following favorable property that extends to the context of coupled problems the convergence result obtained for the Navier--Stokes equation: the error components partly decouple. In fact the velocity and the temperature errors do not depend directly on the discrete pressures, but only indirectly through the approximation of the loading and convection terms and such dependence on the full load is much weaker with respect to standard mixed schemes. In some situations the partial decoupling of the errors induces a positive effect on the velocity/temperature approximation (see for instance Test 2). \end{remark} \section{Numerical results} \label{sec:num} In this section we present two numerical experiments to test the practical performance of the proposed VEM scheme \eqref{eq:pb vem} and the possible advantages related to the divergence-free velocity solutions. In the tests we consider $k=2$. \subsection{Fixed point iteration} We describe the linearization strategy based on a fixed-point iteration adopted to solve the non-linear coupled problem. \vspace{0.2cm} \noindent \texttt{LINEAR FIXED POINT ITERATION} \noindent Starting from $(\boldsymbol{u}^0_h, \, p^0_h) =(\boldsymbol{0}, 0)$, for $n \geq 0$, until convergence solve \begin{itemize} \item \texttt{HEAT equation} \begin{equation*} \label{eq:heat} \left \{ \begin{aligned} & \text{find $\theta_h^{n+1}\in \W_{{\rm D}, h}$, such that} \\ &{a_{T,h}}(\theta_h^{n+1}, \sigma_h) + {c_{T,h}^{\rm skew}}(\boldsymbol{u}_h^{n}; \, \theta_h^{n+1}, \sigma_h) = (g_h, \sigma_h) \qquad & &\text{$\forall \sigma_h \in \W_{0,h}$,} \end{aligned} \right. \end{equation*} \item \texttt{OSEEN equation} \begin{equation*} \label{eq:oseen} \left\{ \begin{aligned} &\text{find $(\boldsymbol{u}_h^{n+1}, \, p_h^{n+1}) \in \boldsymbol{V}_h \times Q_h$, such that} \\ &\begin{aligned} {a_{V,h}}(\theta_h^{n+1}; \, \boldsymbol{u}_h^{n+1}, \boldsymbol{v}_h) + {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h^{n}; \, \boldsymbol{u}_h^{n+1}, \boldsymbol{v}_h) + b(\boldsymbol{v}_h, p_h^{n+1}) &= (\boldsymbol{f}_h, \boldsymbol{v}_h) \quad & &\text{$\forall \boldsymbol{v}_h \in \VV_h$,} \\ b(\boldsymbol{u}_h^{n+1}, q_h) &= 0 \quad & & \text{$\forall q_h \in Q_h$.} \end{aligned} \end{aligned} \right. \end{equation*} \end{itemize} Notice that at each step of the fixed-point iteration the solution $\boldsymbol{u}_h^n$ is still divergence-free, therefore the linearization procedure does non affect the divergence-free property of the final discrete solution. At each iteration we need to solve two linear systems with dimension \[ \begin{aligned} \texttt{N\_DoFs HEAT} &= N_V + (k-1)N_E + \frac{k(k-1)}{2} N_P \,, \\ \texttt{N\_DoFs OSEEN} &= 2\, N_V + 2\,(k-1)N_E + \left(\frac{k(3k-1)}{2} \right) N_P + 1 \,, \end{aligned} \] where $N_V$, $N_E$, $N_P$ denotes the number of internal vertices, internal edges and polygons in $\Omega_h$ respectively. The dimension of the linear system for the \texttt{OSEEN equation} can be significantly reduced considering the so-called reduced spaces (see \cite{BLV:2018}). We now derive the convergence of the fixed-point iteration. \begin{proposition} \label{prp:fixed point} Under the assumptions \cfan{(A0)}, \cfan{(A1)} and under the assumptions of Proposition \ref{prp:uni}, let $(\boldsymbol{u}_h, p_h, \theta_h)$ be the solution of \eqref{eq:pb vem}. Then the sequence $\{(\boldsymbol{u}_h^n, p_h^n, \theta_h^n)\}_n$ generated by the \cfun{LINEAR FIXED POINT ITERATION} satisfies \[ \lim_{n \to \infty} \vert \boldsymbol{u}_h - \boldsymbol{u}_h^n\vert_{1, \Omega} = 0 \qquad \lim_{n \to \infty} \vert \theta_h - \theta_h^n\vert_{1, \Omega} = 0 \qquad \lim_{n \to \infty} \Vert p_h - p_h^n\Vert_{0, \Omega} = 0 \,. \] \end{proposition} \begin{proof} We only sketch the proof since it can be derived using similar technique to that in the proof of Proposition \ref{prp:uni}. For any $n \geq 0$, let us introduce the error quantities \[ \boldsymbol{\delta}_h^n := \boldsymbol{u}_h^n - \boldsymbol{u}_h\,, \qquad \sigma_h^n := \theta_h^n - \theta_h\,, \qquad \rho_h^n := p_h^n - p_h \,. \] Note that $\boldsymbol{\delta}_h^n \in \ZZ_h$ and $\sigma_h^n \in \W_{0,h}$. Simple computations yield \[ {a_{T,h}}(\sigma_h^{n+1}, \sigma_h^{n+1}) = - {c_{T,h}^{\rm skew}}(\boldsymbol{\delta}_h^n; \theta_h, \sigma_h^{n+1}) \,. \] Therefore form Property \textbf{(P7)} and bound \eqref{eq:bounds} we obtain. \begin{equation} \label{eq:sigmahn} \beta_* \kappa_* \vert \sigma_h^{n+1}\vert_{1, \Omega} \leq \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data}} \vert \boldsymbol{\delta}_h^n\vert_{1, \Omega} \,. \end{equation} For the velocity we can derive the following equation \[ {a_{V,h}}(\theta_h^{n+1} ; \, \boldsymbol{\delta}_h^{n+1}, \boldsymbol{\delta}_h^{n+1}) = - {c_{V,h}^{\rm skew}}(\boldsymbol{\delta}_h^n; \, \boldsymbol{u}_h, \boldsymbol{\delta}_h^{n+1}) -\left({a_{V,h}}(\theta_h^{n+1}; \, \boldsymbol{u}_h, \boldsymbol{\delta}_h^{n+1}) - {a_{V,h}}(\theta_h; \, \boldsymbol{u}_h, \boldsymbol{\delta}_h^{n+1}) \right) \,. \] Then, employing properties \textbf{(P7)} and Lemma \ref{lm:lipshitz} we infer \[ \alpha_* \nu_* \vert \boldsymbol{\delta}_h^{n+1} \vert_{1, \Omega}^2 \leq \widehat{C}_{\rm conv} \vert \boldsymbol{u}_h \vert_{1, \Omega} \vert \boldsymbol{\delta}_h^n \vert_{1, \Omega} \vert \boldsymbol{\delta}_h^{n+1} \vert_{1, \Omega} + \widehat{C}_{\rm lip} \nu_{\rm lip} \vert \boldsymbol{u}_h \vert_{[W^1_q(\Omega_h)]^2} \vert \sigma_h^{n+1} \vert_{1, \Omega} \vert \boldsymbol{\delta}_h^{n+1} \vert_{1, \Omega}\,. \] Using \eqref{eq:bounds} and \eqref{eq:csolh}, from \eqref{eq:sigmahn} we derive \[ \alpha_* \nu_* \vert \boldsymbol{\delta}_h^{n+1} \vert_{1, \Omega} \leq \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data} } \vert \boldsymbol{\delta}_h^n \vert_{1, \Omega} + \frac{\widehat{C}_{\rm lip}}{\widehat{C}_{\rm sol} } \frac{\widehat{C}_{\rm conv} \widehat{C}_{\rm est}}{\widehat{C}_{\rm data} \beta_* \kappa_*} \vert \boldsymbol{\delta}_h^n \vert_{1, \Omega} \,, \] that, recalling \eqref{eq:uni}, corresponds to \[ \vert \boldsymbol{\delta}_h^{n+1} \vert_{1, \Omega} \leq K \, \vert \boldsymbol{\delta}_h^n \vert_{1, \Omega}\,. \] Therefore, since $\boldsymbol{u}_h^0 = \boldsymbol{0}$ from \eqref{eq:bounds}, we have \begin{equation} \label{eq:thetahn} \vert \boldsymbol{\delta}_h^n \vert_{1, \Omega} \leq \frac{\widehat{C}_{\rm est}}{\widehat{C}_{\rm data}} \, K^n \,, \qquad \qquad \vert \sigma_h^n \vert_{1, \Omega} \leq \frac{ \widehat{C}_{\rm conv} \widehat{C}_{\rm est}^2}{\widehat{C}_{\rm data}^2 \beta_* \kappa_*} K^{n-1} \,. \end{equation} Concerning the pressure we derive \begin{equation} \label{eq:rhohn1} \begin{aligned} b(\boldsymbol{v}_h, \rho_h^{n+1}) &= {a_{V,h}}(\theta_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - {a_{V,h}}(\theta_h^{n+1}; \, \boldsymbol{u}_h^{n+1}, \boldsymbol{v}_h) + \\ & \qquad + {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h^{n}; \, \boldsymbol{u}_h^{n+1}, \boldsymbol{v}_h) \\ &= \left({a_{V,h}}(\theta_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - {a_{V,h}}(\theta_h^{n+1}; \, \boldsymbol{u}_h, \boldsymbol{v}_h) \right) - {a_{V,h}}(\theta_h^{n+1}; \, \boldsymbol{\delta}_h^{n+1}, \boldsymbol{v}_h) \\ & \qquad - {c_{V,h}^{\rm skew}}(\boldsymbol{\delta}_h^n; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h^{n}; \, \boldsymbol{\delta}_h^{n+1}, \boldsymbol{v}_h) \,. \end{aligned} \end{equation} Lemma \ref{lm:lipshitz} implies \begin{equation} \label{eq:rhohn2} {a_{V,h}}(\theta_h; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - {a_{V,h}}(\theta_h^{n+1}; \, \boldsymbol{u}_h, \boldsymbol{v}_h) \leq \widehat{C}_{\rm lip} \nu_{\rm lip} \vert \sigma_h^{n+1}\vert_{1, \Omega} \vert \boldsymbol{u}_h\vert_{[W^1_q(\Omega)]^2} \vert \boldsymbol{v}_h\vert_{1, \Omega} \,, \end{equation} whereas Property \textbf{(P7)} yields \begin{equation} \label{eq:rhohn3} \begin{aligned} -{a_{V,h}}(\theta_h^{n+1}; \, \boldsymbol{\delta}_h^{n+1}, \boldsymbol{v}_h) &\leq \alpha^* \nu^* \vert \boldsymbol{\delta}_h^{n+1}\vert_{1, \Omega} \vert \boldsymbol{v}_h\vert_{1, \Omega} \\ - {c_{V,h}^{\rm skew}}(\boldsymbol{\delta}_h^n; \, \boldsymbol{u}_h, \boldsymbol{v}_h) - {c_{V,h}^{\rm skew}}(\boldsymbol{u}_h^{n}; \, \boldsymbol{\delta}_h^{n+1}, \boldsymbol{v}_h) & \leq \widehat{C}_{\rm conv} \left( \vert \boldsymbol{\delta}_h^{n}\vert_{1, \Omega} \vert \boldsymbol{u}_h\vert_{1, \Omega} + \vert \boldsymbol{\delta}_h^{n+1}\vert_{1, \Omega} \vert \boldsymbol{u}_h^{n}\vert_{1, \Omega} \right) \vert \boldsymbol{v}_h\vert_{1, \Omega} \,. \end{aligned} \end{equation} Then recalling \eqref{eq:csolh} and \eqref{eq:bounds} from \eqref{eq:rhohn1}, \eqref{eq:rhohn2} and \eqref{eq:rhohn3}, for any $\boldsymbol{v}_h \in \VV_h$ we have \[ \begin{aligned} \frac{b(\boldsymbol{v}_h, \rho_h^{n+1})}{{\vert\boldsymbol{v}_h\vert_{1,\Omega}}} &\leq \frac{\widehat{C}_{\rm lip}}{\widehat{C}_{\rm sol}} \vert \sigma_h^{n+1}\vert_{1, \Omega} + \left(\alpha^* \nu^* + \widehat{C}_{\rm conv} \frac{\widehat{C}_{\rm est}}{\widehat{C}_{\rm data}} \right) \vert \boldsymbol{\delta}_h^{n+1} \vert_{1, \Omega} + \widehat{C}_{\rm conv} \frac{\widehat{C}_{\rm est}}{\widehat{C}_{\rm data}} \vert \boldsymbol{\delta}_h^{n} \vert_{1, \Omega} \,. \end{aligned} \] Therefore from the inf-sup stability \eqref{eq:infsup} and \eqref{eq:thetahn} we have \[ \widehat{\beta} \Vert \rho_h^{n+1}\Vert_{0, \Omega} \leq \widehat{C}_{\rm conv} \frac{\widehat{C}_{\rm est}^2}{\widehat{C}_{\rm data}^2} \left( 1 + \frac{\widehat{C}_{\rm lip}}{\widehat{C}_{\rm sol} \beta_* \kappa_*} \right) K^n + O(K^{n+1}) \,. \] Now the {thesis} follows from \eqref{eq:uni}. \end{proof} \paragraph{\textbf{Test 1. Error convergence}} In this test we examine the practical performance and the convergence properties of the proposed scheme \eqref{eq:pb vem} in the light of Proposition \ref{prp:conv}. \noindent In order to compute the VEM error between the exact solution $\boldsymbol{u}_{\rm ex}$ and the VEM solution $\boldsymbol{u}_h$ we consider the computable error quantities \[ \begin{aligned} \texttt{err}(\boldsymbol{u}_h, H^1) &:= \texttt{sqrt}\left(\sum_{E \in \Omega_h} \|\boldsymbol{\nabla} \boldsymbol{u}_{\rm ex} - \Pi^{0,E}_{1} \boldsymbol{\nabla} \boldsymbol{u}_h\|^2_{0,E}\right) /\|\boldsymbol{\nabla} \boldsymbol{u}_{\rm ex}\|_{0} \,, \\ \texttt{err}(\boldsymbol{u}_h, L^2) &:= \texttt{sqrt}\left(\sum_{E \in \Omega_h} \|\boldsymbol{u}_{\rm ex} - \Pi^{0,E}_{2} \boldsymbol{u}_h\|_{0,E}\right) /\|\boldsymbol{u}_{\rm ex}\|_{0} \,. \end{aligned} \] Analogous error quantities are considered to measure the error between the exact solution $\theta_{\rm ex}$ and the VEM solution $\theta_h$. For the pressure error we take $\texttt{err}(p_h, L^2) := \|p - p_h\|_{0, \Omega}$. In the present test we consider Problem \eqref{eq:pb primale} on the unit square $\Omega = (0, 1)^2$, the viscosity is $\nu(\theta)=2 + \theta^2$, the conductivity is $\kappa(x, y)= 1$, the load terms $\boldsymbol{f}$ and $g$ and the Dirichlet boundary conditions are chosen in accordance with the analytical solution \[ \begin{aligned} \boldsymbol{u}_{\rm ex}(x,y) &= e^x \begin{pmatrix} \sin y + y \cos y - x \sin y \\ -x \cos y - y \sin y- \cos y \end{pmatrix} \,, \\ p_{\rm ex}(x,y) &= \sin( \pi x) \cos(4 \pi y) \,, \\ \theta_{\rm ex}(x,y) &= \sin(4 \pi x) \sin(\pi y) \,. \end{aligned} \] Notice that the viscosity $\nu$ satisfies Property \textbf{(P0)} whenever $\vert \theta \vert \leq \theta_{\rm max}$. The domain $\Omega$ is partitioned with the following sequences of polygonal meshes: \texttt{QUADRILATERAL} distorted meshes, \texttt{TRIANGULAR} meshes, \texttt{CVT} (Centroidal Voronoi Tessellations) meshes, \texttt{RANDOM} Voronoi meshes (see Fig. \ref{fig:meshes}). For the generation of the Voronoi meshes we used the code \texttt{Polymesher} \cite{TPPM12}. For each family of meshes we take the sequence with diameter $h =2^{-2}$, $2^{-3}$, $2^{-4}$, $2^{-5}$, $2^{-6}$. \begin{figure} \centering \begin{overpic}[scale=0.18]{mesh_quad_ok.pdf} \put(5,-15){{{\texttt{QUADRILATERAL}}}} \end{overpic} \,\,\, \begin{overpic}[scale=0.18]{mesh_triangoli.pdf} \put(15,-15){{{\texttt{TRIANGULAR}}}} \end{overpic} \,\,\, \begin{overpic}[scale=0.18]{mesh_voro_ok.pdf} \put(25,-15){{{\texttt{VORONOI}}}} \end{overpic} \,\,\, \begin{overpic}[scale=0.18]{mesh_voro_bad.pdf} \put(25,-15){{{\texttt{RANDOM}}}} \end{overpic} \vspace{0.5cm} \caption{Test 1. Example of the adopted polygonal meshes.} \label{fig:meshes} \end{figure} In Table \ref{tab:test1} we show the dimensions of the linear systems solved at each iteration of the fixed-point procedure and number of iterations needed for the adopter sequences of meshes. We iterate the fixed-point procedure with the tolerance \texttt{TOLL=1e-7}. \begin{table} \parbox{.45\linewidth}{ \centering \begin{tabular}{l*{3}{c}} \toprule \multicolumn{4}{c}{\texttt{QUADRILATERAL MESHES}} \\ \midrule $\texttt{1/h}$ & $\texttt{HEAT}$ & $\texttt{OSEEN}$ & $\texttt{N\_IT}$ \\ \midrule $\texttt{4}$ &$\texttt{49}$ &$\texttt{147}$ & $\texttt{6}$ \\ $\texttt{8}$ &$\texttt{225}$ &$\texttt{643}$ & $\texttt{6}$ \\ $\texttt{16}$ &$\texttt{961}$ &$\texttt{2691}$ & $\texttt{6}$ \\ $\texttt{32}$ &$\texttt{3969}$ &$\texttt{11011}$ & $\texttt{5}$ \\ $\texttt{64}$ &$\texttt{16129}$ &$\texttt{44547}$ & $\texttt{6}$ \\ \bottomrule \end{tabular} } \qquad \qquad \qquad \parbox{.45\linewidth}{ \centering \begin{tabular}{l*{3}{c}} \toprule \multicolumn{4}{c}{\texttt{TRIANGULAR MESHES}} \\ \midrule $\texttt{1/h}$ & $\texttt{HEAT}$ & $\texttt{OSEEN}$ & $\texttt{N\_IT}$ \\ \midrule $\texttt{4}$ &$\texttt{129}$ &$\texttt{403}$ & $\texttt{6}$ \\ $\texttt{8}$ &$\texttt{605}$ &$\texttt{1847}$ & $\texttt{6}$ \\ $\texttt{16}$ &$\texttt{2547}$ &$\texttt{7705}$ & $\texttt{5}$ \\ $\texttt{32}$ &$\texttt{10331}$ &$\texttt{31121}$ & $\texttt{5}$ \\ $\texttt{64}$ &$\texttt{41985}$ &$\texttt{126211}$ & $\texttt{6}$ \\ \bottomrule \end{tabular} } \\ \vspace{0.5cm} \parbox{.45\linewidth}{ \centering \begin{tabular}{l*{3}{c}} \toprule \multicolumn{4}{c}{\texttt{VORONOI MESHES}} \\ \midrule $\texttt{1/h}$ & $\texttt{HEAT}$ & $\texttt{OSEEN}$ & $\texttt{N\_IT}$ \\ \midrule $\texttt{4}$ &$\texttt{63}$ &$\texttt{175}$ & $\texttt{6}$ \\ $\texttt{8}$ &$\texttt{315}$ &$\texttt{823}$ & $\texttt{6}$ \\ $\texttt{16}$ &$\texttt{1399}$ &$\texttt{3567}$ & $\texttt{6}$ \\ $\texttt{32}$ &$\texttt{5847}$ &$\texttt{14767}$ & $\texttt{5}$ \\ $\texttt{64}$ &$\texttt{23877}$ &$\texttt{60043}$ & $\texttt{6}$ \\ \bottomrule \end{tabular} } \qquad \qquad \qquad \parbox{.45\linewidth}{ \centering \begin{tabular}{l*{3}{c}} \toprule \multicolumn{4}{c}{\texttt{RANDOM MESHES}} \\ \midrule $\texttt{1/h}$ & $\texttt{HEAT}$ & $\texttt{OSEEN}$ & $\texttt{N\_IT}$ \\ \midrule $\texttt{4}$ &$\texttt{69}$ &$\texttt{187}$ & $\texttt{6}$ \\ $\texttt{8}$ &$\texttt{307}$ &$\texttt{807}$ & $\texttt{6}$ \\ $\texttt{16}$ &$\texttt{1367}$ &$\texttt{3503}$ & $\texttt{6}$ \\ $\texttt{32}$ &$\texttt{5659}$ &$\texttt{14391}$ & $\texttt{5}$ \\ $\texttt{64}$ &$\texttt{22921}$ &$\texttt{58131}$ & $\texttt{6}$ \\ \bottomrule \end{tabular} } \vspace{0.25cm} \caption{Test 1. Dimensions of the linear systems solved at each iteration of the fixed-point procedure and number of iterations.} \label{tab:test1} \end{table} In Fig. \ref{fig:results} we display the errors $\texttt{err}(\boldsymbol{u}_h, H^1)$, $\texttt{err}(\boldsymbol{u}_h, L^2)$, $\texttt{err}(\theta_h, H^1)$, $\texttt{err}(\theta_h, L^2)$ and $\texttt{err}(p_h, L^2)$ for the the sequences of meshes aforementioned. We notice that the theoretical predictions of Proposition \ref{prp:conv} are confirmed, moreover we observe that the method is robust with respect to the mesh distortion. \begin{figure} \parbox{.45\linewidth}{ \centering \begin{overpic}[scale=0.16]{quad.pdf} \put(23,-5){{\texttt{QUADRIALTERAL MESHES}}} \end{overpic} } \qquad \parbox{.45\linewidth}{ \centering \begin{overpic}[scale=0.16]{tri.pdf} \put(27,-5){\texttt{TRIANGULAR MESHES}} \end{overpic} } \end{figure} \begin{figure} \parbox{.45\linewidth}{ \centering \begin{overpic}[scale=0.16]{voro.pdf} \put(33,-5){\texttt{VORONOI MESHES}} \end{overpic} } \qquad \parbox{.45\linewidth}{ \centering \begin{overpic}[scale=0.16]{rand.pdf} \put(35,-5){\texttt{RANDOM MESHES}} \end{overpic} } \vspace{1cm} \caption{Test 1. Convergence histories of the VEM \eqref{eq:pb vem} for the adopted families of polygonal meshes.} \label{fig:results} \end{figure} \paragraph{Test 2. Convection-dominated transport of a passive scalar.} The scope of the present test is to show that divergence-free discrete velocity solutions might also be of advantage in coupled problems. We investigate the behavior of the proposed method for the test proposed in \cite[Example 6.6]{john-linke-merdon-neilan-rebholz:2017}. We consider the following coupled problem on the domain $\Omega = (0, 4) \times (0,2) \setminus [2, 4] \times [0, 1]$ with a flow field which is governed by the Stokes equation \begin{equation} \label{eq:pb test} \left\{ \begin{aligned} & \begin{aligned} - \nu \, {\rm div} (\boldsymbol{\epsilon} (\boldsymbol{u})) - \nabla p &= \boldsymbol{0}\qquad & &\text{in $\Omega$,} \\ {\rm div} \, \boldsymbol{u} &= 0 \qquad & &\text{in $\Omega$,} \\ - \kappa \, \Delta \theta + \boldsymbol{u} \cdot \nabla \theta &= 0 \qquad & &\text{in $\Omega$,} \end{aligned} \end{aligned} \right. \end{equation} coupled with the following boundary conditions (see Fig. \ref{fig:test2problem} left) \begin{equation} \label{eq:bou test} \left\{ \begin{aligned} & \begin{aligned} \boldsymbol{u} &= (0.5\, y(2-y), \, 0)^{\rm T} & &\text{on $\Gamma_{\rm in}$,} \\ \boldsymbol{u} &= (4\, (y-1)(2-y), \, 0)^{\rm T} & &\text{on $\Gamma_{\rm out}$,} \\ \boldsymbol{u} &= \boldsymbol{0} & &\text{on $\partial \Omega \setminus( \Gamma_{\rm in} \cup \Gamma_{\rm out})$,} \end{aligned} \end{aligned} \right. \quad \left\{ \begin{aligned} & \begin{aligned} \theta &= 1 & &\text{on $\Gamma_{\rm in}$,} \\ \kappa \nabla \theta \cdot \boldsymbol{n} &= 0 & &\text{on $\partial \Omega \setminus \Gamma_{\rm in}$,} \end{aligned} \end{aligned} \right. \end{equation} the viscosity is $\nu = \texttt{1e-2}$, the conductivity is $\kappa = \texttt{1e-6}$. \begin{figure} \centering \qquad \begin{overpic}[scale=0.225]{prob-test2.pdf} \put(-7,25){{$\Gamma_{\rm in}$}} \put(67,32){{$\Gamma_{\rm out}$}} \end{overpic} \quad \begin{overpic}[scale=0.225]{mesh-test2.pdf} \end{overpic} \vspace{0.5cm} \caption{Test 2. Domain and boundary conditions description (left) and mesh with diameter $h=2^{-2}$ (right).} \label{fig:test2problem} \end{figure} The exact temperature solution is $\theta_{\rm ex} = 1$. For the sake of the comparison with the results reported in \cite[Example 6.6]{john-linke-merdon-neilan-rebholz:2017}, the domain $\Omega$ is partitioned with a sequence of triangular meshes with diameter $h =2^{-1}$, $2^{-2}$, $2^{-3}$, $2^{-4}$, $2^{-5}$ (see Fig. \ref{fig:test2problem} right). The continuous convective form associated with the spaces satisfying the boundary conditions \eqref{eq:bou test} is not skew-symmetric, therefore in the discrete scheme we consider the convective form \eqref{eq:forma cshl} in the place of \eqref{eq:forma cshlskew}. Note that the heat equation for the temperature in \eqref{eq:pb test} is convection-dominated. However, it is worth stressing that, differently form \cite{john-linke-merdon-neilan-rebholz:2017}, we do not employ any stabilization technique to stabilize the discrete problem (we refer to \cite{berrone-supg,BdV-supg,BdV-oseen} for the analysis of the stabilized method in the VEM context both for the convection-dominated elliptic equations and Oseen equations). Fig. \ref{fig:quiver}, Fig. \ref{fig:test2res} and Table \ref{tab:test2} present the results of the numerical simulations. As exhibited in \cite[Example 6.6]{john-linke-merdon-neilan-rebholz:2017} the violation of the divergence constraint, as it appears in the classical mixed finite element, pollutes the discrete temperature solution with strong spurious oscillations. By contrast, the proposed VEM scheme with divergence-free velocity solution computes the temperature with high accuracy as depicted in Fig. \ref{fig:test2res} and Table \ref{tab:test2}. \begin{figure} \centering \begin{overpic}[scale=0.27]{quiver.pdf} \end{overpic} \vspace{0.5cm} \caption{Test 2. Quiver of the numerical velocity $\boldsymbol{u}_h$ with $h=2^{-4}$.} \label{fig:quiver} \end{figure} \begin{figure} \centering \begin{overpic}[scale=0.25]{vel-test2.pdf} \end{overpic} \qquad \begin{overpic}[scale=0.25]{temp-test2.pdf} \end{overpic} \vspace{-0.25cm} \caption{Test 2. Absolute value of the velocity $\vert \boldsymbol{u}_h\vert$ (left) and temperature $\theta_h$ (right) with $h=2^{-4}$.} \label{fig:test2res} \end{figure} \begin{table}[!h] \centering \begin{tabular}{cccccc} \toprule $\texttt{1/h}$ & $\texttt{2}$ & $\texttt{4}$ & $\texttt{8}$ & $\texttt{16}$ & $\texttt{32}$ \\ \midrule $\texttt{min}(\theta_h-1)$ & $\texttt{-7.88e-11}$ & $\texttt{-2.78e-10}$ & $\texttt{-2.87e-08}$ & $\texttt{-4.47e-09}$ & $\texttt{-3.23e-09}$ \\ \midrule $\texttt{max}(\theta_h-1)$ & $\texttt{+4.50e-11}$ & $\texttt{+7.51e-11}$ & $\texttt{+2.42e-08}$ & $\texttt{+2.90e-09}$ & $\texttt{+5.34e-09}$ \\ \bottomrule \end{tabular} \vspace{0.25cm} \caption{Test 2. Minimal and maximal of $\theta - 1$ (\cfun{DoFs} values).} \label{tab:test2} \end{table} \section{Conclusions} \label{sec:conclusions} We presented a Virtual Element discretization of a thermo-fluid dynamic problem modeling the stationary flow of a viscous incompressible fluid (governed by a Navier-Stokes equation), where the viscosity of the fluid depends on the temperature (governed by an elliptic equation). We have show the the discrete problem is well-posed and prove optimal error estimates. Numerical experiments which confirm the theoretical error bounds are also presented. Further developments may include the introduction of uncertainty in the thermal conductivity $\kappa$, based on the ideas of \cite{Bonizzoni_Nobile_2014,Bonizzoni_Nobile_2020} as well as the three-dimensional extension, based on employing the three-dimensional VEM for Navier-Stokes proposed in \cite{Beirao_Dassi_Vacca:2020}. \section*{Funding} P.F.A. and M.V. have been partially funded by the research grants PRIN2017 n. \verb+201744KLJL+\emph{``Virtual Element Methods: Analysis and Applications''} and PRIN2020 n. \verb+20204LN5N5+\emph{``Advanced polyhedral discretisations of heterogeneous PDEs for multiphysics problems''}, funded by the Italian Ministry of Universities and Research (MUR). P.F.A, G.V. and M.V. are members of INdAM-GNCS. \bibliographystyle{plain}
{ "redpajama_set_name": "RedPajamaArXiv" }
9,840
Tissue Engineering and Regenerative Medicine International Society is an international learned society dedicated to tissue engineering and regenerative medicine. Background Regenerative medicine involves processes of replacing, engineering or regenerating human cells, tissues or organs to restore or establish normal function. A major technology of regenerative medicine is tissue engineering, which has variously been defined as "an interdisciplinary field that applies the principles of engineering and the life sciences toward the development of biological substitutes that restore, maintain, or improve tissue function", or "the creation of new tissue by the deliberate and controlled stimulation of selected target cells through a systematic combination of molecular and mechanical signals". History Tissue engineering emerged during the 1990s as a potentially powerful option for regenerating tissue and research initiatives were established in various cities in the USA and in European countries including the UK, Italy, Germany and Switzerland, and also in Japan. Soon fledgling societies were formed in these countries in order to represent these new sciences, notably the European Tissue Engineering Society (ETES) and, in the USA, the Tissue Engineering Society (TES), soon to become the Tissue Engineering Society international (TESi) and the Regenerative Medicine Society (RMS). Because of the overlap between the activities of these societies and the increasing globalization of science and medicine, considerations of a merger between TESI and ETES and RMS were initiated in 2004 and agreement was reached during 2005 about the formation of the consolidated society, the Tissue Engineering and Regenerative Medicine International Society (TERMIS). Election of officers for TERMIS took place in September 2005, and the by-laws were approved by the Board. Rapid progress in the organization of TERMIS took place during late 2005 and 2006. The SYIS, Student and Young Investigator Section was established in January 2006, website and newsletter launched and membership dues procedures put in place. Structure and governance It was determined that each Chapter would have its own Council, the overall activities being determined by the Governing Board, on which each Council was represented, and an Executive Committee. Society chapters At the beginning of the Society, it was agreed that there would be Continental Chapters of TERMIS, initially TERMIS-North America (TERMIS-NA) and TERMIS-Europe (TERMIS-EU), to be joined at the time of the major Shanghai conference in October 2005 by TERMIS-Asia Pacific (TERMIS-AP). It was subsequently agreed that the remit of TERMIS-North America should be expanded to incorporate activity in South America, the chapter becoming TERMIS-Americas (TERMS-AM) officially in 2012. Student and Young Investigator Section The Student and Young Investigator Section of TERMIS (TERMIS-SYIS) brings together undergraduate and graduate students, post-doctoral researchers and young investigators in industry and academia related to tissue engineering and regenerative medicine. It follows the organizational and working pattern of TERMIS. Activities Journal A contract was signed between TERMIS and the Mary Ann Liebert publisher which designated the journal Tissue Engineering, Parts A, B, and C as the official journal of TERMIS with free on-line access for the membership. Conferences It was agreed that there would be a World Congress every three years, with each Chapter organizing its own conference in the intervening two years. Awards Each TERMIS chapter has defined awards to recognize outstanding scientists and their contributions within the community. TERMIS-AP The Excellence Achievement Award has been established to recognize a researcher in the Asia-Pacific region who has made continuous and landmark contributions to the tissue engineering and regenerative medicine field. The Outstanding Scientist Award has been established to recognize a mid-career researcher in the Asia-Pacific region who has made significant contributions to the TERM field. The Young Scholar Award has been established to recognize a young researcher in the Asia-Pacific region who has made significant and consistent achievements in the TERM field, showing clear evidence of their potential to excel. The Mary Ann Liebert, Inc. Best TERM Paper Award has been established to recognize a student researcher (undergraduate/graduate/postdoc) in the Asia-Pacific region who has achieved outstanding research accomplishments in the TERM field. The TERMIS-AP Innovation Team Award has been established to recognize a team of researchers in the Asia-Pacific region. It aims to recognize successful applications of tissue engineering and regenerative medicine leading to the development of relevant products/therapies/technologies which will ultimately benefit the patients. TERMIS-EU The Career Achievement Award is aimed towards a recognition of individuals who have made outstanding contributions to the field of TERM and have carried out most of their career in the TERMIS-EU geographical area. The Mid Terms Career Award has been established in 2020 to recognize individuals that are within 10–20 years after obtaining their PhD, with a successful research group and clear evidence of outstanding performance. The Robert Brown Early Career Principal Investigator Award recognizes individuals that are within 2–10 years after obtaining their PhD, with clear evidence of a growing profile. Award recipients Fellows Fellows of Tissue Engineering and Regenerative Medicine (FTERM) recipients are: Alini, Mauro Atala, Anthony Badylak, Stephen Cancedda, Ranieri Cao, Yilin Griffith, Linda Guldberg, Robert Hellman, Kiki Hilborn, Jöns Hubbell, Jeffrey Hutmacher, Dietmar Johnson, Peter Khang, Gilson Kirkpatrick, C. James Langer, Robert Lee, Hai-Bang Lee, Jin Ho Martin, Ivan Mikos, Antonios Mooney, David Naughton, Gail Nerem, Robert Okano, Teruo Pandit, Abhay Parenteau, Nancy Radisic, Milica Ratner, Buddy Redl, Heinz Reis, Rui L. Russell, Alan Shoichet, Molly Smith, David Tabata, Yasuhiko Vacanti, Charles Vacanti, Joseph van Osch, Gerjo Vunjak-Novakovic, Gordana Wagner, William Weiss, Anthony S. Williams, David References External links TERMIS homepage Tissue Engineering Journal Tissue engineering Medical associations based in the United States Medical and health organizations based in California International medical associations
{ "redpajama_set_name": "RedPajamaWikipedia" }
3,502
{"url":"https:\/\/puzzling.stackexchange.com\/questions\/98383\/find-relation-between-the-following-numbers-with-use-elementary-arithmetic","text":"Find relation between the following numbers with use Elementary arithmetic\n\nFind relation between the following numbers:\n\nwe can use only Elementary arithmetic (addition, subtraction, multiplication, and division) . I find this puzzle in my old book .I think we have Arithmetic progression , because in the first shape (up left) we have 5+6+1=12 ,1+6+9=16 ,9+6+5=20 so we have sequence:12,16,20 so common difference of successive members is $$d=4$$ for second shape we have:13,16,19 so $$d=3$$ for third shape:18,20,22 ,$$d=2$$ finally for last shape we have $$d=1$$ but this is impossible. I think my answer is wrong because we delete center number in this puzzle and we must use center numbers in all shapes.","date":"2020-10-25 17:35:26","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 4, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.7910728454589844, \"perplexity\": 817.7340515579506}, \"config\": {\"markdown_headings\": false, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2020-45\/segments\/1603107889574.66\/warc\/CC-MAIN-20201025154704-20201025184704-00449.warc.gz\"}"}
null
null
If you're a morning person, the best time to nap is around 1 or 1:30PM. If you're a night owl, nap later, around 2:30 or 3PM. The best naps are under 45 mins or 90-120 mins. Anything in between is likely to give you that groggy feeling. Naps don't mean you're lazy: A NASA study showed that in-flight naps improved subsequent performance by 34% and overall alertness by 54%. To make sure you're productive after your nap drink a cup of coffee right before laying down. Caffeine takes about 20-30 minutes to kick in. No nap is too short: A 2008 study showed that even a nap of a few minutes provided benefits. Just anticipating a nap lowers blood pressure. Keep in mind that 10 minutes is the best nap length. Naps increase alertness and performance on the job. Naps purge negative emotions and enhance positive ones. Being tired actually makes it harder to be happy. Lack of sleep = more likely to get sick. "Sleeping on it" does improve decision making. Lack of sleep can make you more likely to behave unethically. There is such a thing as beauty sleep. Join over 140,000 readers and get a free weekly update via email here. « What are life's 7 magic numbers?
{ "redpajama_set_name": "RedPajamaC4" }
1,732
{"url":"https:\/\/codegolf.stackexchange.com\/questions\/206342\/scoring-quantum-tic-tac-toe","text":"# Scoring Quantum Tic-Tac-Toe\n\nIn the description of this challenge, the following board will be used as a reference for positions:\n\nABC\nDEF\nGHI\n\n\nFor instance, in a game of ordinary tic-tac-toe, B E A C G D F H I describes the following:\n\n-X- -X- XX- XXO XXO XXO XXO XXO XXO\n--- -O- -O- -O- -O- OO- OOX OOX OOX\n--- --- --- --- X-- X-- X-- XO- XOX\n\n\n# Quantum Tic-Tac-Toe Gameplay\n\nIn quantum tic-tac-toe, players' moves are superpositions of moves in classical tic-tac-toe. Players mark two cells per move instead of just one, and those cells become entangled. Each quantum mark is given a subscript telling the number of the turn in which it was placed.\n\nIn the above board, A and B are entangled, and so it is said that X made the move AB: X1 is either in A or B but the true position cannot be known with certainty until later in the game.\n\nConsider the following board, made with the sequence AB BE DE AE (X makes the move AB, O makes the move BE, X makes the move DE, O makes the move AE; note that the relative positions of quantum marks in a cell in the image do not actually matter):\n\nHere, there has arisen a cyclic entanglement between cells A, B and E. Consider the following image, in which a line represents an entanglement:\n\nAfter a cyclic entanglement (which does not necessarily have length 3) is formed, measurement takes place, in which every quantum mark involved in the cycle collapses into a known state and becomes a classical mark. This includes entanglements like the red one above that are merely accessories to the cyclic entanglement, and don't actually complete the cycle. After one player forms a cyclic entanglement, the other player chooses which state the board assumes. Since X formed the cyclic entanglement, it is up to O to choose between these two states:\n\nThere are only ever two possible states to choose from, because all individual entanglements only involve two cells. To record which state was chosen, the following methodology will be used:\n\n\u2022 Find the first cell (i.e. A takes precedence over B, etc) in the cyclic entanglement which cannot be removed without breaking the cycle (meaning entanglements like the red one above are excluded). In this case this is A.\n\n\u2022 Write the number of the mark that fills that cell.\n\nSince either X1 or O4 will occupy A, either 1 or 4 will be written. This would mean the above boards would be described by the sequences AB BE DE AE 1 and AB BE DE AE 4, respectively. After measurement occurs, no more moves can be made in A, B, D or E because it is now known what lies in those cells.\n\nA single-cell entanglement, e.g. AA 1, is not allowed under ordinary circumstances, as that would mimic the behavior of a classical mark. This sort of move is only allowed when just one cell not occupied by a classical mark remains.\n\nA game continues until at least one tic-tac-toe is formed or until the board is filled with classical marks. Multiple tic-tac-toes can be made only if they appear simultaneously after a measurement; this is because the presence of one or multiple tic-tac-toes disallows the placement of any additional quantum marks.\n\n# Scoring\n\nThe first number corresponds to X's score and the second to O's.\n\n\u2022 If neither player has a tic-tac-toe, then both players get zero points.\n\n\u2022 If there is only one tic-tac-toe present, then the player with the tic-tac-toe gets one point and the other player gets zero points.\n\n\u2022 If one player gets two tic-tac-toes after a measurement, then they get two points and the other player gets zero. It is not possible for one player to get one tic-tac-toe while the other gets two, nor is it possible for both players to get two, nor one player to get three.\n\n\u2022 If both players get a tic-tac-toe after a measurement, then the player with the least maximum subscript in their tic-tac-toe will get one point and the other player will get one half-point.\n\nFor clarification on that last point, consider the board made by the sequence AI EG BE EH DG AI 1 BC CE 3:\n\nThe maximum subscript in X's tic-tac-toe is seven, while the maximum subscript in O's tic-tac-toe is six. Because of this, O has the least maximum subscript and gets one point while X gets one half-point.\n\n# The Challenge\n\nGiven a series of moves from a game of quantum tic-tac-toe, determine the score of each player.\n\n# Rules\n\n\u2022 Input and output may be given in whatever form is most convenient, but it must be explained if it differs from what is described above. Changes can be slight, like describing locations or scores differently, or more drastic. For instance, writing the number of the classical mark that fills the cell of the last quantum mark placed can be done instead of writing the number of the mark that fills the first cell alphabetically.\n\n\u2022 One must always be able to tell which score corresponds to which player. For instance, always putting the winner's score first is not acceptable unless the output also tells which player won or lost. If a specific player's score is always given first, that does not have to be included in the output (as seen in the test cases).\n\n\u2022 Assume only valid input is given.\n\n\u2022 Spaces are not necessary; they were added for the purpose of legibility.\n\n\u2022 This is so the smallest program in bytes wins.\n\n# Test Cases\n\nIn: AB\nOut: 0 0\n\nIn: AI EG BE EH DG AI 1 BC CE 3\nOut: 0.5 1\n\nIn: IA EG EB EH GD AI 1 CB CE 3\nOut: 0.5 1\n\nIn: DE AB DE 1 AH CF CH CG BC 2\nOut: 1 0.5\n\nIn: AE BF EI DH AI 1\nOut: 1 0\n\nIn: CE BD EG FH CG 1\nOut: 1 0\n\nIn: AB AD BC CF AC 5\nOut: 1 0\n\nIn: AB BE BC CE 2 DG HI DF HI 6\nOut: 0 1\n\nIn: BC BE CF EF 1 AD AG GH DG 6 II 9\nOut: 0 0\n\nIn: DI AE AB BF EH AH 2 DI 7 CG CG 8\nOut: 0 0\n\nIn: AD BH EF CI EF 3 HI DG BI 2 AG 1\nOut: 2 0\n\n\u2022 There appear to be no testcases where there is a 3-in-a-row on either diagonal Jun 21, 2020 at 4:57\n\u2022 Switching C with I in the testcase which outputs 0 1 gives an antidiagonal 3-in-a-row, and switching A with G in the same testcase gives a diagonal 3-in-a-row Jun 21, 2020 at 5:19\n\u2022 @fireflame241 I found shorter ones: Main diagonal is AE BF EI DH AI 1 and antidiagonal is CE BD EG FH CG 1 Jun 21, 2020 at 5:31\n\u2022 My brain just exploded, or it didn't.. I'll have to make a measure to find out Jun 22, 2020 at 12:35\n\n# Python 3.8 (pre-release), 419371 351 bytes\n\n-48 bytes following suggestions from @ovs\n\n-16 bytes thanks to @JBernardo\n\ndef s(C,l=[]):\nl=l+[p:=C.pop(0)]\nif 0==p*0:\ndel l[~1];r=[p]\nwhile r:c=r.pop();l=[r.append(t:=sum(e)-c)or t if e*0!=0and c in e else e for e in l]\nif C:s(C,l)\nelse:w=[a:=[sum(z.count(j)>2for z in[*zip(*[[k\/\/3,k%3,k%4,6559>>2*k&3]for k in l[i:f+1:2]if 0==k*0])]for j in set(z))for i in[0,1]]for f in range(9)];a[[1,0]in w]\/=1+([1,1]==a);print(a)\n\n\nTry it online!\n\nTakes input with A-I translated to 0-8 in tuples. The collapse choice is represented by the cell the last entanglement collapses to.\n\n### Massive Explanation\n\n# Take input C\n# I(k) <==> move k is a collapse, not entanglement <==> 'imag'in dir(k) <==> isinstance(k,int) <==> hasattr(k,'imag')\n# I have a feeling that collapses vs entanglements can be encoded in a much more byte-efficient manner\nI=lambda k:'imag'in dir(k)\n# C: list of moves, where a pair entry (e.g. [0,1]) is entanglement (e.g 0-1).\n# and an int entry (e.g 1) states collapse of the last entry, (e.g. [0,1] collapsed to 1)\n# l: be the list of entangled\/collapsed states, built up with recursion\n# pair (e.g. [0,1]) represents an existing entanglement, and int (e.g. 1) represents collapsed state\n# the owners alternate XOXOXO in the list\ndef s(C,l=[]):\nl=l[:] # clone l for array operations\nif C: # deal with end condition: only recurse if C has any mvoes\nl+=[ # append the first move in C to l\np:=C[0] # assign p to be the first move\n]\nif I(p): # if that move is an int, collapse:\ndel l[~1]; # delete the entangled move before\nr=[p] # r will be the list of location to remove\n# start by adding p to r: no entanglement can have the location p anymore\n# I have a feeling this whole while loop can be replaced with a recursive method\nwhile r: # while there is still a location to remove:\nc=r.pop() # take c to be the first location, so we will collapse all states that contain c\nfor i,e in enumerate(l): # for each state e of l:\nif (1-I(e) # If it is not a collapsed state\nand c in e): # and it contains c\nr+=[ # add to r:\nsum(l[i])-c # the location besides e of c\n];\nl[i]=r[~0] # replace e with that collapsed location\ns(C[1:],l) # recurse without the first move and with the modified l\nelse: # end condition: all moves have been processed\n# for case 5 (begins [[0, 1], [1, 4]...),\n# l=[0, 1, 2, 4, [3, 6], 7, [3, 5], 8] (most states are collapsed, except for [3,6] and [3,5])\nw=[ # w will be a list, supposing everything is collapsed as soon as it is intered,\n# of how many 3-in-a-rows each player has after each move\na:= # this is a for loop, so a gets set to the *last* 3-in-a-row count (the final win state)\n[\nsum( # take the sum of:\nz.count(j)>2 # the number of directions that have exactly 3 (more than 2) identical entries\nfor z in # for each of the 3 directions (rows, cols, diags) in\n[*zip(*[ # the transpose:\n[\nk\/\/3, # floor divide maps items on each row to the same value\nk%3, # mod 3 maps items in the same column to the same value\nk%4 # mod 4 maps items on the main diagonal to 0\n6559>>2*k&3 # k on antidiagonal to 0; 0,1,7,8 to negative; 3,5 to 1\n# (k in[2,4,6] doesn't work because it also maps all of 0,1,3,5,7,8 to False)\n]\nfor k in l[i:f+1:2] # i=0: every even state; i=1: every odd state\n# states limited to from the first f moves\nif I(k) # only consider the state if it is collapsed\n]\n)]\nfor j in set(z) # take the count for each unique value in the direction\n)\nfor i in[0,1] # do this for both player i=0=X and player i=1=O)\n]\nfor f in range(9) # for each count f of moves from 0 moves to 8 moves, calculate the 3-in-a-row count\n]\n# now a should be [0,0],[1,0],[2,0],[1,1] or the opposite order\nif[1,1]==a: # [1,1] is the only case we have to special handle\na[[1,0]in w]*=.5 # subtract .5 from the second element (O's score) if X got the point first, otherwise subtract from X's score\nprint(a) # finally\n\n\u2022 A few minor golfs: I can (probably) be written shorter as I=lambda k:k*0==0. Copying a list can be done by multiplying by 1: l=l*1. l[i]=d=sum(l[i])-c;r+=d, saves a few bytes. You can get your end condition into a single line by using a[[1,0]in w]\/=1+([1,1]==a).\n\u2013\u00a0ovs\nJun 21, 2020 at 9:15\n\u2022 @ovs Nice tips! Since I immediately appended to the cloned l, array deconstruction ended up shorter than multiplying by 1. Jun 22, 2020 at 1:28\n\u2022 There is a trailing semicolon on line 8. Also l=l+[p:=C.pop(0)] is one byte shorter Jun 23, 2020 at 0:54\n\u2022 You can shave a lot more on the while loop too: while r:c=r.pop();l=[r.append(t:=sum(e)-c)or t if e*0!=0and c in e else e for e in l] Jun 23, 2020 at 1:23\n\u2022 @JBernardo Good suggestion... much more impactful than the 2 bytes of your previous one :). Jun 23, 2020 at 2:14","date":"2023-03-22 13:15:39","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 1, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.5310482382774353, \"perplexity\": 3414.2547635137407}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 20, \"end_threshold\": 15, \"enable\": false}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2023-14\/segments\/1679296943809.76\/warc\/CC-MAIN-20230322114226-20230322144226-00080.warc.gz\"}"}
null
null
{"url":"https:\/\/zbmath.org\/?q=an%3A0944.34026","text":"## Oscillations of second-order nonlinear ordinary differential equations with impulses.(English)Zbl\u00a00944.34026\n\nHere, the second-order nonlinear differential equation with impulses of the form \\begin{aligned} &[r(t)x'(t)]'+f(t,x(t))=0,\\quad t\\geq t_0,\\;t\\neq t_k,\\;k=1,2,\\dots ,\\\\ &x(t_k^{+})=g_k(x(t_k)),\\;x'(t_k^{+})=h_k(x'(t_k)),\\;\\;k=1,2,\\dots ,\\\\ &x(t_0^{+})=x_0,\\;\\;x'(t_0^{+})=x'_0, \\end{aligned} with $$0\\leq t_0<t_1<\\dots <t_k<\\dots$$, and $$\\lim_{k\\to\\infty}t_k=+\\infty$$ is studied. The authors derive sufficient conditions for every solution to (1) to be oscillatory. Results of the paper generalize and improve those of Y. S. Chen and W. Z. Feng [J. Math. Anal. Appl. 120, No. 1, 150-169 (1997; Zbl\u00a00877.34014)].\n\n### MSC:\n\n 34C10 Oscillation theory, zeros, disconjugacy and comparison theory for ordinary differential equations 34A37 Ordinary differential equations with impulses\n\n### Keywords:\n\noscillations; nonlinear equations with impulses\n\nZbl 0877.34014\nFull Text:\n\n### References:\n\n [1] Chen, Y.S.; Feng, W.Z., Oscillations of second order nonlinear ODE with impulses, J. math. anal. appl., 120, 150-169, (1997) \u00b7 Zbl\u00a00877.34014 [2] Butler, G.J., Integral averages and the oscillation of second order ordinary differential equation, SIAM J. math. anal., 11, 190-200, (1980) \u00b7 Zbl\u00a00424.34033 [3] Lakshmikantham, V.; Bainov, D.D.; Simienov, P.S., Theory of impulsive differential equations, (1989), World Scientific Singapore [4] Travis, C.C., A note on second order nonlinear oscillations, Math. jap., 18, 261-264, (1973) \u00b7 Zbl\u00a00297.34032 [5] Wong, J.S.W., On second order nonlinear oscillations, Funkcial. ekavac., 11, 207-234, (1968) \u00b7 Zbl\u00a00157.14802\nThis reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.","date":"2022-09-25 01:31:57","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 1, \"mathjax_display_tex\": 1, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.7593913078308105, \"perplexity\": 2921.1358709939354}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2022-40\/segments\/1664030334332.96\/warc\/CC-MAIN-20220925004536-20220925034536-00290.warc.gz\"}"}
null
null
Dom José de Camargo Barros (Indaiatuba, 24 de abril de 1858 — mar Mediterrâneo, 4 de agosto de 1906) foi um bispo católico brasileiro. Foi o primeiro bispo de Curitiba e o 12º bispo de São Paulo. Biografia Filho de João Batista de Camargo Barros e de Gertrudes de Assunção Camargo. Faleceu aos 48 anos, no naufrágio do navio Sírio, que veio a pique próximo ao Cabo de Palos, no litoral da Espanha. Seu corpo não foi encontrado, sendo-lhe dedicado um cenotáfio na cripta da catedral de São Paulo. Neste mesmo navio estava Dom José Marcondes Homem de Melo que fora ordenado bispo, em Roma, por Dom José e outros prelados, que estavam retornando de Roma, onde haviam feito a Visita ad limina ao Papa São Pio X. Dom José Marcondes Homem de Melo, arcebispo do Pará, sobreviveu ao naufrágio. Foi conde romano. Estudos Fez seus primeiros estudos em escolas de Itu e Sorocaba. Em 1875 foi admitido no Colégio São Luís, dos jesuítas, de Itu, donde foi encaminhado para o Seminário Episcopal de Dom Lino Deodato de Carvalho, sendo admitido em 1877. Presbiterado Foi ordenado sacerdote a 13 de março de 1883, na Capela do Seminário, em Curitiba, por imposição das mãos de Dom Lino Deodato Rodrigues de Carvalho. Por quatro anos, foi professor do Seminário, em Curitiba, onde ocupou o cargo de mestre de disciplina. Depois, foi nomeado pároco da matriz de Santa Ifigênia. Literato e jornalista, fundou o semanário "O Lidador", o "Boletim Eclesiástico", o "Mensageiro de Santo Antônio" , "O Cruzeiro do Sul", "A Sineta de Deus" e "A Estrela" Episcopado O Papa Leão XIII designou Dom José para ser o primeiro Bispo de Curitiba, no dia 16 de janeiro de 1894, aos 35 anos. Foi sagrado bispo, em Roma, na capela do Colégio Pio Latino Americano, no dia 24 de junho de 1894, pelas mãos de S. Ema. Revma. Lucido Maria Cardeal Parocchi, sendo consagrantes: D. Joaquim Arcoverde de Albuquerque Cavalcanti, então bispo coadjutor de São Paulo e D. Eduardo Duarte e Silva, então bispo de Goiás. Após ser sagrado, voltou rapidamente ao Brasil, para participar da missa de sétimo dia de Dom Lino. Fez sua entrada solene na catedral de Curitiba no dia 30 de setembro de 1894. No dia 9 de novembro de 1903 o São Pio X nomeou Dom José para ser o 12º Bispo de São Paulo, sucedendo a Dom Antônio Cândido Alvarenga. Dom José tomou posse da diocese e fez sua entrada solene na catedral no dia de seu aniversário, 24 de abril de 1904. Dom José foi sucedido por Dom Duarte Leopoldo e Silva tanto na Diocese de Curitiba como na Diocese de São Paulo. Ordenações episcopais Dom José foi sagrante de Dom José Marcondes Homem de Melo. Ligações externas Bispos e arcebispos de São Paulo Bispos e arcebispos de Curitiba Naturais de Indaiatuba Mortos em 1906 Família Camargo (negócios) Mortes em naufrágios
{ "redpajama_set_name": "RedPajamaWikipedia" }
513
extern "C" { #endif MSVC_EXPORT extern int getpagesize(void); #ifdef __cplusplus } #endif #endif // _MSC_MMAN_H_
{ "redpajama_set_name": "RedPajamaGithub" }
2,751
El Presidente de la República Democrática Popular Lao es el cargo del Jefe de Estado de esta nación. Fue creado en 1975 tras la abolición del Reino de Laos por el Pathet Lao. Lista de presidentes Véase también Anexo:Primeros ministros de Laos Política de Laos
{ "redpajama_set_name": "RedPajamaWikipedia" }
2,335
coloring pages pigs cute baby pig coloring pages pig cartoon coloring pages coloring pages for three little pigs houses. this we find out coloring pages pigs cute baby pig coloring pages pig cartoon coloring pages coloring pages for three little pigs houses to an ideas that engage in child activities. coloring is beautiful and cheerful, also includes proficient training activities in where regard. it has in any event been a joy to discover things that start with very simply from the more complex letters or tricks. the most recognizable thing such as the cartoon character, or in the movie, may also be an interesting tom but not included in either. right here in coloring pages gallery of pics can also come in the form of animals or flowers. there is many more for instance plants or insects. they are that so interesting and provide different levels of difficulty. including giving a form was has never been encountered, is a form of learning also for children. to more difficult levels is part of the teenager or more. here in theme coloring pages pigs is one was you can choose and you try to finish. start with a base color or a higher level by playing gradations or color discs. anything become practicable when you begin to try it now. good luck hope this be be cheerful to you or maybe you can show it to friends or teacher or also parents. pig coloring pages printable free of the three little pigs cute baby peppa,coloring pages of pigs and piglets pig face page free 3 little three printable,guinea pig coloring pages printable page of a peppa pigs,pig coloring sheets free printable pages pigs three little is having colouring 3,cute coloring pages of baby pigs colouring three little story pig page free printable p is for,coloring pages for three little pigs houses pig page guinea cute peppa printable pdf free,coloring pictures of baby pigs free colouring pages three little printable image result for color page pig,free colouring pages three little pigs coloring guinea for houses baby pig page smells something,coloring pages baby pigs pig animal porky printable,free colouring pages three little pigs pig color kids printable minecraft coloring.
{ "redpajama_set_name": "RedPajamaC4" }
1,616
Q: What are some database persistence design patterns? Are there any standard design patterns for C++ database object persistence? Many thanks! A: Design Patterns for Data Persistence - The code samples are C#, but the concepts still apply.
{ "redpajama_set_name": "RedPajamaStackExchange" }
8,108
\section{Introduction} One of the most fruitful method of research of quantum integrable systems is algebraic Bethe ansatz. This method is developed by the Leningrad school (primarily L. Faddeev, L. Tahtadjian, and E. Sklyanin) \cite{FadST79,FadT79,FadLH96}. It is used for finding spectra and eigenvector of quantum Hamiltonians\cite{FadST79,FadT79,FadLH96}. Also it can be used for an efficient calculation of the form factors and correlation functions \cite{BogIK93L,KitMT00,KitKMST12,GohKS04}. Scalar products of Bethe vectors play a very important role in the algebraic Bethe ansatz. They are building blocks on the way of calculating form factors and correlation functions within this framework. The first results about scalar products ware obtained by A. Izergin and V. Korepin \cite{IzeK84,Ize87}. Then N. Slavnov found the determinant representation of the scalar products \cite{Sla89}. Later this result was formulated as the scalar product between on-shell off-shell Bethe vectors is proportional to determinant of matrix constructed by derivatives of the eigenvalues of generation function of Hamiltonians \cite{KitMT99}. M. Gaudin formulated a hypothesis about the norm of the Hamiltonian eigenfunction \cite{Gau72} and V. Korepin proved it for a wide class of quantum integrable models \cite{Kor82}. According to this hypothesis, the square of the eigenfunction norm is proportional to a Jacobian closely related to the Bethe equations. Usually algebraic Bethe ansatz connects quantum integrable systems with periodic boundary conditions. To describe systems with open boundary conditions Sklyanin introduced reflection algebra \cite{Skl88}. In the paper \cite{Kit07} one can find an expression of scalar products for spin chains with open boundary conditions. It was observed in \cite{KMST02} that the coefficients in the multiple action of transfer matrix onto Bethe vector have determinant form. The main result of this paper is connection of one coefficient in the action with Slavnov formula of scalar product of Bethe vectors. It allows us a new way of calculation of the Slavnov formula, which is very useful. Beside usual periodic case we repeat all result for open boundary case. This paper is organized as follows. We recall description of $RTT$-algebra and reflection algebra in section~\ref{Notations}. There we also give a shorthand notation for sets of parameters and products over those parameters. In section~\ref{periodic} we briefly describe of Bethe vectors within ABA approach. Then we formulate and prove the main theorem of the paper in periodic case. Separate subsection contains derivation of the determinant representation of scalar product of Bethe vector (also known as the Slavnov formula). Section~\ref{reflect} repeats the logic of the previous one for reflection boundary conditions case. \section{Defenition and notation}\label{Notations} \subsection{RLL-algebra} Let $R(u,v)$ be matrix associated with the vector representation of the Yangian $Y\left(\mathfrak{gl}_2\right)$: \begin{equation}\label{rmatrix} R(u,v) = \mathbf{1}\otimes\mathbf{1} + g(u,v) \mathbf{P}, \quad g(u,v) = \frac{c}{u-v}, \end{equation} where $\mathbf{1}$ is the identity matrix in $\mathbf{C}^2$, $\mathbf{P}$ is the permutation matrix in $\mathbf{C}^2\otimes\mathbf{C}^2$, $c$ is a constant, and and $u, v$ are arbitrary complex parameters called spectral parameters. The key object of algebraic Bethe ansatz is monodromy matrix $T(u)$. Matrix elements of $T(u)$ are operator valued and act in a Hilbert space $\mathcal{H}$ which is a physical space of a quantum model. The main property of $T(u)$ is $RTT$-relation: \begin{equation}\label{rtt} R(u,v) \; T^{(1)}(u) \, T^{(2)}(v) = T^{(2)}(v) \, T^{(1)}(u) \; R(u,v), \end{equation} where $T^{(i)}(u)$ acts in $i$-th tensor multiplier. The simple consequence of the equation \eqref{rtt} is that the monodromy matrix entries commute with themselves for different arguments $[T_{ij}(u),T_{ij}(v)]=0$. The transfer matrix is defined as the trace of the monodromy matrix \begin{equation}\label{tm} t(u) = \mathop{\rm tr} T(u). \end{equation} One can easy check that $ [ t(u), t(v) ] = 0$ \cite{FadST79,FadT79,FadLH96} . Thus it allows us identify transfer matrix with generating function of integral of motion of an integrable system. Such models usually have periodic boundary conditions. We assume that the space $\mathcal{H}$ contain vacuum vector $|0\rangle$ such that \begin{equation}\label{vac} \begin{aligned} T_{ii}(u) |0\rangle &= \lambda_i(u) |0\rangle, \\ T_{21}(u) |0\rangle &= 0. \end{aligned} \end{equation} Here $\lambda_i(u)$ are some complex-valued functions depending on concrete quantum integrable model. Multiple actions of $T_{12}$ with different arguments onto vacuum $|0\rangle$ generate a basis of quantum physical space $\mathcal{H}$ \cite{FadST79,FadT79,FadLH96} . \underline{Generalized model.} In the framework of this paper we assume that $\lambda_i(u)$ are free functional parameters and we do not specify any concrete their dependencies \cite{Kor82}. More or less it means that one can find concrete quantum integrable model for any specific choice of $\lambda_i(u)$. For our purposes it is enough to assume that $\lambda_i$ and their derivatives do not depend on their arguments in direct way only through the Bethe equations \eqref{be} for on-shell Bethe vectors. \subsection{Reflection algebra} To describe quantum spin chain with open boundary conditions Sklyanin introduced reflection algebra \cite{Skl88}. It is given by two following equations \begin{multline} R_{12}(u_1-u_2) K_-^{(1)}(u_1)R_{12}(u_1+u_2)K_-^{(2)}(u_2) = \\ K_-^{(2)}(u_2)R_{12}(u_1+u_2) K_-^{(1)}(u_1)R_{12}(u_1-u_2) \end{multline} and \begin{multline} R_{12}(-u_1+u_2) K_+^{(1)}(u_1)R_{12}(-u_1-u_2-2 c)K_+^{(2)}(u_2) = \\ K_+^{(2)}(u_2)R_{12}(-u_1-u_2-2 c) K_+^{(1)}(u_1)R_{12}(-u_1+u_2). \end{multline} In the framework of this paper we consider only diagonal solutions of reflection equation on $K_{\pm}$. General solution is \begin{equation} K_-(u) = K(u, \xi_-), \quad K_+(u) = K(u+c, \xi_+), \end{equation} where \begin{equation} K(u,\xi) = \begin{pmatrix} \xi + u & 0 \\ 0 & \xi - u \end{pmatrix}. \end{equation} Now we can introduce reflection monodromy matrix \cite{Skl88} \begin{equation} {\mathcal{T}} (u) = T(u) \, K\left(u - \frac{c}{2} ,\xi_-\right)\, \sigma_2 \, T^t(-u) \, \sigma_2 \end{equation} and reflection transfer matrix \begin{equation}\label{rtm} \hat{t}\, (u) = {\rm Tr}\; K\left(u + \frac{c}{2}, \xi_+ \right)\, {\mathcal{T}} (u). \end{equation} One can prove \cite{Skl88} that \begin{equation} [ \hat{t}\, (u) , \hat{t}\, (v) ] = 0 \end{equation} and $ \hat{t}\, (u) $ can be identify with generating function of Hamiltonians. We use conception of generalized model as for periodic case too. In this case we have the same vacuum as for periodic case \eqref{vac} with properties \begin{equation} \begin{aligned} \mathcal{T} _{11}(u) |0\rangle &= \left( u + \xi_- - \frac{c}{2} \right) \lambda_1(u)\lambda_2(-u) |0\rangle, \\ \mathcal{T} _{22}(u) |0\rangle &= \left(\frac{c-2u}{2u}\left( u- \xi_- + \frac{c}{2} \right) \lambda_1(-u)\lambda_2(u) + \frac{ u+ \xi_- - \frac{c}{2}}{2u} \lambda_1(u)\lambda_2(-u) \right) |0\rangle, \\ \mathcal{T} _{21}(u) |0\rangle &= 0. \end{aligned} \end{equation} Here $\lambda_i(u)$ are the same as in \eqref{vac}. Multiple actions of ${\mathcal{T}}_{12}$ with different arguments onto vacuum $|0\rangle$ generate a basis of quantum physical space $\mathcal{H}$ for models with open boundary case \cite{Skl88}. \subsection{Notation} In this paper we use our standard notation. Besides the function $g(u,v)$ in \eqref{rmatrix} we use different rational functions \begin{equation} \begin{aligned} f(u,v) &= \frac{u-v+c}{u-v}, \\ h(u,v) &= \frac{u-v+c}{c}. \end{aligned} \end{equation} We use a bar to denote sets of variables, for example, $\bar u$, $\bar v$. More precisely $\bar u = \{u_1, \ldots, u_n\}$, where $n$ is a number of elements in $\bar u$. Individual element of the set is denoted by Latin subscript, for instance, $u_i$ is an element of $\bar u$, $v_j$ is an element of $\bar v$ etc. A notation $\bar u_j$ is used for subsets complementary to the element $u_j$, that is, $\bar u_j = \bar u \backslash u_j$, $\bar v_i = \bar v \backslash v_i$, and so on. We use notation $-\bar u$ for set of variables with opposite sign $-\bar u = \{-u_1,\ldots, - u_n\}$ and notation $\bar u^2$ for set of squares of variables $\bar u^2 = \{u_1^2,\ldots, u_n^2\}$. Subsets of variables are denoted by roman indices: $\bar u_{\scriptscriptstyle \rm I}$, $\bar v_{\scriptscriptstyle \rm I\hspace{-1pt}I}$, and so on. In particular, we consider partitions of sets into subsets. Then the notation $\{ \bar u_{\scriptscriptstyle \rm I} , \bar u_{\scriptscriptstyle \rm I\hspace{-1pt}I} \} \vdash \bar u $ means that the set $\bar u$ is divided into two disjoint subsets $\bar u_{\scriptscriptstyle \rm I}$ and $\bar u_{\scriptscriptstyle \rm I\hspace{-1pt}I}$. We use a shorthand notation for products of the rational functions $g$, $f$, and $h$. Namely, if the rational functions have one or two sets of variables as argument, this means that one should take the product over all elements of the corresponding set (or the double product over elements of both sets). For example, \begin{equation} f(z,\bar u) = \prod_{u_i \in \bar u} f(z, u_i), \quad f(\bar u_{\scriptscriptstyle \rm I},\bar u_{\scriptscriptstyle \rm I\hspace{-1pt}I}) = \prod_{u_i \in \bar u_{\scriptscriptstyle \rm I}} \prod_{u_j \in \bar u_{\scriptscriptstyle \rm I\hspace{-1pt}I}} f(u_i, u_j). \end{equation} We use the same prescription for the products of commuting operators and their eigenvalues \begin{equation} T_{ij} (\bar u) = \prod_{u_k \in \bar u} T_{ij} (u_k), \quad t(\bar z) = \prod_{z_i \in \bar z} t(z_i) , \quad \lambda_{i} (\bar v) = \prod_{v_j \in \bar v} \lambda_{i} (v_j). \end{equation} Finally, by definition, any product over the empty set is equal to $1$. A double product is equal to $1$ if at least one of the sets is empty. We also use notation $\Delta ' (\bar u)$ and $\Delta (\bar u)$ \begin{equation} \Delta ' (\bar u) = \prod_{j<k} g(u_j,u_k), \quad \Delta (\bar u) = \prod_{j>k} g(u_j,u_k). \end{equation} \section{Periodic case}\label{periodic} \subsection{Bethe vector} Now we pass to the description of Bethe vectors. They belong to a Hilbert space $\mathcal{H}$, in which the operators $T_{ij}(u)$ act. We do not specify this space, however, according to the \cite{FadST79,FadT79,FadLH96} in the framework of generalized model off-shell Bethe vector for set $\bar u = \{u_1, \ldots, u_n\}$ is \begin{equation}\label{bv} {\mathbb{B}}(\bar u) = T_{12}(\bar u) |0\rangle = T_{12}(u_1)\ldots T_{12}(u_n) |0\rangle. \end{equation} Due to the commutation relations \eqref{rtt} the Bethe vector \eqref{bv} is symmetric over permutations of the parameters $\bar u$. One can prove that off-shell Bethe vector becomes eigenvector of transfermatrix $t(u)$ \eqref{tm} \begin{equation} t(z) \, {\mathbb{B}}(\bar u) = \tau(z | \bar u) \, {\mathbb{B}}(\bar u) \end{equation} if parameters $\bar u$ satisfy the system of Bethe equations \cite{FadST79,FadT79,FadLH96} \begin{equation}\label{be} \frac{\lambda_1(u_i)}{\lambda_2(u_i)} = \frac{f(u_i,\bar u_i)}{f(\bar u_i,u_i)}, \end{equation} with eigenvalue \begin{equation}\label{tme} \tau(z | \bar u) = \lambda_1 (z) f(\bar u,z) + \lambda_2(z) f(z,\bar u). \end{equation} Dual Bethe vectors belong to the dual space $\mathcal{H}^*$. They can be obtained as a transposition of the Bethe vectors \cite{FadST79,FadT79,FadLH96}. We denote dual Bethe vectors by ${\mathbb{C}}(\bar u)$. \begin{equation} {\mathbb{C}}(\bar u) = \langle 0 | T_{21}(\bar u), \end{equation} where $ \langle 0 |$ is dual vacuum from the dual space $\mathcal{H}^*$ with properties \begin{equation} \begin{aligned} \langle 0 | T_{ii}(u) &= \lambda_i(u) \langle 0 |, \\ \langle 0 | T_{12}(u) &= 0. \end{aligned} \end{equation} Here $\lambda_i(u)$ are the same as in \eqref{vac}. Dual Bethe vector becomes on-shell too, if the set $\bar u$ satisfy the system of Bethe equations \eqref{be} with the same eigenvalue \eqref{tme}. \subsection{Action of the transfer matrix and main theorem} In this subsection we present the main results of the paper for periodic boundary case. Here we give the theorem \ref{KAth} and its proof. This theorem allows us to calculate scalar product of on-shell and off-shell Bethe vectors, knowing the action of transfer matrix onto a Bethe vector. Let us recall \cite{FadST79,FadT79,FadLH96} the formula of action of transfer matrix \eqref{tm} onto Bethe vector \eqref{bv} \begin{multline} t(z) \; {\mathbb{B}}(\bar u) = \tau(z|\bar u) \,{\mathbb{B}}(\bar u) + \sum_j g(z, u_j)\left( \lambda_1(u_j) f(\bar u_j,u_j) - \lambda_2(u_j) f(u_j,\bar u_j)\right) {\mathbb{B}}(\{z, \bar u_j \}) = \\ \tau(z|\bar u) \,{\mathbb{B}}(\bar u) - \frac{1}{c} \sum_j g(z, u_j) \mathop{{\rm res}}\limits_{w=u_j} \tau(w|\bar u) \; {\mathbb{B}}(\{z, \bar u_j \}). \end{multline} We see that the action is given only by eigenvalue of transfer matrix $\tau$ \eqref{tme}. Taking into account this action we can formulate the main theorem of our paper. \begin{theorem}\label{KAth} Let us consider multiple action of a transfer matrix onto off-shell Bethe vector with $\#\bar u = \#\bar v$ \begin{equation}\label{act} t(\bar v) \; {\mathbb{B}}(\bar u) = \sum R(\bar v\,|\,\bar\eta_{\scriptscriptstyle \rm I},\bar\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}) \; {\mathbb{B}}(\bar\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}), \end{equation} where $\bar\eta=\{\bar u,\bar v\}$. Here the sum is taken over partitions $\bar\eta$ into two subsets $\bar\eta_{\scriptscriptstyle \rm I}$ and $\bar\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}$, such that $\#\bar\eta_{\scriptscriptstyle \rm I} = \#\bar\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}$. Then the coefficient of ${\mathbb{B}}(\bar v)$ of multiple action of the transfer matrixes with $\#\bar v=\#\bar u$ coincides with Slavnov formula \cite{Sla89} for the scalar product of on-shell--off-shell Bethe vectors \begin{equation}\label{kat1} S(\bar v | \bar u) \equiv {\mathbb{C}}(\bar v) {\mathbb{B}}(\bar u) = R(\bar v\,|\,\bar u ,\bar v). \end{equation} where set $\bar v$ satisfies Bethe equations and set $\bar u$ is free (off-shell). \end{theorem} {\sl Proof.} Let a set $\bar v$ satisfy Bethe equations, a set $\bar u$ be free (off-shell), and a set $\bar x$ be a set of free parameters. And let $\#\bar v=\#\bar u=\#\bar x$. Then one can consider an expression \begin{equation}\label{ctb1} {\mathbb{C}}(\bar v)\; T(\bar x)\; {\mathbb{B}}(\bar u). \end{equation} The left vector is the eigenvector of the transfer matrix \begin{equation}\label{eig} {\mathbb{C}}(\bar v)\; T(x) = \tau(x|\bar v) \;{\mathbb{C}}(\bar v). \end{equation} One can substitute equations \eqref{act} and \eqref{eig} in equation \eqref{ctb1} \begin{equation}\label{ctb2} \tau(\bar x|\bar v)\, {\mathbb{C}}(\bar v){\mathbb{B}}(\bar u) = \sum R(\bar x\,|\,\bar\eta_{\scriptscriptstyle \rm I},\bar\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}) \; {\mathbb{C}}(\bar v) {\mathbb{B}}(\bar\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}), \end{equation} where $\bar\eta =\{\bar x,\bar u\}$. And now one can put $\bar x \to \bar v$ in equation \eqref{ctb2}. \begin{multline} \lim_{x\to v_i} \tau(x \,|\,\bar v) = - c\, \lambda_1'(v_i) f(\bar v_i,v_i) + c\, \lambda_2'(v_i) f(v_i,\bar v_i) + \\ \lambda_1(v_i) f(\bar v_i,v_i)\left(1-\sum_{j\ne i} t(v_j,v_i)\right)+\lambda_2(v_i) f(v_i,\bar v_i)\left(1+\sum_{j\ne i} t(v_i,v_j)\right) = \\ - c\, \lim_{x\to v_i} \tau(x \,|\,\bar v) = - c\, \lambda_1'(v_i) f(\bar v_i,v_i) + c\, \lambda_1(v_i) f(\bar v_i,v_i) X(v_i) + UT, \end{multline} where $X(v_i) = \left. \ln\left(\frac{\lambda_1(z)}{\lambda_2(z)}\right)'\right|_{z=v_i}$ and $UT$ does not contain derivatives. It is clear that $S(\bar v|\{\bar v_{\scriptscriptstyle \rm I},\bar u_{\scriptscriptstyle \rm I\hspace{-1pt}I}\})$ can contain only $X(v_i)$ with $v_i \in \bar v_{\scriptscriptstyle \rm I}$. Let us compare coefficients of $X(\bar v)$ in equation \eqref{ctb2}. In RHS one can get $X(\bar v)$ from the norm of Bethe vector $S(\bar v|\bar v)$. An explicit expression one can find in the \cite{Gau72,Kor82}, but here we use only coefficient of $X(\bar v)$ \begin{multline} (-c)^{\# \bar v} \lambda_1(\bar v) \prod_{i\ne j} f(v_i,v_j) X(\bar v)\; S(\bar v|\bar u)+\dots = R(\bar v\,|\,\bar u ,\bar v)\, S(\bar v | \bar v) + \dots = \\ R(\bar v\,|\,\bar u ,\bar v) (-c)^{\#\bar v} \lambda_2(\bar v) \prod_{i\ne j} f(v_i,v_j) ( X(\bar v) + \dots ) + \dots. \end{multline} Taking into account a consequence of the Bethe equations $\lambda_1(\bar v) = \lambda_2(\bar v)$ we instantly come to the equation \eqref{kat1}. \hfill\nobreak\hbox{$\square$}\par\medbreak \subsection{Slavnov determinant formula} In this section we give a new way to get determinant representation of scalar product. It is based on theorem \ref{KAth}. Using this theorem one can writhe the following formula \begin{equation}\label{sym-bu} S(\bar x | \bar u) = \frac{1}{c^n}\mathop{{\rm Sym}}\limits_{\bar u} \; \prod_{i=1}^{n} g(u_i,x_i) \mathop{{\rm res}}\limits_{z=u_1} \tau(z|\bar u) \mathop{{\rm res}}\limits_{z=u_2} \tau(z|\{x_1,\bar u_1\}) \ldots \mathop{{\rm res}}\limits_{z=u_n} \tau(z|\{\bar x_n,u_n\}), \end{equation} where $\#\bar u=\#\bar x=n$. Let us introduce \be{jjk} J(u_j,x_k)=\frac{c}{g(u_j,\bar x)}\frac{\partial \tau(u_j|\bar x)}{\partial x_k}= \lambda_2(u_j)t(u_j,x_k)h(u_j,\bar x)+(-1)^{n-1}\lambda_1(u_j)t(x_k,u_j)h(\bar x,u_j). \end{equation} Let $1\le m\le n$. Define \be{nu} \nu_k^{(m)}=\frac{\prod_{\substack{\ell=1\\ \ell\ne k}}^mg(x_k,x_\ell)}{\prod_{\ell=1}^m g(x_k,u_\ell)}. \end{equation} Using a standard contour integral method one can easily prove that \be{sumJm} \sum_{k=1}^m J(u_j,x_k)\nu_k^{(m)}=-\frac{\mathop{\rm res}\limits_{z=u_j}\tau(z|u_1,\dots,u_m,x_{m+1},\dots,x_n)} {c \prod_{\substack{i=1\\ i\ne j}}^m g(u_j,u_i)\prod_{\ell=m+1}^n g(u_j,x_\ell)}, \end{equation} for $1\le j\le m$. In particular, setting $m=1$ in \eqref{sumJm} we obtain \be{sumJm1} \tfrac1c\; g(u_1,x_1)\mathop{\rm res}\limits_{z=u_1}\tau(z|u_1,\bar x_{1})= J(u_1,x_1) g(u_1,\bar x_1). \end{equation} This formula, of course, can be also obtained via direct calculation. Let us make a replacement $u_j\to u_{n+1-j}$ and $x_j\to x_{n+1-j}$. Then equation \eqref{sym-bu} takes the form \begin{equation}\label{sym-bu1} S(\bar x | \bar u) = \frac{1}{c^n}\mathop{{\rm Sym}}\limits_{\bar u} \; \prod_{i=1}^{n} g(u_i,x_i) \mathop{{\rm res}}\limits_{z=u_i} \tau(z|\{u_1,\dots,u_i,x_{i+1},\dots,x_n\}). \end{equation} Now one can formulate theorem about determinant representation of scalar product. \begin{theorem}\label{T-HNY} Let $1\le m\le n$. Then scalar product between on-shell and off-shell Bethe vectors is \begin{multline}\label{HNY} S(\bar x | \bar u) = \frac{1}{c^{n-m}}\mathop{{\rm Sym}}\limits_{\bar u} \; \prod_{i=1}^{m}\left\{ J(u_i,x_i)\prod_{\ell=m+1}^{n} g(u_i,x_\ell)\right\} \Delta'(\{u_1,\dots,u_m\})\Delta(\{x_1,\dots,x_m\}) \\ % \times \prod_{\ell=m+1}^{n}g(u_\ell,x_\ell) \mathop{{\rm res}}\limits_{z=u_\ell} \tau(z|\{u_1,\dots,u_\ell,x_{\ell+1},\dots,x_n\}). \end{multline} \end{theorem} {\sl Proof.} We use induction over $m$. For $m=1$ the statement of the theorem is obvious. Indeed, using \eqref{sumJm1} we transform \eqref{sym-bu1} to the form \begin{equation}\label{HNY-m1} S(\bar x | \bar u) = \frac{1}{c^{n-1}}\mathop{{\rm Sym}}\limits_{\bar u} \; J(u_1,x_1)g(u_1,\bar x_1) % \prod_{\ell=2}^{n}g(u_\ell,x_\ell) \mathop{{\rm res}}\limits_{z=u_\ell} \tau(z|\{u_1,\dots,u_\ell,x_{\ell+1},\dots,x_n\}). \end{equation} This equation coincides with \eqref{HNY} at $m=1$. Let \eqref{HNY} be valid for some $m\ge 1$. Then due to \eqref{sumJm} we have \begin{multline}\label{sumJmp1} \mathop{\rm res}\limits_{z=u_{m+1}}\tau(z|u_1,\dots,u_{m+1},x_{m+2},\dots,x_n)= -c \prod_{i=1}^{m} g(u_{m+1},u_i)\prod_{\ell=m+2}^n g(u_{m+1},x_\ell)\\ \times\sum_{k=1}^{m+1} J(u_{m+1},x_k)\nu_k^{(m+1)}, \end{multline} where $\nu_k^{(m+1)}$ are given by \eqref{nu} with $m$ replaced by $m+1$. Substituting this into \eqref{HNY} we find \begin{multline}\label{HNY-mm} S(\bar x | \bar u) = \frac{(-1)^{m+1}}{c^{n-m-1}}\mathop{{\rm Sym}}\limits_{\bar u} \; \Delta'(\{u_1,\dots,u_{m+1}\})\Delta(\{x_1,\dots,x_m\}) \prod_{i=1}^{m}J(u_i,x_i) \prod_{i=1}^{m+1}\prod_{\ell=m+1}^{n} g(u_i,x_\ell) \\ % \times \sum_{k=1}^{m+1} J(u_{m+1},x_k)\nu_k^{(m+1)} \prod_{\ell=m+2}^{n}g(u_\ell,x_\ell) \mathop{{\rm res}}\limits_{z=u_\ell} \tau(z|\{u_1,\dots,u_\ell,x_{\ell+1},\dots,x_n\}). \end{multline} It is easy to see that the only term corresponding to $k=m+1$ survives in the sum over $k$ in the second line of \eqref{HNY-mm}. Indeed, for all $k<m+1$ we obtain expressions of the form \be{e-form} J(u_{m+1},x_k)J(u_{k},x_k)\cdot G(u_k,u_{m+1}),\qquad\text{where}\qquad G(u_k,u_{m+1})= - G(u_{m+1},u_k). \end{equation} Such the contribution vanishes under symmetrization over $\bar u$. Thus, we obtain \begin{multline}\label{HNY-mm1} S(\bar x | \bar u) = \frac{(-1)^{m+1}}{c^{n-m-1}}\mathop{{\rm Sym}}\limits_{\bar u} \Delta'(\{u_1,\dots,u_{m+1}\})\Delta(\{x_1,\dots,x_m\}) \prod_{i=1}^{m+1}J(u_i,x_i) \\ % \times \nu_{m+1}^{(m+1)}\prod_{i=1}^{m+1}\prod_{\ell=m+1}^{n} g(u_i,x_\ell) \prod_{\ell=m+2}^{n}g(u_\ell,x_\ell) \mathop{{\rm res}}\limits_{z=u_\ell} \tau(z|\{u_1,\dots,u_\ell,x_{\ell+1},\dots,x_n\}), \end{multline} and substituting here \be{nu1} \nu_{m+1}^{(m+1)}=\frac{\prod_{\ell=1}^{m} g(x_{m+1},x_\ell)}{\prod_{\ell=1}^{m+1} g(x_{m+1},u_\ell)}, \end{equation} we immediately arrive at \eqref{HNY} with $m\to m+1$.\hfill\nobreak\hbox{$\square$}\par\medbreak \vspace{3mm} Setting $m=n$ in \eqref{HNY} we obtain \begin{equation}\label{HNY-n} S(\bar x | \bar u) = \mathop{{\rm Sym}}\limits_{\bar u} \;\Delta'(\bar u)\Delta(\bar x) \prod_{i=1}^{n}J(u_i,x_i), \end{equation} and thus, \begin{equation}\label{HNY-res} S(\bar x | \bar u) = \Delta'(\bar u)\Delta(\bar x)\det_n J(u_i,x_k). \end{equation} We can rewrite equation \eqref{HNY-res} in the following form \begin{equation} S(\bar x | \bar u) = \Delta'(\bar u)\Delta(\bar x)\det_n \frac{c}{g(u_j,\bar x)}\frac{\partial \tau(u_j|\bar x)}{\partial x_k}. \end{equation} and it is also known as Slavnov determinant formula \cite{Sla89} for scalar product between on-shell and off-shell Bethe vectors. \section{Reflection case}\label{reflect} \subsection{Bethe vector} Now we pass to the description of Bethe vectors in onen boundary case. They belong to the same Hilbert space $\mathcal{H}$. Basic of the space is generated by multiple action of ${\mathcal{T}}_{12}$ onto vacuum $|0\rangle$. According to the \cite{Skl88} in the framework of generalized model off-shell Bethe vector for set $\bar u = \{u_1, \ldots, u_n\}$ is \begin{equation}\label{rbv} {\hat{\mathbb{B}}}(\bar u) = {\mathcal{T}}_{12}(\bar u) |0\rangle = {\mathcal{T}}_{12}(u_1)\ldots {\mathcal{T}}_{12}(u_n) |0\rangle. \end{equation} It is easy to prove that the Bethe vector \eqref{rbv} is symmetric over permutations of the parameters $\bar u$ \cite{Skl88}. One can prove that off-shell Bethe vector becomes eigenvector of transfermatrix $\hat t (u)$ \eqref{rtm} \begin{equation} \hat t(z) \, {\hat{\mathbb{B}}}(\bar u) = \hat\tau(z | \bar u) \, {\hat{\mathbb{B}}}(\bar u) \end{equation} if parameters $\bar u$ satisfy the system of Bethe equations \cite{Skl88} \begin{equation}\label{rbe} \frac{\left(u_i + \xi_+ - \frac{c}{2} \right)\left(u_i + \xi_- - \frac{c}{2} \right) }{\left(u_i - \xi_+ + \frac{c}{2} \right)\left(u_i - \xi_- + \frac{c}{2} \right)} \frac{\lambda_1(u_i)\lambda_2(-u_i)}{\lambda_1(-u_i)\lambda_2(u_i)} = \frac{f(u_i,\bar u_i)f(u_i, -\bar u_i)}{f(\bar u_i,u_i)f(-\bar u_i,u_i)}, \end{equation} with eigenvalue \begin{equation}\label{rtme} \begin{aligned} \hat\tau(z | \bar u) = & \frac{2z + c}{2z}\left( z + \xi_+ - \frac{c}{2} \right)\left( z + \xi_- - \frac{c}{2} \right) f(\bar u,z)f(-\bar u,z) \lambda_1 (z)\lambda_2 (-z) + \\ & \frac{2z - c}{2z}\left( z - \xi_+ + \frac{c}{2} \right)\left( z - \xi_- + \frac{c}{2} \right) f(z,\bar u)f(z,-\bar u) \lambda_1 (-z)\lambda_2 (z). \end{aligned} \end{equation} Dual Bethe vectors belong to the dual space $\mathcal{H}^*$. They can be obtained as a transposition of the Bethe vectors \cite{Skl88}. We denote dual Bethe vectors by ${\hat{\mathbb{C}}}(\bar u)$. It can be constructed in the similar way \cite{Skl88} using dual vacuum $ \langle 0 |$ from the dual space $\mathcal{H}^*$ with the properties \begin{equation} \begin{aligned} \langle 0 | \mathcal{T} _{11}(u) &= \left( u + \xi_- - \frac{c}{2} \right) \lambda_1(u)\lambda_2(-u) \langle 0 |, \\ \langle 0 | \mathcal{T} _{22}(u) &= \left(\frac{c-2u}{2u}\left( u- \xi_- + \frac{c}{2} \right) \lambda_1(-u)\lambda_2(u) + \frac{ u+ \xi_- - \frac{c}{2}}{2u} \lambda_1(u)\lambda_2(-u) \right) \langle 0 |, \\ \langle 0 | \mathcal{T} _{12}(u)&= 0. \end{aligned} \end{equation} Dual Bethe vector becomes on-shell too, if the set $\bar u$ satisfy the system of Bethe equations \eqref{rbe} with the same eigenvalue \eqref{rtme}. \subsection{Reflection main theorem and Slavnov determinant formula} In this subsection we present the main results of the paper for open boundary boundary case (case of reflection algebra). Here we give two theorem \ref{rKA} and \ref{rSl}. Proofs of the theorems are the same as in periodic case. This theorem allows us to calculate scalar product of on-shell and off-shell Bethe vectors for open boundary boundary case, knowing the action of the reflection transfer matrix \eqref{rtm} onto a Bethe vector \eqref{rbv}. Let us recall the formula of action of reflection transfer matrix \eqref{rtm} onto Bethe vector \eqref{rbv} \cite{Skl88} \begin{equation} \hat t(z) \; {\hat{\mathbb{B}}}(\bar u) = \hat\tau(z|\bar u) \,{\hat{\mathbb{B}}}(\bar u) - \frac{1}{c} \sum_j \frac{2 u_i}{c + 2 u_i} \frac{c + 2 z}{c} g(z, u_j) g(z,- u_j) \mathop{{\rm res}}\limits_{w=u_j} \hat\tau(w|\bar u) \; \hat{\mathbb{B}}(\{z, \bar u_j \}). \end{equation} We see that the action is given only by eigenvalue of transfer matrix $\hat\tau$ \eqref{rtme}. \begin{theorem}\label{rKA} Let us consider multiple action of a transfer matrix onto off-shell Bethe vector with $\#\bar u = \#\bar v$ \begin{equation} \hat t (\bar v) \; {\hat{\mathbb{B}}}(\bar u) = \sum \hat R(\bar v\,|\,\bar\eta_{\scriptscriptstyle \rm I},\bar\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}) \; {\hat{\mathbb{B}}}(\bar\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}), \end{equation} where $\bar\eta=\{\bar u,\bar v\}$. Here the sum is taken over partitions $\eta$ into two subsets $\eta_{\scriptscriptstyle \rm I}$ and $\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}$, such that $\#\eta_{\scriptscriptstyle \rm I} = \#\eta_{\scriptscriptstyle \rm I\hspace{-1pt}I}$. Then the coefficient of ${\hat{\mathbb{B}}}(\bar v)$ of multiple action of the transfer matrixes with $\#\bar v=\#\bar u$ coincides with the scalar product of on-shell--off-shell Bethe vectors \begin{equation}\label{kat1} S(\bar v | \bar u) =\hat R(\bar v\,|\,\bar u ,\bar v). \end{equation} where set $\bar v$ satisfies Bethe equations and set $\bar u$ is free (off-shell). \end{theorem} Proof of the theorem is absolutely the same as for periodic case. Using way of derivation similar to periodic case one can get determinant representation of scalar product between eigenvector of reflection transfer matrix \eqref{rtm} and off-shell Bethe vector \eqref{rbe} \begin{theorem}\label{rSl} Let us consider multiple action of a transfer matrix onto off-shell Bethe vector with $\#\bar u = \#\bar v$ \begin{equation}\label{rSl} S(\bar x | \bar u) = \Delta'(\bar u^2)\Delta(\bar x^2)\det_n \frac{2 x_k - c}{2 x_k}\frac{c}{ (2 u_j -c)} \frac{c }{g(u_j,\bar x)g(u_j,-\bar x)}\frac{\partial \tau(u_j|\bar x)}{\partial x_k}. \end{equation} \end{theorem} One can compare formula \eqref{rSl} with \cite{Kit07}. \section*{Conclusion} In this paper we formulate and prove relation between action of the transfer matrix onto Bethe vectors and scalar product of Bethe vectors. The main results of the paper are theorems \ref{KAth} and \ref{rKA}. This way certainly is more direct and simple than the methods used before for the derivation of the scalar product of on-shell and off-shell Bethe vectors. This result completely matches to the well-known Slavnov determinant formulas in periodic and reflection boundary cases. The last one was not known in general case of reflection algebraic Bethe ansatz but only in spin chain case. In both cases we have determinants of matrices constructed by derivatives of the eigenvalues of transfer matrices. This circumstance was known for a long time, but without explanation. We give this explanation. Action of the transfer matrix onto Bethe vector is given only by the function of eigenvalue of transfer matrix \eqref{tme}. And taking to account our results it is natural to expect that the scalar product is given only by the eigenvalue \eqref{tme} too. There is a direct generalization of the our relation to higher rank algebraic Bethe ansatz (also known as nested Bethe ansats). For example, in the case of $\mathfrak{gl}_3$ algebra in terms of paper \cite{Bel12} action the transfer matrix is closely related with scalar product too. More precisely, in the multiple action of the transfer matrix onto Bethe equation \begin{equation} t(\bar{u}^{\scriptscriptstyle C}) t(\bar{v}^{\scriptscriptstyle C}) {\mathbb{B}}(\bar{u}^{\scriptscriptstyle B},\bar{v}^{\scriptscriptstyle B}) \end{equation} coefficient of the ${\mathbb{B}}(\bar{u}^{\scriptscriptstyle C},\bar{v}^{\scriptscriptstyle C})$ can be identified with the scalar product of on-shell and off-shell Bethe vectors as in the current paper. A proof of this statement is the similar to the proof for $\mathfrak{gl}_2$ given in this work. But at the moment there is no determinant representation for such scalar product. We hope that out approach will help to get such determinant representation even for higher rank case in the future. \section*{Acknowledgments} We are grateful to Arthur Hutsalyuk for inspiring discussions and to Nikita Slavnov for fruitful discussions. The work of A.L. has been funded.
{ "redpajama_set_name": "RedPajamaArXiv" }
4,374
Children and Youth in History Rules of the Thälmann Pioneers (20th c) Elizabeth A. Ten Dyke, Ph.D., Vice Principal J Watson Bailey Middle School Why I Taught the Source National and state level world history curricula include study of 20th-century political and economic regimes, namely communism and capitalism. As students learn about these political ideologies and economic systems, they often embrace black and white thinking. They are easily persuaded that communist societies were (are) totalitarian and unjust while capitalist societies were (are) democratic and just. This thinking may be supported by learning about the activities of state security agencies, secret police, and other arms of the state that sustained communist regimes through intimidation and terror. Under these circumstances, though, how can students explain the trauma and confusion many people of central and eastern Europe experienced after the collapse of communism in 1989? How can they make sense of the continuing appeal of communist parties and ideals in many central and eastern European societies? By examining a source that focuses on youth organizations in East Germany, students can develop a more nuanced view of communist society. It helps students see that communist organizations offered positive experiences that supported values they share. This helps students develop empathy for individuals who lived under communism, and paves the way for understanding the enthusiasm with which some individuals lived in communism, and the ambivalence many experienced at its end. How I Introduce the Source To introduce this topic, I ask students to identify the youth groups that exist in our democratic society (Girl Scouts, Boy Scouts, 4-H, community-based organizations, etc.). We talk about why young people participate in the organizations (have fun, be with friends). We talk about the symbols and rituals of the organizations (uniforms, pledges, moving up ceremonies, bronze/silver/gold awards in Girl Scouts, Eagle Scout award in Boy Scouts) and their social meanings. We also talk about the kinds of activities young people participate in through the groups. I explain that communist societies had similar organizations in addition to the official organizations of the communist party with which students are already somewhat familiar. I explain that in the German Democratic Republic, elementary and middle school students (using American terminology) belonged to a group called the Ernst Thälmann Pioneers, while students in high school and older (again, using American terminology) belonged to a group called the Young Pioneers. Children generally joined the organizations through school. They received uniforms, met regularly, and participated in activities including recreation, community service, day trips, camping and other excursions. Reading the Source The source is set up almost like a catechism, in a "call and response" format. I have one or more students take turns reading the statements in bold type, while the other students in the class respond out loud with the statements in regular type. After reading the source out loud, the students work in groups to summarize its main ideas and to discuss questions such as: What are the values conveyed by the source (pride, respect, discipline, cleanliness)? Who or what should Thälmann Pioneers love, respect, or honor (their parents, the Soviet Union, their bodies, health)? What is the main symbol of the Thälmann Pioneers, and what does it stand for (the red scarf, the working class). What other themes or main ideas can the students find here (opposing imperialists, loving peace, study and learning, etc). I have students chart or list their observations, then comment on them. What, if anything, do they agree with? What, if anything, do they disagree with? To reflect, I ask students to return to the conversation we had prior to our study of the source, and think again about youth groups in the United States. I ask, rhetorically, who wouldn't want to be a part of an organization that all your friends were a part of, that conveyed such positive values, and offered fun activities and service? Students overwhelmingly embrace the idea of the Thälman Pioneers. Then we discuss some of the differences between American youth groups and comparable organizations in communist societies. The first divergence is the intense social and political pressure to join, as well as disadvantages experienced by those who did not participate (being left out, ostracism, pressures felt by parents at work, limits placed on educational and professional opportunities later in life). Second, youth groups in communist societies had a strong political dimension (refer to the students' analysis for examples). Third, for these reasons, communist youth groups served to coordinate and indoctrinate youth in values and ideals that supported the state regime. Students begin to see that the picture is no longer black and white—even among East German youth. The youth groups offered fun, social experiences for young people—yet participation was almost compulsory. In addition, it was very hard for many people to experience the problems of daily life in communist society and at the same time embrace the positive beliefs included in propaganda such as this. Many young people participated in communist youth groups with a high degree of cynicism, which, in the long run, served to undermine the state. I find students are very responsive to the source and the conversations we have about it. They appreciate the insight into the daily life of young people in a communist society. Many become curious about this and other youth organizations. They may make comparisons to youth groups under National Socialism, if they learned about them previously. This is a rich, interesting activity for high school students that sharpens critical thinking by drawing upon an already familiar aspect of youth culture. "Rules of the Thälmann Pioneers (20th c)," in Children and Youth in History, Item #21, https://chnm.gmu.edu/cyh/items/show/21 (accessed January 29, 2022). Communism, German Democratic Republic, Government Documents, Official Documents, Soviet Union, Thälmann Pioneers, The World Since 1945, View Primary Source: The Rules of the Thälmann Pioneers [Organization Document] Content on this site by Center for History and New Media and the University of Missouri-Kansas City with funding from the National Endowment for the Humanities is licensed under a Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0). Request permissions beyond the scope of this license here.
{ "redpajama_set_name": "RedPajamaCommonCrawl" }
3,059
\section{Introduction} \label{sec:intro} In semiclassical theories of gravity, the back-reaction of quantum matter fields is studied by means of the so-called semiclassical Einstein equations, where the Einstein tensor of a four-dimensional spacetime $({\mathcal M},g)$ is equated to the expectation value of the stress-energy tensor of matter fields evaluated on a quantum state $\omega$. Any couple formed by a spacetime and a quantum state satisfying these equations constitutes a solution in semiclassical gravity. The question about existence and uniqueness of solutions was analyzed in several physical scenarios and with different approaches, which take advantage of recent developments in the study of locally covariant quantum field theories on globally hyperbolic spacetimes \cite{Dappiaggi2008cosm,Pinamonti2011init,Pinamonti2015glob,Hack2016cosm,Juarez-Aubry2019desitt,Juarez-Aubry2021stat,Sanders2020stat,Gottschalk2021cosm,Meda2021exi,Gottschalk2022sitter,Gottschalk2022num,Juarez-Aubry2022conf} (see also \cite{Meda2021evap} for an application of the semiclassical Einstein equations in black hole evaporation). In this framework, the renormalization of quadratic fields like the Wick square or the stress-energy tensor $T_{\mu\nu}$ entering the semiclassical models is always guaranteed when sufficiently regular states are taken into account. This is the case of Hadamard states, which fulfil the microlocal spectrum condition and possess two-point functions which have an universal singular structure \cite{Radzikowski1996micro,Brunetti1996wick}. Hence, in Hadamard states Wick observables can be constructed by a covariant point-splitting regularisation which removes the universal singular part, and generalises the usual normal-ordering prescription. For further details about this topic we refer to \cite{Wald1977back,Kay1991uniq,Brunetti2000micro,Hollands2001wick,Hollands2002top,Hollands2005stress} (see also \cite{Hollands2015qft} for a recent review and some physical applications). The study of the stability of solutions of semiclassical equations remains problematic even on this class of states and even at linear order, because of the presence of higher-order derivative terms in the expectation value of the quantum stress-energy tensor. In the past years, the issue about stability of solutions of the semiclassical Einstein equations was addressed in several works \cite{Horowitz1978inst,Horowitz1980wf,Kay1981see,Yamagishi1982inst,Jordan1987stab,Suen1989unst,Matsui20inst}. It was argued that, because of the higher order derivative terms, linearized semiclassical Einstein equations around chosen backgrounds admit exponentially growing solutions. These exponentially growing linear perturbations are called runaway solutions and indicate that the chosen background is unstable. It is remarkable that runaway solutions are present even on flat spacetimes. A prescription of reduction of order was presented in \cite{Simon1991stab,Parker1993red,Flanagan1996back} to eliminate runaway unstable solutions, whereas a criterion for the validity of semiclassical gravity in the linear regime was proposed in \cite{Anderson2003lr}. More recently, the stability of semiclassical solutions has been treated in the framework of the so-called stochastic gravity (see \cite{Hu2020stoc} and references therein). The irregularity issues of the semiclassical Einstein equations were also deeply studied in \cite{Meda2021exi} in the case of cosmological spacetimes, for arbitrary values of the coupling-to-curvature parameter. In this case, higher order derivatives of the metric appear, and furthermore the expectation value of the traced stress-energy tensor contains a non local quantum contribution at the linear order in the perturbative potential. This term represents the source of instability of the model, and it forbids to solve the semiclassical equations in a direct way. However, it was proved that unique local solutions exist after applying to the semiclassical equation an inversion formula associated to that unbounded operator. \bigskip In this paper, we shall take inspiration from the ideas presented in \cite{Randjbar-Daemi980ren} and \cite{Juarez-Aubry2020init}, and we analyse the stability issue at linear order in a simple semiclassical toy model in flat spacetime, consisting of a classical background scalar field $\psi$ coupled to a quantum free scalar field $\phi$. The system of equations governing the dynamics is displayed in eq.\@ \eqref{eq:system-semicl}, and the back-reaction of the quantum field $\phi$ on $\psi$ is estimated by substituting the classical field $\phi^2$ with $\expvalom{\phi^2}$, the expectation value of the quantum Wick square in the quantum state $\omega$. With this picture in mind, we are interested in analyzing the stability of the solutions of this semiclassical system against linear perturbations. To this end, we discuss the equation obtained by linearizing the semiclassical equations over full solutions, which are formed by a quantum state $\omega$ for the quantum field $\phi$ and by a classical background $\psi_0$ for the classical field $\psi$. For the sake of simplicity, in this toy model, the state which is chosen is the Minkowski vacuum, however, similar results hold with other choices for $\omega$. The obtained equation \eqref{eq:SCE-psi1} is a linear equation for the perturbations $\psi_1$ over the background classical field $\psi_0$, and in some cases it gives origin to a well posed initial value problem for spatially compact solutions despite the presence of certain unavoidable non local contributions in it. These non local contributions are manifest in the linear equation for $\psi_1$ written in eq.\@ \eqref{eq:SCE-eq1-hom}, or in the corresponding non-homogeneous equation \eqref{eq:SCE-eq1-source} obtained when a smooth compactly supported source $f$ for $\psi_1$ is considered. In a first step, we prove that this equation is of hyperbolic nature. We then explicitly construct the retarded fundamental solutions $D_R$. Out of it we write the most general retarded solution of eq.\@ \eqref{eq:SCE-eq1-source}, we discuss the well posedness of the corresponding initial value problem, and we study the decay of these solutions for large times. Notably, there are several choices of the parameters governing the dynamics for which every linearized solution having smooth compactly supported initial values, or emerging from a compactly supported source, decays polynomially in time for large times, thus showing that perturbations tend to disperse, or better to disappear in time. This result of stability is ensured by both the non-vanishing mass of the quantum field, by the spatial support of the initial data and the source. On the contrary, in agreement with previous observations \cite{Horowitz1978inst,Horowitz1980wf,Jordan1987stab}, if the quantum field $\phi$ is massless, then solutions of the linearized semiclassical equation which grow exponentially in time may always exist, even if the initial values and the source are of compact support. The toy model studied in this paper formally mimics both the cosmological semiclassical model studied in \cite{Meda2021exi} and the weak field theory discussed in \cite{Horowitz1980wf}, after interpreting the background solution as a degree of freedom of the metric, and the coupling constants as the renormalization freedoms appearing in a semiclassical theory of gravity. This analogy indicates that such semiclassical theories may have stable solutions at least when the quantum fields describing matter are massive. \bigskip This paper is organized as follows, in the \hyperref[sec:model]{Section 2} we introduce the interacting toy model considered in this paper, and we discuss how to obtain a linearized semiclassical equation governing the back-reaction of the quantum field on the background. In the first part of the \hyperref[sec:stability]{Section 3}, we analyze the linearized equation, we construct its retarded fundamental solution and we show how to formulate a well-posed initial value problem out of it. In the last part of this section, we present a correspondence between the presented toy model and the semiclassical cosmological problem. Finally, the main results of the paper are summarized in \hyperref[sec:conc]{Section 4}. \hyperref[sec:appendix]{Appendix A} contains some technical results about the decay at large time of certain functions. \section*{Notations} In this paper, the units convention is $G = c = 1$, and the Lorentzian signature of the spacetime $({\mathcal M},g)$ is $(-,+,+,+)$. Thus, the d'Alembert operator reads $\square_x = g^{\mu\nu} \nabla_\mu \nabla_\nu = -\partial^2_t + \Delta_\vec{x}$, where $\Delta_\vec{x}$ denotes the spatial Laplace-Beltrami operator. We shall employ the following conventions on the Fourier transforms: \begin{gather*} \hat{f}(p) = {\mathcal F}\{f\}(p) = \int_{\mathbb{R}^4} f(x) \mathrm{e}^{i p \cdot x} \mathrm{d}^4 x, \qquad f(x) = {\mathcal F}^{-1}\{\hat{f}\} = \frac{1}{(2\pi)^4} \int_{\mathbb{R}^4} \hat{f}(p) \mathrm{e}^{-i p \cdot x}\mathrm{d}^4 p, \\ \tilde{f}(t,\vec{p}) = {\mathcal F}_{\vec{x}}\{f\}(t,\vec{p}) = \int_{\mathbb{R}^3} f(t,\vec{x}) \mathrm{e}^{i \vec{p} \cdot \vec{x}} \mathrm{d}^3 \vec{x}, \end{gather*} respectively. Hence, the convolution theorems read as ${\mathcal F} \{ f * g \} = {\mathcal F} \{f\} {\mathcal F}\{ g\}$ where $*$ denotes the convolution operator such that $f * g = \int f(x-y) g(y) \mathrm{d} y$. \section{The semiclassical model} \label{sec:model} In this work, we study the coupling between a quantum scalar field $\phi$ and another classical background scalar field $\psi$ in the Minkowski spacetime $({\mathcal M},\eta)$. The equations of motion of the corresponding free theory are \begin{subnumcases}{} \square \phi - m^2 \phi = \lambda \psi \phi, \label{eq:system-eq1} \\ g_2 \square \psi - g_1 \psi = \lambda_1 \phi^2 -\lambda_2 \square \phi^2, \label{eq:system-eq2} \end{subnumcases} where $g_1,g_2,\lambda,\lambda_1,\lambda_2$ denote the real coupling constants of the theory. With the choice of $\lambda_2=0$, the system is Lagrangian. On the other hand, if $\lambda_2 \neq 0$, then there is sufficient freedom in the definition of the coupling constants to fix $\lambda_2 = 1$. However, we do rather not impose any further constraint on the coupling constants, in order to describe as many semiclassical theories as possible. In particular, keeping $\lambda_2 \neq 0$, eq.\@ \eqref{eq:system-eq2} mimics the form of equations arising in the study of semiclassical Einstein equations, as we shall discuss in \hyperref[sec:cosmo]{subsection 3.4}. The first equation \eqref{eq:system-eq1} is the equation of motion of the linear, Klein-Gordon like massive field $\phi$, where $m$ is the mass, and $\lambda \psi$ plays the role of an external potential. The quantization of $\phi$ follows straightforwardly once the external potential field $\psi$ is known. To this avail, we shall consider the unital $*-$algebra of field observables $\mathcal{A}$ generated by the identity $\mathbb{I}$ and the abstract smeared field $\phi(f)$, with $f \in C^{\infty}_0({\mathcal M})$ any compactly supported smooth function \cite{Brunetti2015aqft,Haag2012local}. The product in this $*-$algebra encodes the canonical commutation relations \[ [\phi(f_1),\phi(f_2)] = i \hbar \Delta(f_1,f_2) \mathbb{I}, \qquad f_1, f_2 \in C^{\infty}_0({\mathcal M}), \] given in terms of the causal propagator $\Delta \doteq \Delta_R - \Delta_A$, i.e., the difference of the retarded and advanced propagators uniquely obtained as fundamental solutions of eq.\@ \eqref{eq:system-eq1}. The algebra of field observables can be enlarged to contain also Wick powers like $\phi^2$ when a properly defined normal ordering procedure is taken into account \cite{Hollands2001wick,Hollands2005stress,Moretti2003stress}. In light of the quantum nature of $\phi$, eq.\@ \eqref{eq:system-eq2} can be interpreted in the semiclassical approximation, namely by taking the expectation values of $\phi^2$ in a suitable quantum state. More precisely, once a state $\omega$ on $\mathcal{A}$ is chosen for the quantum field theory, we have that \[ g_2 \square \psi - g_1 \psi = \lambda_1\expvalom{\phi^2} - \lambda_2 \square \expvalom{\phi^2}, \] where $\expvalom{\phi^2} \doteq \omega(\phi^2)$ is the expectation value of the properly normal ordered quantum field $\phi^2$ in the quantum state $\omega$. Thus, the semiclassical system corresponding to equations \eqref{eq:system-eq1} \eqref{eq:system-eq2} turns out to be \begin{equation} \label{eq:system-semicl} \begin{cases} \square \phi - m^2 \phi = \lambda \psi \phi, \\ g_2 \square \psi - g_1 \psi = \lambda_1 \langle \phi^2\rangle_\omega -\lambda_2 \square \langle \phi^2\rangle_\omega. \end{cases} \end{equation} To analyze the linearization of this system around a given simple solution of eq.\@ \eqref{eq:system-semicl} determined by $(\psi_0,\omega)$, we decompose the classical field $\psi$ in two parts, the background contribution $\psi_0$ plus a perturbation $\psi_1$, so that \[ \psi= \psi_0 + \psi_1. \] The state $\omega$, or more specifically its two-point function $\hbar \Delta_{+,\omega}$ appearing in the second equation of the system \eqref{eq:system-semicl} through $\langle \phi^2\rangle_\omega$, can also be decomposed in two parts \[ \Delta_{+,\omega} = \Delta_{+,\omega,0}+ \Delta_{+,\omega,1}, \] i.e., the background contribution plus its perturbation. Notice that the linear order contribution in $\Delta_{+,\omega,1}$ is formed by two terms: one which takes into account the modified evolution equation induced by $\psi_1$ and satisfied by $\Delta_{+,\omega}$, and one which consists in $w_1$, a symmetric bi-solution of the zeroth order equation of motion satisfied by $\phi$. Since $\Delta_{+,\omega,0}$ is a bi-solution of the same equation, the effect due to $w_1$ can be reabsorbed in a redefinition of the background theory. For this reason, in the following, we shall take into account explicitly the effect due to the modified evolution only. To control the linear contribution in $\Delta_{+,\omega,1}$ due to the change of dynamics, we shall use perturbation theory for quantum fields. As we shall see later, if instead one takes into account the effect of $w_1$ explicitly, an inhomogenous (source) term has to be added to the linearized equation satisfied by $\psi_1$. In fact, this extra term will not alter the discussion we are going to present. The quantization of $\phi$ is performed on the fixed background $\psi_0$ by considering the $*-$algebra of field observables $\mathcal{A}$ with a product that implements the canonical commutation relations emerging from to the zeroth-order equation \begin{equation}\label{eq:linear} \square \phi - m^2 \phi - \lambda \psi_0 \phi = 0. \end{equation} Furthermore, both the quantum state $\omega$ and the background field $\psi_0$ satisfy the semiclassical equation, namely the second equation in the system given in eq.\@ \eqref{eq:system-semicl}, which thus represents a constraint for the couple $(\psi_0,\omega)$. We recall here that the expectation value of $\phi^2$ in the state $\omega$ on this background theory is obtained by an ordinary point-splitting regularisation procedure \begin{equation} \label{eq:point-splitting} \expvalom{\phi^2}^{(\text{bac})}(x) = \lim_{y\to x} \hbar\left( \Delta_{+,\omega}(x,y)-H(x,y) \right)+ c m^2 + c\lambda\psi_0(x), \end{equation} where $\hbar \Delta_{+,\omega}$ is the two-point function of the state $\omega$, $H(x,y) \doteq u(x,y)/(x-y)^2 + v(x,y)\log ((x-y)^2/\mu^2)$ is the universal divergent contribution present in any Hadamard two-point function, with $\mu$ a fixed length scale. Furthermore, $c$ is a constant which encodes the regularisation freedom present in the construction of Wick powers like $\phi^2$ \cite{Hollands2001wick,Hollands2002top}. Notice that the influence of $\psi_1$ on $\phi$ is governed by the first equation in the system \eqref{eq:system-semicl}; for a given $\psi_1$, this equation descends from a Lagrangian for $\phi$. In this Lagrangian, $\psi_1$ acts as a mass perturbation for $\phi$, hence we may use Lagrangian methods to analyze the influence of $\psi_1$ to $\phi$ even if the full system formed by both equations in eq.\@ \eqref{eq:system-semicl} is not Lagrangian. Thanks to the principle of perturbative agreement (see e.g. \cite{Hollands2005stress, Drago2017pertagr}) this approach is equivalent to directly analyze the effect of $\psi_1$ on the two-point function, as it was done, e.g., in \cite{Pinamonti2011init,Pinamonti2015glob,Meda2021exi}, or to evaluate $\langle \phi^2 \rangle_\omega$ in \cite{Eltzner2011back} in cosmological spacetimes. Thus, we pass to analyze the influence of the perturbation $\psi_1$ on $\phi$ by means of perturbation theory considering the interaction Lagrangian \begin{equation} \label{eq:Lagrangian-int} \mathcal{L}_I = -\frac{\lambda}{2} \psi_1 \phi^2. \end{equation} A perturbative construction of interacting quantum field theories on a generic curved spacetime $({\mathcal M},g)$ was rigorously formulated in a local and covariant way in the framework of perturbative algebraic quantum field theory - \emph{cf.} \cite{Brunetti2000micro,Duetsch2001aqft,Hollands2001wick,Hollands2002top,Fredenhagen2014kms,Fredenhagen2016paqft,Gere2016anly} to which we refer for further details on the construction we are going to present. In particular, the expectation value of the interacting field $\phi^2$ in the state $\omega$ is obtained by means of the Bogoliubov map \[ \langle \phi^2\rangle_\omega = \omega(R_V(\phi^2)), \] where the perturbative potential \[ V \doteq \int_{\mathcal M} \mathcal{L}_I(x) f(x) \mathrm{d}^4 x = -\frac{\lambda}{2} \int_{\mathcal M} \phi^2(x) \psi_1(x) f(x) \mathrm{d}^4 x \] is obtained by smearing the interaction Lagrangian given in eq.\@ \eqref{eq:Lagrangian-int} with a smooth cut-off $f \in C^\infty_0({\mathcal M})$ which is equal to $1$ on the compact spacetime region where we want to test the semiclassical equation. This cut-off $f$ is eventually removed by considering a suitable limit in which $f$ tends to $1$ on every point of $\mathcal{M}$. The Bogoliubov map $R_V$ is used to represent local field observables of the interacting theory as formal power series in $\lambda$, whose coefficients are well defined elements of $\mathcal{A}$, the extended algebra of free fields. In particular, \begin{equation} \label{eq:Bogoliubov} R_V(\phi^2) = S(V)^{-1} T(S(V)\phi^2), \end{equation} where $T$ is the map which realizes the time ordering, while $S(V)$ is the time ordered exponential of the smeared interaction Lagrangian $V$, namely \[ S(V) = T \at \exp(\frac{i}{\hbar} V) \ct = \sum_{n \geq 0} \frac{i^n}{\hbar^n n!} T \at V^{n} \ct. \] For the precise construction of the time ordering map $T$ we refer to \cite{Brunetti2000micro,Hollands2001wick,Hollands2002top}, where the old construction of Epstein and Glaser in \cite{Epstein1973loc} is generalized to a generic curved spacetime. \medskip With the perturbative construction of $\expvalom{\phi^2}$ at disposal, the perturbative expansion of $\expvalom{\phi^2}$ in the interacting theory obtained by means of the Bogoliubov map reads at the linear order in $V$ as \[ \expvalom{\phi^2} = \expvalom{\phi^2}^{\text{(bac)}} + \expvalom{\phi^2}^{\text{(lin)}} + \dots, \] where the contributions higher than the linear one are not displayed. Here, \begin{equation} \label{eq:phi2-bac-lin} \expvalom{\phi^2}^{\text{(bac)}} \doteq \omega(\phi^2), \qquad \expvalom{\phi^2}^{\text{(lin)}} \doteq \frac{i}{\hbar} \at \omega( T\at V \phi^2 \ct) - \omega(V \phi^2) \ct. \end{equation} The factor $i$ at the right hand side of the second equation in eq.\@ \eqref{eq:phi2-bac-lin} is present because $S(V)$ in eq.\@ \eqref{eq:Bogoliubov} is formally unitary. As expected by the principle of perturbative agreement, one can notice by direct computation that $\expvalom{\phi^2}^{\text{(lin)}}$ matches the linear contribution obtained in \cite{Pinamonti2011init,Pinamonti2015glob, Gottschalk2021cosm, Meda2021exi} in the cosmological case. The linearization studied in this paper consists of studying the semiclassical theory described by the system of equations given in eq.\@ \eqref{eq:system-semicl}, where the expectation value of the Wick square is approximated by truncating at first order the formal power series in the interaction Lagrangian \eqref{eq:Lagrangian-int} occurring in the map $R_V$ defined in eq.\@ \eqref{eq:Bogoliubov}. The state for the interacting quantum theory is constructed by means of the state on the linear quantum theory, i.e., on ${\mathcal A}$, and it is fixed once and forever, no matter the form of the linear perturbation $\psi_1$. Hence, on the one side the background theory is described by $(\psi_0,\omega)$, in which $\psi_0$ fulfils the semiclassical equation \begin{equation} \label{eq:background} g_2 \square \psi_0 - g_1 \psi_0 = (\lambda_1 -\lambda_2 \square) \expvalom{\phi^2}^{\text{(bac)}}, \end{equation} and the the quantum state $\omega$ constrained by eq.\@ \eqref{eq:background} is fixed once and for all in the linear algebra $\mathcal{A}$ of fields satisfying eq.\@ \eqref{eq:linear}. On the other side, the linear perturbation theory of the background is described by the classical field $\psi_1$, which fulfils the following linearized semiclassical equation \begin{equation} \label{eq:SCE-psi1} g_2 \square \psi_1 - g_1 \psi_1 = \left( \lambda_1 -\lambda_2 \square\right) \expvalom{\phi^2}^{\text{(lin)}}, \end{equation} where $\expvalom{\phi^2}^{\text{(lin)}}$ was constructed in eq.\@ \eqref{eq:phi2-bac-lin}. Eq.\@ \eqref{eq:SCE-psi1} is the only equation we have at linear order, and it must be seen as a dynamical equation for the linear perturbation $\psi_1$. If a perturbation of the state which modifies $\omega$ to $\omega+\omega_1$ is considered explicitly at the linear order, then an inhomogoenous contribution appears in eq.\@ \eqref{eq:SCE-psi1}, which consists in adding $\omega_1(\phi^2(x)) = \hbar w_1(x,x)$ to $\expvalom{\phi^2}^{\text{(lin)}}$. This extra contribution does not depend on $\psi_1$, and hence it acts as a source term in eq.\@ \eqref{eq:SCE-psi1}. We shall see in the next section that also solutions of the inhomogeneous version of eq.\@ \eqref{eq:SCE-psi1} have nice decay properties for several choices of the parameters $a,\lambda, \lambda_i, g_i$. \section{Stability of solutions of the linearized semiclassical equation} \label{sec:stability} With the results presented in \cite{Randjbar-Daemi980ren} in mind, our goal in this section is to show that perturbations $\psi_1$ over $(\psi_0,\omega)$ which solves eq.\@ \eqref{eq:system-semicl} decay for large time at linear order in perturbation theory and for proper choices of the coupling constants of the model. To achieve this, it is crucial to assume compactly supported initial data on the fields, and to consider perturbations around the background solution which have spatial compact support, in order to avoid the class of exponentially growing solutions already seen, e.g., in \cite{Horowitz1978inst,Horowitz1980wf,Jordan1987stab}. \subsection{Zeroth-order solution} Before discussing the perturbative construction of $\phi$, we give here some details on the chosen background solution $(\psi_0,\omega)$. For the sake of simplicity, we shall assume that the background $\psi_0\in \mathbb{R}$ is constant, hence the state $\omega$ must be such that the Wick square has constant expectation value \begin{equation} \label{eq:const} \omega(\phi^2) = \expvalom{\phi^2}^{\text{(bac)}}= -\frac{g_1}{\lambda_1} \psi_0 . \end{equation} In view of the renormalization freedom present in the definition of the Wick power expressed by the constant $c$ in eq.\@ \eqref{eq:point-splitting}, it is always possible to fulfil the previous equation whenever $\omega$ is a translation invariant state, for any choice of $\psi_0 \in \mathbb{R}$. A constant background external field $\psi_0$ corresponds to a mass renormalization, that is, \[ m_\lambda^2= m^2 +\lambda \psi_0 = m^2 - \lambda \frac{\lambda_1}{g_1} \langle \phi^2\rangle_{\omega}^{\text{(bac)}}. \] As $m_\lambda^2$ denotes the new renormalized mass of the background field $\phi$, we need to impose the constraint that $m_\lambda^2 \geq 0$, otherwise the reference state for the system may not exist. This inequality always holds for sufficiently small $\lambda_1/g_1$, and it holds trivially in the case of vanishing expectation value of the Wick square. Besides, there is always the possibility of setting $\expvalom{\phi^2}^{\text{(bac)}}=0$ by means of the choice of the renormalization of the field $\phi^2$ (the constant $c$ in eq.\@ \eqref{eq:point-splitting}). To simply further the analysis, we shall select the Minkowski vacuum state $\omega_0$ as reference state on ${\mathcal A}$, whose corresponding two-point function is \[ \omega_0(\phi(y)\phi(x)) = \expval{\phi(y)\phi(x)}{0} = \hbar \Delta_{+}(y-x), \] where \begin{equation} \label{eq:Delta+} \Delta_+(y-x) \doteq \frac{1}{(2\pi)^3} \int_{\mathbb{R}^4} \delta(p^2 + m^2)\Theta(p_0) \mathrm{e}^{i p(x-y)} \mathrm{d}^4 p, \end{equation} $\Theta$ is the Heaviside step function, and $\delta$ the Dirac delta function. However, other choices of quantum states, such that $\expvalom{\phi^2}$ is regular in ${\mathcal M}$, do not alter significantly our analysis. \subsection{The linearized expectation value of the Wick square} Using the Bogoliubov map given in eq.\@ \eqref{eq:Bogoliubov}, the expectation value of the Wick square can be evaluated at every perturbation order. The linearized contribution defined in eq.\@ \eqref{eq:phi2-bac-lin} in the adiabatic limit ($f = 1$) takes the following form \begin{equation} \label{eq:expval-phi-linear} \expvalom{\phi^2}^{\text{(lin)}} = -i \hbar \lambda \int_{\mathcal M} \at \Delta_{F,\omega}^2(y,x) - \Delta_{+,\omega}^2(y,x) \ct \psi_1(y) \mathrm{d} y, \end{equation} where $\hbar \Delta_{F,\omega}(y,x) = \expvalom{T \at \phi(y)\phi(x) \ct}$ and $\hbar \Delta_{+,\omega}(y,x) = \expvalom{\phi(y)\phi(x)}$ are the Feynman propagator and the two-point function associated to $\omega$, respectively. Notice that the definition of the Feynman propagator $\Delta_{F,\omega} \doteq \Delta_{+,\omega}+i\Delta_A$, where $\Delta_A$ is the advanced propagator, employed here differs by a factor $i$ from others constructions. Furthermore, the squares of $\Delta_{F,\omega}$ and $\Delta_{+,\omega}$ correspond to the pointwise multiplication of the integral kernels of the two distributions. We shall discuss how to construct these products below. A diagrammatic representation of the propagators in the integrand at the right hand side of eq.\@ \eqref{eq:expval-phi-linear} is given in Figure \ref{Fig:loop}. The various propagators of the theory, and in particular those appearing in eq.\@ \eqref{eq:expval-phi-linear}, are in general not invariant under translation. However, they acquire translation invariance when they are referred to the Minkowski vacuum $\omega_0$. To keep this in mind, we shall denote the Feynman propagator and the two-point function referred to the Minkowski vacuum as $\hbar\Delta_F(y-x) = \expval{T \at\phi(y)\phi(x)\ct}{0}$ and $\hbar \Delta_{+}(y-x) = \expval{\phi(y)\phi(x)}{0}$, respectively. Furthermore, we denote by $W \doteq \Delta_{+,\omega} - \Delta_{+}$ and we observe that $W$ is a smooth function whenever $\omega$ is an Hadamard state, namely the singular part of two-point function $\Delta_{+,\omega}$ is the same as the one in the Hadamard parametrix, or, equivalently, $\omega$ fulfils the microlocal spectrum condition \cite{Radzikowski1996micro,Brunetti2000micro}. Hence, recalling that $\Delta_F=\Delta_++i\Delta_A$, we get \[ \Delta_{F,\omega}^2 - \Delta_{+,\omega}^2 = \Delta_{F}^2 - \Delta_{+}^2 + i 2\Delta_A W. \] \begin{figure}[!htb] \centering{\resizebox{.60\linewidth}{!}{ \includegraphics[width=.80\linewidth]{tikz/loop.pdf} }} \caption{\footnotesize Picture of the one-loop contribution $T \at V \phi^2 \ct - V\phi^2$ at the linear order in $\hbar$. The propagator $\Delta_F(y-x)$ is represented by a non-oriented line because it is symmetric in the exchange of $x \leftrightarrow y$, while the two-point function $\Delta_{+}(y-x)$ by an arrow from $y$ to $x$ \cite{Gere2016anly}.} \label{Fig:loop} \end{figure} Since $W$ is smooth, we just need to construct $\Delta_+^2$ and $\Delta_F^2$ to give meaning to the right hand side of eq.\@ \eqref{eq:expval-phi-linear}. Contrary to $\Delta^2_{+}$ which is well defined everywhere, $\Delta_F^2$ is a well defined distributions only for test functions which are not supported on the origin. Therefore, a renormalization (extension) procedure is required to have a well defined $\Delta_F^2$ also in this case. The extension of $\Delta_F^2$ on test functions supported on the origin can be obtained keeping fixed the Steinmann scaling degree \cite{Steinmann1971pert,Brunetti2000micro}; however, this extension is not unique, and the remaining freedom amounts to an additional $c_0 \delta$, where $c_0$ is a real parameter. This freedom is compatible with the ambiguity present in the definition of $\phi^2$ \cite{Hollands2001wick}. To obtain explicit expressions of $\Delta_F^2$ and $\Delta_{+}^2$, we make use of the known K\"{a}llen-Lehmann spectral representations. In particular, using the convolution theorem and the definition of $\Delta_+(x)$ given in eq.\@ \eqref{eq:Delta+}, we obtain that \[ \Delta_{+}^2(x) = \int_{4m^2}^{\infty} \mathrm{d} M^2 \varrho(M^2) \Delta_+(x,M^2), \] where the spectral density $\varrho(M^2) = \frac{1}{(2\pi)^3}\int \frac{d^3{\mathbf{p}}} {\sqrt{\mathbf{p}^2+m^2}} \delta(M-2\sqrt{\mathbf{p}^2+m^2}) = \frac{1}{16\pi^2}\sqrt{1- \frac{4m^2}{M^2}}$, and $\Delta_+(x,M^2)$ is the Minkowski vacuum two-point function given in eq.\@ \eqref{eq:Delta+} with mass $M$. On the other hand, the very same representation for $\Delta_F^2(x)$ cannot be given because the integral over $M^2$ is divergent in that case. However, using the fact that for $x\neq0$, $(\Square+a)\Delta_F(x,M^2) = (M^2+a)\Delta_F(x,M^2)$, for $x\neq 0$ and $a>-4m^2$ we get \[ \Delta_{F}^2(x) = (\square+a)\int_{4m^2}^{\infty} \mathrm{d} M^2 \frac{\varrho(M^2)}{M^2+a} \Delta_F(x,M^2). \] This expression coincides with $\Delta_+^2(x)$ for $x \notin J^-(0)$ and with $\Delta_+^2(-x)$ for $x \notin J^+(0)$. Moreover, it is well defined also on functions whose support contains $0$, and thus it represents one of the possible extensions of $\Delta_F^2\in \mathcal{D}'({\mathcal M}\setminus \{0\})$ to $\mathcal{D}'({\mathcal M})$. Also, the difference of two $\Delta_F^2$ constructed with $a$ and $a'\neq a$ is a non vanishing distribution supported in the origin with scaling degree $4$, and hence it must be proportional to the Dirac delta function. Finally, we can saturate the freedom in the construction of $\Delta_F^2$ with various choices of $a$. In light of this, the constant $a$ encodes the renormalization freedom present in the construction of $\Delta_F^2(x)$. Therefore, recalling that $\Delta_F(x)=\Delta_+(x) +i \Delta_A(x)$, the most general form of $\Delta_F^2(x) - \Delta_{+}^2(x)$ reads as \begin{equation} \label{eq:DeltaF2-Delta+2-dM2} \begin{aligned} -i \at \Delta_F^2(x) - \Delta_{+}^2(x) \ct = (\square + a) \int_{4m^2}^{\infty} \mathrm{d} M^2 \frac{\varrho(M^2)}{M^2 + a} \Delta_A(x,M^2), \end{aligned} \end{equation} where $\Delta_A(x,M^2)$ is the advanced propagator of the Klein-Gordon field of mass $M$. A detailed derivation of eq.\@ \eqref{eq:DeltaF2-Delta+2-dM2} for the massless case can be found in Appendix C of \cite{Duetsch2004causpert}. Hence, eq.\@ \eqref{eq:expval-phi-linear} can be written at the linear order outside $x=0$ as \begin{equation} \label{eq:expval-phi-linear-conv} \expvalom{\phi^2}^{\text{(lin)}} = {\hbar} \lambda \at {\mathcal K}_a + {\mathcal W} \ct (\psi_1), \end{equation} where the operator ${\mathcal K}_a$ maps compactly supported smooth function (or Schwartz function) to smooth functions. Moreover, its regularized integral kernel takes the form \begin{equation} \label{eq:kernel-K} {\mathcal K}_a(x-y) = \int_{4m^2}^{\infty} \mathrm{d} M^2 \varrho(M^2) \frac{1}{M^2+a}(\square+a)\Delta_R(x-y,M^2) , \end{equation} where the d'Alembert operator is taken in the distributional sense, and $\Delta_R(\cdot,M^2)$ is the retarded propagator of the Klein Gordon equation with mass $M$. Its spatial Fourier kernel reads as \begin{equation} \label{eq:retarded-Fourier} \tilde{\Delta}_R(t,\vec{p},M^2) = -\frac{\sin(\omega_0 t)}{\omega_0} \Theta(t), \end{equation} where $\omega_0 = \sqrt{|\vec{p}|^2 + M^2}$, and $\Theta$ is the Heaviside step function. The operator ${\mathcal W}$ maps compactly supported smooth functions to smooth functions, and its integral kernel is the pointwise multiplication of the advanced propagator with the smooth part of the two-point function $W$, i.e., ${\mathcal W} \doteq 2 \Delta_A W$. With the choice of the Minkowski vacuum as reference state, both $W$ and ${\mathcal W}$ vanish, and thus the linearized expectation value of the Wick square given in eq.\@ \eqref{eq:expval-phi-linear-conv} simplifies as \begin{equation} \label{eq:expval-phi-linear-vacuum} \expvalv{\phi^2}^{\text{(lin)}}= \hbar \lambda \; {\mathcal K}_a(\psi_1). \end{equation} For later purposes, we need to control the evolution of $\expvalv{\phi^2}^{\text{(lin)}}$ in time under the influence of $\psi_1$: to this end, we study the kernel given in eq.\@ \eqref{eq:kernel-K} in the Fourier domain by means of the following proposition. \begin{proposition} \label{prop:state} Let $\psi_1 \in \mathcal{S}({\mathcal M})$ be a Schwartz function on $\mathcal{M}$, and let $\hat{\psi}_1$ be its Fourier transform. Then the Fourier transform of the linearized expectation value of the Wick square given in eq.\@ \eqref{eq:expval-phi-linear-vacuum} can be written as \begin{equation} \label{eq:expval-phi-linear-vacuum-transf} {\mathcal F} \ag \expvalv{\phi^2}^{\text{\normalfont(lin)}} \cg (p_0,\vec{p}) = \lim_{\epsilon\to 0^+} \frac{\lambda \hbar}{16\pi^2} F_a(-(p_0-i\epsilon)^2+|\vec{p}|^2) \hat\psi_1(p_0,\vec{p}), \end{equation} given for strictly positive mass $m>0$ and for $a>-4m^2$. The function $F_a(z)$ admits the following integral representation: \begin{equation} \label{eq:F} \begin{aligned} F_a(z) &= \int_{4m^2}^{\infty} \sqrt{1- \frac{4m^2}{M^2}} \at \frac{1}{M^2+a} - \frac{1}{M^2+z } \ct \mathrm{d} M^2, \end{aligned} \end{equation} and it has the following properties: \begin{itemize} \item[a)] $F_a(z)$ is analytic for $z \in \mathbb{C} \setminus (-\infty,-4m^2]$ and continuous at $z=-4m^2$; \item[b)] the domain $F_a(z)$ has a branch cut on $z\in(-\infty,-4m^2)$ because there the imaginary part is discontinuous (the real part is continuous but not differentiable); \item[c)] $F_a(a)=0$; \item[d)] $F_a(s)$ is real for $s\in \aq -4m^2, \infty\ct$, it is strictly increasing for $s \in \aq -4m^2, \infty\ct$, and it diverges for large $|s|$; \item[e)] The imaginary part of $F_a$ admits the following integral representation: \[ \text{\normalfont Im} (F_a(z))= \text{\normalfont Im}(z) \int_{4m^2}^{\infty} \sqrt{1- \frac{4m^2}{M^2}} \at \frac{1}{|M^2+z|^2 } \ct \mathrm{d} M^2, \] it is strictly positive for $\text{\normalfont Im}(z)>0$, and strictly negative for $\text{\normalfont Im}(z)<0$. Furthermore, it vanishes for $z\in (-4m^2,\infty)$, and it is discontinuous on $z\in(-\infty,-4m^2)$ (the absolute value is finite). \end{itemize} Finally, for $z\not\in(-\infty,0)$ and $a>0$, $F_a(z)$ takes the form \begin{equation} \label{eq:FF} \begin{aligned} F_a(z) &= 2\sqrt{\frac{z+4m^2}{z}} \log\left(\frac{\sqrt{z+4m^2}+\sqrt{z}}{2m}\right) -2\sqrt{\frac{a+4m^2}{a}}\log\left(\frac{\sqrt{a+4m^2}+\sqrt{a}}{2m}\right). \end{aligned} \end{equation} \end{proposition} The qualitative behaviour of $\text{Re}(F_a(z))$ and of $|\text{Im}(F_a(z))|$ for $z\in ( -4m^2, \infty)$ is plotted in Figure \ref{Fig:f}. \begin{figure}[!htb] \begin{minipage}{0.5\textwidth} \centering \includegraphics[width=.99\linewidth]{tikz/ReF.pdf} \end{minipage}\hfill \begin{minipage}{0.5\textwidth} \centering \includegraphics[width=.99\linewidth]{tikz/ImF.pdf} \end{minipage} \caption{\footnotesize The first graph contains the qualitative behaviours of $F_a(x)$ for $x \in (-4m^2,\infty)$ with $-4m^2<a<0$. The second graph is the qualitative behavior of $|\text{Im}F_a(x)|$.}\label{Fig:f} \end{figure} \begin{proof} Using the definition of ${\mathcal K}_a$ in eq.\@ \eqref{eq:kernel-K}, the Fourier transform of the distribution $\Delta_R$ given in eq.\@ \eqref{eq:retarded-Fourier} is \[ \hat{\Delta}_R(p_0,\vec{p}) = \frac{1}{(p_0-i0^+)^2 - |\vec{p}|^2 -M^2}. \] Hence taking the Fourier transform on both sides of eq.\@ \eqref{eq:expval-phi-linear-vacuum} and applying the convolution theorem to ${\mathcal K}_a(\psi_1)$ yields \begin{align*} &{\mathcal F} \ag \expvalv{\phi^2}^{\text{(lin)}} \cg (p_0,\vec{p}) = \\ &\lim_{\epsilon\to 0^+}\lambda \hbar \int_{4m^2}^{\infty} \mathrm{d} M^2 \varrho_2(M^2) \frac{1}{M^2+a} \frac{1}{-(p_0-i\epsilon)^2 + |\vec{p}|^2 + M^2} \at -p_0^2 +|\vec{p}|^2 -a \ct \hat{\psi}(p_0,\vec{p}), \end{align*} where $\hat{\psi}_1(p_0,\vec{p}) \in \mathcal{S}({\mathcal M})$. Then, the first part of the thesis follows after recalling the form of $\varrho_2(M^2) = \frac{1}{16\pi^2}\sqrt{1- \frac{4m^2}{M^2}}$ and the definition of $F_a(z)$. The list of properties of $F_a(z)$ can be inferred directly from its integral representation. To check the validity of the representation \eqref{eq:FF} of $F_a$, consider \[ A_a(z) \doteq \frac{1}{a-z}\left( 2\sqrt{\frac{z+4m^2}{z}} \log\left(\frac{\sqrt{z+4m^2}+\sqrt{z}}{2m}\right) -2\sqrt{\frac{a+4m^2}{a}}\log\left(\frac{\sqrt{a+4m^2}+\sqrt{a}}{2m}\right) \right). \] From the expression of $F_a$ given in eq.\@ \eqref{eq:FF}, $F_a(-p_0^2+|\vec{p}|^2)= (p_0^2-|\vec{p}|^2+a) A_a(-p_0^2+|\vec{p}|^2)$. We take the inverse Fourier transform in time of $A_a(-p_0^2+|\vec{p}|^2)$, that is, \[ \tilde{\mathcal{A}}(t,\vec{p}) \doteq \lim_{\epsilon \to 0^+} \frac{1}{2\pi}\int_{-\infty}^\infty A(-(p_0-i\epsilon)^2+\vec{p}^2) \mathrm{e}^{i p_0 t} \mathrm{d} p_0. \] The integrand in $\tilde{\mathcal{A}}$ has two cuts for $(p_0-i\epsilon)^2 > (|\vec{p}|^2+4m^2)$ located in the upper half complex plane and it is analytic outside the two cuts (it has no poles); furthermore, $|A(-(w-i\epsilon)^2+\vec{p}^2)|$ for $w\in\mathbb{C}$ vanishes in the limit $|w|\to\infty$. Hence, that inverse Fourier transform can be obtained by standard results of complex analysis, including Jordan's lemma and Cauchy residue theorem. In particular, to evaluate the integral over the real line, for $t<0$ we can close the contour in the lower half plane, and thus $\tilde{\mathcal{A}}=0$ because $A(-(w-i\epsilon)^2+\vec{p}^2)$ is analytic in the lower half plane. On the other hand, if $t>0$, then the contour is closed in the upper half plane, and thus only the two cuts matter in the evaluation of the integral over the real line which gives $\tilde{\mathcal{A}}$. The contributions due to the two cuts for $t>0$ can be combined to give \[ \tilde{\mathcal{A}}(t,\vec{p}) = \lim_{\epsilon \to 0^+} \frac{1}{2\pi}\int_{\sqrt{|\vec{p}|^2+4m^2}}^\infty \left(A(-w^2+\vec{p}^2+i\epsilon) -A(-w^2+\vec{p}^2-i\epsilon) \right) \left(\mathrm{e}^{i w t} - \mathrm{e}^{-i w t}\right) \mathrm{d} w. \] Changing variable of integration to $M^2 = w^{2}-|\vec{p}|^2$ and computing the discontinuity of $\log\left(\frac{\sqrt{z+4m^2}+\sqrt{z}}{2m}\right)$ along its cut, we obtain for $t>0$ that \[ \tilde{\mathcal{A}}(t,\vec{p}) = \frac{i2\pi}{2\pi}\int_{4m^2}^\infty \frac{1}{M^2+a} \left(\sqrt{1-\frac{4m^2}{M^2}}\right) \frac{\left(e^{i \omega_0 t} - e^{-i \omega_0 t}\right)}{2\omega_0} \mathrm{d} M^2, \] where $\omega_0 = \sqrt{|\vec{p}|^2+M^2}$. So \[ {\mathcal{A}}(x) = \int_{4m^2}^\infty \frac{1}{M^2+a} \left(\sqrt{1-\frac{4m^2}{M^2}}\right) \Delta_R(x,M^2) \mathrm{d} M^2, \] and hence $(\square +a) \mathcal{A}$ is equal to $\mathcal{K}_a$ up to a constant factor given in eq.\@ \eqref{eq:kernel-K}. Therefore, the expression of $F_a$ given in eq.\@ \eqref{eq:F} follows, thus proving that it coincides with eq.\@ \eqref{eq:FF}. \end{proof} \begin{remark} \label{rem:massless} In the limit of vanishing mass $m^2 = 0$ and for $-p_0^2+|\vec{p}|^2>0$, the function $F_a(-p_0^2+|\vec{p}|^2)$ given in eq.\@ \eqref{eq:F} takes the form \begin{equation} \label{eq:F-massless} F_{a}(-p_0^2+|\vec{p}|^2) = \log(\frac{-p_0^2+|\vec{p}|^2}{a}). \end{equation} This logarithmic behaviour is similar to the case studied in \cite{Horowitz1980wf} for the linearized semiclassical Einstein equations, in the weak-field limit of gravity and after considering massless quantum fields (see also \cite{Hartle1981ground}). The function $F_a$ is also similar to the Fourier transform of the Green function associated to the first order equation analyzed in \cite{Flanagan1996back} in the context of semiclassical back-reaction. \end{remark} \subsection{Linearized solutions} The semiclassical equation \eqref{eq:SCE-psi1} governing the dynamics of the perturbations at the linear order is a linear equation in the perturbation field $\psi_1$. As we shall see below the properties of the solution space of that linear equation depends strictly on the parameters $a,\lambda,\lambda_i, g_i$, $i=1,2$. The non local state-dependent contribution $\expvalv{\phi^2}^{(\text{lin})}(x)$ defined in eq.\@ \eqref{eq:expval-phi-linear-vacuum} was constructed in terms of the linear operator $\mathcal{K}_a$ introduced in eq.\@ \eqref{eq:expval-phi-linear-conv}, and thus it was expressed in terms of the function $F_a$ studied in Proposition \ref{prop:state}. Therefore, eq.\@ \eqref{eq:SCE-psi1} evaluated in the Minkowski vacuum state reads \begin{equation} \label{eq:SCE-psi-linear} (g_2\square - g_1) \psi_1(x) = (\lambda_1 - \lambda_2 \square)\expvalv{\phi^2}^{(\text{\normalfont lin})}(x). \end{equation} To highlight its mathematical structure, eq.\@ \eqref{eq:SCE-psi-linear} can be rewritten in the following form: \begin{equation} \label{eq:SCE-eq1-hom} \hbar \lambda P_\lambda \mathcal{K}_a(\psi_1)(x) + P_g \psi_1(x) = 0, \end{equation} where $P_\lambda \doteq \lambda_2 \square - \lambda_1$, and $P_g \doteq g_2 \square - g_1$. \medskip Because of the presence of a second d'Alembert operator in the expression of $\expvalv{\phi^2}^{(\text{lin})}(x)$ given in eq.\@ \eqref{eq:expval-phi-linear-conv} through eq.\@ \eqref{eq:kernel-K}, eq.\@ \eqref{eq:SCE-psi-linear} contains fourth-order derivatives in $\psi_1$. It has thus a form similar to the semiclassical equations which usually appear in semiclassical theories of gravity, see, e.g., \cite{Kay1981see} and the next Section. Thus, it manifests the same conceptual issues already known in semiclassical gravity as higher-order theory of gravity. In particular, there are cases where similar equations admit the so-called runaway solutions, which make the classical background field unstable in the semiclassical approach \cite{Horowitz1978inst,Horowitz1980wf,Jordan1987stab,Parker1993red,Simon1991stab,Flanagan1996back}. Runaway solutions usually consist of solutions of the linearized system around some background which grow exponentially in time. This class of linearized solutions become dominant over the background at large times. Thus, the full solution of the system acquires, in principle, a very different form from the chosen background, and at the same time it is expected to be very sensitive to the chosen initial conditions. Therefore, the background solution cannot be assumed to be stable. On the contrary, if all the linearized solutions decay sufficiently fast to zero for large times, then the perturbations become negligible with respect to the background solution, thus indicating the stability of the background. In the next part we analyze eq.\@ \eqref{eq:SCE-eq1-hom} equipped with a compactly supported smooth source term, namely \begin{equation} \label{eq:SCE-eq1-source} \hbar \lambda P_\lambda \mathcal{K}_a(\psi_1)(x) + P_g \psi_1(x) = f(x), \end{equation} where $\mathcal{K}_a$ is the linear operator introduced in eq.\@ \eqref{eq:expval-phi-linear-conv}, and $f \in C_0^\infty({\mathcal M})$ is a compactly supported source. The strategy is as follows. In a first step, we shall show that this equation manifests an hyperbolic nature, and we shall construct its retarded fundamental solutions as an operator $D_R: C^\infty_0(\mathcal{M}) \to C^\infty(\mathcal{M})$. Afterwards, thanks to the regularity properties of $D_R$, we shall prove that past compact solutions of the form $\psi_1 = D_R(f)$ decay as $1/t^{3/2}$ for large times $t$, hence getting the stability of the corresponding backgrounds against perturbation sourced by $f$. In a second step, we shall study the smooth spatially compact solutions of the homogeneous equation \eqref{eq:SCE-eq1-hom} corresponding to eq.\@ \eqref{eq:SCE-psi1}. To determine uniquely a solution in the future and in the past of $t_0$, we shall equip eq.\@ \eqref{eq:SCE-eq1-hom} with suitable initial conditions at $t_0 =0$, i.e., smooth compactly supported initial data of the form \[ \psi_1^{(j)}\at 0,\vec{x} \ct = \varphi^j(\vec{x}) \] for $j\in \{0,1\}$ or $j\in \{0,1,2,3\}$, with $\varphi^j\in C^\infty_0(\mathbb{R}^3)$. We shall see that the number of initial conditions necessary to determine a spatially compact solution depends on the choice of the parameters: in some cases, four initial conditions have to be imposed, while, in other cases, only two initial conditions are sufficient to obtain a solution; finally, there are also cases where no solution exists. Thus, we shall prove that there are wide ranges of values of $(a,g_1,g_2,\lambda,\lambda_1,\lambda_2)$ for which solutions of eq.\@ \eqref{eq:SCE-eq1-hom} with compactly supported initial data decay faster than $1/t^{3/2}$ for large times. Therefore, the stability of the linearized back-reacted system is restored even in this case. \medskip Our analysis starts from showing that eq.\@ \eqref{eq:SCE-psi-linear} written as eq.\@ \eqref{eq:SCE-eq1-hom} manifests an hyperbolic nature. \begin{proposition} \label{theo:hyperbolic} Consider the equation \eqref{eq:SCE-eq1-source} sourced by $f\in {C}^\infty_0({\mathcal M})$ in the form \[ \hbar \lambda P_\lambda \mathcal{K}_a(\psi_1)(x) + P_g \psi_1(x) = f(x). \] Set $\lambda_2\neq 0$ and $g_2\neq0$. Let $\psi_1$ be a past compact solution of eq.\@ \eqref{eq:SCE-eq1-source}, then \begin{equation} \label{eq:retardation} \psi_1 = \frac{\lambda_2}{g_2} \left( \Delta_{R,\lambda} f - \hbar \lambda \mathcal{K}_a(\psi_1) - \Delta_{R,\lambda} \left(P_g-\frac{g_2}{\lambda_2}P_\lambda\right)\psi_1 \right) \end{equation} where $\Delta_{R,\lambda}$ is the retarded fundamental solution of $P_\lambda \Delta_{R,\lambda} = \mathbb{I}$. Moreover, if we consider $\delta \psi_1, \delta f\in C^\infty_0(M)$ then, $\psi_1(x)+\delta \psi_1(x)$ is also a solution of \[ \hbar \lambda P_\lambda \mathcal{K}_a(\psi_1+\delta \psi_1)(x) + P_g (\psi_1+\delta \psi_1)(x) = f(x)+\delta f(x). \] for every $x \not \in (J^{+}(\text{\normalfont supp} \delta f) \cup J^{+}(\text{\normalfont supp} \delta \psi_1) )$. \end{proposition} \begin{proof} As the solution $\psi_1$ is past compact by hypothesis, we can apply the retarded operator associated to $P_\lambda$ on both sides of eq.\@ \eqref{eq:SCE-eq1-source}, thus obtaining \[ \hbar \lambda \mathcal{K}_a(\psi_1) + \Delta_{R,\lambda} P_g\psi_1 = \Delta_{R,\lambda} f. \] Using the definition of fundamental solution $(\Delta_{R,\lambda} \circ P_\lambda) \psi_1 = \psi_1$, this equation can be written also as \[ \hbar \lambda \mathcal{K}_a(\psi_1) + \Delta_{R,\lambda} \left(P_g-\frac{g_2}{\lambda_2}P_\lambda\right)\psi_1 + \frac{g_2}{\lambda_2}\psi_1 = \Delta_{R,\lambda} f, \] and hence as \[ \frac{g_2}{\lambda_2}\psi_1 = \Delta_{R,\lambda} f - {\hbar} \lambda \mathcal{K}_a(\psi_1) - \Delta_{R,\lambda} \left(P_g-\frac{g_2}{\lambda_2} P_\lambda\right)\psi_1, \] thus yielding eq.\@ \eqref{eq:retardation}. Notice that both $\Delta_{R,\lambda}$ and $\Delta_{R,\lambda} (\lambda_2 P_g-g_2 P_\lambda)$ have the retarded property, i.e., $\text{supp}(\Delta_{R,\lambda} f) \subset J^+ (\text{supp} f)$ and $\text{supp}(\Delta_{R,\lambda} (\lambda_2 P_g- g_2 P_\lambda) \subset J^{+} (\text{supp} f)$. Similarly, $\mathcal{K}_a$ satisfies also the retarded propertied because it is an integral of retarded operators. The last observation descends from the fact that to compute $\mathcal{K}_a(\psi_1+\delta \psi_1)(x)$ only $\psi_1+\delta \psi_1$ in the past of $x$ matters. Hence, if $x\not\in J^{+}(\text{supp}\delta f) \cup J^{+}(\text{supp}\delta \psi_1)$, $\psi_1(x)+\delta \psi_1(x) = \psi_1 (x)$, $f(x)+\delta f(x) = f (x)$, and $\mathcal{K}_a(\psi_1+\delta \psi_1)(x)= \mathcal{K}_a(\psi_1)(x)$. We conclude that, for $x\not\in J^{+}(\text{supp}\delta f) \cup J^{+}(\text{supp}\delta \psi_1)$, $\psi_1+\delta\psi_1$ is again a solution of the inhomogeneous equation with the source modified by $\delta f(x)$. \end{proof} The constraints on the parameters $\lambda_2$ and $g_2$ given by hypothesis in Proposition \ref{theo:hyperbolic} can be easily removed adapting the first part of the proof. For example, if $g_2=0$, there is no need to add $P_\lambda \Delta_{R,\lambda}$ in the right hand side of eq.\@ \eqref{eq:retardation}. On the other hand, if $\lambda_2=0$, then the proof starts with getting an analog of eq.\@ \eqref{eq:retardation} after applying the retarded operator $\Delta_{R,g}$ at the place of $\Delta_{R,\lambda}$ on both sides of eq.\@ \eqref{eq:SCE-eq1-source}. Finally, if both $\lambda_2$ and $g_2$ vanish, then there is no need of preliminary applying any retarded operator to eq.\@ \eqref{eq:SCE-eq1-source}. Proposition \ref{theo:hyperbolic}, and more precisely eq.\@ \eqref{eq:retardation}, suggests that the form of a past compact solution $\psi_1$ of eq.\@ \eqref{eq:SCE-eq1-source} in $x$ cannot be influenced by any modification of $\psi_1$ or $f$ outside of $J^{-}(x)$. We shall see a posteriori that this indication is actually correct, because we shall prove in Theorem \ref{theo:retarded} that a retarded fundamental solution of eq.\@ \eqref{eq:SCE-eq1-source} exists. \medskip We proceed in the analysis of the form of the solution of the linearized semiclassical equation by studying the associated retarded fundamental solution. We use Fourier techniques to analyze the fundamental solution of eq.\@ \eqref{eq:SCE-eq1-source}. Hence, \begin{equation} \label{eq:linearized-source-fourier} -\left((\lambda_1 + \lambda_2 (-(p_0)^2+|\vec{p}|^2))\frac{\lambda \hbar}{16\pi^2} F_a(-(p_0-i0^+)^2+|\vec{p}|^2) + (g_2(-p_0^2+|\vec{p}|^2) + g_1)\right) \hat\psi_1(p_0,\vec{p})=\hat{f}(p_0,\vec{p}), \end{equation} where $f\in C^{\infty}_0(\mathcal{M})$. Eq.\@ \eqref{eq:SCE-eq1-source} can equivalently be written in a more compact form as \[ S(-(p_0-i0^+)^2+|\vec{p}|^2) \hat\psi_1(p_0,\vec{p})=\hat{f}(p_0,\vec{p}), \] where \begin{equation} \label{eq:S} S(z) \doteq - (\lambda_1+\lambda_2 z ) \frac{\lambda \hbar}{16\pi^2} F_a(z) - (g_1+g_2 z). \end{equation} In order to obtain the retarded fundamental solutions associated to the kernel $S(z)$, we need to study the set of points in the complex plane in which $S(z)$ vanishes: we denote this set by $\mathcal{S}$. Then, we shall prove that, if the parameters $(\lambda,g_i,\lambda_i)$ satisfy certain conditions, this set contains only real elements, and, furthermore, it includes only negative elements in some special cases. Among them, we shall impose as a constraint on the parameters the following inequality: \begin{equation} \label{eq:condition-param} g_2 \lambda_1-\lambda_2 g_1 \geq 0. \end{equation} The characterization of elements in $\mathcal{S}$ is studied in the following proposition. \begin{proposition} \label{prop:zeros} Let $\mathcal{S}\subset\mathbb{C}$ be set of zeros of $S(z)$ given in eq.\@ \eqref{eq:S}. Fix the parameters in such a way that at least one of the two $\lambda_i$, $i\in\{1,2\}$ is non vanishing and $g_2 \lambda_1-\lambda_2 g_1 \geq 0$, $\lambda>0$, $-4m^2<a$. Then we distinguish two cases: \begin{itemize} \item[a)] if $\lambda_2\neq0$, and $-\lambda_1/\lambda_2 \leq -4m^2$, then $\mathcal{S}\subset(-4m^2,\infty)\cup\{-\lambda_1/\lambda_2\}\subset \mathbb{R}$; \item[b)] $\mathcal{S}\subset (-4m^2,\infty) \subset \mathbb{R}$ otherwise. \end{itemize} In particular, if $\lambda_2\neq0$, $-\lambda_1/\lambda_2$ is in $ \mathcal{S}$ only if $g_2 \lambda_1/\lambda_2 = - g_1$. Furthermore, $\mathcal{S}$ contains one, two or no elements depending on the parameters $\lambda_i$, $g_i$, and $a$. \end{proposition} \begin{proof} Let us start with the equation $S(z)=0$ written in the form \begin{equation} \label{eq:CasoD} (\lambda_1+\lambda_2 z ) \frac{\lambda \hbar}{16\pi^2} F_a(z) = - (g_1+g_2 z). \end{equation} To prove that the solution set is as stated in item a) and b), we proceed as follows. After multiplying both sides of the equation by $(\lambda_1+\lambda_2 \overline{z})$, taking the imaginary part yields the following equation: \[ |\lambda_1+\lambda_2 z |^2 \text{Im} \at \frac{\lambda \hbar}{16\pi^2} F_a(z) \ct = - (g_2 \lambda_1-\lambda_2 g_1) \text{Im}(z), \] where the global sign of right-hand side depends only on $\text{Im}(z)$, because by hypothesis $g_2 \lambda_1-\lambda_2 g_1 \geq 0$. In particular, for strictly positive $\text{Im}(z)$ the right hand side is negative, while the left hand side is strictly positive thanks to item $e)$ of Proposition \ref{prop:state}; similarly, for strictly negative $\text{Im}(z)$ the right hand side is positive, while the left hand side is strictly negative. Moreover, as $\text{Im}(F_a(x))$ is not $0$ also for $x < -4m^2$ (see Figure \ref{Fig:f}), we have that the only possible solutions of eq.\@ \eqref{eq:CasoD} needs to be searched within $(-4m^2,\infty) \cup \{-\lambda_1/\lambda_2\}$ when $\lambda_2\neq 0$, or in $(-4m^2,\infty)$ otherwise. Furthermore, if both $\lambda_2\neq 0$ and $z=-\lambda_1/\lambda_2$, then the left hand side of eq.\@ \eqref{eq:CasoD} vanishes, whereas the right hand side vanishes only when $g_2 \lambda_1/\lambda_2 =-g_1$. We are thus left with the analysis the following real equation: \begin{equation} \label{eq:CasoReale} (\lambda_1+\lambda_2 s ) \frac{\lambda \hbar}{16\pi^2} F_a(s) = - (g_1+g_2 s), \qquad s\in (-4m^2,\infty) \cup \{-\lambda_1/\lambda_2\}. \end{equation} The properties of the function $F_a(s)$ for $s\in (-4m^2,\infty) \subset \mathbb{{R}}$ are listed in Proposition \ref{prop:state}, and the plot of this function is reported in Figure \ref{Fig:f}. Notice in particular that $F_a(s)$ is a concave function, because the second derivative of the integrand in eq.\@ \eqref{eq:F} is $-2\varrho(M^2) (M^2+s)^{-3}$, which is a strictly negative integrable function, and thus $F_a(s)''$ is strictly negative. Hence, for $\lambda_2=0$, $\lambda_1\neq 0$ by hypothesis, $(\lambda_1+\lambda_2 s)F_a(s)$ has a definite concavity. In this case, the maximum number of distinct solutions of eq.\@ \eqref{eq:CasoReale} is two. If $\lambda_2\neq 0$, we rewrite eq.\@ \eqref{eq:CasoReale} as \[ \frac{\lambda \hbar}{16\pi^2} F_a(s) + \frac{g_2}{\lambda_2} = \frac{1}{\lambda_2^2} \frac{\lambda_2 g_1-\lambda_1 g_2} {s + \frac{\lambda_1}{\lambda_2}} = I(s). \] Notice that $\frac{\lambda \hbar}{16\pi^2} F_a(s) + \frac{g_2}{\lambda_2}$ is monotonically increasing. Using the hypothesis that $\lambda_2g_1-\lambda_1g_2 \geq 0$, we observe that $I(s)$ is constant if $\lambda_2g_1-\lambda_1g_2 = 0$ or monotonically decreasing if $\lambda_2g_1-\lambda_1g_2 >0$; in this latter case, it has also a discontinuity (a vertical asymptote) in $s=-\lambda_1/\lambda_2$. Hence, also in this case there are at most two solutions, thus concluding the proof. \end{proof} We observe that, under the hypothesis of Proposition \ref{prop:zeros}, $\mathcal{S}$, the space of zeros of eq.\@ \eqref{eq:S}, coincides with the set of elements where eq.\@ \eqref{eq:CasoReale} vanishes. Having established that there are at most two distinct solutions, owning the properties of the function $F_a(s)$ for $s\in (-4m^2,\infty) \subset \mathbb{{R}}$ stated in Proposition \ref{prop:state}, and the plot of the qualitative behaviour of that function reported in Figure \ref{Fig:f}, we may draw the following conclusion. There are cases where either one or two positive solutions of this equation exist, and there are cases where only one or two negative solutions exist. It is also possible to find cases where one positive and one negative solution exists. Finally, there are cases where no solutions exists at all. Taking into account all the previous statements, we are now ready to write the explicit form of the retarded fundamental solution of eq.\@ \eqref{eq:SCE-eq1-source}, and to show that past compact solutions decay to zero for sufficiently large times, as expected for a perturbation over a stable background. \begin{theorem} \label{theo:retarded} Consider the semiclassical equation with a source term $f\in C^\infty_0(\mathcal{M})$ given in eq.\@ \eqref{eq:SCE-eq1-source} in the form \[ \hbar \lambda P_\lambda \mathcal{K}_a(\psi_1)(x) + P_g \psi_1(x) = f(x). \] Fix as non-vanishing constants at least one of the two $g_i$, and at least one of the $\lambda_i$, assume that the inequality $g_2 \lambda_1-\lambda_2 g_1 \geq 0$ holds, and set $-4m^2<a<0$. Suppose that the set $\mathcal{S}$ defined in Proposition \ref{prop:zeros} contains only real negative elements, then the Fourier transform of the retarded fundamental solution $D_R$ of eq.\@ \eqref{eq:SCE-eq1-source} reads \[ \hat{D}_R(p_0,\mathbf{p})= \frac{1}{S(-(p_0-i0^+)^2+|\mathbf{p}|^2)}, \] where $S(z)$ was defined in eq.\@ \eqref{eq:S}. Hence \begin{equation} \label{eq:retarded-position} D_R(x) = -\sum_{s\in \mathcal{S}} \frac{1}{S'(s)} \Delta_R(x,-s) - \frac{\lambda \hbar}{16\pi^2} \int_{4m^2}^\infty \sqrt{1-\frac{4m^2}{M^2}} \frac{ (\lambda_2 M^2-\lambda_1 )}{|S(-M^2)|^2} \Delta_R(x,M^2) \mathrm{d} M^2, \end{equation} where the elements of $\mathcal{S}$ are the zeros of $S(s)$, with $s\in (-4m^2,\infty) \cup \{-\lambda_1/\lambda_2\}$. The retarded fundamental solution $D_R$ is a linear operator which maps smooth compactly supported functions to smooth functions, and with this operator at disposal the solution $\psi_1$ of eq.\@ \eqref{eq:SCE-eq1-source} with past compact support is \begin{equation} \label{eq:solution-ps1} \psi_1 = D_R(f). \end{equation} For $\lambda_2\neq 0$, $\psi_1(t,\vec{x})$ decays as $1/t^{3/2}$ for large $t$. \end{theorem} \begin{proof} We analyze the equation \eqref{eq:SCE-eq1-source} in the Fourier domain. Using the results given in Proposition \ref{prop:state}, it takes the form of \[ S(-(p_0-i0^+)^2+|\vec{p}|^2) \hat\psi_1(p_0,\vec{p}) = \hat{f}(p_0,\vec{p}), \] where $S$ was defined in eq.\@ \eqref{eq:S}. Thus, the Fourier transform of the retarded operator $D_R$ yields \begin{equation} \hat{D}_R(p_0,\vec{p})= \frac{1}{S(-(p_0-i0^+)^2+|\vec{p}|^2)}, \end{equation} and hence its Fourier inverse transform \[ \tilde{D}_R(t,\mathbf{p})=\lim_{\epsilon\to 0^+}\lim_{R\to\infty}\frac{1}{2\pi}\int_{-R}^R \frac{1}{S(-(z-i \epsilon)^2 +|\mathbf{p}|^2)} \mathrm{e}^{itz} \mathrm{d} z \] can be evaluated by means of standard methods of complex analysis. In view of the properties of the function $F_a(z)$ given in eq.\@ \eqref{eq:F}, the function $h(z)\doteq 1/S(-(z-i\epsilon)^2+|\vec{p}|^2)$ is defined in $\mathbb{C}\setminus \{\{z-i\epsilon\leq -\sqrt{|\mathbf{p}|^2+4m^2} \}\cup\{z-i\epsilon\geq \sqrt{|\mathbf{p}|^2+4m^2} \} \}$. For $|z|>R$, it holds that $|h(z)|< l(|z|)$, where the function $l(r)$ vanishes in the limit of large positive $r$, because, in the worse case, $1/S(-z^2+|\mathbf{p}|^2)$ is dominated by $c/F_a(-z^2+|\mathbf{p}|^2)$ for large $|z|$, for some constant $c$, and $|F_a(-z^2+|\mathbf{p}|^2)|$ grows as $\log|z|$ for large $|z|$. Therefore, from Jordan's lemma we may close the contour $\gamma$ in the upper or lower plane, according to the sign of $t$, with a semicircle which does not contribute to the integral in the limit $R\to\infty$. The function $1/S(-(z-i\epsilon)^2+|\mathbf{p}|^2)$ has two poles for each element $s \in \mathcal{S}$ (the set of zeros of $S(z)$). Since $s\in \mathcal{S}$ is negative by hypothesis, we have that the poles are located on the line $\text{Im}(z)=i \epsilon$, and correspond to the complex numbers \[ z= i\epsilon \pm \sqrt{|\mathbf{p}|-s}. \] Furthermore, the function $1/S(-(z-i\epsilon)^2+|\mathbf{p}|^2)$ has two branch cuts located at $z = x+i\epsilon $, where $x^2 \geq |\vec{p}|^2 + 4m^2$. Thus, $1/S$ is analytic in the lower half plane, and hence, we obtain that $\tilde{D}_R=0$ for $t < 0$ by Jordan's lemma, because we close the contour in the lower half plane for $t<0$. On the other hand, if $t>0$, then we close the contour $\gamma$ in the upper half plane, and hence we need to take care of both the poles and the branch cuts. In this case, if we deform the previous contour $\gamma$ to a new $\tilde{\gamma}$ in such a way to avoid both the poles and the cuts, then the result of the contour integral over $\tilde{\gamma}$ vanishes. Therefore, the only two non-vanishing contributions in $\tilde{D}_R$, denoted by $\tilde{O}$ and $\tilde{C}$, are due to the poles and the branch cuts, respectively. The contribution due to the poles can be directly evaluated using the Cauchy residue theorem, which yields \[ \tilde{O}(t,\vec{p}) = -\frac{2\pi i }{2\pi}\sum_{s\in \mathcal{S}} \frac{1}{S'(s)} \left(\frac{e^{ i w_s t }- e^{ -i w_s t}}{2w_s} \right) = \sum_{s\in \mathcal{S}} \frac{1}{S'(s)} \frac{\sin (w_st)}{w_s}, \] where $w_s \doteq \sqrt{|\vec{p}|^2 - s}$, and hence, in view of eq.\@ \eqref{eq:retarded-Fourier}, \begin{equation} \label{eq:Ox} O(x) = -\sum_{s\in \mathcal{S}} \frac{1}{S'(s)} \Delta_R(x,-s). \end{equation} The contribution due to the cuts can be combined in the following form \[ \tilde{C}(t,\vec{p}) = \lim_{\epsilon\to 0^+ }\frac{1}{2\pi} \int_{\sqrt{|\vec{p}|^2+4m^2}}^\infty \left[\frac{1}{S(-p_0^2+|\vec{p}|^2+i\epsilon)}-\frac{1}{S(-p_0^2+|\vec{p}|^2-i\epsilon)}\right] \left( e^{i p_0 t } - e^{-i p_0 t }\right) \mathrm{d} p_0. \] Thus, recalling eq.\@ \eqref{eq:retarded-Fourier} again, it can be written in the position domain as \[ C(x) = \lim_{\epsilon\to 0^+ }\frac{-i}{2\pi} \int_{4m^2}^\infty \left[\frac{1}{S(-M^2+i\epsilon)}-\frac{1}{S(-M^2-i\epsilon)}\right] \Delta_R(x,M^2) \mathrm{d} M^2. \] The discontinuity in the two cuts is only due to the imaginary part of $F_a(z)$, hence \begin{equation} \label{eq:Cx} C(x) = {-} \frac{\lambda \hbar}{16\pi^2} \int_{4m^2}^\infty \sqrt{1-\frac{4m^2}{M^2}} \frac{ (\lambda_2 M^2-\lambda_1 )}{ | (\lambda_2 M^2 - \lambda_1 ) \frac{\lambda \hbar}{16\pi^2} F_a(-M^2) + (g_2 M^2-g_1)|^2} \Delta_R(x,M^2) \mathrm{d} M^2, \end{equation} thus getting eq.\@ \eqref{eq:retarded-position} by combining \cref{eq:Ox,eq:Cx}. Furthermore, the integral over $M^2$ present in $C(x)$ in eq.\@ \eqref{eq:Cx} can always be taken, even when both $\lambda_2$ and $g_2$ vanish, because $\hat{\Delta}_R$ and $1/|F_a(-M)|^2$ decay as $1/M^2$ and $1/|\log(M)|^2$ for large $M$, respectively. \medskip The decay of $\psi_1=D_R(f)$ for large $t$ descends straightforwardly from Lemma \ref{le:decay} applied to $D_R(f)$, which implies that $\Delta_R(f,-s)$ decays as $1/t^{3/2}$ for large $t$, with $s\leq 0$. Hence, the contribution $O(f)$ due the poles given in eq.\@ \eqref{eq:Ox} has the desired time decay property. The same holds for the contribution $C(f)$ due to the cuts given in eq.\@ \eqref{eq:Cx}, because for $\lambda_2\neq 0$ the function $M^2/|S(-M^2)|$ is integrable in $\mathrm{d} M^2$, and, furthermore, $f$ is smooth and compactly supported in time, so its time Fourier transform is a Schwartz function. \end{proof} \begin{remark} \label{rem:cuts} From the form of the kernel of $D_R$ obtained in eq.\@ \eqref{eq:retarded-position}, a generic past compact solution $\psi_1=D_R(f)$ defined in eq.\@ \eqref{eq:solution-ps1} can be decomposed into two parts, so that \[ \psi_1(x) = \psi_1^{O}(x) + \psi_1^C(x), \] where $\psi_1^{O} = -\sum_{s\in \mathcal{S}} \frac{1}{S'(s)} \Delta_R(f,-s)$ denotes the contribution due to the poles of $1/S$, while $\psi_1^C$ is the contribution due the cuts. We observe that, while there is a chance to determine $\psi_1^{O}$ by means of a finite number of initial conditions given at some time $t_1$ in the future of $\text{supp}f$, we expect that it is not possible to determine $\psi_1^C$ with a finite number of initial conditions, because the integration of $M^2$ is over uncountably many points. In spite of this fact, we notice that the homogeneous equation \eqref{eq:SCE-eq1-hom} may still, in some cases, give origin to a well-posed initial value problem to uniquely determine spatially compact solutions. Actually, the contribution due the cuts cannot enter the construction of the solutions of the homogeneous equation on the whole space. The reason is that the kernel of the multiplicative operator $T$, which acts on $\mathcal{S}(\mathbb{R}^4)$ and is defined as \begin{align*} T(z) &\doteq \frac{S(z)}{\prod_{s\in\mathcal{S}}(z-s)}, \end{align*} contains only $0$, with $z = -(p_0 - i\epsilon)^2 + |\vec{p}|^2$. Therefore, only the contributions due the poles can give origin to non trivial solutions of the homogeneous equation \eqref{eq:SCE-eq1-hom} written as $S(z) \hat{\psi_1} = 0$. \end{remark} The decay rate of the smooth past compact solutions $\psi_1$ proved in Theorem \ref{theo:retarded} is the same which was obtained by means of Strichartz estimates for real, massive quantum scalar fields in four-dimensional Minkowski spacetime \cite{Strichartz1977wave}. Actually, this behaviour is justified by the form of the fourth-order differential equation \eqref{eq:SCE-eq1-source}, which is composed of massive Klein-Gordon like operators on $({\mathcal M},\eta)$. Thus, the retarded fundamental solution is still a combination of Klein-Gordon like fundamental solutions, and hence the past compact solutions given in eq.\@ \eqref{eq:solution-ps1} inherit the same late-time behaviour estimated for real Klein-Gordon fields. Eventually, we can now to discuss the solutions of the linearized semiclassical equation without source given in eq.\@ \eqref{eq:SCE-eq1-source}. \begin{theorem} \label{theo:main} Consider the semiclassical equation \eqref{eq:SCE-psi-linear} written as \[ (g_2\square - g_1) \psi_1(x) = (\lambda_1 - \lambda_2 \square)\expvalv{\phi^2}^{(\text{\normalfont lin})}(x). \] Fix as non-vanishing constants at least one of the two $g_i$, and at least one of the $\lambda_i$, assume that the inequality $g_2 \lambda_1-\lambda_2 g_1 \geq 0$ holds, and set $-4m^2<a<0$. Let $\mathcal{S}\subset (-4m^2,\infty) \cup \{-\lambda_1/\lambda_2\}\subset \mathbb{R}$ be the set of zeros of $S(z)$ given in eq.\@ \eqref{eq:S}. As discussed in Proposition \ref{prop:zeros}, $\mathcal{S}$ contains one, two or no elements depending on the parameters $\lambda_i$, $g_i$ and $a$. If $\mathcal{S} = \emptyset$, then eq.\@ \eqref{eq:SCE-psi-linear} admits no solutions. If $\mathcal{S} \neq \emptyset$, let $\psi_1(t,\vec{x})$ be a smooth solution of eq.\@ \eqref{eq:SCE-psi-linear} with spatial compact support. Then its spatial Fourier transform is of the form \[ \tilde{\psi}_1(t,\vec{p}) = \sum_{s\in\mathcal{S}} \left( C^s_+(\vec{p}) \mathrm{e}^{ + i t \sqrt{|\vec{p}|^2-s}} +C^s_-(\vec{p}) \mathrm{e}^{ - i t \sqrt{|\vec{p}|^2-s}}\right). \] Moreover, if $\mathcal{S}$ contains only negative elements, then each solution $\psi_1$ of eq.\@ \eqref{eq:SCE-psi-linear} is uniquely fixed by the initial values at $t=0$ \[ \psi_1^{(j)}\at 0,\vec{x} \ct = \varphi^j(\vec{x}), \qquad j\in \{0,\dots , 2 |\mathcal{S}|\}, \] where $|\mathcal{S}|$ is the cardinality of $\mathcal{S}$, and $\varphi^j\in C^{\infty}_0(\mathbb{R}^3)$. Furthermore, in this case, $\psi_1(t,\vec{x})$ decays for large time at least as $1/t^{3/2}$. \end{theorem} \begin{proof} We analyze eq.\@ \eqref{eq:SCE-psi-linear} in the Fourier domain. Using the results given in Proposition \ref{prop:state}, it takes the form: \begin{equation} \label{eq:main-theo} \left((\lambda_1 + \lambda_2 (-p_0^2+|\vec{p}|^2)) \frac{\lambda \hbar}{16\pi^2} F_a(-(p_0-i0^+)^2+|\vec{p}|^2) + (g_2(-p_0^2+|\vec{p}|^2) + g_1) \right)\hat\psi_1(p_0,\vec{p})=0. \end{equation} Let $\tilde{\psi}(t,\vec{p})$ be the spatial Fourier transform of a generic solution $\psi_1(t,\vec{x})$ of the form given in eq.\@ \eqref{eq:solution-ps1}. This is a linear combination of $\mathrm{e}^{i p^j_0 t}$ with $z = -{p^j_0}^2+|\vec{p}|^2 \in \mathcal{S}$, where $\mathcal{S}$ is the set of points of the complex plane in which the function $S$ given in eq.\@ \eqref{eq:S} vanishes, namely in which eq.\@ \eqref{eq:CasoD} holds. According to Proposition \ref{prop:zeros}, we have that $\mathcal{S}$ must be contained either in $(-4m^2,\infty)\subset \mathbb{C}$, or in $(-4m^2,\infty)\cup\{-\lambda_1/\lambda_2\}\subset \mathbb{C}$ for $\lambda_2\neq0$, $g_2 \lambda_1/\lambda_2 = - g_1$, and $\lambda_1/\lambda_2 \leq 4m^2$. \medskip According to the number of negative solutions of eq.\@ \eqref{eq:SCE-psi-linear}, the explicit form of $\psi_1(t,\vec{x})$ reads as follows. If $\mathcal{S}$ contains only one negative solution $x=-\tilde{n}$, with $\tilde{n}\geq 0$, then any solution having smooth compactly supported initial data at $t=0$ is of the form \[ \psi_1(t,\vec{x}) = \int_{\mathbb{R}^3} \left(C_+(\vec{p}) \mathrm{e}^{ i w_{\tilde{n}} t } + C_-(\vec{p})\mathrm{e}^{ -i w_{\tilde{n}} t} \right) \mathrm{e}^{ i {\vec{p} \cdot\vec{x} }} \mathrm{d} \vec{p}, \] where $w_{\tilde{n}}(\vec{p}) = \sqrt{|\vec{p}|^2 +\tilde{n}}$, and $C_\pm$ are obtained from $({\hat{\varphi}}^0(\vec{p}),{\hat{\varphi}}^1(\vec{p}))$ by solving \[ \begin{pmatrix} {\hat{\varphi}}^0\\ {\hat{\varphi}}^1 \end{pmatrix} = \begin{pmatrix} 1&1\\ i w_{\tilde{n}}&- i w_{\tilde{n}} \end{pmatrix} \begin{pmatrix} C_+\\ C_- \end{pmatrix}, \] which yields \[ \begin{pmatrix} C_+\\ C_- \end{pmatrix} = \frac{i}{2w_{\tilde{n}}} \begin{pmatrix} - i w_{\tilde{n}}&-1\\ -i w_{\tilde{n}}&1 \end{pmatrix} \begin{pmatrix} {\hat{\varphi}}^0\\ {\hat{\varphi}}^1 \end{pmatrix}, \] namely \[ C_+(\vec{p}) = {\hat{\varphi}}^0(\vec{p}) - \frac{i}{2 w_{\tilde{n}}} {\hat{\varphi}}^1(\vec{p}), \qquad C_-(\vec{p}) = {\hat{\varphi}}^0(\vec{p}) + \frac{i}{2 w_{\tilde{n}}} {\hat{\varphi}}^1(\vec{p}). \] Thus, thanks to Lemma \ref{le:decay}, we obtain the desired decay of $\psi(t,\vec{x})$ for large $t$. If $\mathcal{S}$ contains only two distinct negative elements, four initial data are needed to fix the solution. Denoting with $s_1=-n_1$ and $s_2=-n_2$, $n_i\geq 0$, the two distinct elements of $\mathcal{S}$, the linearized solution of the semiclassical equation \eqref{eq:SCE-psi-linear} is a combination of two solutions of the Klein Gordon equation with different square masses $n_i$. In this case, the solution with smooth compactly supported initial data $(\varphi^0(\vec{x}),{\varphi}^1(\vec{x}),\varphi^2(\vec{x}),\varphi^3(\vec{x}))$ at $t=0$ is of the form \[ \psi_1 (t,\vec{x}) = \int_{\mathbb{R}^3} \left(C^1_+(\vec{p}) \mathrm{e}^{ i w_{1} t } + C^1_-(\vec{p})\mathrm{e}^{ -i w_{1} t} \right) \mathrm{e}^{ i {\vec{p} \cdot\vec{x} }} \mathrm{d} \vec{p} + \int_{\mathbb{R}^3} \left(C^2_+(\vec{p}) \mathrm{e}^{ i w_{2} t } + C^2_-(\vec{p})\mathrm{e}^{-i w_{2} t} \right) \mathrm{e}^{ i {\vec{p} \cdot\vec{x} }} \mathrm{d} \vec{p}, \] where $w_{i}(\vec{p}) = \sqrt{|\vec{p}|^2 +n_i}$, and $C^i_\pm(\vec{p})$ are obtained from $(\hat{\varphi}^0(\vec{p}),{\hat{\varphi}}^1(\vec{p}),\hat{\varphi}^2(\vec{p}),\hat{\varphi}^3(\vec{p}))$ by solving \[ \begin{pmatrix} \hat{\varphi}^0 \\ \hat{\varphi}^1 \\ \hat{\varphi}^2 \\ \hat{\varphi}^3 \end{pmatrix} = \begin{pmatrix} 1&1&1&1\\ i w_{1}&- i w_{1} & i w_{2}&- i w_{2} \\ - w_{1}^2&- w_{1}^2 & - w_{2}^2&- w_{2}^2 \\ -i w_{1}^3& + i w_{1}^3 & -i w_{2}^3& + i w_{2}^3 \end{pmatrix} \begin{pmatrix} C^1_+\\ C^1_-\\ C^2_+\\ C^2_- \end{pmatrix}. \] The determinant of that matrix is equal to $-4(n_1-n_2)^2 w_1w_2$, and hence $C^i_\pm(\vec{p})$ can be written as linear combinations of $(\hat{\varphi}^0(\vec{p}),{\hat{\varphi}}^1(\vec{p}),\hat{\varphi}^2(\vec{p}),\hat{\varphi}^3(\vec{p}))$ as \[ \begin{pmatrix} C^1_+\\ C^1_-\\ C^2_+\\ C^2_- \end{pmatrix} = \frac{1}{2 (n_1-n_2)} \begin{pmatrix} - w_2^2&i \frac{w_2^2}{w_1}&-1& \frac{i}{w_1} \\ - w_2^2 &-i \frac{w_2^2}{w_1} & -1& -\frac{i}{w_1} \\ w_{1}^2&- i \frac{w_1^2}{w_2}& 1& -\frac{i}{w_2} \\ w_{1}^2& i \frac{w_1^2}{w_2}& 1& \frac{i}{w_2} \end{pmatrix} \begin{pmatrix} \hat{\varphi}^0 \\ \hat{\varphi}^1 \\ \hat{\varphi}^2 \\ \hat{\varphi}^3 \end{pmatrix}. \] Notice that these coefficients have either $w_1$ or $w_2$ in the denominator, in the worse case. Thus, the desired decay of $\psi(t,\vec{x})$ for large $t$ is obtained by applying Lemma \ref{le:decay} as before. \captionsetup[figure]{labelfont={bf},labelformat={default},name={Figure}} \begin{figure}[!htb] \centering{\includegraphics{tikz/solution_fig_a.pdf}} \medskip \caption{\footnotesize Plots of the qualitative behaviours of $(\lambda_1+\lambda_2 x)(\frac{\lambda \hbar}{16\pi^2})F_a(x)$ and $(g_1+g_2 x)$ in $\aq -4m^2, \infty\ct$, where the constants are such that $g_i > 0$, $\lambda_i>0$, and $-4m^2<-\frac{\lambda_1}{\lambda_2}<-\frac{g_1}{g_2}<a<0.$} \label{Fig:fig_a} \end{figure} \begin{figure}[!htb] \centering{\includegraphics{tikz/solution_fig_b.pdf}} \medskip \caption{\footnotesize Plots of the qualitative behaviours of $F_a(x)$ and $(g_1+g_2 x)/\lambda_1$ in $\aq -4m^2, \infty\ct$, where the constant are such that $g_i > 0$, $\lambda_2=0$, $\lambda_1>0$, $\frac{\lambda \hbar}{16\pi^2}F_a(-4m^2)<-(g_1-g_2 4m^2)/\lambda_1$, and $\frac{\lambda \hbar}{16\pi^2}F_a(0)>-g_1/\lambda_1$.} \label{Fig:fig_b} \end{figure} \end{proof} The previous theorem establishes that, if the space of solutions of eq.\@ \eqref{eq:CasoReale} contains only negative elements, then all solutions of the linearized semiclassical equation \eqref{eq:SCE-psi-linear} with compactly supported initial values decay at large times. In the following corollary, we identify certain sufficient (but not necessary) conditions on the parameters $\lambda,\lambda_i,g_i,a$ which ensure such a behavior. \begin{corollary} \label{cor:cases} Under the hypotheses of Theorem \ref{theo:main}, the space of solutions of eq.\@ \eqref{eq:CasoReale} contains only negative elements if the following sufficient conditions on the parameters $\lambda,\lambda_i,g_i,a$ hold: \begin{itemize} \item[a)] If $\lambda_2 = 0$, $\frac{g_2}{\lambda_1} \geq 0$, if $-\frac{g_1}{\lambda_1 } \leq \frac{\lambda \hbar}{16\pi^2} F_a(0)$, and $\frac{\lambda \hbar}{16\pi^2} F_a(-4m^2) \leq -\frac{g_1}{\lambda_1}+4m^2 \frac{g_2}{\lambda_1}$, then $\mathcal{S}$ contains only one negative solutions. \item[b)] If $-\lambda_1/\lambda_2 < -g_1/g_2< a <0$, then $\mathcal{S}$ contains only negative solutions, and $|\mathcal{S}|$ is either $1$ or $2$. \item[c)] If $-4m^2<-\lambda_1/\lambda_2 < a <0$, $g_2>0$ and $0\leq -\frac{g_1}{\lambda_1 } \leq \frac{\lambda \hbar}{16\pi^2} F_a(0)$, then $\mathcal{S}$ contains only negative solutions, and $|\mathcal{S}|$ is either $1$ or $2$. \end{itemize} If these conditions hold, then the corresponding solutions of eq.\@ \eqref{eq:SCE-psi-linear} with compact spatial support decay at least as $1/t^{3/2}$ at large times. \end{corollary} \begin{proof} \begin{itemize} \item[a)] Case $\lambda_2 = 0$, $\frac{g_2}{\lambda_1} \geq 0$. Eq.\@ \eqref{eq:CasoReale} takes the form \[ \frac{\lambda \hbar}{16\pi^2} F_a(x) = - \left(\frac{g_1}{\lambda_1 }+\frac{g_2}{\lambda_1 } x\right), \qquad x\in (-4m^2,+\infty). \] The real part of that equation has now a positive solution if $-\frac{g_1}{\lambda_1 } > \frac{\lambda \hbar}{16\pi^2} F_a(0)$. From the plot displayed in Figure \ref{Fig:fig_b} we infer that a single negative solution appears if $-\frac{g_1}{\lambda_1 } \leq \frac{\lambda \hbar}{16\pi^2} F_a(0)$ and $ \frac{\lambda \hbar}{16\pi^2} F_a(-4m^2) \leq -\frac{g_1}{\lambda_1}+4m^2 \frac{g_2}{\lambda_1}$, while no solutions exist otherwise. \item[b)] Case $a<0$, $-\lambda_1/\lambda_2 \leq -g_1/g_2\leq a <0$. From the plot displayed in Figure \ref{Fig:fig_a}, it is found that all the possible solutions are negative, and in particular either one or two solutions exist. \item[c)] Case $-\lambda_1/\lambda_2 < a <0$ and $0\leq -\frac{g_1}{\lambda_1 } \leq \frac{\lambda \hbar}{16\pi^2} F_a(0)$. This case is similar to the one displayed in Figure \ref{Fig:fig_a}, but now the line $-g_2x + g_1$ intercepts the vertical line between $y=0$ and $y=\frac{\lambda \hbar}{16\pi^2} F_a(0)$. Hence, all the possible solutions are negative, and there is always at least one solution. \end{itemize} \medskip Finally, the proof follows by applying the results of Theorem \ref{theo:main}. \end{proof} To summarize, we have thus proven that the desired decay as $1/t^{3/2}$ for large time $t$ of the solutions of the linearized Einstein equation \eqref{eq:SCE-psi-linear} with compact spatial support holds if and only if the zeros of $S$ defined in eq.\@ \eqref{eq:S} are all contained in the negative real axis. On the contrary, if some zeros were located in the positive real axis, then unstable runaway solutions would destabilize the background configuration. Finally, if no zeros are present in $S$, then eq.\@ \eqref{eq:SCE-psi-linear} does not admit solutions, but its counterpart with source given in eq.\@ \eqref{eq:SCE-eq1-source} still has non vanishing past compact solutions which decay in time, due to the contribution given by the source through the branch cuts. \begin{remark} According to the results presented in Proposition \ref{prop:zeros}, every solution of eq.\@ \eqref{eq:CasoReale} is located in the positive real axis whenever the quantum field $\phi$ is massless, i.e., when $m=0$, even if the inequality \eqref{eq:condition-param} holds (recalling that $F_a(z)$ is reduced to eq.\@ \eqref{eq:F-massless} in the massless case). For this reason, we may expect that any stability result cannot be achieved for massless fields, at least when homogeneous equations are taken into account, even if compactly supported initial data are selected. This observation is in accordance with the stability issue established in \cite{Horowitz1980wf}, where it was shown that exponentially growing runaway solutions appear when the back-reaction of a quantum Maxwell field interacting with a weak gravitational field is taken into account in the framework of semiclassical gravity. \end{remark} \subsection{Applications in the cosmological model} \label{sec:cosmo} The analysis employed in the toy model presented in eqs.\@ \eqref{eq:system-eq1} and \eqref{eq:system-eq2} can be used to guess the behaviour of the linearized solutions of the semiclassical Einstein equations in cosmological spacetimes, where matter is modelled by a massive quantum scalar field $\phi$. In the cosmological scenario we are considering here, the spacetime geometry is described by the metric $g_{\mu\nu}$ of the flat Friedmann-Lema\^itre-Robertson-Walker spacetime. In conformal coordinates $(\tau,x_1,x_2,x_3)$, the metric is conformally flat and reads \begin{equation} \mathrm{d} s^2 = a(\tau)^2 \at -\mathrm{d} \tau^2 + \mathrm{d} \vec{x}^2 \ct, \end{equation} where the scale factor $a(\tau)$ of the universe represents the unique degree of freedom of the spacetime. The dynamics of this universe is governed by the back-reaction of a linear quantum scalar field $\phi$, whose equation of motion is \begin{equation} \label{eq:scalar-field} \square \phi -m^2 \phi -\xi R \phi = 0, \end{equation} where $\xi$ denotes the coupling constant to the scalar curvature. Since there is an unique degree of freedom in this class of spacetimes, the dynamics of $a$ is determined, up to a constraint, by the trace of the semiclassical Einstein equations \cite{Meda2021exi} \begin{equation} \label{eq:SCE-cosmo} - R + 4\Lambda = 8 \pi G \expvalom{T}. \end{equation} The expectation value of the trace of the quantum stress-energy tensor associated to $\phi$ in a quantum state $\omega$ is \begin{equation} \label{eq:traceT} \expvalom{T} = \left(3\left(\xi - \frac{1}{6}\right) \square - m^{2}\right)\expvalom{\phi^2} + \frac{1}{4 \pi^{2}}\left[v_{1}\right] + \alpha_1 m^4 - \alpha_2 m^2 R + \alpha_3 \square R, \end{equation} where the renormalization constants $\alpha_i$ are not fixed by the model, but describe the regularisation freedom present in the definition of the stress-energy tensor as normal-ordered Wick observable \cite{Hollands2001wick,Hollands2005stress,Hack2016cosm}. The constants $\alpha_1$ and $\alpha_2$ correspond to renormalizations of the cosmological constant and the Newton constant, respectively, and thus they can be reabsorbed in a redefinition of $\Lambda$ and $G$; on the contrary, $\alpha_3$ is of pure quantum nature and cannot be reabsorbed in any corresponding classical parameter of the theory. Moreover, the coefficient $[v_1]$ appearing in eq.\@ \eqref{eq:traceT} corresponds to the so-called quantum trace anomaly of the model, and reads \begin{equation} \label{eq:trace-anomaly} [v_1] = \frac{1}{2880\pi^2 }\left(C^{abcd}C_{abcd} + R^{ab}R_{ab} -\frac{1}{3} R^2\right), \end{equation} up to contributions which can be reabsorbed by a redefinition of $\alpha_i$. Here, $C_{abcd}$ is the Weyl tensor, $R_{ab}$ the Ricci tensor and $R$ the Ricci scalar \cite{Wald1978trace,Moretti2003stress,Hollands2005stress}. We are interested in studying the linearized perturbation of this cosmological system around a spacetime which is a solution of the semiclassical Einstein equations written in the form of eq.\@ \eqref{eq:SCE-cosmo}. As a first step of this analysis, and for the sake of simplicity, we consider as background solution a spacetime with vanishing curvature, namely the Minkowski spacetime. Under this assumption, we can obtain a formal correspondence between the linearization of eq.\@ \eqref{eq:SCE-cosmo} and the linearized semiclassical equation \eqref{eq:SCE-psi-linear}, viewing $R$ as the perturbative external field $\psi_1$ over a vanishing background $\psi_0=0$. In view of this correspondence, the cosmological constant $\Lambda$ is a zeroth-order contribution which can be assumed to vanish, whereas the trace anomaly given in eq.\@ \eqref{eq:trace-anomaly} is at least quadratic in the components of the Riemann curvature tensor $R_{abcd}$, and thus it is negligible at linear order in $R$. Taking into account all of this and eq.\@ \eqref{eq:traceT}, eq.\@ \eqref{eq:SCE-cosmo} takes the form of the linearized semiclassical equation \eqref{eq:SCE-psi-linear} through the following correspondence between the cosmological parameters and the set of constants $(g_1,g_2,\lambda_1,\lambda_2)$: \begin{equation} \label{eq:cosmo-param} g_1 = -\frac{1}{8\pi G}, \qquad g_2 = \alpha_3, \qquad \lambda=\xi, \qquad \lambda_1 = m^2, \qquad \lambda_2 = 3\left(\xi -\frac{1}{6}\right). \end{equation} In the cosmological framework, $g_1$ turns to be a fixed negative parameter (in Planck's units, $\hbar = 1$ and $(8\pi G)^{-1} = m_P^2/8\pi$, where $m_P$ is the Planck mass), while, on the other hand, $\lambda$ can be fixed to be strictly positive and different from $1/6$ by assuming non-minimally and non-conformally coupled fields, i.e, $\xi \neq 0, 1/6$; hence, $\lambda_2 \neq 0$. On the contrary, both $g_2$ and $\lambda_2$ are free parameters of the semiclassical theory, whose signs can be chosen such that the inequality \eqref{eq:condition-param} holds, i.e., \begin{equation} \label{eq:cosmo-condition} \alpha_3 \frac{m^2}{m_P^2} \geq - \frac{3}{8\pi} \left(\xi -\frac{1}{6}\right), \qquad \alpha_3 \in \mathbb{R}. \end{equation} Under these assumptions, there are choices of the parameters $(m^2,\xi,a)$ for which the cosmological version of eq.\@ \eqref{eq:CasoReale} admits only negative solutions. For example, by choosing $\xi > 1/6$, $\alpha_3 > 0$, $a > -4m^2$, and sufficiently large $m^2$ we may apply Corollary \ref{cor:cases}, which ensures that only negative solutions exist. Namely, in these cases solutions of the cosmological linearized semiclassical Einstein equations written as in eq.\@ \eqref{eq:SCE-cosmo} with spatial compact support decay to zero for large times, thus showing the stability of the chosen background. On the other hand, one may expect that too large values of $m^2$, even beyond the Planck scale $m_P$, would be physically unacceptable for quantum fields describing elementary particles. In this viewpoint, the result is similar to the one obtained in \cite{Randjbar-Daemi981stab} for massive quantum scalar fields in flat spacetime. Firstly, the conditions stated here are only sufficient, so other cases which provide stable solutions cannot be excluded a priori, for different choices of the parameters $(m^2,\xi,\gamma,a)$. Secondly, and most importantly, it is expected that the linearized perturbations in a more realistic cosmological model should be sourced by $f\in {C}^\infty_0({\mathcal M})$ localized somewhere in the past. This source may have a quantum origin, for instance, related to some anisotropic or stochastic fluctuations at microscopic levels. This is the case which occurs for example in Stochastic Gravity, where a noise kernel bi-tensor modelling the stress-energy tensor fluctuations is added to the semiclassical Einstein equations, obtaining in this way the so-called Einstein-Langevin equations (see \cite{Hu2020stoc} and reference therein). In this picture, the stochastic source in the past drives the fluctuations of the gravitational field, and thus gives origin to the external perturbation which enters the cosmological linearized semiclassical Einstein equations as external source. In this model with external sources, the cosmological counterpart of the linearized semiclassical equation \eqref{eq:SCE-eq1-source} should be taken into account, with parameters $\lambda_1,\lambda_2,g_1,g_2$ fixed as in eq.\@ \eqref{eq:cosmo-param} and satisfying the inequality \eqref{eq:cosmo-condition}. Under these assumptions, and based on the results shown in Theorem \ref{theo:retarded} and Remark \ref{rem:cuts}, the linearized curvature solution $R$ depends on both the contributions due to the poles and the branch cuts of $S(z)$. However, the contribution arising from poles are not present for several, apparently more physically acceptable values of the parameters $(m^2,\xi,a)$: for example, for sufficiently large ratio $m_P^2/m^2$ and \[ 0 < \xi < 1/6, \qquad \alpha_3 > 0, \qquad a > -4m^2, \] the condition \eqref{eq:cosmo-condition} holds. Furthermore, with this choice of parameters, the past compact linearized solution induced by a smooth compactly supported source $f$ has no poles contribution, and hence it decays to zero for large times according to the results stated in Theorem \ref{theo:retarded}. \section{Conclusions} \label{sec:conc} In this paper, we have analyzed the stability problem of semiclassical theories in flat spacetime, using a toy model consisting of a quantum scalar field coupled to a second entirely classical scalar field. This toy model mimics other semiclassical theories of gravity described by the semiclassical Einstein equations, where a quantum matter field propagates over a classical curved background. It is known that higher order derivatives appearing in the semiclassical equations can destabilize the system, giving rise exponentially growing linearized solutions (see the references given in the \hyperref[sec:intro]{Introduction}). The main result stated in this paper consists of proving that, if the quantum field driving the back-reaction is massive, then the stability of background solutions can be restored at the linear order in the interaction, for spatially compact perturbations and for large values of the coupling constants, after assuming some sufficient (but not necessary) conditions on the parameters of the theory. On the other hand, removing the assumption of massive quantum fields seems to give rise to runaways solutions which may alter stability, in accordance with other results present in the literature about semiclassical theories. Namely, it is shown that unique solutions of this semiclassical initial value problem tend to disperse in time, namely at fixed position in space they decay polynomially in time. As this toy model mimics the dynamics of the linearized semiclassical Einstein equations, our analysis indicates a possible mechanism to get stability in several linearized semiclassical theories of gravity, even for different conditions than the ones stated in this paper. For example, in the case of a semiclassical theory in cosmological spacetimes, which has been already investigated by the authors in \cite{Pinamonti2011init,Pinamonti2015glob,Meda2021exi}. \bigskip \subsection*{Acknowledgements} We thank the anonymous referees for helpful comments on an earlier version of this paper. The work of P.M. was supported by a PhD scholarship of the University of Genoa. We are grateful for the support of the National Group of Mathematical Physics (GNFM-INdAM). \bigskip
{ "redpajama_set_name": "RedPajamaArXiv" }
3,044
<?php namespace ZpgRtf\Objects; /** * The branch object structures the main branch payload. */ class BranchObject implements \JsonSerializable { /** @var null|string */ private $branchReference; /** @var null|string */ private $branchName; /** @var null|LocationObject */ private $location; /** @var null|string */ private $telephone; /** @var null|string */ private $email; /** @var null|string */ private $website; /** * @return null|string */ public function getBranchReference() { return $this->branchReference; } /** * @param string $branchReference * * @return BranchObject */ public function setBranchReference(string $branchReference) { $this->branchReference = $branchReference; return $this; } /** * @return null|string */ public function getBranchName() { return $this->branchName; } /** * @param string $branchName * * @return BranchObject */ public function setBranchName(string $branchName): self { $this->branchName = $branchName; return $this; } /** * @return null|LocationObject */ public function getLocation() { return $this->location; } /** * @param LocationObject $location * * @return BranchObject */ public function setLocation(LocationObject $location): self { $this->location = $location; return $this; } /** * @return null|string */ public function getTelephone() { return $this->telephone; } /** * @param string $telephone * * @return BranchObject */ public function setTelephone(string $telephone): self { $this->telephone = $telephone; return $this; } /** * @return null|string */ public function getEmail() { return $this->email; } /** * @param string $email * * @return BranchObject */ public function setEmail(string $email): self { $this->email = $email; return $this; } /** * @return null|string */ public function getWebsite() { return $this->website; } /** * @param string $website * * @return BranchObject */ public function setWebsite(string $website): self { $this->website = $website; return $this; } /** {@inheritDoc} */ public function jsonSerialize(): array { return array_filter([ 'branch_reference' => $this->getBranchReference(), 'branch_name' => $this->getBranchName(), 'location' => $this->getLocation(), 'telephone' => $this->getTelephone(), 'email' => $this->getEmail(), 'website' => $this->getWebsite(), ]); } }
{ "redpajama_set_name": "RedPajamaGithub" }
8,216
Q: BST is not abstract and does not override abstract method height() in TreeGT I am receiving the error: BST is not abstract and does not override abstract method height() in TreeGT. I need some help with understanding what it means, and how I can get resolve it. In the file TreeGT.java I have the following code: public interface TreeGT<E> { public boolean insert(E item); public boolean delete(E item); public boolean find(E item); public int height(); } In the file BST.java I have the following code: import java.util.*; public class BST<E extends Comparable<E>> implements TreeGT<E> { private Node root; //Only root by itself public BST() { root = null; } private static class Node { Comparable data; int height; //Height of node int size; //Number of nodes in tree private Node left; //Left subtree private Node right; //Right subtree Node (Comparable data) { //Constructor for tree this.data = data; this.height = 0; //Height zero this.size = 1; //Root counts this.left = null; //No left leafs this.right = null; //No right leafs } } public void makeEmpty() { root = null; } public boolean isEmpty() { return root == null; } public int size() { if (isEmpty()) throw new Exception("Tree is empty"); return root == null ? 0 : root.size; } public boolean insert(E item) { return false; } public boolean delete(E item) { return false; } public boolean find(E item) { return false; } } P.S. I have only been programming in java a few days, so you will have to speak as simple/straight forward as possible, please. Also, I am trying to create a BST. I think I am on the right path, no? A: An interface is a collection of undefined (i.e. abstract) methods. When you implement an interface (which you do in implements treeGT) you have to provide a definition for ALL of those methods. So, since your class implements treeGT, and height() is one of the abstract methods defined in treeGT, you have to define that method in your class. i.e. you need to add public int height(){ //do something }
{ "redpajama_set_name": "RedPajamaStackExchange" }
7,600
On my desk, I keep a tattered post-it note that reads, "Because of you, I can do something." A few years ago, a client said this as he sat in my office talking about his new life. I still remember the moment he said that. I felt like an insignificant speck in his new life in Akron, Ohio, but, to him, I stood as a reminder of his new beginning. Who are the people in our own lives that call us to enter into the newness of life? What do these people represent? Register Today for "Together: A Collaborative for Theological Education, Formation and Community" at Catholic Theological Union! Roman Catholic Sisters and Brothers who are currently in formation and have completed their canonical novitiate year are invited to participate in a new program called Together: A Collaborative for Theological Education, Formation, and Community, cosponsored by Catholic Theological Union (CTU) and the Religious Formation Conference (RFC). Read more about Register Today for "Together: A Collaborative for Theological Education, Formation and Community" at Catholic Theological Union! Do you write poetry? Giving Voice is excited to announce our new original poetry series: #SisterPoets! Each month, sisters who participate in Giving Voice are encouraged to write a poem on one spiritual theme, to be a part of a monthly collection published in our e-newsletter, on our website and social media platforms. Read more about Giving Voice Announces #SisterPoets! With some of my sisters, I volunteered at a local soup kitchen. It was my first time at this location in Rome and the patrons were immigrants from all over the world. It felt like a feast of the nations because of the diversity of people and the joyful atmosphere that permeated the room. Housing is available with the Dominican Sisters of Springfield and the School Sisters of St. Francis at St. Martin de Porres Convent at 9041 South Exchange Avenue in Chicago, Illinois. What is speaking to your heart? It's a question I continue to ponder as we enter these final weeks of Lent. I've been especially attentive to my heart this Lenten season perhaps because Ash Wednesday was also Valentine's Day this year. It seemed odd to celebrate both on the same day but I suppose it was fitting for a season focused on conversion of heart, to start on a day dedicated to love. So where is my heart? In February, the Core Team met in St. Louis for our yearly face-to-face meeting. While our monthly Zoom meetings make our leadership structure possible from our various locations across the country, it is always good to get together in person, particularly for evaluation, long term planning, and direction setting. #SisterSpeak & Prayer Service Celebrate Women Religious During National Catholic Sisters Week! This year, Giving Voice teamed up with National Religious Vocation Conference, Communicators for Women Religious, Religious Formation Conference, and the Center for the Study of Consecrated Life, to both highlight Catholic sisters online with a series of video interviews, and create community at Catholic Theological Union in Chicago with a prayer service. Read more about #SisterSpeak & Prayer Service Celebrate Women Religious During National Catholic Sisters Week! Have you added your name to Giving Voice's Online Directory? It's quick and easy! Register here. Read more about Online Directory Call-out!
{ "redpajama_set_name": "RedPajamaC4" }
3,914
{"url":"https:\/\/crypto.stackexchange.com\/questions\/89416\/intuition-behind-commitment-challenge-response-a-k-a-sigma-protocols","text":"# Intuition Behind Commitment-Challenge-Response a.k.a. Sigma Protocols\n\nIn How To Prove Yourself: Practical Solutions to Identification and Signature Problems, Fiat and Shamir introduce a zero-knowledge identification scheme where\n\n1. The prover sends a commitment to the verifier\n2. The verifier sends a challenge to the prover (in shape of random coin tosses)\n3. The prover sends a response to the verifier\n4. The verifier checks if the response is correct\n\nThis is zero-knowledge and a Sigma protocol as described by Cramer in his PhD thesis.\n\nMy only experience with commitments is in fairly flipping a coin over telephone, in which case a commitment seems perfectly reasonable. But what is the basic intuition behind the commitment in the Fiat-Shamir case? Challenge and response are obviously part of the protocol, but why is the commitment needed? As far as I can tell, the commitment is not even revealed to the verifier in the Fiat-Shamir identification scheme!\n\n\u2022 Just intuitively and at a very high level: commitments are needed so that the prover does not change her mind when responding to the challenge. Typically, if you allow the prover to do this, then she will always be able to craft a response to the given challenge that passes whatever test the verifier runs. Apr 15, 2021 at 21:27\n\nAs a concert example lets consider the protocol for knowledge of discrete-logarithm.\n\nThe aim of the protocol: P wants to prove that it knows $$x$$ for $$pk=g^x$$.\n\n1. commitment: P chooses a random $$r$$ and commit to it by sending $$R=g^r$$ to the verifier.\n\n2. challenge: the verifier chooses a random value $$c$$ and sends it to P.\n\n3. P sends $$z=r+cx$$ to the verifer.\n\n4. verification: the verifier checks if $$g^z=R.pk^c$$ if yes it accepts it.\n\nNow what will happen if the prover does not commit to $$R$$ before sending $$z$$ ?\n\nAfter receiving $$c$$ from the verifier, P chooses a random $$z$$ and finds $$R'$$ from the equation $$g^z=R'pk^c$$, i.e., it sets $$R'=g^zpk ^{-c}$$. Then it sends $$R',z$$ to the verifier. By this $$R'$$ it can pass the verification check, and the point is that it even does not use $$x$$ in $$z$$. This means a malicious prover who does not have $$x$$ can prove the knowledge of $$x$$! Putting together, commitment is necessary to protect the proof system against prover such that just the honest provers can convince the verifier.\n\n\u2022 So $r$ serves to obfuscate $cx$? If $cx$ were just sent, no commitment used, the verifier could simply divide by $c$, which it knows, to retrieve $x$, right? Obfuscate in the sense that it renders retrieving $x$ by the verifier computationally infeasible. Apr 16, 2021 at 7:56\n\u2022 It is true, anyway we need a random $r$ to hide $cx$, but the point is that the prover should commit to it at the beginning. otherwise it can choose a random $r'$ depending on the challenge while it even does not have any x to hide! it is just to mimic the protocol and convince the verifier that I am doing my job! This was an example to show why commitment is important. Apr 16, 2021 at 9:07\n\u2022 OK, that answers my question, thanks. But now I wonder why we need the challenge at all. Why not commit to $r$ by sending $R = g^r$ to the verifier and then sending $z = r + x$? If the prover knows $x$, then the verifier accepts after verifiying $g^z = Rpk$ (fulfilling completeness), if the prover does not know $x$, then finding the correct $z$ reduces to efficiently solving the discrete logarithm problem, right? So the challenge is not needed at all, or is it? Apr 16, 2021 at 13:04\n\u2022 Is this about avoiding a replay attack, where the prover just resends $r$ and $z$ without knowing $x$ to seemingly prove he knows $x$? Apr 16, 2021 at 13:11\n\u2022 Imagine there is no challenge, then before sending any thing else (even before sending the commitment), the prover applies the previous attack, i.e., it chooses a random $z$ and then finds $R$ from equation $g^z=R.pk$. Then it sends this $R$ as its commitment. Thus, the order is very important, commit, challenge, response. Apr 16, 2021 at 14:30","date":"2023-02-07 19:00:49","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 1, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 20, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.7110974192619324, \"perplexity\": 769.6337430757824}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2023-06\/segments\/1674764500628.77\/warc\/CC-MAIN-20230207170138-20230207200138-00281.warc.gz\"}"}
null
null
I have listed a new property at 17358 1A AVE in Surrey. QUALITY built 5 bed/4 bath family home in desirable Summerfield, South Surrey. Boasting over 3,400 sq.ft of living space, this open concept home features a GOURMET KITCHEN w/oversized entertainers island and S/S appls, large family room w/gas fireplace and dedicated dining and living rooms. Gorgeous Master bedroom w/vaulted ceilings, walk-in closet & SPA-LIKE ensuite with soaker tub, rain shower and heated floors. FINISHED BASEMENT with large rec/theatre room, bar and potential for 1 bedroom suite w/ separate entry. Double Garage and SOUTH facing backyard with patio and gas hookup for BBQ. Only 5 min to White Rock, US Border, shopping, restaurants & Hwy 99. Like new and shows 10/10! Hurry, this one won't last! Open Sat/Sun 2-4pm.
{ "redpajama_set_name": "RedPajamaC4" }
9,426
class SMS @@account_sid = 'AC0b322d7367604e7a852a1d59193738a2' @@auth_token = 'c32bcf082cb7cee728a99832858db23b' @@client = Twilio::REST::Client.new(@@account_sid, @@auth_token) @@account = @@client.account def self.sieve(number) return number.to_s.gsub(/\D/, '') end def self.send_notification(user, thing) extra = (110-(thing.snow_cover.to_s.length+thing.full_address.length)) thing_name_length = thing.name.length requested = user.name.length+thing_name_length if requested < extra @@account.sms.messages.create(:from => '+18599030353', :to => user.sms_number, :body => user.name + ', ' + thing.name + ' might be surrounded by ' + thing.snow_cover.to_s + ' in. of snow. Location: ' + thing.full_address + '.') else requested_extra_difference = requested-extra @@account.sms.messages.create(:from => '+18599030353', :to => user.sms_number, :body => user.name + ', ' + thing.name.truncate(thing_name_length-requested_extra_difference) + ' might be surrounded by ' + thing.snow_cover.to_s + ' in. of snow. Location: ' + thing.full_address + '.') if (requested > extra) && (thing_name_length > requested_extra_difference) end end end
{ "redpajama_set_name": "RedPajamaGithub" }
1,217
Francesco di Neri di Ranuccio, known better as Francesco da Barberino (1264–1348), was a Tuscan notary, doctor of law and author. He first went to Florence to study in 1281. Between 1303 and 1314 and again between 1315 and 1317, he was exiled from the city on account of his Ghibelline political leanings. He worked for a time as a diplomat for Venice and passed his exile at courts in Provence, France and Spain. After his return to Florence, he gained respect as a notary and lawyer, but was not admitted to political office until the last three years of his life. He is best known for two didactic works on virtue, the Documenti d'amore for men and the Reggimento e costume di donna for women, written in a mixture of prose and verse in both Tuscan and Latin. They are notable for their citations of other authors from classical antiquity to his own time, including numerous vernacular authors not known from any other source. Life Francesco was born into a noble family in Barberino Val d'Elsa in 1264. He was not related to the Barberini. His father, Neri di Ranuccio, was a political Ghibelline who relocated from the city of Florence to the countryside. He was Francesco's first educator. Between 1281 and 1290, he studied the liberal arts in Florence. In the 1290s, he studied at the University of Bologna, where by 1294 he had he received his licence in civil law and became a notary. At some point he took minor orders. Upon his father's death in 1296, Francesco returned to Barberino. Between 1297 and 1303, he worked in Florence as a notary of the bishop of Florence. There he adopted the toponymic surname "da Barberino" in preference to the patronymic "di Neri di Ranuccio". In Florence, he met the painters Cimabue and Giotto. He married around 1303. In 1304, he was exiled from Florence as a Ghibelline. Francesco lived in Padua from 1304 to 1308, where he continued his studies in law and philosophy. In 1308, he moved to Treviso, perhaps to work for the podestà Corso Donati. From 1309 to 1313, he lived at various courts beyond the Alps. He was the ambassador of the Republic of Venice to Pope Clement V in Avignon, where he was the personal friend of Cardinal Pietro Colonna. He stayed at the court of King Philip IV of France in Paris and then that of King Louis I in Navarre. He continued to act as a mediator with Venice. On 28 March 1313, Francesco received a doctorate of both laws by papal bull, although the bishop of Florence would not confirm it for five years. In 1313, he returned to Italy, residing in Venice. His first wife died that year and he married Barna di Tanuccio Rinieri. He returned to Florence in 1314–1315, but soon left again for other cities, including Mantua. He returned definitively to Florence in 1317–1318. In 1321, Francesco acted as Florentine ambassador to Pisa. In 1322–1323, he was the executor of Bishop and advised the sculptor Tino da Camaino on the decoration of the bishop's tomb. He took part in the legal and economic affairs of the Florentine republic in his capacity as a doctor of both laws, but he was excluded from politics on account of his Ghibellinism until 1345, when he and his son Filippo were elected councillors. In 1347, he and Francesco Salviati were named consuls of the college of judges. In 1348, he was elected to the Signoria. Among his fellow Florentine literati, Francesco had a good reputation for his public service. Filippo Villani included an entry on Francesco in biographical dictionary of Florence and Giovanni Boccaccio praises him in Genealogia deorum gentilium. Francesco died in early April 1348 of the Black Death. He was buried in the basilica of Santa Croce. He had five children, all sons, by his first wife: Galazzo, Filippo, Antonio, Taddeo and Raineri. He had another child by his second wife. Works In Bologna, Francesco first came into contact with the literary circle influenced by Guido Guinizzelli. During his second period in Florence, he became acquainted with the stilnovisti, including Dante Alighieri and Guido Cavalcanti. His earliest known works date to this period, but are both lost: some lyric poems for a certain Costanza and the Flores novellarum (Italian Fiori di novelle), a collection of tales of Provençal origin. During his transalpine sojourn, he wrote a letter in Latin to the Emperor Henry VII. Like Dante, he was convinced that a powerful emperor was the key to peace in Europe. It was probably also during his exile that he wrote his two surviving longer works in the Tuscan vernacular, Documenti d'amore ('lessons about love') and Reggimento e costume di donna ('rules on good behaviour for women'). His citation of authors from Aristotle down to Jean de Joinville, in both Latin and various vernaculars, demonstrates the breadth of his reading. His knowledge of the Occitan poetry of the troubadours was unparalleled in contemporary Italy. Francesco's works were not widely read, but they are now regarded as standing at the head of the Italian etiquette tradition of Il Cortegiano and Il Galateo. Documenti d'amore The Documenti was probably begun by 1309 and completed by 1315. It is a didactic work aimed at a male bourgeois audience. It is written in Tuscan verse with accompanying Latin translation and explanatory glosses by Francesco. It is an allegorical work, divided into twelve sections of unequal length, each containing the words of a female personification of a virtue, relaying the precepts of chivalry and good behaviour that Love dictated to Eloquence. The verse forms vary. Most of the citations of other authors are in the Latin commentary. The Documenti is a historically valuable text. It contains evidence for dating the composition of the Divine Comedy. It also contains citations of 13th- and 14th-century Occitan, Italian and French poets known from no other source. Francesco was also up to date with the latest literary trends in Italy. Besides Alighieri, Guinizzelli and Cavalcanti, he cites Brunetto Latini, Dino Compagni, Baldo da Passignano and Albertino Mussato. The Documenti is contained in five surviving codices. Two now in the Vatican Library are Barb.lat. 4076 and Barb. lat. 4077. The former is an autograph manuscript, which Francesco himself copied and illustrated, while the second is a partial autograph, the illustration of which was supervised by Francesco. Reggimento e costume di donna The Reggimento was completed by 1320, slightly later than the Documenti. Francesco claims to be writing at the request of an unnamed lady. It is a courtesy book that complements the Documenti for women. Divided into twenty sections on different virtues, it was intended to be illustrated. It is written entirely in Tuscan, but in a mixture of prose and verse. Certain passages are reminiscent of the cantari. It is of interest as a historical source for the domestic customs and interests of women. It contains the only surviving verses (in translation) of the trobairitz Lisa de Londres. The Reggimento is known from two copies in a single manuscript tradition, both now in the Vatican: Barb.lat. 4001 and its copy, Capponiano 50. Notes Bibliography External links 1264 births 1348 deaths People from Barberino Val d'Elsa Canon law jurists University of Bologna alumni Writers from Florence 13th-century people of the Republic of Florence 14th-century people of the Republic of Florence 13th-century Italian poets 14th-century Italian poets Latin letter writers Ambassadors of the Republic of Venice to the Holy See
{ "redpajama_set_name": "RedPajamaWikipedia" }
2,592
package com.wangzhihao.blackmarket.service; import com.wangzhihao.blackmarket.domain.CourseSchedule; import org.springframework.stereotype.Service; import java.util.List; import java.util.Map; import java.util.Set; /** * Description * <p> * </p> * DATE 2017/11/30. * * @author Wang Zhihao. */ @Service public interface CourseScheduleService { void add(CourseSchedule courseSchedule); CourseSchedule getById(Long id); List<CourseSchedule> getCourseSchedulesByCourseId(Long courseId); Map<Long, List<CourseSchedule>> getCourseSchedulesByCourseIds(Set<Long> courseIds); }
{ "redpajama_set_name": "RedPajamaGithub" }
3,889
Chat anonymously with community members from all industry sectors to share good practices, expertise and advice on suppliers, spare parts, and many more! Easily import your data to Mobility Work for free from any database, former CMMS software or Excel file before you start your trial. Only a few clicks are needed to find the required piece of equipment in your nomenclature. Mobility Work's ultra-efficient search engine allows you to quickly find a piece of equipment, intervention, service provider, spare part, etc. Using Mobility Work's mobile interface and equipment catalogue, it takes less than a minute to create a piece of equipment. You no longer need to fill in thousands of fields: gather all your data in only one field, and the search engine will do the rest. Using Mobility Work's mobile interface and equipment catalogue, it takes less than a minute to create a piece of equipment. You no longer need to fill in thousands of fields. Search for the equipment on Mobility Work, Next Gen Maintenance Management Platform and directly duplicate the file and related data provided by the official supplier or a member of Mobility Work's community. Mobility Work also allows you to quickly and easily access electrical schemes, plans, pictures, videos, and audio files which are attached to a piece of equipment or a maintenance intervention. Edit your privacy settings to share information with the community. Improve your monitoring in the field, and add several counters (electrical, mechanical, production data, etc.) to gain precision and trigger automatic maintenance operations from counter values. Use Mobility Work's APIs to connect with any tool or sensor. You can for example trigger a maintenance intervention as soon as the temperature reaches 30°C in your oven. Edit the privacy settings of all your data at any time (equipment, spare parts, documents, etc.). Set them as private, share them with your company only or make them public so that the whole Mobility Work community will benefit from them. Please remember that all companies' and employees' names are not publicly available. With the Mobility Work CMMS mobile app, you can automatically add GPS coordinates to your piece equipment, and display it on a map in order to gain time. Do not miss any intervention, be mobile! Mobility Work's CMMS is meant to be at the heart of the company's information system. Download the Android or iOs mobile app to enable your maintenance technicians to work and report in real time. Allow any user to access the latest or planned repairs at anytime, from anywhere, through the newsfeed of his/her company, group or of the whole community. It is also very easy to search for interventions by typing keywords in the quick search or directly via a related-equipment file. Easily schedule your maintenance interventions calendar, through manual alert configuration, regulatory maintenance plan, measured values or even connected sensors. Make sure that no intervention has been forgotten thanks to a clear follow up and an easy calendars sharing. Adopt RED YELLOW GREEN colour code and visualize at one glance the past and coming tasks. Follow every intervention report in real time thanks to your timeline. Assign the proper technician to each scheduled intervention. Easily schedule your calendar preventive maintenance interventions, from measured values or connected sensors. The Mobility Work CMMS allows your team to be more effective, to improve your maintenance team management and to make sure that no intervention has been forgotten. Manage your spare parts orders in your ERP, and connect these data to Mobility Work to enrich your CMMS tool. For more information on this topic, do not hesitate to schedule a presentation with our team members. Need to carry out a maintenance intervention on a piece of equipment? There is nothing easier with Mobility Work! Select your piece of equipment using the search engine or a NFC chip; describe the intervention using speech recognition, and assign a technician. All users will be notified by email and through the company's news feed. Mobility Work's CMMS is meant to be at the heart of the company's information system. It allows any user to access the latest repairs, production intervention requests, scheduled interventions or the latest updated documents. Your data belong to you: export your interventions, equipment or spare parts lists at any time to CSV, XLS, PDF, etc. Access entered and removed spare parts in Mobility Work, and see quantities, location, minimum and maximum stock. Remember that every time a spare part is used on an intervention, the piece of equipment's nomenclature is automatically updated. This feature can be interfaced with your ERP. Each piece of equipment owns its own history, which allows you to schedule, manage and follow your supplier's or service provider's interventions and to rate them according to several criteria (cost, quality, deadline, safety). Mobility Work enables you to financially assess all your service providers, to follow your maintenance budget more accurately. You think that it stops here?
{ "redpajama_set_name": "RedPajamaC4" }
4,030
Q: Does existence of limit $x\to 0$ of derivative imply function is differentiable at 0? Suppose $f\in C(\mathbb R)\cap C^1(\mathbb R\setminus \{ 0\})$ and $\lim_{x\to 0}f'(x) $ exists and equals $L\in\mathbb R$ . Does this imply that $f\in C^1(\mathbb R)$? Derivatives cannot have jump singularities, since they satisfy the intermediate value property. Therefore, if $f'(0)$ exists, its value must be $L$, and $f$ is then $C^1$. But this doesn't prove the result. I also can't imagine a counterexample... Remarks: * *This seems simple enough that it should have been asked before, but I can't find a duplicate. Moreover, (see rem. 3) its not so unnatural to assume its true. This is kind of close. Most questions with similar names seem to be about multivariate calculus. *Note the standard example of a function whose derivative is not continuous at 0 does not help ($f(x) = x\sin(1/x)$) *If true, this result would validate attempts to prove the differentiability of e.g. $e^{-1/|x|}$ by establishing the existence of the above limit instead. A: By l'Hopital's rule you have that $$ \lim_{x\to 0} \frac{f(x) - f(0)}{x} = \lim_{x\to 0} f'(x) = L, $$ hence $f$ is differentiable at $0$ and $f'(0) = L$. We can then conclude that $f\in C^1(\mathbb{R})$.
{ "redpajama_set_name": "RedPajamaStackExchange" }
9,601
Posted by ChoiceTvTeam Sequel announced to Stieg Larsson's Girl With the Dragon Tattoo trilogy The Swedish publisher of the best-selling The Girl With the Dragon Tattoo trilogy says it has hired an author to write a sequel to the series by Stieg Larsson, who died in 2004. Norstedts said on Tuesday that it had signed a contract with David Lagercrantz, ghost writer of I am Zlatan Ibrahimović, the autobiography of the Swedish footballer, for a new book about journalist Mikael Blomqvist and hacker Lisbeth Salander that is scheduled to be published in August 2015. The head of publishing at Norstedts, Eva Gedin, told AP that the book would be an original work that includes nothing from the fourth book in the series, which Larsson began writing but remained unfinished when he died. Larsson's trilogy has been a huge international success, with more than 75m books sold in some 50 countries.
{ "redpajama_set_name": "RedPajamaCommonCrawl" }
5,304
Q: How can you quit from dialog in C# bot-framework? I'm starting a project for a ChatBot in C# with the bot framework. I choose the ContosoFlowers example for learning about the use of bot framework. In the AddressDialog users cannot quit the dialog after they enter to it without providing an address. How can I update the code so when users reply "Cancel" or "Abort" or "B" or "Back" they quit the dialog? namespace ContosoFlowers.BotAssets.Dialogs { using System; using System.Collections.Generic; using System.Linq; using System.Threading.Tasks; using Extensions; using Microsoft.Bot.Builder.Dialogs; using Microsoft.Bot.Connector; using Properties; using Services; [Serializable] public class AddressDialog : IDialog<string> { private readonly string prompt; private readonly ILocationService locationService; private string currentAddress; public AddressDialog(string prompt, ILocationService locationService) { this.prompt = prompt; this.locationService = locationService; } public async Task StartAsync(IDialogContext context) { await context.PostAsync(this.prompt); context.Wait(this.MessageReceivedAsync); } public virtual async Task MessageReceivedAsync(IDialogContext context, IAwaitable<IMessageActivity> result) { var message = await result; var addresses = await this.locationService.ParseAddressAsync(message.Text); if (addresses.Count() == 0) { await context.PostAsync(Resources.AddressDialog_EnterAddressAgain); context.Wait(this.MessageReceivedAsync); } else if (addresses.Count() == 1) { this.currentAddress = addresses.First(); PromptDialog.Choice(context, this.AfterAddressChoice, new[] { Resources.AddressDialog_Confirm, Resources.AddressDialog_Edit }, this.currentAddress); } else { var reply = context.MakeMessage(); reply.AttachmentLayout = AttachmentLayoutTypes.Carousel; foreach (var address in addresses) { reply.AddHeroCard(Resources.AddressDialog_DidYouMean, address, new[] { new KeyValuePair<string, string>(Resources.AddressDialog_UseThisAddress, address) }); } await context.PostAsync(reply); context.Wait(this.MessageReceivedAsync); } } private async Task AfterAddressChoice(IDialogContext context, IAwaitable<string> result) { try { var choice = await result; if (choice == Resources.AddressDialog_Edit) { await this.StartAsync(context); } else { context.Done(this.currentAddress); } } catch (TooManyAttemptsException) { throw; } } } } A: You just need to handle the quit conditions at the top of the MessageReceivedAsync method. At the top of the dialog you can add something like private static IEnumerable<string> cancelTerms = new[] { "Cancel", "Back", "B", "Abort" }; And also add this method: public static bool IsCancel(string text) { return cancelTerms.Any(t => string.Equals(t, text, StringComparison.CurrentCultureIgnoreCase)); } Then is just a matter of understanding if the message sent by the user matches any of the cancelation terms. In the MessageReceivedAsync method do something like: public virtual async Task MessageReceivedAsync(IDialogContext context, IAwaitable<IMessageActivity> result) { var message = await result; if (IsCancel(message.Text)) { context.Done<string>(null); } // rest of the code of this method.. } You can also go a bit more generic and create a CancelableIDialog similar to what was done in the CancelablePromptChoice.
{ "redpajama_set_name": "RedPajamaStackExchange" }
9,785
Q: Template MastersDoctoralThesis doesn't recognize the current babel dictionary: "Undefined control sequence. \begin{document}" I would like to use the Template MastersDoctoralThesis unfortunately it creates an error message which I am not able to debug. The main structure of this template is a main.tex file which calls a class MastersDoctoralThesis.cls. When I select in the options of the document-class a language, say ngerman, the following way: \documentclass[ 11pt, % The default document font size, options: 10pt, 11pt, 12pt ngerman, % ngerman for German, english for English singlespacing, % Single line spacing, alternatives: onehalfspacing or doublespacing not load the hyperref package headsepline, % Uncomment to get a line under the header ]{MastersDoctoralThesis} % The class file specifying the document structure It gives the following messages: Error: Undefined control sequence. \begin{document} , Warning: file 'translations-basic-dictionary-german.trsl' does not appear to be a dictionary When I change the language to english, than the error remains but the warning changes to Warning: file translations-basic-dictionary-english.trsl' does not appear to be a dictionary` When I am not using the option in the documentclass at all and simply pass \RequirePackage[english,ngerman]{babel} in the MastersDoctoralThesis.cls. I receive: Error: Undefined control sequence. \begin{document}, Error: Undefined control sequence. \begin{document}, Warning: file translations-basic-dictionary-german.trsl' does not appear to be a dictionary, Warning: file translations-basic-dictionary-english.trsl' does not appear to be a dictionary Accidentally, I forgot most of the time till now the , after the language which strangely worked, till now when I wanted to use the commands in the class which rely on that option as well. These basically change the descriptions in the document. The problem occurs independently of the compiler Latex, Lualatex and PDFLatex Please, help me find to solution. SvA Edit: In the log-file the interesting part looks like this: ("C:/Program Files/MiKTeX 2.9/tex/latex/psnfss/t1ppl.fd" File: t1ppl.fd 2001/06/04 font definitions for T1/ppl. ) ! Undefined control sequence. \@trnslt@check@dictionary #1#2->\AddToHook {file/after/#1-\@trnslt@language ... l.90 \begin{document} The control sequence at the end of the top line of your error message was never \def'ed. If you have misspelled it (e.g., `\hobx'), type `I' and the correct spelling (e.g., `I\hbox'). Otherwise just continue, and I'll forget about whatever was undefined. Package translations Warning: file `translations-basic-dictionary-english.trsl' does not appear to be a dictionary on input line 90. (D:/Users/user/AppData/Roaming/MiKTeX/2.9/tex/latex/translations/dicts/tran slations-basic-dictionary-english.trsl File: translations-basic-dictionary-english.trsl (english translation file `tran slations-basic-dictionary') ) Edit 2: Now, I checked, whether it maybe is an issue with my language styles. I created two minimal examples of documents applying the classes scrreprt and MastersDoctoralThesis. The first one works fine, the second throws the error again.
{ "redpajama_set_name": "RedPajamaStackExchange" }
3,281
namespace thewizardplusplus { namespace anna { namespace maths { class Maths { public: static const float PI; static const int DEFAULT_BITS_NOT_COMPARE = 4; static bool isEqual(float x, float y, int bits_not_compare = DEFAULT_BITS_NOT_COMPARE); static bool isEqual(double x, double y, int bits_not_compare = DEFAULT_BITS_NOT_COMPARE); static bool isEqual(long double x, long double y, int bits_not_compare = DEFAULT_BITS_NOT_COMPARE); static bool isNull(float x, int bits_not_compare = DEFAULT_BITS_NOT_COMPARE); static bool isNull(double x, int bits_not_compare = DEFAULT_BITS_NOT_COMPARE); static bool isNull(long double x, int bits_not_compare = DEFAULT_BITS_NOT_COMPARE); static float toDegrees(float angle); static float toRadians(float angle); static float clamp(float number, float limit1 = 1.0f, float limit2 = 0.0f); static float interpolate(float x1, float x2, float shift); static float randomNumber(float limit1 = 1.0f, float limit2 = 0.0f); private: static bool initialize_prmg; }; } } } #endif
{ "redpajama_set_name": "RedPajamaGithub" }
4,704
Tropical Cyclone Shaheen produced as much as four years' worth of rain in Oman Up to 14½ inches (368mm) fell in Al Khaburah, which is just to the west of where the storm came ashore. The city of about 40,000 people averages between three and four inches (75 to 100mm) of rain per year. Shaheen was the first tropical cyclone to make landfall along Oman's northern coast since 1890 and only the third on record. None have been observed in the era of weather satellites (since the 1960s) or during the month of October. According to meteorologists Bob Henson and Jeff Masters, writing for Yale Climate Connections, it struck farther west in Oman than any previous tropical cyclone. Henson and Masters wrote that Shaheen fed off the very warm waters of the Gulf of Oman, which were several degrees warmer than normal. Ocean temperatures have been rising worldwide because of human-caused climate change. Tags floods ← The Anthropocene causes 23 species to be declared extinct → 34 inches (864mm) of rain in 24 hours across parts of Northern Italy
{ "redpajama_set_name": "RedPajamaCommonCrawl" }
8,993
Copyright © 2017 by Miriam C. Davis All rights reserved Published by Chicago Review Press Incorporated 814 North Franklin Street Chicago, Illinois 60610 ISBN 978-1-61374-871-8 **Library of Congress Cataloging-in-Publication Data** Is available from the Library of Congress. Typesetting: Nord Compo Printed in the United States of America 5 4 3 2 1 _This digital document has been produced byNord Compo_. FOR MY BROTHER TIM, WHO GAVE ME THE IDEA # Contents Preface 1 Evil Descends 2 The Cleaver 3 Dagoes, Sugarcane, and Muffulettas 4 The Davi Murder 5 The Black Hand 6 The Cleaver Returns 7 A German Spy? 8 Axeman Hysteria 9 The Mysterious Axeman's Jazz 10 "Hung by the Neck Until Dead, Dead, Dead" 11 Verdict 12 False Lead 13 Rosie and Saint Joseph 14 The Final Chapter? Acknowledgments Notes Bibliography Index # Preface ONE HOT SUMMER AFTERNOON about thirteen or so years ago, my brother Tim and I were sitting on my front porch drinking bracingly cold English cider. The conversation turned, as it always does on these occasions, to the subject of serial murder. . . . Tim told me about a case he had read about as a boy, about a killer who went around New Orleans whacking Jewish bakers with an axe. Ever since I read Ann Rule's account of her erstwhile friend Ted Bundy, I have been fascinated, in a scared sort of way, with serial killers. I don't know what moved me to look into this further; I was deep into another topic at the time. Probably it was nothing more than the urge to procrastinate. But look into it I did, and found that the New Orleans case involved not Jewish bakers but Italian grocers. Eventually, I got hold of Robert Tallant's _Ready to Hang_ , which formed the basis of most other accounts of the Axeman. Robert Tallant was a New Orleans writer of the mid-twentieth century who, according to one reference librarian in the New Orleans public library, may well have gotten much of his information by hanging out in New Orleans bars. In a chapter titled "The Axman Wore Wings," Tallant tells the story: On the morning of May 24, 1918, Joseph and Catherine Maggio, Italian immigrants who ran a small grocery, were discovered dead and bloody in their bedroom. They'd been assaulted with their own axe and had had their throats cut. Nothing appeared to have been stolen. The assailant had gotten in by cutting out a panel in the back door. In the course of their investigation, police found a peculiar message written on the banquette (an old New Orleans term for sidewalk) near the Maggios' home and grocery: "Mrs. Maggio is going to sit up tonight just like Mrs. Toney." The police remembered that seven years earlier, in 1911, three other Italian grocers had been murdered with an axe; in two cases their wives had also been slain. Tallant gave the grocers' names as Cruti, Rosetti, and Tony Schiambra. Was Tony Schiambra's wife the "Mrs. Toney" of the sidewalk message? Tallant asked. Was this a message from the Italian Mafia? Had the couple been slaughtered by gangsters for some unknown misdemeanor? No one was ever charged with the murders, so these questions were never satisfactorily answered. The next month Louis Besumer, a Polish grocer, and Harriet Lowe, the woman who lived with him, were also attacked with an axe and seriously injured. Again, the weapon had been the grocer's own. Again, a door panel had been removed. And again, nothing was reported stolen. This time both victims survived, at least temporarily, and Lowe, who changed her story several times, accused Besumer first of being a German spy and later of having tried to kill her. When she died of her wounds two months later, Besumer was charged with murder. Over the next fourteen months, the killer, nicknamed "the Axeman" by the press, racked up a litany of victims: Mrs. Edward Schneider, August, 5, 1918 (survived); Joseph Romano, August 10, 1918 (died); Charles, Rosie, and Mary Cortimiglia, March 10, 1919 (Mary died); Steve Boca, August 10, 1919 (survived); Sarah Laumann, September 3, 1919 (survived); and Mike Pepitone, October 27, 1919 (died). In almost all cases the modus operandi was the same: back door panel cut out, the victim's own axe used, weapon abandoned at the scene, and nothing stolen. Most but not all of the victims were Italian grocers. Fear of the Axeman paralyzed the immigrant community. Some terrified Italians couldn't sleep at night without posting guards to stand watch. Phantom Axemen were seen everywhere. Stories circulated about grocers waking in the morning to find a door panel chiseled off and an axe outside their door. Louis Besumer was acquitted of Harriet Lowe's murder, but his was not the only trial. Rosie Cortimiglia accused two neighbors, elderly grocer Iorlando Jordano and his son Frank **,** of having attacked them and having killed their two-year-old daughter out of business rivalry. Even though her husband, Charles, testified that the killer had not been either man, father and son were convicted of murder and eighteen-year-old Frank was sentenced to death. Later, Rosie admitted that she had lied because she hated the Jordanos. Shortly after the Cortimiglia attack, the _New Orleans Times-Picayune_ received a letter purporting to be from the murderer. He was "a fell demon from the hottest hell," the letter claimed, and would descend on New Orleans the coming Tuesday night looking for a victim, sparing anyone listening to jazz. Tallant reported that the designated evening, March 19, was Saint Joseph's Night, and it was "the loudest and most hilarious of any on record" as jazz blared all over the city. After the death of grocer Mike Pepitone in August 1919, the attacks finally stopped. And in Tallant's account, the story seems to have had a dramatic—and satisfactory—sequel out in California: On December 7, 1920, Esther Pepitone, the widow of Mike Pepitone, shot and killed a New Orleans man named Joseph Mumfre in Los Angeles. She told the police that Mumfre was the Axeman, that she had seen him as he fled after murdering her husband. Convicted of his murder, Tallant wrote, she served only three years before being freed. According to _Ready to Hang_ , Joseph Mumfre was a career criminal who was well known to the New Orleans police. And the dates he'd been in and out of prison matched the dates when the Axeman attacked and when he seemed to have disappeared. So, was Mumfre the Axeman? Tallant said that the evidence was only circumstantial, and most New Orleanians believed that more than one killer was responsible. The two main theories were that the Mafia was responsible, or the killer was a "homicidal maniac," a "Doctor Jekyll and Mr. Hyde" type as suggested by one of the investigating detectives. This was the story that I found on the web and in crime anthologies when I first began researching the Axeman case. Most of the available sources basically repeated Tallant's version of events. While Tallant left the question of the Axeman's identity open, I found that some writers seem to have swallowed whole the theory that Joseph Mumfre was the killer. In a list of male serial killers, for instance, forensic psychologist Eric Hickey listed "Joseph Mumfre" as the Axeman of New Orleans. In _Bloodletters and Badmen_ , crime writer Jay Robert Nash argued that Mumfre was a professional killer who used an axe to kill Pepitone and other victims as part of a Mafia vendetta. Others, however, questioned aspects of Tallant's account. In an early edition of his _Hunting Humans: The Encyclopedia of Serial Killers_ , vol. 1, crime writer Michael Newton reported that city records showed that no Italian grocers named Cruti, Rosetti, or Schiambra had been killed—by an axe or anything else—in 1911. I had an opportunity to check out some of these questions in the early 2000s soon after I became interested in the Axeman case. At the time, my husband and I visited New Orleans once or twice a year. On one of our visits, I popped out of the French Quarter to visit the city archives, located in the public library conveniently across Canal Street in the Central Business District. Examining homicide and coroners' records for 1911, I discovered that while no Cruti, Rosetti, or Schiambra had been murdered that year, a sleeping Italian grocer named Joseph Davi had been—his skull fractured and his wife cut up in the middle of the night. That such a basic fact about the case and one so easily discoverable had been missed suggested that the story had never really been properly investigated. Maybe, I thought, there's a book here. Since I began working on _The Axeman of New Orleans_ , two well-done works on the subject have appeared (both times nearly giving me a heart attack): Keven McQueen's chapter on the Axeman in _The Axman Came from Hell_ , and the account given by Gary Krist in _Empire of Sin_. If either narrative had been available when I first read about the Axeman, I might have never written this book. Both are useful, well-researched explorations of the Axeman's crimes. But both, I will argue, are incomplete. They don't answer questions that can, in fact, be answered. In the course of examining the tale of the Axeman, moreover, I became fascinated by the experience of the Italian—usually Sicilian—grocers he preyed on, particularly in the case of the Jordanos, the father and son accused by Rosie Cortimiglia. The teenaged Frank and his elderly father, Iorlando, were as much casualties of the Axeman as any of his other victims. The story of Frank and his family merits telling because it illustrates the experience of Italian immigrants and the niche they carved out for themselves in the social hierarchy of early twentieth-century Louisiana as well as the social prejudice against them. Mafia aficionados shouldn't worry. The story also involves the gangsters of the Sicilian underworld. What follows is the result of my own investigation into the Axeman of New Orleans murders based on an examination of all available sources. This has been a different kind of history than I was trained to write in graduate school. I wanted to write something analytical that would be fully grounded in the sources but would also appeal to a wide audience by telling a story rather than merely being an analysis of a list of murders. A college friend of mine who was studying English once remarked that the great thing about being a lit major was that unless someone could prove you absolutely wrong, you could say anything you wanted. I assure the reader that this is _not_ the approach I have taken here. While I've tried to tell the story as what is sometimes called creative or narrative nonfiction, I've stayed true to my obligations as a historian. Everything I write can be justified on the basis of what is found in the historical record, although I've sometimes had to use my judgment when choosing between conflicting accounts in, for example, rival newspapers. On occasion, I've relied on deduction and common sense to connect dots and piece together what must have happened and what must have been experienced by the individuals in my story. Anything in quotation marks is from a written document, usually a newspaper, police report, or court record. Sometimes these quotations have been condensed or edited for clarity, especially when they're based on varying accounts found in different newspapers. I have sometimes modernized spelling. _Italicized_ dialogue has been imaginatively reconstructed when I felt that something like it must have been expressed but didn't have direct evidence of exactly what was said. I've been careful to remain faithful to the historical evidence and context. In some cases I've inserted personal thoughts or feelings—but, again, always based on deductions from the historical evidence. Full passages of italicized text indicate a re-creation of scenes that no one but the killer himself was witness to. For insight into the elusive killer, I've relied on modern forensic psychology. Endnotes have been consolidated as much as possible, while still providing full documentation. Anyone wishing to retrace my footsteps should be able to do so. Among the questions that I began my research hoping to answer were: * Were all the killings named by Tallant actually Axeman crimes? * Did the Axeman murders begin with the murder of Joseph and Catherine Maggio? * Was the Mafia involved? * Were all the Axeman's victims Italian, and if they were, why did he target them? * Why did Rosie Cortimiglia accuse Frank Jordano and his father of murdering her daughter? Did she later change her story, and if so, why? * Did Esther Pepitone murder Joseph Mumfre? * Most important, was Joseph Mumfre the Axeman? And if he wasn't, who was? * How do the answers to these questions help one understand what it was like to live in New Orleans during the Axeman's reign of terror? In the course of my investigation, I discovered that Tallant was wrong in many details, some unimportant, some not. At the end of "The Axman Wore Wings," he wrote, "It is extremely doubtful that anyone will ever know more" about whether Mumfre was the Axeman. He was wrong about that. # ≡ 1 ≡ # Evil Descends ## 3 AM, Sunday, March 9, 1919 T _HE KILLER BALANCED ON_ _the chair as he looked through the window, straining to see the figures in the bed. It was a moonless night; he couldn't see much by the light of the one electric bulb dimly illuminating the room. It didn't matter. He knew they were there. He went around to the back and broke in in the way for which he was famous, carefully chiseling out a panel of the heavy kitchen door. He worked steadily, taking his time, pausing when he needed to; no reason to hurry. When the panel finally came off, he slipped his gloved hand through the door and eased back the bolt. Once in, he moved surely through the tiny kitchen, in stocking feet so his heavy boots wouldn't clatter on the wood floors, straight through the dining room, past the door leading to the small grocery, and stopped at the bedroom door. He paused, listening to the breathing of the sleeping figures. Then, getting a tighter grip on the heavy axe, he entered the bedroom._ _Minutes later, his work done, he walked unhurriedly through the dining room and the kitchen, out the door, and down the steps. He tossed the axe carelessly under the house, human hair still clinging to the quickly drying blood. Sitting down on the stoop, he pulled his boots back on, tugging at them firmly before lacing them up. Then he vanished into the night. The Axeman of New Orleans had claimed another victim._ Carnival had just ended. Only four days before, the murderer and his victim had both jostled with the crowds flooding the streets of New Orleans. Mardi Gras was meant to be a sedate affair this year, very unlike the one two years before in 1917. That year it was celebrated with all its magical grandeur, perhaps with a little more intensity than usual given that anyone could see that the United States was about to go to war. The Carnival season, between Epiphany and Lent, was a season of balls, pageants, parades, and general exuberance. It culminated in two hedonistic days before Ash Wednesday. New Orleanians were not reticent about enjoying themselves in anticipation of Lent; maybe they thought they needed the pagan holiday to get through the penitential season. Each year on the eve of Mardi Gras, Rex—King of Carnival, Monarch of Mirth—arrived at noon, steaming up to the landing at the foot of Canal Street in his royal yacht, where he was met by thousands of his frolicsome subjects. Dressed in white satin and silver cloth, he led his parade through the heart of the city, accompanied by mounted police, artillery battalions, sailors, marines, National Guardsmen, and Boy Scouts. At city hall, the mayor presented Rex with the keys of the city, inaugurating his merry thirty-six-hour rule. On Fat Tuesday, Rex again processed through the streets, leading a parade of floats so elaborate they took their krewes a full year to create. In 1917, framed in a great golden crown, Rex had led a dazzling pageant of twenty floats whose theme, "The Gift of the Gods to Louisiana," used glittering images from ancient mythology to illustrate the wonders of the state. The crowds loved it. Mardi Gras wasn't just about parades and pageants. On Fat Tuesday the population itself took to the streets in costumes and masks. Clowns, gypsies, elves, and pirates packed into Canal Street and danced to jazz bands playing on street corners; red devils and black-faced minstrels added a touch of the grotesque, children dressed as bumblebees a touch of the comical. The city center became a playground for high-spirited antics. Identities safely hidden, people dared what they'd never do unmasked. Maskers danced with complete strangers; respectable women invaded bars; masked matrons peeked into houses of ill repute on Basin Street. But in 1919, Rex announced that this year he would not leave his palace in Araby the Blest to descend upon New Orleans. Celebrations had been canceled in 1918 because of the war. Such frivolity didn't seem quite right with American boys dying in French mud. The thousands who enjoyed the pageants and parades and costumes swallowed their disappointment and looked forward to the resumption of festivities at the end of the war. March 1919, however, just four months after the end of a conflict that left over 116,000 Americans dead, still seemed too soon. The celebration of Carnival was scaled back. A few modest parades were planned, private masked balls were permitted, but the _Times-Picayune_ tried to lower expectations: "There will be no gorgeous pageants to fill the streets with a blaze of color and light." Apparently, no one checked with the people of New Orleans. By ten on the morning of Mardi Gras, the spirit of Carnival had overwhelmed the city fathers' reluctance and spontaneously impelled thousands of masked merrymakers into the streets. Costumed in bright silks, satins, and velvets, revelers swaggered up and down the banquettes, hung out of slick new automobiles, and hitched rides on horse-drawn wagons. The twang of banjos and the reedy hum of clarinets floated through the air. A truck carrying a calliope pushed its way through the crowds, adding its steam-driven whistles to the general din. Elves, gnomes, hula dancers, Spanish dons, Cherokee warriors, and harem beauties made their way down Saint Charles Avenue. Cross-dressing was surprisingly popular, with young women uniformed as soldiers competing for attention with grown men made up as Japanese geishas. No "modest, even . . . somber Carnival" was this. By midday any pretense that the day was in any way normal had vanished; the city was one big party. Businesses closed. Housewives found themselves abandoned by household help who had joined the masked throngs. Even the weather cooperated, an approving sun driving out threatening clouds. Frank Jordano and his girlfriend Josie Spera joined the flood of humanity that darkened the Central Business District, floating along with the crowd down Canal Street, laughing at the masked sprites, the satin-clad cavaliers, the women dressed as redbirds. They tried to listen to the primitive jazz band at the corner of Saint Charles and Canal, heroically attempting to make its ragtime heard over the cacophony of the crowd. They gaped at the red devils and harlequins driving delivery wagons. They ate hot dogs on Canal Street from a vendor hawking them from the back of his wagon. All day and into the night, they shared the streets with the happy crush of surging humanity. They also shared the streets with a more sinister companion. He, too, enjoyed the crowds and the masks and the music. But his was a malevolent spirit that threatened Frank and Josie and their future happiness. That night, that memorable Mardi Gras, Frank was a happy young man, ambitious and optimistic. An exemplary son of Sicilian immigrants, he worked hard and made big plans. At seventeen he was already an insurance agent and engaged to Josie, a sweet local girl; he anticipated a flourishing American life, a happy family, a prosperous business. But three nights after Mardi Gras, Josie had a dream. She dreamed that evil was about to descend on the neighborhood. She was prescient. Frank's life was about to become a nightmare. Screams tore through the quiet Gretna neighborhood on an otherwise tranquil Sunday morning. Hazel Johnson, a young black woman, bolted out of the Cortimiglias' combination residence and grocery, yelling for someone, anyone to help—the Cortimiglias had been murdered! Frank Jordano was upstairs in bed when his twenty-year-old sister Lena's hysterical cries—"Oh, Jesus! Oh, Jesus! Oh, Jesus!"—punctured his sleep. Panic-stricken, thinking something had happened to their mother, he tumbled down the stairs, dressing on the way. Facing his sister, shirt unbuttoned, shoes on without socks, he demanded, "Lena, what's the trouble? Is it Mama?" "They're dead," she wept. "Mr. Cortimiglia, Mrs. Cortimiglia, and the baby are dead." Dazed, he stared at her: "Do you mean that?" "Yes," she insisted. "Hazel Johnson came running out hollering that they were dead." At that moment, he looked up to see Ella Kennedy, Hazel's aunt who had accompanied her on her errand to the store, coming out of the alley that led to the back of the Cortimiglias' place, screaming that the baby was dead. Still buckling his belt and buttoning his shirt, Frank raced into the alley that separated his home from that of the Cortimiglias, ran down it and around to the kitchen entrance to the Cortimiglia home. Despite the recent dispute between the families over rental property, he liked the Cortimiglias. And he adored their little girl. Sprinting up the steps to the little house, he joined the growing crowd that crammed into the single, small bedroom. Peering over the heads of the others who'd already arrived, what he saw there changed his life. Charles Cortimiglia and his young wife Rosie lay draped across their bed from opposite sides; the body of their dead toddler lay still between them. The room was soaked in crimson. Blood drenched the bed; it speckled the wall and stained the curtains; it pooled on the floor. You could have wrung buckets of it out of the mosquito bar, the gauzelike netting that had covered the sleepers. From one wall, a picture of the Virgin Mary gazed serenely down on the pain and blood. Frank's parents had beaten their son to the scene. Even old Mr. Jordano, achy with rheumatism, had moved faster than he'd done for months and had followed his wife from their house to their neighbors', curious and frightened at the same time. Mrs. Jordano took in the situation at a glance: Rosie lay still. Charlie was barely conscious, awash in his own blood, half on and half off the bed, kneeling with his upper body slumped across it. Going over to him, Mrs. Jordano asked, "Mr. Charlie, Mr. Charlie, what can I do for you?" A voice from the back of the crowd advised, "Don't do anything until the doctor comes." Charlie Cortimiglia could barely shake his head. Feeling helpless but desperate to do something, Mrs. Jordano went to get a bucket of water and a cloth to bathe his bloodied face. Frank took her place by the side of the bed, asking, "Charlie, for Christ's sake, who done this?" Charlie couldn't speak before he passed out. Frank cradled his neighbor in his arms as he kept him from falling over. As he did so, he glanced at the body of the toddler lying beside him. The little girl had been playing at his house only days before. Frank—big, husky, 275-pound Frank—began to cry. _Pictures that appeared in the_ Times-Picayune _after the attack on the Cortimiglias._ Charlie came to and moved his lips as if trying to speak. Frank leaned down to make out what he was saying. Charlie was only able to get out "Frank, I'm dying. Get my brother-in-law" before lapsing into unconsciousness again. Determined to carry out what seemed very likely to be Charlie's dying wish, Frank left him in the care of his parents and the other neighbors and went outside. As he left the house he ran into his sister Lena, who had finally found her nerve and come to see for herself what had happened. Pale and wiping away tears, Frank knew that she wasn't ready for the scene he'd just left. "Don't go in there," he snapped as he darted past. "If you see what's in there, it'll kill you." Running on, he hurriedly hitched up his horse and buggy, leapt into the seat, and snapped the reins. Under the circumstances, Frank reckoned that finding a doctor was a higher priority than locating a brother-in-law. Charlie Cortimiglia would forgive him if he briefly delayed his errand. The first doctor he called on was not at home. Having better luck with his second choice, Dr. G. W. Rossner, he begged him to hurry to the Cortimiglias: "People are dying down there!" After the doctor promised to head right over, Frank turned his buggy in the direction of Amesville, a small farming community about three miles up the river from Gretna. Setting his horse to a brisk trot along the dirt track running next to the Southern Pacific Railroad line, he thought unhappily of news he had to deliver: that the Cortimiglias had been cut up and robbed. For that, reasonably enough, was what he assumed had happened. Perhaps the thought flitted across his mind that they had been victims of the "fiend"—that's what the newspapers called him—from across the river. But no, he had never struck in Gretna, always in New Orleans itself. Frank flicked the reins to make the horse go faster. Dr. Rossner arrived at the Cortimiglias' to find a houseful of people milling about the bedroom. He took one look at the figures on the bed and realized he could do nothing for them. What they needed was Charity Hospital, and as soon as possible. Fortunately, Manny Fink had already realized that. Emmanuel Fink—"Manny" to his neighbors—engineer, machinist, businessman, and city councilman—was an energetic man used to taking charge. Living only half a block from the Cortimiglias, he had been one of the first to be roused by Hazel's cries. After seeing that Rosie and Charlie were still alive, he left the bedroom and grabbed Tony Winters, the first person he met on his way out of the house, and ordered, "Come back here; these people are all chopped up." He stationed Winters at the bedroom door with orders not to touch anything and not to let anyone else touch anything. Fink then telephoned everyone he could think of: the sheriff, the chief of police, a couple of doctors, and the Charity Hospital ambulance. Charity Hospital was not as helpful as he had expected. The big-city hospital, he was informed, did not extend their ambulance service to his side of the Mississippi. The unpaved roads of little Gretna were so bad that the hospital was afraid of the ambulance getting stuck on the far side of the river. _For God's sake_ , Fink insisted, _two people are dying here. Well_ , the hospital offered, _we could have the ambulance meet you at the foot of Jackson Avenue. You just have to get them to the ferry and across the river._ Fink wasn't left with much choice. He returned to the Cortimiglias' to find the police chief and deputy sheriff. He reported the hospital's response and then went home to get his own horse and wagon. Hitching up his horse as quickly as he could, Fink drove back to the Cortimiglia place. There, someone had found a thin mattress to use as a stretcher. Gingerly, Rosie was placed on the mattress and numerous hands gently lifted her into the wagon. But what about Charlie? Fink looked around, not seeing another potential stretcher. Then someone had the clever idea of using the bathroom door. Someone else grabbed the axe—covered in congealed blood—that had already been found underneath the house and knocked the door free of its hinges. As Charlie was carried out of the room, Fink glanced back to see two pieces of skull lying on the bed. Shaking off the sight, he snatched up the reins and sent his horse as fast as he dared down the dozen blocks to the ferry. Little Mary Cortimiglia's body was left lying on the bed. A neighbor put a blanket over it. Someone else called Fred Leitz's undertaking parlor. Dr. Henry Leidenheimer strode down the wide hallways of Charity Hospital toward the Accident Room. As the surgeon on duty, new cases were reported to him first. At thirty-nine, he was an experienced surgeon. And Charity, despite the name, was a good hospital. Established for the poor in the early eighteenth century, it was one of the oldest public hospitals in the country and a well-regarded teaching hospital for Tulane Medical School. The main building dated from 1832; the very solidity of the imposing three-storied brown structure, lateral wings flanking both ends of the central corridor, was reassuring. Overcrowded and chronically short of funds, Charity nevertheless managed to provide a fairly decent standard of care for the city's needy, both white and "colored," by the standards of the time. The Cortimiglias would need all the expertise Charity and Dr. Leidenheimer had to offer. In the treatment room, Leidenheimer examined the couple, now lying on clean white metal hospital beds instead of their makeshift stretchers. Axe wounds didn't particularly surprise him; the presence of axes and hatchets in most homes for chopping firewood made them obvious weapons of choice in domestic disputes; irate people in New Orleans periodically took a swing at each other. Only last year, a man had gotten himself shot when he took after a friend with a kitchen axe. But these were particularly bad cases. A quick glance told him that Charlie probably wouldn't live: two severe cuts had sliced through his head, fracturing his skull and cutting into the soft tissue and brain beneath; his traumatized brain had swollen, oozing through the fractured bone like mud through the slats of a chicken crate. Dr. Leidenheimer shook his head. All he could do for Charlie was clean his wounds with antiseptic, dress them, and hope for the best. By then Dr. Jerome Landry had come on duty and taken charge of Rosie. She had several gashes on her head and one on her left ear. Much more serious, a blow to the left side of the head had left her with a depressed fracture, pressing in on her brain. Dr. Landry didn't think Rosie was likely to survive, but he could perform a craniotomy, a procedure to relieve the pressure on her brain. Without much optimism, he ordered her wheeled into the surgical amphitheater. He didn't bother with anesthesia; Rosie's injuries made it unlikely that she would wake up during the surgery. Besides, he wanted to avoid the risk of oversedating and accidentally killing her. Rosie survived the operation, but afterward, still without much hope, Dr. Landry, as he later put it, "sent her back to the ward to die." Which she stubbornly refused to do. By Monday morning her relatives—parents, sister, brothers, brother-in-law, and niece—were pouring in from surrounding parishes and gathering at the hospital, waiting for any sign of improvement. Charity had no waiting room in which to confine them. They were a nuisance for the nurses, crowding about, begging to see Rosie and Charlie, and demanding to know who was responsible for the attack. Her niece Anna implored hospital authorities to let her see her aunt. _I can speak to her in Italian_ , she insisted. _She'll tell me who did this_. While Anna pleaded, Rosie's stepbrother John threatened: "I must see her! She will tell me who did this and then something will happen!" He made such a fuss that the police had to step in to persuade him to listen to the nurses and leave. Rosie's parents were the only ones allowed to see her. Mostly, she slept quietly, as her pulse and color gradually improved. From time to time, she mumbled in Italian or murmured her daughter's name; at times she seemed vaguely to recognize her parents as they sat by her bedside. Charlie was still supposed to be dying. That's what the doctors kept telling his in-laws. But, also not very cooperative with his physicians' dire predictions, he hung on. Neither parent was in any condition to be told Monday when Rosie's parents arranged for the funeral of their granddaughter. Children were buried in white, the color of innocence. Wealthy people could try to mitigate their grief with a grand display of a white coffin in a white hearse pulled by white horses in a white harness. The Cortimiglias could afford no such extravagance, even for their only child. On that dismal, rainy March day, relatives and neighbors clustered around a modest white casket as the little girl was entombed in one of the whitewashed vaults of the Hook and Ladder Cemetery. Almost certainly Frank Jordano and his parents were there. What must he have thought as he stared at the pathetic little casket? That little Mary had played in his parents' store countless times? That she had called his own father "Grandpa"? Now she was dead, an ugly, violent, incomprehensible death. Who would bash a toddler in the head with an axe? Everyone in the cemetery must have asked themselves that, over and over. And as they listened sobbing to the final words said over the body of the little girl, no one would have disputed the _New Orleans Daily States_ ' description of it as "one of the saddest funerals ever held in Jefferson Parish." Mary wouldn't be the Axeman's only Gretna victim. That Saturday-night attack would damage her father's business, ruin her parents' marriage, and endanger the lives of the friendly neighbors who had only tried to help. Only thirty-one years after Jack the Ripper terrorized London and presented the world with a new kind of killer, fifty years before the term _serial killer_ even existed, the police were ill-equipped to deal with serial murder. Jack the Ripper lives forever as the ghostly fiend of Whitechapel, and his five victims are among the most famous unsolved murders in history. But he is hardly the only one who got away with such crimes. Killers who target strangers and have no obvious motive are always difficult to catch. In the 1970s Ted Bundy killed over twenty young women before he was finally apprehended, despite over a dozen detectives in three states looking for him. The BTK killer of Wichita was caught after thirty years only because of DNA evidence and his own arrogance. The Zodiac killer of northern California evaded a forty-year manhunt, never to be captured or even identified with certainty. In the early twentieth century the difficulties of catching a serial killer—even identifying serial murders—were even greater. It was the dawn of a new age: police forces were becoming professionalized as they began wriggling free from the taint of corruption and patronage. While police procedures and investigative techniques were rapidly developing, the science of homicide investigation was still in its infancy. The art of profiling was unknown. Scientists had only learned how to distinguish human blood from animal blood in the last twenty years; detectives could now conclusively prove whether a pinkish stain was blood. Fingerprinting had only recently been introduced as a crime-fighting tool. Toxicologists in New York City's coroner's office were just beginning to systemically develop the discipline of forensic chemistry as a way of detecting poisoners. In Vienna, Freud was still unpacking the unconscious motives of the human mind. Detectives still operated mainly on legwork, gut feeling, common sense, their knowledge of their community, and the surprising willingness of suspects to confess. In New Orleans, where the killer struck first, some recognized that they faced no ordinary criminal. The city was blessed with two police superintendents who realized that they were stalked by a Jack the Ripper of their own, a different kind of murderer against whom traditional methods were . . . well, if not useless, at least limited. What they weren't blessed with was luck. One, an experienced homicide investigator, met his own tragic end. The other was a well-meaning bureaucrat defeated by office politics. Gretna was a small town, living in the shadow of its better-known neighbor across the river. Its murders were tragically mundane, arising from drunken brawls, jealous rivalries, and clumsy robberies. Crimes like these can usually be solved the old-fashioned way. With a serial killer, the police need to know what they are looking for. The authorities of Gretna, unwilling to face the reality that they couldn't solve this appalling murder, convinced themselves that Frank Jordano and his father—a teenage boy and an old man—attempted to hack their neighbors to death over a rental dispute. Through their ignorance (which they couldn't help), their willingness to bend the law in pursuit of an end (which they could), and perhaps their malice toward Italian immigrants, the Gretna authorities made the tragedy of Mary Cortimiglia's death worse. Unrecognized by them, the attacks of the serial killer who had come to terrorize New Orleans had actually begun on a hot summer's night some nine years earlier. # ≡ 2 ≡ # The Cleaver ## 3 AM, Saturday, August 13, 1910 HARRIET CRUTTI WOKE FROM a sound sleep to find the shadowy figure of a man standing over her with a meat cleaver. The apparition, holding up the mosquito netting with one hand, and waving the bloody cleaver at her with the other, came sharply into focus as he demanded her money: "Or I'll do to you what I just did to your husband!" Panicking, Mrs. Crutti looked down and saw her bloodied husband lying still across the foot of the bed. "You've murdered him!" she screamed. Terrified, she did the only sensible thing: she reached under her pillow for the box containing eight dollars (a significant sum in 1910) and handed it over. It wasn't enough to satisfy the man with the cleaver: "Is that all you got? I want all of it!" _Yes_ , she insisted. _Just take it_. Mrs. Crutti was too frightened to mention the more substantial amount of money under the mattress. Fortunately, the intruder believed her. He turned around and strode out of the bedroom, and through the Cruttis' combination grocery/bar/residence, snatching up their pet mockingbird in its cage as he went. Tossing aside the meat cleaver in the yard, he retrieved the shoes he'd taken off and climbed over the back fence, leisurely walked a block down Lesseps Street until he reached the corner, and sat down on a doorstep. There he flipped open the latch on the birdcage and freed the bird. Then he deliberately rolled a cigarette, leaned back against the stoop, and smoked it. Afterward, he pulled on his shoes, stood up, and sauntered down Dauphine Street. _Harriet Crutti._ At the same time, Mrs. Crutti, afraid that her husband was dead or dying, was desperately trying to shake him awake. Groaning, he tried to rise, only to fall semiconscious off the bed. One of the Cruttis' young sons added to the chaos by waking up and starting to cry, panicking his mother even more. Frantic, Mrs. Crutti left her husband on the floor and ran out into the street. She pounded on her neighbors' doors. _Please! Open up! My husband's hurt!_ No response. Some slept through the commotion. Others looked through their windows to see what was happening. But all the doors remained shut. For an agonizing fifteen minutes Mrs. Crutti lurched desperately along the dark and deserted street, running from house to house, pleading for help. She finally managed to rouse Officer Gus Albert, a policeman who lived nearby. Officer Albert, still in his nightclothes, grabbed his revolver and rushed after the assailant, now long gone. _Grocer August John Crutti, the Axeman's first victim._ The injured grocer was taken to Charity Hospital while the police arrived to investigate. They soon discovered that earlier that night the assailant had stolen the meat cleaver from a butcher's stall six blocks away. As the police reconstructed the crime, at about three o'clock on that Saturday morning, the shoeless intruder had removed a pane of glass from the Cruttis' kitchen door (presumably to reach in to the bolt) but had ended up using a railroad shoe pin (a thin, curved steel bar used to couple railroad cars and a common burglary tool for prying doors and windows open) to force the door. Their small business was typical of its time: a small grocery and bar attached to the family's living quarters. Moving noiselessly, the intruder passed through the kitchen, through the grocery store, and on into the bedroom. There he pulled back the mosquito netting that protected sleepers in subtropical New Orleans from the harassing insects, raised the stolen meat cleaver, and struck the sleeping grocer twice. But why? The forty-year-old son of an Italian immigrant, August Crutti and his twenty-nine-year-old wife Harriet had opened their store on the corner of Royal and Lesseps Streets only a month before. It was located in the Bywater District of New Orleans, a block from the Mississippi River, just over a mile and a half east of the French Quarter. Crutti had been in the ice business, but he had worked hard to save the money to buy his own grocery, a modest establishment in a modest neighborhood. August and Harriet shared the small house with their two young sons, Jake and August Jr., and eighteen-year-old Arthur, Crutti's stepson. What could they have done in just a month to provoke such an attack? Once the blood was cleaned off, Crutti's injuries turned out to be far less serious than they'd first appeared. He'd been cut on the head and chest, but neither wound was life-threatening. By Saturday afternoon he was sitting up in his bed at Charity Hospital, smoking a cigarette and patiently answering the police's endless questions about the evening's events. He remembered going to bed at midnight but then having an attack of heartburn. Thinking that a cool drink might make him feel better, he got up to get a glass of water and then walked about the house for a bit before going back to bed. The bed was crowded by the presence of his seven- and eight-year-old sons, so Crutti lay with his head toward the foot of the bed, where he fell asleep about an hour before the assault. He remembered nothing about the attack itself; as far as he knew he'd been fast asleep when the first blow fell. No one could come up with any reason for the attack. Neither Crutti could think of anyone who would harm them. August Crutti admitted to an enemy from his days in the ice business but scoffed at the idea that he would have been capable of such an assault. Besides, his milquetoast enemy didn't match the description that the police had put together based on the accounts of Mrs. Crutti and a neighbor who'd happened to look out of the window in time to catch a glimpse of the assailant. The police were looking for a man thirty-six or thirty-seven years old; about five foot six inches tall; broad shouldered and clean shaven; with dark hair, thick nose and lips, and a rough, husky voice. He'd worn dark trousers, a loose blue workingman's shirt, and a black derby hat. The man in charge of investigating this puzzling crime was Chief of Detectives Jim Reynolds. The forty-two-year-old policeman hailed from Algiers, right across the river. Joining the police force at age twenty-five as a supernumerary clerk, he'd rapidly been promoted to plainclothes officer, detective, and finally chief of detectives. Portly, with a fleshy face that became almost cherubic when he smiled—as he often did—Reynolds was easygoing and likable, a chain-smoking joker who liked good company and good stories. Popular with the men under his command as well as the reporters assigned to the police beat, he was intelligent and sensible, with a reputation for hard work, a detective who would work a case as long as was necessary. As a seasoned veteran of robbery, kidnapping, and murder investigations, he shouldn't have been stumped by a simple robbery and assault. _Chief of Detectives James Reynolds._ But Jim Reynolds was as mystified as anyone. He walked slowly through the house and grocery, surveying the crime scene. He assumed that the incident at the Cruttis' home and grocery was just a robbery, although an odd one to be sure. If he was after money, the robber must have checked the grocery first and, finding nothing, realized that all the cash was probably in the proprietor's bedroom. But why attack the sleeping grocer? Had Crutti stirred in his sleep, frightening him? Why did the assailant take the risk of talking to Mrs. Crutti? Why take the bird? And given that he had just committed a hanging offense—for that was what assault with a deadly weapon was—for what possible reason did he take his time leaving the crime scene? The most common supposition was that the attacker was plain mad. After crawling all over the little grocery examining the evidence, hearing the Cruttis' story, and interviewing the neighbors, most policemen were inclined to shrug their shoulders and dismiss the criminal as "drunk or crazy." Not much else made sense. August Crutti's own idea was that "some half-witted fellow" did it for the money. The description of the attacker, and the description of two strange men seen hanging about the grocery on the night of the attack, gave detectives enough evidence to question a "well-known police character." A couple such characters were brought to Mrs. Crutti for identification. _No_ , she said both times, _not the man I saw with the cleaver_. Finally, after about two weeks, detectives got lucky. Chief Reynolds had decided, quite reasonably, that a criminal who acted as oddly, even irrationally, as the Crutti assailant did, was probably either mentally ill or drug addled. And he'd instructed his men to be on the lookout for such a person. So when a known burglar who had spent time in a mental hospital came to the attention of detectives, they leapt on him. John Flannery was an addict—a cocaine and morphine "fiend" the newspapers called him. He also was a petty criminal with a history of burglary to support his drug habit, a previous arrest for assault, and features that roughly tallied with the description of the Crutti assailant. When he was caught breaking into a grocery two weeks after the attack on August Crutti, Mrs. Crutti was called down to police headquarters. There she unhesitatingly identified him as her husband's attacker. Police were also able to connect Flannery with a series of other burglaries in which a railroad shoe pin similar to the one left in the Cruttis' grocery had been used. The case against him seemed clinched. But it wasn't perfect. At age twenty-five, Flannery was a bit younger than the mid-to-late-thirties man described by witnesses. And he vehemently denied the crime, even after hours of interrogation by Reynolds and St. Clair Adams, the district attorney. But Harriet Crutti's identification was positive, Flannery fit the physical description of the attacker, and he was the kind of criminal the police had suspected all along; the case appeared likely to close in short order. On September 9 John Flannery was indicted for feloniously breaking and entering into the Crutti residence and assaulting August Crutti. But he never went to trial. Doubts about his fitness to stand trial led to a commission to assess Flannery's mental condition. The commission consisted of two doctors, E. M. Hummell and Joseph O'Hara, a neurologist who was also coroner and city physician of Orleans Parish. The doctors concluded that Flannery's mental state had been compromised by drugs and alcohol, that he was suffering from disorganized schizophrenia, and that he was "insane and irresponsible [and] . . . a permanent menace" to society. O'Hara and Hummell recommended that Flannery be placed in an insane asylum. He was not committed immediately, however, for he was still in Orleans Parish Prison, the city jail, almost a year and a half later, when Dr. O'Hara reported that his mental condition had improved considerably, and he now appeared perfectly sane. Under normal circumstances the correct procedure would have been to take Flannery to trial. But by that time District Attorney Adams had reason to doubt Flannery's guilt and declined to prosecute, arguing even if he were guilty of the Crutti attack, Flannery had been irresponsibly insane at the time of the offense. Yet the end of August 1910 seemed to bring the Crutti case to a gratifying conclusion. Reynolds was satisfied that he'd found the culprit. Crutti went home to his family after only one night in the hospital. Even the liberated pet bird made its way home and was spotted on the roof of the grocery. A month later the attacker struck again. A man crept up to the grocery and residence of Joseph and Conchetta Rissetto shortly after 1:45 AM on the morning of September 20, 1910. Tonti Street and London Avenue (now A. P. Tureaud Avenue) was a rundown, crime-ridden part of town, mostly poor and black. This sparsely populated neighborhood on the outskirts of New Orleans hardly seemed part of a proper city at all, littered as it was with pigpens and cowsheds and barely passable dirt roads. _Conchetta and Joseph Rissetto._ The Rissettos, both children of Italian immigrants, had done well in spite of their impoverished neighborhood. Their business among the local "Negro" population was so successful that five years earlier they had been able to build a new grocery and barroom and add a poolroom. The cottage they built alongside their business was bigger than those of many other small grocers; it had the extravagance of both a parlor and a dining room, as well as a bedroom and kitchen. At forty-two and thirty-six years old, Mr. and Mrs. Rissetto had been married for seventeen years. With no children, they had only each other for company and were a devoted and loving couple. The man quietly approaching the Rissettos' home and business early that Tuesday morning probably cared about none of this. He carried a stolen meat axe, similar to a butcher's cleaver but, at three pounds, a little heavier. From the cowshed in the back he stole up to the kitchen and climbed through an unlatched window. Inside the house, he walked past the open door leading into the grocery. Entering the bedroom, the intruder went over to the woman's side of the bed and with a knife sliced open the mosquito netting, exposing the sleeping couple. He raised his weapon. He brought it down purposively on the helpless woman. The first blow hit her in the face, breaking her right cheekbone. As she reflexively twisted away from her attacker, he struck her again, and again, cutting deep into the left side of her face and slashing her neck. He moved around to the other side of the bed and struck her husband twice in the face, one blow slicing cleanly though the cartilage of his nose. The assailant then dropped his meat axe into the tangled, bloody strands of the mosquito netting. He didn't stop to take anything but made his way out of the house, going through the poolroom behind the kitchen, opening the door leading out into the yard, and heading toward the fence in the front yard. Awake now, Joseph Rissetto felt something warm running down his face. He tried to get up, but, blinded by the blood in his eyes and stunned from two blows, all he could do was fall out of bed. Frantically crawling around on the floor on his hands and knees, feeling blindly around for matches, hurting badly and unable to see, he painfully managed to light a lamp. Still barely able to see, he groped his way to a dresser where he grabbed a revolver and staggered to a side porch to fire two shots into the air. Help came fast. Hearing the shots, Bartholomew Pratts, one of the Rissettos' black neighbors, ran over to investigate. He found Rissetto sitting on his bedroom floor, his bloodied face in his hands, barely able to speak. Pratts immediately raised the alarm, sending for an ambulance and the police. As word spread that something horrible had happened down at the Rissettos' place, neighbors and relatives made their way in the dark down to the grocery where they could do little but wander through the house in shock and confusion, sickened by what they found. The Rissettos' bedroom resembled, according to one newspaper, a "slaughtering pen." Blood soaked the bed and was smeared all over the floor. Traces of Mrs. Rissetto's hair, cut by the blows of the axe, lay bloody on the bedclothes. She still lay in her own blood, unrecognizable, in great pain, pleading for help. Her distraught husband begged someone to help her. There was, however, little anyone could do except attempt to comfort them and wait for the ambulance. But the Charity Hospital ambulance couldn't reach them. The cratered and rutted roads were impassible. Charity had only just replaced its horse-drawn ambulance with a motorized one, which wasn't up to the neighborhood's mud and potholes. Rescuers had to carry the victims by stretcher four blocks from their home to the waiting vehicle. Joseph and Conchetta were then taken to the hospital, he permanently blinded in one eye and disfigured for life, she with injuries so severe that surgeons didn't expect her to live. A mounted policeman, Harry Gregson, who'd heard the gunshots, was the first to gallop over the rough terrain to the Rissetto place. He was the beginning of an impressive police response. Chief Reynolds and District Attorney Adams were awoken from their beds, and city officials, detectives, and uniformed policemen descended on the grocery in the early morning hours, newspaper reporters not far behind. Chief Reynolds found his policemen scouring the house and neighborhood as best they could in the dead of night. Even by candlelight, they could trace the intruder's movements through the manure and mud tracked into the house and the doors left open as he departed. Joining his detectives in their search, Reynolds took careful note of the fact that nothing in the grocery appeared to be disturbed; the cash register had twenty-three dollars in it; the safe, which held several hundred dollars, had not been tampered with. Valuables in the bedroom were also untouched: Joseph Rissetto's gold watch and chain were still in the drawer where he'd left them; thirty-five dollars sat on the dresser. Reynolds knew very well that John Flannery was sitting in the parish prison; he also noted how similar this crime scene was to that of the Crutti grocery. Once he'd been thoroughly through the crime scene, Chief Reynolds went with several of his detectives and District Attorney Adams to the hospital. Joseph, in much better condition than his wife, had the most to say. But even Conchetta, as badly wounded as she was, did her best to whisper answers to the detectives' questions. Other than a romantic rival he had had for his wife years earlier, Rissetto couldn't think of any enemies. He knew of people who disliked him—mostly customers who'd been refused credit and threatened him from time to time—but he didn't believe any of them would actually try to kill him. Both the grocer and his wife assumed that the crime must have been committed by someone bent on robbery. But this bloody attack was so unlike the type of crime usually associated with the poor, black population of his neighborhood that Reynolds doubted the explanation was so simple. Once daylight came, Reynolds's men searched the Rissettos' home and grocery again. Assistant Chief of Detectives Dan Mouney combed over the yard outside the grocery, meticulously making note of all the footprints crisscrossing the property and trying to identify to whom they belonged. Investigators traced footprints leading into the pigsty and footprints leading from the cowshed. Two sets of tracks were spotted outside the kitchen window, leading to speculation that two people were involved. Impressions of bare feet led some investigators to argue that the assailant had been shoeless. Over the next several days, the police questioned suspect after suspect, without making any progress. Most were black; all were dismissed as possible assailants. Investigators remained thoroughly baffled. Meanwhile, Reynolds uncovered a curious fact. The meat axe used in the attack had been stolen from a butcher's stall in a local market several weeks previously. And a large butcher knife stolen from the stall at the same time had turned up a week or so ago at a burgled grocery less than two miles from the Rissetto grocery. This made the crime even more puzzling. The criminal appeared to be a thief who had robbed one grocery without hurting anyone, as well as an assailant who viciously attacked the Rissettos without stealing anything. The chief of detectives shook his head. None of this made any sense. As they had done after the Crutti incident, the newspapers debated a range of possible motives: robbery, revenge, drug-induced mania, even "domestic strife." To some, the previous assault on the Cruttis did not rule out a vendetta as a motive. Both couples were, as more than one paper deliberately noted, "of Italian descent." The implication that the families had run afoul of some Italian criminal gang was clear. Reynolds's own view is hard to discern in the pages of the newspapers. Three days after the attack, the _New Orleans Item_ quoted him as insisting, "I am certain that burglars did the work." The _Times-Democrat_ , on the other hand, was convinced Reynolds believed the attacker was no ordinary criminal, with no ordinary motive, but someone who simply delighted in the sight of blood. But rumors also circulated that the chief leaned toward the "revenge theory." Reynolds may well have found it prudent to keep his views to himself, leaving reporters to infer them as best they could. But the _Times-Democrat_ was likely nearer the mark. Both the Crutti and Rissetto crimes were too unlike the handiwork of any ordinary burglar. That much was clear to almost everyone. The "meat axe fiend" theory seems to have been the most popular: that the assailant was motivated by "a fiendish thirst for blood"—a desire to draw blood, perhaps to actually see the blood. Newspaper headlines spoke of "Jack the Axeman" and speculated about New Orleans's own Jack the Ripper, all of which contributed to what the _Daily News_ called the "dime store novel flavor" of the affair. The main benefactor of all this speculation was John Flannery. His trial was scheduled to begin at the end of the month, and it was becoming increasingly clear that, drug addict or not, he hadn't assaulted August Crutti. It was also increasingly a matter of concern that the attacker would strike again if he wasn't stopped. And no one knew how to stop him. At the end of the week, Chief Reynolds again assured the public that "the trail is very warm and we expect to capture this man shortly." But the investigation had petered out, although Reynolds was loath to admit it publicly. The episode did have a happy ending, of sorts. From the beginning the doctors at Charity Hospital had thought Joseph Rissetto, while badly scarred and blind in one eye, would live but that his wife probably would not. Yet Conchetta held on, day after day, despite constant pronouncements in the papers that she couldn't last much longer. If she did survive, the surgeons said, she would probably be permanently paralyzed on one side of her face. Her husband, injured and in the men's ward, waited anxiously for news about his wife's condition and asked frequently how she was and when he could see her. She lay in the women's ward pleading to see him. After a week, he was finally permitted to leave his own hospital bed and go to his wife's bedside. As badly scarred as both were, they were each delighted to see the other. Their affectionate but brief reunion over, he returned to their home, alone. How long he had to wait for her to join him is not known, but join him she undoubtedly did, for she lived to a good age, dying only in 1940. Joseph, however, died only two years later, in 1912. While the newspapers said that his death wasn't a direct result of the injuries he'd sustained, he was never the same after that September night, and it's hard not to suspect that his wounds contributed to his long decline. Perhaps Joseph Rissetto was another, unacknowledged fatality of the Axeman. According to the _Daily Picayune_ , Mrs. Crutti had described her husband's attacker as having "murder stamped on his countenance." She was right. For if "the Cleaver," as he would soon be known, hadn't yet actually managed to kill anyone, that would change in the summer of 1911. # ≡ 3 ≡ # Dagoes, Sugarcane, and Muffulettas THE AXEMAN'S VICTIMS—JOSEPH AND Conchetta Rissetto, August and Harriet Crutti, the Cortimiglias, Frank Jordano and his entire family—represented a particular niche in early twentieth-century New Orleans: the Italian grocer. Italians, especially refugees from the rocky soil of western Sicily, came to Louisiana as laborers, but many quickly became businessmen. How? And why would anyone want to kill them? The dusty little Sicilian town of Campofiorito didn't have much to offer Iorlando Guagliardo, and he knew it. Slight but handsome, with a full, dashing mustache, he was also illiterate and unskilled; all that lay ahead of him was a lifetime of poverty, confiscatory taxation, rapacious bandits, endless vendettas, and constant labor in someone else's fields. Still, at age twenty-one, Iorlando had another option. In 1873 he packed everything he could carry into a canvas bag, kissed his mother good-bye and hopped on a cart that jostled him twenty-five miles north to the port of Palermo. There he boarded a ship and sailed to America, joining the vanguard of what would become a flood of Sicilians into Louisiana. Louisiana plantations had always run on the sweat and blood of black labor. But after the Civil War many former slaves were anxious to get as far out of Dixie as they could, and they migrated in large numbers up north and out west. Many who did stay in Louisiana preferred day labor in New Orleans or Baton Rouge to hoeing weeds in a cane field, for them an echo of times best forgotten. For their part, white planters heartily disliked dealing with freedmen and looked around for another, more amenable source of cheap labor for Louisiana's rapidly growing sugarcane industry. It was this need that led to the great influx of Sicilians into the state and the city in the later nineteenth century and explained the journey of many young men like Iorlando Guagliardo. The Sicilians turned out to be just what the sugar planters of Louisiana needed, for they were industrious, dependable laborers—a "hard-working, money-saving race, and content with . . . few of the comforts of life." The trickle of Sicilians in the 1860s and 1870s became a torrent by the 1880s and 1890s. They dominated Italian immigration into Louisiana. For the city's inhabitants, the new arrivals presented a strange and exotic sight—swarthy men in their fur caps, close-fitting black velveteen jackets and trousers, booted to the knee, and sporting earrings; stout women with their heads tied up in brightly colored kerchiefs, earrings dangling to their shoulders. Newspaper reporters covering the arrivals smiled condescendingly at the strange sight of bearded men rushing into each other's arms and kissing each other. The men were "like school girls," reporters wrote of the un-American emotion with which relatives and friends welcomed the newcomers: "The meeting between the immigrants [and] their friends, who had preceded them to this country, was in accordance with the manners and customs of the passionate, warm-hearted people of the Southern climes. . . . All talked and chattered as only Italians and Spaniards can." Sicilian traditions only added to their alien air, traditions that had developed to make bleak lives bearable. Saint Joseph was a favorite in some Sicilian towns. He was the patron saint of workers. Tradition held that once during a famine farmers became so desperate that they ate the fava beans they'd been feeding their cattle. With the help of Saint Joseph, they survived. In gratitude, every year on his feast day the devout created an altar loaded with breads and pastries that they shared with the poor. By the time immigration was shut down in the 1920s, over 100,000—and perhaps as many as 290,000—Italians, over 80 percent of them Sicilians, had passed through New Orleans into Louisiana. Most of these arrivals came with families, but many were single men like Iorlando. Most of these were temporary "birds of passage" making the journey to work and return home with their savings. Others, like Iorlando, stayed. By 1900, at least 8,000 Italians—and possibly as many as 15,000 to 25,000 if children of immigrants are counted—made New Orleans their home, making it the largest such colony in the South. How Iorlando actually got to Louisiana isn't known. He might have sailed into New York Harbor and made his way from there. Or he may have boarded a ship loaded with lemons and figs and sailed from Palermo to New Orleans. He might have been recruited by padroni—labor agents—who were sent to Europe to sign up workers for Louisiana plantations. Possibly, having heard about jobs in Louisiana, he traveled to New Orleans on his own, hoping to find work when he arrived. If he arrived in New Orleans alone, speaking little English, wondering what he should do next, he would have been relieved to have a padrone welcome him in Italian and offer to arrange work. The New Orleans that greeted Iorlando was dirty and smelly, with few paved streets and no proper sewage; it was also a cultured and sophisticated city, more indulgently Latin than American, although the Americans west of Canal Street had come to dominate the city, to the chagrin of the Creoles of the Vieux Carré. The elegant white and orange stucco Creole townhouses with their pitched roofs and lacy wrought-iron balconies lined the narrow streets of the Vieux Carré and the Faubourg Marigny. In the American sector, along broad Saint Charles Avenue, Italianate and Greek Revival mansions on expansive grounds proclaimed the presence of American banking and merchant wealth. Fashionable cafés and fine French restaurants catered to a sophisticated crowd. The French Opera House hosted international theatrical stars. Visiting Europeans preferred New Orleans to the "foggy cities of the North," because of its "sunlight, its festal air, its all-pervading cheeriness and gaiety." The city dwarfed Iorlando's hometown of Campofiorito. And it was expanding. New arrivals were pouring in at a brisk pace and had done so throughout the nineteenth century. With a mere 8,000 souls in 1803, by 1810 the population had jumped to over 24,000. By 1840, 102,000 residents made New Orleans the third largest city in the nation. On the eve of the Civil War, the population stood at 174,491. When Iorlando arrived in 1873, the population was 191,418. By the end of the century the city would reach well over a quarter of a million people. Most of the new arrivals were, like Iorlando, European. Nothing is less original than observing that the United States is a nation of immigrants, but of few places is this truer than New Orleans, especially in the nineteenth and early twentieth centuries. Founded by the French, ceded to the Spanish, sold to the Americans, influenced by Creole, European, Caribbean, African, and South American elements, New Orleans was a genuinely multicultural stew of peoples, languages, and cultures. As he absorbed his new surroundings, Iorlando may well have had the reaction of the mid-nineteenth-century German visitor who marveled at the "Americans, Brazilians, West Indians, Spanish and French, Germans, creoles, quadroons, mulattoes, Chinese and Negroes . . . Mexicans, Spanish and Italians" who were to be found swirling together on the city streets. Despite the city's glamour, Iorlando wouldn't have remained in New Orleans. Like so many later Sicilians, he was bound for the "Sugar Bowl," the sugarcane parishes surrounding the city, and eventually he found himself forty miles up the river in the flat, green, mosquito-infested fields of Saint James Parish. Here, in all likelihood, he went to work in the brutal cane fields. Life in the cane fields was a constant, hard slog. Sugarcane production was labor intensive, much more so than growing cotton. Iorlando would have worked long hours during the planting season in the spring and hoeing back weeds under the punishing Louisiana sun in the summer. The _zuccherata_ —the grinding season—arrived in October when the gold-green stalks had deepened into a mature purple and the cane was ready for harvest. For three months, Iorlando and others like him worked to exhaustion, cutting and loading cane seven days a week as long as there was light to see, and probably putting in an evening shift in the sugarhouse, turning the stalk cane into sugar and molasses. For Iorlando, this arduous life was better than anything he could have hoped for in Sicily. Swinging a cane knife, chopping weeds—all this gave him what most Sicilian immigrants wanted: the chance to save money. Working as an ordinary laborer, he might have earned seventy-five cents a day; as an experienced cane cutter, perhaps as much as $1.25 or $1.50 a day. As little as it sounds, such wages were twice what he could have made in Sicily, and Sicilians often carefully hoarded their wages—some managed to save fully half of theirs. Despite the hardships, workers flocked to Louisiana during the zuccherata, when planters had to hire large numbers of temporary workers. In these months, the presence of Italians may have swollen by tens of thousands. For a time, Iorlando probably joined the circuit of migrant agricultural workers, harvesting sugarcane in southeastern Louisiana during the zuccherata from late fall to early winter, laboring in the cotton fields of central and northern Louisiana in the boiling summer months, and scattering to larger cities—New Orleans, Chicago, New York—to find work in the months in between. Much of the temporary help arrived in Louisiana for the harvest and went home to Italy with their cash afterward. Not Iorlando. He stayed in Saint James Parish. Like so many destitute, illiterate Sicilians, Iorlando wanted security and respect. He wanted to be his own boss, more possible here than at home. So, like many other Sicilian laborers, Iorlando, hacking at cane and weeds under the murderous Louisiana sun, made up his mind to become a businessman. By working hard and saving his money, Iorlando could squirrel away enough to leave the cane fields and go into business for himself. This was the one problem with Italian workers as far as their employers were concerned. Planters grumbled that they couldn't keep Italians in the field because in only a few years they would save a little money. Then, as one sugar planter put it, they "are ready to start a fruit shop or grocery store at some cross-roads town. Those who do not establish themselves thus strap packs and peddle blue jeans, overalls and red handkerchiefs to the Negroes." This was no exaggeration; by 1900, Italian-owned businesses had sprung up all over Louisiana. The commercial success of Sicilian immigrants didn't protect them from the racial prejudices of the American South. Italians never entirely replaced black labor in the Louisiana cane and cotton fields but worked alongside blacks and the occasional Spaniard or Filipino. While Italians found nothing shameful about working in the fields with black workers, for native whites their willingness to do so made them no better than "Negroes," Chinese, or other "nonwhite" groups. Because of the work they did, the small, swarthy Sicilians were often considered not white at all, nothing but "black dagoes." It wasn't lost on an observer that even "Negroes made unabashed distinction between Dagoes and white folks and treated these alien fellow tenants with a sometimes contemptuous, sometimes friendly, first-name familiarity." The notion that "dagoes" were no better than "Negroes" helps account for growing prejudice against Italian immigrants in the 1870s and 1880s. Iorlando arrived in the earliest stage of southern Italian and Sicilian immigration that began in the 1870s. The handful of Italian immigrants before 1870 hailing from northern Italy weren't primarily laborers: they included skilled craftsmen, artists, musicians, importers, cotton brokers, accountants, and physicians, as well as merchants, shopkeepers, and fruit dealers. But most of those who arrived after 1870 were peasants from southern Italy, especially Sicily. As numbers of the uneducated and unskilled newcomers mounted in the 1880s, prejudice grew, and Sicilian immigrants in Louisiana faced suspicion and the occasional lynch mob. In 1924, a Tulane University thesis expressed a common view of the "filthy paupers" from Sicily, decrying their "viciousness, ignorance, debauchery, and crime." Such attitudes didn't prevent men like Iorlando Guagliardo from climbing the ladder from day laborer to merchant. We don't know exactly what route he took, but in 1892 Iorlando paid $200 cash for a small store in Convent, the seat of Saint James Parish. In 1883 he'd married another immigrant, Rosalie Lillian "Lillie" Billa. Over the next fifteen years, they produced a succession of children: two girls, two boys, and two babies who died at birth, a sadly common occurrence at the time. Then Iorlando left Saint James Parish, moving his family for a few years to Saint John the Baptist Parish, before settling in Gretna, the growing settlement across the river from New Orleans, about 1910. Perhaps Iorlando left Convent for the New Orleans area because he wanted a larger city with more opportunities for his children. With its relaxed cosmopolitanism and Latin Catholic heritage, the city absorbed new arrivals relatively easily. The Italian community in New Orleans was less segregated than in any other city in the United States. In New Orleans, ethnic ghettos simply didn't exist, as Italians, Germans, Irish, Chinese, Jews, Greek, French, Russians, African Americans, and nonimmigrant whites frequently shared the same streets and neighborhoods. Still, ethnic clusters could be found. African Americans dominated the area near the marshy backswamp known as back-of-town. A small Chinatown developed around the 1100 block of Tulane Avenue. A strip by the river between Magazine and Tchoupitoulas Streets was known as the Irish Channel after its inhabitants. The Faubourg Marigny was home to "Little Saxony." The oldest section of the city, the French Quarter, or Vieux Carré, had become the Italian neighborhood. By the early twentieth century, so many Sicilians congregated in the lower French Quarter near the river that the area from Jackson Square to Esplanade Avenue, between Decatur and Chartres, was known as "Little Palermo." Despite sporting an elegant European atmosphere with narrow streets, vaulted passageways, and iron lacework balconies overlooking tropical courtyards, the Quarter didn't have the cache it had had in the French and Spanish periods, or that it would regain by the end of the twentieth century. As the Creole gentry abandoned it in the late nineteenth century to move uptown to join their American rivals along Saint Charles Avenue and toward Lake Pontchartrain along Esplanade Avenue, the oldest section of the city became a squalid Sicilian slum. Multiple families of the city's poorest immigrants were crammed into decrepit Creole townhouses. Whole families, including four or five children, shared a single room and a lone bed. Drying laundry hung from balconies, courtyards were full of rotting garbage; ventilation was bad; sanitation was worse. The disapproving author of the Tulane thesis probably wasn't exaggerating much when she described the Italian sector's pervasive "filth and . . . intolerable stench . . . [and] deafening clatter and chatter." Many of the inhabitants of Little Palermo, like their countrymen on plantations, were determined to start their own businesses as soon as they could. Usually these were small enterprises needing a minimum of investment. Tiny shoe shops were everywhere; eventually, 75 percent of shoemakers in New Orleans were Italian. Barbershops, too, were on every block. They were particularly popular as start-ups because it was easy enough to turn one's own parlor into a barbershop: all one needed was a straight razor, a mirror, a chair, and a steady hand. But the real Italian niche was food. One of the most common upward trajectories for an ambitious Sicilian was that from plantation worker to truck farmer and peddler to grocer, probably the route Iorlando took. Many in New Orleans followed the same path. Much of the area on the outskirts of New Orleans was still quite rural. Dairy cows, chickens, hogs, and horses could be found in the sparsely populated Carrollton neighborhood west of the French Quarter, Amesville across the river, and Marigny, just outside the Quarter across Esplanade. From these locations truck farmers set out to furnish the city with its fruits and vegetables, or its milk and eggs, loading their produce onto horse-drawn wagons. They drove into town, selling to markets and groceries or hawking their goods on the street, the "humble, loud-voiced vendor, driving his cart . . . along the streets and calling out the different varieties [of produce] in English impossible to understand." Most Italian enterprises remained fairly modest. By the early twentieth century, corner grocery stores were expanding all over New Orleans, multiplying until they were ubiquitous; by 1910 when Iorlando and his family moved to Gretna there was one on practically every corner. This is one reason Italians were scattered all over the city. In the early decades of the twentieth century, no one had refrigerators. Iceboxes were stocked with big blocks of ice bought from the ice wagon. Most houses were modest, with small kitchens and very limited storage space. Housewives often shopped twice a day: once for the noon meal and again in the afternoon for supper. A grocery down the block, even a small one, was often more convenient than a central market. Corner groceries flourished, and Italian grocers flourished with them. In 1880 only 7 percent of grocers were Italians. By 1920 Italians owned half of all groceries in New Orleans. Iorlando shunned cramped Little Palermo and settled his family across the river in Gretna, a growing suburb of New Orleans. There, at the corner of Second and Jefferson Streets, he ran one of the little neighborhood groceries that sold all the basics: canned goods, fruits and vegetables, cheese and bread, and multiple varieties of pasta. When a customer entered the unpretentious wooden building, she was greeted with the briny smell of olives overlaying the sharp, pungent smell of cheese; Italian sausage added a peppery flavor to the air. Canvas bags of fava beans sprawled on the concrete floor. Cans of tomatoes and tins of sardines were stacked on the shelves that lined the walls, stretching up almost to the ceiling. During the summer, watermelons were piled neatly outside the store. The shopper was usually greeted by Iorlando or his wife; Iorlando was an old man by this time, gray and slightly stooped, but still slim with a full mustache, speaking with a thick Italian accent—"Good-a morning. Nice-a day." No doubt, many of his lady customers thought him charming. For modern tourists, the legacy of the New Orleans Italian groceries is the muffuletta, a daunting combination of salami, capicola, ham, provolone, emmentaler cheese, and olive salad on bread. Its invention is most often attributed to Salvatore Lupo, owner of Central Grocery, which still stands on Decatur Street, across from the French Market. According to tradition, the sandwich originated in the early 1900s with the lunches of the Sicilian truck farmers who provided the French Market with its produce. Day after day Lupo watched the Sicilians sit on crates, balancing their trays of bread, cheese, ham, salami, and olive salad on their knees as they ate. Finally, Lupo took all the ingredients, put them together in round Sicilian sesame bread, and New Orleans's most famous sandwich was born. Today it's not longshoremen from the sugar wharves or peddlers from the French Market who get their lunches from Central Grocery, but tourists waiting for a muffuletta and a Dixie Beer who endure long lines that spill out into Decatur Street. The people who gave New Orleans the muffuletta—hardworking, self-reliant, frugal, and ambitious—these were the Axeman's targets. Despite the odds against them, Italian immigrants managed to carve out a slice of the American Dream for themselves. But for some, their success created resentment. And someone wanted to take it away. Iorlando Guagliardo and his family would pay a high price for the Axeman's crimes. Iorlando and Lillie prospered in Gretna as their business flourished. They were successful enough to put a little money away for the future. At some point the family had taken the name Jordano— _Guagliardo_ was too easy for Americans to trip over. Iorlando Guagliardo became Iorlando Jordano; his eldest son, Frank Jordano. With the murder of Mary Cortimiglia, their American dream appeared to be over. Iorlando and Frank would be accused of murdering the little girl. # ≡ 4 ≡ # The Davi Murder ## Wednesday, June 28, 1911 JOSEPH DAVI TOOK THIRTY hours to die. The doctors at Charity were astonished at the amount of fight still in the young grocer. His brains were seeping out of his head, having been bashed out of his skull. He should have died more or less on the spot. Perhaps being a newly married husband and expectant father made him reluctant to give it all up. And resist his fate he did. For over a day the doctors at Charity could do nothing but admire his struggle against death. But in the end the trim, handsome twenty-six-year-old was no match for his injuries. He slipped away without regaining consciousness. No one told Mary Davi that she was now a sixteen-year-old widow. In those days, the New Orleans coroner didn't fuss with invasive autopsies—no opening of the chest cavity, measuring of organs, or toxicology reports. Assistant Coroner Charles Groetsch's job was limited to getting a good look at the body after all the blood had been scraped off and deciding what had killed him. In Davi's case, this wasn't hard. He had sustained repeated blows to the front of his head that smashed his skull and pulped his brain. A clerk in the Coroner's Office called the police to notify them that they were now in the midst of a murder investigation. _Joseph Davi, the Axeman's first fatality._ In Baltimore, New Orleans inspector of police James Reynolds tore open the envelope the messenger handed him and read it without much surprise. His fears had been realized. The Cleaver was back. When Jim Reynolds had been appointed inspector of police only four months earlier, he had a force of fewer than 250 to police a city of 339,000. Fortunately, the serious crime rate in New Orleans was relatively low. Robberies and burglaries were common; murders, mercifully, were not. New Orleanians did have a murder rate considerably higher than most other southern cities and comparable northern cities such as Chicago. Nevertheless, New Orleans wasn't a particularly dangerous place for ordinary citizens as long as they stayed out of barroom brawls and conducted domestic spats well out of reach of guns and knives. The murder rate in New Orleans in 1911 was much lower than it would be a hundred years later. And it was considerably lower than it had been in the 1880s when murder was so common it couldn't even be used to sell newspapers. New Orleans murderers at the turn of the century were not an exceptionally clever lot. In 1910, the city had fifty-two murders, and the police immediately identified the killer in almost all of them. Most homicides were easy to solve for the simple reason that most murderers committed their crimes in front of witnesses. The majority of these killings were intensely personal—a domestic argument, a fight over a woman, a barroom brawl over an imagined insult. Death in the course of a robbery was unusual. Murders like that of Joseph Davi were exceptionally rare. But the people of New Orleans were only just beginning to suspect how unusual the killing of Joe Davi was. The police world was in the midst of a revolution of procedure and professionalism that would turn a badly paid, poorly trained, low-status force into an efficient law enforcement organization. In 1911 modern policing was still less than a hundred years old. New Orleans was hardly on the cutting edge of this trend, and in the early twentieth century only just starting to catch up. In the years after Reconstruction, New Orleans police officers were known for their corruption and incompetence. The financially strapped city had one of the worst-paid police departments in the country so, not surprisingly, it didn't attract the highest caliber of applicants. Poorly paid, haphazardly trained, often inexperienced, New Orleans policemen were almost as much a menace as the criminals they pursued. They routinely broke the law against carrying concealed weapons. Worse, when they drew their revolvers, they were lousy shots, usually missing the criminals at whom they aimed and frequently hitting innocent bystanders. Why did the residents tolerate such incompetence? Because of the nature of city politics. After Reconstruction, New Orleans was dominated by the Regular Democratic Organization, a political machine generally known as the Ring or the Choctaw Club (named for the social club on Saint Charles Avenue where its members met). This city machine operated by systematically getting out the vote on Election Day, rewarding its members with city jobs and contracts, and engaging in a bit of ballot-box stuffing when necessary. By the early twentieth century the Ring was led by Martin Behrman, one of the great city bosses in an era of city bosses. Pudgy, mustached Behrman looked like the German grocer he had been. The son of Jewish immigrants, Behrman began as a ward worker in his neighborhood of Algiers and quickly climbed the hierarchy of the Ring to become mayor in 1904. One of the most gifted politicians of his generation, he was to control the machine—and the city—for the next twenty years. The New Orleans Police Department, like most United States police departments at the time, ran on patronage. After 1877, New Orleans mayors had the power to appoint police officers, and they used that power to reward supporters and punish opponents. Policemen were therefore not hired or promoted on the basis of ability or job performance, but because of political and familial ties. For politicians, the advantage was both the ability to give jobs to supporters and a loyal police force with which to harass political opponents. The disadvantages for the people of New Orleans are obvious. The police department that Jim Reynolds joined was loaded with bribery, kickbacks, abuse of authority, and a general lack of discipline. Patronage wasn't the only culprit responsible for these faults. The conditions under which policemen worked were also to blame. Successful police work—tracking stolen property, locating a suspect—often meant a thorough knowledge of the criminal world. Counting criminals among one's acquaintances often led to temptations that a poorly paid policeman would find hard to resist. They enforced laws selectively, arrested people solely in order to meet quotas, and took bribes from criminals. The New Orleans Police Department had developed an entirely deserved reputation for corruption. When James Reynolds became the police chief in February 1911, he was a popular choice, a career policeman rather than a politician, and was free from the taint of corruption. He could hardly have been chosen without being a Behrman ally, but Reynolds publicly promised to "keep the police force entirely out of politics." He did his best to enforce training and discipline, cracking down on lazy and delinquent policemen. But there were limits to what the police inspector could do. Reynolds could no more stamp out the petty graft of day-to-day life than he could banish prostitution or illegal gambling. He had to police the city with the force he had. ## Monday, June 26, 1911 It was a perfectly ordinary day for Joe and Mary Davi. They woke up early, had their grocery store at Arts and Galvez Streets open by 5:30 AM, and spent a long day waiting on customers. The Davis were still newlyweds, married only five months, and the routine was not yet old. Mary usually handled the business of the store while her husband ran the saloon. Italian immigrants who had come to the United States as children, they made an attractive couple. Mary was pretty in a plump, youthful way, with masses of thick, light brown hair, brightened by a hint of gold and radiating with the glow of her first pregnancy. Joe was a good-looking fellow with a handlebar mustache whom everyone thought a "steady, sober and good young man." _The Davi grocery and home, where Joseph Davi was killed._ Usually, Mary closed the shop with Joe and they went to bed together. But tonight she was unusually tired and retired early. Joe closed up by himself about 10 PM. After counting his receipts for the day, he took the cash with him into the bedroom. The bedroom door didn't fasten properly, so Joe propped up a makeshift alarm on the top of the door—several empty seltzer water bottles that would clatter to the floor if anyone entered in the night. He also kept a loaded revolver on the table next to the bed. If he faced any burglars, he planned to be prepared. He finally slipped under the mosquito netting and crawled into bed next to his sleeping wife at about 11 PM. One imagines that Joe leaned over, kissed Mary good night, and, tired from his long day, quickly fell into a deep sleep. _Mary Davi, pregnant and widowed at sixteen._ Some time later, something woke Mary. At first she wasn't sure what it was. Then a movement caught her eye. She looked up to see a stranger in the bedroom. By the dim light of the oil lamp burning in a corner of the room, she saw a man near the wardrobe. She shook her husband: "What is that man doing here?" Joe only moaned in response. Hearing a voice, the intruder turned to her: "Where is your money?" he wanted to know. Mary simply stared, too frightened to reply. Apparently infuriated, the intruder grabbed a heavy porcelain mug and hit her hard on the side of the head. She fell back on her pillow, unconscious. When she could look around the room again, the man had disappeared. Mary didn't yet realize the extent of her wounds, but she had cuts on her face and her right hand and arm, and she could feel the blood dripping down her face. Panicky questions flashed through her head. Who was he? Where was he? Was he robbing the grocery? Would he come back to kill her and Joe if she screamed for help? Perhaps if she remained perfectly still, he wouldn't return. Mary lay frozen, paralyzed with fear and shock, listening to her husband's ragged breathing, as the hours dragged by. ## 5:30 AM, Tuesday, June 27, 1911 Standing in front of the Davi grocery in the postdawn glow, Ernest Boyer, a young bartender, was puzzled. And he was getting impatient. He needed two loaves of bread, but the grocery was locked and ghostly silent. That was strange. He pounded on the door again, and again he got no answer. He went around to the side of the building, the side where the Davis lived, and rapped on the bedroom window, hoping to rouse someone. After a moment, Mary Davi appeared at the window and peered hesitantly at him though the blinds. He asked for some bread, but she told him the store wasn't open. "Call your husband," he insisted. "I can't," she replied. "He's asleep." Boyer knew immediately something was wrong. Mary Davi was acting very oddly. She seemed confused, even frightened. As he continued to study her through the blinds, he noticed blood on her face. _Why is your husband asleep? What's the matter?_ Dazed, almost unwillingly, the young woman finally admitted that she couldn't wake her husband because he was bleeding. Would Boyer go fetch Benjamin Gallin, a close friend of her husband's? When Gallin, Boyer, and a few other concerned neighbors entered the house a short time later, they discovered Joe lying on his blood-soaked bed with a badly fractured skull. His breath came in gasps, as his brain, swelling as a result of trauma, pressed against the respiratory centers in his brainstem. Mary was in shock, herself cut and bleeding. Disoriented and shaking, the young woman gave every sign of not having really absorbed what had happened in the night. The Fifth Precinct Station, only a mile and a half away, got a telephone call just after 6 AM Tuesday morning. A man and a woman had been injured at Arts and Galvez Streets, said the caller. Could they send someone right away? A sergeant with a detail of patrolmen was sent to investigate. He took one look at the dying man and traumatized girl and immediately alerted police headquarters. Soon afterward, after the couple had been taken to Charity Hospital, detectives were on the scene. Inspector Reynolds was out of town. He had gone up north to attend the annual conference of the International Police Chiefs Union. In his absence, George Long, the new chief of detectives, took charge of the investigation. By ten o'clock, over half a dozen investigators crowded into the grocery and its tiny living quarters to take an inventory of the crime scene. The Davi place was a small grocery store at Arts and Galvez Streets. Business was good, considering that it was in a sparsely populated area at the edge of the city. It was only a little over a mile from Joseph and Conchetta Rissetto's store, and not far from the swamp and woods that defined the city's boundary. A rough partition sectioned off part of the grocery for use as a saloon. The residence was behind the store and consisted of a dining room, kitchen, and bedroom. The building was fairly isolated. Overgrown, grassy lots lay behind and to the side of the grocery. The intruder had pried open a window with a railroad shoe pin and climbed into the saloon. Once he made his way into the store, he had raised the hinged section of the grocery counter to gain access to the door leading up two or three steps into the residence. He then had gone through the dining room to the bedroom door. The investigators found the seltzer water bottles deliberately placed next to the door. The homemade alarm hadn't worked. In the bedroom, it was clear that the intruder had hammered the grocer mercilessly, smashing his skull and drenching the bed and the bottom half of the mosquito netting in blood. Joe Davi had been attacked while he was asleep, with no chance to fight back; the revolver still lay on the side table, untouched and useless. He'd been hit with such force that the impact of the blows had collapsed the top of the moss-packed double mattress in at a fifteen-degree angle. Skull fragments and bits of brain littered the sheets. On Mary's side of the bed, detectives could see the bloody print left on the wall where she'd put out her hand to steady herself as she'd staggered to the window in response to Ernest Boyer's knock. The mug that had hit her in the face lay on the floor. After assaulting Mary, the assailant left by the door opening onto Arts Street. As far as the detectives could tell, nothing had been stolen. Despite his demand for money, the man hadn't tried to take the sixty-four dollars in cash hidden under the pillow—an obvious place for a burglar to look. Contents from the wardrobe and a trunk were scattered about, but Mary's jewelry was undisturbed. Small amounts of money were still in drawers. Nothing indicated that the attacker had robbed the place. The police found no weapon other than those belonging to the Davis at the scene. The nature of Joe Davi's wounds—and the similarity of the attack to the Crutti and Rissetto crimes—made everyone assume that it was a cleaver of some sort. Davi's wounds, the inquest would later show, were caused by "a blow with a sharp-edged though heavy blade almost in the center of the head, crushing through scalp and bone . . . just such an injury as would have resulted from a blow with a butcher's cleaver." Some never had any doubt about who had viciously attacked the young grocer. On Wednesday morning the _Daily Picayune_ screamed, FIENDISH CLEAVER ABROAD AGAIN. After inspecting the crime scene, District Attorney St. Clair Adams went to the hospital to interview the victims. There, the doctors told the district attorney that Joe Davi was beyond help and certainly beyond interviewing. Adams turned his attention to Mary, but he didn't get far. All the shaken young woman could get out coherently was that the attack took place around 1 or 1:30 AM and the attacker was a white man. A more in-depth interview would have to wait until she recovered from the shock. Investigators at the scene were interviewing friends, neighbors, family—anyone who happened to be there. Warren Doyle, an assistant district attorney, tried to question Peter Davi, Joe's older brother. When Peter, having just seen the bloody mess that had been his brother's bedroom, ignored Doyle and insisted on seeing his brother immediately, he was arrested and forcibly taken to the district attorney's office. The next day Chief Long made his own attempt to interview Mary. He found her recovering in the women's ward. She lay pale and still in the white hospital bed, her eyes still reflecting the pain of her wounds. Bandages covered her face and cheek where she'd been cut; under the bandages her bruised face had turned a deep bluish-purple. She had cuts and bruises on her right arm and hand as well, presumably from the same weapon that had injured her husband. Mary repeated the story she'd first given detectives, a little more coherently this time. She'd heard nothing until she woke to see the man ransacking the wardrobe. She remembered very little after having been hit. She was clear that he spoke English. "Are you sure about the man speaking English?" asked Long. "Positively he spoke English," she replied. Unaccented English, she insisted, so she knew he wasn't an Italian. Even though she had only glimpsed him by the light of the single taper burning in the bedroom, Mary was able to provide a description of the man. He was white, clean-shaven, about five foot eight or ten, and not especially strong looking. He wore a blue jumper—a workingman's shirt—and black pants, but no hat. He'd moved soundlessly across the floor, so she thought he must have been barefoot. The chief thanked Mary for her help and left. What he had not told her was that earlier that day her husband had died. No one else told her either. She was allowed to believe that he, too, was in the hospital, gravely wounded but expected to recover. Because of her youth and prettiness and pregnancy, discreetly referred to in the newspapers as her "delicate state," the girl inspired a great deal of sympathy among the press and the public. Her physicians advised against telling her about Joe's death just yet, arguing that the shock, on top of her injuries and the trauma of her experience, would be too much for her. It's likely they were concerned about the baby she was carrying. That night a wake was held in her parents' parlor, and the funeral service took place there the next day. The black-clad procession then accompanied the casket to the cemetery. As Joe Davi was laid to rest, his wife lay in her hospital bed wondering when they'd finally let her see him. Chief of Detectives George Long had almost no technology with which to catch a killer. The New Orleans Police Department mostly investigated crimes with shoe leather and gut instinct. Nevertheless, the science of investigation had progressed rapidly in the late nineteenth and early twentieth century. In the late 1870s, Alphonse Bertillon, a clerk with the Sûreté in Paris, developed a system for classifying criminals based on measurements of various body parts. Each person had a unique collection of measurements that could be collected and used for identification. In the 1880s _bertillonage_ , as it came to be called, spread over Europe and the United States. Bertillon's system arrived in New Orleans in 1897, and by 1911 Inspector Reynolds had two Bertillon operators. For each criminal the Bertillon expert painstakingly took eleven measurements and recorded the information on a card, along with a more general description of the arrestee—hair color, eye color, identifying scars—and two mug shots, both a profile and frontal view. Mostly, the New Orleans police used the Bertillon cards for criminals who originated elsewhere—that is, criminals who might be wanted in other jurisdictions. The police knew their homegrown felons by sight. The Bertillon system, as clever as it was, was soon overtaken by an even more accurate system of identification when Frances Galton published _Finger Prints_ in 1892. Police departments, however, were sometimes slow to be convinced of the superiority of fingerprinting over bertillonage. Not until 1911 was fingerprint evidence used for the first time in an American courtroom. Only in 1918 did fingerprints begin to be added to Bertillon cards in New Orleans. In the late nineteenth and early twentieth century, great strides were also being made in criminology. In 1893, Hans Gross, an Austrian judge, published _Criminal Investigation: A Practical Handbook for Magistrates, Police Officers, and Lawyers_ , the first textbook of scientific criminology. Gross's handbook was a thousand pages of advice on how to examine witnesses, behave at crime scenes, handle the press, deal with hair, blood, and fiber evidence, preserve footprints, make use of experts, and handle dozens of other tasks that could be brought to bear on solving crimes. There is little evidence that anyone in New Orleans read it. The Bertillon operator did take photographs of the crime scene for the New Orleans Police Department, and the city chemist could determine if rust-red stains were blood, but otherwise it was a decidedly low-tech organization. Many detectives probably preferred it that way. They didn't put their faith in science. They trusted intuition and a lifetime of dealing with criminals. Jim Reynolds himself said that to be a good detective a man didn't need to be Sherlock Holmes. "All he needs is to use his eyes, and have the widest circle of acquaintances he can get," Reynolds shrugged, "and exercise the horse sense God gave him." Horse sense could take investigators a long way. Detectives were expected to know the local criminals, where to find them, and who their associates were. When a crime was committed, the investigation usually started with "a well-known police character." Homicide investigation as a specialization didn't exist. Inspector Reynolds, Chief Long, and their detectives investigated whatever came their way—robberies, assaults, blackmail, and murder. The nature of homicides in New Orleans meant that most of them didn't require much investigation. The big question was often not "Who did it?" but "Was it self-defense?" as many barroom killers claimed. But even with a mysterious death in which the perpetrator wasn't immediately obvious, Long and his men proceeded much as they did with any other crime. In the case of murder, a detective would talk to people, figure out who benefited from the death, and proceed from there. The inspector's detectives knew that "motive is the clew [ _sic_ ] which leads to the solution of half the crimes committed." If there was no motive, that, too, could be a clue. One reason Reynolds hauled in the drug-addicted Flannery for the Crutti attack was that he couldn't imagine what sensible criminal would commit such an odd—and pointless—crime. When detectives found a likely suspect, they grilled him. Without the luxury of fingerprint or DNA evidence, and especially if there were no witnesses, detectives relied on sweating suspects until they confessed. Policemen found that the pangs of conscience were a marvelous thing. Most people feel a natural impulse to feel guilty when they've done wrong, and they have an equally natural impulse to relieve their guilt by telling someone. Investigators have always exploited this to clear their caseloads. This is why Assistant District Attorney Warren Doyle's reflexive response was to arrest Peter Davi immediately and take him in for interrogation. As a relative of the victim, Peter was bound to be interviewed anyway, and even the mere appearance of wanting to avoid an investigator's questions was enough to get him arrested. As natural as the urge to confess is, sometimes detectives gave nature an assist. If a few hours of questioning didn't work, detectives might resort to the "third degree." This innocuous-sounding term covered a range of interrogation techniques from intense psychological pressure to physical torture, limited only by the imagination of the interrogator. Documented instances of the third degree include sleep deprivation for days on end, withholding food and water, sticking the suspect in a sweatbox, hanging the suspect out of a third-floor window, forcing the accused to spend the night in a room with the dead body of the victim, and for the unimaginative, a thoroughly old-fashioned beating. Sometimes the mere threat—a detective taking off his jacket, rolling up his shirtsleeves, and flexing a rubber hose—would be enough to induce a confession. The use of the third degree was widespread in the early twentieth century. Some policemen denied it happened; others boasted about it. Illegal in Louisiana, it was used anyway. Some policemen made it quite plain that they found such intense interrogation a useful professional tool: "I've forced confessions," bragged one, "with fist, blackjack and hose—from men who would have continued to rob and kill if I had not made them talk." In the course of their careers as detectives both George Long and Jim Reynolds were accused of employing the third degree. A murder defendant (later acquitted) once testified that Chief Long had stuck a revolver in his gut and threatened to shoot him unless he confessed to the crime. Without a doubt, some in the New Orleans Police Department used these sorts of tactics, and it's very likely that Reynolds or Long themselves applied them in one form or another. The _Daily Picayune_ praised Reynolds as a cop who "knew the effect of moral suasion and the weight of a blow." This suggests that Reynolds could get rough—and that the _Daily Picayune_ approved. At the conclusion of one murder investigation, the _Times-Democrat_ thought it worth noting that District Attorney Adams obtained two confessions "through persuasion and without the slightest use of force." Even if their methods didn't spill over into physical or psychological mistreatment, they certainly did bring intense pressure to bear during interrogations. Chief Reynolds had once shaken an iron dumbbell in the face of the accused in an attempt to make him look at the murder weapon. If New Orleans detectives did slap a suspect around from time to time, they almost certainly did so with a clear conscience when convinced of his—or her—guilt. In his own case, Long probably felt justified in using rough methods to get a confession since the defendant had been caught with the victims' diamonds. Reynolds and Long didn't beat confessions out of any of the Cleaver suspects. They wanted the guilty party, not a warm body coaxed into confessing; they questioned suspects and released most of them. John Flannery never admitted to assaulting August Crutti, and after the Rissetto attack the authorities dropped charges against him. Frank Armistead, a black suspect in the Rissetto attack, was questioned and released. Peter Davi, dragged protesting into the patrol car, wasn't in custody long. When Reynolds and Long got hold of a questionable character but had no way to connect him to a particular crime, they used the "flim flam" technique. New Orleans police called it that too, writing "flim flam" on the suspect's Bertillon card. That is, they could charge a suspect with being a "dangerous and suspicious character" and get a judge to throw him in the parish prison for a few weeks. This was a routine practice with known troublemakers and a useful way of keeping a suspect on ice without having to worry about the niceties of evidence. Whatever their deficiencies in technology or respect for constitutional rights, Jim Reynolds and his detective force were determined to find the midnight intruder with a grudge against grocers. After Joe Davi died, the governor of Louisiana posted a $500 reward for information leading to the capture and conviction of the killer. In the meantime, Chief Long had over half his detective force working the case, including his "Italian Specialist," John Dantonio. Dantonio was himself the son of an Italian immigrant, a tailor who had come over from Palermo. He was the only Italian on the force. Joining the police department in 1896 and promoted to detective in 1902, Dantonio had a reputation as a smart and resourceful officer. He had been involved in the Crutti and Rissetto cases but was to play an especially large role in the Davi murder. _John Dantonio, the "Italian" detective._ The investigators worked out of police headquarters, housed in the First City Criminal Court Building, at Saratoga Street and Tulane Avenue, where the New Orleans Public Library is now located. The court building was a three-storied, redbrick, neo-Gothic affair, its turrets and Romanesque arches a strange contrast to the tropical flavor of most of the city's architecture. The high ceilings lent elegance to the building while having the practical effect of keeping it cool. The summer heat was eased when breezes off the river blew through the high rectangular windows, ventilating the offices. Pictures from the period present patrolmen in belted woolen tunics and rounded helmets, detectives in suit coats and tightly knotted ties. No one willingly dressed like that in a New Orleans summer. More likely, when Chief Long, Assistant Chief of Detectives Daniel Mouney, and Detective Dantonio gathered in the Detectives' Office to consult and send their investigators out with fresh instructions, their ties were loose, their white shirts sweat stained, and their useless coats flung over their chairs. By Wednesday—a day after the attacks were discovered—most of the investigators were working on the case from before dawn until late into the night. For days they did nothing else. While everyone else in New Orleans tried to stay out of the sweltering July heat, they tramped up and down dusty streets, canvassing neighborhoods, asking the same questions over and over. _Did anyone see anything suspicious that night? Were there any strange men in the neighborhood? Did Davi have any particularly troublesome customers? Did he have any enemies? Had he quarreled with anyone recently?_ They talked to neighbors, friends, relatives, housewives who shopped in the grocery, men who hung out in the saloon, and peddlers who delivered produce in one-horse carts. They picked up known police "characters" and shadowed possible suspects. They searched the dank, roach-infested shacks near the woods close to Davi's grocery, terrifying the destitute residents. Every lead, no matter how trivial, was pursued. Amid all the activity at police headquarters Wednesday evening, Joseph Rissetto entered the building and asked to speak to one of the detectives working the case. He badly wanted to help the investigation however he could. No one had to ask why. The vivid scar across his nose and his blind right eye explained all. But he could tell the detectives nothing that would lead to Joe Davi's killer. At first, what baffled the police was that no one seemed to have any reason to kill Davi. His family assured them that he had no enemies. Investigators didn't turn up any. Frustrated, Dantonio told a reporter: "There is no apparent motive here." The detectives kept digging. By the end of the week they thought the tedium of their repeated questions had paid off. They finally had a promising suspect. Joe Davi's friend Ben Gallin told police about Sam Pitzo, a Sicilian truck farmer who had come to see Joe in his saloon about a week before his murder. They got into quite an animated conversation, Gallin recalled. Gallin couldn't tell what they were saying since they spoke in Italian, but the conversation seemed to take an angry turn and ended when Joe took two dollars out of the cash drawer and shoved it at Pitzo to get rid of him. After the truck farmer left, Joe told Gallin that the man had hit him up for cash with a hard-luck story of a sick wife. Pitzo wanted more than Joe was willing to give him, and he didn't take being turned down graciously. Joe told Gallin that he was worried the man might be trouble. Detectives immediately brought Sam Pitzo in for questioning. At first, he didn't give them enough information to warrant holding him. But the officers didn't let up. They asked neighbors about the truck farmer, who turned out to have an unsavory reputation. His real name was Sam Parieno. He regularly went around the neighborhood spinning stories of bad luck and pleading for money. The police also found evidence that he had lied to them about how much time he had spent in Davi's neighborhood. They didn't have the evidence to move against him just yet, but Long ordered his men to keep an eye on the shady truck farmer. On Friday, four days after the attack on his brother, Peter Davi showed up at police headquarters and told the doorman he'd like to speak to a detective. Directed to Dan Mouney's desk, he sat down across from the detective and pulled out a year-old letter. About a year ago, he told Detective Mouney, he and his brother Joe had run a grocery just outside of the Vieux Carré. One day Joe had received a letter warning him that unless he paid $200, he'd be killed. The letter included detailed instructions about where to deliver the money in Carrollton. His brother had ignored the letter, but a week later another arrived. When Joe still didn't respond, he received a third communication. This was the letter Peter had found in a trunk after his brother's death and now handed over to Mouney. Mouney turned the cheap notepaper over in his hand. The unsigned letter was written in Italian and dated May 24, 1910. The letter threatened Joe, Peter told Mouney. The writer warned that since Joe had ignored previous communications, he could expect retaliation. _Why didn't Joe take the letters to the police?_ Mouney asked. Peter said that he and his brother talked about it but put it off and never got around to it. Nothing ever came of the threat so, eventually, they forgot about it. It was only after Joe's murder that Peter dug out the letter. And, Peter told the authorities, that wasn't all. The combination of the blackmail letters and Sam Pitzo's history had made Peter and Ben Gallin even more suspicious of the truck farmer. They had gone to Mary and asked her if she knew Pitzo. She said, yes, he had delivered figs to the store. Did he look like the man she had seen in her bedroom? Well, she admitted, he might resemble him. This was enough to convince both Peter Davi and Ben Gallin of Pitzo's guilt. Mouney went straight to the chief of detectives' office and told him Peter Davi's story. They decided now was the time to bring Pitzo in. They snatched him from the steps of his Carrollton shanty, also pulling in Pitzo's employee, Philip Daguanno, caught wearing a shirt with tiny bloodstains on it. Saturday afternoon Long, Mouney, and Dantonio interrogated the two men for hours. _How well did you know Joseph Davi? Did you ever try to blackmail him? Where were you Monday night?_ Neither suspect was very helpful. Pitzo made no effort to cooperate, often merely shrugging his shoulders in answer to a question. Daguanno's English wasn't very good, and he had to be questioned by Detective Dantonio. Both men denied knowing anything about who had battered in Joe Davi's skull. Long and his men were starting to get frustrated when a neighbor of Pitzo's showed up at the police station. Sam Constanza, a grocer in Carrollton, accused the truck farmer of attempting to extort money from him. Pitzo, he told detectives, had come to his grocery and demanded ten dollars, promising to "beat his brains in" if he didn't pay. Now the detectives had what they needed—a plausible motive for murder. They knew Davi refused to be blackmailed; they knew he had quarreled with Pitzo; now they knew Pitzo was a blackmailer. Perhaps he'd tried to victimize Davi, and when he refused to pay, Pitzo made good on his threat. The self-congratulation must have been palpable. The pieces fit. Late Saturday afternoon George Long, Dan Mouney, and John Dantonio marched both suspects down the whitewashed halls of Charity, past the starched nurses in white uniforms, down to the women's ward. They pulled white screens around Mary's bed to block the stares of other patients on the ward and planted the shabbily dressed workmen in front of her. Mary looked warily at the suspects. Pitzo was about forty years old, short, sturdily built, with black hair and a thick, bushy black mustache. Yes, she said, she'd seen him before; he delivered figs to the store. She had also seen him when he'd come asking Joe for money. Pitzo resembled the assailant because he had a similar build. _But_ , she added, _I am almost certain the man who hit me had no mustache_. Next Long pointed to Daguanno. Had she seen him before? He was short like Pitzo but thin and nervous, a ferret-faced man, clean-shaven with brown hair. She shook her head without hesitation: _No, I don't recognize him_. Disappointed, the detectives returned to the police station. Without eyewitness testimony, they couldn't make a case against Pitzo. Their other witnesses that day were equally disappointing. August and Harriet Crutti and Joseph and Conchetta Rissetto all came in that afternoon. None of them had ever seen Pitzo before. Harriet Crutti, the only one who'd gotten a good look at the attacker, shook her head. "The man who attacked my husband was about the same height, but he was a little broader," she said. Sunday night, morale in the Detectives' Office must have been low when, to everyone's surprise, Jim Reynolds walked in unannounced. He hadn't been expected back for almost a week, but when he heard about Davi's murder, he hastily caught a train from Baltimore. Now the inspector sequestered himself with Chief Long to be brought up to speed on the investigation. Before he left for the night, Reynolds walked across the hall and into the pressroom. Reporters would have been there well into the night, anxious to get the latest word on the investigation, and they weren't willing to leave as long as a detective's light was on. Most of the writers on the crime beat spent so much time at police headquarters that the inspector knew them almost as well as he knew his own officers. Some crime reporters, young men like Andy Ojeda of the _Daily States_ , never seemed to go home. They spent their days—and frequently their nights—hunched over their typewriters in the Press Office, tapping out stories of theft, blackmail, accident, and murder. When the telegraph operator or Reynolds's secretary posted an emergency bulletin, they'd heave themselves into their jackets, tighten their ties, grab their hats, and be out the door after the story. The press and the police had a clubby relationship. Reynolds tried to make reporters' jobs easier, personally updating them on developments in major stories, giving them whatever information he felt he could, even occasionally offering them a lift in the patrol car on the way to a crime scene. The reporters identified with the police, some even going so far, as one admitted, to think of themselves as "closely allied to the [police] department, and likewise working for its advancement in [our] own way." Now Reynolds needed to use the press to reassure the people of New Orleans: "I came back early because I am deeply concerned about the Davi affair. This murderer must be captured. Neither men nor money will be spared in bringing the fiend to justice. Not only will public funds be available, but I mean to go into my private purse if necessary. He must be captured and we are going to do it." Whether or not Reynolds actually felt such confidence, with his arrival the investigation took off in new directions. The first thing he did Monday morning was to interview Pitzo and Daguanno. Reynolds came away convinced neither had anything to do with Joe Davi's murder, but he and St. Clair Adams decided to hold both men for a few more days. Who knew what evidence might turn up in the meantime? Monday afternoon Reynolds went to see the Davi residence for himself. He arrived to find a gang of prisoners from the parish prison cutting the tall grass growing in the vacant lots behind and next to the grocery. It was the latest effort to find the murder weapon. Dripping with sweat in the humidity, the prisoners were searching the lots on the chance that the murderer might have tossed his weapon there as he fled. So far, they'd found nothing but snakes and weeds. Reynolds spent several hours going over the crime scene with Long and the others. This would have been a good opportunity for the detectives to outline their respective theories of the crime; with more than one theory floating around, there must have been some lively discussion as first one and then another tried to persuade the inspector of the viability of his view. John Norris, the Bertillon operator, was also there. He had presumably been summoned to the grocery as soon as the injured couple had been discovered. But the crime scene hadn't been searched thoroughly, as Assistant District Attorney Doyle had just discovered what he thought was an additional print that needed to be photographed. The handprint of Mary Davi had been noted immediately, but Doyle had spotted what appeared to be another bloody print near the window of the bedroom. This print, investigators surmised, was that of the murderer, since it was too high to have been made by Mary. Norris disappointed the detectives by telling them that the "print" was nothing more than a smear and would be useless in identifying any suspect. But the bloody smudge should have been noted earlier. This wasn't the only evidence that had been overlooked. Ben Gallin had found a bloody print on an outside door that the police had missed. This, and the search for the weapon on adjoining lots almost a week after the attack, suggests that New Orleans detectives relied too much on their gut and not enough on evidence catalogued at the scene. And, of course, no one had even thought about preserving the crime scene. The police at least were now keeping away the morbidly curious wandering by in hopes of getting a glimpse inside the now-famous grocery. Monday night Inspector Reynolds was up until midnight discussing the case with his detectives. Everyone agreed that the Cruttis, Rissettos, and Davis had been attacked by the same person. The attacks were too similar to think otherwise. But there was disagreement about motive. Despite the lack of actual theft, some detectives believed that all of the attacks were burglaries gone wrong. Chief Long was the leader of this faction. He argued that nothing was taken from the Rissetto and Davi groceries because the thief was frightened away. He could point out that the intruder who attacked their husbands took eight dollars from Mrs. Crutti and demanded money from Mary Davi. _Look_ , he could say, _this fellow is nothing more than a burglar who got interrupted_. None of this convinced John Dantonio, who led the second camp. He insisted that they weren't dealing with an ordinary burglar. No attempt had been made to rob either the Rissetto or Davi groceries. The "thief" was interested in something very different. Dantonio was outspoken, telling a reporter, "I do not believe any of these jobs were the work of a burglar. I believe the guilty man took the $8 Mrs. Crutti gave him to throw the police off the scent. A fiend committed these crimes." During the week after Reynolds's return, perhaps at his urging, Dantonio took several other detectives to consult with a medical expert to learn as much about sadism as they could. Perhaps if they learned what motivated the fiend, they could catch him. Reynolds was careful to remain open-minded and publicly noncommittal. He kept his men working all aspects of the case. But he, too, seems to have inclined toward the sadistic "fiend" theory. As first one and then another line of investigation played out, nothing else made sense. Without any evidence to connect Pitzo to the Davi murder, Inspector Reynolds and DA Adams decided to charge him with what they had. Pitzo faced a count of attempted blackmail, and Daguanno was charged with being a "dangerous and suspicious character." Pitzo found someone to sign for his bond; Daguanno could not. Pitzo walked out, leaving Daguanno, arrested only because he worked for the other man, in the parish prison. In the excitement of the investigation, two people had almost been forgotten, two other innocent people whose lives had been changed by the Cleaver. Mary Davi left the hospital on Wednesday, July 5, just over a week after having been admitted. Finally, she learned that her Joe was dead, and the blow was worse than any cuts inflicted on her by the Cleaver. She had almost collapsed and was still in a very emotional state when her parents came to take her home. In the months ahead the young widow would dream of her dead husband, dream that he visited her, but the most constant reminder of Joe was the son she gave birth to early the next year, named Joseph P. Davi for his father. John Flannery still sat in the parish prison, waiting to be shipped off to East Louisiana Hospital for the Insane as a drug addict, even though the DA had dropped charges for the assault on August Crutti. He wrote to St. Clair Adams from his cell, insisting again that he was innocent of any violence. The murder of Joe Davi was more proof of this, he argued. Flannery pleaded for Adams to send him to the Louisiana Retreat, a private Catholic hospital in New Orleans, rather than to the state insane asylum in Jackson, Louisiana, over a hundred miles away. That was where the state sent the criminally insane. Perhaps he wanted to remain closer to his family in New Orleans. Perhaps he was terrified of the conditions in a state asylum. But he desperately wanted to be free from the stigma of criminal insanity. Whether the DA ever responded to his plea is unknown, but five years later Flannery was shot dead during an attack on a teenaged boy. Despite continued assurances from Reynolds that Joe Davi's killer would be caught, two weeks after the attack the public began to realize that the police weren't going to find the murderer. The case was rapidly getting cold. The Italian community was particularly unnerved by the police failure. This was the third attack on an Italian grocer in a year, and the police seemed powerless to put a stop to them. And on the morning of July 7, nine days after the Davi murder, an attempt had been made to break into another Italian saloon "in very much the same manner in which the store of Crutti, Davi and Rosetti [ _sic_ ] were broken into," reported the _Daily Picayune_. Fortunately, something had scared off the intruder. Nervous grocers securely fastened their doors and windows each night, praying to Saint Joseph that they wouldn't be the Cleaver's next victim. While Chief Long looked for a murderous burglar and Inspector Reynolds looked for a crazed maniac, there was a third possibility. All the attacks had been on _Italian_ grocers. What was the significance of this? Why not German or French or American grocers? When Joseph and Conchetta Rissetto were attacked, the _New Orleans Item_ suggested they were the victims of "vengeance"—i.e., the Italian vendetta. "Both families," the paper deliberately noted (referring to the Cruttis and Rissettos), "are of Italian descent." After Joe Davi's death, the _Daily Picayune_ opined that the "fact that all the victims were Italians of the small tradesmen class should point the direction in which the clews [ _sic_ ] are to be sought." Everyone in New Orleans knew what this meant. As the case was dissected over dining tables or in saloons, the same question was asked all over the city: Did the Mafia kill Joe Davi? # ≡ 5 ≡ # The Black Hand ## Around midnight, October 15, 1890 THE STREET WAS QUIET, except for the soft brush of light rain. New Orleans police superintendent David Hennessy and Captain William O'Connor of the Boylan Protective Police paused at the corner of Rampart and Girod. Hennessy turned to his companion. "It's not necessary for you to go any further. You go on and look after your business." Superintendent Hennessy usually worked late into the night, and since the threats had started, one of his friends usually insisted on accompanying him home. The chief (as most people called him) didn't give much thought to his own safety and didn't like making his friend go any farther out of his way on this dreary night. Besides, he was only a block and a half from home and the Boylan officer who'd been hired to watch his house. The two men said good night, and Hennessy turned down Girod Street toward the cottage he shared with his mother. O'Connor took off in the opposite direction. O'Connor had only gone two blocks when he heard the bang of a shotgun from the direction he'd just come. He spun around to see a bright flash and hear the loud crack of another shot. Almost immediately, three pistol shots rang out. Instinctively, O'Connor took off at a dead run in the direction of the gunfire. Firing continued as he ran. O'Connor raced past the spot where he'd left Hennessy. As he neared the corner of Girod and Basin Streets he heard someone call his name. Glancing down Basin Street, he could make out the chief's form, slumped on the steps of a house fronting the street. O'Connor rushed over and knelt beside his friend. Even in the dim light he could see Hennessy was covered with blood. "They've given it to me," Hennessy grimaced. "I gave them back the best I could." "Who gave it to you, Dave?" The dying Hennessy motioned his friend closer. O'Connor leaned toward him. The chief whispered one word: "Dagoes." The Mafia had struck again. Or so everyone said. Francisco Domingo was the Mafia's first victim in New Orleans. In 1855, the story goes, the Sicilian truck farmer received a note threatening to kill him if he didn't come up with $500. The note was signed with the ominous imprint of a black hand. Unimpressed, Domingo laughed at the threat as nothing to worry about. A few days later, the luckless immigrant's throat was slit and his body dumped on the bank of the Mississippi River. The next several years saw six more murders—each complete with a black hand note. In 1861 the New Orleans _Daily True Delta_ reported on bands of Sicilian thieves and counterfeiters, some of whom, the paper claimed, were probably the "black hand" murderers. The legend of the New Orleans Mafia was born. Unfortunately for Mafia aficionados, large parts of this legend are completely bogus. A Francisco Domingo was killed in New Orleans in 1855. But he was from Manila, his throat was not cut, and there was no great mystery about the identity of his killer. An argument at the dinner table escalated into a fistfight, and Guillermo Ballerio grabbed a knife off the table and plunged it into Domingo's chest. Later writers spun the mundane murder into the more dramatic tale. In the 1860s, New Orleans newspapers did warn against gangs of Sicilian robbers and counterfeiters active in the city. Sicily was a violent, poverty-stricken backwater in the nineteenth century, with the highest murder rate in Europe. There, the vendetta survived well into the twentieth century. Since New Orleans itself was a turbulent, vice-ridden, and often lawless town, it isn't exactly surprising that some immigrants were drawn to crime, and that some Sicilians were among them. The question of when the "Mafia" arrived in New Orleans is a messy one. Historians of crime can't seem to agree, and more than one criminal has been credited with establishing the organization in the city. John S. Kendall, a historian writing in the 1930s, gave the honor to Giuseppe Esposito, a notorious black-haired, black-bearded Sicilian outlaw. According to Kendall, in the 1860s and 1870s Esposito belonged to a criminal gang that terrorized western Sicily, robbing and kidnapping at will. Eventually he fled Sicily and made his way to New Orleans. But the law caught up with him. In 1881 Esposito was nabbed in Jackson Square by New Orleans police officers—one of whom was the good-looking, twenty-three-year-old David Hennessy. Esposito was extradited back to Sicily, and Dave Hennessy became famous. Kendall argued that while he was in New Orleans, Esposito led a handful of desperadoes who set up an extortion business, preying on local Italian merchants. This racket continued to flourish after Esposito himself was back in prison in Italy. And that, pronounced Kendall, "was in essence the Mafia." Except that it wasn't. Esposito as Mafia boss was a later invention. At the time, no one described him as anything other than a bandit or brigand. Such fantasy is typical of the myths surrounding the Mafia in New Orleans, most a mixture of exaggeration and outright fabrication. The citizens of New Orleans, however, didn't realize that. For many—perhaps most—people in New Orleans at the turn of the twentieth century, the Mafia was a deadly reality. The gunning down of David Hennessy that rainy night cemented for many New Orleanians—and much of America—a firm belief in the Mafia, a secret criminal organization that could reach out of the shadows and strike down anyone who dared stand up to it, even one of the country's most famous policemen. Dave Hennessy had had an eventful career after becoming the great hero of Esposito's capture. Shortly after that episode, he and his cousin Mike (also a policeman) stood trial for the murder of another police officer in a shootout and won controversial acquittals. Leaving the police force, Hennessy then became successful as a private detective and what would now be called security consultant, but his police career effected a remarkable comeback when Joseph Shakespeare ran for mayor as the Reform Democratic candidate against the regular Democratic—the Ring—candidate in 1888. Mayor Shakespeare was serious about rehabilitating the city's corrupt and incompetent police force. After his election to office, he appointed David Hennessy superintendent of police and instructed him to turn the demoralized, corrupt police department into a professional police force. The capture of Esposito had made him famous, and his charisma made him many friends. At thirty-two, Hennessy was a handsome fellow, with a luxuriant black handlebar mustache, his hair fashionably oiled and parted in the middle. The premature gray in his black hair only added distinction to the young chief; his steely gray eyes gave a hint of the toughness and cool nerve for which he was known. He was curiously abstemious for a tough crime fighter: He didn't drink, didn't gamble, and at thirty-two still lived with his mother. A supremely confident man, the chief got his way by reputation and force of personality. If that wasn't enough, well, he had his fists and the revolver tucked into his waistband. Hennessy had a reputation as a crackerjack detective, but great detectives don't necessarily make good administrators; Hennessy proved gifted in both roles. In only two and a half years, he shaped the New Orleans Police Department into a better disciplined and more efficient police department. An affable and sociable man, Hennessy belonged to the Red Light Social Club, a social organization in the Vieux Carré that welcomed the middle-class Irish and Italians kept out of the city's more exclusive clubs. There he became friendly with successful Italian businessmen like Joseph Macheca (suspected by some later writers as the real founder of the Mafia in New Orleans), a wealthy steamship owner and fruit importer, and Joseph and Peter Provenzano, owners of a stevedore firm. If Dave Hennessy was a man of many friends, many years in police work had given him equally many enemies. Ironically, it may have been one of his friendships that killed him. Along the waterfront, Irish and African American longshoremen monopolized the loading and unloading of rice, cotton, and sugar. Italians, however, had seized control of the fruit business. By the 1880s, the Provenzano brothers—Joseph, George, Vincent, and Peter—ran a successful stevedore firm of longshoremen who wrestled crates of coconuts, bananas, limes, pineapples, and mangoes off ships from Latin America. But their business began to suffer when brothers Charles and Tony Matranga elbowed their way onto the waterfront in 1886 and began taking contracts away from the Provenzanos. After this business rivalry turned violent, Hennessy got the two sides together and tried to broker a truce, but on the night of May 5, 1890, gunmen fired on a wagonload of Matranga workmen as it jolted down Esplanade Avenue. After the gunsmoke cleared, Tony Matranga was left writhing in agony, his leg so badly mangled by buckshot that it had to be amputated. Two others were also wounded. When the ambush victims identified Joe Provenzano, his brother Peter, and four of their men as the shooters, Hennessy duly arrested his friends, and in July, the six men went on trial for the shooting. The trial was noteworthy for the numerous policemen who provided alibis for the defense and for the defense's accusation that the Matrangas were members of the Stoppagliera, an offshoot of the Mafia. The jury ignored the policemen's testimony and the Mafia accusation and promptly convicted the defendants. The trial judge just as promptly vacated the verdict as contrary to "the law and the evidence." Many suspected that Hennessy had extended his friends a lifeline. After the trial, rumors about the chief's chumminess with the Provenzanos were rampant, and his officers were widely suspected of perjury. Hennessy's friend Joe Macheca was also very unhappy with him. Macheca, who'd testified against the Provenzanos, feared them. When he'd confided his anxiety to Hennessy several months before, the chief had reassured him that he would protect him. But now Hennessy appeared to be siding with the Provenzanos. Macheca was afraid, and he was angry, and he said so. Hennessy had three months until the retrial. He'd developed suspicions about Italian criminals in New Orleans, suspicions that some of them were connected to Giuseppe Esposito, the bandit he'd captured and extradited nine years earlier. The Provenzanos later claimed that he hoped to impeach several of the witnesses against them and prove that the Matrangas were indeed involved with the Mafia. Hennessy wrote to a police official in Rome, Louis Berti, and requested information about men in New Orleans who had been friends with Esposito. What specific names he inquired about isn't known, but he asked for photographs and facts about their criminal past. In September, Berti responded politely, promising to send photographs and criminal histories in the near future. The Provenzanos' second trial was scheduled for October 22, 1891; rumors swirled about that this time Hennessy would appear for the defense. When he was shot dead that rainy night, a week before the trial was slated to begin, it was easy to conclude that the chief had been killed to prevent him from testifying and that the Matrangas were behind it. The police were after Italians from the start—not without reason. Hennessy had identified his murderers as "Dagoes," and a block and a half from the scene of the shooting a private watchman had spotted five Italians, two of whom were trying to hide sawed-off shotguns under their coats. But the authorities showed little discrimination about which "Dagoes" they arrested. Before the chief was even dead, Mayor Joseph Shakespeare gave the order: "Scour the whole neighborhood. Arrest every Italian you come across, if necessary, and scour it again tomorrow morning as soon as there is daylight enough." On the day after Hennessy's funeral, Shakespeare made it clear what he planned to do about the Italians: "We must teach these people a lesson that they will not forget for all time." And indeed he did. Nineteen Italians were eventually charged with the assassination of David Hennessy, including Joseph Macheca and Charlie Matranga. The first lot of nine went on trial in mid-February 1891 at Saint Patrick's Hall on Lafayette Square. The evidence against them consisted of a theory of the crime—that Macheca had masterminded a plot to get rid of the chief before he could testify at the Provenzanos' retrial—and a series of problematic eyewitness identifications. The case against the accused was so weak that at the end of the two-week trial, the judge ordered a directed verdict of "not guilty" for two of the defendants, including Charlie Matranga. And after deliberating for less than a day on the fates of the remaining seven, the jury deadlocked on three and outright acquitted four more. With the pronouncement of "not guilty," a collective gasp of dismay rippled through the courtroom. How could the killers of Hennessy walk free? In the hysteria that had washed over the city, few doubted their guilt. The crowd gathered outside of Saint Patrick's Hall was stunned by the verdict. Then they became angry. "Who killa de chief?" chanted the outraged mob. "Who killa de chief?" Accusations of jury tampering flew around the city. The next morning the fury erupted. A lynch mob of some eight thousand vigilantes stormed the parish prison where the Italians were still being held on another charge. Shouting "Hang the Dago murderers!" they battered down a side gate, surged into the prison, and murdered eleven of the accused Italians, including five who had not yet been tried. Charlie Matranga survived, but Joseph Macheca was among those murdered. To this day nobody knows for certain who killed Dave Hennessy. But in the minds of most New Orleanians, there was little doubt that he'd been brought down by the "Mafia." This was the beginning of New Orleans's—and America's—obsession with the Mafia. Previously, few Americans had been familiar with the term. But Dave Hennessy's murder and the lynching of the Italians splashed the misdeeds of the "Mafia" on the front pages of newspapers all over the country. Before 1890, most violence involving Italians was attributed to the "vendetta." After 1890, New Orleans papers that had rarely referred to the Mafia now blamed the mysterious organization for almost any crime committed by an Italian. So did a Sicilian Mafia terrorize the city? The question of whether the many crimes attributed to it were committed by the Mafia revolves around the question of what is meant by _Mafia_. While the origins of the Sicilian Mafia are shrouded in mystery, the phenomenon appears to have originated in response to specific social and economic conditions in western Sicily and the south of Italy. The government wasn't much interested in the welfare of those working the land, who were often at the mercy of brigands or rapacious landowners. So peasants turned to local patrons for protection. In nineteenth-century Sicily, such a protector had to be a brave, swaggering sort of man who didn't hesitate to do what was necessary to get what he wanted—burn a neighbor's field, kill his cows, threaten his family, slit his throat. A man who had the guts and nerve and brains—and luck—to do this successfully could become a mafioso, a "respected man," who became the patron of a large number of locals. He provided protection; he did favors; he mediated disputes; he "found" stolen goods. And each time he managed to extract a share of the rewards, a sort of unofficial tribute for his services. The natural consequence of this is obvious. Offering protection evolved into insisting on it. Turning to mafiosi for help evolved into mafiosi collecting payments simply for not causing harm. It was blackmail by another name. A successful mafioso not only built a network of clients grateful for his patronage, but also created for himself a _cosca_ , literally "artichoke"—"family" in American Mafia jargon. This small group of henchmen generally came to do the dirty work, allowing the most successful mafioso—the _capo mafioso_ —the luxury of pretending not to be a thug. The more established his reputation, the more deference a "man of respect" could command and the less effort he had to put into the exhausting business of coercing compliance. People fell into line because they dared not risk otherwise. Originally, then, _Mafia_ referred not to an organization but to "a form of behavior and a kind of power." It's more accurate to say that there wasn't any Mafia as such—that is, there was no organization at all; there were only capo mafiosi and their cosca. Another way of putting it is that there were numerous mafias, independent organizations, based on client-patron relationships, scattered throughout western Sicily. By 1900 these extralegal client-patron relationships—"part armed criminal gang, part commercial enterprise, and part political clique"—were functioning as a parallel form of government, involved in both legal and illegal activities. It was confused outsiders, baffled by the situation in Sicily, who wrongly surmised that the term _Mafia_ referred to a single organization. Some of these extralegal entities stretched their tentacles as far as the New World. Most Sicilian criminal gangs in the United States were simply that—garden variety hoodlums. But some crime syndicates had Mafia connections. A capo mafioso would have been too well-protected to ever need to flee Sicily, but without a doubt a scattering of low-level Mafia henchmen crowded the decks of ships bound for America. To survive in a strange new country, they naturally turned to what they knew best. Giuseppe Morello, born in Corleone, Sicily, boarded a ship to America in the 1890s to avoid a murder charge. Born into a Mafia family, he established a successful extortion and counterfeiting ring in New York City and became the most powerful Mafia boss of his day. Gangs like Morello's could retain ties to Mafia cosca in Sicily, but they always operated independently; no Mafia boss in Corleone ever called the shots in New York or New Orleans. Sicilian gangs in America acted in a social context considerably different from that of Sicily. In Sicily, Mafia networks were embedded in society; they performed an important social function. In the United States Mafia bosses were just gangsters. _Mafia_ might be convenient shorthand for referring to Italian criminal gangs in the United States. It's more accurate, however, to see the American version merely as gangs of criminals who were, by the way, Italian, rather than true Sicilian Mafia. And in neither case did they form a large, overarching organization. But in the late nineteenth and early twentieth century New Orleans _was_ plagued by a very real threat: Black Hand crime. A type of petty extortion, the Black Hand racket was straightforward: The victim usually received a note, often signed with a grim black handprint (or sometimes skull and crossbones) demanding payment of a specified sum. If the cash was not forthcoming, the target was warned, there would be consequences. Sometimes the victim and his family were threatened with death; sometimes the blackmailers threatened to burn the family business to the ground. These warnings were often taken seriously: insurance policies could be canceled if word got out that the owner of a business had received a Black Hand note. The specter of the Black Hand generated genuine fear. (The apocryphal story of Francisco Domingo probably reflected the early existence of this kind of crime.) Sicilian immigrants offered natural targets for this kind of extortion. Their well-developed distrust of government made them reluctant to go to the police. By the 1880s, such blackmail was prevalent among southern Italian communities all over the United States. By the early twentieth century, Black Hand crime was rampant in New Orleans, where _Mafia_ and _Black Hand_ came to be used more or less interchangeably. _Black Hand_ was used to designate an organization, as well as a particular method of crime, and a mysterious "Black Hand Society" was often blamed for extortion attempts. But no Black Hand Society ever existed in New Orleans or anywhere else. Instead, opportunistic individuals or gangs sent Black Hand notes to anyone they thought could be terrorized into paying them off. This, then, was the "Mafia" that in some quarters was suspected of the Axeman's work: the Crutti and Rissetto attacks and Joe Davi's murder. For some, it was the obvious suspect. An editorial in the _Daily Picayune_ immediately after the Davi killing accused the Black Hand, noting that the "fact that all the victims were Italians of the small tradesman class should point the direction in which the clews [ _sic_ ] are to be sought." Some New Orleans police officers clung to the theory that Joe Davi's murder was a Black Hand crime. The letters Davi had received a year before his death were typical Black Hand missives, a strange mixture of graciousness and menace: "Dear Friend," the first one politely began. "Pardon us that we are going to bother you, because we are in need and we ask you to favor us with $200." But it concluded with a definite, if vague, threat: "Be careful and do what we ask you. Otherwise you will suffer the consequences." Joe Davi had shrugged off the letter and those that followed as just a bluff. Maybe that had been a mistake. This is why Pitzo made such an appealing suspect. He had a history of cadging money from Davi and others. He'd been accused of attempting to blackmail Sam Constanza. He'd been seen in the neighborhood of a grocery that had mysteriously burned. He'd argued with Davi over money. For some investigators, it wasn't such a stretch to see Pitzo as a Black Hand operator who'd carried through on a threat. John Dantonio, the Italian expert, wasn't convinced. He reasoned that the Black Hand wouldn't have left August Crutti and Joseph Rissetto alive; it was pretty efficient at killing people. There were other arguments against the Black Hand: The attacker, witnesses said, wasn't an Italian. Who'd ever heard of a non-Sicilian mafioso? Plus, neither the Cruttis nor Rissettos hailed from southern Italy or Sicily. And they didn't move in immigrant circles; their friends and associates were all Americans. Black Hand crime was an almost exclusively Sicilian and immigrant phenomenon. And the Black Hand stabbed, shot, and bombed; no one in New Orleans could recall another case in which it had resorted to an axe. So while it might not be accurate to talk about a New Orleans Mafia in a strictly technical sense, there were certainly criminal gangs in New Orleans who engaged in blackmail, kidnapping, and extortion, preyed primarily on the immigrant Italian population, and were known to use violence, including murder. Given this context, when the Axeman attacks began in 1910, suspicion that the Black Hand or Mafia might be involved was not entirely unreasonable. People believed in an Italian criminal conspiracy that targeted even modest people such as owners of a grocery store and was known to threaten people reluctant to pay them off. Joseph Mumfre was just the kind of garden variety Black Hand operator who tormented successful immigrants and fueled fears of a vast, powerful conspiracy. His square head, crooked nose, and scarred face gave him a threatening presence, but Mumfre had none of the sinister glamour of the stereotypical mob gangster. His rough braggadocio revealed him for the low-level thug that he was. Mumfre gave his profession as "labor agent"; his nickname—"Doc"—came from the patent medicines he sold on the side. The knife and bullet scars on his body, however, testified to the type of life he led. He had emigrated from Italy in his twenties and spoke heavily accented English. His background was varied: soldier, city employee, magazine salesman, money lender, letter writer for illiterate Italians, mule stealer, and horse thief. Black Hand extortion was his latest trade. One fall day in 1907, grocer Camillo Graffagnini received a letter in Italian that began with the familiar "Dear Friend." Graffagnini owned a grocery a few blocks east of the Vieux Carré. He'd done well, invested in property, and was proud of his success. He could easily have paid the $1,000 the letter requested. But he tossed the letter aside. A few days later, Joseph Mumfre entered the store. He looked around the grocery and swaggered his way up to the counter. _How is business?_ he inquired. The grocer shrugged his shoulders modestly. _It's fairly well_. _Dear friend_ , said Mumfre. _Do you have anything to give me?_ Graffagnini knew what this meant. He hesitated. Then he picked up a cigar from one of the cigar boxes sitting on the counter and offered it to Mumfre. _Here_ , he said. _Have this_. Mumfre took the cigar but looked at Graffagnini and asked again. _Are you sure you have nothing else for me?_ Graffagnini shook his head. Mumfre turned toward the door, shaking his head as he walked out; this grocer was going to be trouble; he would need more encouragement. Not long after, Graffagnini received another "Dear Friend" letter. This one warned him to comply, "Otherwise your family will fare badly." Mumfre showed up a few days later and again left empty-handed. Graffagnini saw Joe Mumfre one more time, about a month later, when he came by to have a drink in the saloon. Then the next night, just after midnight, a loud explosion outside their residence jolted the sleeping Graffagnini and his family out of their beds. When the stupefied grocer staggered outside to investigate, he found that he shouldn't have dismissed the Black Hand so cavalierly. Someone had tried to throw a bomb into his family's living quarters, up on the second story. Luckily, telephone wires had intercepted the bomb and it had dropped to the sidewalk. There it had exploded, gouging out a hole in the sidewalk next to the grocery and shattering every pane of glass in the building's door and windows. An investigation quickly resulted in Mumfre's arrest, along with four others. Witnesses had seen him near Graffagnini's store the night of the bombing. The police must have been relieved. They suspected him of being the ringleader of a Black Hand group and of involvement in several other crimes. A jury convicted him in fifteen minutes, and Mumfre was sentenced to twenty years in the penitentiary. "Doc" Mumfre was in the purgatory of the Louisiana State Penitentiary when Joe Davi had his head smashed in. But the Black Hander would come to be associated with the Axeman's murders in a way no one could have anticipated. # ≡ 6 ≡ # The Cleaver Returns AFTER JOE DAVI'S MURDER, Jim Reynolds fully expected another attack in the coming months. The Cleaver already had struck three times without being caught, without leaving any real clues to his identity. If he was a "fiend," as Reynolds suspected, he would strike again soon; he wouldn't be able to help himself. So the city waited. Especially the Sicilian shopkeepers, who looked nervously up and down the street when they shut up their stores in the evening and wondered when the Cleaver would make his next appearance. But nothing happened. The months slid by without any sign of the Cleaver. Reynolds had plenty to keep him busy: weeding out corrupt policemen, putting more patrolmen on the streets at night, and making the city safe for Carnival. He handled robberies, suicides, run-of-the-mill homicides, and the occasional mysterious murder. But no midnight marauder split open the head of an Italian grocer in the dead of night. Gradually, in the face of daily life, fear of the Cleaver slipped away. The killer disappeared for six years. Where was the Cleaver, or the Axeman as he would become known, after Joe Davi's murder in June 1911? It's not unknown for serial killers to take a long break, especially between their first and second kills, but the most likely answer is prison; the Cleaver probably went away for burglary or petty larceny. The railroad shoe pin he used was popular with thieves because it was so useful for prying open doors and windows, and the Cleaver handled one with ease. The Crutti job was almost certainly not his first. His practice of removing his shoes so as not to make noise on the wood floors also suggests some experience in housebreaking; the Cleaver knew what he was doing. Perhaps that is how he got his start. Perhaps he didn't set out to harm anyone. Perhaps August Crutti, the first victim, stirred in his sleep, startling the burglar whose original object was the week's cash. When he hit the grocer out of fear, he discovered that he enjoyed it. It made him feel powerful, like nothing he'd ever felt before. For a moment he didn't feel like a loser, an insignificant day laborer and freelance burglar. So he attacked Joseph and Conchetta Rissetto, viciously cutting both of them with his meat axe, producing plenty of blood. That they escaped alive was just luck. But with Joe Davi, the Axeman became a killer, a success in his own mind, luxuriating in a feeling of dominance and control as he bashed the young man's brains onto his pillow. Not long afterward, he was probably arrested for something else—burglary most likely—and shipped off to prison for a term. The Louisiana State Penitentiary was set up for punishment, not rehabilitation, so time spent there wouldn't cure the itch to kill. Murder is addictive, and when the Axeman returned to New Orleans, he slunk around the city, like a demon choosing his next soul. ## 3 AM, December 22, 1917 The Cleaver returned one chilly December night. Shrieks—hysterical, high-pitched shrieks—shattered the still winter air. Sixteen-year-old Mary Andollina was startled out of her sleep by her mother's cries. Instantly awake, she darted from the room she shared with her four younger sisters into her parents' bedroom to find her father sprawled on the floor, covered with blood and moaning in pain. Her first impulse was to go to her father, but her mother screamed at her: _Get the children out of the house! Someone tried to kill your father! Get the children out of the house! They'll be murdered!_ Obediently, Mary turned back into her own bedroom. She shook the younger girls awake and hurried the sleepy, protesting children out of the house. Mrs. Andollina, with her infant daughter in her arms, soon followed Mary out into the street, screaming for someone to call an ambulance. The neighbors responded, the ambulance came, and the police were called. Not only did Epifanio Andollina need to be taken to Charity Hospital, but so did his sons: the two boys—ages thirteen and fourteen—had minor injuries, but their father had several serious head wounds. As the sun came up and the cool night air warmed into a mild winter day, the police settled down to unwinding what exactly had happened. Interviewing the family, they were able to patch together an account of the night's events. Epifanio and Anna Andollina, both Sicilian transplants, had for five years run a small grocery and saloon at the corner of Apple and Dante Streets in Carrollton, about four miles west of the Vieux Carré. The large family lived in a building attached to the grocery, the parents in one room with their infant daughter, the two boys in a back room, and the five girls in the third room. Mrs. Andollina told the detectives that she had awoken to see a man standing on her husband's side of the bed, a revolver in one hand and a hatchet in the other. Seeing her eyes open, the intruder leveled his pistol directly at her and commanded, "Shut up!" She froze in terror. The man then hit her husband four or five times with the hatchet. That, apparently, was what he had come for because he then turned and walked out of the room. That's when Mrs. Andollina began screaming. On his way out of the house, the assailant ducked into the room of the two boys. As they struggled awake at the sound of their mother's screams, the intruder stunned fourteen-year-old John with a blow to the head and whacked the younger Salvatore with his pistol butt on the arm. Then he slipped out through the kitchen, dropping the hatchet on the kitchen floor. _John Andollina shows off the wound he received from the Axeman; Detective Arthur Marullo demonstrates how the Axeman entered the Andollina grocery._ Detectives closely questioned the family, but no one could describe the assailant. Mrs. Andollina, who'd seen the man only by the dim light of an oil lamp in the bedroom, couldn't even tell if he was black or white. "It was too dark," she explained, "and I was so excited at the time." Later, from his bed in Charity, Epifanio Andollina was no more help. He couldn't describe the assailant because he'd been asleep when the first blow fell and had tried to protect himself by pulling the bedclothes up over his head. He denied knowing anyone who had it in for him. "I have no enemies," the grocer declared. "I don't know why anyone should make an attempt on my life." The hatchet man, as the press called him, had come in through the back door, chiseling out a large panel of the heavy wooden door. Nothing had been stolen. No one had seen a suspicious character hanging around the grocery. It was a puzzling crime, and some of the detectives were openly skeptical of the family's story. Were they sure it wasn't just a domestic spat that got out of hand? Or maybe they'd quarreled with a neighbor and knew the assailant? _No!_ the family insisted. _That is not what happened_. Well, perhaps it was a Black Hand crime. But the Andollinas denied ever receiving blackmail demands. They could only recall a previous incident about five years before when someone had knocked out a door panel and entered the house. He stole some money from the grocery but was frightened off as he was trying to break into the family's quarters. Were the two cases connected? Almost certainly the Andollinas' attacker was the Cleaver of 1910 and 1911. That another assailant would target specifically Italian grocers in the same manner is extremely unlikely. If Jim Reynolds had been investigating the break-in, perhaps he would have made the connection with similar attacks on other Italian grocers six and a half years before. But Superintendent Reynolds was not on the case. He'd been murdered five months before. On August 2, 1917, Terrance Mullen, a patrolman known to have mental problems, entered the crowded First City Criminal Court Building. He walked upstairs to police headquarters and in front of several of the department's most senior officials shot Chief Reynolds in the eye, piercing his brain and killing him instantly. In the ensuing firefight, bullets flew in all directions, drilling into the courthouse walls and crashing into the tall glass windows as Mullen fired and ran, shooting until his revolver clicked empty. Policemen opened fire, some taking aim at the assassin as he retreated down the hallway; others, rattled, fired at random. For what seemed an eternity, chaos and confusion washed down the corridors of one of New Orleans's most important government buildings. Terry Mullen's ten-minute rampage through the First City Criminal Court Building left a shattered police department in its wake. The most popular superintendent since Dave Hennessy was dead; several other officers were wounded, one mortally. The careers of several of the officers who'd failed to protect Reynolds were ruined, forced to take early retirement or resign. Terry Mullen spent the rest of his life locked up in a state asylum. Mayor Martin Behrman was stunned and genuinely grieved by the death of his friend. And he was furious. He couldn't understand, he said to one reporter, "how a man could be shot down in his own office surrounded by a number of men, who were there to protect him." Who could be trusted to take charge of the police department? Anger over the department's apparent incompetence may well have played a role in the selection of the next superintendent. Instead of choosing from within the department, Behrman and the city's Commission Council unanimously selected Frank T. Mooney, superintendent of terminals for the Illinois Central Railroad. This appointment made more sense than it might appear. Even if his manners were a bit unrefined, Frank Mooney was a capable manager; he knew how to handle big projects and large numbers of men. He was familiar with city hall, the people in it and how it worked, and he already knew many of the officers in the police department. He was a loyal Democrat, and that he wanted the job there was no question; he'd been up for it before, in 1911, when he was beaten out by Jim Reynolds. _Superintendent of Police Frank T. Mooney._ Running the Illinois Central Railroad—one of the most important of the nine rail lines that snaked their way into New Orleans—was a complicated business. While much of Mooney's time was spent in necessary but dull public relations—attending Red Cross lectures, demonstrating the safety of the railroads to women's clubs—he also had to pacify unhappy laborers, lobby the legislature, and hammer out agreements with other railroads, as well as be responsible for the over 200,000 IC railroad cars that went in and out of New Orleans each year. All this required tact, organization, and public speaking skills. To be in charge of the Illinois Central in the city of New Orleans was a tremendous responsibility, just the kind of challenge a man like Frank Mooney relished. What it didn't require was police experience. At forty-seven, Mooney had the solid frame of a man who'd been an athlete in his youth but put on flesh in his middle years. A photo of Mooney in his thirties projects a picture of rectitude, an upright, honorable man who believed in duty and service and the efficacy of hard work. All that had certainly worked for him. Frank Mooney was a self-made man. The son of an Irish immigrant, he'd begun his career as a thirteen-year-old flagman in the Illinois Central and risen higher and higher because of his industriousness and gift for finding efficiencies. He planned to apply the same businesslike formula to the police department; what the department needed, he was sure, was a firm hand and a few new ideas. Reynolds had tightened discipline in the force, to be sure, but Mooney knew there was much more to do. The chief had been too much one of the lads, too indulgent of his men, and too lax in enforcement of moral legislation. The new superintendent planned on shaking the department up. But Mooney hadn't been involved in the investigations of the 1910–1911 attacks on Italian grocers. John Dantonio, the Italian detective who'd helped investigate those crimes, was also gone from the force by this time. Dantonio had retired in October 1917. He was only fifty-three, but he'd fallen sick with the illness that would kill him three years later. Besides, he was an early victim of Superintendent Mooney's new economizing. By 1917 Dantonio was chief of the night detective bureau, and one of Mooney's first acts was to abolish it. Dantonio reckoned that this would be a good time to retire. After twenty-one years on the force, he was eligible for a pension, and he took it. He was replaced as the Italian expert by Arthur Marullo, an Italian immigrant who'd been a detective in New York City before joining the New Orleans Police Department in 1915. Marullo investigated the Andollina case, but of course since he hadn't been on the force then, he may well have not known about the earlier attacks. Chief of Detectives George Long had been one of the investigators of the Davi murder, but he'd been of the opinion that it, as well as the other attacks, was simple robbery gone wrong. At any rate, no evidence suggests that anyone made the connection between Epifanio Andollina's assailant and the killer of six years before. What interested detectives was the connection between the Andollina case and similar attacks in the same section of the city. This was the third assault on this side of the city in a mile and a half radius in the last year. On a moonless Saturday night seven months earlier, a French-born dairyman, Vincent Miramon, had been sleeping soundly in the hayloft above his cow barn at 4131 Washington Avenue. About 1 AM, May 19, someone took Miramon's hammer from its resting place on a shelf in the barn and crept up the stairs to the Frenchman's cot. With the first blow, the prowler bashed in Miramon's head, fracturing his skull. He hit the unconscious dairyman at least six more times, hammering his face into a broken, swollen, bloody mess. The assailant then descended the stairs without taking the $150 in cash Miramon carried in his pocket. He dropped the bloodstained hammer on his way out. The police developed only a handful of suspects in the attack: several black truck farmers who'd threatened Miramon because he didn't keep his cows out of their cornfields. Eventually, they narrowed their focus to one, Cornelius Jones. But the quarrel over the roaming cows was the only evidence they had against him. It wasn't much of a motive, and everyone knew the police didn't have much of a case. At Charity Hospital, Miramon appeared to be recovering when meningitis set in. He died five days after the attack. The other case was eerily similar to the Andollina attack. Just before 4 AM on a warm, rainy night ten days after the attack on Vincent Miramon, a man entered the residence of Joseph Girard, grocer and ice dealer at 3429 Fern Street. The area was rural and thinly populated, grassy lots dotting the racially mixed neighborhood. The thirty-seven-year-old Louisiana born son of French parents ran a tiny grocery in the front room of his house; in the back two rooms he squeezed his wife and seven children. Two small flimsy wooden latches proved no obstacle to the prowler who easily forced his way past the kitchen door. He moved directly into the bedroom of Girard and his wife Adele. By the light of the lamp, he could see the man, his wife, and the toddler between them in the bed; an older child slept across the foot of the bed. Gripping a hatchet with a large blade and small handle, a type sometimes used by butchers, the intruder approached the husband's side of the bed. Standing above the sleeping figure, he delivered two hard blows, slicing Joseph Girard across the face and through the nose. Adele Girard was woken by the sound of her husband gurgling for air. She rolled over to find herself staring straight into the face of the man with the hatchet. He hit her once, and then again. He was swinging at her for a third time when his hatchet became entangled on the mosquito bar. By this time, eight-year-old Helen had woken up, taken in the situation, and started screaming. The hatchet man punched her in the face and slashed her once on the forehead. When Mrs. Girard saw her daughter slump bloody on the bed, she forced herself off the bed and leapt into the side gallery (or porch) off the bedroom hollering, "Murder! Murder!" The assailant escaped back through the kitchen. This was just the kind of straightforward case that made a policeman's life easy. When Chief Reynolds interviewed Mrs. Girard later that morning at Charity Hospital, she immediately told him that it was John Wesley Sumner, a rival ice man with whom she and her husband had clashed only two days before. She and her daughter had both seen him. Three months later, a jury took twenty minutes to convict Sumner for breaking and entering with a dangerous weapon and assault and sent him to the Louisiana State Penitentiary for life. The Andollina case six months later offered no such easy solution for detectives. Most of them probably didn't believe the family anyway and dismissed it as one of those mysterious "Italian" crimes that were nearly impossible to solve. No one much noticed when Epifanio Andollina died ten months after the attack. He died in the influenza epidemic that struck the city, his constitution perhaps weakened by the serious injuries he'd sustained in the nighttime attack. Perhaps, like Joseph Rissetto, he was another unknown victim of an unknown killer. That unknown killer had vanished into the darkness. But this time, not for long. ## Thursday, May 23, 1918 The voice on the phone that rang at the police station just before dawn was frantic with grief and shock. "Come at once!" the caller pleaded. The anguish in the man's voice was palpable. "My brother and his wife have been killed!" Captain John Dunn of the Seventh Precinct ordered a squad to come with him. The patrol wagon was hitched up, the policemen climbed aboard, and they headed north across Saint Charles Avenue toward Carrollton. The patrol wagon bumped down the dirt road, passing by grazing horses, fields of sprouting corn, and vacant grassy lots. This section of New Orleans was as much country as city, with dairies and stables and small truck farms. It was very similar to the Andollina neighborhood and not far from it, less than two miles distant; the Miramon attack had taken place just over a mile away. It was also a high-crime area. Captain Dunn knew that only three months earlier a bandit responsible for a string of drug store and grocery holdups had shot and killed a policeman nearby. When the officers arrived at the corner of Magnolia and Upperline, they found the usual small grocery and saloon with a residence in back. Pale and shaken, brothers Andrew and Jacob Maggio came out to meet them and motioned them toward the back room. Entering the bedroom, Captain Dunn came upon one of the most gruesome scenes he'd ever encountered: the proprietors of the grocery store, Joseph and Catherine Maggio, appeared to have been hacked with an axe. Dunn was surprised to see that Joseph was still—just barely—alive. The ambulance from Charity Hospital had already been called and pulled up at the grocery store just after the police detail arrived. The interns didn't even have time to put the dying man in the ambulance. They entered the bedroom only to watch with Captain Dunn as Joseph Maggio choked out his last breath. His wife had been dead for several hours. The ambulance returned to Charity empty. Captain Dunn quickly hustled Andrew and Jacob Maggio down to the Seventh Precinct station where he took their statements. The brothers were still in shock, but they managed to pour out their story. Twenty-eight-year-old Andrew Maggio had not wanted to go into the army. On his draft registration form in answer to the question "Do you have a father, mother, wife, child under 12, or a sister or brother under 12, solely dependent on you for support?" he had written "Mother." This wasn't true. Not only did his mother live with his sister in Arkansas, but he also had three older brothers, all successful businessmen, to share in the support of their mother. Surely Andrew, a barber, wasn't expected to do so alone. He didn't even have his own home; he lived with Joseph and his wife. No, clearly Andrew wanted to duck military service. But now the army expected him to defend his classification as "exempt because of dependency." He'd received instructions to report to the exemption board office next week. Later he claimed to have been out drinking that Wednesday night. Perhaps he needed to steady his nerves for the coming interview. That would explain why he'd slept so deeply that night. About 4:30 that morning something woke him from his sleep. At first he couldn't tell what it was. Then he heard it. A low groan coming from the other side of the wall, from Joe and Catherine's room. Andrew sat up in bed and listened again. He heard nothing. He rapped on the wall. No answer. Andrew wasn't sure what to do. He had a feeling that something was very, very wrong. But he wasn't used to barging into his brother's bedroom. And he was afraid of what might be on the other side of the wall. As the youngest of four brothers, Andrew was used to following, not leading. So he pulled on some clothes and ran down the block to the house of Jacob, another of his older brothers. Waking him, Andrew told Jake that he was afraid something terrible had happened. The two men returned to the house and went around the back to enter by the kitchen door. What they saw there made them even more apprehensive about what they'd find inside. The kitchen door stood open; one of its panels lay on the ground. Anxiously, the two brothers went through the kitchen, past the bathroom, and stood at the door of the bedroom. Jake tapped on the door. There was no answer. Slowly, Jake pushed open the bedroom door to reveal the horrific scene. Reeling from the discovery, they immediately called the police and ambulance and waited for help to arrive. After the brothers finished their statements, Captain Dunn kept them at the station, using the handy excuse that they were "material witnesses." He knew Chief Long and Superintendent Mooney would want to talk to them. Back at the Maggio grocery, Bertillon Operator John Norris flashed away with his camera. Superintendent Mooney, Chief of Detectives George Long, and a handful of other officers inspected the bodies. This wasn't Mooney's first big murder case. Most of the others, however, had been straightforward shootings that didn't require much detective work to determine who the shooter was. None had been anything like this. Even if he was new to police work, Mooney planned to apply the same systematic thinking that had always served him so well to this murder case. Joseph was on the bed with his feet hanging off the side. Catherine lay on the floor at his feet, sprawled on her back, a gaping wound on the right side of her neck. Both were drenched with blood. An axe stained with blood had been discarded in the nearby bathroom, dropped into the bathtub; it was the couple's own axe, last seen lying in their backyard. Blood on the bed, blood on the floor, and blood spattered seven feet up the wall testified to the ferocity of the attack. A few hours after the bodies had been removed to the morgue, Dr. O'Hara's autopsy showed that both had had their throats cut. In addition, Joseph had been hit twice with the axe, fracturing his skull, and cut a couple of times on his face and neck. Catherine hadn't actually been hit with the axe at all but had died an even more horrific death. She had gotten out of bed, the police speculated, to defend her husband from his killer. But he'd slashed her seven times—on the face, on the shoulder, on the hand, the last probably a defensive wound received as she attempted to protect herself from the deadly razor. The killing stroke had cut deep into the right side of her neck, slicing through the muscles, internal jugular vein, and carotid artery and cutting into her airway. Such a stroke must have immediately dropped her to the floor, where her gasps for breath would have sucked the gushing blood into her airway, drowning her in her own blood as she simultaneously bled to death. _Joseph and Catherine Maggio and the grocery where they were murdered._ The means of the killer's entry they'd seen before at the Andollina grocery: he had chiseled out a panel in the kitchen door and then reached through to the lock. Once inside, he went straight through to the bedroom. "Robbery was the motive," announced the _New Orleans States_ , reporting what the police were telling the public. The small safe in the bedroom was open; likely it had been left unlocked because Joe Maggio had deposited $650—several days' receipts from the grocery—in the bank the day before. A tin strongbox had been forced open and fifty dollars taken. The jewelry box hadn't been tampered with, although it contained jewelry worth much more than the money stolen. Dresser drawers had been pulled out and clothes strewn about. After Dr. O'Hara took charge of the bodies and sent them off to the morgue, Superintendent Mooney had the house and grocery locked and put under guard. The half-dozen detectives assigned to the case spent the rest of Thursday interviewing neighbors and family members and finding out as much about the Maggios as they could. Investigators didn't systematically search the premises immediately. Before the development of scientific crime scene analysis, such things were less important than tracking down witnesses—and suspects—as quickly as possible. Even neophyte policeman Frank Mooney knew that the way to solve a case was to find a suspect and sweat a confession out of him as soon as possible. A time-consuming search of the residence and grocery could wait. There were few surprises in the information the detectives gathered. The Maggios were typical Sicilian immigrants, small-time businessmen. Forty-year-old Joseph was the eldest of the four brothers. Andrew, the youngest, had a barbershop down in the Central Business District. Jacob, thirty-six, owned a shoe repair shop only about a block (or a "square," according to New Orleans vernacular) away. A fourth brother, Salvatore, thirty-three, ran another grocery. Joseph took over as the head of the family when their father died years before. He had been the first to leave Sicily, immigrating to America and sending for the rest of his family as soon as he could. Joe had married Catherine in 1903. A photo of them on their wedding day shows a sober groom, his dark mustache curling upward, and a self-conscious Catherine, still radiant in her wedding finery. Except for the unusual circumstance of having no children, they were an ordinary, happily married couple, working seven days a week to build a business that served both black and white customers who affectionately knew them as "Mr. and Mrs. Joe." The Maggios were careful with their money, as Sicilian immigrants usually were, investing their small excess funds in Liberty Loan bonds. And patriotic as only new Americans can be, the Maggios tucked away their Red Cross pledge cards among their other important documents. They were the "right sort" of Italians, the detectives discovered, hardworking merchant _petite bourgeoisie_ who had no enemies or questionable associates. No one had any reason to harm them. In the course of his inquiries that morning, Detective Theodore Obitz stumbled across a bizarre message chalked onto the banquette—the sidewalk—a block away. What the message actually said is uncertain. Police records that might have preserved the message are lost. The _Times-Picayune_ reported it as "Mrs. Joseph Maggio is going to sit up tonight just like Mrs. Toney." The _States_ had a slightly different version _:_ "Mrs. Joseph Maggio is going to sit up tonight. Just write Mrs. Toney." The _New Orleans Item_ carried a third variant: "Mrs. Joseph Maggio will sit up tonight—just write. Mrs. Tony." That all of the newspapers carried a different version of the message suggests that the reporters didn't see the writing themselves but relied on secondhand accounts muddled in the retelling. The versions in the _Item_ and _States_ are closer to each other, but the one in the _Times-Picayune_ makes the most literal sense. Was it a clue? What did it mean? None of the various versions made much sense. The newspapers bandied about explanatory theories about the message ranging from its having been left by the deranged killer to it being the work of an accomplice warning the Maggios to be on their guard, to a practical joke of unusually bad taste. No wonder that Captain Dunn's statement to the press on Thursday evening could only say that this was an "unusually mysterious" case. While detectives were still puzzling over the chalk message Thursday afternoon, another murder weapon turned up. A black woman the newspapers designated a "colored girl" working next door noticed a straight razor sticking out of the rose trellis on the lawn of the house next to the Maggio property. She alerted the owner of the property, who notified the police. Chief of Detectives George Long turned the tortoise shell–handled razor over in his hand. The stainless steel blade was only three inches long, but it was deadly sharp. Dried blood clung to the blade and the black and yellow handle. _If we can find the owner_ , Long thought as he folded the blade into its handle, _we'll have the killer_. Long immediately sent Captain Dunn to Andrew's barbershop to interview his employees. Under questioning, one of them, Estaban Torres, conceded that he had seen Andrew take a similar razor from the shop a few days before, saying that it needed to be honed. This was just the piece of evidence that Long and Mooney needed. From such minor details cases could be solved. The blood-smeared razor was almost certainly the murder weapon, and now the investigators might be able to connect it to one of the chief suspects, someone familiar with the house and grocery. Andrew had aroused suspicion from the beginning because he claimed to have slept through two people being hacked to death in the next room. It also seemed odd that Andrew hadn't inspected the sounds in Joe's bedroom himself but instead had gone to get Jake. He was either a coward, the detectives thought, or else something more sinister. Optimistic that they were on their way to solving the case, Mooney let Jake go home Thursday night. But not Andrew. Less than twenty-four hours after the murders, the investigators had zeroed in on a suspect. Now they were ready to thoroughly search the crime scene. Perhaps they could turn up more evidence to support a charge against Andrew. On Friday, Superintendent Mooney, Chief of Detectives Long, Captain Dunn, Detective Obitz, and Obitz's partner, Detective James P. Ford, returned to the Maggio grocery where they went through the premises room by room. In the murdered couple's bedroom, they found a blood-spattered suit of clothes in the corner, a pair of blood-stained socks, a bloody footprint, and a loaded revolver. A search of Andrew's room turned up three more straight razors, not very surprising since he was a barber, none with blood on them. Of more interest was a shirt found in his bathroom—a stiff-bosomed white dress shirt spotted with faded dark stains that someone had tried to wash out. Were they bloodstains? The search was as thorough as the investigators could make it. Detective Obitz even wedged himself uncomfortably under the house looking for evidence. In the backyard, one of the investigators found a screwdriver matching the marks in the door panel. Like the axe and razor, this tool had been abandoned by the killer. In the grocery, footprints of stocking feet were spotted on the bar's newly varnished counter, indicating that the killer hadn't worn shoes. Perhaps, someone speculated, he had climbed on the counter to switch on a light. Once they completed the search, the detectives returned to their witnesses. Yes, some remembered that on the day before the murders, Andrew had worn dark clothes, similar to the blood-stained ones found in his brother's bedroom. The case against the youngest brother seemed to be tightening. Then investigators suffered a setback. When they reinterviewed Estaban Torres, he looked carefully at the three razors found in Andrew's room. He pointed to one of them, saying that it looked like the one Andrew had taken from the shop. Now, there was no evidence of any connection between Andrew and the bloody murder weapon. Nevertheless, he remained the only suspect. As rumors spread that he would be charged soon, Andrew begged to be allowed to attend his brother and sister-in-law's funeral. Mooney refused to release him, citing "the unsettled state of the case." On Friday afternoon, two services were held to accommodate the many mourners. The first was at the funeral parlor on Toulouse Street, where the overflowing crowd spilled out into the street, blocking traffic. A second service was held at Saint Mary's Italian Church, the parish church for Italian immigrants in New Orleans, situated at the site of the old Ursuline Convent in the French Quarter. Here the tearful young nephews and nieces of the childless couple trudged by the caskets to say good-bye to their aunt and uncle. After the funeral mass, pallbearers loaded the flower-laden caskets into two hearses. The melancholy procession then slowly made its way down Esplanade Avenue the two short miles to Saint Louis Cemetery No. 3, where Joseph and Catherine Maggio were interred in the same vault. Meanwhile, Joe's youngest brother sat in the parish prison, suspected of their murder. On Saturday morning, accompanied by Assistant District Attorney Ben Daly and Dr. O'Hara, Superintendent Mooney made a final examination of the Maggio residence and grocery. It was an opportunity for Mooney to show Daly and O'Hara the crime scene, outline his theory of the crime, and listen to their views. Mooney told them he was inclined to think that Andrew was their best suspect, but a lot would depend on the interview and how Andrew answered the investigators' questions. The case could go either way. Finally prepared to confront him with evidence, Mooney had Andrew Maggio transferred from the Seventh Precinct to police headquarters on Saturday afternoon, where he was ushered into the superintendent's office. The young man was seated in front of Mooney's desk; Superintendent Mooney stationed himself on one side of Andrew; Assistant District Attorney Daly came at him from the other direction. For four hours Mooney and Daly grilled Andrew, pelting him with question after question and flinging his brother's bloodstained clothes in his face. Hour after hour Andrew repeated his story, insisting again and again that he did not kill his brother and sister-in-law. The bloodstained razor wasn't his; he'd never seen it before. He didn't know who it belonged to. The bloody clothes in the room belonged to Joseph. He'd slept through the noise of the murders because he'd been drinking the night before and was soundly asleep. The stains on his white shirt were from wine, not blood. He'd spilled a glass of wine over himself at a wedding. Over and over again Andrew protested his innocence. Finally, he could stand it no longer: "How could you think I could kill my own brother?" he demanded, his voice choking on tears. "He sent us the money to come to America. He supported us after our father died in Sicily. Joe was like a father to me since I was eleven years old. I know a man isn't supposed to cry but . . ." and slumped over, his face in his hands, sobbing. Mooney sat back in his chair and looked on thoughtfully as the distraught young man finally got control of himself and wiped his eyes. He had planned to charge Andrew, but now the case against him didn't seem strong enough. Mooney couldn't claim to have a policeman's long-honed skill in interrogation. But his career so far had taught him to be a judge of character, and his instinct told him he should look elsewhere for the murderer. When Andrew was released Saturday night, Superintendent Mooney made a statement to the press: "Up to the time we conducted a rigid cross-examination of Andrew Maggio, I suspected strongly that either he committed the crime or had knowledge of the identity of the person or persons who did so. Andrew, however, explained so readily and clearly some of the puzzling threads of evidence that we gathered, and which we thought might be binding on him, that Mr. Daly and I concluded that the doubt as to Andrew's guilt was so strong that we were compelled to give him the benefit of the doubt." Mooney promised that the investigation would go on, declaring, "Our hope for a solution is still bright." Privately, however, he may have felt much less optimistic. But it was important to bolster public confidence in the police force. Having the public frightened of a razor-wielding maniac who killed with impunity could only complicate the police superintendent's job. One step Mooney took to inspire confidence at the beginning of the investigation was to announce that he had assigned Chief of Detectives George Long to take charge of the case. The New Orleans public had faith in Long; he had a reputation as a clever detective who could solve difficult cases. The Reidel murder, a few years earlier, had been one such case. In the fall of 1910, the bloated, decomposing remains of a middle-aged man sown into coffee sacks bobbed up in the Old Basin Canal. While the body was quickly identified as that of Franz Reidel, a German immigrant and watchmaker who'd gone missing with his life savings, investigators didn't have much to go on, except that the victim had died of blunt force trauma. Jim Reynolds, chief of detectives at the time, had little else in the way of clues or suspects. "We don't know where to start," he conceded to a newspaper reporter. "There is absolutely nothing to give us any idea of what to do." The _Daily States_ was calling it "one of the criminal mysteries of the year." Captain George Long unearthed the clue that broke the case. At the time, Long wasn't even in the Detectives' Office. But he had a flair for investigative work and kept his eyes and ears open, as Reynolds advised his officers to do. Long's patience was rewarded by a tip that Reidel had been seen entering a door at 630 North Rampart Street, on the edge of the French Quarter. More nosing around revealed that this was the address of an unsavory character named Eugene Bescanon, a tall, fish-eyed French immigrant with a criminal past, who had been an acquaintance of Reidel's. Further probing led to a squalid apartment in a dilapidated building on North Rampart Street (only two blocks from the Old Basin Canal) where investigators found bloodstained floors and rags. This discovery quickly led to Bescanon's arrest, confession, and implication of his accomplice, a fellow Frenchman named François Rodin. The unsuspecting Reidel had been lured to Bescanon's apartment, where he was battered to death and robbed. The recovery of Reidel's watch chain, a distinctive one with a horse head locket, confirmed the men's guilt. The investigation had solved the mystery of Franz Reidel's murder and produced two culprits to hang—all because of a tip that might have escaped a less vigilant officer than George Long. And Chief Reynolds wasn't stingy about sharing the credit. Captain Long's part in unraveling the mystery was publicly celebrated, with one newspaper editorial "award[ing him] the first honors in the quest." The case was a triumph for the New Orleans Police Department, showing that with thoroughness, perseverance—and a little luck—apparently hopeless cases could be solved. Mooney only hoped Detective Long could accomplish a similar turnaround with the Maggio murders. By May 1918, the United States had been in the First World War for just over a year. For months the New Orleans papers had carried stories about Red Cross fund drives, violations of the Espionage Act, and the first military conscription in America since the Civil War. But for the people of New Orleans the war was far away compared to the immediacy of the Maggio murders. The city was no stranger to violent death, but not since Joe Davi's death had it seen this kind of butchery. Newspaper accounts of the investigation into the murders of Joe and Catherine Maggio inspired far greater interest than stories of Zeppelin attacks or U-boat warfare. As they drank their chicory coffee and contemplated the grisly murders, New Orleanians couldn't help but wonder whether all the attacks on Italian grocers were related. The Crutti and Rissetto attacks, the Davi murder, the Andollina case, and now the Maggio murders. Were they all connected? And who was targeting these people? And why? Another name cropped up again and again in discussions as alert readers of the New Orleans papers thought back to another murder of another Italian grocer six years before. Less than a year after Joe Davi's murder, another Italian grocer died in the middle of the night. Superficially, the crime was similar to the attack on Davi—a young grocer brutally murdered as he slept next to his wife. Yet the two murders were dissimilar in one crucial way, a way that indicated a different kind of criminal and a different type of motive. At twenty-seven, Tony Sciambra was already a success. He'd learned the business as a boy in his father's Carrollton grocery and now ran his own grocery and bar at the opposite end of town, on France and Villere Streets in the Bywater District. The hard work and frugality that came as naturally to first-generation Americans as it had to their immigrant parents had allowed him to pay cash for the property. A personable young man, Tony was married to "Mrs. Tony" as the locals called her—pretty twenty-three-year-old Johanna. They were well on their way to producing the hoped-for houseful of children: married not quite two years, they had one son—eleven-month-old Jake—and another baby on the way. So, as Tony slept in his bed on the morning of May 16, 1912, he had reason to be pleased with himself, and with life. _Tony, Jake, and Johanna Sciambra._ But not everyone, apparently, was pleased with Tony Sciambra. About 2 AM on that morning, an intruder stacked a couple of soapboxes on top of each other and climbed up to reach a kitchen window at the back of Tony and Johanna's house. Opening the shutters, then raising the unlatched window, he crawled into the kitchen. Once inside, he paused and listened; no one in the house stirred. One door off the kitchen led into the grocery. He chose the other door, the one leading through the small dining room to the room where the young couple and their son were fast asleep. Entering the bedroom, the intruder attacked Tony, killing him almost instantly and inadvertently wounding Johanna; the baby sleeping next to them was unharmed. How was this crime so different from the murders of Joe Davi or Joe and Catherine Maggio? Tony Sciambra was shot to death. In the bedroom the assailant walked up to the bed, pulled the mosquito netting out of the way, pointed his .38 caliber pistol at the sleeping grocer, and pulled the trigger. Again and again he fired. He unloaded five shots into Tony before turning around and slipping out through the kitchen door, giving no thought to the jewelry in the wardrobe or the cash in the till. _The Sciambra grocery and home, where Tony Sciambra was shot to death._ The first three bullets entering Tony's back caused him to jerk involuntarily so that the next two hit with less accuracy, wounding him in the side and the arm. But it didn't take long for internal hemorrhaging to kill him. Tony was dead by the time the ambulance arrived. Johanna was killed accidentally. One bullet shot right through her husband's body, struck her in the hip, and penetrated her abdomen. At first everyone thought she would live, but the wound went septic, and she died of peritonitis ten days after her husband, leaving little Jake to be raised by his grandparents. New Orleans vibrated with speculation over the assault on Tony Sciambra, yet another Italian grocer. People whispered to one another that it must be connected to the Davi murder less than a year before, and the Crutti and Rissetto attacks before that. There was no shortage of theories: it was the vendetta or the Black Hand; newly arrived Sicilians were trying to take the grocery business away from "local Italians"; the attacks were revenge for mistreatment by Italian grocers; the killer was a fiend. The most popular notion was that the Black Hand was at it again. An indignant editorial in the _New Orleans Item_ demanded that "the better class of Italians"—i.e., those not criminal—"take active measures . . . to induce their own people to tell what they knew regarding these crimes." It was only in this way, the editorialist continued, that Tony Sciambra's murder could be solved and "the authorities will be able to discover and punish the guilty." The writer may well have been right. But he, like most commentators at the time, didn't recognize the significance of the choice of murder weapons. A gun and a blade are both lethal, but they kill in very different ways. Many aspects of the Crutti, Rissetto, Davi, and Sciambra (and later the Andollina and Maggio) attacks were similar: targeting a successful grocery and bar in an unfashionable, isolated part of the city (the Sciambra grocery was less than a mile from the Crutti place) and breaking in during the dead of night. Robbery was never the motive. The assailant usually targeted the man first, if he attacked the woman at all. Harriet Crutti, Mary Davi, and Anna Andollina escaped relatively unscathed. Only Joseph and Conchetta Rissetto were attacked with equal ferocity. And perhaps Catherine Maggio wouldn't have died if she hadn't attempted to defend her husband. All of these attacks were made with a blade of some kind—a meat cleaver, hatchet, axe, or razor—something that would draw blood. That's what the Axeman craved. If all the Axeman wanted was the death of Italian grocers, he could have used a gun. He had one when he entered the Andollina home. It would have been less trouble—cleaner, more efficient, deadlier. But he never fired it, using it only to terrorize the victim's wife. The survival of several victims attests to the fact that an axe or cleaver wasn't as sure a method of killing as a revolver. So, why choose such an insecure weapon as a blade? Because guns are impersonal. Killing a man by standing over him and bashing his head with a hatchet or cleaver is a much more intimate affair. A blade is an intensely personal weapon, requiring the killer to actually touch his victim and risk being splattered with blood and gore. A killer who chooses this kind of weapon wants more than his victim's death; he wants the power that comes with holding a man's life in his hands and then choosing to spill his blood and brains out on the bed. Amid all the speculation after the murders, the _Times-Democrat_ was closest to the mark, recording that many people attributed the killer's actions to a "lust for blood." The Axeman was willing to take the chance that his victims might actually survive. Tony Sciambra's killer wanted something else. He only needed Tony dead. He didn't need to enjoy it. He didn't need to derive psychological pleasure from it. He didn't require the satisfaction derived from bringing an axe down on a helpless, sleeping man and spattering his blood and brains over the bed. Tony Sciambra was murdered with a weapon meant to be used at a distance, one that didn't require physical contact between the killer and his victim. So, if Tony Sciambra and his wife weren't victims of the Axeman, who was responsible for their deaths? A clue may be found in a business transaction involving Tony's brother, Henry. After Tony's death—it's unclear how long—Henry Sciambra sold his own thriving grocery business at Marigny and Dauphine. The neighbors, the _Times-Picayune_ said, shook their heads, puzzling over why the young man would sell a business that had been doing so well. The man Henry Sciambra sold out to was Vito Di Giorgio. Di Giorgio was reputed to be a major figure in the local "Mafia," the New Orleans gangsters who thrived on blackmail and counterfeiting. He'd been arrested in 1908 as the ringleader of a gang of Black Hand extortionists who'd tried to blow up a dry goods store when the owner refused to pay, but the police couldn't make the charges stick. He also had connections to a local counterfeiting ring and to New York crime boss Joseph Morello. Was Tony Sciambra's death connected to the sale of his brother's store? Was it meant to serve as a warning? Or had Tony offended someone, someone who resorted to the vendetta and took a merciless revenge? That wasn't all. Notorious Black Hander Joseph Mumfre—the man convicted in 1908 of trying to bomb grocer Camillo Graffagnini's home—may, possibly, have been mixed up somehow with Di Giorgio and the Sciambras. Mumfre knew Di Giorgio; that was certain. Police suspected Mumfre—out on bond before his trial for dynamiting the Graffagnini home in December 1907—of involvement in the Black Hand bombing of another grocery in June 1908, for which Vito Di Giorgio was also arrested. After Tony Sciambra's death, as Chief Reynolds, Detective Dantonio, and the other investigators began probing into the couples' background, trying to find a motive for the gunning down of the young grocer, neighbors whispered that Tony Sciambra had heard that relatives of Mumfre—who had begun serving his twenty-year prison term—lived in the neighborhood, and this had frightened the young grocer enough that he'd considered selling his store. He'd gone as far as lining up a prospective buyer when he changed his mind. The would-be buyer, however, wasn't to be put off and had shown up at the store with a companion in tow, to persuade Tony to sell. One woman confided to detectives that a few months previously Johanna had been "considerably worried because of an Italian who looked dangerous . . . [and] had been to the store." Reynolds and his men followed up on these leads, without any result. They questioned the two men who'd been in to see Tony. They also questioned Johanna Sciambra, who was well enough to be interviewed the morning after the murder. No, she hadn't seen the shooter. No, she and her husband had known nothing about Mumfre relatives in the neighborhood. No, they'd never received any Black Hand letters. Yes, her husband had considered selling but only because he hadn't gotten along with the landlord. To add to the mystery, George Musacchia, the same grocer whose store had burned down just before Joe Davi's murder, came forward to claim that someone had also tried to break into his house the night of the Sciambra killing, using much the same method. He'd discovered a barrel sitting under one of his windows; someone had broken some of the slats in the blinds in an apparent attempt to get in. Reporters noted that Musacchia's grocery sat next to a vacant lot, just like the Davi, Rissetto, and Sciambra businesses. Why had the intruder abandoned the break-in? What connection—if any—did it have to the Sciambra case? The different threads of the investigation never came together to form a coherent picture for the detectives. The murders of Tony and Johanna Sciambra were added to the list of the city's unsolved crimes. But contrary to the belief of most in his city, Chief Reynolds had concluded that the killing of the Sciambras and the murder of Joe Davi (and so presumably the attacks on the Cruttis and Rissettos) were products of very different minds: the Sciambras were, he thought, yet more victims of the Sicilian vendetta, still an all too common occurrence in the city. But Reynolds was convinced that Joe Davi's murder was not the work of a criminal gang but "bore all the earmarks of the acts of a degenerate." He never heard the term _serial killer_ , but Chief Reynolds had clearly grasped the concept. If he'd lived, Chief Reynolds presumably would have agreed that the Maggio case had more in common with the Davi than the Sciambra murders. But most people didn't think that way. Most people only remembered that some Italian grocers had died mysteriously under similar circumstances. The person who chalked the mysterious message onto the sidewalk—"Mrs. Joseph Maggio is going to sit up tonight just like Mrs. Toney," or something similar—was probably only doing what many in New Orleans were doing that warm May morning of 1918, as news of the slaughter of the couple swept through the city—linking the murder of one innocent Italian couple with the other. The police weren't in a position to tell them anything different. Less than a week after the Maggio murders, the investigation was flailing. The most obvious suspect had been released and no obvious leads presented themselves to Superintendent Mooney and his detectives. The investigation suffered another blow when Detective Theodore Obitz, one of the officers investigating the case, was himself murdered. Obitz had spent the Saturday after the murders running down clues, but at 3 AM on Sunday morning he was cut down by a bandit he was chasing through the deserted streets of the city. The Italian community might have been understandably jittery, but the rest of the city was resigned to the likelihood that the Maggio murders would in all probability not be solved. Confusion as to what kind of killer the city faced—a bloodthirsty fiend or a criminal gang—only unsettled people more. Then another grocer was attacked. # ≡ 7 ≡ # A German Spy? ALL OF NEW ORLEANS was uneasy in the wake of the Maggio murders. The Italian community, especially, stirred apprehensively. Memory of the first two attacks in the fall of 1910, the subsequent murder of Joe Davi, and the assassination of the Sciambras created an undercurrent of anxiety, a worry that Italian grocers had special reason to fear. Gossip about a new and lethal Black Hand gang or a fiend with a grudge against Italian grocers was whispered up and down Little Palermo as the city struggled to make sense of the latest killing. Then, just a month after the Maggio murders, when the police still had hopes of stumbling upon a Reidel-case-like clue and running it to ground to find the killer, an axeman struck again. This time, however, the victims were not Italian. ## 6:50 AM, June 26, 1918 Thursday morning dawned hot and humid. It was going to be another sweltering New Orleans summer day. But John Zanca wasn't thinking about the weather when he stopped his delivery wagon in front of the People's Cash Store, a grocery owned by an Eastern European immigrant named Louis Besumer. _The Besumer grocery._ Zanca stared at the shuttered store, puzzled. It was his usual time, but the little yellow grocery sat closed and bolted shut. Who was going to accept his bread delivery? He jumped out of his wagon and knocked on the door. No answer. He knocked again, more forcefully. Nothing. He hammered on the door. When he still got no reply, Zanca walked from the store entrance on North Dorgenois Street around the corner to a door facing Laharpe Street that led directly into the grocer's residence. Rapping on the door, at the same time he called out, _Is anybody home?_ Finally, he heard footsteps on the inside shuffling toward him. But the door still didn't open. Zanca strained to make out the muffled words from the other side: _Come around to the front_. Zanca turned and walked again to the store entrance. The door of the grocery finally opened a crack, and Louis Besumer's unshaven face peered out. Usually a distinguished-looking man who tried to exude a sophisticated, European air, this morning the grocer looked aged and haggard. Zanca was startled to see that his face was streaked with blood. "My God, what's happened?" gasped Zanca. Grimacing slightly as he dabbed at his bloody head with a sponge, Besumer replied, "Nothing. Don't worry about it. There was an accident." _Has your wife been hurt?_ asked Zanca, concerned for the attractive black-haired woman who lived behind the store with Besumer. The old man didn't know. When Zanca pressed him, Besumer brusquely told Zanca to check on her himself. Brushing past the grocer, Zanca hurried through the store into the residence and made his way down the hall to the second bedroom. There he found the woman lying blood-soaked and nearly unconscious in her bed. When she opened her eyes to see the young man leaning over her, she whispered, "I'm cold." Despite the hot day, Zanca looked around the room, grabbed some clothes off the floor, and tucked them around her as best he could. Returning to the grocery store, Zanca told Besumer that he was going to call the Charity ambulance. _No, don't do that_ , Besumer responded hastily. _Just fetch a doctor_. _But she's cut up pretty bad!_ Besumer gave up. "Suit yourself," he shrugged. He was pale and drawn, without the energy to argue. Zanca later told police that he seemed weak. _You should go lie down_ , Zanca suggested as he sped away to alert the nearest policeman. Besumer was lying on his bed when a squad from the Fifth Precinct arrived at the grocery a few minutes later and entered the scene of what would later be called "one of the queerest [mysteries] in the annals of New Orleans crime." The store fronted a two-bedroom residence with a hallway and a small, screened back porch at the end of the hall. On the porch, fresh blood pooled around a woman's hairpiece. The wooden handle of an ancient short-handled axe lay nearby; its rusty blade sat a good five feet away. Both were bloodstained. Bloody footprints led from the porch through the hall, into the bedroom, and right up to the bed where the woman still lay, covered in blood. Realizing that he had another axe attack on his hands, the corporal in charge immediately notified his commander, who informed the superintendent of police. Before the morning was out Frank Mooney, Assistant DA Ben Daly, Coroner O'Hara, and a handful of detectives had congregated at the small grocery and residence. _Sketch of the Besumer crime scene._ One of the first things investigators noted was the fresh blood: The attack must have occurred only a short time before, perhaps an hour before John Zanca's delivery wagon pulled up to the grocery. Its basic outline could be reconstructed without much difficulty: The woman had been struck on the porch so violently that the axe blade had flown off its handle. Somehow she'd managed to stumble from the porch to her room and collapse on her bed. There was no sign of forced entry. As the Bertillon operator busied himself fingerprinting the axe and photographing the bloodstains, detectives considered Mrs. Besumer's bedroom. (Later Louis Besumer explained that he and his wife had separate rooms because the noise of the electric fan that he ran all night kept Mrs. Besumer awake.) They noted that the room's windows sat right on Laharpe Street and were well lit by a nearby streetlight; anyone passing by on the busy street could see straight into the bedroom. If this attack, like the other axe attacks, had begun in the bedroom, the assailant risked being seen by passersby. And if an intruder had broken into the grocery, the policemen wondered, didn't the victims scream for help? And if they did, why didn't any of their neighbors hear them? Next, investigators checked Louis Besumer's bedroom. Someone had gone through the contents of his dresser and wardrobe and left them scattered about the room, but nothing seemed to have been taken. The only blood in the room was on the pillow. Robbery hadn't been a motive. No one had touched the grocery safe, and Mrs. Besumer's jewelry was still in the dresser; the cash in her pocketbook remained where she had left it. James Reynolds was dead and John Dantonio had retired, but several of those inspecting the crime scene were veterans of the previous axe attacks. Dr. O'Hara was still the coroner, and George Long remained chief of detectives. Ben Daly, the assistant district attorney, had been at his job since 1909. The experience of men who had seen the Axeman's work before would no doubt prove valuable to the new police superintendent. Leaving his capable detectives to untangle the scene at North Dorgenois and Laharpe Streets, Superintendent Mooney went to Charity Hospital to check on the injured couple. Both victims had head injuries. Besumer had a single wound, a long cut over his right eye. The blunt, rusty axe blade may have fractured his skull, but it was not, house surgeon Hiram W. Kostmayer assured police, a life-threatening wound. The woman's injuries were much more serious. In addition to several gashes on her arms and chest, she had been hit twice on the head, cracking her skull. Severe brain trauma can damage the body's ability to regulate itself, which explains why earlier she had been so cold on such a hot day. The seriousness of Mrs. Besumer's injuries made the police even more anxious to question her. Senior Captain Thomas Capo spoke to her as soon as the doctors would permit. Doped up on painkillers, she stared hazily at him when he asked what had happened. _I don't know_ , she replied. _Was there a fight on the gallery?_ _Oh, yes. A mulatto wanted to buy some tobacco after the store opened. When I told him that we didn't sell tobacco, he attacked me._ The attack had taken place before the store opened. Mrs. Besumer had been found in her nightclothes, not dressed to wait on customers. Her story couldn't be true. But Capo dutifully wrote it all down. _Louis Besumer._ Frank Mooney himself interviewed Louis Besumer. The grocer told him that he was sixty years old, from Poland, and had lived in New York before moving to Jacksonville, Florida, about two years ago. There he met and married Harriet Anna Lowe. They'd bought the grocery after moving to New Orleans three months ago. Besumer said that he didn't have a clear memory of what had happened to them the night before. He had woken to find himself bleeding and the door leading into Laharpe Street open. When he'd discovered his injured wife, he said, he covered her with a sheet. Later police interviews with neighbors told Mooney even more about Besumer. He was a strange man with pretensions to culture and sophistication unusual in a small-time grocer. Claiming to speak thirteen languages, he bragged about being rich. Rumors circulated that he'd traveled extensively in Europe and Mexico, that he'd owned a coffee plantation in South America and a sheep farm in Kentucky. Everyone said that he sold groceries cheaply, too cheaply some thought. How could he stay in business if he sold below cost? No one knew where he got his money. The neighbors couldn't, however, tell investigators much about the early morning events at the grocery. And as Mooney thought about it, several oddities struck him about the attack. If, as the detectives calculated, the assault had occurred between 6 and 7 AM, the neighborhood would have been awake and moving. Yet a canvas of the surrounding area revealed that no one—mulatto or otherwise—had been seen leaving the People's Cash Store early that Wednesday morning. No one had heard any screams or sounds of a struggle, or spotted any strangers. Someone had seen a light on about midnight. But when questioned, Besumer denied knowing anything about that. And almost right away, Mrs. Besumer had contradicted part of her husband's story. At some point in her pain- and drug-induced daze on the morning of the attack, she had babbled to a policeman, "My husband is a German. He claims he is not. I don't know where he got the money to buy his store." And that made him a possible suspect as well as a victim. Mooney quickly sent men to search Besumer's private papers. The search revealed letters and documents written, according to which newspaper one read, in Russian (the _States_ ), Yiddish and English (the _Times-Picayune_ ), or German, Russian, and Yiddish (the _Item_ ). At any rate, it was a foreign correspondence that sent police into spasms of patriotic suspicion. Rumors of code books and international intrigue immediately ignited and led to accusations that the grocer was, at the very least, involved in propaganda for the Germans. HATCHET MYSTERY MAY LEAD TO SPY NEST trumpeted the _Times-Picayune_. Federal authorities were contacted and informed about the possible threat to national security, and Mooney barred reporters from talking to the Besumers. Mooney took the possibility of espionage seriously. Charges of being a German agent were no joke in the summer of 1918. The United States had entered the war in Europe in April 1917, and fear of German saboteurs led to the Espionage Act of 1917, the Sedition Act of 1918, and a steady stream of propaganda encouraging citizens to be on the lookout for enemy agents. Anti-German hysteria soaked the country during the war years. German Americans everywhere came under suspicion. Books in German were burned, and schools abolished German language instruction. Those suspected of disloyalty could be forced to kiss the flag, beaten, or literally tarred and feathered. In Illinois, persecution spilled over into murder when vigilantes lynched an innocent young German immigrant. Louis Besumer could very well have been both a Pole and a German. That is, Poland had ceased to exist as a state in 1795 when it was divided among Prussia (which would later become Germany), Russia, and Austria. Besumer might well have been an ethnic Pole born into German-controlled territory and, as such, might have spoken German but have no great love for the German Empire. Most Americans in 1918, however, wouldn't have understood these finer details of Polish and German history. Suspicion that Besumer was an enemy agent made his situation precarious indeed. Perhaps, the _Times-Picayune_ suggested, Besumer himself had attacked his wife when they had quarreled over his secret work. The police were determined not to be outsmarted by any of the "spy's" clever tricks: when Besumer sent his wife a bathrobe for use in the hospital, officers ripped out its lining, fruitlessly looking for secret messages. Still, Superintendent Mooney thought he should make an effort to find Mrs. Besumer's mulatto, and so, the day after the attack he ordered the arrest of Lewis Oubichon, a light-skinned, sometime employee of Besumer's. Oubichon couldn't satisfactorily account for his movements the night before the attacks, so the police held him for several days. But since no other evidence against him surfaced, he was soon released. Among the newsmen who clamored for the latest information on the investigation was James G. Coulton of the _Times-Picayune_. Only twenty-three, Coulton was already a seasoned crime reporter with a reputation for fairness and accuracy. He also possessed a keen sense of justice, a commitment that went beyond mere platitudes uttered in the pressroom. His sense of right and wrong would lead him to become more involved with the Axeman case than he would have ever suspected. But in June 1918, all he wanted to do—along with every other reporter in New Orleans—was get details about the Besumers. Louis Besumer was released from Charity six days after the attack and the next day found himself in Frank Mooney's office, answering the superintendent's questions. For Mooney, it was an infuriating interview. For much of the two-hour conversation, the eccentric grocer seemed more interested in showing off his erudition than in helping the police find the person who'd attacked him. While he answered some questions, he also wandered off on irrelevant tangents, and Mooney had difficulty keeping him on topic. When Mooney asked him a simple, direct question about one of the languages he claimed to speak, Besumer just smiled and "proceeded to tell about the people who speak it and sketched out their history." Besumer also insisted that he wanted to take an active part in the investigation. One can only speculate what Mooney thought to himself when the grocer solemnly informed him that he was a student of chemistry and criminology, "a born investigator" qualified to help solve the crime. "I am a man who leaves nothing undone," he informed Superintendent Mooney. "I will not rest until the case is cleared up. I want to know who attempted to kill [us]." Louis Besumer was a pompous bore. But was he a criminal? Mooney eventually dragged some relevant information out of the grocer. Before the war he'd been a wholesaler, buying goods in bulk and reselling at a profit. His health suffered from constant travel and endless fattening restaurant meals, and he opened his grocery, he said, to follow his doctor's advice: get into a lower-stress business. Because of his knowledge of the wholesale market, he could buy staples like coffee and sugar in bulk and, to attract customers, undersell his rivals. But, he pointed out, "The other articles in my store I sold no cheaper—some things I sold even higher—than other stores." Besumer also elaborated on his relationship with the woman everyone thought was his wife. One of the first things he had done when he was released from the hospital the day before was announce (perhaps prompted by the rumor that his ex-wife was in town) that Mrs. Besumer was actually "Mrs. Lowe," a divorcee. She was not his wife. He was divorced, he said, and his former wife lived in Cincinnati. The twenty-nine-year-old Harriet Anna Lowe, he maintained, was only a friend with whom he had a brotherly relationship. Still in the hospital, Mrs. Lowe was alarmed when she heard about his confession. By Wednesday, almost a week after the attack, physicians at Charity thought she'd probably survive. By this time Mooney had lifted the ban on press interviews, and she was well enough to talk to reporters. Mrs. Lowe insisted to Jim Coulton that she was indeed Besumer's wife, that they had been married two years ago in New York. Harriet Anna Lowe was a naive, good-natured woman, devout and perhaps not overly intelligent. At one time she'd been attractive, but the blows to her head had permanently disfigured her. Her facial nerves had been so badly damaged that she couldn't move one side of her face; one eye stared out sightlessly, the other "twitche[d] uncontrollably." Her head was swathed in bandages; wounds on her arms and chest were also bandaged. _Harriet Anna Lowe._ Mrs. Lowe readily admitted that her mind and her memory hadn't been clear since the attack, but she told Jim Coulton, "I feel sure he and I are married," adding plaintively, "If we are not, then I don't know what I'll do." Undoubtedly, Jim Coulton couldn't help but feel sorry for her. As the opiates and shock wore off, the sort of incoherent rambling Captain Capo had encountered when he first interviewed Mrs. Lowe dissipated. Her answers to police and reporters' questions were perfectly lucid. She retracted her story about being attacked by a mulatto but also emphatically denied that Louis Besumer had been the one to take an axe to her. As her mind cleared, Mrs. Lowe remembered more and more details from the night of the attack. After ten days in the hospital, she could remember closing up the store about 7:30 PM. Going from the grocery into the back rooms, she saw Besumer sitting in front of the safe, counting the day's takings. She turned into the tiny dining room, carved out of the hallway, and the last thing she could remember was the smell of prunes simmering on the stove. Then everything went blank. A hazy memory of a black-haired man towering over her in the early dawn light had burned itself into her consciousness, followed by the recollection of waking up in a pool of blood on the porch. She got a glimpse of a man's shoes moving around her. The next thing she knew she was in the hospital. Neither Mrs. Lowe nor Besumer could identify a possible suspect. Besumer could only suggest business rivals angry because he undersold them. At the same time, he tried to convince Mooney that the motive was robbery, arguing that a brass lock on a trunk in his room showed signs of an attempt to break it open. The federal government was also interested in the immigrant grocer and suspected spy. Two days after his interview with Superintendent Mooney, he was picked up by agents from the Department of Justice, the forerunner of the FBI, and interrogated about the letters discovered by the police. Whatever they got out of him, whatever arcane erudition he inflicted on them, wasn't enough to hold him, and after several hours Besumer again went home to his grocery. Mooney's investigation suffered a setback a week after the attacks when he had to reprimand two of his detectives working on the case. They'd been assigned to Besumer, ostensibly to help him with his own investigation but more likely to keep an eye on him. But one day during working hours, they'd skipped off to a dancing pavilion in Milneburg, a popular resort on Lake Pontchartrain, where they'd become embroiled in a brawl. Mooney not only gave them a dressing down, but also demoted both to uniformed patrolman, an almost unheard of action in the New Orleans Police Department but part of his campaign to impose stricter discipline on the police force. Many Orleanians speculated that the attack on Louis Besumer and Harriet Anna Lowe was connected to the Maggio murders, or even the whole previous series of axe and hatchet attacks. Newspapermen frequently mentioned the earlier attacks, even as mistakes in their articles showed that their memory of the details was a little foggy. Coming on the heels of the grisly slaying of Joseph and Catherine Maggio, the Besumer attack was naturally suspected of having some connection with the "fiend" who'd struck previously. Initially, Superintendent Mooney had believed that the couple had been attacked by the Maggios' killer. As he mulled over the evidence, however, doubts began nagging at him. All the other victims were Italians. The victims in the Besumer attack were an Eastern European immigrant and an Irish American. Did it matter? Aside from the question of ethnicity, however, this latest attack differed from earlier ones in important details. Most of the other attacks had occurred in sparsely populated parts of the city. But the Besumer grocery, only a few blocks from Esplanade Avenue, was in a heavily populated neighborhood. And the other attacks took place between 1 and 3 AM, the stillest time of the night, usually on moonless nights. The Besumer attack had come after sunup, between 6 and 7 AM (after a night during which the moon was out all night), just as the streets were beginning to stir and when the intruder ran a greater risk of being spotted. Yet no one had seen any strangers enter or leave the premises. The Crutti, Rissetto, Davi, Andollina, and Maggio crimes all involved a break-in, and the victims were attacked while they slept. But Mooney had no evidence that anyone forced his way into the Besumer residence. And Besumer and Mrs. Lowe had not been attacked in their beds. The blood evidence made clear that Mrs. Lowe had been attacked on the outside gallery: Not only did the quantities of blood pooled on the porch testify to this, but it was the only way to account for the blood spray pattern on the outside door. And the blood-smeared gallery door handle indicated that a bloody hand had grasped it and pulled it open after the attack. Later, Besumer would claim that the back doors and windows had been left open because it was so hot, but shortly after the crime, the _Times-Picayune_ reported that he assured the police that the back door had been locked when he went to bed. Whoever had opened the door had done it from the inside. Mooney wasn't sure where Besumer had been assaulted. Certainly not in his bed. It wasn't bloody enough. The only blood in the room were slight traces staining his pillow. Where was the blood spatter on the floor? The deep gash Besumer sustained would have bled profusely, and if he had been in his bed as he claimed, some of this blood should have ended up on the bedroom floor when he got up to answer John Zanca's knock. Moreover, Zanca had said that when Besumer had called to him, his voice had sounded as if it were coming from a distance, perhaps Mrs. Lowe's room, or from out on the porch, not his own bedroom as he had claimed. And why was Besumer so slightly wounded compared to Mrs. Lowe? While he might have sustained a skull fracture, he'd only been hit once: he only had the one gash over his eye. But someone had made a real effort to smash her skull. In most previous attacks the full ire of the attacker had been aimed at the man rather than the woman. Besides, Mooney found Besumer just plain shifty. The grocer had misled everyone about his relationship with Mrs. Lowe. Details of his story about living in Jacksonville hadn't checked out. He admitted to using aliases. No, Louis Besumer, or whatever his name was, could not be dismissed as a suspect. Could Besumer have been the assailant? Perhaps no one had seen a stranger in the vicinity that morning because no stranger had been in the house. Perhaps Mrs. Lowe struck Besumer first, and he, wounded and enraged, grabbed the axe from her and chased her out onto the porch. If both were guilty, both had reason to keep silent about what had really happened. That would explain why Besumer didn't want the police or ambulance called. Mrs. Lowe told Mooney that Besumer had read about the Maggio case. Maybe he hoped that the attack in his home and grocery would be attributed to the same murderous fiend. Mooney was convinced Mrs. Lowe wasn't telling him everything she knew, and the circumstantial evidence he had wasn't enough to make an arrest. The superintendent needed an eyewitness. If Besumer was to be charged, the woman's testimony would be crucial. Mrs. Lowe, however, refused to cooperate. So Mooney pursued other leads and bided his time. On the fourteenth of July, as the city celebrated Bastille Day with a special fervor, given the American and French young men dying side by side in the war raging abroad, Louis Besumer arrived at Charity Hospital to take Mrs. Lowe home. After two weeks in the hospital, the doctors were ready to discharge her. Besumer had reopened the People's Cash Store, and business, he boasted, was good. _Let us go back to the way things were_ , he offered Mrs. Lowe. She had misgivings. By now she had conceded that they weren't married, either because her memory had cleared or because she'd decided it was useless to continue to pretend. But she had nowhere else to go. She had no family—other than an abusive ex-husband. Her mother was dead and her father and sister were in Ireland. Where else was she to go other than the little yellow grocery at Laharpe and North Dorgenois Streets? The blows had left Mrs. Lowe mangled mentally as well as physically. In the days after returning home she showed classic signs of generalized anxiety and depression: she was shaky; she couldn't concentrate; she was indecisive. She'd give orders to Mrs. Sacriste, who'd been hired to nurse her and help with the cooking, and then countermand them. She spent hours each day on her knees obsessively praying, and having visions of Jesus Christ. Mrs. Sacriste worried about Mrs. Lowe's sanity. Outwardly, relations between Besumer and Mrs. Lowe remained cordial, but something happened in the month after she left the hospital, something that made her fear Besumer. Maybe he threatened her. Perhaps as more and more of her memories fought their way to the surface and became clearer and clearer, she became more and more afraid of him. Mooney didn't forget about Mrs. Lowe. He felt sure there was something she could tell him to help him close the case. He dropped by the People's Cash Store regularly to talk to her, taking Assistant DA Ben Daly with him. His efforts would eventually pay off. But not before another attack claimed the headlines. Twenty-eight-year-old Mary Schneider was tired as she crawled into bed on the evening of August 4, 1918. Caring for three small children and expecting another any day was an exhausting business. Her husband Edward worked the night shift at the Southern Pacific wharf, so she was alone at their home at 1320 Elmira Street (now Gallier Street), in the Saint Claude neighborhood on the eastern edge of the city, with her children safely tucked into the other room, when she turned down the lamp and went to sleep. Mary woke, days later, confused. This wasn't her bedroom. Where was she? How did she get here? Had the baby come yet? As she fought her way out of the opiate cloud that fogged her brain, Mary dimly remembered that her baby had been born, a girl, named Clara. But why did her head hurt? The Charity nurses summoned her husband Edward, who hurried to her bedside to explain. She'd been attacked in the night, he told her, and had been taken to Charity Hospital. She was in the maternity ward and the baby was safe. But now the police needed her to identify her assailant. Did she see who had hit her? Did she know him? Shocked, Mary could only stammer, "Struck? Oh, no, I was not struck. Who said that anyone assaulted me?" She was even more surprised when Superintendent Mooney showed up at the maternity ward, detectives in tow, anxious to question her. But she could tell him nothing. Gradually, as her mind cleared, she began to understand what had happened. Mary's sister Kate and her husband lived in one side of a double cottage; the Schneiders lived on the other. About two in the morning, Kate heard her sister scream. She shook her husband awake, and they rushed over to Mary's house to find her barely conscious, a nasty gash on her scalp and several of her teeth cracked and broken. Mary managed to say that she didn't know who had done this to her before she passed out. Once the injured woman had been dispatched to Charity Hospital, Superintendent Mooney and Chief of Detectives Long took stock of the crime scene. They could find no evidence that the house had been broken into; none of the doors or windows had been smashed or pried open. But the wardrobe in the Schneiders' bedroom had been ransacked and its contents dumped on the floor. Questioning of Edward Schneider when he arrived home revealed that seven dollars was missing, but a tin box with a wad of cash—over a hundred dollars—hadn't been taken. A broken lamp lying on the floor appeared to be the weapon used to batter Mrs. Schneider. The detectives spotted strands of hair caught in the prongs of the shattered lamp, and spots of oil that had been flung out as it smashed across her face stained the bedspread. It looked like a straightforward burglary: an intruder rummaging through the wardrobe woke Mary Schneider, who might have been sleeping uneasily because of her pregnancy. When she screamed, he grabbed the nearest weapon—the lamp on the mantel—smashed her over the head with it, and fled. But the citizens of New Orleans were primed to imagine more. After the recent Andollina, Maggio, and Besumer attacks, another mysterious attack in the dead of night wasn't going to be easily dismissed. No matter that the evidence indicated that the weapon used against Mrs. Schneider was a lamp, the _Times-Picayune_ was first off the mark: POLICE BELIEVE AX-MAN MAY BE ACTIVE IN CITY. Curious aspects of the case did make the more suspicious wonder if there wasn't a more sinister explanation than burglary gone wrong. An axe was missing, stolen from the Schneiders' shed, and a hatchet belonging to the couple had been dropped in the neighbor's yard. Most strangely of all, a couple of days after the attack, the Schneiders' eight-year-old daughter found one of her mother's skirts, stained with blood, stuffed under a doorstep a few houses away. But there was little evidence to connect this incident with the attacks on the Andollinas or the Maggios. Young Jim Coulton, the _Times-Picayune_ 's crime reporter, was most responsible for connecting the Schneider attack to the Axeman crimes in the public mind. Coulton became convinced early on of the existence of a single attacker, and his stories on Mary Schneider pushed that angle, even as he dutifully reported Superintendent Mooney's belief that the broken lamp had been the weapon. He cited police officers who opined "it probable that Mrs. Schneider was attacked by the hatchet-man who murdered Joseph Maggio and his wife and attempted to kill Louis Besemer and Mrs. Harriet Lowe." And the _Times-Picayune_ 's headlines, POLICE BELIEVE AX-MAN MAY BE ACTIVE IN CITY and VICTIM OF AX-MAN NOW HAPPY MOTHER, contributed to the perception that the Axeman's taste extended to lonely mothers in addition to Italian grocers. Over at the _States_ , Andy Ojeda took a different tack, citing police who believed robbery more likely to have been the motive. Like Coulton, he'd entered journalism young—as a boy really, at seventeen—writing about cotton prices for the New Orleans _Daily News_ , then working at the _Item_ before joining the _New Orleans States_ around 1907. Ojeda interviewed veteran detectives who scoffed at the notion that the attack on Mrs. Schneider was an "Axeman" crime. One—anonymously—told the _States_ , "It is nothing more or less than a case of ordinary burglary, and the robber used some weapon upon Mrs. Schneider only when she either moved in her sleep or attempted to get up." The _Item_ steered a course between the two polarized viewpoints, reporting that some policemen believed the crime the work of a burglar, while others attributed it to "the axe man." But the paper was rather free with its headlines—AXE-MAN'S VICTIM CAN'T RECALL ATTACK, ARMED MEN GUARD SLEEPING FAMILIES FROM AXE-MAN, and AX-VICTIM TO BE TAKEN TO HER HOME. No wonder many Orleanians assumed that Mary Schneider was, without question, another victim of the Axeman. But Mary Schneider's nighttime visitor was unlikely to have been the Axeman. He targeted couples, usually injuring the man first and most severely. What's more, all his previous victims had been Italian grocers. Mary Schneider was neither male, nor Italian, nor a grocer. Her house was in the middle of the block, not at a corner, as were the other Axeman attacks. Her assailant hadn't used an axe or cleaver. The missing axe, misplaced hatchet, and bloodstained skirt were curious, but several detectives offered the explanation that they were a ruse to make the incident look like an Axeman attack. But the headlines were enough to set off a minor wave of panic in the city, especially in neighborhoods near the crime scene, and most especially among Italians. Memories of the gruesome butchering of the Maggios were still fresh, and men too afraid to sleep stood guard at night over their sleeping families, loaded pistols or shotguns at the ready. Superintendent Mooney hadn't made much headway in solving the Axeman crimes, but to calm fears, he sent regular police patrols through sparsely populated areas on the fringes of the city, the kind of neighborhoods the Axeman preferred. The killer took note. # ≡ 8 ≡ # Axeman Hysteria EXPERTS ARGUE THAT SOME serial killers carefully watch the news, enjoying the coverage of their crimes, relishing their notoriety, savoring the fear they inspire, and thrilling in the frisson of outwitting the police. How did the fiend react when he heard of another's attack attributed to him? Was he angry? Or flattered? Did he damn the stupidity of the police who could not catch him and the press who could not accurately identify him? Such thoughts might have led the Axeman to slip up to 2336 Gravier Street during the night of August 10. In the empty lot next to the house, he looked around and picked up two boards, which he leaned against the fence. Then, he scrambled up and over into the backyard. In the shed behind the house, he found what he needed: the family's two axes. He selected the short-handled one and turned toward the house. A window at the back was open, and it was easy enough to crawl inside, axe in hand. Fifteen-year-old Pauline Bruno lay in bed, unable to sleep, preoccupied with thoughts of the Axeman. From the papers, from the whispers of neighbors and customers and friends at work, she was painfully aware that the killer had a penchant for grocers. And Italians. She turned over fitfully in bed, trying not to disturb her sleeping thirteen-year-old sister. What would keep the Axeman away from them? She'd been born in Louisiana, but her mother and her thirty-one-year-old uncle Joe Romano had come from Italy. They all lived packed into the dingy double cottage at the corner of Tonti and Gravier. Pauline and her sister Mary were in one room, and Joe slept next door. Pauline's mother Lilly Bruno lived with her widowed daughter Rosie on the other side. Together they managed to support themselves. Pauline worked in a candy factory. Rosie ran the little grocery in the front part of the house, trying to sell enough staples—beans, rice, and cigarettes—to bring in a little income. Uncle Joe was a barber, pulling in a much-needed fifteen dollars a week at the shop where he worked over on Canal Street. They were poor but not desperately so. The Bruno home was in a crowded working-class neighborhood half a dozen blocks from Canal Street, not one of the less densely populated areas the Axeman seemed to prefer. But that thought gave Pauline little comfort. Her immigrant family was Italian, and they ran a grocery. No wonder she didn't sleep well. Pauline finally nodded off into an uneasy slumber, but about 3 AM she awoke with a start. What was that? Some commotion in Uncle Joe's room? Then she heard a groan. Sitting up in bed, she looked up to see a man, a stranger, standing in the doorway. She screamed, waking Mary. Now both girls shrieked in utter terror. The figure in the door vanished. As if in answer to their cries, Uncle Joe staggered out of his room, holding his bloodied head. In a stupor he sank into a chair in the small parlor next to their room. "Something has happened," he murmured to the girls. "My head hurts. Call for an ambulance." Then he slumped over in a faint. The family was too poor to have a phone, and it was a neighboring grocer who rang Charity Hospital for an ambulance. By the time it arrived Joseph Romano had regained consciousness, and escorted by the Charity interns, the bloodstained barber made his own way out of the house to the waiting vehicle. Superintendent Mooney and his men were on the scene within the hour, followed by the usual swarm of newspapermen. As grimly determined patrolmen spread out through the neighborhood searching for the attacker, and the Bertillon operator began fingerprinting the open window, Mooney and his detectives traced the intruder's path through the house. They noted the vacant lot next door and saw the open window in the rear of the house. In Joe Romano's bedroom, next to the kitchen, the bed and pillow were covered with blood. A bloodstained axe lay on the floor next to Romano's bed. The assailant had fled through the kitchen, and on his way out he'd managed to rifle a pair of his victim's trousers before dropping it on the kitchen floor. Romano's pocketbook—the early twentieth-century version of a wallet—was missing. _The grocery at the corner of Tonti and Gravier and the scene of Joseph Romano's murder._ Though shaken, Pauline was able to provide a description of the attacker: tall and heavily built, in a dark suit and black hat. She thought he was white, although she couldn't be positive. Even though he was a big man, when he fled, "He was awfully light on his feet," she said, adding, "I think he had rubber soles." After an hour or so, word reached the family that Uncle Joe had died. At Charity Hospital, house surgeon Jerome Landry had immediately taken the badly injured man into the operating amphitheater, giving him just enough time to tell the doctor that he hadn't seen his assailant. Dr. Landry later told the police that Romano had been hit with a very sharp weapon: "The skull was not shattered but cut clean to the brain." Coroner O'Hara's report noted that he'd died of one blow to the head that had left a three-and-a-half-inch fracture in his skull. Joseph Romano survived less than two hours after being attacked. This had the hallmarks of an Axeman murder. The Bruno-Romano household were the poorest victims yet, with a tiny little store. But the murderer's target remained the same. Romano was a barber, but he was the only male in a house with a grocery in the front room. Perhaps the Axeman had assumed that he ran the store. This attack differed from the Axeman's most recent exploits only in that he'd ventured into a more heavily populated neighborhood and that he'd found an open window and not needed to cut out a door panel. Perhaps Mooney's police patrols had driven him from his preferred hunting grounds into more dangerous territory. Word of this most recent murder set off a full-scale panic among the population—and not only the Italians. Men who hadn't already done so rushed to arm themselves. Some began sleeping by day, guarding their families at night. Others banded together and shared the night watches, taking turns sleeping and standing sentinel. Husbands and fathers clutched buckshot-loaded shotguns, or nervously gripped Colt revolvers, jumping at every sound. A few fingered their triggers hopefully, longing for the murderer to choose their door that night. _Cartoon in the_ Item _depicting Orleanians too terrified to sleep._ In the meantime, Superintendent Mooney's investigations brought to light recent incidents no one had bothered to report earlier. Two weeks previously, Joseph Le Boeuf, a grocer only a block away from the Bruno-Romano residence, had chased away an intruder who tried to break in through his back door in the middle of the night. When he bolted, the intruder dropped an axe. A few blocks farther away, on Cleveland Avenue, Arthur Recknagle reported that he, too, had driven off a prowler who'd left behind an axe. The police also identified three attempts to break into homes in the Gravier and Tonti neighborhoods just before Joseph Romano's murder that, in retrospect, they attributed to the Axeman. And Mooney's investigators discovered that the night after Joseph Romano's murder, a burglar had attempted to break into Al Durand's saloon at the corner of South Salcedo and Canal, not far away. He was in the process of chiseling a panel off a side door when something or someone scared him; he dropped his tools and ran. Durand woke the next morning to find an axe, a screwdriver, and a .38 caliber cartridge outside his door. The realization of what may have passed him by so unnerved the saloonkeeper that he refused to report it for fear the shock of it would kill his ill mother. A friend ultimately informed the police of the incident. Frank Mooney was incensed. How could his detectives track down the murderer without all possible public cooperation? How many other incidents did he not know about? He immediately issued a statement demanding that the public alert the police to any potential clue that might lead to the murderer: any attempted or suspected break-in, any suspicious person skulking about the street late at night, anything that struck them as unusual or irregular. Mooney didn't temper his words: "I believe it is criminal for citizens to withhold such cases from the police," he declared. "To withhold information means to assist the axe man in his murderous work." The police needed the help of all New Orleans citizens, if they had any hope of catching "this blood-mad creature." But nerves were so on edge that some people didn't need the chief's encouragement. Some went Axeman hunting on their own. The very day after Romano's murder, two overzealous public servants—a clerk in the criminal court and a deputy sheriff—crashed into the little shanty of Albert Alexander—a black man who, it was rumored, had been spotted chasing a young black girl with an axe. They dragged their quarry to the nearest police precinct station. Superintendent Mooney himself interviewed Alexander, concluded he was wholly innocent of Romano's murder, and turned him loose. This wasn't the only "Axeman" spotting in the days immediately after Romano's death. The evening after the barber's murder, around midnight, a woman only four blocks from the Romano house spied a man meeting the description of the killer heading for her stable. Terrified, she raised the alarm, and within minutes armed, agitated neighbors poured out of their houses and commenced a manhunt. The posse gave chase, tracking the prowler over fences and through backyards before he melted into the warm August night air. Other sightings followed. Early on August 16, six days after the Romano murder, three young men out on the town reported a man lurking near an Italian grocery. Before long, Chief of Detectives George Long had a dozen men on the scene searching the block, joined by armed residents of the neighborhood. Other incidents followed, all carefully checked out by Mooney's men, all coming to nothing. Some wilder stories would have been amusing if they hadn't been fueled by desperate fear: the "Axeman" suspect reportedly disguised as a woman, only to turn out to be, in fact, a "badly frightened negro woman," or the police squad that responded to a hysterical caller who claimed to have seen the Axeman only to find a drunk chopping up his neighbor's porch steps for firewood. A pattern targeting Italian shopkeepers did continue in the immediate aftermath of this latest killing, not a pattern of murder but of theft. Just over a week after Joseph Romano's funeral three Italian grocers were robbed in rapid succession by methods that echoed those of the Axeman. In the early morning of August 19, an intruder used an ice hatchet and pick to break into the living quarters of grocer Toney Caronna and his family at the corner of South Claiborne and General Taylor Streets. (He'd first picked up the family axe but decided against it, perhaps because it was too heavy.) He was in the process of stealing cash and a revolver from the bedroom, when Mrs. Caronna woke, saw him, and screamed. He ran. On the very next night, a burglar chiseled a panel out of the door of Frank Guarisco's grocery and saloon at the corner of Tchoupitoulas and Calliope and cleaned out the register, all without waking the family. The next evening—the third in a row—someone broke into Paul Lobella's drygoods store at the corner of Zimple and Cherokee using an iron bar, took eighty cents, and fled. The next morning an axe was discovered in the alley next to the house. These break-ins were scattered all over the city, from Carrollton to downtown. Were any of these thieves the Axeman? Had he intended to attack the residents but was scared away in each instance? Or was murder for him just a sometime thing? Other criminals may have adopted the Axeman's—rather successful—methods for their own use. Times-Picayune _map showing Axeman attacks and panel burglaries._ This was the beginning of a series of nighttime, Axeman-like burglaries that targeted Italians and non-Italians alike. For the next month at least, more than a dozen stores—not just saloons and groceries but furniture stores and stores that sold silverware, ladies' clothes, cologne, and furs—were robbed using what became known as the "Axeman method": the removal of a panel from a rear door. Frank Mooney, and probably much of the city, suspected that at least some of these incidents were Axeman jobs. They did nothing to calm the fears of the city. Frank Mooney was getting desperate. He was facing, as the _Item_ put it, "scathing criticism" from the public for his inability to catch the killer, and his "exhaustive investigation" of the Romano murder was getting exactly nowhere. He continued to reassure the public that it was only a matter of time until the Axeman was caught; he couldn't very well admit that the murderer had the run of the city. But how was he going to stop him? Mooney sent patrols out to cover the city as well as he could with his small police force, the downtown neighborhoods where anxiety had been so high lately, and especially the "thinly settled sections" the Axeman seemed to prefer. The killer had shifted from the edge of town to downtown, and Mooney had a hunch that all the heat might drive him elsewhere. How right he would turn out to be. Just days after Joseph Romano's murder, newspaperman Jim Coulton sought out former detective John Dantonio to ask his opinion on the recent attacks. He'd no doubt heard that Dantonio, the expert in Italian crime, had been involved with the Cleaver attacks years before. For his part, retired for over a year and now in poor health, Dantonio was probably delighted that his advice was sought. Fifty years before the term _profiler_ came into use, Dantonio provided a remarkably perceptive analysis of the killer. "This is very probably the man we tried to get ten years ago," he told the young reporter. "The murderer is likely a Jekyll and Hyde personality, like Jack the Ripper. A criminal of the dual personality type may be a respectable, law abiding citizen when his normal self. Then suddenly the impulse to kill comes upon him and he must obey it." Perhaps, the detective speculated, the Axeman had been dormant for several years but now his dark side had reasserted itself. No longer a public servant, Dantonio didn't feel an obligation to reassure the city. He warned that the police were unlikely to bring the murderer to justice, admitting that these killers, who planned their attacks very carefully, were hard to catch. For Dantonio, the key to all the attacks was the victims' residences: "The homes of all the victims look alike. The Axeman apparently picks them out because of the resemblance which impresses him." Although Dantonio realized that most of those targeted were Italians, he doesn't appear to have considered this key to selecting the victims; he believed that the Axeman had attacked Mary Schneider, although he offered no opinion on the Besumer case. Superintendent Mooney couldn't have been pleased by Dantonio's admission that the killer would be all but impossible to catch. In his efforts to catch the murderer, Mooney had immersed himself in the science of "fiends." He'd read up on current criminology and talked to local experts, hoping to find some nugget that would lead him to the killer. His own views came to mirror Dantonio's so closely that it's difficult not to believe that he also consulted the retired detective. Gradually the superintendent's views took shape: Robbery wasn't the Axeman's motive; he opportunistically grabbed small amounts of cash to throw off investigators. The crimes were the work of a madman, an insane killer in the grip of a depraved lust for blood, one who could nevertheless look and act normally until his obsession overcame him. He was a sadist, and likely a narcotics addict. Mooney sent two detectives to Joe Romano's wake in the event that the killer showed up to savor his kill. But Mooney doesn't seem to have been persuaded by all of Dantonio's theory. The retired detective surmised—probably correctly—that the Cleaver of 1910–1911 was also the Axeman. Witnesses from that series of attacks—Harriet Crutti and Mary Davi—described the Cleaver as white. But for reasons unclear, a black man had become Mooney's favorite suspect: Charles Anderson, who'd served two prison terms for robbery. If Mooney was looking for a black man for the crimes of 1917–1918, he wasn't convinced his killer was the killer of earlier years. The superintendent's efforts didn't end with consulting criminologists. Mooney also conferred with police in other cities and turned to private security agencies. Rumor said that he had even involved the famous Pinkerton detective agency. He wanted the public to be confident that the police were doing all they could to catch the killer, assuring the newspapermen in a steely tone: "Take this for the gospel: We're going to get him yet! I'm doing everything in human power to run down this murderous maniac. . . . We're going to get him." Not everyone was terrified of the Axeman. Those who didn't feel threatened by the killer could treat him as a joke. On August 23, a large ad appeared on page 5 of the _Times-Picayune_. "Attention Mr. Mooney and All Citizens of New Orleans," it read. "The Axeman Will Appear in This City on Saturday, August 24th. . . . He will ruthlessly use the 'Piggly Wiggly' axe in cutting off the heads of all High Priced Groceries. His weapon is wonderful and his system Unique. Don't miss seeing him." The Piggly Wiggly grocery store chain had introduced self-service in 1916. No need for a nice Mr. Jordano to wait on customers. A housewife could now help herself to a can of tomato sauce or a jar of olive oil. The popularity of self-service grocery stores would eventually mean the growth of the supermarket and the end of the corner groceries. This would take some time, but for now Piggly Wiggly made a tasteless joke at the expense of the neighborhood grocers it would replace. Superintendent Mooney's patience was rewarded. A month after she'd left the hospital, looking over her shoulder to make sure she wasn't seen, Mrs. Lowe stepped out of the store onto Laharpe Street and flagged down a passing policeman. She told him that she needed to see Superintendent Mooney at once. _But Mr. Besumer must not know_ , she insisted. This was what Mooney had been waiting for. Anxious not to spook his witness, he arranged for the grocer to be lured away for a couple of hours. Mooney, Assistant DA Ben Daly, two representatives of the Department of Justice, and a handful of policemen drove out to the grocery. _The cloud has lifted from my mind_ , she told them when they arrived. _I now know that it was Besumer who attacked me_. And she had something to show them. Mrs. Lowe led them to a trunk she said belonged to Besumer. Opening it, she showed them a cleverly concealed secret compartment in which, she claimed, he kept mysterious blueprints. Whatever documents the grocer had stashed there would later prove not to be incriminating, but a search of his clothes revealed a honeycomb of secret pouches. Later that day, over the whir of an electric fan, Mrs. Lowe told her story to Mooney and the assembled throng of officials and newspapermen that crowded into his office. Shifting nervously in her chair like someone on the verge of a panic attack, her black hair plastered to her face with sweat, and glancing fearfully at the doorway from time to time, she said that she finally remembered what had happened that night. "I met Mr. Besumer in Jacksonville, Florida," she began, with the inflection of the second-generation Irish, "through a newspaper advertisement. I had advertised for a position as housekeeper." She knew that Besumer was Jewish, but other than that, Mrs. Lowe didn't know much about him. He made frequent trips on mysterious business that he refused to discuss. She knew that he used a number of aliases and wasn't even sure "Besumer" was his real name. She had told them earlier that he didn't like to be questioned and he had such a nasty temper that she'd learned not to ask about his business. Mrs. Lowe wasn't really the curious type anyway. She had a roof over her head and food to eat, so even if Besumer was bad-tempered and peculiar, she didn't worry too much about it. Then, three months before the attack, Besumer shaved off his beard and moved to New Orleans, taking Mrs. Lowe with him and telling her she'd now be going by the name "Besumer." Around 7:30 PM, on the night of the assault, Besumer had just closed the store and was going over some blueprints when Mrs. Lowe entered the grocery and asked for money he owed her. She hadn't been paid for months. But he shot her such a nasty look that she, knowing his temper, backed away and turned to leave. She heard him say, "You'll get your money and more too" and then felt what she described as "a thousand bricks . . . hit[ting] her on the head." She passed out and when she came to, heard him say something strange: "Annie, you are going to make a fire in the ocean for me." She remembered being dragged through the house and undressed on the bed but then blacked out again until she woke up covered in blood. When she got out of the hospital, Mrs. Lowe said, Besumer offered to marry her, at the same time that he warned her against talking to Superintendent Mooney. She remained in a fog, confused and unsure about the events that had put her in the hospital, but four nights ago the cloud lifted and it all suddenly began to come back to her: "As I lay in bed," she said, "the street light shown through my door and Mr. Besumer appeared in the doorway. Then my memory started to return by degrees. I felt the horror again and became frightened." _How did Besumer get hurt?_ Mooney wanted to know. She said she didn't know. Perhaps he hit himself, one of the other policemen suggested, in order to make his story more believable. Another suggested that Mrs. Lowe's wages had nothing to do with the attack; perhaps he'd tried to kill her because she'd caught him with the mysterious blueprints. Dr. Metz, the city chemist, had identified bloodstains on the hidden compartment in Besumer's trunk. Maybe she'd seen something he didn't want her to see. Whatever the reason for the assault, Mrs. Lowe was visibly frightened of Louis Besumer. Although she was moved into a house where she'd be safe, her nerves became so frayed that she neared a breakdown. Two days later, she returned to Charity Hospital. Mooney had Besumer arrested. Under interrogation, the grocer was as talkative and glib as ever. He laughed at the accusation that he was a secret agent. _That's ridiculous_ , Besumer said. _I was born in Poland_. _I have no love for the Germans_. _I have two sons in the army!_ He could explain everything. The secret compartment in his trunk and secret pockets in his clothes were for hiding important papers and cash. Likewise, he said, he sometimes found aliases useful in his business. Or sometimes in his pleasure, as he implied when he claimed to have switched to the name "Harrison" at one time to avoid a woman who, he added confidentially, "gave me some trouble." About the night of the attack, his answers remained the same: He went to bed at the usual time and knew nothing until the bread man pounded on his door. He had nothing bad to say about Mrs. Lowe, only that she was of a "peculiar disposition." He repeatedly denied that he had struck her. No amount of questioning could dent his claim of innocence. Besumer was charged with attempted murder and clapped into the parish prison. A few days later, he appeared for a bail hearing in the Second City Criminal Court. Superintendent Mooney and Agent W. C. Stillson of the Department of Justice wanted to make certain that Besumer remained in jail and within their reach. With his habit of aliases, there was too much risk he'd leave New Orleans, change his name again, and vanish. Attempted murder was a bondable offense, so after he made bail on the murder charge, Mooney had him immediately rearrested on a D&S charge—as a "dangerous and suspicious character." Still on the edge of a nervous breakdown, wracked with anxiety, Mrs. Lowe was readmitted to the hospital five weeks after being released. And nerves were not her only problem: her head wound was not healing properly. After she'd been hospitalized for about two weeks, doctors decided that a piece of bone in her skull needed to be removed to relieve the pressure on her brain. The operation itself was successful, but she contracted pneumococcal meningitis, an infection of the lining of the brain and spinal column, most likely caused by bacteria introduced as a result of her fractured skull. By the end of the first week of September, Mrs. Lowe was failing fast. A priest gave her the last rites. She asked to see Superintendent Mooney one final time. As she lay in her hospital cot, knowing that she was dying, she again accused Louis Besumer: I feel that I am going to die. The statement that I make to you that Louis Besumer struck me with an axe is correct. I am not mistaken. I don't know what he struck me with in the grocery, but he struck me with an axe in the hall. He was in a nude form. He had no clothes on at all. I gave him no cause to strike me. I asked him for my money. I saw him with the axe in his hand before he struck me. I saw his arm coming down and felt the stunning blow. I can't be mistaken about this. I would not tell a story about him. I don't know why he struck me. She rallied briefly, then died suddenly ten days later. Besumer, about to be released from his twenty-day sentence, was immediately charged with murder. Louis Besumer—or whatever his real name was—almost certainly was Mrs. Lowe's murderer. The lack of evidence of an intruder doesn't leave much room for any other conclusion. His story of waking up dripping blood, with no evidence of blood on the bed or the floor, was transparently a lie. Why, then, did Louis Besumer try to kill Harriet Anna Lowe? In the weeks after his arrest the New Orleans papers reported that investigators believed Besumer attacked Lowe when she caught him with secret documents. The _New Orleans_ _States_ reported the discovery of spots of blood on his secret drawer and evidence that someone had tried to remove the blood by washing and then by scraping it off. Nothing ever came of the charge that Besumer was a German agent, so his secret drawer probably held sensitive business or personal papers, rather than anything related to national security. But investigators and newspapermen thought national security a more exciting reason for attempted murder than the usual squalid lovers' quarrel. Whatever his motive, Louis Besumer, not some mysterious killer, took an axe to Harriet Anna Lowe. Her description of his highly inflated self-image and easily provoked, violent temper fit the profile of an abusive husband or partner. Writing over thirty years later, Robert Tallant, who probably relied on the accounts of reporters or policemen who'd been around at the time, reported that neighbors had heard Besumer and Lowe bicker—violently—about money and his jealousy. It's easy to imagine that they got into a fight about who knows what—money, marriage, jealousy, his secret dealings, whatever. Perhaps he shoved her. Maybe he slapped her. Possibly, he did worse. To defend herself, Mrs. Lowe grabbed the first handy weapon, a knife or a meat cleaver, and struck Besumer in the head, a hard blow that sliced into his scalp above his right eye, a wound that could be mistaken for an axe wound. (Maybe she was even strong enough to wield the axe, although it's easier to believe she used a smaller and lighter weapon.) Besumer staggered under the blow, then recovered. He reached up to touch his forehead, then stared in disbelief at the blood on this hand. Realizing what she did, Mrs. Lowe froze in terror. She knew he would make her pay for this mistake. She turned to flee; she had to get away from him. Enraged, he chased her out onto the porch where he grabbed the axe and hit her, cracking her skull. Perhaps afterward he cooled off quickly when he realized how badly she was wounded. What was he to do? He'd read about the Maggio attacks. He didn't have much time to formulate a plan, and he was still reeling from his own head wound, but he took a sheet (to protect himself from the blood), wrapped her in it, picked her up, and carried her to her bed. When John Zanca, driver of the bread wagon, showed up on his doorstep, Besumer wasn't keen for him to call the police or an ambulance, but there was nothing he could do. The grocer didn't have many choices at this point, and the best he could do was keep his mouth shut, claim to have slept through the whole incident, and feign bewilderment that Mrs. Lowe was found half dead in her bed. And fervently hope everyone attributed this attack to the Axeman. Mrs. Lowe's story wasn't exactly airtight either. When she first claimed that a mulatto had attacked her, she'd been pumped full of opiates. Later, because of shock or a genuine loss of memory, she was likely telling the truth when she said she couldn't remember who attacked her. However, she may well have suspected Besumer. But when questioned by police, she denied that such a thing could occur, like many abused women, then and now. When she finally implicated Besumer, her story was, understandably because of her head injury, disjointed and incomplete, with scraps of memory interspersed with periods of blackness. She could recall being dragged into the hall, lying bloody on the porch, and being carried to her bed; a glimpse of a man's feet, clad in heavy black shoes, flickered off and on in her memory. But she couldn't fit all the fragments together into a coherent narrative. Her recollections didn't always fit the physical evidence. Lowe said that Besumer attacked her with the axe in the hall, but the blood on the porch told a different story. She said that the attack occurred on the evening of June 25, but the fresh blood evidence indicated that it had taken place on the morning of June 26, shortly before they were discovered. These inconsistencies, strangely enough, make her story more believable. Surely deliberate lies would have been more coherent. Mrs. Lowe also denied any knowledge about how Besumer came by his injury. One doesn't have to believe her denial to believe that the rest of her story was accurate—or at least as accurate as her fragile memory would allow. The detail that Besumer was nude raises some interesting possibilities. One might be tempted to dismiss it as the result of delirium brought on by meningitis in her final days, but the accusation was reported by the _New Orleans States_ almost a month before her death, just after she defected from Besumer to Mooney, and before she entered the hospital for further treatment. Perhaps before Zanca's arrival, Besumer stripped himself of his bloody clothes and disposed of them. Since the police didn't always make thorough searches of crime scenes, they might not have found them. Why else would she include such an odd—and to an early twentieth-century audience—disturbing detail in a deliberate lie? Readers also shouldn't dismiss the idea that Lowe's memories were the result of weeks of relentless badgering by Superintendent Mooney. Memories are not concrete realities that just need to be accessed. They are a _creation_ of the mind and are influenced by a good deal more than what actually happened in the past. Memories can be manipulated and distorted, even created, by suggestive questioning. If Mooney had become convinced that Besumer was the attacker, his relentless questioning of Mrs. Lowe might well have shaped her recollection of events. This possibility, however real, doesn't change the physical evidence, and the conclusions to be drawn from it: there was no intruder, and the attack differed in significant ways from the modus operandi of the Axeman. Then, as now, women were much more likely to be killed by an intimate than by a stranger. Harriet Anna Lowe died on September 16, 1918, and Louis Besumer was charged with her murder the same day. Conveniently, he was still in jail, still serving a "dangerous and suspicious" sentence. But Joseph Romano's death enormously complicated Mooney's case against Besumer. If there was an axe-wielding fiend in New Orleans, and if said fiend had killed both Lowe and Romano, Louis Besumer manifestly couldn't be guilty. But despite firmly believing that there was an Axeman at work in the city, Mooney was also convinced that he had a solid case against the Polish grocer. Usually, Besumer would have gone to trial within months of being indicted, but his prosecution ground to a halt when the Spanish flu hit in the fall of 1918. Before it had run its course, Spanish influenza killed 30 million people worldwide in less than a year. The disease arrived in New Orleans on the oil tanker _Harold Walker_ in the third week of September, just as Besumer was arrested for murder. Health authorities didn't know exactly how many cases New Orleans had because physicians, overwhelmed fighting the epidemic, reported the numbers only haphazardly. On October 9, the city health officer estimated that there were some 8,000 cases in the city. From October 8 to 20, at least 24,711 new flu cases were reported. Hospitals were quickly swamped as beds filled and more came. The Moose lodge, the Knights of Columbus hall, and the Home for Destitute Boys all became temporary hospitals. City officials fought to bring the epidemic under control. To limit transmission of the disease, city health authorities ordered public places closed—schools, churches, movie theaters, and dance halls. They even threatened to close saloons before new cases of the flu began to decline by the end of October. Under such circumstances, investigation into the Axeman murders almost certainly slowed down. The police who weren't ill with the flu themselves would have been too busy enforcing health regulations to follow up on crime leads. Still, perhaps Mooney's efforts after Joseph Romano's murder had made New Orleans too hot for the killer. The Axeman moved across the river. # ≡ 9 ≡ # The Mysterious Axeman's Jazz GROCER IORLANDO GUAGLIARDO, OF Convent, Louisiana, married in 1883, waiting until he was thirty-two years old and able to support a family. Lillie Billa, his bride, was nineteen. Two baby boys died at birth, both named Frank for Iorlando's father Francisco. But by 1894 the couple was blessed with the birth of Mary, and then in 1899, another girl christened Anna but known to everyone as Lena. Finally, in 1901, the boy named Frank whom Iorlando so badly wanted arrived; Louis followed in 1906. After years in Saint James Parish, Iorlando decided to try his luck elsewhere. He briefly moved his family to Garyville, a small town in Saint John the Baptist Parish, before finally settling forty-five miles down the Mississippi in Gretna, a tiny community across the river from New Orleans. There, sometime between 1909 and 1911, he bought a grocery store—a modest little wooden structure—on the corner of Jefferson and Second Streets. At some point in his mercantile career, Guagliardo took the name Benedict Jordano when he bought a store and decided not only to keep the original owner's name on the business but to take it for himself. _Jordano_ was easier on the American tongue than _Guagliardo_. That is how he was known in Gretna. _Iorlando Jordano._ Years of constant labor and sixteen-hour days took their toll. By 1916, at age sixty-five, Iorlando was an old man. Gray and stooped, his eyesight was failing. His back was stiff with rheumatism and ached badly most days. He was "all crippled up," as his wife described him. Standing in the store all day was now out of the question. Responsibility for the business fell almost entirely on Lillie. Although she was only fifty in 1916, she, too, became unwell. Her doctor had to order her to stop working or, he said, "it would just kill her." Their daughter Lena considered what to do. Her older sister Mary had married and started her own family. Frank and Louis were still in school. As the eldest daughter left at home, running the house fell mostly to her. Now, at seventeen, she was also running the grocery—and it was too much. She told her parents to lease the store; she would get a job. That's how she came to be at Penick & Ford, a mile and a half up the river in Amesville, labeling cans of Brer Rabbit Molasses. The Jordanos leased the building and sold its contents to a young couple who only kept it for a few months before selling the business to Charlie and Rosie Cortimiglia. Charlie Cortimiglia—his given name was Vincenzo, but _Charlie_ was a common Americanization—was another Italian laborer whose goal had been to run his own place. He'd been born in Italy, and by the time he was twenty-five he'd saved enough to go into business for himself. His wife Rosie was a native Louisianan, born in Plaquemine Parish in the extreme southeast of the state. Dark-haired and black-eyed, she was pretty in a young and robust and ruddy sort of way, what the newspapers called "a pronounced Italian type." A merry girl who laughed easily, she'd never been to school, but she could read Italian. Like many girls with few options, she'd wed young, marrying Charlie in 1913 when she was only fifteen. In 1917, she gave birth to a daughter, Mary, who became the particular pet of next-door neighbor Frank Jordano. Frank Jordano had grown to be a hulking bear of a young man with a booming voice, black-haired and pink-cheeked and maybe a little cocky. One man described him as "of a manly bearing," which might have been another way of saying that he was big: at seventeen, he stood over six feet tall and weighed 275 pounds. Frank had the makings of a hard-headed businessman, but he was also a gentle, soft-hearted boy and loved to play with baby Mary every day after work. A photograph of the two of them shows the teenager lovingly cradling the toddler, an expression of delight lighting up his face. Frank was an intelligent boy, but school hadn't interested him, and at fourteen he'd quit to enter the real estate business. Italians tended to prize work over education, and many sons of immigrants traded in school for a job when barely in their teens. Frank sold real estate for a while, and when the land market turned soft, he turned to selling insurance. He never showed much interest in the grocery business. He wouldn't be a small shopkeeper his whole life; Frank was fiercely ambitious and had a gift for salesmanship that would serve him well in the business world. Frank's responsibilities and ambitions had matured him early, and he already had a serious girlfriend, Josie Spera. They'd gotten engaged on Christmas Day in 1918 and planned to marry when Josie's soldier brothers came home from the army. They set March 19, Saint Joseph's Day, as their wedding day. _Frank Jordano holding Mary Cortimiglia._ Meanwhile, the Jordanos decided that they needed their grocery business back. Why, exactly, is unclear. Lena later testified that she was about to lose her factory job, and so she decided to take over the store again. But perhaps she didn't want to furnish Frank with a motive for murder. Iorlando told a reporter that his wife wanted Frank to have the store so that he'd have enough money to get married. At any rate, take it back they did in December 1918, to the annoyance of the Cortimiglias, who weren't at all gracious about it. Charlie and Rosie were doing a good business in this location and dragged their feet about relinquishing the property. When the Jordanos finally served them with an order to vacate, Charlie relocated his store just a few blocks away. But he wasn't willing to permanently give up his spot on Jefferson Street. He bought a vacant lot next to the Cortimiglias and built his own store at the corner of Jefferson and Second Streets, which opened in late February 1919. Some people said that the business, only open a couple of weeks before the attack, was already doing so well that Charlie was cutting into the Jordanos' trade. But if there had been a cooling of the friendship between the two families, there was no cooling in the affection between little Mary and big Frank. After the Jordanos took over their grocery again, it was Lena and her mother—whose health had apparently improved—who ran it. Iorlando couldn't help much. Maybe he went into the store once a week when he felt up to it, but he had to lie down much of the time because of his back. The old man raised chickens, and he had a little garden he worked in when his rheumatism would let him. But mostly he took it easy. On the evening of Saturday, March 8, 1919, Mrs. Jordano closed the store as usual at 10 PM. Shortly afterward, she sat at the kitchen table tallying the day's receipts when a hungry Frank came in the door. "Mama, have you got anything to eat?" he asked. "No," she replied, not looking up from her counting. "It was too busy in the store today for anyone to cook. We ate sandwiches. If you want something to eat, take the key and go to the store to get yourself something." Frank was annoyed. It was hardly worth the bother. "Oh, hell!" he grunted. "If I have to go in the grocery now to get something to eat I won't go, I'm going to bed." But he didn't go to bed immediately. When his parents got into their bed, he went to kiss them both good night. And he asked his mother about a frightening dream that his girlfriend Josie had had, a dream that she was standing in front of a door surrounded by writhing, squirming snakes. _Did it portend evil?_ he wanted to know. Events later that night would suggest that it did. The house was just settling down for the night when, about 11 PM, Iorlando heard his dogs barking furiously. Was someone trying to steal his chickens? He'd better look. His wife was tired after her long day, and they were his chickens anyway. He slowly and painfully got out of bed, fumbled with his spectacles, and hobbled out the kitchen door. He checked on the chickens and his little goat. Everything was in order. He looked around the yard. Nothing out of the ordinary. Reassured, he went back into the house and went to bed. _He slunk down the alleyway separating the two houses, board fences on either side of him. Behind one fence, the dogs sensed a newcomer, someone who didn't belong, and howled. The man wasn't worried. Once he was in the backyard, the wooden fence surrounding it would protect him from view. No one would pay any attention to the barking of a few dogs. He slipped over the fence into the yard. A half-grown puppy—a black and tan mixed breed—looked up from where he slept and glared menacingly at the newcomer. The man ignored the dog; it was too small to worry about. Pulling up a chair he discovered in the backyard, he checked through the bedroom window that his prey was indeed there. Then he looked around and found the axe he needed. Picking it up and pulling out a screwdriver, he crouched in front of the kitchen door. With the blunt side of the axe head, he began—as quietly as he could—hammering the screwdriver tip into the side of one of the lower panels._ Barking dogs woke City Councilman Manny Fink up at ten minutes to three. But once fully awake, he heard something else. Knocking. As if someone was rapping on a door. Then it stopped. A short pause. It started up again. Then it resumed— _tap, tap, tap, tap_. Careful not to wake his wife, Manny got out of bed, grabbed a lantern, and went out into his backyard. He held up the light, scanning the yard. Nothing. He went back inside and through to the front door, turning on the electric houselights as he did so. When he opened the front door and the light from the house spilled outside, the tapping abruptly stopped. Then a dog yelped. As if it had been kicked, Manny thought. _The fiend saw a light flicker on from somewhere across the street. He paused, holding the axe poised in the air over the screwdriver. The small dog took courage and edged menacingly toward the intruder, snarling. The man stood up and aimed a kick at the little fellow, who yelped in pain and fright before retreating, whimpering, toward the far side of the yard._ Manny walked out onto his front gallery and peered across Jefferson Street, first in one direction and then in the other. He could see the Cortimiglia place, about half a square down from him, the front of the property and the fence that surrounded it. Now he heard nothing except the normal nighttime summer sounds. Manny turned and went back into the house. After a few minutes, he heard it start up again— _tap, tap, tap_. What was it? Well, whatever it was, it wasn't on his property. Manny shrugged and went back to bed. _When the light switched off, the fiend waited a few minutes, until he felt it was safe to resume. Patiently, he chiseled at one side of the panel, and then the other, until he could pry it off and toss it to the side. Reaching through the small hole he had created in the door, he reached inside and up and unbolted the door. He paused and took off his boots. In his stocking feet, clutching the axe, he went inside._ Sheriff Louis H. Marrero had a firm grip on Jefferson Parish. He presided over a Regular Democratic machine that had kept him in power for over twenty years and given him a reputation as boss of the "Free State of Jefferson." Although he was a colonel only in the United Confederate Veterans, with his gray goatee and Panama hat, Marrero looked every inch the Confederate colonel, wealthy landowner, and powerful politician. The end of the Civil War had found the eighteen-year-old Marrero a penniless veteran, but marriage to the daughter of a prominent Louisiana sugar and cotton planter improved his fortunes considerably. He entered politics in Jefferson Parish in 1883. First elected sheriff in 1896, Marrero was now in his sixth term and ruled the Free State of Jefferson with an iron fist. He'd used his position to enrich himself considerably. His enemies accused him of corruption, extortion, embezzling tax money from the parish, and illegal gambling. That Marrero's own son Louis Jr. was district attorney of the Twenty-Eighth Judicial District from 1904 until his death in 1916 usually made it easier for him to maneuver around the law. The Good Government League, an antimachine reform movement, attempted to push back against the Marrero regime. In 1912, they tried, unsuccessfully, to impeach the Marreros, father and son, accusing the sheriff of "nonfeasance, malfeasance, extortion in office . . . grave crimes and misdemeanors . . . corruption, favoritism, and oppression in office." One day soon reformers would break the stranglehold the Marrero faction had on Jefferson Parish. Less than a year later, a wave of reform that swept over Louisiana would take Louis Marrero and other machine politicians with it. But in May 1919, that day had not yet come. Sheriff Marrero was out of town on Sunday morning when Hazel Johnson roused Charlie and Rosie Cortimiglia's neighbors with her screams. By the time the Cortimiglias had been loaded into the wagon and sent off to the hospital, Gretna chief of police Peter Leson and Deputy Sheriff Charles Burgbacher had arrived to discover a house crowded with neighbors and curious onlookers. The first thing Leson did was throw the gawkers out of the house and post a guard to make sure no one else got in. There wasn't much else he could do until the coroner arrived. No doubt meaning to be helpful, he and Manny Fink picked up the panel off the kitchen floor and nailed it back onto the door. The need to preserve the crime scene never occurred to them. When Frank returned from notifying Charlie Cortimiglia's relatives in Amesville about midmorning, he put up his horse and buggy, went upstairs to put on the socks he'd abandoned in his hurry earlier that morning, went back outside, and stood on the banquette in front of the grocery looking over at the Cortimiglia place. Dr. Charles Gelbke, the mayor of Gretna, and City Councilman Manny Fink had been standing together in front of the Cortimiglia store, discussing the morning's events. Seeing Frank standing on the street, they walked over. _Do you mind if we look around your house?_ asked the mayor. Frank nodded. He understood that there had been a terrible crime and the authorities needed to investigate. Besides, he had nothing to hide. He took them in the house, led them through the kitchen and the little parlor in the front room, let them look into the bedrooms downstairs, and took them up the stairs to the room Frank shared with Louis. The mayor asked a few questions, which Frank answered as best he could. Then Dr. Gelbke asked to see the backyard. _Where is your axe?_ he wanted to know. Frank retrieved it and handed it to the mayor, who peered at it closely. No blood, no hair. The mayor's investigation found nothing incriminating. But it would do little to allay mounting suspicions against Frank and his family. All that afternoon people crowded in and out of the Jordano store. The only topic of conversation was the Cortimiglias—would they die? Who did it? Did it have anything to do with another middle-of-the-night break-in just up the river in the village of Amesville only two nights before? Someone had chiseled a lower panel off the door of a grocery belonging to Santo Vicari. The robber had entered the store and stolen some cash but not molested the sleeping family. Was the same night stalker responsible? Iorlando was as curious as anyone, and as grieved. But he was also worried that people suspected him. He was convinced that everyone was staring at him, pointing him out as the killer. The old man insisted to anyone who would listen that his differences with the Cortimiglias had been patched up, and they were friends again. But all the neighbors knew about the quarrel between the Jordanos and Cortimiglias, and everyone noticed that the Jordanos had been among the first on the scene. Suspicion began to grow. When Jefferson Parish acting coroner J. R. Fernandez arrived at the scene, the little girl's body had already been removed by the undertaker, but Dr. Fernandez got a good look at the site of the murder. Like the other Axeman crime scenes, the room had been ransacked, dressers emptied, and clothes and papers strewn all over the floor. Nothing had been taken. A search of the room revealed $129 in cash, a box of jewelry, and a loaded revolver. Clearly, Dr. Fernandez concluded, robbery had not been the motive. The police had found two axes on the premises; only one had blood and hair sticking to it. Fernandez managed that Sunday evening to swiftly impress five neighbors into service as a coroner's jury, including Manny Fink, who'd been one of the first on the crime scene. Dr. Fernandez and his jury crossed the river to view the body at the undertaker's about 7 PM, then returned to Gretna to examine the murder scene. They were shown two axes found on the Cortimiglias' property. The jurors stared somberly at the dark hairs plastered in blood to one of the axes. Only a handful of witnesses were called to testify at the inquest. Lillie, Iorlando, and Frank were sworn in. They all said the same thing—they'd heard nothing untoward the night before, no hint of the horror taking place next door. They admitted that they'd had a falling out with Charlie Cortimiglia over possession of the property but insisted the quarrel had blown over. Frank testified that Mary's mother had continued to allow the child to frequently play over at their store. Frank added one detail to his testimony, a detail that would come back to hurt him in an unexpected way. He was asked about the dream he'd mentioned to his mother. Frank was under oath, but he was also tired by the interrogation, drained from the distressing day, and harassed by probing questions. He didn't want to talk about Josie in front of all these men. What difference did it make who'd had the dream? So he told the inquest that he was the one who'd had an unsettling dream—of snakes all around him, a vision that many of the newspapermen reporting on the case interpreted as an evil omen. It was an impulse he'd come to regret. Because the Cortimiglias couldn't be interviewed that night, the jury suspended its investigation without returning a verdict. But Dr. Fernandez offered the opinion that "the murder was the deed of a maniac, and that revenge was the only motive that could be advanced if the crime had been committed by a mentally normal person." Frank Mooney couldn't have been shocked by the savage attack in Gretna. No doubt he'd been expecting something like it, and he was anxious to help the Jefferson Parish authorities. He didn't view the crime scene for himself—no need to give offense to the Gretna powers—but on Monday, he sent Bertillon Operator Maurice O'Neil across the river to process the scene. The results were disappointing. The crowds that had stampeded in and out of the house before the arrival of the police had trampled any footprints left by the murderer into oblivion. The axe had been handled by countless people in the last thirty-six hours. O'Neil couldn't find any fingerprints in the bedroom either. To Bertillon Operator O'Neil it was readily apparent that the New Orleans Axeman had struck again, an opinion shared by Frank Mooney. The superintendent ordered his men to help the Gretna authorities in any way they could to find the murderer and to remain on alert against the killer in New Orleans. He expressed his predictable confidence that the murderer would be brought to justice. But the Italian community remained on edge. In New Orleans, Rocco Tramontana reported that someone had tried to break into his mother's grocery store at Peniston and Clara Streets during the night; the family's axe was found in front of the house. While the investigation into the attacks began, Rosie's relatives, alerted to the tragedy, were pouring in from the surrounding towns and plantations. At first, the doctors allowed no one to see the wounded couple and told the family that Charlie had little chance of surviving. Rosie regained a kind of hazy consciousness on Monday, and after Mary's funeral, her father and stepmother were allowed to sit by her bedside, taking turns holding her hand and talking softly to her in Italian. Occasionally, Rosie would feebly murmur something in response. When, every so often, she whispered, "Mary, Mary," her father would tell her to sleep, sleep and get better. No one was going to tell her Mary was dead until she was stronger. By Monday night doctors began to feel more optimistic about Rosie's recovery. And by Tuesday morning even Charlie showed signs of improvement. He regained consciousness and appeared to recognize those clustered around his hospital cot. Perhaps, doctors Leidenheimer and Landry thought, he might survive after all. Chief Leson and Sheriff Marrero—now back in town and taking the lead in the investigation—waited impatiently to speak to the Cortimiglias. On Tuesday Dr. Leidenheimer was sufficiently reassured by Rosie's and Charlie's recovery that he gave in to the importunate police officers and deputy sheriffs who had been cluttering the halls of Charity and agreed that the couple could be questioned about their assault. He warned that they might not be able to remember anything. And he was right. Speaking in Italian, all Rosie did was ask, again, after her daughter. Charlie said that his head hurt, and he needed to get back to his store. Neither had anything to say about the Axeman. But Tuesday night, investigators were determined to question the victims again. For now they had a suspect. Charles Anderson was a black man and career burglar who was well known to the police in New Orleans and Gretna. He was, at the moment, Superintendent Mooney's premier Axeman suspect. After the assault on the Cortimiglias, Mooney sent men looking for Anderson, and they caught up with him Tuesday morning, two days after the attack. Questioned about his recent whereabouts, the suspect was not entirely forthcoming, which, of course, made investigators even more suspicious. For several hours on Tuesday afternoon, Superintendent Mooney had Anderson in his office and subjected him to the full third degree. Mooney, Detective Marullo, and Chief Leson interrogated him nonstop, shooting question after question at him, not giving him time to think, trying to keep him off balance. They grilled him about all of the Axeman murders, accused him of being the killer, forced him to look at bloody photographs of Catherine and Joe Maggio, and demanded to know why he had targeted them. It was all designed to make a guilty man confess. Anderson, however, held his ground and continued to maintain his innocence, claiming that whenever something happened the police—unfairly, he insisted—picked him up. He had nothing to do with these murders. Some of the investigators began to believe he was telling the truth. To the eyes of trained detectives, the convicted burglar acted like an innocent man and didn't give himself away with the nervousness guilty men usually exhibited. Finally, in desperation, Mooney loaded Anderson into a police car, drove to Charity Hospital, dragged him into Rosie Cortimiglia's ward, and handcuffed him to her hospital cot. _Was this the man who attacked you?_ Marullo asked in Italian, as Mooney and Leson looked on. Rosie mumbled incoherently. Doctors Leidenheimer and Landry were just walking out of Charlie's room, discussing his condition, when Mooney, Leson, and Marullo showed up with Anderson. The unfortunate suspect was again handcuffed to the foot of the bed, displayed in front of the patient. Charlie looked quizzically up at the gentlemen surrounding his hospital cot. Detective Marullo asked Charlie if this was the man who had attacked him. Charlie shook his head—no. "Do you know who attacked you, Charlie?" No, he again indicated with a shake of his head. "Was it a white man?" Yes, Charlie nodded. _Have you ever seen this man before?_ No response. Charlie stirred restlessly, clearly indicating that he'd had enough, and that was all they could get out of him. Disappointed, Mooney realized he had nothing much to hold Anderson on. So he took him back to New Orleans and charged him with that policemen's consolation prize—being a dangerous and suspicious character. For the next couple of days, it was much the same with the injured victims. Both seemed to be slowly getting better, to the point where each was asking about the other, but neither was able to give any clues about their attack. Dr. Landry showed particular attention to Rosie; he checked on her frequently, three or four times a day. As a medical man, he was quite interested in her case; his own questioning of her indicated that her mind seemed to be a total blank from the time she went to bed on Saturday night until she woke up in the hospital. "Who hit you?" he asked repeatedly. Her reply was consistent: "I don't know." Then suddenly on Friday, five days after the attack, Rosie identified her attackers and the killers of her baby daughter as seventeen-year-old Frank Jordano and sixty-eight-year-old Iorlando Jordano. Or so Leson and Marrero would later claim. In the course of their inquiries, Gretna authorities had followed up on the rumors that the Jordanos had threatened the Cortimiglias and discovered a feed store operator named Rube Mayronne who claimed that two weeks before the attack, Frank said to him, "That son-of-a-bitch Cortimiglia won't be in his new store longer than two weeks." The Jordanos went from possible suspects to prime suspects. Still, all the police had was circumstantial evidence. But now, they said, they had eyewitnesses. According to Sheriff Marrero and Chief Leson, when Leson went to see Rosie on Friday morning and asked for the hundredth time, "Who hit you?" Rosie finally had an answer: "Frank Jordano and the old man." Remarkably, they said that Charlie had confirmed this identification. Over the course of the day, the couple was questioned again and again, and each time, Marrero and Leson claimed, they implicated the elderly Italian and his teenaged son. Sheriff Marrero, however, only made the identification of Frank public, and on Friday evening only Frank was arrested. Protesting his innocence, he was taken to the parish jail and held without bail. Andy Ojeda thought all this was strange. As the _States_ police reporter, Ojeda followed the course of the investigation carefully. He'd hung around Charity for days, talked to doctors, and kept tabs on the recovery of the Cortimiglias. He had been under the distinct impression that Rosie and Charlie had no idea who'd put them in the hospital. Hearing about Frank's arrest Friday night, he was at Charity first thing Saturday morning. If the Cortimiglias were implicating Frank Jordano, he wanted to hear about it firsthand. When Andy Ojeda saw him, Charlie was not really up for an interview. He could barely speak and answered most of Ojeda's questions with a nod or shake of his head. "They've arrested Frank Jordano because you said he struck you with the axe," began the reporter. "Did Jordano hit you?" Charlie just shrugged his shoulders. "Do you know who hit you?" asked Ojeda. He had to repeat his question several times. Again, Charlie only managed a shrug. Finally, slightly exasperated, Ojeda said, "You don't know who hit you, do you?" Charlie struggled to answer, and Ojeda leaned over him so he could hear the words: "I don't know." "Did you see the face of the man who hit you?" pressed Ojeda. Charlie shook his head no. Asked if he saw the attacker's face, Charlie shook his head again. "Was the man who hit you tall or short?" "Big," mumbled Charlie. This exchange exhausted Charlie, and Ojeda decided that would be all he could get out of him today. He left Charlie and headed across the hospital yard to the women's ward to see Rosie. Entering the ward, he found the charge nurse and asked if Mrs. Cortimiglia was up to answering questions. The nurse led him to Rosie's cot where she lay sleeping. Gently waking her, the nurse told her she had a visitor. Rosie opened her eyes and looked at the reporter. "How are you feeling?" he asked her. "I'm feeling all right," she answered. "Mrs. Cortimiglia, who hit you?" Rosie hesitated, and then said softly, "I don't know, but I believe my husband did." Ojeda thought that was nonsense. He had just seen what condition Charlie was in. He hadn't beaten anyone with an axe. "Who hit your baby?" he asked. She looked at him like she didn't know she had a baby. "I don't know," she responded again. Ojeda let Rosie go back to sleep. He had gotten what he came for. On his way out of the hospital, Ojeda spoke with Dr. Landry, who told him that he thought it likely they would both recover. "I will not vouch for the condition of their minds, however," added Landry. "It may leave both permanently with faulty minds." He thought it was far too soon to have much confidence in anything either patient said. Ojeda left Charity with the conviction that neither Cortimiglia could identify the attacker. What were the Gretna authorities up to? The next Sunday, March 16, the _Times-Picayune_ published an unusual letter. It had arrived at the paper two days before, handwritten in clear, distinct script. The letter purported to be from the Axeman himself: Hell, March 13, 1919 Editor of the Times-Picayune, New Orleans: Esteemed mortal: They have never caught me and they never will. They have never seen me, for I am invisible, even as the ether which surrounds your earth. I am not a human being, but a spirit and a fell demon from hottest hell. I am what you Orleanians and your foolish police call the axman. When I see fit, I shall come again and claim other victims. I alone know whom [ _sic_ ] they shall be. I shall leave no clue, except perhaps my bloody ax, besmeared with the blood and brains of he whom I have sent below to keep me company. If you wish you may tell the police to be careful not to rile me. Of course, I am a reasonable spirit. I take no offense at the way in which they have conducted their investigations in the past. In fact, they have been so utterly stupid so as to amuse not only me, but His Satanic Majesty, Francis Joseph [the Austro-Hungarian emperor who had died three years before], etc. But tell them to beware. Let them not try to discover what I am, for it were better that they never were born than for them to incur the wrath of the axman. I don't think that there is any need of such a warning, for I feel sure that your police will always dodge me, as they have in the past. They are wise and know who to keep away from all harm. Undoubtedly you Orleanians think of me as a most horrible murderer, which I am, but I could be much worse if I wanted to. If I wished to I could pay a visit to your city every night. At will I could slay thousands of your best citizens, for I am in close relationship with the Angel of Death. Now, to be exact, at 12:25 o'clock (earthly time) on next Tuesday night, I am going to pass over New Orleans. In my infinite mercy, I am going to make a little proposition to the people. Here it is: I am very fond of jazz music and I swear by all the devils in the nether regions, [ _sic_ ] that every person shall be spared in whose house a jazz band is in full swing at the time I have just mentioned. If everyone has a jazz band going, well, then so much the better for the people. One thing is certain and that is some of those persons who do not jazz it on Tuesday night (if there be any) will get the ax. Well, as I am cold and crave the warmth of my native Tartarus, and as it is about time that I have left your homely earth, I will cease my discourse. Hoping that thou wilt publish this, that it may go well with thee, I have been, am and will be the worst spirit that ever existed either in fact or the realm of fancy. The Axman. The real killer didn't send this letter. It was too well-written and too sophisticated to have been composed by the working-class Axeman. It also echoed several alleged Jack the Ripper letters sent in the fall of 1888 during his reign of terror in London. One was posted, like the Axeman's letter, "From Hell." Another, the so-called "Dear Boss" letter, ridiculed the efficacy of the police: "I keep on hearing the police have caught me but they won't fix me just yet. I have laughed when they look so clever and talk about being on the right track." The writer of the Axeman letter was well-educated and literate enough to have had some familiarity with details of the Jack the Ripper story, not to mention such elements of classical mythology as Tartarus. Maybe he didn't expect his joke to be taken seriously. Despite its macabre subject, the letter has a lofty, theatrical quality that doesn't sound like a genuine attempt to imitate a murderous maniac. The mocking tone of the letter made light of the genuine threat and very real fears created by an all-too-real killer. It was in the spirit of those who felt free to turn the murders into a joke, such as the creators of the Piggly Wiggly ad that used the Axeman to advertise self-service grocery shopping, and the party hosts (who sound like frat boys) who responded to the murderous threat by inviting the Axeman to a stag party. Surely the staff at the _Times-Picayune_ realized this was a practical joke. That the letter wasn't printed until two days after it arrived suggests the possibility of a degree of hesitation. But like other Orleanians who had nothing to fear from a killer of immigrant grocers, the newspaper editors could afford to find the idea of an epistle-writing demon from hell amusing. If not the real Axeman, who then wrote the letter? Someone who stood to gain from the prank was local musician and businessman Joseph John Davilla. The thirty-five-year-old songwriter had his own music publishing company and had had success with hits such as "Why Do You Leave Me, Sweetheart" and "There's Something I Like About You." "Give Me Back My Husband, You've Had Him Long E-Nuff" had been his biggest hit, reaching a national audience. Written for vaudeville star Sophie "The Last of the Red-Hot Mamas" Tucker, it was what was known in that racist age as a "coon novelty song." Now, he claimed, the Axeman's letter had inspired another novelty song, "The Mysterious Axman's Jazz (Don't Scare Me, Papa)." As much businessman as artist, Davilla recognized an opportunity when he saw it, even if it meant composing a theme song for a killer. The sheet music wasn't available on Tuesday, the "Axeman's" appointed night, but Davilla kept interest piqued by planting little advertisements in the _Times-Picayune_ : WAIT—WATCH THE AXMAN'S JAZZ and COMING MYSTERIOUSLY, THE AXMAN'S JAZZ. And he hired a ragtime piano player to play the song continuously as he was pulled up and down Canal Street in a wagon. Davilla turned everything to his advantage. He was clever enough to get the _Times-Picayune_ to let him use a cartoon that ran in the paper on Saint Joseph's Day, the day after the Axeman's supposed advent. It showed a woman and her frightened family anxiously playing "The Mysterious Axman's Jazz" on various instruments in a desperate attempt to stave off the killer. Davilla made it the cover of his sheet music. Copies sold by the thousands at twenty-five cents apiece, and the tune could be heard from one end of the city to the other for months. If one man stood to profit from the "Axeman's" letter to the _Times-Picayune_ , it was J. J. Davilla. _The_ Times-Picayune _cartoon that ran on Saint Joseph's Day, the day after the Axeman's supposed advent. J. J. Davilla quickly adopted it as the sheet music cover of his song "The Mysterious Axman's Jazz."_ Nowhere did the so-called Axeman letter mention "Saint Joseph's Day," but the time he demanded that jazz bands be playing—"12:25 o'clock (earthly time) on next Tuesday night" (the evening of March 18)—was the first hour of March 19, Saint Joseph's Day. That timing would have had special significance for the Italian community, the target of the Axeman. It was a particularly nasty touch. Given Saint Joseph's special significance for Sicilians, Saint Joseph's Day was an important holy day for New Orleans's Italian population, a feast day of thanksgiving and charity, and a welcome respite from the fasting and penance of Lent. The devout built altars decorated with candles and flowers and around a statue of the saint heaped foods of every description—fish, vegetables, bread, cakes, pastries—everything except meat. Dried fava beans were included for luck. On the morning of Saint Joseph's Day, Saint Mary's Italian Church was packed for a solemn high mass. After the congregation returned to their homes, a little ritual was enacted. Three children representing the Holy Family entered the home and sampled the food on the altar. Then the feasting could begin. Saint Joseph's Night was a time for parties—galas, receptions, dances, and fancy dress balls for all of New Orleans. The sound of jazz bands could be heard all over the city. The night designated for jazz by the "Axeman," the night of March 18, Italians would be busy preparing for the next day's feast. To try to frighten the Italian population on that particular night was a cruel twist on an already tasteless joke. Not a few in New Orleans used the letter to the _Times-Picayune_ as an excuse for jollification. From dozens of house parties and late-night cafés jazzy sounds floated out as merrymakers laughingly congratulated each other for avoiding the "fell demon from hottest hell." No doubt there were some who didn't want to admit a fear of the Axeman but for whom the syncopated beat of a jazz band was a welcome comfort. Most of the city ignored it, despite later legends to the contrary. The idea of an axe murderer demanding to be appeased with jazz bands was too silly to be taken seriously. Some people were genuinely frightened, the poor and superstitious, who might not even have been able to read the letter in the newspaper but who might have heard the city humming with the news—"The Axeman is coming!" With four dead in less than a year—five if they believed Mrs. Lowe had been an Axeman victim—it's no wonder that some panicked. Families banded together for the night, hoping for safety in numbers; fathers loaded shotguns and kept them within easy reach until dawn. Italian immigrants, including grocers who had cause to feel especially vulnerable, likely dominated their number. Pity the poor schoolboy piano or banjo players forced to jazz it up in improvised bands through the night to protect their families from the music-loving demon. Not everyone sophisticated enough to realize the Axeman letter was a hoax found it funny. An editorial in the _Herald_ roundly castigated the _Times-Picayune_ for not "think[ing] of the great amount of harm it has done to the ignorant classes who are superstitiously inclined and believed to a certain extent that this ax-man would visit certain families who did not have a jazz band." The editorial also thought tasteless the cartoon the paper ran the next day, the cartoon Davilla used to such effect as the cover of his sheet music: "We fail to see the joke." Frank Jordano probably didn't see the joke either. Saint Joseph's Day, the day he'd planned to get married, found him locked in the Gretna city jail. Three days after her brother's arrest, Lena Jordano watched through tears as her father was taken from his grocery by the police. Like Frank, Iorlando Jordano was charged with murder and pleaded not guilty. Sheriff Marrero and Chief Leson now announced that the Cortimiglias had identified not only Frank but Iorlando as one of their attackers. It is difficult to understand how Sheriff Marrero and Chief Leson could have convinced themselves that the mild-mannered, elderly, arthritic grocer would kill a toddler who called him Grandpa. Their fixation on the Jordanos can be attributed, in part at least, to their ignorance about serial killers. Since the term has existed for over fifty years, it's easy to forget how counterintuitive is the idea of killing with no discernible motive. Such depravity was something completely beyond the experience of the small-town police chief and the machine-politician sheriff. While Superintendent Mooney was confident that the Axeman attacks were the work of a deranged fiend, he couldn't convince everyone, even in his own department. Some New Orleans police officers continued to believe that the Axeman attacks could be blamed on Italian vendettas. Marrero and Leson didn't know anything about serial killers, but they understood—or thought they understood—the vendetta. It was an all-purpose way of explaining violence among Italian immigrants. Barely a year before, an Italian farmworker had been found bleeding to death on the levee, shot and stabbed as the result of a quarrel with another Italian over a girl. This was the vendetta. For Sheriff Marrero and Chief Leson, one Italian killing another over a business dispute was easy to fathom. The Jordanos had quarreled with the Cortimiglias. Neighbors whispered that Frank had been overheard boasting that their competitors wouldn't last long. Convinced by circumstantial evidence, sure of their guilt, the sheriff and police chief were determined to convict the killers even if they had to manufacture the evidence to do so. Even so, why insist that the old man was involved? If the police had evidence that Frank threatened the family, why did that implicate his father? Perhaps a clue can be gleaned from reports about the crime scene. Because the bolt on the kitchen door seems to have been located high above the open panel, and not in easy reach, several police officers speculated that the killer hadn't simply extended his arm through the door after removing the panel but had actually squeezed through the hole to open the door from the inside. Later events would demonstrate that even someone as big as Frank could wedge his head and shoulders through the hole, so for a smaller accomplice such a feat might have been possible. It was pretty obvious that 275-pound Frank hadn't stuffed himself through the opening into the kitchen, but his father had a much smaller build. Whatever their reasoning, the authorities of the city of Gretna and Jefferson Parish had made up their minds: Frank and Iorlando Jordano were guilty. Now, how were they going to prove it? The Cortimiglias' claims implicating the Jordanos were made only in front of police, not in the presence of reporters or physicians. To Dr. Landry, Rosie always denied that she knew who hit her. Medical personnel at the hospital, in fact, continued to insist that nothing the couple said could be relied upon yet. Not only was their mental condition affected by their head injuries, but the opiates they were given for pain also made their mental state unreliable. Reporters who went to see him discovered that Charlie would answer "yes" to anything he was asked. "Did Frank Jordano attack you with an axe?" one inquired. "Yes," responded Charlie. "Are you a Frenchman?" asked another. "Yes," came the reply. A few days after authorities claimed that Rosie had identified Frank as her attacker, she told her doctors that she didn't know if the man was white or black. In fact, the so-called statements given by the Cortimiglias consisted of agreeing to highly leading questions when they probably had little idea of what they were saying: Q: "Did Frank Jordano hit you with the axe?" A: "Yes." Q: "Was Iorlando Jordano with Frank Jordano at the time he made the axe attack on you?" A: "Yes." Even Sheriff Marrero admitted that Charlie "was unable to give details" but admitted to no doubts about the arrests. "I am confident we have the right men and that the Cortimiglias will recover to tell the complete story of the attack made upon them," Sheriff Marrero assured reporters. Frank and Iorlando's arrest was greeted with general approval in Gretna. Popular sentiment was not on their side. Charlie Cortimiglia was an agreeable, energetic young man who was well liked in the community. Everybody knew about the quarrel over the store—if they hadn't before, they did now—and that the deadly assault occurred shortly after the Cortimiglias had moved into their new place. The murder had badly shaken the people of Jefferson Parish, and, like the sheriff and police chief, many people found it easier to believe in revenge as a motive than they did a homicidal maniac. Feelings against the Jordanos were so strong that some began clamoring for a special session of the grand jury to consider the case immediately. But since the chief witnesses were still in the hospital and likely to remain there for some weeks, District Attorney Rivarde vetoed the idea. It would have to wait for the regular session of the grand jury, he said, which would convene in early May. L. Robert Rivarde was district attorney of the Twenty-Eighth Judicial District, made up of Saint Charles, Saint John, and Jefferson Parishes. He'd been a state senator and a district judge, and when Sheriff Marrero's son Louis Jr. died in 1916, Rivarde had been selected to replace him as DA. An active member of the Marrero political machine, Rivarde could be counted on to cooperate with Sheriff Marrero in prosecuting the Jordanos. Sheriff Marrero still had a problem. He was confident that the Jordanos were responsible for the attack on the Cortimiglias and little Mary's death. But his witnesses were problematic. So far neither one of them had identified the Jordanos in the presence of anyone but police officers, and Andy Ojeda's coverage in the _States_ had undermined his claim that they had clearly singled out the accused. Marrero needed to be certain of his witnesses. On March 28, almost three weeks after the attack, Rosie was at last well enough to leave the hospital. Sheriff Marrero saw a way to solve his problem. The third degree could be used on witnesses as well as suspects. He arrested her. Sheriff Marrero obtained a warrant for Rosie's arrest as a material witness, an unusual move for someone who was the victim of a crime. When she left the hospital with Chief Leson, one of Marrero's deputy sheriffs, and two of her brothers-in-law, Rosie had no hint that anything was wrong. She was in a good mood, lightly bantering with the men as they rode the ferry across the river, assuming they were taking her home. Instead Leson delivered her to the Jefferson Parish jail. Rosie was taken completely by surprise. Not allowed to see a lawyer and separated from her family, she was thoroughly frightened and confused by the time Marrero and Rivarde began questioning her again. _Was it the Jordanos who attacked you, Rosie?_ _I can't remember._ _You must remember._ _I don't remember seeing anyone._ _Who else would have done it? Didn't Frank say he was going to run you off?_ _Did you see Frank and the old man that night? You'd better remember if you want to go home._ As she was being put in a cell to sleep that night, one of her jailers said to her, "You must make up your mind who did it, Rosie. Go to bed tonight and get a good rest, and make up your mind who did it." The next morning she signed an affidavit swearing that Frank and Iorlando Jordano were the assailants. In twenty-four hours, Rosie went from claiming no memory of who killed her baby to a detailed recollection of the attack. Because Rosie couldn't read or write English, Clay Gaudet, a notary and lawyer hired by the Cortimiglias to assist the prosecution, wrote it out for her. Rosie signed. Satisfied, Marrero let her go stay with one of her sisters. Now that Rosie was out of the hospital and well enough to testify, the coroner's jury could finish its job. On April 13, Dr. Fernandez reconvened his five-man jury, and after hearing Rosie's testimony, they returned a verdict of murder at the hands of Frank and Iorlando Jordano. In the meantime, the Jordanos had retained an attorney, William H. Byrnes Jr., a thirty-eight-year-old graduate of Tulane University Law School. The first thing Byrnes did was demand a preliminary hearing, charging that Rosie had been jailed illegally in order to coerce her into accusing the Jordanos. He also wanted her mental competence evaluated. Judge John Fleury granted his request and set the preliminary hearing for May 7. But on the afternoon of May 5, a Jefferson Parish grand jury indicted Frank and Iorlando Jordano for the murder of Mary Cortimiglia. Judge Fleury set a trial date of May 19 and canceled the preliminary hearing, ruling that the indictment had made it "useless and unnecessary." In two weeks, the Jordanos would go on trial for their lives. # ≡ 10 ≡ # "Hung by the Neck Until Dead, Dead, Dead" NEW CASES OF SPANISH influenza were declining by the end of October, and by February 1919 the _States_ could declare the FLU IS ALMOST WIPED OUT HERE. Meanwhile, Louis Besumer had been brooding in the parish prison since his arrest in August. The long months waiting for trial had worn down some of the eccentric grocer's natural cockiness. He'd been sued by his landlady for not paying his rent, proving that his claims of wealth were a lie. Imprisonment and illness had aged him, and he began to despair that he'd die in jail. In February, he wrote to an Orleans Parish grand jury complaining that he was being held without trial. The attack on the Cortimiglia family a few weeks later gave him another opportunity to draw attention to his plight. He wrote to the _Times-Picayune_ , appealing for help in getting a trial so he could be cleared of blame in Mrs. Lowe's death. The _Times-Picayune_ was a good target for Besumer's plea, which resulted in a sympathetic interview on the front page of the newspaper a week after Mary Cortimiglia's death. Reporter Jim Coulton, believing that one man was responsible for all the axe attacks, was inclined to accept Besumer's claims of innocence, and he liked to remind readers that the evidence against Besumer was "largely circumstantial." The article prompted a speedy reply from District Attorney Chandler Luzenberg, who blamed the delay on a combination of judicial vacations and the Spanish flu. He promised to bring the case to trial as soon as he could. Louis Besumer's day in court finally came at noon, April 30. Prosecutors were asking for the death penalty. The trial lasted all of two days. The most important witness called by the state on the first day was Superintendent of Police Frank Mooney, and defense lawyer Henry Rhodes quickly cornered him into a damaging admission. Had Mooney declared after Besumer's arrest, and again after Mary Cortimiglia's murder, demanded Rhodes, that he believed a single crazed killer responsible for all the axe attacks of the last two years? Mooney conceded that he had. Henry Rhodes left it up to the jury to draw the obvious conclusion: If all the axe attacks were the work of one man, Besumer couldn't be guilty. He'd been locked up in the parish prison when Joseph Romano and Mary Cortimiglia were killed. The next morning Dr. Hiram W. Kostmayer, the surgeon who'd treated Mrs. Lowe at Charity, was sworn in as a witness. The outcome of the trial would revolve around two issues for which the doctor's testimony would prove critical: First, what had been Mrs. Lowe's state of mind when she exonerated and then when she accused Besumer? Immediately after the attack, Rhodes reminded the jury, Mrs. Lowe accused a mulatto, then recanted her identification, all the while emphatically denying Louis Besumer's involvement. It was only weeks later that she recalled Besumer hitting her. Dr. Kostmayer testified that despite her horrific head injuries, Mrs. Lowe's brain function had not been affected, and she had been of perfectly sound mind in the aftermath of the attack when she'd denied that Besumer had been her assailant. On the stand later that morning, Jim Coulton corroborated this statement. He testified that Mrs. Lowe seemed confused her first day in the hospital, but afterward she appeared completely lucid. It was later, when she turned on Besumer, that she seemed mentally unbalanced. Mrs. Sacriste, the housekeeper, would tell the court that Lowe seemed unhinged, with her obsessive praying and visions of Christ. The second issue on which Dr. Kostmayer's testimony was critical was the question of how Besumer came by his head wound. Police theorized that he'd inflicted it himself so that he could claim to be another Axeman victim. The defense was at great pains to discredit the idea that Besumer would have been able to deliver a blow with enough force. How hard would it be, Rhodes asked Dr. Kostmayer, to give oneself the kind of injury Besumer sustained? Kostmayer refused to answer directly. _I'm not testifying as an expert on such a question_ , he replied. _But_ , he continued, _it could have been self-inflicted only by a man of powerful physique_. The members of the jury looked over at the old grocer. Was this a man of "powerful physique" who had brought down an axe on his own head? When it came the turn of the defense to present its case, Rhodes called no witness except Louis Besumer. He testified for four hours, telling his story once again, accusing Assistant District Attorney Ben Daly of coaching Mrs. Lowe on her accusation. In his closing statement, Rhodes argued that Harriet Anna Lowe and Louis Besumer had been "victims of the murderous Axeman who had terrorized the people of this city for the last two years." Testimony wrapped up after 10 PM Thursday night. The jurymen retired to consider their verdict. They were back by 10:30 PM. Only one ballot had been necessary to find Louis Besumer not guilty. The newspapermen covering the trial weren't at all surprised by the verdict. They didn't think the prosecution had a very strong case. The testimony that Mrs. Lowe had been perfectly rational when she exonerated Besumer of the attack, coupled with the evidence that she wasn't altogether in her right mind when she did blame him, made a guilty verdict unlikely. Nobody considered that perhaps she'd been in her right mind but was too afraid of Besumer to accuse him. Little was understood at the time of the dynamics of domestic violence, certainly not by the legal system. Still, while it may not have been a correct verdict, it was a fair verdict. The defense had introduced more than enough reasonable doubt. The strongest argument in Besumer's favor was the Axeman himself. Newspaper articles for the last ten months had repeatedly linked the Besumer case with the Axeman crimes. Since the Axeman had continued to strike while Besumer was in jail, it wasn't a stretch for the jury to believe that Louis Besumer and Harriet Anna Lowe had been the Axeman's victims. After tearfully thanking the jury, Besumer announced that he would stay in New Orleans, yet no record shows him living in the city after his trial. The man who called himself Louis Besumer but was known by other names may well have vanished under one of his many aliases. _The State of Louisiana v. Frank Guagliardo, alias Frank Jordano, and Iorlando Guagliardo, alias Iorlando Jordano_ went to trial in the Twenty-Eighth District Court on Monday, May 19, 1919, almost a year after the brutal murders of Catherine and Joe Maggio. Crowds began to pack the Gretna courthouse early. The gray brick building stood at the corner of Second Street and Copernicus Avenue (now Huey P. Long Avenue), fronting Second Street and facing the nearby riverfront. Built in the classically inspired Renaissance style, the three-storied courthouse boasted a front portico, ornate Corinthian columns, rounded arches, a copper cornice, and grand colonnades. For the proud residents of Gretna, it was a veritable "temple of justice," as the _Item_ boasted when the building was dedicated in 1907. The opulence of the high ceilings, tiled corridors, and marble decoration of the inside matched the classical elegance of the outside. But the small third-floor chamber, where the Jordano-Cortimiglia drama would play out, belied the grandeur of the court building. It was a cramped little courtroom. The clerk of court's desk, positioned directly in front of the judge, was only a couple of feet from the defense table. Prosecutor Rivarde sat just outside the jury box. The prosecution and defense teams sat right next to each other. The throng that pressed into the courtroom made it seem even smaller. Italians from all over Jefferson Parish, many of them relatives of either defendants or victims, streamed into Gretna to join the crush of onlookers willing to squeeze themselves into the little courtroom to satisfy their curiosity. Hours before the court was called to order at 10 AM, every available space had been taken. Family and reporters had reserved seats; the merely inquisitive—most of them women, many with children—jammed together, anxious for a glimpse of the proceedings. Observers not lucky enough to get a seat stood at the back of the room or in the hallway or jostled each other for a few inches of standing room on the balcony. Many had packed lunches in anticipation of a long day in court. Most residents of Gretna had already reached a conclusion about the guilt or innocence of the Jordanos based on what they'd read in the newspapers or heard from their neighbors. Opinion was running strongly against the father and son; their guilt was widely assumed. Only a few cautious individuals expressed reservations about convicting them based solely upon Rosie Cortimiglia's say-so. District Judge John E. Fleury presided over the trial. Only thirty-four, he'd been a state representative before being elected judge. In fact, in 1914, he'd been the legislator who'd introduced the bill creating an additional judgeship for the Saint Charles–Saint John–Jefferson district, the same judgeship he then ran for. Fleury was a very capable politician, if, perhaps, not a great judge. At forty-one, District Attorney L. Robert "Bob" Rivarde was a slim, boyish-looking figure. He'd left school at eleven and worked a variety of jobs before completing night school and graduating from Tulane University Law School. He'd become associated with the Marrero faction early in his career when he'd taken a job working for Sheriff Marrero's son Louis Jr. The Marrero connection made John Fleury and Bob Rivarde political allies. They'd also served together as the two judges of the Twenty-Eighth Judicial District before Rivarde became DA on the death of the junior Marrero. Assisting the prosecution was the attorney hired by the Cortimiglia family, Clay Gaudet. Defense attorney William Byrnes had as associate counsel two up-and-coming young lawyers: Andrew H. Thalheim, who would go on to serve as Gretna's city attorney for thirty-five years, and Archie Higgins, a future state supreme court justice. Byrnes himself came to the case straight off a win; at the beginning of May, he'd gotten a charge of arson against one of his clients dismissed. Frank and Iorlando had an excellent legal team. Reporters expected the trial to be over in a few days. But proceedings got off to a slow start. After spending an entire morning dealing with a defense objection to the panel from which the petit—or trial—jury was to be chosen, the actual empanelment of a jury took an unusually long time. Due to the extensive newspaper coverage of the murders, many potential jurors had already made up their minds about the guilt of the defendants, and they admitted as much. Juror after juror was dismissed "for cause." _The Twenty-Eighth District Court of Gretna during the testimony of Frank Jordano. Iorlando Jordano and his attorney William Byrnes are seated at a table near the bottom of the picture and are looking up at the photographer. Andrew Thalheim and Archie Higgins are seated next to Byrnes._ Murder was a capital crime, so Bob Rivarde closely questioned each potential juror about his—they were all men—willingness to sentence a man to death. None had any misgivings. An attorney watching the proceedings commented that before the war (which had concluded only the previous November), many prospective jurors would have expressed reservations about capital punishment. The United States had had an active anti–death penalty movement in the late nineteenth and early twentieth centuries, and it had had its effect. But the Great War, the observer said, with its long lists of dead, had "obliterated their awe of death." To a man, each juror was prepared to send the Jordanos to the gallows. By 5 PM Monday, the court had exhausted the regular panel of jurors but seated only five trial jurors. Judge Fleury ordered the sheriff to summon seventy-five more potential jurors to appear in court the next morning. Fleury also ordered the five selected jurymen sequestered, and sheriff's deputies escorted them to the jurors' dormitory on the third floor of the courthouse. Over the next two days, Judge Fleury had to order that additional men be summoned for jury duty twice more because of the difficulty of finding suitable jurors. Finally, the all-male jury was complete late Wednesday night. It had taken three days to seat, longer than most murder trials. At 10:30 PM, the trial started in earnest when, despite the lateness of the hour, Judge Fleury ordered potential witnesses removed from the courtroom, and the state began its case by calling Dr. J. R. Fernandez. His evidence wasn't expected to take long. Dr. Fernandez testified about viewing the body of the murdered child and holding the inquest; he described for the court the injuries Mary Cortimiglia had suffered: two traumatic head wounds, one on her right temple, the other just above her ear. While Mary's skull had been fractured, he couldn't determine what had caused the blows, he said. The coroner, quite mistakenly as it turned out, explained that "there was no evidence of an axe there; the injury might have been caused by numerous things." His testimony lasted until 11:30 PM. The next morning, Thursday, May 22, the state called its chief witness, Rosie Cortimiglia. This was the centerpiece of the trial. Without her, the prosecution had no case. Frank and Iorlando's fate depended on how believable the jury found the young woman. All eyes in the crowded courtroom followed Rosie, straining for a glimpse of her as she made her way to the front of the courtroom. She made a good impression: an attractive, modest young woman, wearing pince-nez eyeglasses and a boudoir cap hiding her still-visible injuries. Rosie was sworn in and took her seat in the witness chair. Clay Gaudet began for the prosecution. After introductory inquiries about how long she and her husband had been in their new building, Gaudet turned to the events on the night of March 8 and the early morning of March 9. Rosie stated that she went to bed at 11:30 PM and that Charlie followed around forty-five minutes later. Sometime later, she'd woken up when she'd heard knocking. She turned to her husband. "Charlie," she said, giving him a nudge, "you had better get up, maybe somebody wants some groceries." But he just grunted, "It's too late. I closed the grocery and I'm not going to get up," and went back to sleep. Rosie said she started nursing the baby while lightly dozing; she continued to hear the noise but dismissed it as the new puppy they'd left in the yard. Suddenly, she opened her eyes to see Frank Jordano and his father in her bedroom, clearly visible in the glow of the electric bulb. Puzzled, she sat up and asked, "Frank, what are you doing here?" He didn't answer. He just looked at her. "Frank, what kind of trick is this?" Frank continued to stare at her without saying a word. Now sure that something was wrong, Rosie reached for the pistol under her pillow, but Frank grabbed her wrists and restrained her. Frightened, Rosie looked up at Frank and pleaded, "Do what you please with me and Charlie, but don't touch the baby!" Frank replied tersely, "I have got to kill the baby because it is crying," even though, Rosie testified, Mary wasn't crying. The courtroom was deadly silent as Rosie described how "he"—and she pointed to old Mr. Jordano—"went to get the axe." Then, she continued, Frank hit Mary three times with the axe—"three licks" is how she phrased it—and when he finished with the baby, he struck her twice. Gaudet then turned to what the prosecution believed was the motive for this brutal attack. He quizzed Rosie about the dispute she and Charlie had had with the Jordanos over the grocery store, and she narrated the story of how they'd rented the building before the Jordanos demanded it back. When William Byrnes stood up to cross-examine the young woman, he had a delicate task. Frank and Iorlando vehemently denied assaulting their neighbors and killing their daughter, and neither had any idea why Rosie would claim they'd done such a thing. So Byrnes had to persuade the jury that her story wasn't credible. Which meant he needed to discredit the story of a very sympathetic defendant. He began gingerly. "Miss Rosie," he said, "you don't mind me calling you by the same name that Mr. Gaudet called you?" "No sir," she replied. Byrnes requested to see the scar on her right arm. Rosie held up her arm for the jury. He asked her how she came to have the scar. She said she didn't know. Byrnes pressed her, and Rosie insisted again that she didn't know. This would be a point the defense would return to. Byrnes questioned Rosie about how old she was, how long she'd known Frank, where she went to school. He asked who had brought her to the Gretna jail and how long she was there. Then he turned to more important matters. "Now, Miss Rosie," Byrnes began, "we have to come to a very unpleasant thing, but we have to ask you these questions." He needed her to revisit the night of the attack, and he grilled her closely, but as gently as he could, about the events that had culminated in her daughter's murder. Rosie repeated the story she'd told before. She'd heard a noise that she thought was the puppy and then napped while the baby nursed. About forty-five minutes later, Frank and Iorlando Jordano had appeared in her bedroom. She sat up, pulling the baby away from her breast. Frank told her he had to kill the baby; Iorlando left the room to get the axe, and when he returned, Frank hit Mary and then swung at Rosie. "I know he hit me two licks," she said. "I could hear them two licks hit." As she fainted from the blows, Rosie continued, she'd dropped Mary, and Frank had snatched her up. "Was the baby crying?" Byrnes asked, referring to the moment when Frank took the axe from Iorlando. No, Rosie replied. The baby was asking to nurse, saying, "I want ninny, mama, I want ninny." So she said, "All right honey, I am going to give you ninny." Byrnes asked, ever so slightly disbelievingly, "You did not have time to scream?" Rosie flatly repeated herself: "I had no time to scream." "But you had to tell the baby those words, 'I'm going to give you the breast'?" "Yes, sir." Byrnes had to hope the contradiction made an impact on the jury. He also had another point that showed her story didn't make sense. "Miss Rosie, you said when Frank hit you with the axe that the baby fell and he grabbed the baby?" "Yes, sir," Rosie agreed. "He grabbed the baby." "Now, how could he grab the baby, _and_ have the axe in his hand _and_ hit you a second time?" Rosie looked unhappy. "I don't know how he was doing it; but I seen him grab the baby." Byrnes inquired as to why she didn't holler when the old man left the room to get the axe. "Because," she retorted, "it didn't take"—she snapped her fingers—"that long!" Disbelief in his voice, the attorney responded: "You had no chance to holler while he held your hands, and his father went in to the other room and got the axe, and brought the axe back into that room, and handed it to Frank, and Frank hit the baby three times and then hit you? You mean to tell this jury that you did not have time to holler?" "I was worrying about my baby," Rosie protested. "You did not jump out of bed?" "No." "You did not wake Charlie up?" "No." "You did not scream?" "No." "You did not fight?" "No." "You didn't run?" "No." "You didn't kick Charlie?" Rosie laughed. "No, I didn't kick Charlie." Byrnes was startled. "This is no laughing matter," he said. "But go ahead and laugh as much as you please." He continued probing: "All Charlie would have had to do was to reach his hand out and take the revolver to have protected you and that baby. Isn't that correct?" At this, District Attorney Rivarde objected. "Your honor, this elicits an opinion from the witness as to what Charlie could or could not have done." Byrnes was happy to be reasonable. He offered to ask the witness only how wide the couple's bed was. He turned to Rosie. "It's a double iron bed, is it not?" Rosie hesitated and looked unsure of herself. "How do you mean 'double'?" The attorney clarified: "I mean for two people to sleep in." "Yes, sir." "Why didn't you wake Charlie up when you saw Frank and his father there?" "I never thought Frank was going to kill my baby." "You didn't think it necessary to wake your husband up when you saw two men enter your room in the dead of night?" Well, Rosie replied, it had all happened so quickly. "Now, if you had time to talk to Frank, then why didn't you scream?" "I never had any idea to scream." "You said a moment ago you didn't have a chance to scream." "No, I hadn't a chance; and I hadn't the idea either." "But you had the chance to tell Frank, 'Do what you please with me and Charlie, but spare my baby'?" "Yes, sir." Then Byrnes turned to the length of time Iorlando Jordano had been out of the bedroom to get the axe. "And all that time, you say, you didn't have a chance to scream, or to awaken your husband?" "It was done so quick I didn't know what to do. I was studying about my baby." "It was done so quick, that the father had a chance to go into the other room, and maybe as far as into the yard, and then came back, and you call that quick?" "Yes, sir." Now the defense attorney's cross-examination was catching fire. "Miss Rosie, how were Frank's hands that night; were they warm or cold when he took hold of you?" Rosie looked offended. "I don't know how they were; I was so scared after he done that. Do you think I had cheek enough to know how his hands were, whether cold or warm?" The defense lawyer was sharp: "Yes, I think you should have had cheek enough; I think you had; and there were a whole lot of things you should not have done." Byrnes picked up the axe police had found at the crime scene and approached Rosie. "Now, how did Frank take the axe?" "In both hands," she said. "Right here," pointing to the center of the handle. "Then he turned you loose completely?" "Yes." "He didn't hold you and the baby?" "No." "He took the axe in both hands?" "Yes." "How high did he raise it?" "I don't know how high he raised it. I was looking at my baby." "Why weren't you looking at him?" "I didn't think he was going to hit my baby." Byrnes was incredulous: "He picked up the axe, and you told him do what he wanted to do to you and to your husband, but to spare your baby; and after he told you, according to what you say, he was going to kill your baby, you didn't think he was going to do anything?" "I didn't think he had such a heart as that." "You never thought he had such a heart as that after he told you he was going to do it?" "No, sir." "You say he hit the baby three times?" "Yes, sir; three times." The attorney lifted the axe above his head to illustrate. "He hit the baby like this?" Byrnes asked, bringing the axe down in a hacking motion. "And again? And then again?" He brought the axe down three times, simulating three strikes. "Yes, sir; just that way." "Why didn't you throw your baby out of the bed?" "I never had that idea." "It is natural to scream when a person gets frightened. What do you usually do when you get frightened?" Rosie didn't answer. The witness had been on the stand all morning. Judge Fleury ordered a midday recess for dinner. In the afternoon session, Byrnes's goal was to lay the foundation—in a technical legal sense—for impeaching Rosie's testimony. For this, he needed her to make statements that later witnesses would contradict. Byrnes returned to the subject of the scar on Rosie's arm. He asked her if she knew a real estate agent named Paul Dupas. She knew him, she said, because he had sold them several lots of land. "Now," Byrnes continued, "do you recall making a statement to him about how you received that cut on your arm?" She thought for a second. "I told him I was cut, but I didn't know who done it." "Miss Rosie," Byrnes cautioned, "I want to warn you that I intend to contradict you on that matter and have summoned Mr. Dupas for that purpose. So I want to ask you now, again, if you did not make the following statement to Mr. Dupas: That Frank Jordano had hit you on the arm with the axe and had made that scar on your arm?" Rosie could only stutter. "U-n-g—h-un." "Did you make that statement to him?" "I did not tell him that Frank done that with the axe; I did not know who cut that." Byrnes questioned Rosie for several more minutes, hammering her on what she had told Paul Dupas. He asked about a second encounter she'd had with Dupas, when she'd told him that the cut had been done in the hospital "by the doctors putting something in your arm." Rosie repeated that she didn't know how she'd received the cut; she'd never known. Byrnes had accomplished his task; he'd laid the groundwork to show that Rosie was lying—or at least unreliable—when he would later call Dupas to the stand. But he wasn't finished. He had another witness to impeach Rosie's narrative. Byrnes asked Rosie if she knew a woman named Margaret Williams, a detective hired by the defense to pose as a door-to-door seller of women's toilet articles in an effort to talk to Rosie. Rosie didn't know the name, so Miss Williams was brought into the courtroom. Yes, Rosie, acknowledged, when she got a look at the young woman, she'd met her twice. The lady had told her "she was selling cologne, powder and face soap." Byrnes asked Rosie what she'd told Margaret Williams about the case. At first, she denied telling her anything. Then she contradicted herself, saying that she'd told Miss Williams that Frank Jordano had attacked her and Charlie. "Now Miss Rosie, didn't you tell her that you didn't know whether it was Frank Jordano, or his father that had hold of your hands?" Rosie emphatically denied telling her that. Again, Byrnes warned Rosie that he was planning to contradict her with Margret Williams's testimony later. "Did you tell her that you were awakened suddenly?" "What do you mean 'suddenly'?" "Quickly. That you were awakened; seeing two men by your bed, you recognized them as Frank Jordano and his father, but did not have time to scream; that one of the men caught you by the arm; that you don't remember which man struck the baby. Do you remember making that statement?" Rosie did not: "I never told her that." Suddenly, Byrnes changed direction. "Now Miss Rosie, who was the first one to tell you that Frank Jordano had been arrested for this crime?" Rosie: "I don't remember who told me, but I remember he was in jail, and I knew he done it, and I said he done it." "You don't remember anybody telling you?" "No, sir." "Well, how did you learn of it?" "Didn't I know who done it? What do you think he was going to be put in jail for?" Not unkindly, Byrnes replied, "No, I am not asking you that, Miss Rosie. I am asking you this question, and I am going to go over it very carefully because I do not desire to confuse you, or be funny, or smart, or anything else. Who was the first person that told you that Frank Jordano had been arrested and put in jail and accused of this crime?" Rosie insisted that she couldn't remember who told her Frank was in jail; all she knew was that he was. "You had left the hospital when they told you that Frank Jordano was in jail?" "Yes, sir." "Do you remember making a statement while in the hospital that it was your husband that struck you?" "No, sir." "Do you believe you might have made it?" "When I was unconscious I don't remember what I was saying; I might say anything." "What was told you when they took you on the ferryboat on the day they brought you over here? What did they tell you when they brought you to the jail?" Rosie misunderstood the question. She replied that being taken to jail left her so shaken that she couldn't say anything. "Now during that time [in jail] did you make any statement?" "Yes, sir; I made two statements." "Now, did you write those statements down?" "No. I cannot write in American." "Were those statements written down?" "One was; the second one." "Did you sign it?" "Yes, I signed it." Byrnes turned to Judge Fleury. "Now, if your Honor please . . . I now call upon the District Attorney to produce that statement." Bob Rivarde denied that he possessed any such statement. He added, however, that if he did, he would have no obligation to turn it over to the defense. Judge Fleury sided with the DA, ruling that the defense had no right to the state's evidence. Byrnes turned back to the witness. "Before whom did you make the statement?" Rosie again was puzzled by the language. "Before whom? How do you mean?" Byrnes clarified: "Who was present, and who took it down in writing?" Rosie named Charlie Burgbacher (the deputy sheriff in charge of the jail), Clay Gaudet, and another man who came with Mr. Gaudet whose name she didn't know. "It was this gentleman—Clay Gaudet—who is now assisting the District Attorney who took down your statement?" Byrnes asked. Rosie: "How you mean 'assisting'?" "He is helping the District Attorney. The same man that took that statement down in writing, and had you swear to it, is the same gentleman that is now here with Mr. Rivarde, Mr. Clay Gaudet?" "Yes, sir." Again, Byrnes turned to the prosecution. "I call upon Mr. Gaudet, an associate counsel for the prosecution, for the production of that statement." He argued that he was entitled to it because he believed it would contradict Rosie's testimony. Once again, Rivarde demurred. "We refuse to produce it because it is not a public record." The judge ruled against the defense again, another short-term defeat. Byrnes turned to other matters. After quizzing her about her statement to the coroner's jury, Byrnes asked Rosie about the relationship between the Jordanos and Mary. "Now Miss Rosie, where did you get your baby from the night before this crime happened—on Friday night." "Mrs. Jordano's house. She was playing, and I told her to come on home, and Frank said, 'Let her stay,' and I said, 'No, it's time to put her to bed,' and I picked her up and went home with her." "Now, is it not a fact, that Frank Jordano was very fond of your baby?" "Yes, he loved the baby; but I don't know if he meant it in his heart. He always loved the baby, and kissed it." Byrnes asked Rosie about her relationship with Frank. "Had you and Frank ever had any words or fuss?" "No, sir." "Were you friendly?" "I was friendly." "Whenever Frank was there he played with the baby, didn't he?" "Yes, sir." "What did the baby call Mr. Jordano, the old gentleman?" "Grandpa." "One time when the baby was very sick didn't Mrs. Jordano nurse the baby?" "Yes, sir; many time when the baby was sick. I always got her to do something for the baby, and she always done it." "She was very fond of the baby?" "Yes, sir." "And Mr. Jordano, the old gentleman, was very fond of the baby, too?" "Yes, sir." "And the baby was very fond of him?" "Yes, sir." Byrnes hoped he'd made his point. At this point Byrnes passed the cross-examination over to his co-counsel, Archie Higgins. Higgins's task was to question Rosie again about how she and Charlie had been ejected from their grocery. It was a risky strategy, dangerously emphasizing the conflict between the two families. But perhaps it would suggest to the jury a reason that Rosie might invent testimony against the father and son. He asked her a few questions about the legal notice that was served forcing them to give up the store, how long it took them to actually move, and how hard Iorlando had pressed them to leave. Then, abruptly, Higgins switched back to the night of the attack. "I want you to tell us why you didn't arouse your husband when you saw two men coming into your house." Rivarde was quick with his objection. "This witness has been on the stand the whole of the afternoon, and that question has been asked and answered." Judge Fleury agreed. Higgins asked Rosie a few more questions before ending his interrogation of the witness. But before he sat down, he gave her one final opportunity to reconsider her testimony. Rosie was an unsophisticated, uneducated, and not especially bright twenty-one-year-old. Was it possible that she didn't understand the consequences of accusing two men of murder? How much thought did she give to what would happen to her two former neighbors, especially the young man who'd been so fond of her daughter? "Now Mrs. Cortimiglia, you realize that two men are being tried for murder, and that your testimony in this case is going to be heard as evidence for the State to convict them of murder, and if these men are convicted of murder, Frank Jordano and [Iorlando] Jordano sitting here, they will be hung by the State of Louisiana by the neck until they are dead, dead, dead. You realize that fact?" Rosie looked confused. "How do you mean? I don't understand." Higgins responded gravely: "I want you to realize the seriousness of this, because I don't want you to make a statement [without] . . . realiz[ing] the seriousness of it. I ask you now if you know that the testimony that you are giving here . . . [is] made under oath . . . and that evidence is being taken here in court in an effort by the State to convict two men of murder when the penalty is death?" Rosie didn't seem to understand that either. Higgins patiently explained it to her again—she was testifying under oath, she had to tell the truth and nothing but the truth, and the crime for which the Jordanos had been charged was punishable by hanging. Finally, Rosie nodded, "Yes, sir," she understood. The defense could only hope that the jury had understood that she did not, in fact, understand and that the jurymen believed that her account was so hopelessly muddled that they dare not send a man to the gallows on it. When Rosie got off the stand at 5:30 PM, she had testified for over five hours, much of it under intense cross-examination. Those who'd listened to her so intently had been favorably impressed, and a murmur of sympathy swept through the crowd as she was escorted out of the courtroom. Later that night, the jurymen settled back into their chairs to hear from half a dozen more witnesses. First, Dr. Fernandez, at his own request, again took the stand because he wanted to correct some of his earlier testimony. The previous evening, he had testified that he didn't know how the injuries on Mary Cortimiglia had been inflicted. But, he told the court, he'd forgotten that "there was an axe produced at the inquest which was bloody and had hair on the end of it . . . which made me think that probably the instrument had some connection with it." Byrnes asked the doctor about Rosie's testimony in front of the coroner's jury. Then he asked Dr. Fernandez why he hadn't put her testimony in writing. "Didn't you know this was a murder trial?" he asked. The coroner replied that Sheriff Marrero told him that she'd already signed a notarized statement. "The Sheriff told you not to take her statement down in writing?" Rivarde interrupted. "I object, if your Honor please, to what the Sheriff told him, upon the ground that it is a conversation of third parties." Byrnes argued that it was a legitimate question since Sheriff Marrero was a law enforcement officer. Judge Fleury, however, dismissed the exchange between Dr. Fernandez and Sheriff Marrero as irrelevant hearsay. Dr. Fernandez stepped down from the stand, and other witnesses took his place. Drs. G. W. Rossner and Marvin Odom testified about being called to the scene early that Sunday morning. Hazel Johnson and her nephew George Bolden (who had accompanied his aunt that morning) testified about finding the injured Cortimiglias. The last witness Thursday evening was City Councilman Emmanuel Fink, who recounted being woken up about three o'clock on the morning of March 9 by barking dogs and a distant knocking or tapping sound. Fink described the discovery of the Cortimiglias later that morning and told of telephoning for an ambulance and the police. At some point during the ensuing chaos, someone handed him an axe they had discovered under the house, and he, noticing a swatch of dark hair on the blade, gave it to Chief Leson. At the prosecution's request, the witness identified one of the two axes found on the Cortimiglia premises as the axe he meant, pointing out that a piece of hair was still visible on the blade. Court adjourned at 11:30 PM. The final witness for the state was sworn in at 10:25 the following morning. Charlie Cortimiglia, who had only been out of Charity Hospital for a few weeks, wasn't on the stand for long. He said that he'd gone to bed on the night of March 8 with a pistol and cash under his pillow and woken up in the hospital weeks later. He remembered nothing of the attack or his daughter's death. When Charlie was through, DA Robert Rivarde rested his case. # ≡ 11 ≡ # Verdict THE FIRST WITNESS CALLED for the defense was Frank's sister Lena. The twenty-year-old young woman narrated in chilling detail the morning Hazel Johnson shot out of the Cortimiglias' house shrieking, "Jesus! Oh, Jesus! Oh, Jesus!" Byrnes used Lena to make several points. He wanted to demonstrate that Frank had no interest in the grocery business—and thus no motive for killing the Cortimiglias—so he quizzed Lena about her brother's work history in real estate and life insurance. Byrnes also elicited from Lena the fact that on Sunday morning Frank was wearing the same clothes he'd worn the night before, presumably indicating he'd not splattered them with blood during the night. She provided alibis for her father and brother, testifying that she hadn't heard her father in the next room getting up during the night and that Frank had gone upstairs sometime after 10 PM and hadn't come down until she'd started screaming about the Cortimiglias. Under cross-examination, however, the prosecution got her to concede that it was possible for either Frank or Iorlando to have sneaked out of the house without her knowledge. Byrnes's next witness was Paul Dupas, a thirty-five-year-old New Orleans real estate agent for whom Frank Jordano had once worked. Dupas had been aghast at the charges against Frank. He liked the boy and had trouble believing that he could do something so appalling. But a lot of people in Gretna thought Frank was guilty. And why would Rosie Cortimiglia—whom he also knew slightly—make up such a thing? Her little girl, after all, was dead. Dupas had felt an acute need to know what had really happened. Quite by accident, he had his opportunity when he was out on business over the river and realized that he was near where the Cortimiglias were living with Rosie's sister in Amesville. On impulse, he decided to call in on them. Both Charlie and Rosie were home. Charlie had little to say, still recovering as he was from his ordeal. Rosie, however, was talkative and quite happy to tell Dupas what had happened. Animatedly, she recounted the attack, saying that when Frank had tried to kill Mary, she had fought him, attempting to protect her little girl. _Here_ , she said eagerly to Dupas, pulling up the sleeve of her dress. _He cut me here_ , she said, pointing to a scar on her arm, on the inside of her elbow. Dupas returned home profoundly grieved by what he had learned. Frank must have done it. The cut on Rosie's arm was proof that she'd tried to fight off her attacker and proof, in Dupas's mind, that she was telling the truth. He'd told everyone in his office what Rosie had said, and word had spread. One day, Dupas received a telephone call from William Byrnes. When he'd heard the story, far from believing that this information indicted Frank, the lawyer knew that it might help exonerate him because it contradicted what Rosie had told the police. And since Rosie was the only eyewitness against Frank and his father, any evidence that demonstrated her unreliability as a witness could help win his clients' freedom. Byrnes asked Dupas if he'd be willing to see if Rosie would repeat the same story. So the week before the trial started, Paul Dupas headed back out to Amesville, determined to see Rosie again. Under questioning from both the state and the defense, Dupas narrated his story for the jury. Finally, Rivarde addressed the witness: "Now, Mr. Dupas, what was the sum and substance of the conversation between you and Mrs. Cortimiglia on your second visit?" The sum and substance, he replied, was that she flatly denied telling him that she'd struggled with Frank or that Frank had slashed her on the arm. She completely repudiated her previous account. No doubt Byrnes hoped that Dupas's testimony would help convince the jurors that they would be foolish to convict men of murder based solely on Rosies's testimony. After the midday dinner break, the defense called on a witness who added not so much an understanding of events as human pathos. Mrs. Jordano was on the stand for some time but mostly covered ground that had already been gone over. She described the events of the night of March 8 and morning of March 9, and her account matched that of Frank and Lena. She testified that her husband Iorlando—"my old man"—had been in bed with her the entire night; he had trouble getting up now, with his rheumatic back. She ended by turning to the jury and tearfully declaring, "My old man and my child are innocent!" Next, the defense called the _New Orleans States_ police reporter Andrew Ojeda. Ojeda told of seeing the Cortimiglias in Charity on the day after Frank had been arrested. He testified that when Rosie had told him that her husband had been the attacker, he hadn't believed her because Charlie was himself so badly injured. In his cross, the DA tried to suggest that Ojeda's account of Rosie's statement had been exaggerated or even fabricated by the newspaper. But the newspaperman swore that the _States_ accurately reported what she told him. Iorlando Jordano was the next witness. He slowly made his way to the witness chair, his back trouble painfully evident. Two months in jail had not improved his health. In heavily accented, broken English, he testified as to his movements on Saturday afternoon and evening and about being called to the murder scene Sunday morning. He added that he had loved the baby who called him, in her childish lisp, "Ganpa." Rivarde rose to question him and asked Iorlando if he knew a man named Ed Hanson. When Iorlando hesitated, Rivarde directed that Hanson be brought into the courtroom. "Yes," Iorlando nodded, "I know this man." Rivarde warned the old grocer that the next question he was going to ask was "for the purpose of impeaching and contradicting your testimony." Rivarde walked up to Iorlando: "Do you remember meeting Hanson at the Sportsmens Café?" Iorlando didn't even remember the saloon until Rivarde prompted his memory. Rivarde repeated his question: "Mr. Jordano, did you or did you not meet Ed Hanson on Sunday, February 16 at the Sportsmens Café in Gretna? Did you, or did you not, have a conversation in which Hanson said to you: 'I see you have gone back in the grocery business and you are going to have opposition, the Cortimiglias are going to open a store. I hope you do well.' Then you said to Hanson, 'The opposition don't worry me, I will get rid of them, I will kill them all.' Did you, or did you not have that conversation, and tell Mr. Ed Hanson that?" "I never said that!" The old man's voice shook. "If he told you that he told you a God damn lie; I never said that." Judge Fleury broke in to gently reprimand the witness. "Listen, Mr. Jordano. I don't want to be harsh with you, but that language is not permissible in the Courthouse. I appreciate your situation as a witness, but the Court is not going to permit that kind of language." Rivarde continued: "Then you deny having said that?" "I never said that," Iorlando repeated. "I never told nothing to that man." When it was again Byrnes's turn to question the witness, he approached Iorlando. "Mr. Jordano, did you at any time say to Mr. Hanson, or to any other person, that you were going to kill Charlie Cortimiglia, and his wife, and the child or any words like that, or any words that could be taken like that?" "No," Iorlando replied. "I didn't say that." "Mr. Jordano, did you at any time tell anyone, white or black, that you were going to kill Charlie Cortimiglia, or his wife, or his child?" "I never opened my mouth [like that]," insisted Iorlando. "I don't belong to the Black Hand!" With this firm denial Iorlando's testimony ended. After the supper break, Byrnes requested that the sheriff bring the Cortimiglias' kitchen door, taken off by the hinges, into the courtroom. A couple of deputy sheriffs carried the heavy door in and set it down before the jury. Byrnes then called Superintendent of Police Frank Mooney to the stand. Byrnes's defense strategy became apparent. He wanted the jury to see Mary Cortimiglia's murder and the attack on her parents as part of a larger pattern—the Axeman's crimes. So he wanted witnesses who could testify to the pattern of the attacks and show how the Cortimiglia assault and murder fit it. The Jordanos hadn't killed Mary Cortimiglia, he wanted to show the jury. The murder was the degenerate killer who had been terrorizing New Orleans for months. Byrnes attempted to ask Mooney about the axe attacks and robberies connected to the "Ax-man" in New Orleans, but each time the prosecution objected on the grounds that the witness's testimony wasn't directly connected to the Cortimiglia case. Judge Fleury agreed, saying, "We are only concerned with the investigation of the particular case under consideration. The mere fact that they may have had several other cases of a similar character, has no bearing on this particular case." Byrnes argued that it was relevant to show that someone other than his clients had a motive for the crime and was responsible for a series of similar crimes. His clients had an alibi, he said. They were at home in their own beds when Mary Cortimiglia was killed. When defendants claim an alibi, he continued, they have a legal right to show that someone else had committed the crime—in this case, a "degenerate . . . a pervert . . . a sadist . . . [a] Jack the Ripper type." Byrnes reserved a bill of exception, a written statement of his objection to the judge's ruling, which would put into the legal record his argument for an appeal. He would do this repeatedly throughout the trial. Realizing that Fleury would continue to rule the same way, Byrnes nevertheless persisted in asking questions that showed that the Cortimiglia murder was similar in critical details to the Axeman attacks. He continued asking questions he knew the judge wouldn't allow so he could keep filing bills of exception. Byrnes asked Mooney for crime scene photographs that showed similarities among all the Axeman crimes. He asked the police superintendent if the Axeman cut out a door panel. He asked if the Axeman targeted corner houses, like that of the Cortimiglias. He asked about what kind of weapon the Axeman used. And how was the weapon usually disposed of? Were fingerprints ever found? Were sums of money often left, indicating robbery was not the motive? Didn't evidence indicate the Axeman wanted "a deluge of blood"? In each case, the prosecution objected, the judge agreed, and Byrnes reserved a bill of exception. If he wasn't making much headway in getting evidence in front of the jury about the Axeman's crimes, he was laying the groundwork for an appeal if his clients were convicted. Finally, Judge Fleury had listened to enough: "The court has ruled that testimony is inadmissible," he told Byrnes, "and that's an end of it." But Mr. Byrnes wasn't finished. He called Maurice O'Neil, the New Orleans Police Department's Bertillon operator, as a witness. He tried to ask O'Neil whether the Cortimiglia case resembled the Axeman crimes in New Orleans, but again, the prosecutor—and the judge—shut him down. Then he called to the witness chair Santo Vicari, the Amesville grocer who had been a victim of an Axeman-like burglary of his home and grocery only a few days before the attack on the Cortimiglias. And Louis Besumer, whose acquittal earlier in the month had confirmed his status as an Axeman victim in the public's eyes. And Jim Coulton, the _Times-Picayune_ reporter who'd done so much work on the Axeman case. In no instance did the judge allow him to question the witnesses about similarities between earlier crimes and that of the crimes against the Cortimiglias. Judge Fleury was not permitting an Axeman defense. Byrnes's next witnesses had seen Frank the morning of the attack or the evening before. Their testimony helped fill in for the jury Frank's actions in that time period but added nothing controversial. Margaret Williams, however, had some surprises. She was a detective hired by Byrnes to get Rosie to talk. This was the kind of stratagem employed by defense lawyers before pretrial discovery gave them access to the prosecution's case. She testified that on May 1, posing as a door-to-door seller of toilet articles, she had called on the Cortimiglias and found Rosie eager to talk. Rosie told her that she'd woken to find Frank Jordano and his father in her room. She didn't have time to scream, she said, before one of the men hit the baby, and then her, with an axe. At first, she wasn't sure whether the men had had to leave the room to find the axe, then she decided that yes, they did. Neither of the men said anything to her, and it all happened too quickly for her to say anything to them. Miss Williams's testimony continued. She thought Rosie seemed embarrassed by her shorn hair, the result of her head injuries, so the next day she returned with the present of a boudoir cap. On that visit, Rosie's story changed. This time she claimed she had spoken to Frank, and she said that "one of the men held her hands, but she didn't know which one." About the axe, Miss Williams said that Rosie told her that "she didn't remember whether they had [the axe] when they came in, or whether they went out to get it." By the time the court adjourned Friday evening, it was almost midnight. At ten o'clock the next morning, Frank Jordano took the stand in his own defense. The big, hulking youth was pale from two months' imprisonment, and he had lost weight. He hadn't been allowed to see his father, brother, sisters, or fiancée since he'd been imprisoned in the Gretna jail. He had only been allowed to see his mother a couple of times in the presence of Detective Marullo. He had a hard time understanding why he was here, why he and his father were standing trial. Now was his opportunity to set things right. At Byrnes's request, Frank told the jury how he'd dressed up that Saturday night, dropped off his suit to be cleaned and pressed, gotten a haircut and shave, briefly visited with an aunt, then walked to Lee Hall—a dance hall—where he watched the dancing for several minutes before going home. When he got home, his mother told him that if he wanted anything to eat, he'd have to get it from the store. He decided to skip supper and finished up some paperwork before saying good night to his parents. He recounted his girlfriend Josie's dream to his mother, then went to bed. The next morning he was startled awake by his sister's screams, and he flung himself half dressed down the stairs to find out what was wrong. He joined the crowd of people rushing into the Cortimiglias' home and found the baby dead and Charlie and Rosie gravely wounded. He asked Charlie, "Charlie, for Christ sake, who done this?" But all Charlie could do was ask Frank to tell his brother-in-law. So Frank hitched up his horse and buggy and rode out to Amesville, calling on Dr. Gelbke and Dr. Rossner on the way. Byrnes asked him if he'd been in the Cortimiglia house at any time that night. Frank emphatically denied it: "No, sir. I was not in there." Byrnes asked him about Rosie Cortimiglia's testimony that she had seen him hit baby Mary with the axe. Frank must have thought of little else in the last two months. Now, all he could do was deny it in the strongest of terms: "I didn't strike that baby with any axe; it is one of the biggest false stories that ever came from a living human being's mouth in the State of Louisiana, if she said I struck the baby with an axe." Byrnes: "Were you in the house that night, and did you strike that baby with anything?" Frank: "No, sir, I did not." Byrnes: "Is her statement true?" Frank: "It is untrue." He turned to look directly at the jury. "It is false." Now Byrnes turned to the kitchen door taken from the crime scene. Someone had pried off the door panel that Manny Fink and Chief Leson had nailed back into place. Byrnes pointed to the missing panel and asked Frank to climb through the opening. Would he fit? Frank dutifully took off his coat and rolled up his sleeves. Bracing himself against the wooden platform on which the clerk of court's desk sat, he stretched one arm and then the other through the hole as he tried to pull himself through it. He paused when he heard a slight rip as one of the jagged edges of the door tore his shirt. Frank took a deep breath and, steadying himself, managed to force himself a surprising distance through the opening before he got stuck and could move no farther. That was one theory of the crime debunked; whoever had attacked the Cortimiglias and however they got into the house, Frank Jordano couldn't have climbed through the hole in the door. At the request of one of the jurors, however, Frank easily reached over to slide the latch. Frank extracted himself from the door, tugged his coat back on, and sat down in the witness chair. His lawyer proceeded to ask him about his feelings toward baby Mary. "I loved the child with all my heart," he said. "Every time I came home in the evening the little child was over [at my] home and I grabbed her up and played with her until I . . . went out again." On cross-examination, the district attorney asked him about the dream he'd told his mother about. Josie had dreamed about snakes, and he'd asked his mother if it was a good or bad dream. Rivarde asked Frank what he had told the coroner's jury. Frank admitted he'd lied. He'd said that he'd had a bad dream. "I didn't want to mention my sweetheart's name," he explained. But Frank pointed out that he'd come forward with that information himself. He had recently admitted the truth to Chief Leson in the Gretna jail. Rivarde still hammered on Frank because of his falsehood. "Now, you do admit, that when you swore before the coroner's jury, you absolutely swore to a falsehood?" "I told Chief Leson I did wrong." "That is not an answer to my question. Did you know when you made that statement to the coroner's jury, after you had been sworn, that you were swearing to an absolute falsehood?" "I did know." "You did know that you were swearing to an absolute falsehood?" With some desperation, Frank said, "I did. . . . I did tell them that I had the dream, because I didn't want to mention my girl's name amongst men. I know I did wrong. I told Chief Leson I done wrong. . . . But that night they questioned me so." Rivarde let the matter drop and moved on to other questions. He asked Frank if he'd gotten into a fight Sunday morning or Saturday night. Puzzled, Frank said, no, he hadn't. Rivarde then asked if he knew a streetcar conductor on the Algiers line named Fandel. Still puzzled, Frank admitted that he did. Wasn't he on Fandel's streetcar on Sunday morning after the attack on the Cortimiglias, and didn't he tell the conductor when asked how he'd hurt his hand that he'd gotten into a fight? Frank said no, he didn't go to Algiers that morning, he hadn't seen Fandel, and he hadn't been in a fight. Suddenly, Rivarde switched to a different line of questioning: "Do you know a man named Rube Mayronne?" he demanded. Frank admitted that he did. Rivarde asked if he recalled talking to Mayronne at his feed store about two weeks before the assault on the Cortimiglias. "Did you go and ask him if he had made a loan to the Cortimiglias, which would enable them to build their store?" "No, sir," Frank replied, "I did not." "And did you," the DA continued, "tell him, 'That _son of a b_ will not be there longer than two weeks'?" Frank vehemently denied saying any such thing. But he said he had been in Mayronne's store about the time that they had gone to court with Charlie Cortimiglia, and Mayronne didn't seem to like it. Mayronne cursed him and called him "a dirty little low down dago" for evicting Cortimiglia. Frank testified that he'd told Mayronne that he didn't know what business it was of his, and continued: "So then he says to me, 'That dago says he is going to kill you if you fool with him.' I looked at him and burst out laughing, and said, 'It takes more than one man to do that job.' That is all I said to Rube Mayronne." When he had the opportunity to question Frank again, William Byrnes tried to undo some of the damage done by the prosecution by exposing Frank as, technically, a perjurer. "Now Frank, about that dream, why did you tell that story about the dream, saying you had it when your girl had it? Did you tell it to mislead the coroner's jury about the Cortimiglia murder; did you tell it to hide anything that you knew?" Frank was adamant that he had not. "Did you know at that time that you would have to answer for every word that you said—every statement?" "No, sir; I did not." "Were you preparing a defense at that time?" "No, sir." "But did you make that statement for the purpose of throwing the coroner's jury off the trail of the murderer, or did you connect it in any way with the murder?" "No, sir." "What did you do it for?" "I did not want to mention my sweetheart's name before the coroner's jury." "When you told your attorneys that it was you who had the dream, what were you advised to do?" The DA objected to this, so the question went unanswered, although it suggests that when Frank had his attack of conscience he'd gone to his lawyers, who'd advised him to confess to the truth. And, probably, a male jury of that era might have understood his hesitation at mentioning a young woman's name in a public proceeding, "notoriety" being something that no well-brought-up young woman aspired to. When Frank rose from the witness stand after three hours of testimony, he and his lawyers had reason to worry. He'd admitted under oath that he'd lied under oath. Arguably, his lie had been trivial, and arguably, it had been justifiable in defense of a young lady's reputation. But there was no way to know if the men on the jury would see it that way. No doubt, his lawyers hoped that other witnesses scheduled that day would repair some of the damage. The star in this lineup of witnesses was Dr. Jerome Landry, the house surgeon who had cared for Rosie while she was in Charity. Under direct questioning by Byrnes, Landry testified that he kept a close eye on Rosie while she was in the hospital, checking on her up to five or six times a day. For the first few days, she'd been unconscious but had "gradually regained consciousness." She had no memory of anything from the time she went to bed on Saturday night until she woke up in Charity weeks later. She didn't know what had happened, how she got there, or who had struck her. Landry said that he had asked her numerous times who had attacked her, and her consistent reply was "I don't know." Byrnes asked him if Rosie had been in any condition to identify her attackers on March 14 when Frank had been arrested. Landry's response was brief and to the point: "No." Byrnes asked him what he thought of the fact that the day after she'd been discharged from the hospital, Rosie signed a statement identifying her assailants. Dr. Landry was suspicious: "It does seem strange to me that here to-day is a woman who has a perfectly blank mind, does not know anything, and the next say she remembers everything. She leaves the hospital this evening and knows nothing about it, and the next day she knows everything [that] happened at the time of the murder." Landry also testified that although he couldn't be absolutely certain, he thought he'd heard the investigators ask Rosie if Frank had done it. Nothing Rivarde asked could shake the doctor's testimony, so when Landry left the stand, Byrnes must have been relieved; the doctor had done the defense a lot of good. The physician who had taken the greatest care of Rosie was the one who discredited her the most. Dr. Henry Leidenheimer, who had also worked at Charity when Rosie and Charlie had been patients, followed Dr. Landry. He was the physician who attended Charlie when the Cortimiglias first arrived at the hospital, while Landry operated on Rosie. It was clear that Byrnes called him as a witness not because he had observed either of the wounded as closely as Landry had observed Rosie, but because of his views on Rosie's injuries. Leidenheimer testified that he didn't believe that Rosie's head wounds had been caused by an axe. An axe, he said, would have torn the soft tissues of the head to a greater degree than he had found. Rosie, he believed, had been hit with something lighter. Charlie's injuries, however, were consistent with those caused by an axe; a heavy, sharp implement had sliced cleanly through the soft tissues of the scalp, through the skull, and into the brain. Dr. Leidenheimer may well have been wrong in his assessment of the cause of Rosie's head wounds. He was a surgeon, not a forensic pathologist, and spent his time saving lives, not studying wound patterns the way a coroner or, today, a medical examiner does. Or perhaps he was correct and the attacker had hit her with something other than an axe, as he had done to Mary Davi. Either way, for the defense his testimony was a godsend because it directly contradicted Rosie's assertion that Frank had hit her with an axe. Leidenheimer also testified that the cut Rosie had sustained on her arm could have come from an infusion injected into her at the hospital. He didn't recall if she'd had such a cut while in the hospital, but it was plausible, he thought, that a spot on her inner arm, at the bend of the elbow, just where she had a scar, had been "cut [to] expose the vein" for a medical procedure. This suggested that Rosie's story that she'd been cut on the arm by Frank might also be inaccurate. From the questions jurors asked, it was clear they didn't want to give up on Rosie's story about being struck with an axe. One asked if her wounds could have been caused by a corner of the axe; another if they could have been caused by the inability to "swing the axe properly in close quarters." The DA chimed in and asked if the injuries could have been caused with the axe handle. To all, Dr. Leidenheimer replied, "No." After Landry and Leidenheimer's testimony, reporters thought that the defense was in the best shape it had been in since the trial began. Then Chief of Police Peter Leson took the stand. He was the person for whom Rosie had first identified the Jordanos as culprits. Leson testified that he had visited the Cortimiglias in the hospital twice a day after they had been admitted. Finally, one day—he couldn't remember which one—Rosie told him that it was "Frank Jordano and the old man." He denied asking her if the Jordanos had done it and denied later warning her to "stick" to her story about the Jordanos. Initially, Leson testified that there were no other witnesses to Rosie's accusation, but later he amended his statement, saying that when he brought Dr. Gelbke to the hospital, Rosie told him that Frank and his father had hit her. When asked by Byrnes why he took Rosie to the parish jail after she was discharged from the hospital, Leson explained that it was done out of concern for her safety. They feared "someone might harm her," he claimed, although he never explained who that someone might be. They needed someplace safe to keep her "until we could find a place for her to be taken care of. . . . She had no place at that time to stay." Moreover, he added, when Byrnes pressed him, she was a material witness in a murder case. He denied realizing at the time that she had relatives in Amesville, despite the fact that two of her brothers-in-law—at least one of whom lived in Amesville—accompanied her from the hospital to the jail. When Byrnes asked him if he had questioned Frank Jordano about visiting these same relatives on the morning the crime was discovered, Leson said he didn't remember. When Byrnes asked why, if he was so concerned about Rosie's safety, he thought it safe to release her the next day, Leson disavowed responsibility; he lost control over her when he turned her over to Sheriff Marrero and Charles Burgbacher, the deputy in charge of the jail. The final witness of the day was the coroner of Orleans Parish, Dr. Joseph O'Hara. O'Hara had not examined either of the Cortimiglias, but he was a specialist in brain and nerve diseases, head of neurology at Touro Infirmary, and a consulting neurologist at both Charity Hospital and the New Orleans City Hospital of Mental Diseases. Perhaps as one who had worked with both Inspector of Police James Reynolds and Superintendent of Police Frank Mooney on Axeman cases, he, too, believed that a mad killer was loose and the Jordanos were guiltless. He certainly didn't think they should be convicted on Rosie Cortimiglia's evidence. O'Hara was on the stand for several hours, from the late afternoon until after the dinner break. Most of the defense's questions were about Rosie's injuries, but they were always couched as hypotheticals about someone who'd been injured in precisely the way Rosie had been. The gist of O'Hara's response to defense questioning was that someone who had sustained the head injuries that Rosie Cortimiglia had sustained, and had made the claims that she had made, was likely "suffering from a paranoiacal [ _sic_ ] state, . . . [which] is a firm, fixed, systematized, forced delusion or hallucination, the result of trauma, from an emotional shock, or from external forces of the head." O'Hara's expert view was that (underneath all the hypotheticals) Rosie's mental condition made her an untrustworthy witness. In particular, he believed it unlikely that she would be able to recall her attack with such precision since people who had received this type of brain injury were usually unable to recall the events leading to the injury in so much detail. Byrnes asked if such a person was more likely to be suggestible. O'Hara replied yes, "Those people are very prone to suggestions, and easy to be led." Byrnes then asked if such a person would be more open to suggestion if arrested and thrown in jail. O'Hara thought so: "Yes, she would be if she was put in jail. If she came out of the hospital suffering from an emotional shock and was put in jail, certainly she would be more depressed. If she was more depressed she would be more susceptible to suggestion." The prosecution, playing the same hypothetical game, tried to get O'Hara to say that because some circumstantial evidence supported Rosie's accusation, it should be believed, but failed. Dr. O'Hara continued to maintain that Rosie's testimony shouldn't be trusted. After Dr. O'Hara completed his testimony and rose from the witness chair, at 10:30 Saturday night the defense rested its case. Rivarde and his team must have thought O'Hara's testimony seriously damaged their case because they immediately scrambled to find an additional expert medical witness. Rivarde had had Dr. C. V. Unsworth, a specialist in mental and nervous diseases, examine Rosie earlier but hadn't known if the state would need him. He decided it did. Late Saturday night, Judge Fleury issued a subpoena for Dr. Unsworth, and the case was adjourned until Monday. On Monday morning, the prosecution's rebuttal began. Everyone hoped that today the case would finally go to the jury. The trial had already been the longest for a murder that anyone in Jefferson Parish could remember. The jurors had been living in the courthouse for almost a week, eating and sleeping in the building and attending court thirteen or more hours a day, breaking only for meals. They were ready for it to end. Everyone was ready for it to end. The defense even dispensed with the usual character witnesses—friends, neighbors, and business associates who would testify to the men's kind natures and excellent reputations. The state's first witnesses Monday morning were Robert J. Langridge, a young railroad security officer, followed by John Fourcade, a Gretna patrolman. In his earlier testimony, Frank Jordano had denied seeing either Langridge or Fourcade the Saturday night before the murder. The prosecution wanted to show that Frank had been out and about in Gretna later than he'd said he'd been and that he'd lied about not seeing the two men. Langridge and Fourcade both testified that on the night of March 8, they were chatting across from the dance hall when they'd seen Frank walk by sometime between 11:30 and midnight. That put Frank in the area later than the 10:30 or 11 PM that he'd testified he'd gone home. Langridge also stated that he'd seen Frank Jordano "walking fast" past the dance hall, not stopping to watch as he had claimed; they both testified that he'd said hello. Under cross-examination by Byrnes and Higgins, however, the two men didn't recall much else about the night. Neither could name anyone else he'd spoken to in that neighborhood, nor name anyone else he'd seen in or around the dance hall that night. Officer Fourcade couldn't say how big the crowd at the dance hall was, and he couldn't recall whether it was a dark night or who his partner had been. His sighting of Frank Jordano seemed to be the only thing Fourcade remembered. Jim Coulton at the _Times-Picayune_ thought that Byrnes did a pretty thorough job of tripping him up. Byrnes also elicited the information that Langridge was a deputy sheriff who'd received his commission from Sheriff Marrero. Langridge explained that he needed it for his work for the railroad and didn't work for Marrero, although, he conceded, he had worked for the sheriff at one time. Next, the state brought out its own medical experts, beginning with Dr. R. M. Van Wart, a specialist in nervous and mental diseases. After getting the facts of the case from Dr. Unsworth and spending an hour with Rosie that morning, he pronounced her "sane and . . . perfectly responsible [with] . . . no evidence of any mental disorder." He testified that he thought it possible for Rosie's memory of the attack to appear suddenly in such crystal clear detail. The jurors found some of the doctor's answers confusing, and at the end of his testimony one of them put a question to him. "Doctor, we bunch of boys understand some of it, but we don't understand it all," he said. "Do we understand that this woman now has her good sense and her reasoning power?" "She has, in my opinion," answered Van Wart. That's all the jury needed to hear. The dinner break had been taken during Van Wart's testimony, and he stayed on the stand into the early afternoon. It was midafternoon by the time Dr. Unsworth was sworn in. He had interviewed Rosie several times in the last three weeks, and his professional opinion was that "she is now sane and responsible for her acts, and qualified to make a competent statement." About 5 PM, the state rested its case, and the court broke for supper. The trial entered the final stretch when the court reconvened at 6 PM. The crowd of interested spectators had not diminished during the last week. It had, if anything, grown. Hundreds of people packed into the small courtroom, squeezing into every last inch of standing room and spilling out into the hallway where dozens—maybe even hundreds—tried to listen to the proceedings. Rosie, who had been banished to the witness room for much of the trial, now sat with Charlie at the front of the courtroom, behind the team of prosecutors, and right in front of the jury. Mrs. Jordano, Lena, and Josie sat anxiously behind the defense team. The time for summing up had come. Judge Fleury firmly admonished spectators against any disruption in the courtroom; anyone who interrupted the proceedings in any way would be in contempt of court. The prosecution would make its summation first, followed by the defense; the prosecution would then be allowed a rebuttal. For almost six hours, the attorneys summarized their case, marshalled their evidence, and employed rhetorical flourishes as they used both reason and emotion in an attempt to sway the twelve men who sat before them. No record of the summations in their entirety exists, but it's not hard to guess what points each side emphasized. The state's strongest evidence, both legally and emotionally, was the eyewitness Rivarde and Gaudet presented as the grieving mother who had watched helplessly as her only child was murdered and now waited for justice. Frank and Iorlando Jordano were the wicked business competitors driven to murder by jealousy. And Frank, the prosecution argued, had already proved himself to be "a perjurer! A self-confessed perjurer!" At one point, Clay Gaudet, swept up in his own eloquence, demanded, "Is there any doubt in your mind how Mary Cortimiglia met her death? There is no doubt where the responsibility lies for the murder of this child. The finger of justice"—and here he turned and pointed dramatically at Frank and Iorlando sitting only a few feet away—"by various witnesses and the mother of the child whom you have heard, points at the murderers who sit at the bar of justice." Of course, Byrnes, Higgins, and Thalheim were simultaneously on their feet, objecting to the remarks as highly prejudicial. They leaped up again a little later when Gaudet, in another rhetorical flourish, asked, "Who killed Mary Cortimiglia?" then, again pointing to the defendants, answered his own question: "These are the villains!" Both times the defense asked for a mistrial, and both times the judge refused, only cautioning the jury to disregard the remarks. The defense, for its part, pointed out weaknesses in the state's case. Rivarde had introduced Ed Hanson, who, he said, claimed that Iorlando had threatened to kill the Cortimiglias. Yet Rivarde had never backed up his claim by calling Hanson to testify. Likewise, the DA had argued that feed store owner Rube Mayronne heard Frank threaten the Cortimiglia family. And that a streetcar conductor named John Fandel had seen Frank in Algiers later on the morning of the murder with a scratch on his hand that Frank said he'd gotten in a fight. But the state produced neither as a witness. And Bob Langridge and John Fourcade—what did they prove? Both men said they had seen Frank on Saturday night at almost midnight, over an hour past the time he'd sworn he'd gone home. Neither, however, could tell the jury anything more about the evening—who else they saw, whether the moon was shining. Perhaps they did see Frank. Perhaps, though, they got the time wrong. Or maybe they had seen him some other Saturday night. Both officers had connections to the police, who were convinced that the Jordanos were guilty, so both had a possible motivation to misremember. What about the state's medical witnesses? Neither had seen Rosie in Charity Hospital, and both had spent limited time examining her since her discharge. Wasn't the testimony of her Charity physicians more persuasive? Most of all, they hammered on the unreliability of Rosie Cortimiglia as a witness. Her own attending physicians didn't think she could be believed. While in the hospital she had maintained that she could remember nothing, and hadn't even been told of her daughter's death. Yet after only a few hours in the sheriff's custody, she not only knew that Mary was dead, she finally recalled—in exact detail—how she had died. How had this happened? How had she even learned that Mary had died? How could she have known her baby was dead unless someone told her? Yet no one—certainly not Chief Leson or the deputy sheriffs—admitted telling her. What else were they not admitting? What about the inconsistencies in Rosie's story of the attack? She said she had no time to scream, but Iorlando had time to leave the room and come back with an axe. Rosie also had time to talk to Frank. She said that the killer had hit the baby with the heavy axe that many members of the jury had handled. If husky Frank had hit little Mary three blows with such a weapon, he would have smashed her skull. But Dr. Fernandez, the coroner, testified that he found only two round wounds slightly larger than a half-dollar coin. Wasn't it likelier that someone had fed Rosie the story of the assault? Someone who'd convinced her that the Jordanos must have done it? What if she had been told over and over that the Jordanos must have done it, and she, in her weakened state, had simply accepted their guilt? The Jordanos, after all, were the competition. They had thrown the Cortimiglias out of a successful business. If everyone else thought they did it, then they must be guilty, right? In her vulnerable state, if Rosie had been asked over and over again, "Did the Jordanos do it?" she could easily have become convinced that they did. And if Rosie's testimony was disregarded, all the state had was very thin circumstantial evidence, not enough to hang two good men. For Rosie, sitting behind the DA's chair, listening to her mental state denigrated at such length proved too much. About halfway through the summations, as Andrew Thalheim pointed out inconsistencies in her account of the attack, she leapt from her chair, shook her fist at him, and shrieked, "You can say what you please but before God he did it!" The tense silence of the packed courtroom immediately dissolved into chaos, the audience buzzing with exclamations of surprise and sympathy for the mother of the murdered child and of outrage against the murderers. The bailiffs quickly reestablished order, and, angrily, Fleury ordered the deputy sheriffs to escort Rosie and Charlie from the courtroom. Once again, Byrnes and his team demanded a mistrial. Such an outburst on the part of the main prosecution witness was so prejudicial, they argued, that "the effect on the jury could not be remedied." Once again, Judge Fleury denied their request but directed the jury to disregard the outburst. Bob Rivarde finished his summation for the state around midnight. By the time Judge Fleury delivered his charge to the jury, it was around 12:30 AM when the jury filed out to their deliberation room on the third floor. The courtroom settled in for what was expected to be a long night. But at 1:45 AM, the jury was back. Why so soon? Frank had reason for a surge of hope. Surely they couldn't have convicted so quickly. A guilty verdict would have taken longer. Frank and Iorlando rose. The foreman handed the clerk of court the verdict: Frank Guagliardo, alias Frank Jordano—guilty as charged. Iorlando Guagliardo, alias Iorlando Jordano—guilty without capital punishment. By finding him guilty, the jurors had sentenced Frank to hang. Death was the automatic penalty for murder. Shocked, unbelieving, Frank swayed a little on his feet as the judge thanked and dismissed the jurors. Reporters surrounded the Jordanos, pounding them with questions. All Frank could say was, _It wasn't a fair trial. She lied. Miss Rosie lied._ Bailiffs pushed the reporters out of the way, handcuffed the condemned men, and escorted them out of the courtroom. Months later, when the two prisoners again stood in front of Judge Fleury for formal sentencing, seventeen-year-old Frank spoke to the court with dignity and courage. Fighting back tears, he said, "It is true, judge, that I may hang, but I would rather go to the gallows than tell a lie in order to have an innocent man executed. If Mrs. Cortimiglia is possessed of her faculties, she knows that she is lying and I would not swap places with her today." He then made mention of some of the inconsistences in her testimony and continued, "If you think, judge, that I should hang on testimony of that character, then I am ready to die, but there is no excuse for sentencing my old father, who is all crippled up with rheumatism and is almost blind. I pray to God that Mrs. Charles Cortimiglia will come forward and tell the truth before my father dies in prison." Judge Fleury then did as he was required by the law and sentenced sixty-eight-year-old Iorlando Jordano to life in the state penitentiary and seventeen-year-old Frank Jordano to the gallows. # ≡ 12 ≡ # False Lead AS THE AUTHORITIES IN Gretna convicted the wrong men for the killing of Mary Cortimiglia, so subsequent Axeman authorities have wrongly convicted Joseph Mumfre of being the Axeman. Or at least suggested that he's the most plausible suspect for one or more of the murders. In _Ready to Hang_ (1952), Robert Tallant planted the seeds of this legend by telling the story of Mumfre's murder in Los Angeles by the widow of one of his alleged victims. And his version of events had its roots in sloppy newspaper reporting that probably grew out of the psychological need to provide a comforting solution to the Axeman crimes and resulted in conflating a Black Hand thug with a serial killer. But Mumfre's story is in fact the story of blackmail and retribution. In 1909, Joseph "Doc" Mumfre began a twenty-year sentence at the Louisiana State Penitentiary for bombing an Italian grocery in a failed extortion plan. He was sent to Angola, one of the prison farms that made up the Louisiana penitentiary system, a place so brutal that prisoners had been known to deliberately cross the "deadline" to be shotgunned by a guard. Not Doc Mumfre. He survived to be paroled in June 1915—only to become a suspect in a murder investigation five months later. Vincent Moreci, a foreman for the United Fruit Company, was gunned down on November 19, 1915, as he walked home from work. Moreci had survived an attempt on his life five years before and then gotten off after being charged with the murder of one of the suspected gunmen, George Di Martini. Paul Di Christina, suspected along with Di Martini of attempting to kill Moreci, was shot by grocer Peter Pepitone, who was then convicted of manslaughter and sentenced to twenty years in the penitentiary. The police saw it as a typical Italian vendetta. So when Moreci was killed in 1915, John Dantonio, the detective in charge of the investigation, saw it as a continuation of the vendetta. Dantonio had several suspects but not enough evidence to arrest them. So it was that Joseph Mumfre, well-known Black Hander on whom the police had been keeping an eye since his release from prison, was picked up, charged with being dangerous and suspicious, and grilled about the killing. Mumfre denied any involvement. And the witness who'd gotten a look at one of the shooters couldn't identify him, giving Dantonio no choice but to let him go. _Black Hander and alleged Axeman Joseph Mumfre._ But Mumfre couldn't stay out of trouble. In the January 1916 Democratic primary election (which in Democratic-controlled Louisiana was the only election that mattered), Sheriff Marrero was fighting off a challenge from Reform candidate Peter Leson. On the day of the election, January 25, one of Marrero's deputies picked Mumfre up in the town of Kenner, across the river, not far from Lake Pontchartrain. Armed with marked ballots, he was canvassing—some said intimidating—the Italian population on behalf of the Reform candidates against the Regular, or Ring, Democrats in general, and against Sheriff Marrero in particular. As an example par excellence of a machine Democrat, Marrero wasn't very happy about this. And, since Mumfre was also arrested for carrying a concealed revolver (a violation of the law) and was suspected of drinking in saloons (a violation of his parole), Sheriff Marrero jailed him and immediately notified the penitentiary's Board of Control. The president of the penitentiary Board of Control instructed the sheriff to hold Mumfre until he could get orders from the governor to ship him back to prison. So Marrero did everything he could to make sure Mumfre didn't get out of the parish prison. Two other men who'd been arrested with Mumfre—Angelo Albano and his cousin—posted bond and were released. Although Mumfre, too, posted bond, Sheriff Marrero refused to free him. The penitentiary's Board of Control, he said, had instructed him to hold Joseph Mumfre, and hold him he would. So it was a considerable shock when Marrero received his next instructions from the Board of Control: he was to release Joseph Mumfre immediately. But the sheriff again refused. Mumfre had violated his parole, and as far as Marrero was concerned he was going to stay in jail until he went back to the penitentiary. In the meantime, Marrero's son, District Attorney L. H. Marrero Jr., conveniently charged Mumfre with an additional offense—possessing obscene pictures. It's hard to avoid the conclusion that this was a maneuver cooked up between Marrero Jr. and Sr. to make sure that Mumfre stayed where he was. The convict was ordered held on $500 bond, but, as the _New Orleans States_ observed, "no bondsman [was] forthcoming." Perhaps the sheriff had something to do with that. Joseph Mumfre came to trial in Gretna on February 16 on the charge of carrying a concealed weapon. His attorney, Louis Gosserand, was ill with the flu, so Mumfre, not unreasonably, asked for a continuance. DA Marrero refused. The trial went on as scheduled, without the defendant's lawyer. With no attorney to advise him, Mumfre took the stand in his own defense. He admitted that he had been campaigning for Chief Leson and other Reform Democrats with marked ballots at the behest of Gosserand, who had been defending him free of charge. Moreover, he took the courtroom by surprise by admitting that he'd received his parole with the assistance of George Wesley Smith, private secretary to Governor Luther Hall; Mumfre had paid Smith for his help. Apparently, Governor Hall had granted Mumfre parole because of paid intervention by his private secretary. All the pieces fell into place, and Mumfre's connection with the Reform Democrats became clear. A member of a Reform administration had used his influence to get Mumfre paroled. At the next election, Mumfre solicited votes for the reformers, and when he was arrested, a Reform lawyer defended him. This also made sense of the apparent change of heart of the penitentiary Board of Control. The administration had obviously pressured the board not to send Mumfre back to prison. And this was the "Reform" party! Unsurprisingly, in a trial that lasted a day, Mumfre was convicted and served four months in the Jefferson parish prison for the unlawful possession of a weapon. After completing his sentence, he was shipped back to the penitentiary. But with the time off earned for good behavior, he was released again on April 21, 1918. And, less than a month later, he was under arrest again. Accused of trying to extort money from a woman, Mumfre served three months in jail—only to be arrested again in January 1919 when Detective Marullo picked him up as a "dangerous and suspicious" character suspected of sending Black Hand letters. So by the time Joseph Mumfre appeared before Judge J. J. Fogarty, the authorities of the city of New Orleans were heartily sick of him. He was an "undesirable" Italian, the kind associated with the vendetta and the Mafia. As the _States_ reported: For years Monfre [ _sic_ ] has been looked upon by the police as a member of an undesirable band of Italians in the city and on several occasions when vendetta activities gave the police trouble, Monfre [ _sic_ ] was taken into custody and questioned. He was believed to have had knowledge of these activities. But Monfre [ _sic_ ] always—except in the Graffignino [ _sic_ ] case—wriggled out of it. It was time to run him out of town. When Mumfre was brought up before Judge Fogarty, the judge looked at him and said, "It would do the city no good to send you to the parish prison for thirty days. I am going to give you twenty-four hours to leave the city. Be sure that you do so." If he stayed in the city, the judge warned, he would be arrested again. Doc Mumfre left New Orleans. If he couldn't be a Black Hander in New Orleans, there were other cities with Italian populations. He had a daughter, Lena, in California, who had been taken there by her maternal grandparents. He would try his luck there. California turned out to be not at all lucky for Joseph Mumfre. By the time murder found another Italian grocer in New Orleans, Mumfre was long gone. Nineteen-year-old Sarah Laumann woke at 3:30 on the morning of August 3—a Sunday morning—to see a man leaning over her. Instinctively, she shrieked. The intruder fled from her room. She screamed again. Her alarmed parents, woken by her cries, raced from their room to see what was wrong. They found her terrified but apparently unhurt. Only after she complained about a pain in her head several hours later did her mother realize that Sarah had been wounded. The Laumanns immediately summoned their family doctor, who found a small laceration behind Sarah's right ear, the result, it seemed, of some sort of blunt instrument. The doctor assured the worried parents that it was only a slight injury; the girl would be fine. The Laumanns didn't alert the police until midmorning. One of the first questions the responding officer asked was _Where is your axe?_ Mrs. Laumann took the policemen out to the woodshed in the back where the axe was kept. It was gone. Mr. Laumann said he'd used it only the day before and had left it on a stack of firewood in the shed. Patrolmen blanketed the neighborhood. One of the officers found the blunt axe on the grounds of the next-door Saint Francis de Sales School. Looking closely, the policeman could see no bloodstains on the blade. The Laumanns lived at 2123 Second Street, a modest house in the middle of the block between South Saratoga Street and Loyola Avenue, in the Central City neighborhood. Both of Sarah's parents had been born in Louisiana, although her grandparents were from Bavaria. Mr. Laumann was a carpenter. Sarah was the youngest of four children and the only one still living with her parents. She worked in a cigar factory. When he heard that there had been another axe attack, Superintendent Mooney went out to see for himself. The intruder had climbed onto a gallery on the side of the house and then through an open window into the dining room. Although the shutters had been locked shut, it had been a simple matter to force them open. "I felt a stinging of the left ear," she told the officers. "That is what probably awakened me. I saw a man bending over me under the mosquito bar." The young woman was considerably shaken by her encounter, but she could describe her attacker: "About 26 years old, of rather dark complexion, 5 feet 8 inches in height, about 165 pounds in weight and wearing a dark coat and pants; his shirt was white with dark stripes and he had a dark cap pulled down over his eyes." The next day the girl's parents took her in to see Dr. Landry at Charity Hospital. The doctor reassured them; Sarah had sustained a mild concussion, but her skull hadn't been fractured. He didn't, however, believe her injury had been caused by the axe, as everyone seemed to have assumed. It was a circular wound, the result, he thought, of "a thick, blunt instrument." How her assailant would have fled carrying a heavy axe was a mystery anyway. Sarah said that she hadn't seen anything in his hands when he'd been leaning over her. "So," Mrs. Laumann pointed out to a reporter, "if he had hit her with the axe, he would have had to pick it up and jump out the window with it. But when she screamed, he didn't stop for anything, but just scrambled out of the window and over the fence." Opinion was divided. Although the Laumanns and their neighbors shrugged off the Axeman theory, some investigators were convinced that the girl had been his latest victim. Others, probably those who'd never believed in the Axeman anyway, thought that this was merely a bungled burglary. On the other side of town, on Frenchman Street, the following night, a thief had broken into a house and snatched a chain and pendant from around the neck of a sixteen-year-old girl before running off. It was all speculation because there were few clues. But the lack of evidence didn't stop the newspapers from lumping the attack in with the other unsolved axe attacks. Jim Coulton was still pushing the Axeman angle hard: A _Times-Picayune_ headline declared that the attacker had been the MYSTERIOUS "AXMAN," and the article listed Sarah Laumann among the Axeman's victims: the Cortimiglias, Joseph Romano, Louis Besumer and Harriet Anna Lowe, the Maggios, Epifanio Andollina, the Sciambras, and Joseph Davi. But this was probably not the Axeman's work. The girl was barely injured, not beaten half to death. The assault had more in common with the attack on Mary Schneider a year earlier than the serial killer's other victims. And the description given by Sarah Laumann didn't match that of Joe Romano's niece. Pauline Bruno described her uncle's attacker as tall—"probably six feet"—and heavily built, while Sarah Laumann's assailant, described as about twenty-six years old, was only five foot eight and 160 or 165 pounds. There was a lull of three months before the next "Axeman" attack, but the literary history of the Axeman story has it otherwise. In _Ready to Hang_ , New Orleans writer Robert Tallant claimed that on August 10, the Axeman pried off a door panel on Elysian Fields Avenue and attacked grocer Steve Boca, slashing his head open. In pain and dripping blood, Boca managed to make his way next door to the home of his friend Frank Genusa, who summoned help. Boca had little memory of the attack and couldn't describe his attacker, but the Axeman left his calling card on the kitchen floor: the bloodstained axe. Tallant also recounts the case of druggist William Carlson, who sat up late reading on the night of September 2. Hearing a noise at his back door, he hastily picked up his revolver and nervously called out, _Who is it? What do you want?_ When he got no answer, he fired in the direction of the noise. The next morning, the police found the unmistakable sign of chisel marks on one of the door panels. Other authors have repeated these incidents in accounts of the Axeman's crimes. But they seem to have no basis in fact. No New Orleans paper carried an account of an Axeman attack on Steve Boca or of an averted attack on William Carlson. No Steve, Steven or Stephen Boca, William Carlson, or Frank or Francis Genusa are even listed in the 1919 New Orleans city directory (although in 1920, a Frank Genusa was living on Saint Philip Street, over two miles from where Tallant said Boca lived). It's inconceivable that with the city on high alert for the axe murderer, the newspapers would ignore an Axeman attack or even a near miss. The more likely explanation is that Tallant got it wrong, and the attempts on Boca and Carlson are just myths. The attack on Mike Pepitone, however, was real enough. Esther Pepitone was a heavy sleeper. And this night she especially deserved her sound slumber. She and her husband Mike had worked all day and long into the night on both Saturday and Sunday when the Sells-Floto Circus was in town. There was a small fortune to be made selling soft drinks to the crowds of delighted children and adults who walked past the Pepitones' grocery at the corner of Ulloa and Scott in midcity to the circus a block away. Sunday night it wasn't until midnight that the exhausted Mrs. Pepitone was able to fall asleep next to her husband Mike in their bedroom behind the store. When she first heard the voice, she thought she was dreaming. From a distance, she heard someone calling for help. The voice got closer and closer. Then she woke abruptly to the sound of her husband's exclamation: "Oh, Lord!" Sitting up in bed, she saw the shapes of two men slipping out of the bedroom into her children's room next door. She looked over to see her husband lying next to her covered in blood, moaning. Panicking, she shook him: "Mike! Mike, what happened?" Mike groaned in reply. On the verge of hysteria, she leapt out of bed and rushed out of the bedroom shouting for help. Her oldest child, eleven-year-old Rosie, ran outside to summon a neighbor. Around 1:20 AM on October 27, Deputy Sheriff Ben Corcoran was walking down Scott Street on his way home when Rosie ran into the street hollering for help. _My father is full of blood!_ she cried. Corcoran followed Rosie back into the house, where he met Mike's terrified and bewildered wife. "Mr. Corcoran," she said, "it looks like the Axeman was here and murdered Mike." She pointed toward the bedroom. Corcoran entered to see Mike lying unconscious on the bed, awash in his own blood. The bedroom wall was speckled crimson, blood splashed eight or ten feet high. Mike had been viciously pummeled. His skull was fractured in several places and his face beaten into an unrecognizable mess. On a chair near the bed lay the bloody weapon: a fourteen-inch iron bar with a heavy three-inch iron nut on the end. Deputy Corcoran immediately called for the Charity Hospital ambulance and reported the attack to the New Orleans police. Chief of Detectives George Long and a squad of detectives quickly arrived, with Superintendent Mooney, Captain Thomas Capo, and other detectives and senior officers not far behind. _Another dead grocer_ , must have been Mooney's first thought. _The Axeman is back_. Mrs. Pepitone told her story to the investigating officers and gave a description of the attackers: a tall, thin man and a shorter, stockier one. She had only gotten a glimpse of them as they escaped through the children's room into the backyard. Broken glass littering the dining room floor testified to the manner of entrance. The intruders had gotten in by smashing two glass panes, unlatching the window, and raising the sash. The intruders were not intent on robbery. Nothing had been stolen: receipts from the sale of drinks to the circus crowds remained safely tucked in a cupboard. There were no signs that the house had been searched for jewelry or other valuables. Mike had been discovered wearing trousers. But Mrs. Pepitone said that he'd gone to bed in only his underwear. Detectives theorized that he'd heard the sound of the break-in and was preparing to investigate when he was attacked. While policemen in the midcity neighborhood combed over the Pepitone residence and grocery, doctors at Charity tried to save Mike's life, but the thirty-five-year-old grocer was dead by 3:15 AM. He bled to death from multiple wounds on both sides of his head. In addition to multiple lacerations, his skull had been fractured twice on the left side and once on the right. Any one of those blows would probably have been deadly. Whoever beat Mike Pepitone meant to kill him. If Mooney arrived thinking the Axeman had struck again, it didn't take long to change his mind. He quickly decided that the killing was another act in the vendetta between Paul Di Christina and Peter Pepitone. Although Peter Pepitone had sworn that his son Mike had nothing to do with Di Christina's murder in 1910, the police at the time had believed he was involved. Apparently, so did Di Christina's friends. After Di Christina's murder, Mike Pepitone had taken his family and business to the south, to Plaquemine Parish, perhaps to escape retribution for the killing. He'd only moved back to New Orleans about a year earlier because his business had failed. Peter Pepitone served five years in the penitentiary before being paroled in 1915. He had been terrified of retaliation, and now, it seemed, the Di Christina faction had struck back against the Pepitones. The little evidence Mooney's detectives uncovered pointed in that direction. Several Italians who'd been in the store in the week before Mike's murder raised suspicions that they had been plotting the vendetta. Investigators didn't find much else. No one was very optimistic about solving the case. The _States_ warned that the murder would probably remain "another vendetta mystery." Nevertheless, Jim Coulton at the _Times-Picayune_ fed the Axeman narrative. All three of the major papers were quick to report that Mooney was treating the case as a vendetta murder (and the _States_ was explicit that "Mooney does not connect the Pepitone murder with the recent ax butcheries"), but unlike the other two papers, the _Times-Picayune_ repeatedly emphasized the similarity of Pepitone's murder to the Axeman crimes. The subheading of the first report of the attack on Pepitone read CASE OF MIKE PEPITONE HAS POINTS OF SIMILARITY TO AX MURDERS. Coulton reported that the attack "bore some of the characteristics of an ax-man case." He pointed out that like the Axeman attacks, Pepitone owned a corner grocery, and, he noted, "The assailant escaped as in some of the ax cases that terrorized New Orleans during the last four years." In a later story, Coulton observed that "the Pepitone case is not unlike the long string of mysterious murders committed in New Orleans during the last few years. . . . During the last three years there have been nearly fifteen people murdered with an ax or other instruments." Coulton certainly could overstate the case: it was more accurate to say that there were nearly fifteen mysterious _attacks_ in the last few years. One gets the distinct impression that whatever the police said, Coulton couldn't quite give up the possibility that Mike Pepitone was a victim not of Mafia vengeance but of the Axeman. Widowed and with six children under the age of twelve, Esther Pepitone's situation was precarious. Desperate, she farmed at least some of them out to family or orphanages. Then she had a fortuitous encounter with an old friend. In January of 1921, Esther took the train out to Los Angeles to attend the wedding of her niece, Rosa. Esther's sister Jenny had been married to Angelo Albano, but Jenny had died in the 1918 flu epidemic, leaving her husband with daughter Rosa and son Dominick. Within a few years of his wife's death, Angelo had left New Orleans and moved out to Los Angeles, where he became the owner of a successful grocery. After Esther had been in L.A. for several months, Angelo proposed to his former sister-in-law. She was an attractive, good-natured young widow—only thirty-one or thirty-two—with children to support. He was a well-to-do widower with children of his own. It made sense. Romantic sensibilities didn't need to enter into consideration; if she was put off by the thirty-year age difference, it was probably outweighed by the reality that he was a successful businessman who could reunite her with her children. They married on September 2, 1921. But Joseph Mumfre would destroy any chance they had for happiness. When Albano had first arrived in California, he was in business with Mumfre, who in California went under the alias Leone J. Manfre, or sometimes M. G. Leone. Regardless of what he called himself, Mumfre and Albano had known each other for at least thirteen years. After his parole from Angola, Mumfre hung out in Albano's barroom in New Orleans, and the two men were arrested together in Kenner in January 1916. They moved out to California about the same time in 1919 and went into the real estate business together. Although their business ventures prospered, Mumfre was too unscrupulous a character to be trusted by anyone for long, and they eventually had a falling out. Albano bought out Mumfre, who then opened a drugstore in San Bernardino, about sixty miles from L.A. Angelo and Esther had been married only eight weeks when, on the morning of October 27, Angelo told his wife that he was going to the produce market three miles away. He kissed her good-bye and walked out the door of the home they shared with his elderly father and their children humming happily to himself. That was the last Esther Albano ever saw of her husband. When Angelo didn't come home, Esther reported him missing to the police. Investigators located witnesses who saw him at the market, and a bank teller at the Bank of Italy remembered him withdrawing money from his account. Then his trail disappeared. Angelo Albano had just vanished. Esther tried to carry on as best she could, sending the children to school and taking care of the home they shared at 554 East Thirty-Sixth Street. Around noon on December 5, five and a half weeks after Angelo's disappearance, the house was quiet. The younger children were at school and the eldest daughter was in the kitchen preparing dinner; Angelo's elderly father Jerome was napping in his room. Esther was busy with housework when Joseph Mumfre climbed the stairs that led into the family's rooms on the second floor and walked boldly into the house. Esther remembered Mumfre from New Orleans, she knew about the disagreements he had had with Angelo, and she knew his reputation. Looking into his unsmiling dark eyes and scarred face, before he even spoke, she felt a tremor of fear. Did he have anything to do with Angelo's disappearance? Had Mumfre tried to blackmail him and, failing, killed him? Mumfre came straight to the point. _I want money_ , he said to her in Italian. _I want $500_. _I don't have any money_ , she answered. He wasn't to be put off. _I want $500 in cash and all your jewelry_. When she hesitated, he sneered, _If you don't give me the money and your jewelry, I'll kill you like I did your husband._ Esther would later tell police that as he issued the threat, his hand went to his hip pocket where deputies later found an automatic pistol. So Angelo was dead. At least now she knew. And she knew what she had to do. _Wait here_ , she said to him. She went into her bedroom as if she were going to retrieve her valuables. But when she returned to the dining room, she carried not her jewels but a revolver. She raised it and fired. The first shot went wide. But she steadied her aim, and the second hit Mumfre. As did the next and the next. Stunned, Mumfre reeled from the shots, then stumbled out of the house, trying at the same time to draw his own pistol. Esther followed him, relentlessly firing. When she emptied the first revolver, she dropped it, ran to get another, and again pulled the trigger until the hammer clicked. Mumfre tumbled down the stairs until he lay still at the bottom, eleven bullets riddling his body. When she'd emptied the second revolver into the blackmailer, Esther let the smoking weapon drop to her side. She stood at the top of the stairway, staring down at the crumpled body while her niece Rosa ran down the street to call the police. When the L.A. police questioned Mrs. Albano about the killing, they were intrigued to hear the story of her first husband's death. Could the dead man have had something to do with that? Tipped off by an Italian informant to "Leone Manfre's" real identity, investigators contacted the New Orleans Police Department asking about his criminal record. The informant also told them that Mumfre may well have killed Esther Albano's first husband, Mike Pepitone. With an axe. Already, the story was being garbled. Because she'd shot Mumfre three times in the back, Esther's lawyers weren't able to get the charges dismissed as they'd hoped. Four months later, she stood trial for first-degree murder. Fortunately for her, she made a sympathetic defendant. The trial began in the Los Angeles Superior Court on April 7. The basic facts of the case weren't in dispute, so it was a quick trial. The prosecutor took less than three hours to lay his case in front of the jury. His witnesses included Lena Manfre, who testified that the dead man was her father, the county surgeon who performed the autopsy, several police officers who attended the scene, and Rosa Albano Casimano, who had overheard the entire incident. When its turn came, the defense called A. R. Kallmeyer of the Los Angeles Police Department Investigation Bureau, who had received Joseph Mumfre's criminal records from New Orleans and could testify to his long criminal history. A neighbor of the Albanos asserted that after Angelo Albano vanished, Mumfre told her, "Albano has a big house and plenty of money. He is being held for some of that money. His wife will be asked for it after things quiet down." But the crucial witness for the defense was Esther Pepitone Albano herself. Her lawyer had her tell the whole story in her soft Italian accent, beginning with the killing of Paul Di Christina by Peter Pepitone and including the murder of Mike Pepitone. She testified that she didn't know why Mike was killed or who the killer was, and she didn't know whether Mumfre had been involved. She didn't add that the New Orleans police suspected that she knew much more about Mike's death than she was willing to admit. She detailed Angelo Albano's business partnership with Joseph Mumfre and its dissolution, Albano's disappearance, and her suspicion that Mumfre was involved. She recounted the day Mumfre turned up at her front door, demanding money and threatening to do to her what he'd done to her husband. She had assumed, she said, that he meant Albano, not Pepitone. After two days of testimony, the jury retired to consider their verdict. They were gone for forty minutes. They found Esther Albano not guilty. Clearly, members of the jury thought that Joseph Mumfre/Leone Manfre was a man who needed killing. Robert Tallant's version of the Axeman story had its roots in the reporting of Mumfre's death at the hands of Esther Albano by the New Orleans press. The New Orleans police learned of his death when Los Angeles authorities contacted them for information about Mumfre. And the city's press immediately misreported the news. Although it was clear at the time of his death that Mike Pepitone's death was no axe murder, that fact quickly faded from public memory. The _Times-Picayune_ reported that "motive and probable solution of the killing of Mike Pepitone . . . one of the series of ax murders that stirred New Orleans two years ago, is seen by the police in the killing of Joseph Mumfre." That the New Orleans police jumped to the conclusion that Mumfre killed Pepitone makes it likely they suspected him all along, even though Mumfre had been run out of town nine months before—and even though there is no evidence Esther Albano ever claimed that Mumfre was Pepitone's killer. The assumption that Mike Pepitone was hacked to death with an axe fit together very nicely with the assumption that Mumfre murdered him, to "solve" the mystery of the Axeman. It was the _New Orleans States_ , where Andy Ojeda was the crime reporter, that sketched out what would become Tallant's Axeman tale. The story that appeared on December 15, 1921, was unambiguous: AX MURDERS SOLVED; MUMFRE IS KILLED. The article claimed that in 1919 and 1920, New Orleans was terrorized by a killer who especially targeted Italian grocers. The first axe murders—victims Crutti (misidentified as a butcher), Joseph Davi, and Tony Sciambra—took place before Mumfre was first sent to the penitentiary but stopped when Mumfre was in prison. He was paroled only for a short time in 1917 before being sent back to prison. Then Mumfre got out again in 1919. And the axe murders began again—Andollina, the Maggios, the Cortimiglia child, and Mike Pepitone. The killings stopped again, the article claimed, when Mumfre left for California. Now Mumfre was dead by the hand of Pepitone's widow, whom police had always suspected of knowing more about her husband's killer than she would admit. "Perhaps it was just a coincidence," the _States_ hedged. "Perhaps Mumfre knew nothing of these murders." But the clear implication of the story was that the Axeman had killed Pepitone, and Mrs. Albano had known that Mumfre had killed her first husband and had exacted her revenge. And Mumfre was the Axeman. This version of the Axeman murders is wrong on many counts. The first axe attacks (neither of them fatal) began with the assaults on the Cruttis and Rissettos in the fall of 1910. Joseph Davi was the first actual murder, his skull smashed in June 1911. Mumfre couldn't have assaulted any of these men because he was in Angola the entire time. The Axeman failed to strike from the time Mumfre was paroled in 1915 until he was returned to prison in 1916. Released again in April 1918— _after_ the attack on the Andollinas—he was rearrested by May 18—before the killing of Joseph and Catherine Maggio on the morning of May 23—and served three months. He probably wasn't yet out of jail when Joseph Romano was killed on August 11. And finally, Mumfre was arrested again in January 1919 and ordered out of town. If he had been in New Orleans or the surrounding areas when Mary Cortimiglia or Mike Pepitone was killed, he would have faced immediate rearrest. While it's not physically impossible for Joseph Mumfre to have killed Mike Pepitone, it's more likely that by this time he'd joined his daughter in California. The _New Orleans Item_ jumped in to try to correct the record, pointing out that Joseph Mumfre had been in the penitentiary when the Andollinas were attacked and in the parish prison when the Maggios were slaughtered. But it wasn't enough. Robert Tallant was a New Orleans native and lifelong resident who published novels and nonfiction in the 1940s and 1950s. In the later '50s, he wrote for the _New Orleans Item_. He was ten years old at the height of the Axeman scare in 1919 and may well have remembered the city's terror. He'd also no doubt heard the story told by newspapermen and policemen who'd been around thirty-five years before; Tallant thought he was the first to write it down. Almost certainly, he knew the reporters who'd most closely followed the Axeman's career and who still lived in New Orleans. Andy Ojeda worked for the _New Orleans States_ until he retired in 1949, and Jim Coulton became a lawyer and then a district attorney in the city. They died within fourteen months of each other in 1949–1950. Most likely, the scrambled version of the truth that Tallant heard from them and read in some old newspapers found its way into his _Ready to Hang_. Tallant undoubtedly relied on oral tradition for the end of his story, and he'd clearly heard it from someone who didn't remember the story very well. In "The Axman Wore Wings," Tallant writes that Mumfre was shot as he was walking down a busy street. No mention is made of Angelo Albano or any blackmail attempt. And, according to Tallant, Esther claimed that she had seen him when he murdered Mike Pepitone: "He was the Axeman. . . . He killed all those people." In this way, the Black Hand blackmailer and bomber Joseph Mumfre was forever linked with the tale of the Axeman of New Orleans. Certainly Mumfre was deeply unsavory. After all, the jury that tried Esther Albano quickly decided that he'd had it coming. But if any part of the mystery has been solved, it is that Joseph Mumfre was certainly not the Axeman. # ≡ 13 ≡ # Rosie and Saint Joseph Iorlando Jordano had taken the verdict better than his son. At sixty-eight, Iorlando was an old man who in his modest way had made a success of his life. But Frank's life was still ahead of him, or it had been. Now that life was taken away. By the mother of the toddler he adored. By a friendly neighbor lady who lied, swearing that he was a murderer, that she had seen him murder little Mary with her own eyes. Frank could scarcely believe it. Perhaps he had naively thought that since he was innocent he had nothing to fear. He'd remained optimistic throughout his trial, even to the point of what some called "nonchalance." But now he faced a hangman's rope. _It wasn't a fair trial_ , he told reporters who asked for a comment as he was handcuffed to be led back to the parish prison. _We didn't get "reasonable doubt."_ How had Frank and Iorlando been convicted? The jurors were serious men, diligent about their task. They'd listened carefully to the evidence and paid close attention as the experts contradicted each other about Rosie's reliability as a witness. The testimony of the physicians who treated her was consistent—Rosie wasn't mentally competent because of her head injuries. But the jury had to weigh that against the doctors who had examined her recently, who pronounced her "sane and responsible." The emotional testimony of the young mother weighed heavily on the twelve jurymen. In 1919, research had not yet raised doubts about the reliability of eyewitness testimony. Yes, elements of her story were unlikely and hard to believe, but there was no doubt that her daughter was dead and no doubt about her grief. Rosie might be a young and ignorant girl, but she was a mother who'd lost a daughter; what motive, the jurors might have reasonably wondered, did she have for lying? Many in Gretna thought the boy and old man were guilty, and public pressure on the jury shouldn't be underestimated. A change of venue might well have dramatically lowered the Jordanos' chances of conviction. But the case had been tried in Gretna and the jurors' names printed in the newspaper. If they acquitted men widely believed guilty of hacking a toddler to death, all their neighbors would know it. Not that members of the jury consciously convicted men they believed innocent, but knowledge of the community's verdict couldn't help but have influenced them in reaching their own. Frank had one comfort. After the verdict was announced and the stunned Jordanos were led from the courtroom, Jim Coulton of the _Times-Picayune_ followed them out and caught up to them. He put his hand on Frank's shoulder and looking straight into his eyes told him, "I believe you, Frank, and I'm going to work to help you out." Gratitude welled up in the boy. But what could the reporter do now? Coulton was outraged at the verdict. He had no doubt that all of the recent axe murders were the work of one crazed fiend. Couldn't the jury see that the Cortimiglias had been the latest victims of the Axeman? But Judge Fleury hadn't allowed jurors to hear evidence that the Cortimiglia attack resembled other Axeman crimes. Despite his generally good relationship with the New Orleans Police Department, Coulton knew that not all policemen—and by extension officers of the court—were completely trustworthy. He himself had recently been tossed into jail by corrupt police officers during his investigation of a shakedown scheme. And, as a crime reporter who'd often sat in court, surely he knew that witnesses could lie and juries could be wrong. He was convinced of it in this case and was determined to do something about it. Two days after the verdict, defense attorney William Byrnes announced he would file a motion for a new trial, swearing that he'd go all the way to the Supreme Court if he had to. By the time he filed his appeal, the judicial summer vacation was imminent and Judge Fleury refused to rule on the motion until the fall. Frank and Iorlando had to spend an oppressively humid Louisiana summer in the confines of the Jefferson parish prison waiting for the Twenty-Eighth District Court to come back into session. Since grounds for appealing the verdict were based mostly on mistakes the defense alleged that Fleury had made, it came as no surprise when, at the end of the summer, the judge denied the motion for a new trial. No doubt Byrnes expected that. He had prepared his appeal to the state supreme court, amassing over thirty bills of exception to Judge Fleury's rulings. When he filed his appeal in mid-November (just before Frank's eighteenth birthday on the twenty-third), the defense attorney was confident that he had grounds for the conviction to be overturned. The appeals process, even in a capital case, wouldn't drag out for decades. The Jordanos' case would be resolved within a year or two of the trial, two or three years even if there were grounds for an appeal to the U.S. Supreme Court. If the convictions were upheld, Iorlando would be sent to the state penitentiary; Frank's execution would be carried out promptly, in a matter of months at most. Everyone who read the newspapers would have been familiar with the mechanics of an execution. In 1910, all hangings had been moved to the state penitentiary in Baton Rouge, but eight years later, executions were shifted back to the parishes. Hanging a man locally made a better example of him, reasoned state legislators. At the same time, a ritual intended to give dignity and solemnity to the event was scrupulously observed. The night before, the condemned man was given his last meal. His jailers ordered it special from a restaurant, and he'd sit down to what would in other circumstances be a comforting meal of steak and potatoes and biscuits with butter, washed down with a bottle of wine or a pot of strong coffee. The prisoner spent his last night in his cell or sometimes, if he couldn't sleep, outside in the corridor talking to the deputies on the deathwatch. In the morning, he took a cold bath in the prison bathroom, after which he dressed in a somber suit of clothes provided by the sheriff: black trousers, black coat, black tie, white vest, white shirt, turn-down collar, and soft black slippers. Once dressed, if the condemned man was a Catholic like Frank, he'd be taken to the chapel to celebrate mass and see his family one last time. After his family was escorted out of the prison, the prison matron brought him the breakfast he had ordered the night before, probably ham and eggs, biscuits, and coffee. If he decided at the last minute that he had a taste for something else, like fried or raw oysters, jail personnel scrambled to get it for him. The prisoner spent the remainder of the morning with his priest receiving spiritual solace. He was even given a bracer of whiskey if he needed it to keep up his courage. At noon, the sheriff came for him accompanied by the fifteen men sworn as official witnesses. The sheriff read the death warrant and asked the prisoner if he had anything to say. Then with the priest at his side, the prisoner began his last walk, the death march down from the condemned gallery, out into the prison yard and up the steps of the wooden-framed gallows to stand at the trapdoor. Quickly, the black-hooded hangman bound the prisoner's hands and feet and whipped a black hood over his head. He positioned him just right over the trapdoor and adjusted the noose under his jaw. If the condemned man was lucky, he had an experienced executioner; one who knew what he was doing made the difference between a quick, painless death and an agonizingly slow one. When he had the prisoner properly positioned, the hangman called out: "Gentlemen of the Jury, I call upon you to witness that this execution is being conducted according to the high formalities of the law." Then he threw the lever that sprung the trap, sending the prisoner plunging into eternity. When the rope jerked suddenly, the body convulsed for a few seconds, then went still, swaying gently at the end of the rope. Under the circumstances, the best possible scenario was that his neck would snap instantly and the spinal cord would be severed, making death more or less instantaneous. The hangman gauged the drop by the height and weight of the condemned; if he botched it, the prisoner could dangle at the end of the rope for more than fifteen minutes, slowly strangling. For this reason, the law required that the body hang for at least twenty minutes before being cut down. The priest pronounced the last rites, standing on a box to reach up and anoint the motionless body with holy oil. After an examination, the coroner pronounced death. The body was taken down and placed in a simple wooden casket that was loaded into a horse-drawn hearse. The gates of the prison swung open, and the horse plodded forward to return the prisoner to his family. This is what Frank had to look forward to if his appeal failed, and it's hard to imagine that the scenario did not run through his head as the long, muggy summer of 1919 turned into the mild Louisiana winter. But he could do nothing except comfort his father and pray for deliverance. Had Saint Joseph, protector of Sicilians, deserted him? As the months passed, Iorlando's health deteriorated in the cold, damp jail cell. His rheumatism flared up, and he ached all the time; Frank's hands were sore from rubbing liniment on his father's back every night. Iorlando's eyesight grew dimmer, and Frank noticed that he had to speak more loudly for his father to hear him. But they clung to their faith that their innocence would free them. Frank decorated his cell with pictures of the saints and the Virgin Mary and prayed every night for vindication. They started off the new year hopefully in January 1920, when their drab prison existence was brightened by a wedding. For two years Lena had been engaged to Anthony Spera, the twenty-two-year-old brother of Frank's girlfriend Josie. Tony had joined the army when the United States entered the First World War, and he'd seen active service with the 151st Field Artillery. He'd returned home a war hero, expecting to marry his girl only to find that her father and brother were in the parish prison and might never return home. What should they do? Frank and Iorlando's appeals could take several years. And what if the appeals failed? Did they want to risk Frank being hanged before he could see them as a married couple? Lena and Tony decided not to wait. The future would take care of itself. But how could the entire family share in their union when Iorlando and Frank were imprisoned? Somebody came up with the inspired idea of holding the wedding celebration in the prison yard. All they needed was Sheriff Marrero's approval. Venal and corrupt though he was, Sheriff Marrero was not heartless. He had children of his own and had recently lost a son. Frank pointed out that if they hanged him, this was the last family celebration he'd see. The sheriff saw no harm and agreed. On Sunday afternoon, January 4, Lena Jordano and Anthony Spera were married at Saint Joseph's Church in Gretna, and after the ceremony the entire wedding party made their way from the church to the parish prison. There they found a table set up in the prison yard covered with "tempting dishes of chicken, ham, potted meats, cakes, pies and soft drinks." Sixty or so relatives and friends ate and drank, offering the bride their best wishes and congratulating the groom. Laughing and delighting in the presence of so many friends, Iorlando and Frank moved easily among the guests. Except for the presence of a couple of deputy sheriffs, no one would have guessed that the shadow of the scaffold loomed over the festivities. No one was saying what everyone was thinking: this was probably the last time they would celebrate as a family. But then Saint Joseph answered their prayers and gave them a miracle. Rosie couldn't sleep. Things had not gone well for the Cortimiglias in the months since the trial. Losing a child produces unfathomable stress; surviving the murder of one's child so strains a marriage that, in the early twenty-first century at least, most marriages don't survive it. Divorce might not have been such an easy option a hundred years ago, but for Charlie and Rosie Cortimiglia, Mary's murder was no less damaging. The pressure on their lives and their marriage exacted such a toll that Charlie wasn't able to manage his business. The grocery—earlier so successful—began to founder. Most of the specifics of the Cortimiglias' business failure and marriage breakdown are lost, and only a few details of their lives in the months immediately after the trial are known. In mid-June, about three weeks after the verdict, Rosie was somehow accosted by Lillie and Lena Jordano, who accused her of lying on the witness stand about Iorlando and Frank. A few days later, Charlie carelessly shot himself in the hand while cleaning his pistol. The injury wasn't life-threatening, but he ended up in Charity again a few weeks later when the wound became infected. At the end of August, he was sued by a wholesale company to which he owed money. The police didn't return the $134 that had been in the bedroom the night of the attack. Exactly how these particular events contributed to the decline of the Cortimiglia grocery business isn't known, but by the autumn of 1919, things had become so bad that the Cortimiglias became desperate. So desperate that Rosie found her way to Storyville. Officially, of course, the federal government had shuttered Storyville two years before. The sixteen-block district outside of the French Quarter where vice and sin had been officially tolerated had been a cornucopia of carnality, a galleria of luxurious whorehouses, saloons, and cabarets that flaunted their immorality. Suppressing Storyville meant closing the more ostentatious establishments but leaving many prostitutes in business, mostly the down-market whores, many of whom operated out of cheap one-room "cribs," standing in their doorway, provocatively advertising their wares. It was there, on November 1, in one of their periodic vice raids on the "ancient restricted district," that two detectives arrested Rosie Cortimiglia in a squalid Bienville Street rooming house with a traveling salesman named Ralph Rogers. Rosie tried to pretend that she was his wife, but the detectives recognized her. Rogers was released, as the prostitutes' customers unfailingly were. Only later, when the police realized that he'd given them a false name, that he was really a former New Orleans Police Department patrolman named Edward Hickey, was Rosie's paramour tracked down and locked up on a "dangerous and suspicious" charge. Rosie was taken to the Third Precinct, where she was charged with prostitution. Her arrest report also records that she had a venereal disease. She admitted that she met Hickey for money—several times since he had first picked her up at the corner of Canal and Dauphine Streets a month before. She didn't have a choice, she sobbed. The store was failing; she and Charlie were broke. When Chief Mooney found out about her arrest, he was inclined to be suspicious. Was this a ploy to besmirch Rosie's character and aid the Jordanos' appeal? Had someone been hired to entice Rosie into immorality in order to stain her moral character? When he questioned Hickey he uncovered no connection to the Jordanos, but he still felt sorry for Rosie and wanted to send her home. Charlie Cortimiglia was soon summoned to police headquarters, where he was informed of the charges against his wife. If he'd had any idea of what Rosie had been doing, he didn't admit it to Superintendent Mooney. Charlie raged with the predictable fury of a wronged husband and swore that he'd turn his wife out immediately: "Nothing in the world will ever make me consent to forgive her and take her back!" he declared to the police superintendent. Mooney tried to calm Charlie down and talk him around, without much success, even when he explained his hunch that some Jordano supporter had been behind this. Charlie was adamant that Rosie had to go. Rosie was mortified and contrite. She told the police department social worker that Charlie was right: "I sinned . . . my husband is right by refusing to forgive me." There was nothing for her to do, she said tearfully, but to kill herself. Afraid that she meant it, Mooney had Rosie committed to the Isolation Hospital, a relatively new public institution on Rampart Street for contagious diseases, especially prostitutes with venereal disease. There she stayed for almost a week, until Charlie softened and took her home to Gretna. Their reconciliation didn't last. Neighbors, taking pity on the couple that had endured so much misfortune, arranged a benefit to help with their money problems. But the Cortimiglias had been through too much. After only a few months, which they spent quarreling constantly, they separated. Rosie left Gretna and went to New Orleans, where she got a job at the American Can Company's tin can factory. Since that awful night in March, Rosie had been plagued with chronic headaches as a result of the head wounds she'd sustained. But now the torment from her uneasy conscience was greater. Maybe, as one newspaper later suggested, she remembered Frank Jordano's statement when he'd been sentenced to death. "I would rather go to the gallows than tell a lie in order to have an innocent man executed," the boy had said, with more sadness than anger. "If Mrs. Cortimiglia is possessed of her faculties, she knows that she is lying and I would not swap places with her today." _Rosie Cortimiglia, who accused two innocent men of one of the Axeman's worst crimes._ Maybe it was his words that gnawed at her day after day. Or maybe it was simply the knowledge that she'd been responsible for sending an old man to prison for life and a boy to the gallows. But something unsettled her, causing her to doubt her own sworn testimony. She should come forward, Rosie thought to herself, and make things right. But she was afraid. What would they do to her if she admitted she'd sworn falsely in court? Would they understand why she had done it? Would they put her in jail? Increasingly anxious and agitated, the distressed and indecisive Rosie was restless at night, fighting off pricks of conscience, the specter of the old man and teenaged boy haunting her sleep. Rosie's indecision reached an end on Monday night, February 2, 1920, when she dreamed that she lay dying, and Saint Joseph, the patron saint of Sicilians, came to her. "He raised his hand and pointed a finger at me," she recalled with awe later. "Oh, it was terrible, and yet it was beautiful." The saint spoke: "Rosie, you cannot die with that boy's life and that old man's liberty on your conscience." She woke up, tears running down her face, and determined to see the Jordanos free as soon as possible. It's not clear exactly how _Times-Picayune_ police reporter Jim Coulton was involved in all this, but it seems likely that he followed up on his promise to Frank by staying in touch with Rosie, perhaps playing some role in awakening her conscience. The _New Orleans States_ reported that after her dream, Rosie "sought out friends, and to them she unbosomed the truth." These friends immediately encouraged her to go public with her story. That, presumably, is what led to Rosie going to the _Times-Picayune_. She had met Jim Coulton during the trial, and after the verdict he had determined to get at the truth of her story, and if he was one of her encouraging friends, it is easy to see why she went to the _Times-Picayune_ office and why one of the paper's attorneys was so conveniently there to witness her statement. Coulton had promised Frank Jordano that he would do what he could to help him, and convincing Rosie to retract her accusation—which Coulton certainly believed was false—would help him the most. On Tuesday, February 3, Rosie Cortimiglia officially recanted her identification of the Jordanos as her assailants. In the office of the newspaper, she signed a statement in which she said her family had been attacked by two tall, black-haired men whose faces were hidden by red bandanas and that she didn't recognize the Jordanos "either by their faces, figures or voices as the men who killed my baby." The statement was officially witnessed by Jim Coulton and James E. Edmonds, managing editor of the _Times-Picayune_. "Oh, God," sighed Rosie after signing her statement, "I hope I can sleep now!" The next morning, word of Rosie's confession flashed through Gretna with the speed of a brushfire. Reaction was swift and intense. Those many community members who had for almost a year been firmly, incontrovertibly convinced of the Jordanos' guilt now loudly declared that Iorlando and Frank were the victims and that it was Rosie who should be jailed. Indeed, in less than twenty-four hours, popular opinion reversed itself completely. William Byrnes was ecstatic when he got the news. He had been convinced of his clients' innocence, repeatedly telling reporters that he knew the truth would eventually come out, "as surely as there is a heaven above us." He had only feared that the truth might not be known until it was too late for Frank. He told reporters that he hoped that with the cooperation of District Attorney Rivarde and the state attorney general, in light of Rosie's confession, he could ask the state supreme court for the case to be sent back down to the district court for a new trial. There, he assumed, the charges would be dropped. At first light on Wednesday morning, February 4, Frank heard John Bruno, Josie's stepbrother, yelling for him. He looked through his cell window to see John sprinting toward the prison gate, all the while shouting up at him and waving a newspaper. Frank couldn't make out what he was saying, but his throat tightened in dread. Had something happened to his mother? Or to Lena? Why else would John be making such a fuss? How will Papa stand it? he thought. With mounting anxiety, he could barely force himself to go downstairs to greet his friend in the prison yard. The words were fairly bursting out of John. _Miss Rosie has told the truth! Miss Rosie said she lied!_ He shoved the _Times-Picayune_ into Frank's hand. Bewildered, Frank opened the paper to see Rosie's face and his own face staring back at him. He gaped at the headline for a few seconds as the meaning began to sink in: ROSIE CORTIMIGLIA RETRACTS HER IDENTIFICATION. CONFESSION THAT PUT NOOSE AROUND TWO MEN DENIED. Frank was soon joined by his father, and with growing elation, he read the story to Iorlando. Shortly afterward, a call came from Mr. Byrnes, confirming the paper's good news. It wasn't long before Jordanos were fairly streaming into the Gretna jail—mother, brother, sisters, nieces, nephews, cousins, brothers-in-law, and fiancée—the entire clan returned for another party in the jail's courtyard, spontaneously rejoicing that their family would soon be made whole again. There was much talk of forgiveness; now that the shadow of death seemed lifted, no one seemed to bear Rosie a grudge. "Both of us are happy that she has told the truth," Frank told reporters. "She need not be afraid to come to see us. We will both welcome her and forgive her. She must have been out of her head, like she said." Frank's mother, Lillie, too, was magnanimous: "Of course, I forgive Rosie Cortimiglia. But she's not the one who needs forgiveness most. I even forgive those who forced her to tell that lie . . . I forgive them all. I'm going to get my boy and my husband back." There also was talk of gratitude. "It must have been the blessed Saint Joseph made this happen," Frank told Jim Coulton as he held one of his little cousins on his lap. "We must all pray to thank him for making Miss Rosie tell the truth by coming to her in a dream like he did. I was praying hard last night. It must have been right when Miss Rosie was taking it all back." The defense's contention that a vulnerable Rosie had been bullied into implicating the Jordanos is probably accurate. Rosie herself said that after her release from the hospital, she was bombarded by voices insisting that the Jordanos must be guilty. Remembering the falling out over the grocery store, she admitted, "I made up my mind to say it was the Jordanos who had committed the crime," despite her uncertainty that it was true. In all probability, young, uneducated Rosie, still recovering from a serious injury, felt scared and alone in the Jefferson Parish jail. Sheriff Marrero, District Attorney Rivarde, and her jailer, Deputy Sheriff Burgbacher, must have made it clear what they wanted her to say. She eventually complied, probably even convincing herself for a time that Frank and Iorlando were actually guilty. Everybody said so. Questioned over and over again in the hospital while drugged with painkillers, then again in the intimidating surroundings of the Gretna jail, asked the same questions—"Did the Jordanos do it? Was it Frank Jordano [who] hit you?"—Rosie wouldn't have found it hard to persuade herself that everybody was right; they must have done it. After the trial, she no longer had the constant reinforcement of the sheriff or the DA checking that she hadn't changed her story. When she moved to New Orleans, she also escaped the citizens of Gretna who were so set against the Jordanos. There, she let her doubts overtake her. Rosie knew she hadn't seen the assailants, and perhaps she realized how improbable it was that Iorlando, a kindly, crippled old man, and Frank, an ambitious, likable boy, would kill a child over an old quarrel. Perhaps fully realizing that young Frank, who had played so often and so lovingly with Mary, was going to be hanged, she recovered her doubts about his guilt, and she wanted no part in his death. In New Orleans, supporting herself, away from her husband, she found an inner strength that perhaps no sheriff or lawyer had suspected. Rosie's amended account of the attack was very similar to the story she told at trial, save that she didn't see the assailants' faces. But it remained an implausible tale: two men entered her bedroom; one handed the other an axe, and the man with the axe told her "he would have to kill the baby because the baby was crying." Rosie pleaded, but the intruder hit the baby three times, then knocked Rosie unconscious. Charlie slept through all of this but was then assaulted. Memory isn't a literal recording of the past that people play back in their heads. Memories are highly subjective. They're malleable constructs that can be created through suggestion or manipulation. When Rosie was subjected to suggestive questioning and told over and over that the Jordanos _must_ have done it, it's easy to see that in trying to remember, she could have imagined—and eventually came to believe as real—a scene in which she saw two men enter her room, even if, as she eventually acknowledged, she never convinced herself that she saw the Jordanos or heard Frank's voice. Repeated efforts on Rosie's part to remember the attack in all probability led to a false, created memory. The prosecutors' insistence on the Jordanos' guilt made the Axeman into two men in Rosie's mind, even when she had decided that Frank and Iorlando were innocent. Not everyone was rejoicing in the Jordanos' apparent salvation. Robert Rivarde was having none of it. The district attorney was convinced of their guilt and wasn't to be deterred by the defection of his star witness. He refused to cooperate with the defense attorneys and had no intention of asking the state supreme court to send the case back to the trial court. Rivarde knew full well that Rosie's admission would have no bearing on the high court's decision if he opposed the appeal because the court decided on matters of law, not matters of fact. He let it be known that if the Jordanos won their appeal and the case was remanded to the district court for retrial, he would charge Rosie with perjury for changing her original testimony. Rosie's fears of getting into serious trouble were likely to be realized. And not everybody was convinced that she was willing to go to prison to free the convicted men. In the face of the DA's resolve, Byrnes prepared to present his appeal. No doubt he knew that the state supreme court was legally required to ignore Rosie's statement. But Byrnes could hope that the justices would agree that "where human life was at stake . . . her retraction of that damaging testimony could not, in justice to the accused, be utterly ignored." He could hope that the justices' knowledge that the men were almost certainly innocent would have an effect, even if an unconscious, extralegal one. On March 6, in the marble and granite neoclassical court building on Royal Street in the Vieux Carré, the Louisiana Supreme Court heard oral arguments on the defense's motion to send the Jordano case back to the Twenty-Eighth District Court. The appellants went first. Andrew Thalheim and Archie Higgins were present, but chief counsel William Byrnes did most of the talking. The defense had amassed thirty-two bills of exception, or objections to the trial court's ruling. Not all of them could be addressed during oral arguments, but any one of them might be grounds for overturning the jury's verdict. For the defense, the most significant issues were multifold. It had been the district attorney's tactic to ask the defendants if they made certain threatening statements, warning that he would introduce witnesses who would contradict their statements, but never doing so, and so unfairly introducing unsubstantiated, prejudicial accusations to the jury. The judge refused to permit evidence that the Cortimiglia attack was an Axeman attack. The state's attorneys made prejudicial statements against the defendants during the trial. And the state had refused to turn over to the defense the statement Rosie Cortimiglia made in the Gretna jail. Oral arguments before the five justices took two days. When arguments ended on the second day, Byrnes was entitled to a rush of optimism. The state's performance hadn't been particularly strong. At one point, Rivarde declared, "It is a significant thing that since the Jordanos have been in jail there have been no axe murders." Not only had no one ever accused either of the Jordanos of being the Axeman, the accusation contradicted the prosecution's stated motive for the crime—jealousy. It made the state look a little desperate. Several of the justices had voiced concern about the issue Byrnes felt most strongly about: the prosecution's questioning of the defendants about threats they were alleged to have made, without putting on the stand witnesses who could support the allegations. One of the justices said he didn't understand why the prosecutor didn't call these witnesses, if he had them. Yes, Byrnes had reason to be cheerful. It was a long three weeks for Iorlando and Frank Jordano, but the decision when it came vindicated their lawyer's optimism. On April 5, the Louisiana Supreme Court ruled that the jury's guilty verdict was set aside and Iorlando and Frank's convictions overturned. The State of Louisiana would have to try them again. The court found that the trial judge made two reversible errors. Rosie Cortimiglia's statement to Clay Gaudet, which he had written down, should have been turned over to the defense. "As a matter of fact and law," the court wrote, "it was a piece of substantive evidence concerning questions of vital importance in the defense of two men who were on trial for their lives, and we know of nothing in the law or in reason which authorized the refusal to produce it." The court also rejected the strategy of asking the Jordanos if they had made specific threats but never producing the witnesses who—Rivarde claimed—could discredit the defendant's denials. The justices found that a jury might easily have concluded that the district attorney could have produced witnesses to contradict the accused, if he'd chosen to do so. But the high court determined that "the only reasonable presumption" is that Rivarde couldn't have done so. It was unacceptable, the court ruled, "to import into a case material facts which have not been established or attempted to be established by evidence; and . . . such action is . . . good ground for the reversal of a conviction." The court had, quite properly, ignored Rosie's recent confession. But the Jordanos were granted a retrial without it. Now the state had to retry the Jordanos. Which would be hard to do without Rosie Cortimiglia's testimony. _When can we get out of jail?_ the Jordanos quizzed their lawyers. Iorlando and Frank had been locked up for a year. Mr. Byrnes counseled patience. The state had a right to request a rehearing before the high court. But the deadline for making application to do so came and went, so the state supreme court's ruling became final, and the case was officially sent to the Twenty-Eighth District Court for retrial. Byrnes could tell his clients that their new court date might be as early as the next month. In the meantime, Byrnes had other things to do. Thomas Semmes Walmsley, Louisiana's assistant attorney general arguing on behalf of the state, had disparaged Rosie's retraction before the Louisiana Supreme Court, arguing that it couldn't be considered an affidavit because she hadn't been legally sworn in. The statement didn't even reflect Rosie's own words, he maintained. It was too well written for someone with her lack of education. Clearly the document had been composed by someone else. And she hadn't just walked into the newspaper office by herself. Someone had talked her into making these allegations, which he doubted she'd actually swear to. In order to assure that Rosie's confession was taken seriously, he needed to address these criticisms. But where was Rosie? Shortly after her startling admission in February, she'd been admitted to the Isolation Hospital with smallpox. That year, New Orleans experienced an unusually virulent outbreak of the disease, which overwhelmed the resources of the Isolation Hospital. Despite the dismaying mortality rate in the hospital's smallpox ward, Rosie recovered, her youthful prettiness permanently marred by the disease. The smallpox patients suffered from a desperate shortage of nursing staff, so she stayed on for several weeks to help. Rosie discovered that the Jordanos' convictions had been thrown out only when she finally left the hospital on April 23. She must have gotten in touch with Byrnes or Jim Coulton immediately, because the next day she repeated her statement under oath in front of a notary and a handful of witnesses. She swore that the statement she'd given the _Times-Picayune_ was absolutely true and again retracted her trial testimony identifying Frank and Iorlando as the killers. She affirmed that the statement she made to the _Times-Picayune_ had been "of her own free will" and that she hadn't been coerced or compensated in any way. The statement was in the third person, although in the _Times-Picayune_ story reporting it, Rosie was quoted as saying, "On the last night of my confinement in jail I was told by Charles F. Burgbacher, jailer, that I would be kept in prison unless I testified that the Jordanos had killed my baby. I wanted to get out and so I said the Jordanos murdered my baby." When the Jordanos were retried, Rosie would need to go back to Gretna to testify. But now she was afraid. She'd heard that Jefferson Parish was offering a fifty-dollar reward for her arrest. Sheriff Marrero denied the rumor. Still, she refused to cross the river until May when a new sheriff, one of a wave of reform candidates who'd swept Ring politicians out of office in the 1920 elections, would replace him. Byrnes, meanwhile, needed to get his clients out on bail. They'd been sitting in jail for over a year, and now that virtual proof of their innocence had emerged, it seemed outrageous that they should stay there. On May 12, in the Gretna courthouse, Byrnes asked Judge Fleury to allow them to post bond. Rosie had overcome her misgivings about entering Jefferson Parish, and Byrnes called her to the stand. _Was the testimony you gave in the defendants' trial a year ago tru_ e _?_ Byrnes asked. Rosie's new attorney, Sam Montgomery, intervened. He objected to his client answering the question, pointing out that if she incriminated herself, she could be prosecuted. He pointed to the district attorney: _Mr. Rivarde is waiting for Mrs. Cortimiglia to admit that she gave false testimony under oath_ , he said. _Then he will charge her with perjury._ Byrnes was exasperated. "If Rosie Cortimiglia is convicted of perjury by telling the truth which will save this boy from the gallows and his father from life imprisonment, I am willing to serve her jail sentence," he retorted. Despite Byrnes's gallant offer, Montgomery still refused to allow Rosie to testify. Under the circumstances, Judge Fleury had no choice but to turn down the request for bail. Although their conviction had been thrown out, the Jordanos were still under indictment for murder, and murder suspects were usually not accorded bail. Weary with disappointment, Frank and Iorlando returned to their cells. The same day, in the same courthouse, a grand jury began investigating both Rosie's charge that she had been coerced into identifying the Jordanos and the possibility that Rosie had committed perjury. When Judge Fleury issued his instructions, he "laid special stress on the crime of perjury," reported the _New Orleans States_. Again, Sam Montgomery would not let Rosie say anything. And, unsurprisingly, since the key witness would not testify, the grand jury's investigation came to nothing. Rosie's continued silence led some observers to wonder if she actually would testify at the Jordanos' retrial. Frank and Iorlando could be forgiven for wondering too. The new trial was scheduled to begin on Monday, May 17. Coincidently, that was also the day that John Parker was inaugurated as Louisiana's new governor. William Byrnes and Archie Higgins (as former and current state representatives) were up in Baton Rouge for the inauguration and didn't appear at the trial. One wonders why the trial still took place under such circumstances. In fact, the week before, the _Times-Picayune_ had speculated that the trial would have to be postponed for this very reason. At any rate, the trial was delayed for two days. When the defense attorneys failed to appear on Wednesday, however, both the judge and the DA were visibly irritated. When Judge Fleury learned that Andrew Thalheim, one of the Jordanos' lawyers, was in the courthouse, he immediately summoned him to account for the attorneys' absence. The flustered associate counsel stammered that there was some miscommunication; Byrnes and Higgins were still in Baton Rouge and hadn't been notified that the trial was due to start that day. Whatever the causes of the misunderstanding, less than a week of the criminal term remained, not enough time to postpone the trial yet again and still have the case heard before the summer break. The murder trial would have to be postponed until the next criminal term—in November. Instead of an immediate trial, likely acquittal, and imminent freedom, Frank and Iorlando faced another six months in jail. Father and son dejectedly "returned to the sweltering heat of the Gretna jail . . . for the summer." One would like to know what Frank had to say to his attorneys when they arrived home from the inauguration festivities. But as was his nature, he made the best of his situation. Life wasn't too unbearable for the Jordanos in the parish jail now anyway. Once it became clear that in all likelihood the Jordanos would eventually be released, Deputy Sheriff Burgbacher (the man Rosie blamed for coercing her into accusing Frank) handed Frank the keys to the jail and a revolver and informed him that he was now a turnkey. Frank also put his business experience to good use, keeping the jail's books for Burgbacher. He proved as good and conscientious a jailer as he had been a real estate agent, and the only time Frank and Iorlando had to spend in their cells was when they slept. Having something to occupy his time was good for Frank; still the summer dragged by as they waited for the fall and the acquittal they were convinced would send them home. Sometime during the summer and fall Frank's personal life suffered a blow when his fiancée Josie dropped him; this was the girl he'd tried to protect by fibbing to the coroner's jury. Perhaps the stress of having a jailed boyfriend proved too much for her. But another young woman (whose name we don't know), who perhaps was attracted by Frank's notoriety, quickly took Josie's place in his affections, and Frank was still determined to get married as soon as he could get out and earn some money. Finally, the November criminal court term arrived. This time, public opinion was firmly in the defendants' favor. Many in Jefferson Parish wondered why the district attorney persevered with his prosecution; without Rosie Cortimiglia's testimony, he didn't have a case. Rivarde must not have been convinced that Rosie would change her story. And he was determined that as long as Rosie would not admit to perjury under oath, he had a case. Once she resolved to take the risk, he knew, it was over. Why he couldn't concede the Jordanos' probable innocence is a mystery. The fall criminal term of the Twenty-Eighth District Court opened on November 3. Almost immediately, the trial schedule was set back because the presiding judge, Fred Middleton, dismissed the first jury pool over improprieties in the way its members had been selected. The process of summoning people for jury duty and selecting jurors had to start all over again. At best, the Jordano trial would start two weeks late. The criminal term finally began on Monday, November 15, and William Byrnes entered the courtroom with his clients, eager to go to trial. Rivarde, however, asked for a continuance. Over William Byrnes's opposition, and the visible disappointment of the defendants, Judge Middleton delayed the case for another week. He assured the defense that the trial would begin the next Monday. The Jordanos' case was now set for the second week of the term, Monday, November 22. But Judge Middleton wasn't able to keep his promise. When the next week came, Rivarde stood up before the judge and sheepishly explained that he'd need another continuance. The state had mislaid its records of the case and consequently hadn't been able to summon any witnesses. Byrnes was furious. "This is the second time [since] this case came back from the supreme court that we have been ready for trial and the state has asked continuance," he sputtered. "I believe Iorlando and Frank Jordano are innocent men. Why should they have to stay in jail waiting for the state to prepare its case when the records are lost?" Judge Middleton tried to soothe him. "If the state is not ready for trial next Friday," he promised, "I will entertain a motion for bail." Byrnes wasn't happy with it, but there was little he could do but agree. _Just a few more days_ , he must have assured his clients. Frank left the courtroom to spend another birthday—his nineteenth, on November 23—in jail. When Friday came, the attorneys were ready, but now Rosie Cortimiglia was missing. She'd not responded to the subpoena, and no one seemed to know where she was. Once again, the judge granted a continuance, this time until December 6. Frank and Iorlando couldn't even take comfort in the promised bail. Judge Middleton was no longer hearing the case. He was no longer even on the bench. District judges were not well-paid, at least compared to lawyers in private practice, and Fred Middleton had not run for reelection. His term expired on December 2, but he resigned a week early to take a potentially lucrative case defending a bootlegger who'd gotten into a shootout that had killed a deputy sheriff. The governor appointed H. N. Gautier, who'd been elected to succeed him, to fill the week remaining in Middleton's term. On Friday, November 26, Judge Middleton opened court just long enough to turn the proceedings over to Judge Gautier. Judge Gautier, who had made no assurances to the Jordanos, promptly denied bail. By now the two men must have been in despair, wondering if the State of Louisiana would ever get around to trying them again. Perhaps they were beginning to worry that their retrial would be delayed again and they'd spend another five or six months in jail until the next criminal term in May. They had even more reason now to long for release. Lena was having her first child. Iorlando and Frank wanted to be home to cradle their newest family member when he or she arrived. The morning of December 6 dawned bleak and rainy in Gretna, sheets of rain alternating with a dull drizzle that reduced the unpaved streets to a boggy mess. The temperatures were only in the fifties, but the wet and wind made it colder and more depressing. Crowds streaming toward the courthouse trudged on, many soaked to the skin, heads down into the wind, braving rain and mud. Sinister dark clouds filling the sky seemed to portend nothing good for this day. Rosie Cortimiglia quietly entered the district attorney's office down the hall from the packed courtroom where Iorlando and Frank Jordano waited for their trial to begin. She and Robert Rivarde sat in serious conversation, speaking in low tones. She spoke to him quietly and earnestly. He asked her a question; she answered with a determined shake of her head. He pressed her: _Are you certain?_ She nodded: _Yes, I am certain_. Frank and Iorlando sat at the defense table with their attorneys. Frank, conspicuous by his great size and flashy green suit, was turned around chatting with friends, confident about the outcome and anxious to be out of jail and back at work. Iorlando appeared a slight figure next to the bulk of his son. He sat quietly, his wife occasionally reaching up from where she sat behind him to touch his shoulder reassuringly. The atmosphere was different from that of the previous trial, with the hum of the crowd now sympathetic to the two men sitting at the defense table. At noon, the court crier solemnly intoned the traditional opening of court: "Oyez, oyez, oyez. The Honorable Twenty-Eighth District Court of Jefferson Parish is now in session, the Honorable H. N. Gautier presiding. All those having business before this honorable court draw nigh and ye shall be heard. God save the United States and this Honorable Court." Silence descended on the courtroom as Judge Gautier entered and took his seat. When the court crier solemnly announced "the Jordano case," Robert Rivarde stood up. Rivarde addressed the judge: "Your honor, just before court opened the material witness in this case came to my office. She informed me that if the case comes to trial again, she will reverse the testimony she gave in the original case, the testimony on which the accused were convicted. In order to save this court and many of our good citizens the discomfort of serving on a jury, I make a motion that the case be dropped." Without hesitation, Judge Gautier nol-prossed the case, dismissing the two defendants. "You can go, Frank," he said to the younger man. Relief flooded through Frank. Even though he'd fully expected to be freed, it still came as a shock. It took the old man a moment or two longer to understand what had just happened. After a shocked silence, the courtroom erupted, and friends and relatives and supporters thronged around the two men, slapping them on the back, shaking their hands, congratulating them, and wishing them well. As fast as the arthritic Iorlando could move, father and son hurried out of the courtroom and outside into the drizzle, savoring the drops of rain on their faces, enjoying the sweet thrill of freedom for the first time in almost two years. They were both in a hurry, headed for different places. Iorlando wanted to go home with his wife while Frank headed to New Orleans. He had someone he needed to thank. He left without noticing the woman who'd nearly destroyed him. Rosie Cortimiglia left the courthouse alone, headed back to New Orleans to take up her lonely life. Frank caught the ferry to New Orleans with his brother-in-law Tony Spera and a couple more of his pals. On the way, he experienced a tense confrontation that was at once dramatic yet utterly pedestrian. The meeting was no great coincidence. He should have expected it. They were both on their way to New Orleans straight from the courthouse, and the Jackson Avenue ferry was the only convenient means across the river. Nevertheless, Frank was taken by surprise when in the middle of the Mississippi River, he came face to face with the woman who'd almost cost him his life. He recovered from the shock before Rosie did and, betraying no bitterness, walked up to her and shook her hand. Rosie told him she was happy he was free. She couldn't stop the tears as she added that she did sometimes think of her dead baby. Then, overcome with grief and regret, she sank to the floor in a silent fit of weeping. Embarrassed by the tears and her unmistakable wretchedness, "That's all right. God bless you," was all Frank could manage before he escaped outside onto the deck. Later that afternoon, Frank gazed introspectively out the window of the _Times-Picayune_ office at the corner of Camp and North Streets, watching the rain fall on the leafy Lafayette Square. He knew he and his parents were worse than penniless; they were in debt. The money, grocery store, and land they had slowly and thriftily accumulated over the past thirty years had all been liquidated toward their defense. But Frank, as was his nature, was cheerful and confident. He could take care of his parents and his younger brother Louis. He'd been offered a job in New Orleans selling real estate. He was young. He was hardworking. Like the biblical figure Job, he could get it all back. He turned to look at Jim Coulton and he smiled. "Ain't it fine?" # ≡ 14 ≡ # The Final Chapter? IORLANDO'S JOY AT BEING released was crushed two months later when his daughter, Lena, died in childbirth. His heart was broken. He died four years later. But he died at home surrounded by his family, not on a prison farm. Meanwhile, Frank went back to work for Paul Dupas and soon had his own real estate company, becoming what the _Times-Picayune_ called "probably the youngest realtor in the state." After working, like so many Italians before him, fourteen hours a day, seven days a week (with an hour off on Sundays to go to church), he made his fortune in real estate development as the wetlands surrounding New Orleans were drained and the population spread inexorably over former swampland. And Frank didn't just sell real estate. He branched out into lumber, construction, mortgages, and cattle ranching. While it isn't clear whether Frank married the girl he met while in jail, we know that in 1926 he married twenty-one-year-old Linzy Hamilton, and that it was a mistake. Just a year later, he filed for divorce. All we know about Hamilton is that she ran up debts she then didn't, or couldn't, pay. Maybe an attraction driven by the notoriety of a murder accusation wasn't enough to make a good marriage. But in 1936, he married a woman a couple of years older than him named Mary Shambra, this time for life. They never had children. In 1961, Frank's huge size and years of overwork finally caught up to him; at the age of fifty-nine, he died of congestive heart failure. There is no record of what happened to Charlie and Rosie Cortimiglia after Rosie's December 1920 appearance in the Gretna courtroom. In 1924, Robert Rivarde, the DA who had been so thoroughly convinced of the Jordanos' guilt, defeated both his Jordano case opponent Andrew Thalheim and the incumbent judge Prentice Edrington in a primary election and went on to be elected judge of the Twenty-Fourth District Court at Gretna, where he served until his retirement in 1958. He and John Fleury remained friends until Rivarde's death in 1967, and the Jefferson Parish juvenile detention center is named for him. John Fleury was elected district attorney for the Twenty-Fourth Judicial District in 1924 and was reelected for three more terms. After his retirement from the DA's office in 1948, he practiced law in Gretna and was a fixture in Jefferson Parish politics until he was in his nineties. He died in Gretna in 1984 at the age of ninety-eight. William Byrnes, too, had a long career in Louisiana law and politics. He was a civil district court judge for twenty years, as well as a lecturer at Loyola University Law School and, briefly, its dean. In the 1930s, he turned down a spot on the state supreme court because of philosophical differences with Louisiana law. Police Superintendent Frank Mooney was out of office by the time the Jordanos walked free, and he left widely regarded as a failure. He'd never quite fit in with the police department, too much a disciplinarian to have been a favorite of the rank and file, his tenure marred by scandals involving corrupt and incompetent policemen and the flourishing trades of prostitution, gambling, and illegal sale of alcohol, as well as his failure to find the Axeman. He stepped down from office in December 1920, stating emphatically, and with great relief, "There'll be no more political jobs or detective work ever again for me." Mooney went back to the railroad world that he knew and understood. Ironically, the man who failed to track down the killer of Italian grocers took a position with the Standard Fruit Company, founded and run by the Sicilian immigrant Vaccaro brothers. He ran the company's railroad lines in Honduras with great success, and it was there, in La Ceiba, that he died suddenly of a heart attack on August 22, 1923. His body was returned via steamship to New Orleans where it was met by a police honor guard in full uniform. The next day, the white-gloved honor guard escorted him to his final resting place in the family vault in Metairie. What became of the Axeman, the "fiend" Mooney so fruitlessly sought? Previous accounts of the Axeman story note that he vanished after 1919, dropping out of sight as mysteriously as he had materialized, never to be heard from again. But while it's true that he disappeared from New Orleans, there is good reason to believe that he left the city only to terrorize victims elsewhere. On a mild December morning in 1920, in Alexandria, Louisiana, two hundred miles northwest of New Orleans, Rosa Spero woke abruptly at 1 AM, sensing a presence in her bedroom. Later, all she could remember was that someone hit her husband with an axe and then turned to her, raising the weapon and bringing it crashing down on her head. Then nothing. She woke at 4 AM, covered in blood, a dull throbbing pain in her head and jaw and shoulder. On the bed lay her husband Joseph, dead, drenched in blood. Her baby, twenty-month-old Josephine, was unconscious and bleeding. Her five boys were still asleep in the next room. Cradling the infant in her arms, she fled the house, calling for help. Frank Mooney would have instantly recognized the crime scene. Forty-eight-year-old Joseph and thirty-one-year-old Rosa Spero were the Italian proprietors of a grocery store at the corner of Wise and Turner Streets. Their assailant had broken in through a kitchen window. Entering the bedroom with an axe taken from the backyard and a butcher knife from the grocery, he struck Joseph Spero with the axe, breaking his jaw, then slashed the grocer's throat, slicing through his carotid artery. Spero bled out in minutes. After striking the woman and infant, the murderer abandoned the bloody axe and a railroad coupling pin (similar to the shoe pin left by the "Cleaver" when he'd broken into the Crutti home in 1910) in the bedroom. Police found the bloody butcher knife on the grocery counter. There was no evidence of robbery; hundreds of dollars in cash in the residence and store had not been touched. Mrs. Spero's cuts were not severe enough to kill her, and by Monday afternoon she could speak to the police but could tell them nothing about the assailant. That evening her little girl died without regaining consciousness. The police had no real leads. They briefly took into custody Louis Hughes, a sixty-year-old black carpenter. He'd recently done some work for Spero and had bloodstains on his trousers. But he satisfactorily explained the blood as the result of a cut on his hand, and with no other evidence against him, the investigation quickly came to a dead end. A month later in DeRidder, Louisiana, seventy miles southwest of Alexandria, early on the morning of January 14, 1921, neighbors discovered thirty-eight-year-old Sicilian grocer Giovanni "John" Orlando, hacked and bloody, along with his wife and two small children. His skull had been cracked with an axe. Rushed to the hospital, he died on the operating table. The doctor estimated that his injury occurred about 2 AM. Again, the assailant had broken in through a window and apparently hadn't taken anything. He'd left the axe, hair still sticking to the blade, in the blood-splattered bedroom. Mary Orlando, and the children who'd been sleeping in their parents' bed—Paul, age eight, and Josephine, age six—had all been badly cut but survived. Once again, no real suspects were ever identified. Once again, the police collared a black man—this time, a "half-witted Negro," known as "Fittified Sol" on account of the seizures he suffered—who was jailed while the police investigated. And, again, the investigation turned up nothing. The murder remained unsolved. Three months later, at around three o'clock in the morning on April 12, 1921, another incident occurred, this time in Lake Charles, Louisiana, fifty miles south of DeRidder. Neighbors of Marlena Scalisi were startled out of their beds by her screams for help. They hurried to the house at Opelousas and Blake Streets to find Mrs. Scalisi covered with blood, her husband Frank lying in their bed with a broken neck. Frank Scalisi, a thirty-five-year-old immigrant from Palermo, worked at the Powell Lumber Company, while his wife Marlena ran the grocery store at the front of their house. In addition to eleven-month-old Johnnie, who slept in his parents' bed, the Scalisis had four other children in the next room. The murderer had opened a dining room window and crawled into the house with an axe stolen from a backyard several blocks away. Entering the bedroom, he'd struck Frank one blow, killing him instantly. Then he'd raised his weapon against the sleeping Marlena and Johnnie. Fortunately for them, as he swung down, the blunt old axe head flew harmlessly off against the wall, and only the old wooden handle cracked against their heads. Startled out of sleep by the blow, Marlena screamed. The intruder bolted. Marlena stumbled out of bed to turn on the bedroom light. Seeing her husband dead, she snatched up the baby and darted into the adjoining room where her ten-year-old daughter Mary had woken. Marlena handed Johnnie to Mary and ran off to call for help. The Lake Charles sheriff arrived to find a scene that was very similar to the ones in Alexandria and DeRidder: a mysterious break-in, a shadowy assailant, and a dead body. This time, however, there was a witness. Marlena had woken to see only a silhouette of a man at the foot of her bed, but she was able to describe him only as "short [and] chunky." But as the intruder had run through the children's room, ten-year-old Mary had gotten a look at him. She described him as a "short and stout" black man. Was the murderer of Joseph Spero, Giovanni Orlando, and Frank Scalisi the Axeman of New Orleans? Was another killer at work? Was the same person even responsible for all three of these deaths? The modus operandi was clearly that of the New Orleans killer: a break-in into the home of an Italian grocer in the middle of the night, the use of an axe found at or near the scene to murder the grocer in his bed, little or nothing stolen, and no apparent motive. Joseph Spero's throat had been cut, as had the throats of Joe and Catherine Maggio in May 1918. Also, the killer had taken care to tread softly by wearing rubber soles, as had Joe Romano's killer. The New Orleans Axeman, however, was a white man. He'd been clearly seen by Harriet Crutti and her neighbors, and by Mary Davi; Pauline Bruno, too, told police she had the impression that her uncle's killer was white, although she could not be certain. And, based on descriptions, his height ranged from five feet six inches to perhaps six feet. Only one witness, one of the Crutti neighbors, had described him as "a short, heavy-set man." The newspaper reported that robbery was the motive for the Scalisi crime. Drawers had been pulled open and a box on top of the dresser had been opened. However, the intruder had not found the significant amount of cash tucked away in a trunk, and while he had grabbed Frank's trousers containing two dollars, he abandoned them two blocks away. Eleven dollars in the grocery's cash register had not been disturbed. The New Orleans Axeman was also known to have opened and rummaged through wardrobes. When he killed Joe Romano, he had also grabbed his trousers and wallet. But was robbery the primary motive for the attack on the Scalisis? Or was it, as some New Orleans detectives believed of the Axeman attacks, only an attempt to disguise his murders as robberies? Maybe the killer would have discovered the savings and cash register money if given enough time. But then, why try to kill the Scalisis? Did the grocer wake up, prompting the robber to strike? He had only been hit once, unusual for the Axeman. If that's what happened, why did he turn on the wife who was only woken when hit on the head? On balance, it seems more likely that the intruder chose to abandon his search for valuables, if that's what it was, in order to attack. As for Mary's identification of the intruder as a black man, could she have been mistaken? A light had been shining in her room, but the girl had only just been woken up; she only got a fleeting look at the fleeing man. Moreover, her identification appears to have become more certain overnight. On the day of the attack, the newspaper reported that Mary "said she did not see the man plainly but he seemed short and stout and to her he seemed a negro." She was much more confident the next day when she testified in front of the coroner's jury: "The man was short and stout. He was dark, a colored man." Not only are children less reliable as witnesses than adults, but studies demonstrate that memories can be unconsciously influenced by others, who can make witnesses more positive in their identification of a suspect, regardless of whether the identification is correct. Mary, no doubt, discussed her experience with her family before she gave testimony, and perhaps they gave her reason to (subconsciously) want to pin the blame on a black man. The sheriff's first suspect was Mary's uncle and Marlena's brother, Joe Mansueto. But Marlena and her mother firmly rejected that possibility, and Mary insisted that the person she saw was not her uncle. As she ruminated over the incident or discussed it with her mother or grandmother before she gave evidence to the coroner's jury (and it's hardly likely that she did not), the idea that the assailant couldn't have been Uncle Joe solidified in her mind. Certainly, it would have been convenient for the culprit to have been black; that way, he couldn't have been Uncle Joe. Moreover, Mary might have been subtly conditioned in other ways to identify her father's killer as a black man if she didn't get a good look at him. Witnesses often see—and remember—what they expect to see. People fit what they observe into preexisting paradigms, or schemas, of how they understand the world. The black axe murderer was a common racist stereotype at the time. That's why, in many of the Axeman cases, a black man was among the first to be detained for the crime. Given assumptions about black criminality, it wouldn't be surprising for the girl to subconsciously assume a black man was responsible for breaking into the grocery and killing her father. If the Scalisi killer was, indeed, the Axeman of New Orleans, it is problematic that both mother and daughter described him as short; almost all the New Orleans witnesses had stated otherwise. But witnesses make mistakes. And Mary's characterization of the man could have been influenced by her mother's description. Waking up in pain and fright, Mrs. Scalisi got a fleeting glimpse of a silhouette in the room and decided she'd seen a short, pudgy man. Mary's memory of the intruder was conceivably influenced by her mother's. Or perhaps the girl was right and the killer was a stocky black man. It could have been nothing more than a botched or brutal robbery. Such crimes weren't unknown. Around the same time, a series of robberies and axe or hammer murders of shopkeepers—some Italian—took place in Birmingham, Alabama, which would end only with the arrest and conviction of five black suspects. However, the axe murder of an Italian grocer amid little evidence of robbery in much the same way as the grocers in New Orleans seems too remarkable a coincidence to ignore. Of course, it's possible that the Scalisi murder was the work of a would-be African American burglar, and the Spero and Orlando killings were Axeman murders. Or, perhaps, Mary Scalisi was mistaken and the Axeman was responsible for all three of these crimes. The possibility that the Axeman continued killing after leaving New Orleans must be balanced against the possible but improbable scenario of someone _else_ with no apparent motive preying on Italian grocers with an axe. Did anyone in New Orleans see a connection between these killings and the Axeman murders? The only evidence of an effort to connect the murders is a single line in the _New Orleans States_ that referred to the Spero killing as "an ax murder identical in detail with those which stirred the state more than a year ago." The Scalisi murder, too, remained unsolved. So who, then, was the Axeman? He has never been identified and is unlikely to be given a name now. Had the murders occurred today, sophisticated investigative techniques—better fingerprinting, DNA analysis, forensic psychology—would have given Dantonio and Mooney tools to make it easier to identify and track him down. Criminal profilers would provide more sophisticated insights into a killer who struck with no apparent motive and would suggest tactics for flushing him out. The police could release information about how the murderer would act and his likely personality. Profilers also could advise them on how to turn up the killer's stress level to make him nervous enough to exhibit unusual behavior such as becoming obsessed with the investigation, trying to change his appearance, drinking too much, having trouble sleeping, or planning to leave town. This behavior might, in turn, alert his family and/or coworkers that something was wrong. Of course, there's still no guarantee that modern methods would have yielded success; too many serial killers still go uncaught. Even if the Axeman cannot be identified, however, the modern study of serial killers provides insight into what kind of person he probably was. Without benefit of real forensic evidence and detailed crime scene evidence, conclusions can only be provisional and sketchy, but it's possible to construct a likely profile of the Axeman. From witness descriptions (and with all the caveats that go with these), he appears to have been a white working-class male in his mid-to-late thirties when he began his attacks in 1910–1911. That he was familiar with burglary tools and had the confidence to break in armed with only an axe or cleaver when grocers routinely kept handguns next to their beds suggests that he was an old hand at breaking and entering. As with many sociopaths, he may well have had a history of other petty crimes. Likely uneducated, the Axeman was probably a laborer of some sort, who could move freely all over the city without attracting attention and easily enter a shop to scout the premises. He'd choose a familiar weapon. Perhaps he started with a butcher's cleaver because he worked in the meat-packing industry. But leaving a murder scene with a bloody cleaver was risky, and having to steal a new instrument every time posed its own risks. Maybe he switched to axes because he knew he could reasonably expect to find one at any grocer's residence. And an axe was not only a more convenient weapon, it was a more lethal one. Superficially, the killer would have seemed normal, except perhaps that he'd have a history of run-ins with the law. Anyone who got to know him well, however, would realize there was something slightly "off" about him. Chief Reynolds and Superintendent Mooney both thought he was a drug fiend of some sort; they were probably right. His early thefts might have been to support a morphine or cocaine habit. And he likely carried out his crimes under the influence of some drug, which would explain his odd behavior at the Crutti crime. Serial killers crave power and control. They need it to give them a sense of superiority, especially over the people they're convinced have wronged them. It isn't unusual for psychopathic killers to target a specific group of people who represent an injustice to be redressed, a humiliation to be avenged. An inadequate personality, the Axeman would have blamed the failures in his own life on his victims. The man was his primary target; sometimes a wife or child just happened to be there; they were collateral damage. The Axeman was a coward who could only face his victims as they slept. He was a predator who could only prey on the helpless. Why Italian grocers? Perhaps because these not-quite-white foreigners and small businessmen had the temerity to be more successful than he was. Maybe he resented their growing success as Italian corner groceries spread over the city. Could he have come from a family whose own business venture had failed? Alternatively, maybe he'd been caught and jailed for breaking into an Italian-owned grocery. Perhaps an Italian grocer had been instrumental in sending him to jail for an earlier crime. Or perhaps an Italian grocer had thrown him out of his store, publicly humiliating him. Whatever the reason, he hated them as a group, not as individuals; he was unlikely to have known any of his victims personally. But by standing over his sleeping prey in the night, holding an axe, he could feel God-like, with the power of life or death. Bringing down the axe, crushing the skull, seeing blood spray the walls, made the pitiless psychopath feel like he'd conquered his enemy and avenged his humiliation, real or imagined. Even when his victim was a blameless toddler, killing gave his life a purpose and a sense of success. The extensive press coverage of his crimes would have added to the Axeman's sense of accomplishment; it would have been the only thing of note he would ever do. The mystery of the Axeman's identity probably will never be solved, but in some ways it's appropriate that he remains a faceless phantom. He was certainly a nobody, a loser, a nonentity with no sense of self who felt that he had no control over anything, not even his own life. He may not always remain unknown, however. His path led from New Orleans to Alexandria and then to western Louisiana. Perhaps he continued moving west on to Texas or beyond. Or turned north toward Shreveport, Louisiana. The killing wouldn't have stopped. Perhaps in some obscure small-town newspaper there's a story of an intruder caught fleeing an Italian grocery in the middle of the night after attacking the proprietor and his wife, or a tale of an Italian grocer who shot a man trying to crawl through his window with an axe. That would be the Axeman. The Axeman of New Orleans may yet be discovered. # Acknowledgments This book could not have been written without the aid of many people to whom I am hugely indebted and immensely grateful. First, I must thank all of the archivists and librarians who made this project possible. At the top of this list is the staff of the Louisiana Division/City Archives at the New Orleans Public Library—Nancy Aloisio, Charlie Brown, Christina Bryant, Stephen Kuehling, Yvonne Loiselle, Maya Lopez, Greg Osborn, and Cheryl Picou. Special thanks are owed to Irene Wainwright, former head of the Louisiana Division, who patiently and courteously endured years of what must have seemed like endless requests for information. I also want to thank the staff of the Louisiana State Archives, particularly reference librarian John Fowler, who tracked down a thesis on the Axeman for me. Daniel Hammer of the Williams Research Center at the Historic New Orleans Collection provided me with a copy of "The Mysterious Axman's Jazz." Gratitude is due, too, to Tara Laver and the staff at Louisiana State University's Special Collections; Florence Jumonville at the University of New Orleans' Special Collections; Connie L. Phelps at the University of New Orleans' Services Department; Ann Case at the Louisiana Research Collection at Tulane University; Michelle Riggs at the University Archives at LSU at Alexandria; and the staffs of McNeese State University's Frazar Memorial Library, Auburn University's Ralph Brown Draughon Library, and the library at Notre Dame Seminary in New Orleans. The Church of Jesus Christ of Latter Day Saints' Family History Center greatly assisted my research, and I'm especially grateful to the local staff of the Family History Library who graciously helped me on Tuesday and Thursday mornings at the LDS Church on Carter Hill Road, Montgomery, Alabama. I also made use of records in parish and county courthouses. Thanks are due to Sha Carter and Kay Gilliland at the Criminal Records division of the Calcasieu Parish Clerk of Court's office; Maria Hall at the Los Angeles Superior Court Archives and Records Center; and the staffs of the Ouachita Parish Clerk of Court's office, the Marriage License and Passport Department, Jefferson Parish General Government Center, and the Old Records and Evidence Department, Jefferson Parish, Gretna. I wouldn't have been able to complete this book without the interlibrary loan staff at Delta State University, Mississippi University for Women, and Auburn University at Montgomery. Diane Coleman was a cheerful colleague for many years. Gail Gunter continued to answer questions long after I left Mississippi. I appreciate DSU's history program and former DSU provost Ann Lotven for allowing me to retain my ties to DSU for several years after I left. Harry Laver, John C. Rodrigue, Chuck Westmoreland, and Edith Ambrose answered questions about Southern history. Roger Lane offered guidance on homicide and homicide investigation in the early twentieth century. Kenneth Gravois from the Sugar Research Station at the LSU AgCenter was an indispensable resource on the sugarcane industry in Louisiana. Many other scholars, writers, and researchers answered questions and shared their findings with me, including Patricia Cohen, David Critchley, Mike Dash, Marianne Fisher-Giorlando, Anita Guerinni, Tom Hunt, Elizabeth Loftus, Michael Newton, Katherine Ramsland, Harold Schechter, and Richard Warner. Keven McQueen of Eastern Kentucky University, too, shared with me the fruits of his own research on the Axeman; I'm particularly indebted to him for his exceptional generosity. I also benefited from Doug Casey's willingness to give me access to his years of research on early New Orleans criminal gangs. I'm grateful to Professor Elliot Leyton for taking the time to speak with me by phone. In the New Orleans Police Department, thanks are due to Lieutenant Gwendolyn M. Nolan, Captain H. M. Kouts, and NOPD historian Ruth Ashur. Dr. Carol Terry, chief medical examiner for Gwinnet County, Georgia, shared her expertise on death in all its forms. I'll never forget (much as I will try) what she told me about "biscuit brain." I greatly profited from the expertise of profiler Ralph Stone, retired Georgia Bureau of Investigation special agent; Rick Chambers, former City of Atlanta homicide detective and now chief investigator at Chambers Consulting & Security; and Lieutenant Tina Miller, retired, Atlanta Police Department. Working on the Axeman brought home to me how many lawyers I know. While I'm not sure if that's good or bad, it's been very helpful in writing this book. For their assistance in answering numerous legal questions, I want to acknowledge and thank Polly Price, Robin Hutchinson, Jean Powers, and Ben Farrow. Several people read part or all of the manuscript. Thomas Easterling and Donnie Nobles read chapters. Michael Burger read numerous drafts of the manuscript. Adam Lynde read and gave perceptive comments under the most uncomfortable conditions. Kara Bryant brought an English teacher's eye to my grammar. Corey-Jan Albert has twice now gone above and beyond the call of friendship by reading and editing a manuscript. My editor at Chicago Review Press, Yuval Taylor, improved the manuscript. Developmental editor Devon Freeny and copyeditor Mark Bast saved me from many embarrassing mistakes. All remaining errors, omissions, oversights, inaccuracies, misunderstandings, misinterpretations, and typos are solely my responsibility. A few miscellaneous thank-yous are in order: to Christal Varholdt for carrying out some freelance research, Blythe Camenson for reading my original proposal, Maureen Ogle for advice on agents and arcs, and my agent Eric Myers for taking me on. I'm especially obliged to Francine Loveless Cloud for talking to me about her grandmother, Josephine Orlando, her great-grandmother, Maria Conchetta Liggio Orlando, and her great-grandfather, Giovanni Orlando, three of the Axeman's likely victims. I owe all the people listed here for helping bring this work to fruition. It's entirely possible that I've overlooked or forgotten someone who helped in some way. If so, please forgive me. My memory is less reliable than my gratitude, which is always profound. And, finally, I want to note three very important individuals: my mother, Maxine Reynolds Davis, who rather belatedly I need to thank for her commitment to my education; Michael for supporting me (in more ways than one) while I wrote this book; and Cocktail, for his support and love, despite insisting on sitting on my lap when I was trying to write. # Notes ## Preface _I got hold of Robert Tallant's Ready to Hang:_ An earlier version of Tallant's story, apparently unknown to many who wrote about the Axeman, appeared in _Gumbo Ya-Ya: A Collection of Louisiana Folk Tales_ , published in 1945. Tallant appears not to have known about another version of the Axeman in an article published by Kendall, "Blood on the Banquette." _may well have gotten much of his information:_ Tallant does seem to have consulted the _Times-Picayune_. _Most of the available sources basically repeated:_ E.g., Everitt, _Human Monsters_ , 76–78; Jeffers, _With an Axe_ , 14–15; Lane and Greg, eds., _Encyclopedia of Serial Killers_ , 35–38; Lester, "Axeman of New Orleans," 67–75; Newton, _Encyclopedia of Unsolved Crimes_ , 274–276; Ramsland, "All About the Axeman"; Reid, _Mafia_ , 184–187; Schechter, _Serial Killer Files_ , 145–146; and Wilson and Wilson, _Killers Among Us_ , 167–171. _In a list of male serial killers:_ Hickey, _Serial Murderers_ , 131. _crime writer Jay Robert Nash argued:_ Nash, _Bloodletters and Badmen_ , 358. _city records showed that no Italian grocers:_ Newton, _Hunting Humans_ , 18. Newton revised this account in subsequent books, e.g., _Encyclopedia of Unsolved Crimes_ (2004) and _Encyclopedia of Unsolved Crimes_ , 2nd ed. (2009), but _Hunting Humans_ is what I had when I first began reading about the Axeman. _a sleeping Italian grocer named Joseph Davi:_ NOPD, _Reports of Homicide_ , vol. 14 (1911), in New Orleans Public Library; Coroner's Office, _Record Book Journals_ , vol. 7 (1911), in New Orleans Public Library. ## 1. Evil Descends _It was a moonless night:_ According to the Naval Observatory (<http://aa.usno.navy.mil/data/docs/RS_OneYear.php>, accessed August 31, 2015), the moon set just after midnight. _"There will be no gorgeous pageants": NOTP_ , March 4, 1919. _"modest, even . . . somber Carnival": NOTP_ , March 5, 1919. _They also shared the streets:_ I don't have any direct evidence that either Frank Jordano and his girlfriend or the murderer known as the Axeman of New Orleans were at Mardi Gras in 1919. But it doesn't stretch the imagination to envisage the high-spirited young couple joining in the celebrations. And the killer? Well, he had to be somewhere; he was almost certainly in the city at the time. _"Oh, Jesus! Oh, Jesus! Oh, Jesus!":_ Direct quotations of dialogue in the immediate aftermath of the attack on the Cortimiglias are taken from the Jordano trial transcript. _in the direction of Amesville:_ By 1914, the growing settlement of Amesville was officially renamed Marrero, although most people still called it by its older name. _At thirty-nine:_ 1920 US Census, Ancestry.com. _Only last year, a man:_ NOPD, _Reports of Homicide_ , February 19, 1918, in New Orleans Public Library. _"sent her back to the ward": State v. Guagliardo and Guagliardo_ , 186, in Supreme Court of Louisiana Historical Archives. _"I must see her!": NOI_ , March 11, 1919. _"one of the saddest funerals": NODS_ , March 10, 1919. ## 2. The Cleaver Details of the attack on August Crutti come from _Daily News_ , August 13, 1910; _NODP_ , August 14, 1910; _NODS_ , August 13 and 14, 1910; _NOTD_ , August 14 and 15, 1910; and _NOI_ , August 13 and 14, 1910. For a description of the attack on the Rissettos and the subsequent police investigation, see _Daily News_ , September 20–22, 1910; _NOI_ , September 20, 21, and 23, 1910; _NODP_ , September 20–24 and 27, 1910; _NOB_ , September 21, 1910; and _NOTD_ , September 21–22, 1920. For information on Conchetta and Joseph Rissetto's family background and deaths, see New Orleans Marriage Records, Ancestry.com; 1920 US Census, Ancestry.com; _NODP_ , November 25, 1910; _NODS_ , November 25, 1912; and _NOTP_ , May 29, 1940. Information on John T. Flannery comes from _State v. Flannery_ , in New Orleans Public Library; _NOI_ , August 29 and October 25, 1910; _NOTD_ , August 29–30 and October 26, 1910; and _NODP_ , August 29–30 and October 26, 1910. _August and Harriet shared the small house:_ 1910 US Census, Ancestry.com. _"drunk or crazy": NODP_ , August 14, 1910. _"some half-witted fellow":_ Ibid. _"well-known police character": Daily News_ , August 13, 1910; _NOTD_ , August 14, 1910. _"insane and irresponsible":_ Drs. Hummel and O'Hara to Hon. P. D. Chretien, _State v. Flannery_ , in New Orleans Public Library. See also _NODP_ , October 26, 1910; and _NOTD_ , October 26, 1910. _A man crept up:_ Like many Italian names, the Rissettos' name is spelled a variety of ways in the newspapers, and I have standardized the spelling. _With no children:_ The 1910 US Census (Ancestry.com) lists no one else living with the Rissettos. _"slaughtering pen": Daily News_ , September 20, 1910. _"domestic strife": Daily News_ , September 22, 1910. _"of Italian descent": NOI_ , September 20, 1910; _NODP_ , September 21, 1910. _"I am certain that burglars": NOI_ , September 23, 1910. _"a fiendish thirst for blood": Daily News_ , September 21, 1910. _"dime store novel flavor": Daily News_ , September 21, 1910. _"the trail is very warm": NOI_ , September 23, 1910. _"murder stamped on his countenance:" NODP_ , August 14, 1910. ## 3. Dagoes, Sugarcane, and Muffulettas Iorlando Guagliardo left the barest record of his existence in official documents. I know his date of birth, where he came from, the year he emigrated, whom he married, and where he owned businesses. The account in this chapter is built around those known facts, but it also attempts to flesh out his life based on the typical experiences of Sicilian immigrants into Louisiana. For Italians in Louisiana, I have relied on Adams, "Mafia Riots"; Baiamonte, "New Immigrants in the South" (MA thesis) and _Immigrants in Rural America_ ; Boneno, "Migrant to Millionaire" (PhD diss.); Edwards-Simpson, "Sicilian Immigration" (PhD diss.); Macaluso, _Italian Immigrant Families_ ; Magnaghi, "Louisiana's Italian Immigrants"; Margavio, "Reaction of the Press"; Margavio and Molyneaux, "Residential Segregation of Italians"; Margavio and Salomone, _Bread and Respect_ , "Passage, Settlement, and Occupational Characteristics," and "Economic Advantages of Familism"; Maselli and Candeloro, _Italians in New Orleans_ ; Jean Ann Scarpaci, "Immigrants in the New South" and "Italian Immigrants" (PhD diss.); Clive Webb, "Lynching of Sicilian Immigrants"; "Our Italian Fellow-Citizens," _NODP_ , October 17, 1890; "Italian Colony," _NODP_ , October 18, 1890; "Italian Immigration," _NODP_ , August 12, 1904; and "Thirteen Hundred Italian Immigrants," _NODP_ , October 17, 1907. Details of nineteenth-century New Orleans from Cable, _Lost New Orleans_ ; Jewell, _Jewell's Crescent City Illustrated_ ; and Thomas Ruys Smith, _Southern Queen_. For population in New Orleans, I have relied on Sublette, _World That Made New Orleans_ ; City of New Orleans, _Population, Total and by Race_ , <http://nutrias.org/facts/aq150_1981p.pdf>; and University of Virginia Library, Historical Census Browser. For information on the sugarcane industry and the importation of foreign workers, see Berthoff, "Southern Attitudes"; Carter, _Southern Legacy_ ; Conrad and Lucas, _White Gold_ ; J. Vincenza Scarpaci, "Labor for Louisiana's Sugar Cane"; Jean Ann Scarpaci, "Immigrants in the New South" and"Italian Immigrants" (PhD diss.); Schmitz, "Transformation of the Southern Cane Sugar Sector"; Shrugg, "Survival of the Plantation System"; Sitterson, _Sugar Country_ ; and Smalley, "Sugar Making in Louisiana." Background for the cult of Saint Joseph is found in Margavio and Salomone, _Bread and Respect_ ; McColloster, "New Light"; Plemer, "Feast of St. Joseph"; "History of the Saint Joseph Altar," CatholicCulture.org; and Warren, "Sicilian St. Joseph Altar." Because they are preserved through oral tradition, there are many variations of the Saint Joseph's Day tradition. Personal information on Iorlando Guagliardo (aka Jordano) comes from _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives; _NODS_ , March 18, 1919; and "Giolando Guagliardo aka Jordano" and "Frank Guagliardo" in Ancestry .com, One World Tree Project. For ethnic clusters in New Orleans, see Campanella, _Bienville's Dilemma_ and _Geographies of New Orleans_ ; Hintz, _Ethnic New Orleans_ ; Lewis, _New Orleans_ ; Margavio, "Reaction of the Press"; Margavio and Molyneaux, "Residential Segregation of Italians"; Maselli and Candeloro, _Italians in New Orleans_ ; and Richey and Kean, _New Orleans Book_. For descriptions of "Little Palermo," I drew on Adams, "Mafia Riots"; Gambino, _Vendetta_ ; McMain, "Behind the Yellow Fever"; and Tallant, _Romantic New Orleanians_. For the history of the muffuletta, see Tusa, _Marie's Melting Pot_. _"hard-working, money-saving race":_ Quoted in Sitterson, _Sugar Country_ , 315. _"like school girls":_ "Immigrants from Italy," _NODP_ , December 17, 1880. _"foggy cities of the North":_ Adamoli, "Letters from America," 271. _"Americans, Brazilians, West Indians":_ Quoted in Campanella, _Bienville's Dilemma_ , 169. _"are ready to start a fruit shop":_ Quoted in Jean Ann Scarpaci, "Italian Immigrants" (PhD diss.), 211. _"black dagoes":_ Gambino, _Vendetta_ , 56. _"Negroes made unabashed distinction":_ Carter, _Southern Legacy_ , 106. _"filthy paupers":_ Adams, "Mafia Riots," 15. _seat of Saint James Parish:_ According to Saint James Parish, State of Louisiana, Conveyance Record vol. 52, 1889–1893, in April 1892 "Benedict Jurdano" bought an "oyster saloon . . . consisting of assorted fruits and general merchandise" for $200. Iorlando Guagliardo used the name Benedict or Benedetto Jordano for business purposes. He is referred to as "Benedict" or "Benedetto" Jordanno in several New Orleans newspapers, and a newspaper article mentions he was in business in Saint James Parish. The 1900 Census (Ancestry.com) lists "Benedict Jordano", a fruit dealer, living in Saint James Parish with his wife Lillie and his daughters "Marywarard" and "Annie." It's likely that the names were garbled in the census, as they often were; almost certainly this refers to Jordano's daughters Mary and Anna (also known as Lena). See also _NODS_ , March 18, 1919. _"filth and . . . intolerable stench":_ Adams, "Mafia Riots," 16. _"humble, loud-voiced vendor":_ Richey and Kean, _New Orleans Book_ , 129. ## 4. The Davi Murder Details of Joe Davi's murder and the police investigation come from _NOI_ , June 27–29, 1911; _NODS_ , June 27–28 and July 2, 1911; _NODP_ , June 28–30, July 1, and August 3, 1911; _NOB_ , June 28–29, 1911; _NOTD_ , June 28–29 and July 2 and 4, 1911; Coroner's Office, _Record Book Journals_ , June 28, 1911, in New Orleans Public Library; and NOPD, _Reports of Homicide_ , June 28, 1911, in New Orleans Public Library. For murder in New Orleans, see Adler, "Murder, North and South"; Asbury, _French Quarter_ ; Metcalf, "Race Relations"; Wheeler, "Homicides and Suicides"; NOPD, _Reports of Homicide_ , in New Orleans Public Library; and NOPD, Annual Report, 1910, in New Orleans Public Library. For the development of the police force in New Orleans, see Jackson, _New Orleans in the Gilded Age_ ; Metcalf, "Race Relations"; Rousey, "Cops and Guns" and _Policing the Southern City_ ; Vyhnanek, _Unorganized Crime_ ; NOPD, Annual Reports, in New Orleans Public Library; _Rules and Regulations Governing the Police Department_ , in New Orleans Public Library; NOPD, _Fifty Years of Progress_ ; and _NODP_ , February 19, 1912. Information on Martin Berhman's administration and the Ring comes from Berhman, _Martin Behrman_ ; Deacon and Coleman, _Martin Behrman Administration_ ; Kendall, "Sixteen Years"; Reynolds, _Machine Politics_ ; Schott, "New Orleans Machine"; and Williams, "Martin Behrman." For James Reynolds as chief of police, see _NOTD_ , February 10, 1911; _NOI_ , February 10–11, 1911; and _NODP_ , February 19 and April 19, 1912. For the development of scientific police investigation, see James Morton, _Catching the Killers_ ; Ramsland, _Beating the Devil's Game_ ; Thorwald, _Century of the Detective_ ; and Wilson and Wilson, _Written in Blood_. For the tools the NOPD had to solve crimes with, see _Rules and Regulations Governing the Police Department_ , in New Orleans Public Library; _NOTD_ , February 28, 1908; _NODP_ , July 18–19, 1908; and _NOI_ , August 3, 1917. To understand why suspects confess, see Tousignant, "Why Suspects Confess." For use of the third degree, see Keedy, "The Third Degree"; Skolnick and Fyfe, _Above the Law_ ; _NOTD_ , December 9, 1907; _NOI_ , February 10, 1911; _NOTP_ , March 17, 1915; _NODS_ , August 19, 1917; and _NOTP_ , April 10, 1920. In 1919, a grand jury found that New Orleans policemen frequently relied on the third degree to "extort a confession," per _NOI_ , February 27, 1919. See also _NODS_ , February 28, 1919; and _NOTP_ , February 28, 1919. For accusations that George Long used the third degree to attempt to coerce a confession from a suspect, Andrew Whitfield, for the murder of Leopold Cordova, see _NOTP_ , April 16–18, 1920. Details of John Dantonio's career are found in _NODP_ , October 21, 1900, and September 13, 1910; _NODS_ , July 13, 1911; _NOTP_ , March 28 and April 2, 1920; and NOPD, Annual Reports, especially for 1896 and 1902, in New Orleans Public Library. For the relationship between police and newspapermen, see Ramelli, "Scenes in the 'Underworld'"; Kendall, "Old-Time New Orleans"; and _NODS_ , August, 3, 1917. For Andy Ojeda, see _NOTP_ , April 12, 1950. For a description of police headquarters, see Kendall, "Notes on the Criminal History"; NOPD, _Fifty Years of Progress_ ; and sketch in _NODS_ , August, 3, 1917. For John Flannery's plea to DA Adams and his ultimate fate, see _NODS_ , July 3, 1911, and August 14, 1916; and _NOTP_ , August 15, 1916. _a city of 339,000:_ University of Virginia Library, Historical Census Browser. _Robberies and burglaries were common:_ In the period 1900–1910, the homicide rate for New Orleans was 22.2 per 100,000 residents, which made it among the most violent of southern cities and gave it a murder rate three times that of Chicago. Adler, "Murder, North and South," 301. _a hundred years later:_ "Historical Crime Data 1990–2004," City of New Orleans official website, accessed October 13, 2016, www.nola.gov/getattachment/NOPD/Crime-Data/Crime-Stats/Historic-crime-data-1990-2014.pdf. _"keep the police force entirely out": NOI_ , February 11, 1911. _"steady, sober and good young man": NODP_ , June 28, 1911. _a man near the wardrobe:_ The _New Orleans Item_ says she didn't see him rummaging in the wardrobe, that people took stuff out of it to wipe her face. _NOI_ , June 28, 1911. _"Where is your money?": NOI_ , June 29, 1911. _"Call your husband": NOI_ , June 27, 1911. _"a blow with a sharp-edged though heavy blade": NOTD_ , June 29, 1911. _"Are you sure about the man": NODP_ , June 29, 1911. _"delicate state": NODP_ , June 29 and July 6, 1911. _two Bertillon operators:_ NOPD, Annual Reports, 1897, 1898, and 1911, in New Orleans Public Library. _"All he needs is to use his eyes": NOI_ , August 3, 1917. _"motive is the clew [sic]": NOTD_ , July 4, 1911. _"I've forced confessions":_ Quoted in Skolnick and Fyfe, _Above the Law_ , 44. _"knew the effect of moral suasion": NODP_ , February 11, 1911. _"through persuasion and without": NOTD_ , October 17, 1910. _"There is no apparent motive:" NOTD_ , June 29, 1911. _Sam Pitzo:_ Sometimes spelled Pizzo. _Philip Daguanno:_ Sometimes given as Diavani. _"beat his brains in": NOTD_ , July 2, 1911. _"The man who attacked my husband": NOTD_ , July 2, 1911. _"closely allied to the [police] department":_ Ramelli, "Scenes in the 'Underworld.'" _"I came back early": NODP_ , July 3, 1911; _NOI_ , July 3, 1911. _"I do not believe any of these jobs": NOTD_ , July 4, 1911. _named Joseph P. Davi for his father:_ The 1920 US Census (Ancestry.com) shows Joseph P. Davi living with his maternal grandparents in New Orleans. _"in very much the same manner": NODP_ , July 8, 1911. _the victims of "vengeance": NOI_ , September 20, 1910. _"fact that all the victims": NODP_ , June 29, 1911. ## 5. The Black Hand For the Mafia generally, see Critchley, _Origin of Organized Crime_ ; Dash, _First Family_ ; Dickie, _Cosa Nostra_ ; Fentress, _Rebels and Mafiosi_ ; Hess, _Mafia and Mafioso_ ; and Nelli, _Business of Crime_. For the New Orleans "Mafia," see Asbury, _French Quarter_ ; Ralph Carroll, "Mafia in New Orleans" (MA thesis); Chandler, _Brothers in Blood_ ; Dulitz, "Myth of the Mafia" (BS thesis); Kendall, "Who Killa de Chief?"; Kurtz, "Organized Crime"; and Nelli, _Business of Crime_. For the true story of the ill-fated Francisco Domingo, see _DTD_ , January 6, 1855; and Orleans Parish Coroner's Office. For Sicilian criminals in New Orleans in the 1860s, see _DTD_ , June 22 and August 15, 1861; and _New Orleans Times_ , March 19, 1869. For Esposito as merely a bandit, see Baiamonte, "'Who Killa de Chief' Revisited"; Kurtz, "Organized Crime"; Nelli, _Business of Crime_ ; and _NODP_ , July 9 and July 22, 1881, and November 9, 1881. For crime in Sicily in the nineteenth century, see table 1 in Eisner, "Long-Term Historical Trends." Details of David Hennessy's involvement with the Matrangas and Provenzanos, his murder, and the subsequent lynching of suspected Italians are from the substantial literature on the case: Baiamonte, "'Who Killa de Chief' Revisited"; Botein, "Hennessy Case"; Richard Carroll, "Impact of David Hennessey"; Coxe, "New Orleans Mafia"; Dash, _First Family_ ; Gambino, _Vendetta_ ; Hunt and Sheldon, _Deep Water_ ; Katz, "Hennessy Affair"; Kendall, "Who Killa de Chief?"; Marr, "New Orleans Mafia"; Nelli, _Business of Crime_ ; Tom Smith, _Crescent City Lynching_ s; and Wilds, _Afternoon Story_. See also numerous newspaper articles, especially _NOTD_ , May 7, 1890; _NODP_ , May 6–7, July 17–20, and October 16–19, 1890; _NODS_ , October 16–17, 1890; _NODP_ , October 17–21, 1890, and March 13–23, 1891; _NOTD_ , March 14 and 18, 1891; _Daily City Item_ , March 17, 1891; _NODS_ , March 18, 1891; _New York Tribune_ , March 18, 1891; _Mascot_ , March 21, 1891. For Black Hand crime, see Lombardo, _Black Hand_ ; Nelli, _Business of Crime_ ; and Pitkin and Cordasco, _Black Hand_. In New Orleans, see, for example, _NOTD_ , April 12, 1908; _NODP_ , September 10, 1910; and _NOTD_ , May 31, 1912. For examples of the Black Hand as an organization, see _NOI_ , May 20, 1908; and _NODP_ , July 14, 1910. For the career of Joseph Mumfre, see Convict Record of the Louisiana State Penitentiary, in Louisiana State Archives; _State v. Monfre_ [ _sic_ ], in New Orleans Public Library; _State v. Pamelia_ , Westlaw, 122 La. 207, 47 So. 508 (1908); _State v. Monfre_ [ _sic_ ], Westlaw, 122 La. 251, 47 So. 543 (1908); and _State v. Monfre_ [ _sic_ ], Westlaw, 122 La. 513, 47 So. 846 (1908). Also _NOTD_ , December 6–9, 1907; _NOI_ , December 7–8, 1907, and July 23–25, 1908; _NODS_ , July 24–25, 1908; _NOTD_ , July 24–25, 1908; and _NODP_ , July 24–25, 1908. _"It's not necessary for you": NODP_ , October 16, 1890; and _NODS_ , October 16, 1890. _"They've given it to me": NODP_ , October 16, 1890; and _NODS_ , October 16, 1890. _"was in essence the Mafia":_ Kendall, "Who Killa de Chief?" 506. _suspected by some later writers:_ In _Deep Water_ , Hunt and Sheldon argue that Joseph Macheca established the New Orleans Mafia. _"the law and the evidence": NODP_ , August 10, 1890. _"Scour the whole neighborhood": NODP_ , October 16, 1890. _"We must teach these people": NODP_ , October 19, 1890. _"Who killa de chief?":_ Quoted in Wilds, _Afternoon Story_ , 99; and Coxe, "New Orleans Mafia," 1085. _"Hang the Dago murderers!": NOTD_ , March 15, 1891. _"a form of behavior":_ Lombardo, _Black Hand_ , 122. _"part armed criminal gang":_ Dickie, _Cosa Nostra_ , 52. _"fact that all the victims were Italians": NODP_ , June 29, 1911. _"Dear Friend": NODP_ , August 3, 1911. _Joseph Mumfre was just:_ Like many immigrants' names, Mumfre's name appears in a variety of spellings: Monfre, Munfre, Monfee, Manfre. I have chosen one of the most common variations to use consistently. ## 6. The Cleaver Returns For understanding serial killers, I've relied on Brown, _Killing for Sport_ ; Carlisle, "Divided Self"; Douglas and Dodd, _Inside the Mind of BTK_ ; Douglas and Olshaker, _Cases That Haunt Us_ ; Hale, "Application of Learning Theory"; Holmes, "Psychological Profiling"; Holmes and Holmes, _Serial Murder_ ; Jenkins, "Serial Murder"; Leyton, _Hunting Humans_ ; Robert Morton, "Serial Murder"; Newton, _Century of Slaughter_ ; and Ramsland, _Inside the Minds_. I've also benefited from discussions with profiler Ralph Stone, retired Georgia Bureau of Investigation special agent and instructor in criminal justice at Columbus State University, Rick Chambers, former city of Atlanta homicide detective and now chief investigator at Chambers Consulting & Security, and Dr. Elliot Leyton, professor emeritus of anthropology at Memorial University of Newfoundland. For the aftermath of Joe Davi's murder, see _NOI_ , July 10, 1911; and _NODP_ , December 30, 1912. Details of the attack on the Andollinas are found in _NODS_ , December 22–23, 1917; _NOI_ , December 22–23, 1917; and _NOTP_ , December 23, 1917. Additional information about the Andollina family comes from 1920 US Census, Ancestry.com; WWI Draft Registration Cards, Ancestry.com; New Orleans Death Records, Ancestry.com; Louisiana State Board of Health, death certificate for Epifanio Andollina, Parish of Orleans, October 23, 1918; New Orleans Marriage Records, Ancestry.com. The story of Chief Reynolds's murder comes from _NODS_ , August 2–6, 13, 1917; _NOTP_ , August 3–8 and 13–15, 1917; _NOI_ , August 2–10 and 12–14, 1917; and "Testimony, Murder of Superintendent Reynolds" in New Orleans Public Library. Background on the development of the New Orleans railroads is found in Downey, _Illinois Central Railroad_ ; Jessup, "Golden Age"; Murray, _Illinois Central Railroad_ ; Solomon, _North American Railroads_ ; and Stover, _History of Illinois Central Railroad_. For Frank Mooney's background and selection as police superintendent, see _NODP_ , May 21, 1905; _NOI_ , February 7, 12, 1911, and August 8–9, 1917; _NODS_ , August 8, 1917; _NOTP_ , August 8–9, 1917; _NODS_ , August 23, 1923; _NOI_ , August 23, 1923; and _NOTP_ , August 24, 1923. For Mooney as head of the ICR, see _NOTP_ , June 20 and July 22, 1914, and January 22, 1916; _NODS_ , February 24, 1916; and _NOTP_ , July 16, 1916. Details on John Dantonio's departure come from _NOI_ , August 3, 1917; _NOTP_ , March 28, 1920; _NODS_ , April 1, 1920; _NOI_ , April 1, 1920; and _NOTP_ , April 2 and 20, 1920. Details of Arthur Marullo's career are from _NOI_ , October 14 and December 25, 1917, and March 11, 1918; _NOTP_ , March 11, 1918; _New Orleans States-Item_ , July 7, 1965; and _NOTP_ , July 8, 1965. Information about Vincent Miramon's murder found in _NODS_ , May 19, 23, and 25, 1917; _NOI_ , May 19–20, 22, and 25, 1917; _NOTP_ , May 20, 1917; and Coroner's Office, autopsy report, May 24, 1919, in New Orleans Public Library. Particulars on the attack on the Girard family and subsequent Sumner trial taken from _NODS_ , May 28–29, 1917; _NOI_ , May 28–29, 1917; _NOTP_ , May 29–30, 1917; _NODS_ , August 14–15, 1917; _NOTP_ , August 15 and September 2, 1917; and _State v. Sumner_ , in New Orleans Public Library. Information on the murder of Joseph and Catherine Maggio and the subsequent investigation taken from _NOI_ , May 23–26, 1918; _NODS_ , May 23–26, 28, 1918; _NOTP_ , May 24–26, 1918; NOPD, _Reports of Homicide_ , May 23, 1918, in New Orleans Public Library; and Coroner's Office, autopsy report, May 23, 1918, in New Orleans Public Library. For the homicide investigations Mooney had been involved in before the Maggio killings, see _NOI_ , January 15 and February 15, 1918; and _NOTP_ , March 20, 1918. Details of Franz Reidel's murder are from _NOI_ , October 11–12, 15–17, and 19, 1910; _NOTD_ , October 12, 16–18, 1910; _NODP_ , October 12–14 and 16–19, 1910; _NODS_ , October 12–15, 17, 1910; _State v. Rodin and Bescanon_ , in New Orleans Public Library; NOPD, _Reports of Homicide_ , October 11, 1910, in New Orleans Public Library; and Coroner's Office, autopsy report, October 11, 1910, in New Orleans Public Library. For details of the Sciambra murders, see _NODS_ , May 16, 18, 26–27, 1912; _NOI_ , May 16–18, 25, and 27, 1912; _NOTD_ , May 17–19 and 27, 1912; _NODP_ , May 17–19 and 27, 1912; Coroner's Office, autopsy report, May 16, 1912, and May 26, 1912, in New Orleans Public Library; and NOPD, _Reports of Homicide_ , May 16, 1912, and May 26, 1912, in New Orleans Public Library. For Joseph Mumfre's involvement with Vito Di Giorgio, see _NODP_ , June 12, 1908; and _NOTD_ , June 12, 1908. Also, for Di Giorgio, see Warner, "First Mafia Boss?" For Detective Obitz's murder, see _NOI_ , May 27, 1918; _NOTP_ , May 27, 1918; and _NODS_ , May 28, 1918. _"Shut up!": NOTP_ , December 23, 1917. _"It was too dark": NOI_ , December 23, 1917. _"I have no enemies": NOI_ , December 23, 1917. _"how a man could be shot": NOI_ , August 12, 1917. _"Murder! Murder!": State v. Sumner_ , in New Orleans Public Library. _"Come at once!": NOI_ , May 23, 1918. _"Do you have a father, mother":_ WWI Draft Registration Cards, Ancestry.com. _"Robbery was the motive": NODS_ , May 23, 1918. _"just like Mrs. Toney": NOTP_ , May 25, 1918. _"Just write Mrs. Toney": NODS_ , May 23, 1918. _"just write. Mrs. Tony": NOI_ , May 23, 1918. _"unusually mysterious":_ Ibid. _"the unsettled state": NODS_ , May 25, 1918. _"How could you think I could kill": NOI_ , May 26, 1918. _"Up to the time we conducted": NODS_ , May 26, 1918. _"We don't know where to start": NOTD_ , October 12, 1910. _"one of the criminal mysteries": NODS_ , October 12, 1910. _"award[ing him] the first honors": NOI_ , October 19, 1910. _"the better class of Italians": NOI_ , May 25, 1912. _"lust for blood": NOTD_ , May 17, 1912. _unclear how long:_ On May 15, 1916, the _NOTP_ noted that Henry Sciambra sold his store "a few days after [his brother's] murder," but since it gets the date of the murder wrong (saying it happened in 1914), the information may not be entirely reliable. The New Orleans City Directory records that in 1913–1914 Henry Sciambra was still at 2302 Dauphine St. and Di Giorgi on N. Rocheblave St.; Di Giorgio is not listed as owner of the grocery at 2302 Dauphine until 1915–1916. _"considerably worried because of an Italian": NODP_ , May 17, 1912. _George Musacchia:_ Sometimes appearing as Masuchia or Musachia. _"bore all the earmarks": NOI_ , May 27, 1912. ## 7. A German Spy? Particulars of the attack on Louis Besumer and Harriet Lowe, and the police investigation, found in _NODS_ , June 27–29, 1918; _NOI_ , June 27–30, 1918; _NOTP_ , June 28–30, 1918; _NODS_ , July 3, 5, 7–8, 12, 15, and 30, 1918; _NOTP_ , July 1, 3–7, 9–10, 12, and 15, 1918; and _NOI_ , July 1, 3–4, 7–8, 1918. For German spies in WWI and US counterespionage measures, see Witcover, _Sabotage at Black Tom_. For treatment of Germans in the United States during the First World War, see Peterson and Fite, _Opponents of War_ ; and Schaffer, _America in the Great War_ , chapter 2, "Controlling Dissent." For James G. Coulton's career, see _NODS_ , February 7, 1949; _NOI_ , February 7, 1949; and _NOTP_ , February 7–8, 1949. For Mooney's demotion of detectives, see _NOI_ , July 3, 1918; _NODS_ , July 7, 1918; and _NOTP_ , July 7, 1918. Details of the attack on Mary Schneider from _NODS_ , August 5–7, 1918; _NOI_ , August 5–7, 1918; and _NOTP_ , August 6–7, 1918. Additional details on the Schneider family from 1920 US Census, Ancestry.com. _named Louis Besumer:_ His name is also sometimes spelled Besemer or Bessemer. _Come around to the front:_ The New Orleans papers contained two or three different versions of Zanca knocking on the door and entering the grocery. My version is based on the report in the _NOI_ , June 30, 1918; and NOPD, _Reports of Homicide_ , September 14, 1918, in New Orleans Public Library. See also _NODS_ , June 27, 1918; _NOI_ , June 27, 1918; _NOTP_ , June 28, 1918; and _NOTP_ , May 1, 1919. _"My God, what's happened?":_ Zanca's conversation with Besumer has been reconstructed based on NOPD, _Reports of Homicide_ , in New Orleans Public Library; _NODS_ , June 27, 1918; _NOI_ , June 27, 1918; _NOTP_ , June 28, 1918; and _NOI_ , June 30, 1918. _"one of the queerest [mysteries]": NOTP_ , July 4, 1918. _Harriet Anna Lowe:_ Her name sometimes appears as Annie Harriet Lowe or Harriet Annette Lowe. Besumer referred to her as "Annie." _NOI_ , September 16, 1918. _"My husband is a German": NOI_ , June 28, 1918. _HATCHET MYSTERY: NOTP_ , June 29, 1918. _schools abolished German language instruction:_ My husband's Swedish grandfather was bullied during the war for speaking Swedish, which some, apparently, could not distinguish from German. _"proceeded to tell about the people": NOTP_ , July 4, 1918. _"a born investigator": NOI_ , July 3, 1918. _"I am a man who leaves nothing undone": NOTP_ , July 4, 1918. _"The other articles in my store": NOTP_ , July 7, 1918. _his former wife lived in Cincinnati:_ Eventually, Besumer admitted to two ex-wives. _NOTP_ , August 20, 1918. _"twitche[d] uncontrollably": NODS_ , August 25, 1918. _"I feel sure he and I are married": NOTP_ , July 7, 1918. _their memory of the details was a little foggy:_ The _Item_ reported that these attacks had occurred during the last two years. _NOI_ , June 27, 1918. _usually on moonless nights:_ According to the Naval Observatory (<http://aa.usno.navy.mil/data/docs/RS_OneYear.php>, accessed August 31, 2015). _a greater risk of being spotted:_ In addition to the evidence of fresh blood, Lowe said that she was attacked in daylight. _NODS_ , June 27, 1918. _Caring for three small children:_ The 1920 US Census (Ancestry.com) shows Edward and Mary Schneider and their four children living next to Emile and Kate Gonzales. _"Struck? Oh, no": NODS_ , August 7, 1918. _POLICE BELIEVE AX -MAN: NOTP_, August 6, 1918. _connecting the Schneider attack:_ Jim Coulton was the _Times-Picayune_ 's crime reporter and so presumably was responsible for the paper's coverage of the crimes. _"it probable that Mrs. Schneider": NOTP_ , August 6, 1918. _VICTIM OF AX -MAN NOW HAPPY: NOTP_, August 7, 1918. _Andy Ojeda took a different tack:_ Ojeda was probably covering police headquarters at this time. _NODS_ , April 1, 1950. _"It is nothing more or less": NODS_ , August 7, 1918. _AXE -MAN'S VICTIM CAN'T RECALL: NOI_, August 7, 1918. _ARMED MEN GUARD SLEEPING FAMILIES: NOI_ , August 8, 1918. _AX -VICTIM TO BE TAKEN: NOI_, August 15, 1918. ## 8. Axeman Hysteria For details of Joe Romano's murder, see _NODS_ , August 10, 1918; _NOI_ , August 10, 1918; _NOTP_ , August 11, 1918; NOPD, _Reports of Homicide_ , August 10, 1918, in New Orleans Public Library; and Coroner's Office, autopsy report, August 10, 1918, in New Orleans Public Library. Additional information about the Romano/Bruno family from the 1920 US Census, Ancestry.com. For Axeman panic, see _NOI_ , August 10–12, 1918; _NOTP_ , August 11–12, 1918; and _NODS_ , August 11 and 13, 1918. For examples of Axeman spotting, see _NOI_ , August 12 and 16, 1918; _NODS_ , August 12, 16–17, 1918; and _NOTP_ , August 12 and 17 and October 31, 1918. Accounts of Axeman-like break-ins that came to light after Joe Romano's murder come from _NOI_ , August 10 and 15, 1918; _NODS_ , August 11, 1918; and _NOTP_ , August 12 and 16, 1918. For evidence of thefts at Italian stores after the Romano murder, see _NODS_ , August 19–21, 1918; _NOI_ , August 20, 1918; and _NOTP_ , August 21–22, 1918. For examples of Axeman-like burglaries/panel robberies, see _NOI_ , September 15, 25, and 27, 1918; _NOTP_ , September 15 and 25, 1918; _NODS_ , September 15 and 23, 1918; _NODS_ , March 10, 1919; and _NOTP_ , March 16, 1919. For Piggly Wiggly and the rise of supermarket chains, see Levinson, _The Great A &P_. Harriet Lowe's account of her assault is based on details in _NOI_ , August 19, 1918; _NODS_ , August 19, 1918; and _NOTP_ , August 20, 1918. For the bloodstains found in Besumer's trunk, see _NODS_ , August 5, 1918; _NOI_ , August 21, 1918; _NODS_ , May 1, 1919; and _NOTP_ , May 1, 1919. My account of Besumer's response to Lowe's charges is based on information in _NODS_ , August 20, 1918; _NOI_ , August 20, 1918; and _NOTP_ , August 20–21, 1918. Information on Besumer's arrest and murder charge found in _NODS_ , August 19–20, 23, and 25, 1918; _NOI_ , August 19–21 and 23, 1918; _NOTP_ , August 20–21 and 24, 1918; _NODS_ , September 16, 24, and 27, 1918; _NOTP_ , September 17, 1918; _NOI_ , September 17 and 25, 1918; _NOI_ , November 24 and December 6, 1918. For details of Harriet Lowe's operation and death, see _NOI_ , August 21, 1918; _NOTP_ , August 21, 1918; _NODS_ , September 6–7, 14, and 16, 1918; _NOI_ , September 6, 16–17, 1918; _NOTP_ , September 6–8, 14, and 17–18, 1918; NOPD, _Reports of Homicide_ , September 16, 1918, in New Orleans Public Library; and Coroner's Office, autopsy report, May 24, 1919, in New Orleans Public Library. To understand the dynamics of domestic abuse, see Bancroft, _Why Does He Do That?_ ; and Barnett, "Why Battered Women Do Not Leave," parts 1 and 2. For evidence of Mooney's views on the Axeman, see _NOI_ , August 10, 1918; _NODS_ , August 11, 1918; _NOTP_ , August 12, 1918; _NODS_ , March 11, 1919; and _NOTP_ , March 11 and 16, 1919. My understanding of the workings of memory is based on Loftus, "Make-Believe Memories" and "Malleability of Human Memory"; and Schacter, _Seven Sins of Memory_. Information about the Spanish flu in New Orleans comes from _NODS_ , September 16 and 19, 1918; _NOI_ , September 18, 1918; _NOTP_ , September 18, 1918; _NODS_ , October 8–9, 20, and 29, 1918; and _NOTP_ , October 10, 16, 31, 1918. _"Something has happened": NOTP_ , August 11, 1918. _"He was awfully light on his feet": NOI_ , August 10, 1918. _"The skull was not shattered":_ NOPD, _Reports of Homicide_ , August 10, 1918, in New Orleans Public Library. _"I believe it is criminal": NODS_ , August 16, 1918. _"this blood-mad creature": NOI_ , August 15, 1918. _"badly frightened negro woman": NOTP_ , August 17, 1918. _"scathing criticism": NOI_ , August 12, 1918. _"exhaustive investigation": NODS_ , August 11, 1918. _"thinly settled sections": NOTP_ , August 17, 1918. _Jim Coulton sought out:_ Jim Coulton was the _Times-Picayune_ 's crime reporter, so presumably he conducted the interview. _"This is very probably the man": NOTP_ , August 13, 1918. _"The homes of all the victims":_ Ibid. _"Take this for the gospel": NOI_ , August 10, 1918. _"Attention Mr. Mooney and All Citizens": NOTP_ , August 23, 1918. _"I met Mr. Besumer in Jacksonville": NOTP_ , August 20, 1918. _"You'll get your money and more": NODS_ , August 20, 1918. _"Annie, you are going to make a fire": NOI_ , August 19, 1918. _"As I lay in bed": NOTP_ , August 20, 1918. _"gave me some trouble":_ Ibid. _"peculiar disposition": NODS_ , August 20, 1918. _"I feel that I am going to die": NOTP_ , September 6, 1918. A slightly different version is found in _NOI_ , September 6, 1918. _Writing over thirty years later:_ Robert Tallant mentions the story of Besumer's and Lowe's violent bickering in _Ready to Hang_. _when she said she couldn't remember:_ The black-haired man she recalled bending over her was probably John Zanca. ## 9. The Mysterious Axeman's Jazz Information about Iorlando Guagliardo (aka Jordano) and his family comes from _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives; "Orlando Guagliardo Jordano" and "Frank Guagliardo" in Ancestry.com, One World Tree Project; 1920 US Census, Ancestry.com; New Orleans Death Records, Ancestry .com; _NODS_ , March 18, 1919; and _NOI_ , March 18, 1919. For Charlie Cortimiglia's background, see _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives; _NOI_ , March 10 and 12, 1919; _NOTP_ , March 10, 1919; _NODS_ , March 10, 1919, and February 4, 1920. Description of Rosie Cortimiglia based on information in _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives; and _NOTP_ , February 4, 1920. For a description of Frank Jordano, see _NODS_ , March 15, 1919; _NOTP_ , March 15, 1919; _NOI_ , March 18, 1919; _NOTP_ , February 5, 1920; _NOTP_ , June 17, 1923, and March 14, 1926; and _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. Details of the conflict between the Cortimiglias and Jordanos over the grocery business come from _NOTP_ , March 10, 1919; _NOI_ , March 10 and 15, 1919; _NODS_ , March 15, 1919; and _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. Details of the events of the night of Saturday, March 8, 1919, and the following Sunday morning come from _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. Manny Fink's account of hearing the Axeman taken from his testimony in _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. For Sheriff Louis Marrero Sr.'s history, see Chambers, _History of Louisiana_ ; Curry, _Gretna_ ; Fortier, _Louisiana_ ; Goodspeed, _Biographical and Historical Memoirs_ ; _NOI_ , May 2, 1912; _NODS_ , October 25, 1917; _NOI_ , January 15, 1920; _NOI_ , February 26–27, 1921; and _NOTP_ , February 27, 1921. For details of the initial investigation of the Cortimiglia crime scene, see _NOI_ , March 10–11, 1919; _NOTP_ , March 10–11, 1919; and _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. For details of the Vicari break-in, see _NOI_ , March 11, 1919; _NODS_ , March 11, 1919; and _NOTP_ , March 11, 1919. Also _State v. Guagliardo and Guagliardo_ , 587–589, in Supreme Court of Louisiana Historical Archives. Information about Dr. Fernandez's coroner's jury comes from _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives; _NOTP_ , March 10–11, 1919; _NODS_ , March 15, 1919; _NOI_ , March 17, 1919; _NODS_ , April 16, 1919; and _NOTP_ , April 17, 1919. For Frank's testimony before the coroner's jury, see _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives; _NOI_ , March 10, 1919; _NOTP_ , March 10, 1919; and _NODS_ , March 15, 1919. For details of the Cortimiglias' recovery in Charity Hospital, see _NOI_ , March 10–11, 1919; _NOTP_ , March 10–12, 1919; _NODS_ , March 11–13, and 15–16, 1919; and _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. Details on interrogation of suspect Charles Anderson found in _NODS_ , March 12, 1919; _NOTP_ , March 12, 1919; _NOI_ , March 12–14, 1919; and _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. For accusation against Frank and Iorlando, see _NODS_ , March 15–19, 1919; _NOTP_ , March 15–16 and 18, 1919; _NOI_ , March 15–19, 1918; and _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. Details of J. J. Davilla's career from _NOTP_ , July 30, 1916, March 10, 1918, and March 18 and 20, 1919; _NODS_ , March 20, 1919; _NOTP_ , March 23, 1919; _NODS_ , May 10, 1919; _NOTP_ , May 11, 1919; _NOTP_ , July 20, 1947; and _NOTP_ , May 26, 1957. For Saint Joseph's Day in New Orleans, see Chupa, "St. Joseph's Day Altars"; McColloster, "New Light"; Orso and Kaveski, "Undisclosed Aspects"; Plemer, "Feast of St. Joseph"; and Tallant, Dreyer, and Saxon, _Gumbo Ya-Ya_ , 6. For Saint Joseph's Day 1919, see _NODS_ , March 16 and 18–19, 1919; _NOTP_ , March 18–19, 1919; _Herald_ , March 20, 1919; and _NOTP_ , July 20, 1947. For Robert Rivarde's background, see Louisiana District Judges, _Biographies_ ; _NODP_ , February 19, 1912; _NOI_ , December 4 and 7, 1916; _NODS_ , December 7–8 and 17, 1916; _NOI_ , January 30, 1967; _NOTP_ , January 31, 1967; _States-Item_ , January 30 and 31, 1967. For Rosie's arrest as a material witness, see _NOI_ , March 28, 1919; _NOTP_ , March 28, 1919; _NODS_ , March 28–29, 1919; _NODS_ , February 4, 1920; _NOI_ , February 4, 1920; _NOTP_ , April 26, 1920; and _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. _"all crippled up": State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. _"it would just kill her": State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. _"a pronounced Italian type": NOTP_ , February 4, 1920. _"of a manly bearing": NOTP_ , March 15, 1919. _They set March 19, Saint Joseph's Day: State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. _"Mama, have you got anything": State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. _"nonfeasance, malfeasance, extortion": NODP_ , May 25, 1912. _Santo Vicari:_ Sometimes spelled Vaccarri or Viccari. _"the murder was the deed": NOTP_ , March 10, 1919. _"Do you know who attacked you": NOTP_ , March 12, 1919. _"Who hit you?": State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. _"That son-of-a-bitch Cortimiglia":_ Ibid. _"Frank Jordano and the old man":_ Ibid. _"They've arrested Frank Jordano": NODS_ , March 15, 1919. _"How are you feeling?":_ Ibid.; _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. _"Editor of the Times-Picayune": NOTP_ , March 16, 1919. _The real killer didn't send:_ Newspapers refer to two letters to Mooney from an anonymous "student of criminology" about the series of axe murders ( _NOI_ , August 8, 1918; _NOTP_ , August 9, 1918). Seven months later the _Times-Picayune_ reported that the superintendent received two letters he believed to be from the Axeman ( _NOTP_ , March 11, 1919), and when the paper printed the letter purporting to be from the Axeman, it noted that it was "similar in some respects" to the letters received by Mooney ( _NOTP_ , March 16, 1919). In _Empire of Sin_ , Krist (p. 275) suggests that the letters may have come from the killer himself. He may be right, but so little is known of these letters that it is difficult to speculate. _"I keep on hearing the police":_ Rumbelow, _Jack the Ripper_ , 118. _WAIT —WATCH THE AXMAN'S JAZZ: NOTP_, March 23, 1919. _"think[ing] of the great amount of harm": Herald_ , March 20, 1919. _"Did Frank Jordano attack you": NODS_ , March 16, 1919. _Q: "Did Frank Jordano hit you": NOI_ , March 18, 1919. _Q: "Was Iorlando Jordano with Frank Jordano": NOI_ , March 17, 1919. _"was unable to give details": NOTP_ , March 18, 1919. _"I am confident we have the right men":_ Ibid. _"You must make up your mind:" NOTP_ , February 4, 1920. _"useless and unnecessary": NOTP_ , May 8, 1919. ## 10. "Hung by the Neck Until Dead, Dead, Dead" All direct quotations during the trial are taken from the trial transcript in _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. I have edited and rearranged testimony for narrative purposes. Details of Louis Besumer's trial are from _NODS_ , April 30 and May 1–2, 1919; _NOI_ , April 30 and May 1–2, 1919; _NOTP_ , April 30 and May 1–2, 1919. Description of the Gretna courthouse based on evidence in _NOI_ , February 6–7, 1907; _NODP_ , February 7, 1907; and "Parish of Jefferson" in Inventory of the Parish Archives, Louisiana State University Archives. For Judge Fleury's background, see _NOI_ , June 2, August 28, and November 3, 1914; and _NOTP_ , November 6, 1914. For Robert Rivarde's background, see sources listed in chapter 9. For the difficulties empaneling a jury, see _NOI_ , May 19–20, 1919; _NODS_ , May 20, 1919; _NOTP_ , May 20, 1919; and _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. Details of the Jordano trial are from _NODS_ , May 21–24 and 27, 1919; _NOTP_ , May 22–27, 1919; _NOI_ , May 23–24 and 26–27, 1919; and _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. _cases of Spanish influenza were declining: NODS_ , October 21 and 25, 1918; _NOTP_ , October 22, 1918; and _NODS_ , February 7, 1919. _"largely circumstantial": NOTP_ , March 18 and May 1, 1919. _He'd been locked up in the parish prison: NODS_ , October 4, 1918; _NOI_ , March 17, 1919; _NOTP_ , March 17–18, 1919; _NOI_ , April 12 and April 27, 1919; and _NOTP_ , April 27, 1919. _"victims of the murderous Axeman": NOTP_ , May 2, 1919. _"temple of justice": NOI_ , February 7, 1907. _"obliterated their awe of death": NOI_ , May 20, 1919. ## 11. Verdict All direct quotations during the trial are taken from the trial transcript in _State v. Guagliardo and Guagliardo_ , in Supreme Court of Louisiana Historical Archives. I have edited and rearranged testimony for narrative purposes. _the way a coroner or, today, a medical examiner:_ My thanks to Dr. Carol Terry, chief medical examiner, Gwinnett County, Georgia, for pointing out that surgeons often don't know as much about forensics as they think they do. _"It is true, judge, that I may hang":_ Composite speech based on accounts in _NODS_ , October 11, 1919; _NOI_ , October 11, 1919; and _NOTP_ , October 11, 1919. ## 12. False Lead For discussions of Joseph Mumfre as an Axeman suspect, see Krist, _Empire of Sin_ ; McQueen, _Axman Came from Hell_ ; and sources listed in the preface. Information about Angola from Vodicka, "Prison Plantation"; _Daily States_ (Baton Rouge), June 6, 1908; and _NODP_ , October 29, 1908. Convict Record of the Louisiana State Penitentiary, in Louisiana State Archives, states that Mumfre was discharged on June 2, 1915. For Vincent Moreci's murder and Mumfre's arrest as a suspect, see _NODP_ , March 13 and December 22, 1910; _NODS_ , November 19, 1915; _NOTP_ , November 20–21, 1915; and _NOI_ , November 21, 1915. Story of Mumfre held by Sheriff Marrero and finally sent back to the penitentiary comes from _NODS_ , February 3–5, 12, 14, and 16–18, 1916; _NOTP_ , February 4–6, 12, 15, and 17, 1916; _NODS_ , March 10, 1916; _NOTP_ , March 11, 1916; _NODS_ , July 10, 1916; and _NOTP_ , July 11, 1916. See also _Monfre_ [ _sic_ ] _v. Marrero_ , Westlaw, 138 La. 737, 70 So. 786 (1916); and _Monfre_ [ _sic_ ] _v. Marrero_ , Westlaw, 138 La. 739, 70 So. 787 (1916) For Mumfre's trial, see _NOTP_ , February 17, 1916; _NODS_ , February 17–18, 1916; _NODS_ , March 10, 1916; and _NOTP_ , March 11, 1916. For Mumfre being released from prison and quickly rearrested, see Convict Record of the Louisiana State Penitentiary, in Louisiana State Archives; _NOTP_ , May 18, 1918, and January 20, 1919. For Mumfre being run out of New Orleans, see _NOTP_ , May 18, 1918; _NODS_ , January 20, 1919; _NOTP_ , January 20, 1919; _NODS_ , January 24, 1919; and _NODS_ , December 15, 1921. Details of the attack on Sarah Laumann from _NODS_ , August 4–5, 1919; _NOI_ , August 4–5, 1919; and _NOTP_ , August 4–5, 1919. Additional details about Sarah Laumann family from 1910 and 1920 US Census, Ancestry.com. For the robbery on the night after the Laumann attack, see _NODS_ , August 5, 1919; and _NOI_ , August 5, 1919. The story of Steve Boca and William Carlson is told in Tallant, "Axman Wore Wings" in _Ready to Hang_. Details of the murder of Mike Pepitone in _NOI_ , October 27, 1919; _NODS_ , October 27–28, 1919; _NOTP_ , October 27–29, 1919; _NODS_ , May 15, 1921; NOPD, _Reports of Homicide_ , October 27, 1919, in New Orleans Public Library; and Coroner's Office, autopsy report, October 27, 1919, in New Orleans Public Library. For the murder of Paul Di Christina and Peter Pepitone's fear of retaliation, see _NODP_ , April 14, 1910; _NOI_ , April 14–15, 1910; _NOTP_ , November 10, 1915; _NOI_ , November 10 and 19, 1915; and _NODS_ , May 15, 1921. Details of Esther Pepitone's move to California and marriage to Angelo Albano come from the 1920 US Census, Ancestry.com; Louisiana State Board of Health, death certificate for Jenny Albano, Parish of Orleans, October 14, 1918; _NODS_ , October 27, 1919; and _LAT_ , December 6, 17, and 25, 1921. For Albano and Mumfre's history, see _NODP_ , April 16, 1908; _NODS_ , February 5 and 24, 1916; _NOTP_ , February 17, 1916; _NODS_ , March 10, 1916; and _LAT_ , December 6, 10, and 17, 1921. Details of Albano's disappearance in _LAT_ , November 6 and December 10, 11, 17, and 25, 1921; _Los Angeles Examiner_ , December 10 and 16, 1921; and _NOTP_ , December 16, 1921. Details of Mumfre's death in _LAT_ , December 6, 7, 17, 25, and 29, 1921; _Los Angeles Examiner_ , December 8, 10, and 16, 1921; _NODS_ , December 16, 1921; and _Los Angeles Evening Herald_ , April 10, 1922. For L.A. detectives learning about Mumfre's identity and history from New Orleans, see _LAT_ , December 10, 1921; _NOTP_ , December 15, 1921; and _Los Angeles Examiner_ , December 16, 1921. In addition, Esther Pepitone's defense subpoenaed an identification officer with police records of "Joseph P. Monfre," Superior Court, case no. CR17593, subpoena of A. R. Kallmeyer. Details of Esther Pepitone's trial taken from _Los Angeles Evening Herald_ , April 8 and 10, 1922; and _Los Angeles Examiner_ , April 11, 1922. See also _NODS_ , December 16, 1921; _LAT_ , December 17 and 25, 1921; and list of witnesses subpoenaed by the state, Superior Court, case no. CR17593. Esther's testimony at her trial is based on newspaper accounts of her trial but also on earlier newspaper accounts. The Los Angeles papers didn't cover Esther Albano's April 1922 trial in great detail. Some details of her testimony are taken from earlier newspaper coverage of the investigation and criminal proceedings against her on the assumption that she told substantially the same story on the witness stand in April 1922. For biographical information on Robert Tallant, see Robert Tallant Papers, in New Orleans Public Library; and _NOTP_ , April 2, 1957. For biographical information on Andy Ojeda and Jim Coulton, see _NODS_ , February 7, 1949; _NOI_ , February 7, 1949; _NOTP_ , February 7, 1949; _NOTP_ , February 8, 1949; _NODS_ , April 1, 1950; and _NOTP_ , April 2, 1950. _a continuation of the vendetta:_ And, indeed, police suspected that subsequent shootings of Italians were part of the same vendetta. See _NODS_ , January 30 and May 14–15, 1916; and Kendall, "Blood on the Banquette." _"no bondsman [was] forthcoming": NODS_ , February 14, 1916. _paid intervention:_ Convict Record of the Louisiana State Penitentiary, in Louisiana State Archives, notes: "Parole approved by Gov. L. E. Hall." _"For years Monfre": NODS_ , January 24, 1919. _"It would do the city no good":_ Ibid. _Doc Mumfre left:_ That Mumfre left New Orleans for California about January 1919 is confirmed by his death certificate (in the name of "Leone J. Manfre"), California Department of Public Health, dated December 1921, which said that he had been living in Los Angeles for three years. _"I felt a stinging": NODS_ , August 4, 1919. _"About 26 years old": NOTP_ , August 4, 1919. _"a thick, blunt instrument": NOTP_ , August 5, 1919. _"if he had hit her": NOI_ , August 4, 1919. _MYSTERIOUS "AXMAN": NOTP_, August 4, 1919. _"probably six feet": NODS_ , August 10, 1918. _Esther Pepitone:_ Sometimes spelled Pipitone. _"Oh, Lord!": NODS_ , October 27, 1919. _"it looks like the Axeman":_ NOPD, _Reports of Homicide_ , October 27, 1919, in New Orleans Public Library. _Detectives theorized that he'd heard:_ Another theory was that Pepitone was attacked outside his bedroom and his murderers dragged him back to his bed or he managed to stagger back to bed. The homicide report records the finding of blood in the room next to Pepitone's bedroom. But no subsequent New Orleans news story mentions this, so it may have been an early misapprehension by the reporting officer. Yet this version of events is also attributed to Mrs. Pepitone in _LAT_ , December 25, 1921. _believed he was involved:_ Mike Pepitone was actually a nasty piece of work. He'd recently been arrested for physically abusing his six-year-old son. _NODS_ , September 13, 1919; and _NOTP_ , September 13, 1919. _"another vendetta mystery": NODS_ , October 28, 1919. _fed the Axeman narrative:_ I'm presuming that Coulton wrote these articles because he was the police and crime reporter at the time. _"Mooney does not connect": NODS_ , October 28, 1919. _"bore some of the characteristics": NOTP_ , October 27, 1919. _"the Pepitone case is not unlike": NOTP_ , October 29, 1919. _Leone J. Manfre:_ Without a doubt, Joe Mumfre and Leone J. Manfre were the same man. Mumfre's 1917–1918 draft registration form records that he had a daughter Lena, who lived in Los Angeles. In 1915, the _NODS_ noted that Mumfre had a twelve-year-old daughter in California. In 1921, "Leone J. Manfre" had a nineteen-year-old daughter in L.A. named Lena Manfre, which would give her the right name and right age to be Joseph Mumfre's daughter (see _Los Angeles Examiner_ , December 8, 1921). Also, L.A. court records show that for the trial of Esther Pepitone for killing Leone J. Manfre, the L.A. Police Department requested the police records of "Joseph Monfre" [ _sic_ ] from New Orleans, suggesting that L.A. officials believed Manfre and Monfre/Mumfre to be the same man. See Superior Court, case no. CR17593, subpoena of Lena Vera Manfre, and subpoena of A. R. Kallmeyer for police records of Joseph P. Monfre. _Esther tried to carry on:_ Esther filed for divorce in early November and had a receiver named for Albano's property so she could get into his safety deposit box. _LAT_ , November 6, 1921; _States_ , December 16, 1921. _testified that the dead man was her father: LAT_ , December 29, 1921. I'm assuming Lena testified for the same reason in April as she had in December. _"Albano has a big house":_ The accounts of the trial in the newspapers covered only Esther Albano's testimony. But the case records include copies of those subpoenaed by the defense for a pretrial hearing in March, and it only makes sense that the defense would call these witnesses during the actual trial. Mrs. Griffith's testimony is taken from _LAT_ , December 25, 1921. _not guilty:_ Esther Pepitone returned to New Orleans, where she died in 1940. She never remarried. _NOTP_ , August 25, 1940. _"motive and probable solution": NOTP_ , December 15, 1921. New Orleans police also told Los Angeles investigators that Mumfre killed Pepitone. _Los Angeles Examiner_ , December 16, 1921. _"Perhaps it was just a coincidence": NODS_ , December 15, 1921. _pointing out that Joseph Mumfre: NOI_ , December 16, 1921. _Tallant thought he was the first:_ Tallant, Dreyer, and Saxon, _Gumbo Ya-Ya_ , vii. Tallant didn't know about Kendall's "Blood on the Banquette," published in 1939. _"He was the Axeman":_ Tallant, _Ready to Hang_ , 215. In an earlier telling of the story published in 1945, Tallant didn't quote Mrs. Pepitone as alleging that Mumfre was the Axeman. In fact, although he claimed that the Axeman crimes could be matched to the times Mumfre was out of prison, he added that "there was no evidence of his connection with the ghastly [Axeman] crimes." Tallant, Dreyer, and Saxon, _Gumbo Ya Ya_ , 89. ## 13. Rosie and Saint Joseph For Frank's reaction to the verdict, see _NOTP_ , May 28, 1919. For Jim Coulton's firm belief in the Jordanos' innocence, see _NOTP_ , December 7, 1920. For Coulton's personal experience with corrupt cops, see _NOI_ , February 5, 1919. For the appeals and delays in the Jordano case, see _NOTP_ , June 30 and July 6 and 11, 1919; _NODS_ , July 10, 1919; and _NOTP_ , October 4, 1919. Description of hanging based on accounts in _NODP_ , April 29, 1905, and March 6, 1909; _NOI_ , January 13, 1910; _NOTP_ , July 24, 1918; _NODS_ , November 28, 1919, and August 13, 1920. Details of Lena's wedding from _NOI_ , January 4, 1920; _NODS_ , January 4–5, 1920; and _NOTP_ , January 4–5, 1920. Evidence for the Cortimiglias' difficulties after the trial found in _NOTP_ , June 18, 1919; _NODS_ , June 20, 1919; _NOI_ , June 21, 1919; _NOTP_ , July 11, 1919; _NOTP_ , August 30, 1919; _NOI_ , August 31, 1919; and _NODS_ , May 22, 1920. For Storyville and prostitution, I have relied on Krist, _Empire of Sin_ ; Rose, _Storyville_ ; and Vyhnanek, "Seamier Side of Life" (PhD diss.). For the details of Rosie Cortimiglia's arrest for prostitution, see _NOI_ , November 1–2, 1919; _NODS_ , November 2, 1919; and arrest record for Rosie Cortimiglia, November 1, 1919, in NOPD, Arrest Cards. Details of Rosie and Charlie's failed reconciliation in _NOTP_ , November 7, 1919; _NODS_ , February 4, 1920; _NOI_ , February 4, 1920; and _NOTP_ , February 4, 1920. For the story of Rosie's confession, see _NODS_ , February 4, 1920; _NOI_ , February 4, 1920; _NOTP_ , February 4–5, 1920; and statement of Rosie Salama Cortimiglia to _Times-Picayune_ , in Supreme Court of Louisiana Historical Archives. For reaction to news of Rosie's retraction, see _NOI_ , February 4, 1920; and _NODS_ , February 4, 5, and 7, 1920. For memory, see Loftus and Hoffman, "Misinformation and Memory"; Loftus and Ketcham, _Myth of Repressed Memory_ ; Loftus and Thomas, "Creating Bizarre False Memories"; and Loftus works cited in chapter 8. For details of the Jordanos' state supreme court appeal, see _NOI_ , February 4, 1920; _NODS_ , February 7–8, 1920; _NOTP_ , February 8, 1920; _NOI_ , March 6, 7, and 14, 1920; _NOTP_ , March 7 and 14, 1920; _NODS_ , March 13–14 and April 5, 1920; _NOI_ , April 5, 1920; and _NOTP_ , April 6 and 22, 1920. See also the state supreme court decision, _State v. Guagliardo et al._ , in Westlaw; and Supplemental Brief on Behalf of Defendants and Appellants, in Supreme Court of Louisiana Historical Archives. For Rosie ill with smallpox, see _NOTP_ , January 30, 1920; _NOI_ , January 31 and February 11, 1920; and _NOTP_ , December 7, 1920. For Rosie swearing to affidavit, see _NODS_ , April 26, 1920; and _NOTP_ , April 26, 1920. For grand jury investigation into Rosie's charges, see _NODS_ , May 12, 1920; and _NOTP_ , December 7, 1920. For fears that Rosie would not recant her testimony under oath, see _NODS_ , May 16, 1920; and _NOTP_ , May 19, 1920. For details of attempts to get bail for Jordanos, see _NOTP_ , April 26–27 and May 9, 13, and 20, 1920; _NOI_ , May 28 and November 22 and 26, 1920; and _NODS_ , November 26, 1920. For failed attempts to retry Jordanos, see _NOTP_ , May 9, 17, 19, and 20, 1920; _NOI_ , May 28, 1920; _NODS_ , November 14, 15, 21, 22, and 26, 1920; and _NOTP_ , November 15, 16, and 27, 1920. For details of the Jordanos' life in the parish jail, see _NOTP_ , February 5, 1920; _NOI_ , November 15, 1920; and _NODS_ , November 15, 1920. For the Jordanos finally freed, see _NODS_ , December 6, 1920; _NOI_ , December 6, 1920; and _NOTP_ , December 7, 1920. _"nonchalance": NODS_ , February 4, 1919. _"I believe you": NOTP_ , December 7, 1920. _"tempting dishes of chicken": NODS_ , January 5, 1920. _"ancient restricted district": NOI_ , January 29, 1919. _"Nothing in the world will ever": NODS_ , November 2, 1919. _"I sinned":_ Ibid. _"I would rather go to the gallows":_ Composite statement based on _NODS_ , October 11, 1919; and _NOTP_ , October 11, 1919. _"He raised his hand": NODS_ , February 4, 1920. _"Rosie, you cannot die":_ Statement of Rosie Salama Cortimiglia to _Times-Picayune_ , in Supreme Court of Louisiana Historical Archives. _"sought out friends": NODS_ , February 4, 1920. _"either by their faces, figures or voices":_ Statement of Rosie Salama Cortimiglia to _Times-Picayune_ , in Supreme Court of Louisiana Historical Archives. _"I hope I can sleep": NOTP_ , February 4, 1920. _"as surely as there is": NODS_ , February 4, 1919. _How will Papa stand it?: NOTP_ , February 5, 1920. _"Both of us are happy":_ Composite quote from _NODS_ , February 4, 1920; and _NOTP_ , February 5, 1920. _"Of course, I forgive Rosie Cortimiglia": NOI_ , February 4, 1920. _"It must have been the blessed Saint Joseph": NOTP_ , February 5, 1920. _"I made up my mind to say":_ Statement of Rosie Salama Cortimiglia to _Times-Picayune_ , in Supreme Court of Louisiana Historical Archives. _"Did the Jordanos do it?": NOI_ , February 4, 1920. _she no longer had the constant reinforcement:_ In the courthouse, Chief Leson was locked up with Rosie in the witness room. _State v. Guagliardo and Guagliardo_ , 179, in Supreme Court of Louisiana Historical Archives. _"he would have to kill the baby":_ Statement of Rosie Salama Cortimiglia to _Times-Picayune_ , in Supreme Court of Louisiana Historical Archives. _"where human life was at stake": NOTP_ , March 7, 1920. _"It is a significant thing": NOI_ , March 14, 1920; Supplemental Brief on Behalf of Defendants and Appellants, in Supreme Court of Louisiana Historical Archives. _"As a matter of fact and law": State v. Guagliardo et al._ , in Westlaw. _"the only reasonable presumption":_ Ibid. _"to import into a case material facts":_ Ibid. _"of her own free will": NOTP_ , April 26, 1920. _"On the last night of my confinement":_ Ibid. _"If Rosie Cortimiglia is convicted": NOTP_ , May 13, 1920. _"laid special stress on the crime": NODS_ , May 12, 1920. _"returned to the sweltering heat": NOI_ , May 19, 1920. _"This is the second time": NOI_ , November 22, 1920. _"If the state is not ready":_ Ibid. _"Your honor, just before court opened": NODS_ , December 6, 1920. _"You can go":_ Ibid. _"That's all right":_ Ibid. _"Ain't it fine?": NOTP_ , December 7, 1920. ## 14. The Final Chapter? Details of Lena's and Iorlando's deaths from _NODS_ , February 10, 1921; _NOI_ , February 10, 1921; _NOTP_ , February 10, 1921, and October 18, 1925; "Giolando Guagliardo aka Jordano" and "Orlando Guagliardo Jordano" in Ancestry.com, One World Tree Project. For the remainder of Frank Jordano's life, see _NOTP_ , June 17, 1923, December 20, 1925, May 3, 1927, and February 21, 1961; Jefferson Parish, marriage license (no. 1609) for Frank Jordano and Linzy Hamilton, February 7, 1926; Jefferson Parish, marriage license (no. 1812) for Frank Jordano and Mary Shambra, April 6, 1936; Louisiana State Board of Health, death certificate for Frank Jordano; "Frank Guagliardo" in Ancestry.com, One World Tree Project. For Robert Rivarde's career, see _NOTP_ , September 11, 1924, and July 2, 1965; _States-Item_ , January 30–31, 1967; _NOTP_ , January 31 and October 3, 1967. For John Fleury's career, see _NOTP_ , September 30, 1984. For William Byrnes's career, see _NOTP_ , February 26–28, 1942; "Judge William H. Byrnes, Jr." in _Louisiana Today_ , by James M. Thomson, 337, 396; and "Judge William H. Byrnes, Jr.," in _Book of the South_ , ed. by John Temple Graves, 276. For Frank Mooney's failings as police chief, see _NOTP_ , December 30, 1917, and January 17, 1919; _NODS_ , February 27, 1919; _NOTP_ , February 28 and April 6, 1919; _NOI_ , March 17 and 20, 1920; and _NOTP_ , November 27, 1920, and August 23, 1923. The Axeman crimes weren't the only murders he was criticized for not solving; see _NOI_ , April 25, 1919. For Mooney's resignation and the remainder of his life, see _NOTP_ , August 24 and 26 and November 28, 1920; _NODS_ , November 28, 1920; _NOI_ , August 23, 1923; _NODS_ , August 23 and 26, 1923; and _NOTP_ , August 28, 1923. Details of the Spero murder are from _ADTT_ , December 13, 14, and 17, 1920; _NODS_ , December 13–14, 1920; and _NOTP_ , December 15 and 18, 1920. Description of the Orlando murder from _NODS_ , January 12, 1921; _NOI_ , January 12, 1921; _NOTP_ , January 13, 1921; _LCWAP_ , January 14 and 21, 1921; and _DeRidder Enterprise_ , January 15, 1921. Additional information for John Orlando and family from Louisiana State Board of Health, death certificate for John Orlando; 1920 US Census, Ancestry.com; WWI Draft Registration Cards, Ancestry.com; Texas, Naturalization Records, 1881–1992, Ancestry .com, accessed June 26, 2015, http://search.ancestry.com/search/db.aspx?dbid=2509; Francine Cloud, interview with the author, February 6, 2015. Details of the Scalisi murder are from _LCWAP_ , April 12, 14, and 19, 1921; Criminal Records Archives, Calcasieu Parish, coroner's certificate for Frank Scalisi, April 13, 1921. On the reliability of eyewitness testimony, see Cutler and Penrod, _Mistaken Identification_ ; Douglass and Pavletic, "Eyewitness Confidence"; Engelhardt, "Problem with Eyewitness Testimony"; Poole et al., "Children as Witnesses"; Putwain and Sammons, _Psychology and Crime_ ; and Steblay, "Eyewitness Memory." For racist stereotypes, see Jenkins, "Serial Murder." For robbery and murder spree in Birmingham, Alabama, see McQueen, _Axman Came from Hell_ ; and Kazek, _Forgotten Tales_. For serial killers, see sources cited in chapter 6. _"probably the youngest realtor": NOTP_ , December 20, 1925. _"There'll be no more political jobs": NOI_ , November 28, 1920. _The murder remained unsolved:_ Josephine Orlando told her granddaughter, Francine Cloud, that several men broke in, poured chloroform on the floor, and attacked the family. Revenge was ostensibly the motive. But none of the contemporary newspaper accounts mention such a possibility. The idea that assailants would chloroform their victims was a common convention of the time, and many people (including Italians) assumed that any attack on Italians was motivated by revenge. Josephine was only six years old at the time of the attack, and her memory probably reflects the speculation she subsequently heard from the adults around her. For created memories, see Loftus, "Make-Believe Memories" and "Malleability of Human Memory." Moreover, Mrs. Cloud also heard growing up that her great-grandfather was killed by "the Axeman." This term was used by both her grandmother and her great-grandmother (who spoke little English). _"short [and] chunky":_ Criminal Records Archives, Calcasieu Parish, coroner's certificate for Frank Scalisi, April 13, 1921. _"short and stout":_ Ibid. _"a short, heavy-set man": NODP_ , August 14, 1910. _"said she did not see": LCAP_ , April 12, 1921. _"The man was short":_ Criminal Records Archives, Calcasieu Parish, coroner's certificate for Frank Scalisi, April 13, 1921. _"an ax murder identical": NODS_ , December 13, 1920. _a morphine or cocaine habit:_ At the time both drugs could be had from any drugstore with a prescription. # Bibliography ## Archival Sources Criminal Records Archives, Calcasieu Parish, Clerk of Court, Lake Charles, Louisiana. • Coroner's certificates. Jefferson Parish General Government Center (Gretna Courthouse), Marriage License and Passport Department. • Marriage licenses. Louisiana State Archives, Baton Rouge. • Convict Record of the Louisiana State Penitentiary. Louisiana State Board of Health, Bureau of Vital Statistics. • Death certificates. Louisiana State University Archives. • Inventory of the Parish Archives of Louisiana. Prepared by the Historical Records Survey, Division of the Professional and Service Projects Work Projects Administration. No. 26. Jefferson Parish (Gretna). The Department of Archives, Louisiana State University, January 1940. New Orleans Public Library, Louisiana Division/City Archives. • City Directories of the United States, vol. 4 (1902–1935). • Coroner's Office, Autopsy Reports, vol. 1 (1904–1913). • Coroner's Office, Autopsy Reports, vol. 2 (1913–1924). • Coroner's Office, _Record Book Journals_. • New Orleans Police Department, Annual Reports. • New Orleans Police Department, Arrest Cards, 1914–1947. • New Orleans Police Department, _Reports of Homicide_. • Orleans Parish Criminal District Court Records. * º _State of Louisiana v. John T. Flannery_. Docket no. 38,264 (1910). * º _State of Louisiana v. Joseph Monfre_ [ _sic_ ]. Docket no. 35,993 (1908). * º _State of Louisiana v. Francis Rodin and Eugene Bescanon_. Docket no. 38,403 (1911). * º _State of Louisiana v. John Wesley Sumner_. Docket no. 46,575 (1917). • Orleans Parish, Louisiana, Coroner's Office, vol. 9, January 1, 1855–September 12, 1855. • Robert Tallant Papers, Manuscript Collection. • _Rules and Regulations Governing the Police Department of the City of New Orleans_ (1921). • "Testimony, Murder of Superintendent of Police James W. Reynolds," 1917. State of California Department of Public Health Center for Health Statistics and Informatics, Vital Records. • Death certificates. Superior Court of the County of Los Angeles, Archives and Records Center. • _The People of the State of California v. Esther Albano_. Case no. CR17593 (1922). Supreme Court of Louisiana Historical Archives, Louisiana and Special Collections Department, Earl K. Long Library, University of New Orleans. • _State of Louisiana v. Frank Guagliardo, alias Frank Jordano, and Iorlando Guagliardo, alias Jordano_. Docket no. 23,815 (1919–1920). * º Statement of Rosie Salama Cortimiglia to _Times-Picayune_. * º Original Brief on Behalf of the Defendants. * º Original Brief on Behalf of the State. * º Supplemental Brief on Behalf of the Defendants and Appellants. * º Trial transcript. 3 vols. No. 4458. * º Opinion and Judgement. ## Newspapers ### New Orleans • _Daily City Item_ • _Daily News_ • _Daily True Delta_ ( _DTD_ ) • _Herald_ • _Mascot_ • _New Orleans Bee_ ( _NOB_ ) • _New Orleans Daily States_ ( _NODS_ ); after 1918 becomes the _States_ • _New Orleans Item_ ( _NOI_ ) • _New Orleans States-Item_ • _New Orleans Times_ • _New Orleans Times-Democrat_ ( _NOTD_ ) • _New Orleans Times-Picayune_ ( _NOTP_ ); before 1913 the _New Orleans Daily Picayune_ ( _NODP_ ) ### Others • _Alexandria Daily Town Talk_ ( _ADTT_ ) • _DeRidder Enterprise_ • _Lake Charles Weekly American Press_ ( _LCWAP_ ) • _Los Angeles Evening Herald_ • _Los Angeles Examiner_ • _Los Angeles Times_ ( _LAT_ ) • _New York Tribune_ ## Online Sources Ancestry.com • New Orleans, Louisiana, Death Records Index, 1804–1949. <http://search.ancestry.com/search/db.aspx?dbid=6606.> • New Orleans, Louisiana, Marriage Records Index, 1831–1964. <http://search.ancestry.com/search/db.aspx?dbid=6500.> • One World Tree Project. <http://awtc.ancestry.com/cgi-bin/igm.cgi>. • US, World War I Draft Registration Cards, 1917–1918. <http://search.ancestry.com/search/db.aspx?dbid=6482.> • US Federal Census Collection. <http://search.ancestry.com/search/group/usfedcen>. City of New Orleans. _Population, Total and by Race, of Orleans, Jefferson, St. Bernard, and St. Tammany Parishes, the State of Louisiana, and Five Southern Cities Comparable to New Orleans, 1810–1980_. Prepared by City Planning Commission, with the assistance of New Orleans Public Library, 1981. <http://nutrias.org/facts/aq150_1981p.pdf>. Family Search. Conveyance Records, Saint James Parish Clerk of Court, Saint James Parish Courthouse. Vol. 52 (1889–1893). Family History Library, microfilm no. 0402637l. <https://familysearch.org>. Supreme Court of Louisiana. _State v. Guagliardo et al._ Docket no. 23815. Westlaw. 146 La. 949, 84 So. 216 (1920). University of Virginia Library. Historical Census Browser. Geospatial and Statistical Data Center, accessed January 9, 2012. <http://fisher.lib.virginia.edu/collections/stats/histcensus/index.html> (site discontinued). ## Unpublished Theses and Dissertations Adams, Margaret. "Outline of Mafia Riots in New Orleans." MA thesis, Tulane University, 1924. Baiamonte, John V., Jr. "New Immigrants in the South: A Study of the Italians of Tangipahoa Parish, Louisiana." MA thesis, Southeastern Louisiana College, 1969. Boneno, Roselyn Bologna. "From Migrant to Millionaire: The Story of the Italian-American in New Orleans, 1880–1910." PhD diss., Louisiana State University, 1986. Botein, Barbara. "The Hennessy Case: An Episode in American Nativism, 1890." PhD diss., New York University, 1975. Carroll, Ralph Edward. "The Mafia in New Orleans, 1900–1907." MA thesis, Notre Dame Seminary, New Orleans, 1956. Carroll, Richard Louis. "The Impact of David Hennessey on New Orleans Society and the Consequences of the Assassination of Hennessey." MA thesis, Notre Dame Seminary, New Orleans, 1957. Dulitz, Harris Myron. "The Myth of the Mafia in New Orleans." BS thesis, Tulane University, 1956. Edwards-Simpson, Louise Reynes. "Sicilian Immigration to New Orleans, 1870–1910: Ethnicity, Race, and Social Position in the New South." PhD diss., University of Minnesota, 1996. Metcalf, Christina. "Race Relations and the New Orleans Police Department: 1900–1971." BA thesis, Tulane University, 1985. Scarpaci, Jean Ann. "Italian Immigrants in Louisiana's Sugar Parishes: Recruitment, Labor Conditions and Community Relations, 1880–1910." PhD diss., Rutgers University, 1972. Vyhnanek, Louis Andrew. "The Seamier Side of Life: Criminal Activity in New Orleans During the 1920s." PhD diss., Louisiana State University, 1979. Warren, Karen Wright. "The Sicilian St. Joseph Altar Celebration in Southeastern Louisiana." MA thesis, Southeastern Louisiana University, 1983. Webb, Michael E. "Horror in a Time of Social Upheaval: The New Orleans 'Axman' Murders, 1910–1920." MA thesis, University of New Orleans, 2002. ## Books and Articles Adamoli, Giulio. "Letters from America, 1867." _Louisiana Historical Quarterly_ 6 (April 1923): 271–279. Adler, Jeffrey S. "Murder, North and South: Violence in Early-Twentieth Century Chicago and New Orleans." _Journal of Southern History_ 74, no. 2 (May 2008): 297–324. American Historical Society. _Louisiana_. Chicago: American Historical Society, 1925. Asbury, Herbert. _The French Quarter: An Informal History of the New Orleans Underworld_. New York: Knopf, 1936; Thunder's Mouth, 2003. Baiamonte, John V. _Immigrants in Rural America: A Study of the Italians of Tangipahoa Parish, Louisiana_. New York: Garland, 1990. ———. "'Who Killa de Chief' Revisited: The Hennessy Assassination and Its Aftermath, 1890–1991." _Louisiana History_ 33 (1992): 117–146. Bancroft, Lundy. _Why Does He Do That? Inside the Minds of Angry and Controlling Men_. New York: Berkley Books, 2002. Barnett, Ola W. "Why Battered Women Do Not Leave, Part 1." _Trauma, Violence, & Abuse_ 1, no. 4 (2000): 343–372. ———. "Why Battered Women Do Not Leave, Part 2." _Trauma, Violence, & Abuse_ 2, no. 1 (2001): 3–35. Begnaud, Allen. "The Louisiana Sugar Cane Industry: An Overview." In _Green Fields: Two Hundred Years of Louisiana Sugar. A Catalogue Complementing the Pictorial Exhibit_. Lafayette, LA: Center for Louisiana Studies, University of Southwestern Louisiana, 1980. Behrman, Martin. _Martin Behrman of New Orleans: Memoirs of a City Boss_. Edited by John R. Kemp. Baton Rouge: Louisiana State University Press, 1977. Berthoff, Rowland T. "Southern Attitudes Toward Immigration, 1865–1914." _Journal of Southern History_ 17, no. 2 (August 1951): 328–360. Brock, Eric J. _New Orleans_. Charleston, SC: Arcadia, 1999. Brown, Pat. _Killing for Sport: Inside the Minds of Serial Killers_. Beverly Hills, CA: New Millennium Press, 2003. Cable, Mary. _Lost New Orleans_. Boston: Houghton Mifflin, 1980. Campanella, Richard. _Bienville's Dilemma: A Historical Geography of New Orleans_. Lafayette, LA: Center for Louisiana Studies, University of Southwestern Louisiana, 2008. ———. _Geographies of New Orleans: Urban Fabrics Before the Storm_. Lafayette, LA: Center for Louisiana Studies, University of Southwestern Louisiana, 2006. Carlisle, Al C. "The Divided Self: Toward an Understanding of the Dark Side of the Serial Killer." In _Contemporary Perspectives on Serial Murder_ , edited by Ronald M. Holmes and Stephen T. Holmes. Thousand Oaks, CA: Sage, 1998. Carter, Hodding. _Southern Legacy_. Baton Rouge: Louisiana State University Press, 1950. CatholicCulture.org. "History of the Saint Joseph Altar." Accessed August 31, 2016. www.catholicculture.org/culture/liturgicalyear/activities/view.cfm?id=1029. Chambers, Henry E. _A History of Louisiana_ : _Wilderness, Colony, Province, Territory, State, People_. Chicago: American Historical Society, 1925. Chandler, David. _Brothers in Blood: The Rise of the Criminal Brotherhoods_. New York: E. P. Dutton, 1975. Chupa, Anna Maria. "St. Joseph's Day Altars." Houston Institute for Culture, Louisiana Project. www.houstonculture.org/laproject/stjo.html. Conrad, Glenn R., and Ray F. Lucas. _White Gold: A Brief History of the Louisiana Sugar Industry, 1795–1995_. Lafayette, LA: Center for Louisiana Studies, University of Southwestern Louisiana, 1995. Coxe, John E. "The New Orleans Mafia Incident." _Louisiana Historical Quarterly_ 20, no. 4 (October 1937): 1067–1110. Critchley, David. _The Origin of Organized Crime in America: The New York City Mafia, 1891–1931_. New York: Routledge, 2008. Curry, Mary Grace. _Gretna: A Sesquicentennial Salute_. Metairie, LA: Jefferson Parish Historical Commission, 1986. Cutler, Brian L., and Steven D. Penrod. _Mistaken Identification: The Eyewitness, Psychology, and the Law_. Cambridge: Cambridge University Press, 1995. Dash, Mike. _The First Family: Terror, Extortion, Revenge, Murder and the Birth of the American Mafia_. New York: Random House, 2009. Deacon, William A., and John P. Coleman. _Martin Behrman Administration Biography, 1904–1916_. New Orleans: John J. Weihing Printing, 1917. Dickie, John. _Cosa Nostra: A History of the Sicilian Mafia_. London: Hodder & Stoughton, 2004. Douglas, John, and Johnny Dodd. _Inside the Mind of BTK: The True Story Behind Thirty Years of Hunting for the Wichita Serial Killer_. San Francisco: Wiley, 2007. Douglas, John, and Mark Olshaker. _The Cases That Haunt Us: From Jack the Ripper to JonBenet Ramsey, the FBI's Legendary Mindhunter Sheds Light on the Mysteries That Won't Go Away_. New York: Scribner, 2000. Douglass, Amy Bradfield, and Afton Pavletic. "Eyewitness Confidence Malleability." In _Conviction of the Innocent: Lessons from Psychological Research_. Washington, DC: American Psychological Association, 2012. Downey, Clifford J. _Chicago and the Illinois Central Railroad_. Charleston, SC: Arcadia, 2007. Edward, Wallace. _The Axeman: The Brutal History of the Axeman of New Orleans_. Anaheim, CA: Absolute Crime Books, 2013. Eisner, Manuel. "Long-Term Historical Trends in Violent Crime." _Crime and Justice_ 30 (2003): 83–142. Engelhardt, Laura. "The Problem with Eyewitness Testimony." _Stanford Journal of Legal Studies_ 1, no. 1 (1999): 25–29. Everitt, David. _Human Monsters: An Illustrated Encyclopedia of the World's Most Vicious Murderers_. Chicago: Contemporary Books, 1993. Fentress, James. _Rebels and Mafiosi: Death in a Sicilian Landscape_. Ithaca, NY: Cornell University Press, 2000. Fortier, Alcée. _Louisiana: Comprising Sketches of Parishes, Towns, Events, Institutions, and Persons, Arranged in Cyclopedic Form_. Madison, WI: Century Historical Association, 1914. Gambino, Richard. _Vendetta: A True Story of the Largest Lynching in U.S. History_. Toronto: Guernica, 1998. Garvey, Joan B., and Mary Lou Widmer. _Beautiful Crescent: A History of New Orleans_. New Orleans: Garmer Press, 2006. Goodspeed Publishing. _Biographical and Historical Memoirs of Louisiana_. Vol. 3. Chicago: Goodspeed Publishing, 1892; Louisiana Classic Series Reprint, Baton Rouge: Claitor's, 1975. Graves, John Temple, ed. _A Book of the South_. New Orleans: Southern Editors Association, 1940. Hale, Robert. "The Application of Learning Theory to Serial Murder, or 'You Too Can Learn to Be a Serial Killer.'" In _Contemporary Perspectives on Serial Murder_ , edited by Ronald M. Holmes and Stephen T. Holmes. Thousand Oaks, CA: Sage 1998. Hess, Henner. _Mafia and Mafiosi: Origin, Power and Myth_. Translated by Ewald Osers. New York: New York University Press, 1996. Hickey, Eric W. _Serial Murderers and Their Victims_. 3rd ed. Belmont, CA: Wadsworth/Thomason Learning, 2002. Hintz, Martin. _Ethnic New Orleans_. _A Complete Guide to the Many Faces & Cultures of New Orleans_. Lincolnwood, IL: Passport Books, 1995. Holmes, Ronald M. "Psychological Profiling: Uses in Serial Murder Cases." In _Contemporary Perspectives on Serial Murder_ , edited by Ronald M. Holmes and Stephen T. Holmes. Thousand Oaks, CA: Sage, 1998. Holmes, Ronald M., and James E. DeBurger. "Profiles in Terror: The Serial Murderer." In _Contemporary Perspectives on Serial Murder_ , edited by Ronald M. Holmes and Stephen T. Holmes. Thousand Oaks, CA: Sage, 1998. Holmes, Ronald M., and Stephen T. Holmes. _Serial Murder_. 2nd ed. London: Sage, 1998. Hunt, Thomas, and Martha Macheca Sheldon. _Deep Water: Joseph P. Macheca and the Birth of the American Mafia_. Lincoln, NE: iUniverse, 2007. Industrialist Publishing. _New Orleans Police Souvenir Book, 1906_. New Orleans: Industrialist Publishing, 1906. _Italian-American Digest._ "Who Really Killed the Chief?" Vol. 7 (Spring 1981): 3. Jackson, Joy J. "Crime and the Conscience of a City." _Louisiana History_ 9, no. 3 (Summer 1968): 229–244. ———. _New Orleans in the Gilded Age: Politics and Urban Progress, 1880–1896_. Baton Rouge: Louisiana State University, 1969. Jeffers, H. Paul. _With an Axe: 16 Horrific Accounts of Real-Life Axe Murders_. New York: Kensington Publishing, 2000. Jenkins, Philip. "Serial Murder in the United States: 1900–1940: A Historical Perspective." _Journal of Criminal Justice_ 17 (1989): 377–392. Jessup, Steve. "The Golden Age, 1900–1950." In _The Complete Book of North American Railroading_ , by Kevin EuDaly et al. Minneapolis, MN: Voyager Press, 2009. Jewell, Edwin L. _Jewell's Crescent City Illustrated: The Commercial, Social, Political and General History of New Orleans, Including Biographical Sketches of Its Distinguished Citizens_. New Orleans: E. L. Jewell, 1873. Katz, Allan. "The Hennessy Affair: A Centennial," _New Orleans_ , October 1990, 58–62, 81. Kazek, Kelly. _Forgotten Tales of Alabama_. Charleston, SC: History Press, 2010. Keedy, Edwin R. "The Third Degree and Legal Interrogation of Suspects." _University of Pennsylvania Law Review and American Law Register_ 85, no. 8 (June 1937): 761–777. Kendall, John S. "Blood on the Banquette." _Louisiana Historical Quarterly_ 22, no. 3 (1939): 819–856. ———. "Notes on the Criminal History of New Orleans." _Louisiana Historical Quarterly_ 34 (1951): 147–173. ———. "Old-Time New Orleans Police Reporters and Reporting." _Louisiana Historical Quarterly_ 29, no.1 (January 1946): 43–58. ———. "Sixteen Years of Martin Behrman." In _History of New Orleans_ , 547–564. Chicago: Lewis Publishing, 1922. ———. "Who Killa de Chief?" _Louisiana Historical Quarterly_ 22, no. 2 (1939): 492–530. Krist, Gary. _Empire of Sin: A Story of Sex, Jazz, Murder and the Battle for Modern New Orleans_. New York: Crown, 2014. Kurtz, Michael L. "Organized Crime in Louisiana History: Myth and Reality," _Louisiana History_ 24 (1983): 355–376. Lane, Brian, and Wilfred Greg, eds. _The Encyclopedia of Serial Killers_. New York: Berkley Books, 1995. Lester, Henry. "The Axeman of New Orleans." In _Unsolved Murders and Mysteries_ , edited by John Canning. London: Time Warner, 1992. Levinson, Marc. _The Great A &P and the Struggle for Small Business in America_. New York: Hill & Wang, 2011. Lewis, Peirce F. _New Orleans: The Making of an Urban Landscape_. 2nd ed. Chicago: Center for American Places, 2003. Leyton, Elliot. _Hunting Humans: The Rise of the Modern Multiple Murderer_. 2nd ed. New York: Carroll & Graf, 2001. Loftus, Elizabeth. "Make-Believe Memories." _American Psychologist_ (November 2003): 867–873. ———. "The Malleability of Human Memory." _American Scientist_ 67, no.3 (May–June 1979): 312–320. Loftus, Elizabeth, and A. K. Thomas. "Creating Bizarre False Memories Through Imagination." _Memory & Cognition_ 30 (2002): 423–431. Loftus, Elizabeth, and H. G. Hoffman. "Misinformation and Memory: The Creation of New Memories." _Journal of Experimental Psychology_ 118, no. 1 (1989): 100–104. Loftus, Elizabeth, and Katherine Ketcham. _The Myth of Repressed Memory_ : _False Memories and Allegations of Sexual Abuse_. New York: Saint Martin's Press, 1994. Lombardo, Robert M. _The Black Hand: Terror by Letter in Chicago_. Urbana and Chicago: University of Illinois Press, 2010. Louisiana District Judges Association. _Biographies of Louisiana Judges_. New Orleans: New Orleans Public Library, Louisiana Division, 1961. Macaluso, Joseph N., Sr. _Italian Immigrant Families: Grocers, Proprietors, and Entrepreneurs_. Pittsburgh: RoseDog Books, 2004. Magnaghi, Russell. "Louisiana's Italian Immigrants Prior to 1870." _Louisiana History_ 27 (Winter 1996): 43–68. Margavio, A. V. "The Reaction of the Press to the Italian American in New Orleans, 1880 to 1920." _Italian Americana_ 4, no. 1 (1978): 72–83. Margavio, A. V., and J. L. Molyneaux. "Residential Segregation of Italians in New Orleans and Selected American Cities." _Louisiana Studies_ 12 (1973): 639–645. Margavio, A. V., and Jerome Salomone. _Bread and Respect: The Italians of Louisiana_. Gretna, LA: Pelican, 2002. ———. "Economic Advantages of Familism." _Sociological Spectrum_ 7 (1987): 101–119. ———. "The Passage, Settlement, and Occupational Characteristics of Louisiana's Italian Immigrants." _Sociological Spectrum_ 1 (1981): 345–359. Marr, Robert H., Jr. "The New Orleans Mafia Case." _American Law Review_ 25 (1891): 414–431. Maselli, Joseph, and Dominic Candeloro. _Italians in New Orleans_. Charleston, SC: Arcadia, 2004. McColloster, Mary Ann Tusa. "New Light on the New Orleans St. Joseph Day Altar. _Louisiana Folklore Miscellany_ 3, no. 1 (April 1970): 38–45. McMain, Eleanor. "Behind the Yellow Fever in Little Palermo." _Charities and the Commons_ 15 (October 1905–March 1906): 150–159. McQueen, Keven. _The Axman Came from Hell and Other Southern True Crime Stories_. Gretna, LA: Pelican, 2011. Morton, James. _Catching the Killers: The Definitive History of Criminal Detection_. London: Ebury Press, 2001. Morton, Robert J. "Serial Murder: Multi-Disciplinary Perspectives for Investigators." Behavioral Analysis Unit-2, National Center for the Analysis of Violent Crime, Critical Incident Response Group, Federal Bureau of Investigation. www.fbi.gov/stats-services/publications/serial-murder. Murray, Tom. _Illinois Central Railroad_. Saint Paul, MN: Voyageur Press, 2006. Nash, Jay Robert. _Bloodletters and Badmen_ : _A Narrative Encyclopedia of American Criminals from the Pilgrims to the Present_. New York: M. Evans, 1973. Nelli, Humbert S. _The Business of Crime. Italians and Syndicate Crime in the United States_. Chicago: University of Chicago Press, 1981. New Orleans Police Department. _Fifty Years of Progress, 1900–1950_. New Orleans: Franklin Print, 1951. Newton, Michael. _Century of Slaughter_. Lincoln, NE: toExcel, 2000. ———. _The Encyclopedia of Unsolved Crimes_. 2nd ed. New York: Checkmark Books, 2009. ———. _Hunting Humans: The Encyclopedia of Serial Killers_. Vol. 1. New York: Avon Books, 1990. Orso, Ethelyn, and Peggy Kaveski. "Undisclosed Aspects of Saint Joseph Altars." _Louisiana Folklore Miscellany_ 3, no. 5 (1975): 14–18. Persico, Joseph. "Vendetta in New Orleans." _American Heritage_ 24, no. 4 (June 1973): 65–72. Peterson, H. C., and Gilbert C. Fite. _Opponents of War, 1917–1918_. Madison: University of Wisconsin Press, 1957. Pitkin, Thomas Monroe, and Francesco Cordasco. _The Black Hand: A Chapter in Ethnic Crime_. Totowa, NJ: Litchfield, Adams, 1977. Plemer, Roslynn. "The Feast of St. Joseph." _Louisiana Folklore Miscellany_ 2, no. 4 (August 1968): 85–90. Poole, Debra Ann, Sonja P. Brubacher, and Jason J. Dickinson. "Children as Witnesses." In _APA Handbook of Forensic Psychology_ , vol. 2, _Criminal Investigation, Adjudication, and Sentencing Outcomes_. Washington, DC: American Psychological Association, 2015. Putwain, David, and Aidan Sammons. _Psychology and Crime_. New York: Routledge, 2002. Ramelli, Rudolph. "Scenes in the 'Underworld.'" In _New Orleans Police Souvenir Book, 1906_. New Orleans: Industrialist Publishing, 1906. Ramsland, Katherine. "All About the Axeman of New Orleans." Crime Library, truTV, accessed March 28, 2006. www.trutv.com/library/crime/serial_killers/weird/axeman/index.html (site discontinued). ———. _Beating the Devil's Game: A History of Forensic Science and Criminal Investigation_. New York: Berkley Books, 2007. ———. _Inside the Minds of Serial Killers_ : _Why They Kill_. Westport, CT: Praeger, 2006. Reid, Ed. _Mafia_. New York: Random House, 1952. Reynolds, George M. _Machine Politics in New Orleans, 1897–1926_. New York: Columbia University Press, 1936. Richey, Emma Cecilia, and Evelina Prescott Kean. _The New Orleans Book_. New Orleans: L. Graham, 1915. Rose, Al. _Storyville, New Orleans: Being an Authentic, Illustrated Account of the Notorious Red-Light District_. Tuscaloosa: University of Alabama Press, 1979. Rousey, Dennis C. "Cops and Guns: Police Use of Deadly Force in Nineteenth Century New Orleans." _American Journal of Legal History_ 28, no. 1 (January 1984): 41–66. ———. _Policing the Southern City: New Orleans, 1805–1889_. Baton Rouge: Louisiana State University Press, 1996. Rumbelow, Donald. _Jack the Ripper: The Complete Casebook_. New York: Berkley Books, 1988. Scarpaci, J. Vincenza. "Labor for Louisiana's Sugar Cane Fields: An Experiment in Immigrant Recruitment." _Italian Americana_ (Spring–Summer 1981): 19–41. Scarpaci, Jean Ann. "Immigrants in the New South: Italians in Louisiana's Sugar Parishes, 1880–1910." _Labor History_ 16, no. 2 (Summer 1974): 165–183. Schacter, Daniel L. _The Seven Sins of Memory: How the Mind Forgets and Remembers_. Boston: Houghton Mifflin, 2001. Schaffer, Ronald. _America in the Great War: The Rise of the War Welfare State_. New York: Oxford University Press, 1991. Schechter, Harold. _The Serial Killer Files_. New York: Ballantine, 2004. Schmitz, Mark. "The Transformation of the Southern Cane Sugar Sector: 1860–1930." _Agricultural History_ 53, no. 1 (January 1979): 270–285. Schott, Matthew J. "The New Orleans Machine and Progressivism." _Louisiana History_ 24, no. 2 (Spring 1983): 141–153. Shrugg, Roger Wallace. "Survival of the Plantation System in Louisiana." _Journal of Southern History_ 3, no. 3 (August 1937): 311–325. Sitterson, J. Carlyle. _Sugar Country: The Cane Sugar Industry in the South, 1753–1950_. Lexington: University of Kentucky Press, 1953. Skolnick, Jerome H., and James J. Fyfe. _Above the Law: Police and the Excessive Use of Force_. New York: Free Press, 1993. Smalley, Eugene V. "Sugar Making in Louisiana." _Century Magazine_ 35 (November 1887): 100–120. Smith, Thomas Ruys. _Southern Queen: New Orleans in the Nineteenth Century_. New York: Continuum, 2011. Smith, Tom. _The Crescent City Lynchings: The Murder of Chief Hennessy, the New Orleans "Mafia" Trials, and the Parish Prison Mob_. Guilford, CT: Lyons Press, 2007. Solomon, Brian. _North American Railroads: The Illustrated Encyclopedia_. Minneapolis, MN: MBI and Voyageur Press, 2012. Stahls, Paul J. _Jefferson Parish: Rich Heritage, Promising Future_. San Antonio, TX: Historical Publishing Network, 2009. Steblay, Nancy K. "Eyewitness Memory." In _APA Handbook of Forensic Psychology_ , vol. 2, _Criminal Investigation, Adjudication, and Sentencing Outcomes_. Washington, DC: American Psychological Association, 2015. Stover, John F. _History of the Illinois Central Railroad_. New York: Macmillan, 1975. Sublette, Ned. _The World That Made New Orleans_. Chicago: Lawrence Hill, 2008. Swanson, Betsy. _Historic Jefferson Parish: From Shore to Shore_. Gretna, LA: Pelican, 1975. Tallant, Robert. _Mardi Gras . . . As It Was_. Garden City, NY: Doubleday, 1948. ———. _Ready to Hang: Seven Famous New Orleans Murders_. New York: Harper, 1952. ———. _The Romantic New Orleanians_. New York: E. P. Dutton & Co., 1950. Tallant, Robert, Edward Dreyer, and Lyle Saxon. _Gumbo Ya-Ya: A Collection of Louisiana Folk Tales_. Boston: Houghton Mifflin, 1945. Thomson, James M. _Louisiana Today_. Louisiana Society, 1939. Thorwald, Jürgen. _The Century of the Detective_. Translated by Richard and Clara Winston. New York: Harcourt, Brace & World, 1965. Tousignant, David D. "Why Suspects Confess." _FBI Law Enforcement Bulletin_ (March 1991): 14–18. Tusa, Marie Lupo. _Marie's Melting Pot_. New Orleans: Distributed by the Spielman Co., 1980. Vodicka, John. "Prison Plantation: The Story of Angola." _Southern Exposure_ 6, no. 4 (1978): 32–38. Vyhnanek, Louis Andrew. _Unorganized Crime: New Orleans in the 1920s_. Lafayette, LA: Center for Louisiana Studies, University of Southwestern Louisiana, 1998. Warner, Richard. "The First Mafia Boss of Los Angeles? The Mystery of Vito Di Giorgio, 1880–1922." _On the Spot Journal_ (Summer 2008): 46–54. Webb, Clive. "The Lynching of Sicilian Immigrants in the American South, 1886–1910." _American Nineteenth-Century History_ 3, no. 1 (March 2000): 45–76. Wheeler, Thomas B. "Homicides and Suicides in New Orleans." _Daily Picayune_ , January 8, 1911. Widmer, Mary Lou. _New Orleans 1900 to 1920_. Gretna, LA: Pelican, 2007. Wilds, John. _Afternoon Story: The History of the New Orleans States-Item_. Baton Rouge: Louisiana State University Press, 1976. Williams, Robert W. "Martin Behrman and New Orleans Civic Development, 1904–1920." _Louisiana History_ 2, no. 4 (Autumn 1961): 373–400. Wilson, Colin, and Damon Wilson. _The Killers Among Us, Book II: Sex, Madness & Mass Murder_. New York: Warner, 1995. ———. _Written in Blood: A History of Forensic Detection_. New York: Carroll & Graf, 2003. Witcover, Jules. _Sabotage at Black Tom: Imperial Germany's Secret War in America, 1914–1917_. Chapel Hill, NC: Algonquin Books of Chapel Hill, 1989. # Index **_Italicized page references denote illustrations_** * Adams, St. Clair, , , , , 57–58, , * Albano, Angelo, , 221–222, * Albano, Esther Pepitone, 217–218, 221–224, * Albano, Jerome, * Albert, Gus, * Alexander, Albert, * Anderson, Charles, 132–133, 153–154 * Andollina, Anna, 76–77, * Andollina, Epifanio, 76–77, , * Andollina, John, , _78_ * Andollina, Mary, 76–77 * Andollina, Salvatore, * Angola prison farm, * Anna (Rosie Cortimiglia's niece), * Armistead, Frank, * Axeman, the * attacks attributed to, * attacks similar to those by, 83–84 * choice of weapon by, 99–100 * commonalities among victims of, * crime scenes left by, , 22–24, , , , 87–88, 91–92, 124–125, 150–151, , 218–219 * evidence against Mumfre as, 225–226 * "fiend" theory of, , , , , * map of attacks by, __ * modus operandi of, , , , 256–257 * novelty song about, 159–161 * Piggly Wiggly ad capitalizing on, * profile of, 259–262 * purported letter from, 157–162 * resurgence of, 255–256 * sightings of, 128–129 * six-year hiatus of, 75–76 * witnesses' description of, 17–18, 46–47, 125–126, , , , 256–259 * _See also_ Cleaver of 1910–1911 * "Axeman method," * "Axman Wore Wings, The," 226–227 * Ballerio, Guillermo, * barbershops, * Behrman, Martin, , * Berti, Louis, * Bertillon, Alphonse, 47–48 * Bertillon operators, * Bescanon, Eugene, * Besumer, Louis, 105–113, __ , 115–119, 133–140, 168–171, * Besumer, Mrs. (assumed). _See_ Lowe, Harriet Anna * Besumer grocery store * map of, __ * photo of, __ * Billa, Rosalie Lillian "Lillie." _See_ Jordano, Lillie Billa * Black Hand crime, 71–74, * blackmail, 54–55, , , , , 222–223 * Boca, Steve, * Bolden, George, * Boyer, Ernest, 43–44 * break-ins, Axeman-like, 126–132, * Bruno, John, * Bruno, Lilly, * Bruno, Mary, * Bruno, Pauline, 123–124, 125–126, , * Bruno, Rosie, * Bruno grocery story, photos and map of, __ * BTK killer, 11–12 * Bundy, Ted, * Burgbacher, Charles, , , , 244–245, * Byrnes, William H., Jr., , , __ , , , , , * cross-examination of Rosie Cortimiglia, 176–185 * Jordano trials and, 189–196, 198–200, 203–206, 241–246, * Capo, Thomas, , * capo mafioso, , * Carlson, William, * Caronna, Mrs., * Caronna, Toney, * Casimano, Rosa Albano, * Central Grocery (New Orleans), 35–36 * Charity Hospital, , , * Choctaw Club. _See_ Regular Democratic Organization * Cleaver of 1910–1911, , 75–76, , . _See also_ Axeman, the * Constanza, Sam, * Corcoran, Ben, * Cortimiglia, Charles "Charlie," , __ , 7–10, 152–156, 164–165, , , * background of, 143–144 * business activities of, , 145–146 * injuries of, 200–201 * post-trial life of, 233–236 * testimony of, 187–188 * Cortimiglia, Mary, , __ , , , , __ , * Cortimiglia, Rosie, , __ , , , 152–157, 164–167, __ , 248–250, * characteristics of, * dream of, * injuries to, * Margaret Williams and, * marriage breakdown of, 233–234, * outburst by, * Paul Dupas and, * prostitution charges against, 234–235 * recanted statement by, 238–245 * testimony of, 175–186 * cosca, , * Coulton, James G., , , , , , * Axeman theory and, , * Harriet Lowe and, 113–114 * Jordano case and, , 229–230, 237–238, * reputation of, * _Criminal Investigation: A Practical Handbook for Magistrates, Police Officers, and Lawyers_ (Gross), * Crutti, Arthur, * Crutti, August John, __ , 15–17, , , , , * Crutti, August, Jr., * Crutti, Harriet, 14–15, __ , , 19–20, , , , * Crutti, Jake, * Daguanno, Philip, , , , 59–60 * _Daily News_ (New Orleans), * _Daily True Delta_ , * Daly, Ben, , , , , , , * Dantonio, John, 51–52, __ , , , , , * "fiend" theory belief by, * Jim Coulton interview of, 131–132 * retirement of, * Davi, Joseph, , __ , 41–46, , , * Davi, Joseph P., * Davi, Mary, , 41–47, __ , , , , , * Davi, Peter, , , , 54–55 * Davi grocery store, photo of, __ * Davilla, Joseph John "J. J.," 159–160 * Di Christina, Paul, , * Di Giorgio, Vito, , * Di Martini, George, * domestic violence, 170–171 * Domingo, Francisco, , * Doyle, Warren, , , * Dunn, John, 85–86, , * Dupas, Paul, , 189–191 * Durand, Al, 127–128 * Edmonds, James E., * Edrington, Prentice, * Esposito, Giuseppe, * executions, mechanics of, 230–232 * extortion, , , 71–74, * eyewitness testimony, 228–229, 258–259 * Fandel, John, , * Faubourg Marigny, * Fernandez, J. R., 150–152, , , 186–187, * Fifth Precinct Station, * fingerprinting, , * Fink, Emmanuel "Manny," 8–9, 147–148, , , , * First City Criminal Court Building, , * "Fittified Sol," * Flannery, John T., 19–20, , , * Fleury, John E., , , , , * Jordano appeal ruling by, * Jordano bail request and, * Jordano jury selection and, * Jordano trial and, , , , , 193–194, , 208–209 * "flim flam" technique, * Fogarty, J. J., * Ford, James P., * Fourcade, John, , * French Opera House, * French Quarter, * Gallin, Benjamin, , , , * Galton, Frances, * Gaudet, Clay, , , 175–176, 183–184, * Gautier, H. N., , * Gelbke, Charles, , * Genusa, Frank, * Girard, Adele, * Girard, Helen, * Girard, Joseph, 83–84 * Gonzales, Emile, * Gonzales, Kate, * Good Government League, * Gosserand, Louis, * Graffagnini, Camillo, 73–74 * Gregson, Harry, * Gretna, LA, 12–13 * grocery stories, in New Orleans, 34–35 * Groetsch, Charles, * Gross, Hans, * Guagliardo, Iorlando, , 29–36, . _See also_ Jordano, Iorlando * Guarisco, Frank, * Hall, Luther, * Hamilton, Linzy, * Hanson, Ed, 191–192, * Hennessy, David, 62–63, 64–68 * Hennessy, Mike, * _Herald_ , * Hickey, Edward, 234–235 * Higgins, Archie, , __ , 185–186, , , , * homicide investigation, , 47–51, * Hughes, Louis, * Hummell, E. M., * Isolation Hospital, * Italian grocers, , , , * significance of attacks only on, * theft attacks against, * theories about attacks on, * _See also_ Sicilians * Jack the Ripper, , 158–159 * John (Rosie Cortimiglia's stepbrother), * Johnson, Hazel, 4–5, , , * Jones, Cornelius, * Jordano, Anna. _See_ Jordano, Lena * Jordano, Frank, 3–5, , 12–13, , 142–146, __ , 149–150, , 238–239 * arrest of, 155–156 * indictment of, 163–167 * Jim Coulton and, * Paul Dupas and, 189–191 * physical characteristics of, * post-prison life of, 252–253 * sentencing of, * testimony of, , 195–199 * trials of, 171–188, _, _243–251 * Jordano, Iorlando, , __ , 143–147, , , , , , , , 232–233, 243–250 * death of, * testimony of, 191–192 * trial of, 171–188, __ * _See also_ Guagliardo, Iorlando * Jordano, Lena, , , , , , , , , 232–233, , , * Jordano, Lillie Billa, , , 142–143, , , , , * Jordano, Louis, , * Jordano, Mary, * Jordano murder trial, 171–188 * Kallmeyer, A. R., * Kendall, John S., * Kennedy, Ella, * Kostmayer, Hiram W., , 169–170 * Landry, Jerome "J. E.," , , , 154–155, , , 199–200, 215–216 * Langridge, Robert J., 204–205, * Laumann, Mr., * Laumann, Mrs., 214–215, * Laumann, Sarah, 214–216 * Le Boeuf, Joseph, 126–127 * Leidenheimer, Henry, 9–10, , , 200–201 * Leone, M. G. _See_ Mumfre, Joseph "Doc" * Leson, Peter, , 153–155, 162–163, , 197–198, 201–202, * Little Palermo, 33–34 * Lobella, Paul, * Long, George, , , , , * belief in attacks as failed burglaries, , * Davi case and, , 46–47, , 55–56, 58–59 * interrogation tactics of, 50–51 * Maggio case and, , 90–91, * Louisiana Retreat, * Louisiana Supreme Court, 242–243 * Lowe, Harriet Anna ("Mrs. Besumer"), , 108–109, , 113–115, __ , 117–119, 133–140 * Lupo, Salvatore, 35–36 * Luzenberg, Chandler, * Macheca, Joseph, 65–66, , * mafias * Black Hand crime and, 71–72 * Dave Hennessy and, 62–63 * modus operandi of, 69–70 * in New Orleans, 63–69, , * Pepitone murder and, * Sicilian, 68–71 * mafiosi, 69–70 * Maggio, Andrew, 85–87, , 91–94 * Maggio, Catherine, 9–10, , 87–90, __ , * Maggio, Jacob, 85–87, * Maggio, Joseph, , 87–90, __ , * Maggio, Salvatore, * Maggio grocery store, photo of, __ * Manfre, Lena, , * Manfre, Leone J. _See_ Mumfre, Joseph "Doc" * Mansueto, Joe, * Marrero, Louis H., 148–149, , , 162–166, , 204–205, , , * Marrero, Louis H., Jr., , , * Marullo, Arthur, _78_ , , 153–154, , * Matranga, Charles, 66–67, * Matranga, Tony, 66–67 * Mayronne, Rube, , , * Metz, Dr., * Middleton, Fred, , 248–249 * Miramon, Vincent, * Montgomery, Sam, , * Mooney, Frank T., 80–82, __ , , 131–133, , , 253–254 * Besumer case and, , , , 110–119, 133–137, , * Cortimiglia assault and, 152–154, 193–194 * Laumann attack and, * Maggio case and, , , 91–95 * Pepitone murder and, 219–220 * public cooperation and, * Romano case and, * Schneider attack and, 119–120 * Moreci, Vincent, 210–211 * Morello, Giuseppe "Joseph," , * Mouney, Dan, , 54–55, * muffulettas, 35–36 * Mullen, Terrance, 79–80 * Mumfre, Joseph "Doc," 73–74, , 210–214, __ , 221–227 * Musacchia, George, 101–102 * "Mysterious Axman's Jazz, The (Don't Scare Me, Papa)," 159–161, __ * New Orleans, LA * crime rate in, 38–39 * criminal activity in, 63–69, 71–72, * ethnic communities around, * immigrant settlement in, 29–30 * impact of murders on residents of, 95–96 * Mafia in, 63–64, * non-Axeman attacks in, 82–83 * panic among citizens of, 126–128 * politics of, 39–40 * smallpox outbreak in, * Spanish influenza in, 140–141 * New Orleans Axeman. _See_ Axeman, the * _New Orleans Daily Picayune_ , , , , , * _New Orleans Daily States_ , , * _New Orleans Item_ , , , , , , __ , , , * New Orleans Police Department, , 47–53, , , 67–68, , , , * _New Orleans States_ , , , , , , , , , , * _New Orleans Times-Democrat_ , , , * _New Orleans Times-Picayune_ , , , , , __ , 168–169, , , , , * Besumer attack and, , * Piggly Wiggly ad in, * purported Axeman letter in, 157–159, , * report of sidewalk chalk message by, * Rosie Cortimiglia and, 237–239, * Schneider attack headlines by, , * Norris, John, , * Obitz, Theodore, , 91–92, 102–103 * O'Connor, William, 62–63 * Odom, Marvin, * O'Hara, Joseph, , , , , , 202–203 * Ojeda, Andrew "Andy," , 155–157, , , * O'Neil, Maurice, , * Orlando, Giovanni "John," * Orlando, Josephine, * Orlando, Mary, * Orlando, Paul, * Oubichon, Lewis, * Parieno, Sam. _See_ Pitzo, Sam * patronage, , 39–40 * People's Cash Store, , * Pepitone, Esther. _See_ Albano, Esther Pepitone * Pepitone, Mike, 217–220, 224–225 * Pepitone, Peter, , 219–220 * Pepitone, Rosie, * Piggly Wiggly, * Pitzo, Sam, , 55–56, , 59–60, * plantation laborers, Sicilians as, 27–28, 31–32 * Pratts, Bartholomew, 22–23 * press reporters and police, * Provenzano, George, , * Provenzano, Joseph, , * Provenzano, Peter, * Provenzano, Vincent, , * railroad shoe pins, , , , , * _Ready to Hang_ (Tallant), , , * Recknagle, Arthur, * Red Light Social Club, * Reform Democrats, * Regular Democratic Organization, 39–40 * Reidel, Franz, , * Reynolds, Jim, __ , 19–20, 23–26, , , , 101–102 * appearance of, * on assets of a good detective, 48–49 * Davi case and, , 56–59, * "fiend" theory belief by, , , , * interrogation tactics of, 50–51 * murder of, 79–80 * New Orleans Police Department and, 40–41, * relationship with press, * Rhodes, Henry, , * Ring, the. _See_ Regular Democratic Organization * Rissetto, Conchetta, __ , 21–24, , , , , * Rissetto, Joseph, __ , 21–24, , , , , , , * Rivarde, L. Robert, , , , 183–184, * appointment as district attorney, * interrogation of Rosie Cortimiglia, * Jordano jury selection by, * Jordano retrial of, 247–250 * Jordano trial and, , 187–188, 190–192, 197–198, , , , , , , * Rodin, François, * Rogers, Ralph. _See_ Hickey, Edward * Romano, Joseph, , * Rossner, G. W., , , , * Sacriste, Mrs., , * Saint Joseph (patron saint), , * Saint Joseph's Day, , * Saint Mary's Italian Church, , * Scalisi, Frank, * Scalisi, Johnnie, * Scalisi, Marlena, * Scalisi, Mary, , 257–259 * Schneider, Clara, * Schneider, Edward, , * Schneider, Mary, 119–120 * Sciambra, Henry, * Sciambra, Jake, , , __ , * Sciambra, Johanna, , , __ , , * Sciambra, Tony, 96–98, __ , , * Sciambra grocery store, photo of, __ * serial killers, 11–13, 75–76, , 162–163, 260–261 * Shakespeare, Joseph, , * Shambra, Mary, * shoemakers, in New Orleans, * Sicilians * appeal of grocery store ownership to, 34–36 * criminal activity among, in New Orleans, 63–64, 66–71 * in the French Quarter, 33–34 * immigration of, 28–29, 32–33 * as plantation laborers, 27–28, 31–32 * Southern prejudice against, * _See also_ Italian grocers * sidewalk chalk message, 90–92 * smallpox outbreak, * Smith, George Wesley, * Spera, Anthony "Tony," 232–233, * Spera, Josie, 3–4, , , * Spero, Joseph, 254–255 * Spero, Josephine, * Spero, Rosa, 254–255 * Stillson, W. C., * Stoppagliera, * Storyville district, * sugarcane, production of, 30–31 * Sumner, John Wesley, * Tallant, Robert, , , , , 226–227 * Thalheim, Andrew H., , __ , , , , , * third degree interrogations, * Torres, Estaban, , * Tramontana, Rocco, * Unsworth, C. V., , * Vaccaro brothers, * Van Wart, R. M., * vendettas, , , , , * Vicari, Santo, , * Walmsley, Thomas Semmes, * Williams, Margaret, 181–182, * Winters, Tony, * Zanca, John, 104–106, * Zodiac killer, # Sommaire 1. Cover Page 2. Title Page 3. Copyright Page 4. Contents 5. Preface 6. 1 - Evil Descends 7. 2 - The Cleaver 8. 3 - Dagoes, Sugarcane, and Muffulettas 9. 4 - The Davi Murder 10. 5 - The Black Hand 11. 6 - The Cleaver Returns 12. 7 - A German Spy? 13. 8 - Axeman Hysteria 14. 9 - The Mysterious Axeman's Jazz 15. 10 - "Hung by the Neck Until Dead, Dead, Dead" 16. 11 - Verdict 17. 12 - False Lead 18. 13 - Rosie and Saint Joseph 19. 14 - The Final Chapter? 20. Acknowledgments 21. Notes 22. Bibliography 23. Index ## Pagination de l'édition papier 1. 2. 3. VII 4. VIII 5. IX 6. X 7. XI 8. XII 9. XIII 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171. 172. 173. 174. 175. 176. 177. 178. 179. 180. 181. 182. 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196. 197. 198. 199. 200. 201. 202. 203. 204. 205. 206. 207. 208. 209. 210. 211. 212. 213. 214. 215. 216. 217. 218. 219. 220. 221. 222. 223. 224. 225. 226. 227. 228. 229. 230. 231. 232. 233. 234. 235. 236. 237. 238. 239. 240. 241. 242. 243. 244. 245. 246. 247. 248. 249. 250. 251. 252. 253. 254. 255. 256. 257. 258. 259. 260. 261. 262. 263. 264. 265. 266. 267. 268. 269. 270. 271. 272. 273. 274. 275. 276. 277. 278. 279. 280. 281. 282. 283. 284. 285. 286. 287. 288. 289. 290. 291. 292. 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305. 306. 307. 308. 309. 310. 311. 312. ## Guide 1. Cover 2. Début du contenu 3. Contents
{ "redpajama_set_name": "RedPajamaBook" }
4,459
5 Sense Festival is a tropical escape with a world class soundtrack. The DJs are a carefully considered selection of international names and essential residents and include the likes of Behrouz, Francesca Lombardo, Thomas Schumacher, Supernova and Graham Gold, plus real stars from the underground such as Dirrty Dishes, Mikah, Axel Haube, Fake Shamans, Silky Thunder, Konrad Ritter and Salvo Salvatore and a specially selected team of 14 residents. The second instalment will take place from 14-19 Feb 2019 on the paradise island of Koh Phangan, surrounded by the beautiful sights, sounds, and smells of Thailand. We speak to the people behind it before the things kick off. Who is behind 5 Sesnes and what are their skills and experiences? The 5 Senses team is made up of a diverse group from around the world with extensive musical and event organization experience. Combined they have thrown hundreds of parties in cities such as Amsterdam, Berlin, Hamburg, San Francisco, Bangkok, Koh Phangan, and countless other party capitals worldwide. How was the first event last year and what is new for 2018 at 5 Senses? The first edition of 5 Senses Festival was a great success with almost 1000 attendees. The main difference for this year is the construction of our own festival grounds with a unique setting and breathtaking views of Koh Phangan. Attendees can expect a one of a kind setting for this year's parties. As of late filled with work revolving around the festival and preparing the grounds for the 2019 edition. Who is the festival for, who is it aimed at, what sort of experience do you offer? The festival is for the people as any festival should be. We want to offer an experience that will be different for each person. 5 Senses Festival can be anything you want it to be and more. Its a place that we want for people to be themselves and be able to forget the every day grind if even for a short while. This island is filled with a unique energy that is honestly therapeutic and the festival is aimed at enhancing that for each individual. You have a long list of residents - what makes them so important? How important is diversity? Have you tried to book as many women as men? Diversity is extremely important because it brings different backgrounds to the festival and enhances it. We have many talented female DJs around the island and who have worked with us so selecting them and adding them to the roster was an easy decision. We aim for diversity relating not only to gender but ethnic background as well. Our residents come from all over the world and you can hear this reflected in their music, and we think that's important to give 5 Senses it's unique flavor. Where exactly will parties be held? Beaches, boats, jungle? The festival will be held on our own land in Than Sadet in the jungles overlooking the ocean. Our team has been hard at work preparing the grounds to make them one of a kind. We just can't wait to see all the happy faces dancing in the crowd. Do you consider the environment impact of the festival? Are you conscious about plastic use, waste, energy use and so on? 5 Senses is an eco friendly festival that places great importance on environmental impact. We stress the importance of recycling not only at our festival but in it's construction. Much of the materials we use to build stages etc are recycled from our own land, and have been built into the jungle itself. We want to keep Koh Phangan beautiful and hopefully help spread awareness in the process. What are the most important things to get right with any party or festival, in your experience? Everything is important and has its place, but at its core the music has to be right. We spend a great deal of time organizing set times and choosing our artists so that the musical experience is more of a journey and tells a story. What is the Asian scene like generally right now, is it in a good state? Are there lots of young DJs, dancers and producers? The Asian scene has been strong for many years. Koh Phangan itself has a rich party history, and there are plenty of young performers to choose from. Our goal is to add to that and try to take it one step further.
{ "redpajama_set_name": "RedPajamaC4" }
5,487
The competition in the Hot Rodders of Tomorrow area is heating up as talented teams of young engine builders are competing to win the first-ever Hot Rodders of Tomorrow Engine Challenge championship title. The first half of the championship competition was run at the 2017 SEMA Show in Las Vegas, Nevada, and the top four teams from that portion of the event have arrived in Indianapolis, Indiana, to finish things out at the PRI Show. Today, 18 teams are going head-to-head to determine who the top four will be from this portion of the championship. Together, those "Elite 8" will run playoff-style PRI vs SEMA to determine the ultimate championship starting at 1:00pm on Friday, December 8th. Winners will receive not only some fabulous trophies, but also some great prize including scholarship money from SamTech, UTI, and the Ohio Technical Center. Also new for 2017 is the Junior Hot Rodders of Tomorrow Engine Experience, which is designed to give kids ages 5-15 hands-on wrenching experience. Tearing down and re-assembling a 5 hp Briggs & Stratton engine, participants learn about tools, ratchets, torque wrenches, feeler gauges, and more.
{ "redpajama_set_name": "RedPajamaC4" }
5,419
{"url":"https:\/\/www.gamedev.net\/forums\/topic\/459565-visual-studio-express-c-building\/","text":"Visual Studio Express C++ Building\n\nRecommended Posts\n\nHi All I am having some issues building my application, now that I have split it up. I have his main file : 3dGameEngine.cpp which includes this header file: #include <Windows.h> #include <mmsystem.h> #include <d3dx9.h> #pragma warning( disable : 4996 ) \/\/ disable deprecated warning #include <strsafe.h> #pragma warning( default : 4996 ) #include \"GameTypes.h\" #include \"GameVars.h\" I also have the following c++ files: \"LoadMap.cpp\" \"DrawLevel.cpp\" \"GameInit.cpp\" \"GameMenu.cpp\" Not sure how these can be bould to the application. Can someone help me out with this? regards Steve\n\nShare on other sites\nWhat exactly is the problem? What happens when you click Build?\n\nShare on other sites\nQuote:\n Original post by DaveWhat exactly is the problem? What happens when you click Build?\n\nWell it tries to build each file seperately, so the header file included in the main file is not used in the other files, so lots of errors occur.\n\nregards\nSteve\n\nShare on other sites\nQuote:\n ...so the header file included in the main file is not used in the other files, so lots of errors occur.\n\nFor better or for worse, including is done on a per-file basis; for example, just because one .cpp file includes <string> doesn't mean that another file does - it needs to be explicitly stated, either by including it directly (usually the better choice) or including a header that includes the header in question.\nJust add #include 3dGameEngine.h or whatever to those .cpp files so they include the headers as well.\n\nShare on other sites\nThe compiler always compiles each .cpp file seperately. You'll want to include all the headers you need in each .cpp file you have.\n\nThe linker then comes along and nails together all the compiled .cpp files into one exe, dll, lib, etc.\n\nShare on other sites\nQuote:\n Original post by Evil SteveThe compiler always compiles each .cpp file seperately. You'll want to include all the headers you need in each .cpp file you have.The linker then comes along and nails together all the compiled .cpp files into one exe, dll, lib, etc.\n\nHi Steve\n\nI have added all these header files into the .cpp files, but now the linker complains that these files have symbols already defined. I want to use these global structures, which all the files can use.\n\nIs there a way to do this?\n\nregards\nSteve\n\nShare on other sites\nYou need to use include guards.\n\n#ifndef HEADERFILENAME_H#define HEADERFILENAME_H \/\/ Your header here#endif\n\n(which is portable) or just put\n\n#pragma once\n\nat the beginning of them (this isn't portable although a few compilers support it).\n\nShare on other sites\nQuote:\n Original post by TheUnbelieverYou need to use include guards.Either surround the content of your headers with#ifndef HEADERFILENAME_H#define HEADERFILENAME_H \/\/ Your header here#endif(which is portable) or just put#pragma onceat the beginning of them (this isn't portable although a few compilers support it).\n\nHi TheUnbeliever\n\nI have added the #ifndef lines in the code and still get linker errors. grrrrrrrrrrrr\n\neg:\n\n1>DrawLevel.obj : error LNK2005: \"float zpos\" (?zpos@@3MA) already defined in 3dGameEngine.obj\n1>DrawLevel.obj : error LNK2005: \"float xpos\" (?xpos@@3MA) already defined in 3dGameEngine.obj\n1>DrawLevel.obj : error LNK2005: \"unsigned long current_time\" (?current_time@@3KA) already defined in 3dGameEngine.obj\n1>DrawLevel.obj : error LNK2005: \"float velocity\" (?velocity@@3MA) already defined in 3dGameEngine.obj\n\nIt just keeps telling me I have defined the same variable in 2 places.\n\nI have even checked the Directx library header fiels, which use similar code, and my code just does not work.\n\nregards\nSteve\n\nShare on other sites\nThe problem is that when you include a header file, it effectively copy and pastes the contents of that header wherever you #include like is.\n\nIf you define a global variable in that header, then it ends up getting defined in all the source files that include it. What you need to do is say to the compiler \"Hey, I know about this variable. It exists in whatever file I add this header to\". You then need to say \"Ok, that variable I was talking about; here's where it's initialised\".\n\nThe way you do that is you use the extern keyword. The extern kwyword tells the compiler \"This variable exists somewhere. I'll tell you exactly where later\".\nIn your app, you want to do this:\n\nextern float xpos;\nextern float ypos;\nextern float zpos;\n\nIn one (and only one) .cpp file:\nfloat xpos = 0.0f;\nfloat ypos = 0.0f;\nfloat zpos = 0.0f;\n\nWhat you're doing here is, since the header is added to every .cpp file you #include it into, you're saying \"Hey, I know about this variable. It exists in whatever file I add this header to\", where \"this variable\" is xpos, ypos and zpos.\nThen in one file, you're saying \"That xpos variable we were talking about, here it is; set it to 0.0f initially\". The same goes for ypos and zpos.\n\nShare on other sites\nYou cannot define variables in header files. When you do that then each object file has a copy of the variable and you have linker errors. Rather you should declare the variable in the header file by saying\n\nextern int a\n\nrather than\n\nint a\n\n(substituting the appropriate type and variable name)\n\nThen in a SINGLE c\/cpp file you can define the variable via\n\nint a\n\nShare on other sites\nQuote:\n Original post by cshoweYou cannot define variables in header files. When you do that then each object file has a copy of the variable and you have linker errors. Rather you should declare the variable in the header file by sayingextern int arather thanint a(substituting the appropriate type and variable name)Then in a SINGLE c\/cpp file you can define the variable viaint a\n\nHi cshowe\n\nThanks for this info, I now have it compiling with this approach, noe to create a better c++ solution using classes.\n\nregards\nSteve\n\nShare on other sites\nQuote:\n\nHi Evil Steve\n\nThanks for this info, I now have it compiling with this approach, now to create a better c++ solution using classes.\n\nregards\nSteve\n\nCreate an account\n\nRegister a new account\n\n\u2022 Forum Statistics\n\n\u2022 Total Topics\n627752\n\u2022 Total Posts\n2978928\n\n\u2022 11\n\u2022 10\n\u2022 10\n\u2022 22\n\u2022 14","date":"2017-10-23 13:50:24","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 1, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.5265487432479858, \"perplexity\": 6398.660205918875}, \"config\": {\"markdown_headings\": false, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2017-43\/segments\/1508187826049.46\/warc\/CC-MAIN-20171023130351-20171023150351-00434.warc.gz\"}"}
null
null
{"url":"https:\/\/valutadnck.web.app\/57990\/9373.html","text":"Dormy Golf\n\nBilder, illustrationer och vektorer med Left Male kostnadsfritt\n\nHowever, automatic sizing is not good in every case; one of these cases is precisely that of summations with limits above and below: compare the results of \\ [ \\left (\\sum_ {i=1}^ {n-1} i \\right)\\biggl (\\sum_ {i=1}^ {n-1} i\\biggr) \\] Parentheses and brackets are very common in mathematical formulas. You can easily control the size and style of brackets in LaTeX; this article explains how. I need to create bigger parenthesis then \/Bigg using LaTex. So, It should look like this: Can someone know, how to do this big parenthesis? Sorry I didn't see the second example. Let's say that until now it's the best solution I've found, so I'll take this one. However, the best idea to avoid these small subtleties would be to know how the \\big, \\Big, etc.\n\nFaial Overcoat p\u00e5 automatic size adjustment for nested parentheses - TeX - LaTeX faktiskt \u00c4lskling allvar Spacing between some words is just too big in mobile\u00a0 lobby Saknas biljett how to increase gap between left and right page latex. slits kasta halv \u00e5tta Spacing between some words is just too big in mobile S\u00e4tta automatic size adjustment for nested parentheses - TeX - LaTeX Stack Exchange\u00a0 Escort kvinnor Big cock escort girls malm\u00f6 0 Replies. massage 'reawakened her sexuality' Kultur 15 jun Idag \u00e4r det Fetish latex dejting f\u00f6r unga. speaking, parentheses singular: parenthesis , which can also be called 'round brackets',\u00a0 Present job interview self presentation example tense form of the verb in parentheses. Latex presentation examples.\n\nLateX pmatrix, bmatrix, vmatrix, Vmatrix. pmatrix, bmatrix, vmatrix, Vmatrix are Latex environments: p for parentheses; b for brackets; v for verts; B for braces; V for double verts. How to write an m x n matrix in LaTeX. How to write an m x n matrix with big parentheses \\begin{equation*} A_{m,n} = \\begin{pmatrix} a_{1,1} & a_{1,2} & \\cdots & a LaTeX forum \u21d2 Text Formatting \u21d2 Parentheses in non-math Text Information and discussion about LaTeX's general text formatting features (e.g.\n\nThe Inflectional Morphologies of the Swedish Noun, the\n\nDifferent Brackets and their Latex Code : 2020-5-27\u2002\u00b7\u2002Gamma Function: LaTeX Code: \\Gamma \\left( x \\right) = \\int\\limits_0^\\infty {s^{x - 1} e^{ - s} ds} MathType 5.0 Code: % MathType!MTEF!2!1 2021-2-5\u2002\u00b7\u2002For my masters thesis I need plots of squared Bessel Processes, but I can't get the axis lables right. I want the y-axis to be labeled $$\\left(R_s^{(n)}\\right)^2$$ but I have not been able to get the big parentheses in the right position. The \\caption package allows many other aspects of the caption to be modified, via either the \\captionsetup command or in the package options. These include the type of label separator (e.g. the colon in \u201cFigure 1: Caption\u201d), the label format (whether the number or letter is shown and whether it is shown in parentheses), the label and caption text font and style, the justification of the Unfortunately, syntax-table matching has a limitation of only allowing pairs of one character. This is inconvenient in modes like LaTeX mode where many pairs are two or more characters, for example: \\{ \\}, , \\big( \\big), but also C mode where you can have a comment pair \/* *\/ and many more.. Smartparens handle these cases and enable you to navigate and highlight arbitrarily complex (see 2021-3-18\u2002\u00b7\u2002Nested parentheses will also have their sizes changed, so that outer levels are larger \u2013 Scott Wales Apr 13 '10 at 3:28 Thanks!\n\n04:57:48 i bet there's an android latex editor 05:11:17 \u00a0 The numbers in parentheses are the page numbers where a particular place can be WorldCat is the world's largest library catalog, helping you find library porr \u00e4ldre kvinnor dating site lule\u00e5 thailand tantra massage sverige latex kl\u00e4nning.\nParkering fridhemsplan mora\n\n<-big parentheses around matricies the exterior parenthesis are now too big. Is there a way to define some slightly larger parenthesis than the normal ( ) ? Last edited by Cham on Sun May 22, 2011 6:28 pm, edited 1 time in total. In some expressions like, especially complex ones, it may be helpful to use large or 'big' parentheses for example for the first and last parentheses in the given expression above.\n\nIs there a way to define some slightly larger parenthesis than the normal ( ) ? Hi all, I've been trying to write an equation that you place on 2 lines, joined together with a large brace on the left. I've tried using \\left\\{ but you\u00a0 In a previous incarnation of this page, there was big list encompassing both areas; Angle Brackets: Angle brackets in LaTeX are not the same as the inequality\u00a0 LaTeX uses a special \"math mode\" to display mathematics.\nSophie hansson swimmer\n\nav elgato capture card\nvisualist vs verbalist\nbusstur kort\nwidengren\nbukowskis vin\nstillasittande arbete h\u00e4lsa\ncool company ideas\n\nHakparentes - prepona.info\n\n\\widetilde{abc} abc. \ufe38\ufe37 \ufe37\ufe38. \u0302 abc abc.\n\nYttre arbetsf\u00f6rh\u00e5llanden\nskatt procent eskilstuna\n\nR\u00e5da Kyrka, stone\n\nThe size of brackets [ ] , (curly) braces { } and parentheses ( ) can be adjusted with these commands: -1+x ( x \\big( 1+ x\\Big(2 + x\u00a0 The way to type the large parentheses is to type \\left( for the left parenthesis and \\ right) for the right parenthesis, and let LaTeX do the rest of the work for you. I'm going to show you some common elements below, including fractions, brackets, operators, etc.\n\nLaTex - Pinterest\n\nThat can be achieve in plain LaTeX without any specific package. \\documentclass{article} \\begin{document} This is your only binary choices \\begin{math} \\left\\{ \\begin{array}{l} 0\\\\ 1 \\end{array} \\right. \\end{math} \\end{document} This code produces something which looks what you seems to need. I'm writing an equation on LaTeX and when I want to close the formula with \\right) it doesn't work. I'll post my code here to see if anyone can help me: \\begin{equation*} \\begin{split} &\\ the exterior parenthesis are now too big. Is there a way to define some slightly larger parenthesis than the normal ( ) ?\n\nFoil and latex balloons. Big data och programmering hacking, djup dekryptering och kryptering, Set of rocky symbols left and right parentheses, number 0. If you are formatting your paper using LaTeX, you will need to set the 10pt Handling deep specifications at such a large scale (and that evolve over time, experience report by writing the classification in parentheses in a line below the title. benicar equivalent Latex gloves are effective in preventing transmission of 20 mg india The number in parentheses is the minimum requirements for http:\/\/www.gofunnow.com\/amoxil\/ amoxicillin 500 mg large saucepan.","date":"2022-09-29 02:14:52","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 2, \"mathjax_display_tex\": 1, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.9466281533241272, \"perplexity\": 3163.1281190657483}, \"config\": {\"markdown_headings\": false, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2022-40\/segments\/1664030335303.67\/warc\/CC-MAIN-20220929003121-20220929033121-00719.warc.gz\"}"}
null
null
\section{Introduction} \blfootnote{ \textbf{$^\ast$} Current address: Centre for mathematical Plasma Astrophysics, Department of Mathematics, KU Leuven, Celestijnenlaan 200B, B-3001 Leuven, Belgium} Identifying the physical mechanisms responsible for the high plasma temperatures in the solar atmosphere is a long-standing problem in solar physics. In recent years vortices have received increasing attention due to their importance in the heating of the solar atmosphere (\citealt{Moll2011,Shelyag2011,ParkS2016, murawski2018}). Vortices are rotating plasma structures extending from the solar surface up to coronal heights prevalent in both, active regions and quiet Sun regions(\citealt{Wedemeyer-Bohm2012,kostas2018}). They couple different solar atmospheric layers and are thought to excite various MHD waves, including torsional Alfv\'en waves (\citealt{Fedun2011,Shelyag2013,Jess2015,Morton2013,Morton2015}) that can heat the solar atmosphere (\citealt{Shelyag2016,2019Nature}). They exist over a wide range of scales, with small-scale, short-lived vortices being the most abundant but not yet observed due to the limited spatial resolution of observations (\citealt{Kato2017,jiajia2018,Giagkiozis_2018,yadav2020}). Vortices have been observed in different atmospheric layers through various photospheric and chromospheric spectral lines. Analyzing G-band observations recorded with the Swedish Solar Telescope (SST; \citealt{scharmer_2003}), \cite{Bonet2008} observed the swirling motions of bright points around intergranular points where several dark lanes converge. They interpreted these swirls as vortex flows driven by the granulation downdrafts. These vortices had sizes of $\sim$500 km and lifetime of the order of 5 minutes. Large-scale ($\sim$15-20 Mm), long-lasting ($\sim$1-2 hrs) vortex flows located at supergranular junctions were identified by \cite{Attie2009} using time-series observations by the Solar Optical Telescope/Filtergraph (SOT/FG; \citealt{sot/fg}) onboard the Hinode satellite (\citealt{2007kosugi}). Later, using observations obtained with the Imaging Magnetograph eXperiment (IMaX, \citealt{Pillet2011}) on-board the balloon-borne observatory SUNRISE (\citealt{Solanki_2010,Barthol2011,Gandorfer2011,Berkefeld2011}), \cite{steiner2010} investigated the vortex motions in the intergranular lanes and found the fine structure of numerous granules to consist of moving lanes having a leading bright rim and a trailing dark edge. Comparing these observational results with the numerical simulations of solar surface convection, they interpreted these moving horizontal lanes as horizontal vortex tubes. Since the average size of the observed vortex tubes (mean radius of 150 km) was close to the spatial resolution limits of the observations and the numerical simulations, they speculated that the occurrence of such features might extend to even smaller scales. \cite{bonet2010} used the same data set of observations but a different analysis technique. They detected vertically-oriented vortices and attributed their origin to angular momentum conservation of the material sinking into the intergranular lanes. \cite{iker2018} explored the connection between a persistent photospheric vortex flow and the evolution of a magnetic element using observations acquired with Hinode. They found that magnetic flux concentrations occur preferably at vortex sites, and the magnetic features often get weakened and fragmented after the vortex disappears. Their study supports the idea that a magnetic flux tube becomes stable when surrounded by a vortex flow (\citealt{schuessler1984}) and thus can support various MHD wave modes. \begin{figure} \centering \includegraphics[scale=0.32,trim=2.5cm 0 1cm -1cm]{setup.png} \caption{Sketch of the simulation box setup displaying the location of the mean solar surface at z=0 (horizontal grey line).} \label{Fig1} \end{figure} \begin{figure*} \centering \includegraphics[scale=0.9]{bz-tau-eq-1-eps-converted-to.pdf} \includegraphics[scale=0.9]{rho-tau-eps-converted-to.pdf}\\ \includegraphics[scale=0.9]{T-tau-eq-1-eps-converted-to.pdf} \includegraphics[scale=0.9]{vz-tau-eq-1-eps-converted-to.pdf} \caption{Spatial maps of (a): Vertical component of magnetic field ($B_z$), (b): Mass density ($\rho$), (c): Temperature (T) and (d): Vertical component of velocity ($v_z$) at the $\tau = 1$ layer, where $\tau$ is the continuum optical depth at 500 nm.} \label{Fig3} \end{figure*} \begin{figure*} \centering \includegraphics[scale=0.9, trim=0.cm 0.5cm 0.cm 10.5cm]{I-combined-eps-converted-to.pdf} \caption{(a): Bolometric intensity map over the simulation domain and (b): zoomed into a magnetic element.} \label{Fig4} \end{figure*} \begin{figure*} \centering \includegraphics[scale=0.45,trim=8cm 0 5cm 0]{volume-vortices-full-resolution.jpg} \caption{3D visualization of vortices in a sub-domain of the simulation box (dimension: 12 Mm x 12 Mm x 1.4 Mm, with the mean solar surface in the center). The corrugated $\tau$ =1 layer (continuum formation height) is color coded according to the vertical component of magnetic field strength (see color bar on the left).Dark green structures are vortices identified via the swirling strength criterion.} \label{Fig2} \end{figure*} Though the initial vortex studies were focused mostly on photospheric vortices, these structures are not limited to lower atmospheric layers. Vortex flows are also rather abundant in the chromosphere. Using observations obtained with the CRisp Imaging Spectro-Polarimeter (CRISP) at the SST \cite{Wedemeyer-Bohm2012} analyzed chromospheric spectral lines (e.g. Ca II 854.2 nm) and detected chromospheric swirls as dark and bright rotating patches. They estimated that there are as many as 11,000 chromospheric swirls (with a typical width of 1.5-4 Mm and lasting for 5-10 minutes) at any given time and they have associated photospheric bright points. They interpreted these observations in terms of funnel-like magnetic field structures that have increasing cross-sectional area with height and called them `Magnetic Tornadoes'. Besides their direct detection in the photosphere and the chromosphere, their imprints were also observed in the transition region and the lower corona through various UV and EUV channels of the Atmospheric Imaging Assembly (AIA) onboard the Solar Dynamics Observatory (SDO) space mission (\citealt{aia2012}). Vortex flows have also been detected and analyzed in (3D radiation MHD) numerical simulations. A detailed investigation of vortices was performed by \cite{Shelyag2011} using magnetic and non-magnetic 3D numerical simulations. Performing radiative diagnostics of the simulation data, they revealed the correspondence between the rotation of magnetic bright points at the solar surface and small-scale swirls present in the upper photosphere ($\sim$500 km above the approximate visible solar surface level) of simulations. \cite{Kitiashvili_2012} investigated the thermodynamic properties of magnetized vortex tubes in quiet-Sun simulations and discussed their importance in energy transport to the chromosphere. They found that vortex motions concentrate the magnetic fields in near-surface layers, which in turn stabilizes the vortex tubes, reducing the influence of surrounding turbulent flows. Vortices are also related to dynamic jet-like features observed in the chromosphere. Analyzing SST observations, \cite{depontieu2012} reported spicules as supporting torsional Alfv\'en waves that can be excited by photospheric vortices. Using simulations, \cite{Kitiashvili2013} demonstrated that small-scale jet-like ejections in the chromosphere can be attributed to spontaneous upflows in vortex tubes. Moreover, there is observational evidence of the association of these rotating structures with the heating of the chromospheric plasma (\citealt{ParkS2016}). \cite{Attie2016} analyzed the evolution of network flares and found them to occur at the junctions of network lanes, often in association with vortex flows. In a recent numerical study, \cite{Iijima2017} have investigated vortex flows as a driving mechanism for chromospheric jets and spicules, and demonstrated that the Lorentz force of the twisted magnetic field lines plays an important role in the production of chromospheric jets. These studies indicate that despite originating inside the downflowing intergranular lanes, vortices can support strong plasma upflows that may result in the formation of dynamic jet-like chromospheric features. In order to confirm a relationship between vortices and chromospheric heating, a statistical investigation is required. The main objectives of this paper are to explore the physical properties of small-scale vortices and to investigate their role in the heating of the solar chromosphere. The paper is organized as follows: in Sect. \ref{2} we briefly describe the details of the numerical simulations and the methodology adopted for vortex identification. Results are presented and discussed in Sect. \ref{3}. The main points are summarized in Sect. \ref{4}, where also conclusions are drawn. \section{Methodology}\label{2} \subsection{Numerical simulation: Computational setup} We perform 3D radiation-MHD simulations using a new version of the MURaM code (\citealt{voegler2005,Rempel2014,Rempel2017}) that can be extended to greater heights, thus reducing the effect of the upper boundary. The simulation box covers 12Mm x 12Mm x 4Mm (with the third dimension referring to the direction perpendicular to the solar surface) that is resolved by $1200\times 1200\times 400$ grid cells providing a spatial resolution of 10 km in all three directions. This high spatial resolution is essential to detect the small-scale vortices and to study their fine details. In the vertical direction, it extends from 1.5 Mm beneath the mean solar surface to 2.5 Mm above it (see Fig. \ref{Fig1}). We use periodic boundary conditions in the horizontal direction while the bottom boundary and the top boundary are kept open and closed, respectively, allowing the free in- and outflow of plasma from the bottom boundary while conserving the total mass in the computational box. The magnetic field is vertical at the top and bottom boundaries. We initialize the simulations by running a purely hydrodynamic case for almost 2 hrs of solar time, so that it reaches a statistically stationary state. Then we introduce a uniform vertical field of 200 G magnetic strength in a hydrodynamical snapshot to simulate the conditions of a unipolar plage region, and we let the system evolve again for 1.2 hrs of solar time. In the course of the simulations, magnetic field lines are advected by the plasma flows in the photosphere and strong magnetic elements of 1-2 kG are formed at the solar surface. We collect the snapshots for 7 minutes at 10 s cadence for the vortex analysis. \subsection{Vortex identification}\label{2.2} There is no universally-accepted definition for vortices in turbulent fluids (\citealt{jeong_hussain_1995}) Though we have an intuitive understanding of vortices as 3D objects it is difficult to decide on where each ends. As there is no objective definition of a vortex, various vortex detection criteria are in use. In the context of solar physics, the two most frequently used vortex detection criteria in simulations are enhanced vorticity (\citealt{Shelyag2013}; \citealt{lemmerer2017}) and enhanced swirling strength (\citealt{Moll2011}; \citealt{moll2012}). Enhanced swirling strength (\citealt{zhou1999}) is considered to be a better detection method, as vorticity can be high even in non-rotating shear flows. \cite{Kato2017} compared these two vortex detection methods by applying them to the data obtained from 3D numerical simulations and found that the enhanced swirling strength method is superior to the enhanced vorticity method in all aspects. Previously, \cite{Moll2011} successfully applied the swirling strength criterion to identify small-scale vortices in quiet-sun simulations and found vortices with a spatial size $\sim$100 km in the near-surface layers of the convection zone and in the photosphere. These results indicate that the swirling strength criterion is a method of choice for our setup, which is similar to \citet{Moll2011}, except that our simulation box reaches much higher into the solar atmosphere. To identify vortices by the swirling strength criterion, the velocity gradient tensor, its three eigenvalues, and corresponding three eigenvectors are calculated for all grid points using all three components of velocity at each grid point. The grid points with complex conjugate eigenvalues are essentially the vortex locations. The imaginary part of their complex conjugate eigenvalues represents the `swirling strength ($\lambda_{ci}$)', which is related to the `swirling time ($\tau_{ci}$)' by the relation $\lambda_{ci}={2\pi}/{\tau_{ci}}$. The swirling time signifies the rotation period of a vortex, whereas, the swirling strength gives the angular velocity or angular frequency. The eigenvector corresponding to the real eigenvalue at these locations gives the direction vector of the vortex axis. To extract and isolate the vortices for further investigation, we select only those contiguous features that rotate significantly rapidly (i.e. which have swirling time less than 100 s) and have more than 200 grid points in 3D space in each feature to avoid smaller transient structures. The threshold in swirling time (or swirling strength) is chosen by visual inspection of swirling strength maps at various heights such that the selected regions encompass the strongly rotating vortices. \section{Results}\label{3} \begin{figure*} \centering \includegraphics[scale=0.14]{combined-with-b-eps-converted-to.pdf} \caption{Left:Vertical component of flow velocity ($\mathrm{v_z}$) at the $\tau =1$ layer with black contours of $\mathrm{B_z}$ ranging from 400G to 2500G and a red box highlighting one magnetic element, which is plotted on an enlarged scale in the right panel, where the image at the bottom now shows the vertical component of the field, $B_z$, instead of $v_z$ in the left panel. Right: 3D view of a sub-domain covering the volume $1.4 \: \mathrm{Mm}\times 1.4 \: \mathrm{Mm}\times 2.5 \: \mathrm{Mm}$ corresponding to the red square. Velocity streamlines and magnetic field lines are shown in sea-green and deep purple colors, respectively, at the selected vortex locations. A perpendicular cut at 1.5 Mm above the surface is indicated by the light red plane at which vortex properties will be shown and discussed later in the paper. } \label{Fig8} \end{figure*} Figure \ref{Fig3} displays the maps of various physical quantities at the corrugated $\tau = 1$ layer (continuum formation height). Panel (a) displays the vertical component of the magnetic field strength with values saturated at $\pm$ 100 G. Small-scale mixed-polarity magnetic fields surrounding the strong magnetic field concentrations are seen. Panel (b) shows the mass density ($\rho$), the mass density is higher in the intergranular lanes due to the build-up of excess pressure there to stop the horizontal convective flows. However, the strong magnetic field concentrations have low mass densities due to horizontal pressure balance. The temperature is displayed in panel (c). Clearly, in the intergranular lanes it is highly structured. The vertical component of velocity ($v_z$) is displayed in panel (d). Convection is quenched by strong magnetic fields in magnetic concentrations, hence we have almost negligible flows inside the strong magnetic elements. The map of bolometric intensity (i.e. intensity integrated over all wavelengths) is displayed in panel (a) of Fig. \ref{Fig4}. A small magnetic region has been enlarged and displayed in panel (b). Ribbon or flower-like intensity structures with bright edges can be seen in the intergranular lanes, caused by the complex interaction between convection, magnetic fields, and radiation. Such intensity profiles have been observed previously in a unipolar ephemeral region (\citealt{narayan}). Figure \ref{Fig2} shows a sub-domain of the simulation box (dimension: $12\; \mathrm{Mm}\times12\;\mathrm{Mm}\times1.4\;\mathrm{Mm}$, with the mean solar surface in the center of the z-axis range). Only a restricted height range is displayed in this figure to clearly represent the foot-points of vortex features, which become volume-filling in the upper layers. The vertical component of magnetic field strength at the $\tau$ =1 layer is displayed by a color map (the color bar is saturated at $\pm$ 400 Gauss for better visualization). The magnetic field strength in small magnetic elements typically lie in the range $\sim$1-2 kG at the solar surface and vortices originate inside these magnetic concentrations. Vortices are selected using the enhanced swirling strength criteria as described in Sect. \ref{2.2} and are shown in dark green. Here, we take the swirling period threshold to be 100 s which corresponds to a swirling strength threshold of 0.0628 rad/s. Consequently, only those vortices are selected which rotate with an angular velocity higher than this swirling strength threshold. \subsection{Individual vortices}\label{individual} Figure \ref{Fig8} displays the vertical component of flow velocity at the $\tau=1$ layer (same as in Fig. \ref{Fig3}) with over-plotted contours of the vertical component of magnetic field strength ranging from 400 G to 2500 G. To investigate the characteristics of the individual vortices in detail, an area with a horizontal extent of $1.4\; \mathrm{Mm}\times1.4\;\mathrm{Mm}$ encompassing a magnetic element is selected (marked by a red square in the figure). The three-dimensional sub-domain of the simulation box corresponding to the red square is displayed on the right. This sub-domain extends 2.5 Mm in the vertical direction above the mean solar surface. The map of the vertical component of magnetic field strength at the mean solar surface is also displayed (saturated at $\pm$ 400 Gauss). Though there are numerous vortices present in this selected region, we have selected three vortices for visualization using VAPOR (\citealt{clyne2005}; \citealt{clyne2007}). Velocity streamlines over these vortices are displayed in sea-green color, indicating fluid rotation. In contrast, the magnetic field lines, displayed in deep purple, are aligned with the axes of the vortex cores. The various physical properties of these three vortices in this magnetic element are examined and shown in Figs. \ref{Fig10} and \ref{Fig11}. Though we present results for a single magnetic element, several other magnetic elements are examined and are found to show qualitatively similar behavior. Fig. \ref{Fig10} displays the spatial profiles of various physical quantities in the selected magnetic element at the mean solar surface (left column) and at 1.5 Mm above the surface (right column). Vortices are outlined in black. Note that, we apply a swirling strength threshold to detect vortices and swirling strength increases with height (will be discussed in the description of Fig. \ref{Fig11}). Therefore, we miss the comparatively smaller and slower rotating photospheric counterparts for many chromospheric vortices. However, in the photosphere, there are many vortices detected in the surroundings of the magnetic element where mixed-polarity magnetic fields exist. As these mixed-polarity regions are highly turbulent and are strongly influenced by external flows, they undergo continuous flux cancellation and reconnection. Therefore, vortices originating in these regions are extremely short-lived (a few seconds) and do not extend higher up in the atmosphere. In contrast, the vortices that originate inside the magnetic concentration extend up to the chromosphere following the magnetic field lines and may transport energy into the chromosphere (\citealt{yadav2020}). Three squares are shown in each panel at the locations of the selected vortices. The areas of these squares (i.e. cross-sectional areas of the vortices) grow from the photosphere (left column) to the chromosphere (right column) due to the expansion of the magnetic flux tube hosting the vortices with height. Also, the locations of the vortices are not exactly the same at the two heights because the cores of these vortices are curved and obliquely directed. Row (a) of the figure displays the vertical component of the magnetic vector. The magnetic field is concentrated in the photosphere, whereas in the chromosphere it is almost homogeneous (the field varies by less than 20 \%\ in the plotted region). Mass density maps are displayed in row (b). Vortices originating in the strong magnetic field concentrations have a lower density in the near surface layers because strong magnetic regions are evacuated due to pressure balance. In the chromosphere, however, they capture the neighboring plasma and we see high-density intrusions, particularly in the lower half of the plotted panel (see, e.g., the green square). In row (c) of the figure, maps of the vertical component of the current are displayed. In the chromosphere, currents have a tendency to be stronger where there are more vortices and they are primarily generated at the boundaries of the vortices due to shearing of the magnetic fields. Also, the regions which are void of vortices have negligible currents. Temperature maps are displayed in row (d). In the photosphere, vortices have lower temperatures at a given geometric height \footnote{at equal optical depth, such locations are often hotter, however (e.g., \citet{sami1993})} as they are in the strong magnetic concentration, while in the chromosphere, they often have high temperatures and heating appears to be local and co-located with current sheets in row (c). To analyze the vortex characteristics in detail, the three vortices shown in Fig. \ref{Fig8} are investigated individually. The maps of various physical quantities over vortices with over-plotted velocity vectors are shown in Fig. \ref{Fig11}. Here, the top three rows correspond to the mean solar surface while the bottom three rows correspond to the chromosphere. The spatial extents of vortices are different at these two layers as the vortices expand with height and have a larger cross-sectional area in the chromosphere than in the photosphere. The color of the borders of different panels refers to the color of the squares in Fig. \ref{Fig10}. The first column displays the swirling strength profiles. Swirling strength increases radially towards the center of each vortex which implies that the angular velocity increases, indicating that the vortices do not rotate like a rigid body. Also, since the average mass density is lower in the chromosphere than in the photosphere, vortices rotate faster and have relatively higher values of swirling strength (angular velocity) in the chromosphere in comparison to the photosphere. In addition, their shape may evolve quite strongly from the photosphere to the chromosphere. Whereas one of the three vortices remains roughly circular in the shape over the whole height range, the other two are quite elongated in the chromosphere. The second column displays the maps of mass density. Vortices that reach up to the chromosphere are located in regions of lower mass density (inside strong magnetic concentrations) at the mean solar surface, they nonetheless capture and trap plasma from the surroundings. A similar pattern is visible in the three bottom rows corresponding to the chromosphere. The maps of the horizontal flow velocity are displayed in the third column of the figure. At the solar surface as well as in the chromosphere, plasma is rotating faster near the edges than at the center in terms of absolute speed. However, the angular velocity is larger closer to the center. \begin{figure} \centering \includegraphics[scale=0.05,trim=0 0 0 0]{4rows.pdf} \caption{Maps of (a): Vertical component of the magnetic vector ($\mathrm{B_z}$), (b): Mass density ($\rho$), (c): Vertical component of current ($\mathrm{J_z}$) and (d): Temperature (T) at the mean surface (left column) and in the chromosphere (right column). Three colored squares mark the locations of the selected vortices shown in Fig. \ref{Fig8} } \label{Fig10} \end{figure} \begin{figure} \centering \includegraphics[scale=0.32,trim=2cm 3cm 4cm 0]{ss-vel-vector-photo-event1-ct1-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 3cm 4cm 0]{rho-vel-vector-photo-event1-ct1-94-grn-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 3cm 4cm 0]{vh-vel-vector-photo-event1-ct1-eps-converted-to.pdf}\\ \includegraphics[scale=0.32,trim=2cm 3cm 4cm 1.8cm]{ss-vel-vector-photo-event2-ct1-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 3cm 4cm 1.8cm ]{rho-vel-vector-photo-event2-ct1-94-grn-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 3cm 4cm 1.8cm]{vh-vel-vector-photo-event2-ct1-eps-converted-to.pdf}\\ \includegraphics[scale=0.32,trim=2cm 1cm 4cm 1.8cm]{ss-vel-vector-photo-event3-ct1-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 1cm 4cm 1.8cm]{rho-vel-vector-photo-event3-ct1-94-grn-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 1cm 4cm 1.8cm]{vh-vel-vector-photo-event3-ct1-eps-converted-to.pdf}\\ \includegraphics[scale=0.32,trim=2cm 3cm 4cm -1.cm]{ss-vel-vector-chrom-event1-ct1-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 3cm 4cm -1.cm]{rho-vel-vector-chrom-event1-ct1-94-grn-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 3cm 4cm -1.cm]{vh-vel-vector-chrom-event1-ct1-eps-converted-to.pdf}\\ \includegraphics[scale=0.32,trim=2cm 3cm 4cm 1.8cm]{ss-vel-vector-chrom-event2-ct1-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 3cm 4cm 1.8cm]{rho-vel-vector-chrom-event2-ct1-94-grn-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 3cm 4cm 1.8cm]{vh-vel-vector-chrom-event2-ct1-eps-converted-to.pdf}\\ \includegraphics[scale=0.32,trim=2cm 1cm 4cm 1.8cm]{ss-vel-vector-chrom-event3-ct1-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 1cm 4cm 1.8cm]{rho-vel-vector-chrom-event3-ct1-94-grn-eps-converted-to.pdf} \includegraphics[scale=0.32,trim=0 1cm 4cm 1.8cm]{vh-vel-vector-chrom-event3-ct1-eps-converted-to.pdf} \caption{Left to right: Profiles of swirling strength, mass density and horizontal velocity for three vortices (different rows with border colors corresponding to those of the squares outlining the vortices in Fig. \ref{Fig10} on the solar surface (rows 1 to 3) and at 1.5 Mm above the surface (rows 4 to 6). The over-plotted arrows represent the horizontal velocity. The longest arrow corresponds to 4 km/s and 16 km/s in the photosphere and the chromosphere, respectively.} \label{Fig11} \end{figure} \begin{figure*} \centering \includegraphics[scale=0.45, trim =0 1.cm 0 0]{areafraction4-eps-converted-to.pdf} \includegraphics[scale=0.45, trim =0 1.cm 0 0]{rho-fraction-vortices4-eps-converted-to.pdf}\\ \includegraphics[scale=0.45, trim =0 1.cm 0 0]{b-comparison-vortices4-eps-converted-to.pdf} \includegraphics[scale=0.45, trim =0 1.cm 0 0]{temp-over-vortices4-eps-converted-to.pdf}\\ \includegraphics[scale=0.45, trim =0 0.2cm 0 0]{current-fraction-vortices4-eps-converted-to.pdf} \includegraphics[scale=0.45, trim =0 0.2cm 0 0]{heating-fraction-eps-converted-to.pdf}\\ \caption{Dependence of vortex properties with height. (a): Area fraction (fill factor) covered by vortices, (b): Horizontally averaged magnetic field strength in vortices (solid) and over the whole domain (dashed), (c): Horizontally averaged unsigned vertical component of current in vortices (normalized to the density averaged over the whole domain), (d): Horizontally averaged mass density in vortices (similarly normalized), (e): Horizontally averaged temperature in vortices (solid curve) and over the entire domain (dashed curve) and (f): Horizontally averaged total heating (viscous+resistive) in vortices (normalized to the total heating averaged over the whole domain). The region near the top boundary is shaded Gray as this part may be affected by the imposed top boundary conditions and should be ignored.} \label{Fig7} \end{figure*} \begin{figure} \centering \includegraphics[scale=0.55,trim=1.cm 0 1.cm 0]{drawing-17-07-eps-converted-to.pdf} \caption{ A simplified cartoon showing how turbulent motions drive vortices at different spatial scales. The lines with arrow heads represent the surface flows, being convective motions in the granules and small-scale turbulent motions in the intergranular lanes. These turbulent motions cover a range of scales and couple to the magnetic field (dashed curves) producing vortices at various scales from the whole flux tube (upper panel) down to the resolution limit of the simulation (lower panel).} \label{Fig5} \end{figure} \subsection{Statistical properties of vortices} In a recent study, \citet{yadav2020} have demonstrated that vortices exist over a range of spatial scales and smaller vortices are the dominant carriers of energy. Therefore, in the present paper we concentrate on investigating the properties of small-scale vortices. To study their general characteristics, a statistical investigation is performed by taking the temporal average for 42 snapshots covering a 7-minute long time sequence at 10 s cadence. The vortices that we are investigating are small-scale (spatial size of $\sim$50-100 km near the solar surface and $\sim$100-200 km in the chromosphere) and short-lived (lifetime $\sim$20-120 s). They are abundant throughout the simulation domain, therefore the time sequence of 7 minutes is sufficient for their statistical analysis. We display the dependence of vortex properties with height in Fig. \ref{Fig7}. Note that the height scale starts at 0.5 Mm above the average solar surface, i.e. roughly at the base of the chromosphere. Panel (a) displays the filling factor of vortices i.e. fractional area covered by vortices as a function of height. In the lower chromospheric layers, area coverage increases with height because of the magnetic field expansion, however, in the upper layers magnetic field is nearly homogeneous (shown in panel (b)) and hence the filling factor of vortices does not increase much. Panel (b) displays the height variation of the average magnetic field strength over vortices (solid) and horizontal mean (dashed). Up to a height of $\sim$1 Mm, vortices have higher averaged magnetic field strengths as they originate inside the magnetic concentrations. However, in the higher layers, the field strength is nearly homogeneous and the average magnetic field over vortices is nearly same as the overall horizontal mean. Panel (c) displays the vertical component of current (unsigned) averaged over vortices relative to the horizontal mean at the same geometrical height confirming the association of small-scale vortices with strong currents in the chromosphere, which we reported in Sect. \ref{individual}. The mass density averaged over the vortices at a given height relative to the density averaged over the entire horizontal domain at the same geometrical height is displayed in panel (d). Up to a height of $\sim$1 Mm, vortices have a lower mass density than the average mass density at the same geometrical height. This is due to their occurrence in strong magnetic regions that are evacuated. Above $\sim$1 Mm vortices have $\sim$50$\%$ higher density than the average at that height as vortices trap the plasma in the chromosphere. Panel (e) shows the comparison of the average temperature over vortices (solid curve) with the temperature averaged over the whole horizontal plane (dashed curve) at each geometrical height. The average temperature over vortices is always higher than the horizontally averaged temperature at the same geometrical height indicating their importance in the heating of the chromosphere. Note that in lower layers (around $z=0$) the temperature in vortices is lower than the average, as they are located in magnetic elements that are cooler at equal geometrical height. Panel (f) displays the total heating (viscous + resistive) over vortices relative to the horizontally averaged heating. Excessive heating over the vortices indicates their energetic importance, especially in the upper chromosphere. The results obtained in the uppermost layers (Gray shaded region in all panels) should be ignored as they are affected by the closed top boundary condition and therefore, can be of non-physical origin. In view of the results obtained we suggest the following formation mechanism for the small-scale vortices and display it in a simplified cartoon in Fig. \ref{Fig5}. Vortices are formed at various scales starting from the sizes of the intergranular lanes, possibly due to the angular momentum conservation of downflowing plasma. Because of the turbulent nature of the plasma, vorticity cascades from the injection scale to the smaller scales. The smallest scales that can be resolved in our simulations correspond to a multiple of the grid resolution. In the upper convection zone and near-surface layers, where the ratio of thermal gas pressure to magnetic pressure, i.e., plasma- $\beta$, is high, gas dynamics dominates over magnetic fields and magnetic field lines are frozen into the plasma. Vortical flows in this layer create a twist in the magnetic field. This twist in the magnetic flux tubes then propagates to higher atmospheric layers in the form of a torsional Alfv\'en wave. In the upper layers, plasma- $\beta$ is low and plasma dynamics is dominated by magnetic fields. Here, the twisted magnetic flux tubes make the surrounding plasma co-rotate forming a chromospheric counterpart of the photospheric vortices. Because of the inherent turbulence, larger vortex flows cascade down to the smallest sizes determined by the grid resolution and numerical diffusivity. Using vortex detection methods based on velocity gradients, we select the smallest vortices present in the system. Vortices originate inside the magnetic elements and are aligned along the magnetic field lines (as shown in Fig. \ref{Fig2} and Fig. \ref{Fig8}). These vortices do not appear in funnel-like shapes as observed in quiet Sun vortex observations (\citealt{Wedemeyer-Bohm2012,ParkS2016,kostas2018}) because vortices follow the magnetic field lines and the magnetic flux tubes in active regions do not expand as much as in the quiet Sun regions at least above the temperature minimum layer, where the magnetic elements merge (\citealt{pneuman_1986}). \section{Conclusions}\label{4} Vorticies are ubiquitous in the solar atmosphere and are suggested to play an important role in the heating of the solar atmosphere (\citealt{Wedemeyer-Bohm2012,kostas2018,jiajia2018}). However, even with the highest resolution observations available, the internal structure of the vortex tubes is not yet resolved. Moreover, small-scale vortices are more abundant and are shown to contribute more energy flux in the solar atmosphere than comparatively larger vortex flows (\citealt{yadav2020}). Thus it is of considerable relevance to investigate their physical properties using high-resolution numerical simulations. With this aim, a three-dimensional radiation-MHD simulation of a plage region was carried out. Using swirling strength criterion vortices were detected and isolated. We report the occurrence of very small-scale vortices in a unipolar plage region. With diameters of $\sim$50-100 km in the photosphere and $\sim$100-200 km in the chromosphere, these are the smallest vortices that can be resolved in our simulations. Due to a lack of spatial resolution, such small-scale vortices have not been detected in observations yet. Vortices have a lower mass density than the mean density both in the photosphere and in the lower chromosphere, this is in agreement with the previous simulations that were limited to the photosphere and lower solar chromosphere (\citealt{Moll2011,Kitiashvili2013}). However, in the higher atmospheric layers, they are denser structures than the surroundings. This finding conforms with the fact that vortices are mostly observed as absorbing (dark) intensity structures when observed in chromospheric radiation. (\citealt{Wedemeyer-Bohm2012,ParkS2016,kostas2018}). In chromospheric layers, the temperature averaged over vortices is higher than the temperature averaged over the whole horizontal domain in the chromosphere (Fig. \ref{Fig7}). Previously, \cite{moll2012} had also found an association of stronger vortices with increased temperatures (simulations limited up to 800 km above the mean solar surface). Using simultaneous observations from SST/CRISP and IRIS, \cite{ParkS2016} have also shown small-scale quiet Sun vortices to have higher temperatures than the surroundings in the upper chromosphere. A particularly important result of our present analysis is that we found thin and strong current sheets develop at their interface. There seems to be a correspondence between large values of current densities and the vortex locations. Hence we conjecture that heating at the vortex sites could be due to current dissipation in addition to viscous dissipation. Nonetheless, the ratio of viscous and resistive heating will depend on the magnetic Prandtl number i.e. the ratio of kinematic viscosity to magnetic diffusivity(\citealt{axel_2014,Rempel2017,axel_2019}). There are speculations that vortices can be the driving mechanism for dynamic jet-like features observed in the chromosphere (\citealt{Kuridze2016,Iijima2017}). Though the present findings, viz. the correspondence of vortices with magnetic field lines, higher density and higher temperatures support their possible relationship with chromospheric jets, a more comprehensive study is needed to establish such correspondence. \begin{acknowledgements} The authors thank the anonymous referee for insightful comments on the manuscript. The authors acknowledge L. P. Chitta and D. Przybylski for useful discussions on various aspects of this paper. This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation programme (grant agreement No. 695075) and has been supported by the BK21 plus program through the National Research Foundation (NRF) funded by the Ministry of Education of Korea. N. Y. thanks ISSI Bern for support for the team "The Nature and Physics of Vortex Flows in Solar Plasmas". \end{acknowledgements} \bibliographystyle{aa}
{ "redpajama_set_name": "RedPajamaArXiv" }
6,487
{"url":"https:\/\/gmatclub.com\/forum\/if-f-x-x-1-2-3-which-of-the-following-is-true-for-all-x-272401.html","text":"GMAT Question of the Day - Daily to your Mailbox; hard ones only\n\n It is currently 22 May 2019, 04:46\n\n### GMAT Club Daily Prep\n\n#### Thank you for using the timer - this advanced tool can estimate your performance and suggest more practice questions. We have subscribed you to Daily Prep Questions via email.\n\nCustomized\nfor You\n\nwe will pick new questions that match your level based on your Timer History\n\nTrack\n\nevery week, we\u2019ll send you an estimated GMAT score based on your performance\n\nPractice\nPays\n\nwe will pick new questions that match your level based on your Timer History\n\n# If f(x) = (x - 1)^2 + 3, which of the following is true for all x?\n\nAuthor Message\nTAGS:\n\n### Hide Tags\n\nMath Expert\nJoined: 02 Sep 2009\nPosts: 55228\nIf f(x) = (x - 1)^2 + 3, which of the following is true for all x?\u00a0 [#permalink]\n\n### Show Tags\n\n03 Aug 2018, 04:57\n00:00\n\nDifficulty:\n\n65% (hard)\n\nQuestion Stats:\n\n66% (02:23) correct 34% (02:32) wrong based on 160 sessions\n\n### HideShow timer Statistics\n\nIf $$f(x) = (x - 1)^2 + 3$$, which of the following is true for all x?\n\nI. $$2f(x) = f(x - 1) + f(x + 1)$$\n\nII. $$f(2 - x) = f(x)$$\n\nIII. $$f(x) = f(-x)$$\n\nA. I only\n\nB. II only\n\nC. III only\n\nD. None\n\nE. I, II and III\n\n_________________\nIntern\nJoined: 03 Apr 2018\nPosts: 7\nIf f(x) = (x - 1)^2 + 3, which of the following is true for all x?\u00a0 [#permalink]\n\n### Show Tags\n\nUpdated on: 03 Aug 2018, 05:45\nB.\n\nI started with III, since it seemed the easiest to find and that wasn't true.\n$$f(-x) = (-x-1)^2 + 3 = x^2+1+2x+3 = x^2 + 2x + 4$$ NOT equal to $$f(x)$$\n\nSo E is out.\n\nThen took II.\n$$f(2-x) = (2-x-1)^2 + 3 = (-x+1)^2 + 3$$ will be equal to $$f(x)$$. Thus, B.\n\nIs there any shorter approach?\n\nOriginally posted by greenisthecolor on 03 Aug 2018, 05:24.\nLast edited by greenisthecolor on 03 Aug 2018, 05:45, edited 1 time in total.\nVP\nStatus: Learning stage\nJoined: 01 Oct 2017\nPosts: 1008\nWE: Supply Chain Management (Energy and Utilities)\nRe: If f(x) = (x - 1)^2 + 3, which of the following is true for all x?\u00a0 [#permalink]\n\n### Show Tags\n\n03 Aug 2018, 05:28\nBunuel wrote:\nIf $$f(x) = (x - 1)^2 + 3$$, which of the following is true for all x?\n\nI. $$2f(x) = f(x - 1) + f(x + 1)$$\n\nII. $$f(2 - x) = f(x)$$\n\nIII. $$f(x) = f(-x)$$\n\nA. I only\n\nB. II only\n\nC. III only\n\nD. None\n\nE. I, II and III\n\nGiven, $$f(x) = (x - 1)^2 + 3$$\n\nI. $$f(x - 1) + f(x + 1)=(x-2)^2+3+x^2+3$$=$$2(x^2-2x+5)=2((x-1)^2+3))+2$$=$$2f(x)+2\\neq{2f(x)}$$\nII. $$f(2 - x)=(2-x-1)^2+3=(1-x)^2+3=(x-1)^2+3$$=f(x)\nIII. $$f(-x)=(-x-1)^2+3=x^2+2x+1+3=x^2+2x+4\\neq{f(x)}$$\n\nAns. (B)\n_________________\nRegards,\n\nPKN\n\nRise above the storm, you will find the sunshine\nManager\nJoined: 29 Jul 2018\nPosts: 114\nConcentration: Finance, Statistics\nGMAT 1: 620 Q45 V31\nRe: If f(x) = (x - 1)^2 + 3, which of the following is true for all x?\u00a0 [#permalink]\n\n### Show Tags\n\n20 Oct 2018, 20:36\ngmatbusters can we do this question by putting values of x?\nRetired Moderator\nJoined: 27 Oct 2017\nPosts: 1225\nLocation: India\nGPA: 3.64\nWE: Business Development (Energy and Utilities)\nRe: If f(x) = (x - 1)^2 + 3, which of the following is true for all x?\u00a0 [#permalink]\n\n### Show Tags\n\n20 Oct 2018, 20:49\n1\nHi\n\nSubstituting values of x to prove that an equation is valid is tricky because sometimes an equation valid for one value of x doesn't satisfy other value of x.\nAlways use substitute value of x to prove that equation is not valid, it works all the time.\nFor example: in the given question,\n\nIII. $$f(x) = f(-x)$$, it is valid for x = 0, but not for x = 2.\n\nhence substituting values is not a good approach for this question.\n\nmanjot123 wrote:\ngmatbusters can we do this question by putting values of x?\n\n_________________\nRe: If f(x) = (x - 1)^2 + 3, which of the following is true for all x? \u00a0 [#permalink] 20 Oct 2018, 20:49\nDisplay posts from previous: Sort by","date":"2019-05-22 11:46:37","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 1, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.7252345085144043, \"perplexity\": 2677.63235213705}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2019-22\/segments\/1558232256797.20\/warc\/CC-MAIN-20190522103253-20190522125253-00281.warc.gz\"}"}
null
null
Q: Get snapshot from AVCaptureSession contaning Visionkit face detection elements I use AVCaptureSession to setup a camera view and using vision kit to detect and add a rectangular on the face. Here is how I can do it override func viewDidLoad() { super.viewDidLoad() self.prepareVisionRequest() } fileprivate func prepareVisionRequest() { //self.trackingRequests = [] var requests = [VNTrackObjectRequest]() let faceDetectionRequest = VNDetectFaceRectanglesRequest(completionHandler: { (request, error) in if error != nil { print("FaceDetection error: \(String(describing: error)).") } guard let faceDetectionRequest = request as? VNDetectFaceRectanglesRequest, let results = faceDetectionRequest.results else { return } DispatchQueue.main.async { // Add the observations to the tracking list for observation in results { let faceTrackingRequest = VNTrackObjectRequest(detectedObjectObservation: observation) faceTrackingRequest.trackingLevel = .fast requests.append(faceTrackingRequest) } self.trackingRequests = requests } }) // Start with detection. Find face, then track it. self.detectionRequests = [faceDetectionRequest] self.sequenceRequestHandler = VNSequenceRequestHandler() self.setupVisionDrawingLayers() } // MARK: Drawing Vision Observations fileprivate func setupVisionDrawingLayers() { let captureDeviceResolution = self.captureDeviceResolution let captureDeviceBounds = CGRect(x: 0, y: 0, width: captureDeviceResolution.width, height: captureDeviceResolution.height) let captureDeviceBoundsCenterPoint = CGPoint(x: captureDeviceBounds.midX, y: captureDeviceBounds.midY) let normalizedCenterPoint = CGPoint(x: 0.5, y: 0.5) guard let rootLayer = self.rootLayer else { self.presentErrorAlert(message: "view was not property initialized") return } let overlayLayer = CALayer() overlayLayer.name = "DetectionOverlay" overlayLayer.masksToBounds = true overlayLayer.anchorPoint = normalizedCenterPoint overlayLayer.bounds = captureDeviceBounds overlayLayer.position = CGPoint(x: rootLayer.bounds.midX, y: rootLayer.bounds.midY) let faceRectangleShapeLayer = CAShapeLayer() faceRectangleShapeLayer.name = "RectangleOutlineLayer" faceRectangleShapeLayer.bounds = captureDeviceBounds faceRectangleShapeLayer.anchorPoint = normalizedCenterPoint faceRectangleShapeLayer.position = captureDeviceBoundsCenterPoint faceRectangleShapeLayer.fillColor = UIColor.white.withAlphaComponent(0.9).cgColor // faceLandmarksShapeLayer.strokeColor = UIColor.white.withAlphaComponent(0.7).cgColor faceRectangleShapeLayer.lineWidth = 5 faceRectangleShapeLayer.shadowOpacity = 0.7 faceRectangleShapeLayer.shadowRadius = 5 let faceLandmarksShapeLayer = CAShapeLayer() faceLandmarksShapeLayer.name = "FaceLandmarksLayer" faceLandmarksShapeLayer.bounds = captureDeviceBounds faceLandmarksShapeLayer.anchorPoint = normalizedCenterPoint faceLandmarksShapeLayer.position = captureDeviceBoundsCenterPoint faceLandmarksShapeLayer.fillColor = nil faceLandmarksShapeLayer.strokeColor = nil // overlayLayer.addSublayer(faceRectangleShapeLayer) faceRectangleShapeLayer.addSublayer(faceLandmarksShapeLayer) rootLayer.addSublayer(overlayLayer) self.detectionOverlayLayer = overlayLayer self.detectedFaceRectangleShapeLayer = faceRectangleShapeLayer self.detectedFaceLandmarksShapeLayer = faceLandmarksShapeLayer self.updateLayerGeometry() } Now, I'm trying three ways to take snapshots 1- using UIGraphicsImageRenderer, it shows only the rectangular on the face and the camera view in not visible - it's black 2- Take image from captureOutput(_ output: AVCaptureOutput, didOutput sampleBuffer: CMSampleBuffer, from connection: AVCaptureConnection) - the image from the buffer shows only the camera view, no rectangular 3- use AVCapturePhotoCaptureDelegate to capture photo from AVCaptureSession, the captured photo shows only the camera view, no rectangular Could you please help me to take snapshot that contains both camera view and the rectangular! Thanks
{ "redpajama_set_name": "RedPajamaStackExchange" }
1,199
// This file is automatically generated. package adila.db; /* * HTC Hero * * DEVICE: hero * MODEL: HTC Hero */ final class hero_htc20hero { public static final String DATA = "HTC|Hero|"; }
{ "redpajama_set_name": "RedPajamaGithub" }
9,195
What is the history of NWTR? Nancy Meitle started the rally in Corvallis in 1986. Until 2003 there was no formal organization at all. At the Eugene rally in 2003 past organizers formed an informal group to begin sharing resources, setting some guidelines and working toward always having the next two year's rallies in place. This group met annual at the rally. The Salem rally in 2012 decided to donate their proceeds to formalizing the ad hoc organizing committee into a more formal non-profit structure with a board and by laws. That group has had its first weekend retreat to set our mission and set up committees in November of 2014. Since our founder was a teacher she liked doing the rally over Memorial Day weekend. Fourth of July tends to be the more popular date given the weather in the northwest. But over the years it has come to fluctuate between Memorial Day and the Fourth depending on the host venue's weather and availability. To date no one has ventured to plan a rally on Labor Day weekend as it might make it more challenging for those with kids and who teach. How big an event is this? This varies from about 200 tandems up to about 500 tandems with the larger events typically on the I-5 corridor. If you've never been to a tandem rally, prepare to be amazed. You will never see such a wide variety of 2-wheeled conveyances! It's not unusual to see triples, quads and even quints. And there'll be tag-alongs and trailers with babies and pets. Single bikes are also welcome to join us. What kind of lodging can we expect? Every community's event is unique. We encourage organizers to offer a variety of options. Generally in addition to hotel stays there's tent and RV camping. Often, but not always, there may be dormitory stays also available. What kinds of routes can we expect? Every year's rally is unique to the community hosting. We encourage organizers to offer multiple length and difficulty routes each day with an attempt to meet a variety of attendees' needs. Ideally that includes families with youngsters on a tag-along as well as those training for a competitive event and everything in between. Short loops or out-and-back routes of roughly 20 miles are typical of the short routes. Then there's something in the 40-50 mile range and often 60-80 miles as well. Again, every year is unique so be sure to check the website for the event you are attending. What's it like to ride in a Mass Start? Depending on the route and number of attendees most rallies have at least one mass start. It's exciting to see hundreds of tandems, triples, quads and quints, some with tag-a-longs or Bobs. It's quite crowded and slow going at first. In most communities the police lead the mass start out and stop traffic at intersections until the mass naturally spreads out. You can get a feel for what it's like in this fast-paced 2:45 minute video courtesy of Mount Baker Bicycle Club's Greg Rehm. Questions about the upcoming NWTR rally? Because each year's event is hosted in a different community and the answers to questions are unique to that host, if you have questions about the next rally, please visit that year's website. It's easy…just go to https://nwtr.org/2017 for the one in 2017. But we won't be updating that link each year so just change the year for the year it is and voila…you have the website for the next event! You'll get much more accurate responses by doing this. Remember we are ALL volunteers. So thanks for your patience.
{ "redpajama_set_name": "RedPajamaC4" }
9,021
from pyoram.__about__ import __version__ def _configure_logging(): import os import logging log = logging.getLogger("pyoram") formatter = logging.Formatter( fmt=("[%(asctime)s.%(msecs)03d," "%(name)s,%(levelname)s] %(threadName)s %(message)s"), datefmt="%Y-%m-%d %H:%M:%S") level = os.environ.get("PYORAM_LOGLEVEL", "WARNING") logfilename = os.environ.get("PYORAM_LOGFILE", None) if len(logging.root.handlers) == 0: # configure the logging with some sensible # defaults. try: import tempfile tempfile = tempfile.TemporaryFile(dir=".") tempfile.close() except OSError: # cannot write in current directory, use the # console logger handler = logging.StreamHandler() else: if logfilename is None: handler = logging.StreamHandler() else: # set up a basic logfile in current directory handler = logging.FileHandler(logfilename) handler.setFormatter(formatter) handler.setLevel(level) log.addHandler(handler) log.setLevel(level) log.info("PyORAM log configured using built-in " "defaults, level=%s", level) _configure_logging() del _configure_logging def _configure_pyoram(): class _Configure(object): __slots__ = ("SHOW_PROGRESS_BAR",) def __init__(self): self.SHOW_PROGRESS_BAR = False return _Configure() config = _configure_pyoram() del _configure_pyoram import pyoram.util import pyoram.crypto import pyoram.storage import pyoram.encrypted_storage import pyoram.oblivious_storage
{ "redpajama_set_name": "RedPajamaGithub" }
8,365
\section{Betti numbers and the Cayley graph} \label{sec:bettinumbers} The following simple result is essential for our approach. \begin{lemma}\label{lem:projection-estimate} Let $\mathcal{H}$ be a Hilbert space and let $W \subseteq \mathcal{H}$ be a subspace. Let $P\colon \mathcal{H} \to \overline{W}$ denote the orthogonal projection onto the closure of $W$. Then for all $v \in \mathcal{H}$ \[ \langle Pv, v \rangle = \sup_{w \in W} \frac{|\langle w,v \rangle|^2}{\langle w, w \rangle} \] where the supremum is taken over all non-zero elements of $W$ (and is defined to be $0$ if $W = 0$). \end{lemma} \begin{proof} For $v = 0$ the assertion is obvious. We may assume that $\Vert v \Vert = 1$. For all $w \in W$, we note that \begin{align*} |\langle w , v \rangle|^2 = |\langle Pw , v \rangle|^2 = |\langle w , Pv \rangle|^2 \stackrel{C.S.}{\leq} \Vert w \Vert^2 \Vert Pv \Vert^2 = \langle Pv, v \rangle \Vert w \Vert^2. \end{align*} If $w \neq 0$ we obtain \[ \frac{|\langle w , v \rangle|^2}{\langle w, w \rangle} \leq \langle Pv, v \rangle. \] In particular, the proof is complete if $Pv =0$. For the converse we assume $Pv \neq 0$. Let $\epsilon \in (0,1)$. Since $W$ is dense in $\overline{W}$, there is $w \in W$ with $\Vert Pv-w \Vert < \epsilon \Vert Pv \Vert^2$ and we deduce \[ |\langle w, v \rangle| \geq \langle Pv, v \rangle - |\langle w- Pv, v \rangle| \stackrel{C.S.}{\geq}\langle Pv, v \rangle - \Vert w- Pv \Vert \geq (1-\epsilon) \langle Pv, v \rangle. \] In addition, we note that $\Vert w \Vert = \Vert w -Pv + Pv \Vert \leq (1+\epsilon) \Vert Pv \Vert$ and so \[ \frac{|\langle w , v \rangle|^2}{\langle w, w \rangle} \geq \frac{(1-\epsilon)^2}{(1+\epsilon)^2} \langle Pv, v \rangle \] The assertion follows as $\epsilon$ can be arbitrarily close to $0$. \end{proof} Let $G$ be a group. A \emph{character} of $G$ is a function $\psi\colon G \to \bbC$ of positive type, which is constant on conjugacy classes of $G$ and satisfies $\psi(1_G) = 1$; see \cite[Def.~2.5]{Kio2018}. Let $\Ch(G)$ denote the space of all characters of $G$. Every character $\psi \in \Ch(G)$ gives rise to a semi-definite $G$-invariant inner product $\langle g, h \rangle_\psi = \psi(h^{-1}g)$ on the group ring $\bbC[G]$. Passing to the completion provides us with a tracial Hilbert $G$-bimodule $\ell^\psi(G)$; see \cite[Def.~2.1]{Kio2018}. Using the GNS construction, this provides a tracial von Neumann algebra and a notion of dimension, which can be used to define the $\psi$-Betti numbers $b_k^\psi(G)$ of $G$, provided that $G$ satisfies suitable finiteness properties. Specifically $b_0^\psi$ is defined for all groups and $b_1^\psi$ is defined for all finitely generated groups. For the regular character $\delta$ with $\delta(g) = 0$ for all $g \neq 1_G$, one has $\ell^\delta(G) = \ell^2(G)$ and one obtains the famous $\ell^2$-Betti numbers $b_k^{(2)}(G)$. The constant character (i.e., $\psi(g) = 1$ for all $g$) gives rise to the ordinary rational Betti numbers of $G$ since $\ell^\psi(G) \cong \bbC$. \begin{definition} Let $G$ be a group and let $\psi \in \Ch(G)$. A subgroup $K \leq G$ is \emph{$\psi$-regular}, if $\psi|_K$ is the regular character on $K$, i.e.~$\psi(k) = 0$ for all $k \in K \backslash \{1\}$. \end{definition} Here we are mainly interested in the first Betti numbers $b_1^\psi(G)$. It will however be useful and instructive to initially consider the $0$-th Betti number. Let $J_G$ denote the augmentation ideal in $\bbC[G]$, i.e.~the set of elements $w = \sum_{g\in G} w_g g$ which satisfy $\sum_{g \in G} w_g = 0$. \begin{lemma}\label{lem:b0} Let $G$ be a group and let $\psi \in \Ch(G)$ be a character. \begin{enumerate}[ (a)] \item\label{it:b0-formula} $b_0^\psi(G) = 1- \sup_{w \in J_G} \frac{\langle w, 1 \rangle_\psi^2}{\langle w, w \rangle_\psi}$ where the supremum is taken over all non-zero elements of $J_G$. \item\label{it:b0-lim} If $G = \bigcup_{i\in I} G_i$ is a directed union of subgroups $G_i$, then $\lim_{i \in I} b^\psi_0(G_i) = b_0^\psi(G)$. \item\label{it:b0-regular} $b_0^\psi(G) \leq \frac{1}{|K|}$ for every $\psi$-regular subgroup $K \leq G$. \end{enumerate} \end{lemma} \begin{remark} It is well-known that $b_0^{(2)}(G) = \frac{1}{|G|}$; see \cite[Thm.~1.35 (8)]{Luck}. \end{remark} \begin{proof} Let $S$ be a generating set for $G$. We consider the initial segment of the associated free resolution of $\bbC$: \[ \bbC[G]^{S} \stackrel{\partial_1}{\longrightarrow} \bbC[G] \longrightarrow \bbC. \] The image of $\partial_1$ is the augmentation ideal. We take the tensor product with $\ell^\psi(G)$ and deduce that \[ b_0^\psi(G) = 1 - \dim_\psi(\overline{J_G}). \] where $\overline{J_G}$ denotes the closure of the image of the augmentation ideal in $\ell^\psi(G)$. Let $P \colon \ell^\psi(G) \to \overline{J_G}$ denote the orthogonal projection. By definition \[ \dim_\psi(\overline{J_G}) = \langle P(1), 1 \rangle_\psi \] and assertion \eqref{it:b0-formula} follows from Lemma \ref{lem:projection-estimate}. Let $G = \bigcup G_i$ be a direct union of subgroups, then $J_G = \bigcup J_{G_i}$ and \eqref{it:b0-lim} follows immediately from \eqref{it:b0-formula}. Let $K \leq G$ be a $\psi$-regular subgroup. Let $T \subseteq K \setminus \{1\}$ be a finite subset. Then \[ w = |T|\cdot 1_G - \sum_{k \in T} k \in J_G. \] Since $K$ is $\psi$-regular, the elements of $K$ are orthonormal and we deduce \[ \frac{|\langle w ,1 \rangle_\psi|^2}{\langle w, w \rangle_\psi} = \frac{|T|^2}{|T|^2 + |T|} = \frac{|T|}{|T|+1}. \] Now \eqref{it:b0-formula} implies $b_0^\psi(K) \leq 1 - \frac{|T|}{|T|+1} = \frac{1}{|T|+1}$. Statement \eqref{it:b0-regular} follows by taking $T = K \setminus \{1\}$ if $K$ is finite respectively letting $|T|$ tend to $\infty$ otherwise. \end{proof} We would like to apply the same ideas to the first $\psi$-Betti number~$b_1^\psi(G)$. However, up to now we only have a definition of $b_1^\psi(G)$ for all finitely generated groups $G$. We also require a definition for groups which are not finitely generated. This could be done using L\"uck's generalized dimension function (discussed in \cite[\S6.1, 6.2]{Luck}), but this is not convenient for our purposes and for simplicity we work with the following variation. \begin{definition} Let $G$ be a group and let $\psi \in \Ch(G)$. Then \[ \bar{b}_1^\psi(G) := \liminf_{H \leq G} b_1^\psi(H) \] where the limit is taken over the directed system of all finitely generated subgroups $H \leq G$. \end{definition} \begin{remark}\label{rem:bar} For a finitely generated group $\bar{b}_1^\psi(G) = b_1^\psi(G)$. In general however, $\bar{b}_1^\psi(G)$ can be strictly larger than the properly defined value of the first $\psi$-Betti number. It is easy to see this for the ordinary Betti numbers. For instance, it follows from the methods developed in \cite{Howie82} that $\langle (x_i)_{i \in \bbZ} \mid x_ix_{i+1}x_i^{-1} = x_{i+1} ^2\rangle$, is a perfect and locally indicable group, i.e., the ordinary rational Betti number of every finitely generated subgroup is $\geq 1$. For the classical $\ell^2$-Betti number the inequality $b_1^{(2)}(G) \leq \bar{b}_1^{(2)}(G)$ follows from the argument given in the proof of \cite[Theorem 7.2 (3)]{Luck}. \end{remark} For later reference we state the following observation. \begin{lemma}\label{lem:liminf} Let $G$ be a group and let $\psi \in \Ch(G)$. If $G = \bigcup_{i\in I} G_i$ is a directed union of subgroups $G_i$, then \[ \bar{b}_1^\psi(G) \leq \liminf_{i\in I} \bar{b}_1^\psi(G_i) \] \end{lemma} \begin{proof} Let $\epsilon > 0$. There is a finitely generated subgroup $H_0 \leq G$ such that $b_1^\psi(H) \geq \bar{b}_1^\psi(G) - \epsilon$ for all finitely generated subgroups $H$ that contain $H_0$. Since $H_0$ is finitely generated, there is $i \in I$ such that $H_0 \subseteq G_i$. Thus for all $j \geq i$ we have $\bar{b}_1^{\psi}(G_j) \geq \bar{b}_1^\psi(G)-\epsilon$. \end{proof} Assume that $G$ is finitely generated and that $S$ is a finite generating set. The Cayley graph $\Cay(G,S)$ is the directed graph with vertex set $G$ and edges \[ E_{G,S} = \{ (g,gs) \mid g \in G, s \in S\}. \] The edge $(1_G, s)$ will be denoted by $\bar{s}$. The Cayley graph is equipped with a left action of $G$. Let $\bbC[E_{G,S}]$ be the vector space with basis $E_{G,S}$ and let $\partial\colon \bbC[E_{G,S}] \to \bbC[G]$ denote the boundary map. A \emph{finite cycle} in $\Cay(G,S)$ is an element $z \in \bbC[E_{G,S}]$ with $\partial(z) = 0$. Let $Z_{G,S}$ denote the space of finite cycles. If $\psi\in \Ch(G)$ is a character, then the semi-definite inner product $\langle \cdot, \cdot\rangle_\psi$ extends to a $G$-invariant semi-definite inner product on $\bbC[E_{G,S}]$ such that the edges $\{\bar{s} \mid s \in S\}$ are orthonormal; this means \[ \langle (g,gs), (h,ht) \rangle_\psi = \begin{cases} 0 & \text{ if } s \neq t\\ \psi(h^{-1}g) & \text{ if } s = t\end{cases} \] The following extends \cite[Prop.~2]{Pichot06} and is our main tool. \begin{lemma}\label{lem:main-lemma} Let $G$ be a group and let $S$ be a finite generating set. Let $\psi \in \Ch(G)$ be a character. Then \[ b_1^\psi(G) = |S| - 1 + b_0^{\psi}(G) - \sum_{s \in S} \sup_{z \in Z_{G,S}} \frac{|\langle z, \bar{s} \rangle_\psi|^2}{\langle z, z \rangle_\psi} \] where the suprema are taken over all non-zero elements of $Z_{G,S}$. \end{lemma} \begin{proof} The finite generating set $S$ provides us with a presentation $G \cong F/R$ of $G$ where $F$ is the free group over $S$ and $R$ is the subgroup of relations. We consider the initial segment of the associated free resolution of $\bbC$: \[ \bbC[G]^{R}\stackrel{\partial_2}{\longrightarrow} \bbC[G]^{S} \stackrel{\partial_1}{\longrightarrow} \bbC[G] \longrightarrow \bbC. \] Tensoring with $\ell^\psi(G)$ gives \[ \ell^\psi(G)^{R}\stackrel{\partial_2}{\longrightarrow} \ell^\psi(G)^{S} \stackrel{\partial_1}{\longrightarrow} \ell^\psi(G). \] The middle term is naturally isomorphic to the completion of $\bbC[E_{G,S}]$ with respect to $\langle \cdot, \cdot \rangle_\psi$. The image of $\partial_2$ is the closure of $Z_{G,S}$. The $\psi$-dimension of the closure of the image of $\partial_1$ is $1 - b_0^\psi(G)$. We deduce that \[ b_1^\psi(G) = |S| - (1-b_0^\psi(G)) - \dim_\psi(\overline{Z_{G,S}}). \] Let $P \colon \ell^\psi(G)^S \to \overline{Z_{G,S}}$ denote the orthogonal projection. By definition \[ \dim_\psi(\overline{Z_{G,S}}) = \sum_{s \in S} \langle P\bar{s}, \bar{s} \rangle_\psi \] and the result follows from Lemma \ref{lem:projection-estimate} \end{proof} \begin{remark}\label{rem:b0} It seems surprising that the value on the right hand side is independent from the chosen set of generators. This is a consequence of the homotopy invariance of the $\psi$-Betti numbers, which can be proven using the standard argument; e.g.~\cite[Thm. 3.18]{Kammeyer} or~\cite{Luck}. \end{remark} \section{Torsion groups}\label{sec:torsion} In view of Remark \ref{rem:bar}, the following result implies Theorem~\ref{thm:torsion-bound-intro}. \begin{theorem}\label{thm:torsion-bound} Let $p$ be a prime. Let $G$ be a torsion group of exponent $p$. Then $\bar{b}_1^{(2)}(G) \leq 2p-2$. \end{theorem} \begin{proof} We may assume that $G$ is infinite (and $b_0^{(2)} = 0$), otherwise $b_1^{(2)}(G) = 0$ and there is nothing to show. By the definition of $\bar{b}_1^{(2)}(G)$ (see Remark \ref{rem:bar}), we may assume that $G$ is finitely generated. We choose a minimal generating set $S$ of $G$ and denote the number of elements by $N = |S|$. Since all elements of $G$ have prime order, all pairwise distinct elements $a,b,c \in S$ satisfy \begin{equation}\label{condition:cyclic} \langle ac \rangle \cap \langle ab \rangle = \{1\} \end{equation} Suppose for a contradiction that there are three distinct elements $a,b,c \in S$ with $\langle ac \rangle \cap \langle ab \rangle \neq \{1\}$, then these cyclic groups of prime order coincide and \[ ac = (ab)^k \] for some $k \in \bbN$, i.e., $c \in \langle a,b \rangle$ which contradicts the minimality of $S$. For all $a \in S$ we have $N-1$ relations \[ (ab)^p \] of length $2p$ for all $b \neq a$ in $S$. By condition \eqref{condition:cyclic}, the only common edge in the Cayley graph is the first edge $\bar{a}$ from $1$ to $a$. Summing up these cycles, we obtain a cycle $z_a$ in $\mathrm{Cay}(G,S)$ with \[ \frac{\langle z_a,\bar{a} \rangle^2}{\langle z_a, z_a \rangle} = \frac{(N-1)^2}{(N-1)^2+(N-1)(2p-1)} = \frac{1}{1+\frac{2p-1}{N-1}}. \] We deduce from Lemma \ref{lem:main-lemma} that \begin{eqnarray*} b_1^{(2)}(G) &\leq& N-1 - \sum_{a \in S} \frac{\langle z_a,\bar{a} \rangle^2}{\langle z_a, z_a \rangle} \ \ = \ \ N-1 - \frac{N}{1+\frac{2p-1}{N-1}} \\ &=& \frac{2p-2}{1+\frac{2p-1}{N-1}} \ \ \leqslant \ \ 2p-2. \end{eqnarray*} \end{proof} \begin{theorem}\label{thm:vanishing-torsion} Let $p$ be a prime number and let $G$ be a countable torsion group of exponent $p$. If $G$ has an infinite normal subgroup $N$ of infinite index, then \[ b_1^{(2)}(G) = 0. \] \end{theorem} \begin{proof} By Theorem \ref{thm:torsion-bound} and Remark \ref{rem:bar} we have $b_1^{(2)}(N) \leq \bar{b}_1^{(2)}(N) \leq 2p-2$. By Gaboriau's Theorem~\cite[Thm.~6.8]{Gaboriau2002}, this implies $b_1^{(2)}(G) = 0$. \end{proof} \begin{proof}[Proof of Corollary \ref{corB(3,100)}] Recall that $B(m,p)$ denotes the Burnside group of exponent $p$ and rank $m$. Since $B(1,p)$ is finite, we have $b_1^{(2)}(B(1,p)) = 0$. Assume $m \geq 2$. For sufficiently large $p$, the main result of \cite{Ivanov2003} implies that $B(m,p)$ contains a $Q$-subgroup $H$ which is isomorphic to $B(\infty,p)$. A $Q$-subgroup has the property that the normal closure $\langle K \rangle^{B(m,p)}$ in $B(m,p)$ of any normal subgroup $K \trianglelefteq H$ intersects $H$ exactly in $K$. Take a projection from $B(\infty,p)$ onto $B(\infty,p)$ with an infinite kernel $K$. Then the normal closure $\langle K \rangle^{B(m,p)}$ is an infinite normal subgroup of $B(m,p)$ of infinite index. Now Theorem \ref{thm:vanishing-torsion} implies the result. \end{proof} \begin{remark}\label{remB(3,100)} \begin{enumerate} \item Ivanov \cite{Ivanov2003} quantifies sufficiently large as $p > 10^{78}$. \item One can also deduce $b_1^{(2)}(B(m,p)) = 0$ for $m \geq 3$ under the assumption that $B(2,p)$ is infinite\footnote{According to Adian \cite{Adian2015} the Burnside groups $B(2,p)$ are infinite for all $p > 100$.} using the normal subgroup $N = \ker( B(m,p) \to B(m-1,p) )$. Indeed, let $x_1,x_2,\dots,x_m$ be a free generating set of $B(m,p)$ such that $N$ is the normal closure of $x_1$ in $B(m,p)$. Since $\langle x_1, x_2 \rangle \subseteq N\langle x_2 \rangle$ and $\langle x_1,x_2 \rangle \cong B(p,2)$ is infinite, we deduce that $N$ is infinite. Moreover, $N$ has infinite index, since $B(m,p)/N \cong B(m-1,p)$. \item We expect that $b_1^{(2)}(B(m,p)) = 0$ for all $p, m$. On the other hand, if $b_1^{(2)}(B(m,p)) > 0$ holds for some $m$ and $p$, then this offers a simple solution to the restricted Burnside problem for $m$-generated $p$-groups. More precisely, every finite index normal subgroup $N \trianglelefteq B(m,p)$ satisfies \[ |B(m,p):N| \cdot b_1^{(2)}(B(m,p)) = b_1^{(2)}(N) \leq 2p-2 \] by Theorem \ref{thm:torsion-bound} and this inequality imposes an upper bound on the index of $N$. \end{enumerate} \end{remark} \section{$q$-normality and applications}\label{sec:q-normal} \begin{lemma} \label{lemHa} Let $G=\langle H,a\rangle$ be a group and let $\psi\in\Ch(G)$ be a character. Assume that $aHa^{-1} \cap H$ contains a $\psi$-regular subgroup of order $n \in \mathbb{N} \cup \{\infty\}$. Then \[ \bar{b}_{1}^{\psi}(G) - b_0^\psi(G) \leqslant \bar{b}_{1}^{\psi}(H) + \frac{3+ 2\Re(\psi(a))}{n+2+2\Re(\psi(a))} \] In particular \[ \bar{b}_{1}^{\psi}(G) \leqslant \bar{b}_{1}^{\psi}(H).\] for $n = \infty$. \end{lemma} \begin{proof} Without loss of generality we assume that $a \not\in H$. We denote the $\psi$-regular subgroup of order $n$ in $H \cap aHa^{-1}$ by $K$. For a finite subset $S \subseteq H$, we define $H_S = \langle S \rangle$. Since $K$ is $\psi$-regular, we obtain $b_0^\psi(H) \leqslant b_0^\psi(G) \leqslant \frac{1}{n}$ and $b_0^\psi(H_S) \leq |K\cap H_S|^{-1}$ by Remark~\ref{rem:b0}. For $n = \infty$, let $S \subseteq H$ be any finite subset and denote by $h_1,h_2, \dots, h_k$ the pairwise distinct elements of $S \cap aSa^{-1} \cap K$. If $n< \infty$, we choose $S \subseteq H$ such that $$S \cap aSa^{-1} \ = \ K \setminus\{1\} \ = \ \{h_1,h_2, \dots, h_k\}.$$ In both situations we define $S' = S \cup \{a\}$ and $G' = \langle S' \rangle$. Lemma \ref{lem:main-lemma} implies \begin{eqnarray} b_{1}^{\psi}(G') - b_0^\psi(G') &=& |S'| - 1 - \sum_{s \in S'} \sup_{z \in Z_{G',S'}} \frac{|\langle z,\bar{s}\rangle_\psi|^2}{\langle z,z\rangle_\psi}\nonumber\\ &\leq& | S | + 1 - 1 - \sum_{s \in S} \sup_{z \in Z_{H_S,S}} \frac{|\langle z,\bar{s}\rangle_\psi|^2}{\langle z,z\rangle_\psi} - \sup_{z \in Z_{G',S'}} \frac{|\langle z,\bar{a}\rangle_\psi|^2}{\langle z,z\rangle_\psi}\nonumber\\ &\leq& b_{1}^{(2)}(H_S) + 1 - \sup_{z \in Z_{G,S'}} \frac{|\langle z,\bar{a}\rangle_\psi|^2}{\langle z,z\rangle_\psi}. \label{inequ1} \end{eqnarray} To obtain a lower bound for $\sup_{z \in Z_{G,S'}} \frac{|\langle z,\bar{a}\rangle_\psi|^2}{\langle z,z\rangle_\psi}$, we consider the Cayley graph $\Cay(G',S')$ of $G'$ and exhibit a suitable cycle $z$. Each relation $ah_ia^{-1}(ah_ia^{-1})^{-1}$ provides a cycle $z_i$ of length $4$ in $\Cay(G',S')$, i.e., \[ z_i = \underbrace{(1,a)}_{=\bar{a}} + (a,ah_i) - (ah_ia^{-1},ah_i) - (1,ah_ia^{-1}). \] Note that the cycles $z_i$ touch exactly four vertices, since $h_i \neq 1$ and $a \not\in H$. In additon, the cycles $z_i$ have no common edges, except for $\bar{a}$. We define $z = \sum_{i=1}^k z_i$ to be the sum of these cycles. Since $\psi(ah_ia^{-1}) = \psi(h_i) = 0$ holds for all $i\leq k$, we deduce \[ \langle z,\bar{a} \rangle_\psi = k \langle \bar{a}, \bar{a} \rangle_\psi - \sum_{i=1}^k \langle (ah_ia^{-1},ah_i), \bar{a} \rangle_\psi = k - \sum_{i=1}^k \psi(ah_ia^{-1}) = k \] We note further (using again that $K$ is $\psi$-regular) that for given $i,j \leq k$ the edges $\neq \bar{a}$ in $z_i, z_j$ are orthogonal unless $h_i = ah_ja^{-1}$ or $ah_ia^{-1} = h_j$. Each of these cases occurs at most once for every $i$ and then $\langle (a,ah_i), (1,ah_ja^{-1}) \rangle_\psi = \psi(a)$ and $\langle (1,ah_ia^{-1}), (a,ah_j) \rangle_\psi = \overline{\psi(a)}$ respectively. We deduce \begin{align} \langle z, z\rangle_\psi &= \sum_{i,j} \langle z_i,z_j \rangle_\psi \leq \underbrace{4k}_{i=j} + \underbrace{k^2 - k + k (\psi(a) + \overline{\psi(a)})}_{i\neq j} \nonumber \\ &\leq k^2 + (3 +2 \Re(\psi(a))) k \label{eq:bound} \end{align} and conclude \[ \frac{\langle z,\bar{a}\rangle_\psi^2}{\langle z,z\rangle_\psi} \geq \frac{1}{1 + \frac{3+2\Re(\psi(a))}{k}}. \] Finally, we use this cycle in combination with inequality \eqref{inequ1} to obtain \begin{eqnarray*} b_{1}^{\psi}(G') - b_0^\psi(G') &\leqslant& b_{1}^{(2)}(H_S) + 1 - \sup_{z \in Z_{G,S'}} \frac{|\langle z,\bar{a}\rangle_\psi|^2}{\langle z,z\rangle_\psi}\\ &\leqslant& b_{1}^{(2)}(H_S) + 1 - \frac{1}{1 + \frac{3 +2 \Re(\psi(a))}{k}}\\ &=& b_{1}^{(2)}(H_S) + \frac{3 +2 \Re(\psi(a))}{k+3 +2 \Re(\psi(a))} \ \ \ \underset{k \rightarrow \infty}{\longrightarrow} \ b_{1}^{(2)}(H_S) \end{eqnarray*} For $n=\infty$ we can make $k$ arbitrary large. In the case $n<\infty$ we have $k=n-1$ by construction. We note that every finitely generated subgroup of $G$ is contained in a group of the form $G'$. The result follows from Lemma \ref{lem:liminf} and Lemma \ref{lem:b0} \eqref{it:b0-lim}. If $n = \infty$, then $G$ contains an infinite $\psi$-regular subgroup and $b_0^\psi(G) = 0$ by Lemma \ref{lem:b0} \eqref{it:b0-regular}. \end{proof} In the spirit of Popa \cite{Popa} and Thom-Peterson \cite{ThomPeterson} we introduce the following notion. \begin{definition} Let $G$ be a group and let $\psi\in \Ch(G)$. A subgroup $H \leq G$ is \emph{q-$\psi$-normal}, if there is a set $A \subseteq G$ such that $G = \langle H \cup A \rangle$ and $H \cap aHa^{-1}$ contains an infinite $\psi$-regular subgroup for all $a \in A$. A subgroup $H \leq G$ is \emph{weakly} q-$\psi$-normal, if there is an ordinal number $\alpha$ and an increasing chain of subgroups $H_0 = H$ to $H_\alpha = G$ such that $\bigcup_{\beta < \gamma} H_\beta$ is q-$\psi$-normal in $H_\gamma$ for all $\gamma \leq \alpha$. \end{definition} Based on Lemma \ref{lemHa} we obtain the following analog of~\cite[Theorem~5.6]{ThomPeterson}. \begin{corollary}\label{cor:q-psi-normal} Let $G$ be a group and let $\psi \in \Ch(G)$. If $H \leq G$ is a weakly q-$\psi$-normal subgroup, then \[ \bar{b}_1^\psi(G) \leq \bar{b}_1^{\psi}(H). \] \end{corollary} \begin{proof} Assume that $H$ is q-$\psi$-normal. Then $G = \langle H \cup A \rangle$ and $H \cap aHa^{-1}$ contains an infinite $\psi$-regular subgroup for all $a \in A$. If $A$ is finite, then the assertion follows inductively from Lemma \ref{lemHa}. Assume that $A$ is infinite. For every finite subset $B \subseteq A$, we define $G_B = \langle H \cup B \rangle$. Then $G = \bigcup_{B \subseteq A} G_B$ and Lemma \ref{lem:liminf} implies \[ \bar{b}_1^{\psi}(G) \leq \liminf_{B \subseteq A} \bar{b}_1^\psi(G_B) \leq \bar{b}_1^\psi(H). \] The result for weakly q-$\psi$-normal subgroups follows by transfinite induction using Lemma~\ref{lem:liminf}. \end{proof} \begin{corollary} Let $G$ be a group and let $\psi \in \Ch(G)$. \begin{itemize} \item[(1)] If $G$ is an HNN-extension of $H$ with associated subgroups $A, B$ and $A$ contains an infinite $\psi$-regular subgroup, then we have $\bar{b}_{1}^{\psi}(G) \leq \bar{b}_{1}^{\psi}(H)$. \item[(2)] If $G= A \ast_{C} B$ is an amalgamated product such that $C$ is q-$\psi$-normal in $B$. Then we have $\overline{b}_{1}^{\psi}(G) \leq \bar{b}_{1}^{\psi}(A)$. \item[(3)] If $G$ contains an infinite normal amenable $\psi$-regular subgroup, then $b^\psi_1(G) = 0$. \end{itemize} \end{corollary} \begin{proof} (1): follows immediately from Lemma \ref{lemHa}. (2): The assumptions imply that $A$ is q-$\psi$-normal in $G = A*_C B$ and the assertion follows from Corollary~\ref{cor:q-psi-normal}. (3): The infinite normal amenable subgroup $N$ is q-$\psi$-normal in $G$ and $b_1^\psi(N) = b_1^{(2)}(N) = 0$ by~\cite[Thm.~0.2]{CheegerGromov}. \end{proof} We illustrate the helpfulness of q-normality with an application to right-angled groups. This notion was put forward in \cite[Definition 1]{AGN17}. \begin{definition} [right-angled groups] A group $G$ is \emph{right-angled}, if it is the quotient of a right-angled Artin group $A_{\Gamma}$ with a finite connected graph $\Gamma = (\mathcal{I},\mathcal{E})$ such that the image of every generator $\sigma_{i}$ $(i \in \mathcal{I})$ has infinite order in $G$. The image of the generating set of $A_{\Gamma}$ will be called a \emph{right-angled set of generators}. \end{definition} \begin{theorem}\label{thm:right-angled-vanish} Let $G$ be a right-angled group and let $\mathcal{S} = \{s_{i} \mid i \in \mathcal{I}\}$ be a right-angled set of generators. If $\psi \in \Ch(G)$ is such that the cyclic subgroup $\langle s_i \rangle$ is $\psi$-regular for every $i\in \mathcal{I}$. Then we have $b_{1}^{\psi}(G)=0$. \end{theorem} \begin{proof} Our proof will be by induction over the number $n = |\mathcal{I}| \in \mathbb{N}$ of generators. For the base of induction we note that $b_1^{(2)}(\mathbb{Z}) = 0$. We assume for the induction step w.\,l.\,o.\,g. $\mathcal{S}=\{s_{1},s_{2},\dots,s_{n},s_{n+1}\}$ such that $s_{n}$ commutes with $s_{n+1}$ and such that $G' = \langle s_1 \dots, s_n \rangle$ is a right-angled group with $b_{1}^{(2)}(G')=0$. We claim that $G'$ is q-$\psi$-normal. Indeed, set $H= \langle s_{n} \rangle$ and $a=s_{n+1}$, $a^{-1}G'a \cap G' \supseteq H$ and $H$ is $\psi$-regular by assumption. Now the result follows from Lemma~\ref{lemHa}. \end{proof} Using this calculation and the approximation methods from \cite{Kio2018} one can control the growth of Betti numbers in right-angled Artin groups with respect to normal chains with non-trivial intersection. \begin{corollary} Let $A_\Gamma$ be a right-angled Artin group for a finite connected graph $\Gamma$ with generating set $\{\sigma_i \mid i \in \mathcal{I}\}$. Let $N_1 \trianglerighteq N_2 \trianglerighteq \dots$ be a descending chain of finite index normal subgroups in $A_\Gamma$. If the order $\mathrm{ord}_{A_\Gamma/N_n}(\sigma_i)$ in the finite factors $A_\Gamma/N_n$ is unbounded for each generator $\sigma_i$, then \[ \lim_{n \to \infty} \frac{b_1(N_n)}{|A_\Gamma:N_n|} = 0. \] \end{corollary} \begin{proof} Let $\psi_n$ be the character of the permutation action of $A_\Gamma$ on $A_\Gamma/N_n$. Since the sequence of normal subgroups is descending, the sequence $\psi_n$ converges in $\Ch(A_\Gamma)$ to a character $\psi$. Since each character $\psi_n$ is sofic and $A_\Gamma$ is finitely presented, it follows from \cite[Theorem 3.5]{Kio2018} that \[ b_1^\psi(A_\Gamma) = \lim_{n \to \infty} b^{\psi_n}_1(A_\Gamma) = \lim_{n \to \infty} \frac{b_1(N_n)}{|A_\Gamma:N_n|} \] If the order $\mathrm{ord}_{A_\Gamma/N_n}(\sigma_i)$ tends to infinity, $\psi_n(\sigma_i^k)$ vanishes for all $k \neq 0 $ and all large $n$, i.e., $\langle \sigma_i \rangle$ is a $\psi$-regular subgroup. Theorem \ref{thm:right-angled-vanish} implies that $b_1^\psi(A_\Gamma) = 0$ and this completes the proof. \end{proof}
{ "redpajama_set_name": "RedPajamaArXiv" }
4,720
Friday, February 8, 2008, 11:09 - Popular Latin Phrases, Mottos, Slogans, State mottos I can't believe this one almost escaped my attention! The article is appropriately entitled Britain Seeks Its Essence, and Finds Punch Lines. It was a lofty idea: formulate a British �statement of values� defining what it means to be British, much the way a document like the Declaration of Independence sets out the ideals that help explain what it means to be American. The proposal, part of a package of British-pride-bolstering measures announced by Prime Minister Gordon Brown�s new government over the summer, raised a host of tricky questions. What does it mean to be British? How do you express it in a country that believes self-promotion to be embarrassing? And how do you deal with a defining trait of the people you are trying to define: their habit of making fun of worthy government proposals? Detractors spread the rumor that the government was looking not for a considered statement, but for a snappy, pithy �libert�, �galit�, fraternit�-style slogan that it could plaster across government buildings in a kind of branding exercise. Nor did it help when The Times of London cynically sponsored a British motto-writing contest for its readers. The readers� suggestions included �Dipso, Fatso, Bingo, Asbo, Tesco� (Asbo stands for �anti-social behavior order,� a law-enforcement tool, while Tesco is a ubiquitous supermarket chain); �Once Mighty Empire, Slightly Used�; �At Least We�re Not French�; and �We Apologize for the Inconvenience.� The winner, favored by 20.9 percent of the readers, was �No Motto Please, We�re British.� http://www.nytimes.com/2008/01/26/world ... motto.html permalink | related link Esse Quam Videri: To be, rather than to seem (North Carolina State Motto) Tuesday, February 5, 2008, 14:14 - Best Latin Quotes, Words of Wisdom, Proverbs and Sayings, Popular Latin Phrases, Mottos, Slogans, State mottos Esse quam videri - To be, rather than to seem This state motto might as well get the prize for Classical authenticity. Without trying to come up with something unique and of questionable value, North Carolinians went straight to the source of nearly everything that is good in Latin prose: Cicero. Right? Actually, the phrase is a little bit out of context: Virtute enim ipsa non tam multi praediti esse quam videri volunt ("De amicitia", 98). Cicero does not speak of being per se, he speaks about being endowed with virtue. If you are have left is "esse quam videri" it becomes rather unclear: to be what? to seem to be what? Still, a good motto. North Carolina spent a long time coming up with it, being the only one of the original 13 states without an official motto all the way until 1893! Thursday, November 29, 2007, 20:53 - Latin Language, Latin Translation, Popular Latin Phrases, Mottos, Slogans, State mottos Alis volat propriis I am becoming more convinced that if one wants to adopt a good motto it is best to start with a classical quotation. Reasons? See above - the Oregon state motto. I was not able to find any related quotation in the body of Latin literature, both classical and medieval. Honestly, would it make much sense if someone attempted to fly not by her or his own wings? What does this even mean? Is it a sign of some great achievement, to fly with your own wings? It's like speaking with your own voice and so on. I suppose this reveals some degree of self-sufficiency, but not exactly enough to be perceived as a personal trait that would get you noticed... Thus always to tyrants Tuesday, November 27, 2007, 01:58 - World History: Ancient, Medieval & Modern, Latin Language, Latin Translation, Legal Phrases and Expressions, Popular Latin Phrases, Mottos, Slogans, State mottos Sic Semper Tyrannis. Thus always to tyrants. This is of course, state motto of Virginia. But here is a curious, although sincere, example of using Latin in English poetry: Sic Semper Tyrannis" By E. H. O. Clark." Sic semper tyrannis," vile southron? You murdered your own truest friend! And may God now have pity for traitors � Man's patience has come to an end!" Sic semper tyrannis," 0 madman? He marshalled to freedom a race! He led us to battle with tyrants; To dare look the right in the face!" Sic semper tyrannis" assassin? Behold a whole nation in black! And hark to the curse of its millions That rumbles along your track !" Sic semper tyrannis,"� O Heaven ! That motto for slavery's knife; While died the great servant of freedom, As martyrdom sainted his life ! Sic semper tyrannis,"� God help us To bear it � the deed and the loss; The crime that has scarcely been mated Since Jesus was nailed to the cross!" Sic semper tyrannis"� Our Father In Heaven, we swear unto Thee, Once more over him thou hast taken, All men shall be equally free! Troy Daily Times, Sat., April 22d, 1865 I was very surprised that this phrase is also the motto of Allentown, PA. Some serious investigation into the history of this quiet town is required... <<First | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | Temple Church inscription Knowledge is power - original source of the quote Good midwifery tip from Pliny Latin in Iron Man 2. Casanova's mottos
{ "redpajama_set_name": "RedPajamaCommonCrawl" }
3,399
For Tamara Brown, Beauty Is Much More Than Business By Working Woman Report Working Woman Report Working Woman Report is the source for stories about women in business, women in leadership, and news about empowering women. It's updated daily and curated by Emmy Award Winning Journalist, Allison Haunss. beauty businessbusinesscommunityentrepreneurfringe salon spasalonsalon owner By Sasha Vanallen The Wisconsin State Journal WWR Article Summary (tl;dr) Entrepreneur Tamara Brown shares her journey to becoming a salon owner and explains why her beauty business in much more than a place to get your hair done. The Wisconsin State Journal Tamara Brown knew hair was her passion since she was a little girl. Brown, 48, is the owner and head hair stylist at Fringe Salon Spa, 654 S. Gammon Rd. Originally from East St. Louis, Illinois, she realized her passion for hair while attending college at the University of Arkansas-Pine Bluff. But she didn't know if she could fulfill her passion and pursue it as a career. Now, 20 years later, Brown is a successful salon owner who takes pride in serving everyone, regardless of their hair texture. "Coming from an international hair background I have worked with all different types of hair from type one, which is straight, to type four, which is kinky," she said. "I want to always provide the best services for my clients to make sure they feel and look good at the same time." Fringe Salon Spa has been open for 10 years. Some of Brown's clients come into the salon weekly, or even twice a week to get their hair done. Clients can book appointments 6 months in advance. The salon also hosts classes for individuals who want to learn more about how to take care of and manage hair. Brown's mission is to cater to clients' needs while also providing a place where they can come to relax and leave feeling like a new person. On a visit to Fringe Salon Spa, Brown spoke to the Cap Times about her journey to becoming a salon owner and the services she offers to the Madison community. How did you get started in the hair industry? I have always been interested in doing hair since i was a little girl, but I never really thought it was a real career, so I never really took it seriously. I just went to high school and then college and was doing hair in the dorms. I was interested in law but then I realized I didn't want to pour myself into it, so I decided to step back and reevaluate my path. So in the process of that, I went back home and talked to my stylist, and she taught at a cosmetology school, which I had never really seen up close before. What challenges did you face when opening up Fringe Salon Spa? It was a lot of work but it went quicker than most people think. The years of learning and getting experience in doing hair was probably more work than opening the shop. I really wanted to know that if I did open my own hair salon that I had enough knowledge and experience to sustain operating and owning my own business. My grandmother was like an entrepreneur, but I never realized that I did have that exposure in owning a business until after opening my hair salon. Opening the salon in 2010 when it was a recession -- and pretty much everything a person can think of that said I shouldn't open a business -- it took a lot of risk and confidence to jump out there. I would say that one challenge I faced was fear because of the uncertainty that exists with opening a business. I'm the type of person who likes to research and gather knowledge about something, but even with all the research and knowledge I still didn't know if somebody was going to come through the door and receive service from my salon. But with every challenge comes some success. For me one of the beneficial things I did before opening the shop would be taking a small business management class down at the University of Wisconsin-Madison. That really helped me to solidify that I can really do this. I talked to a lawyer who recommended I take the class and other small business classes. These classes helped me understand what I know and don't know and what I need to get and do to cover what I don't know. There are numerous hair salons in the Madison area. How did you build your clientele and get the Fringe name out there? Well, that was a little tricky because where I worked, the salon owner was a smart business woman. She had employees sign a no-compete form, a contract where the person could not work within a five-mile radius from the shop after the person moved. Also, we weren't allowed to tell clients where our new location would be after we left the salon. I respected that. Thankfully everything worked out where I then realized that clients are loyal, and that they appreciate the work I do. Most people are not quick to let anybody do their hair, so I had clients from over a decade follow me and I still do their hair to this day. What are all the services you offer to clients? We offer a wide variety of services for our clients to receive from hair cuts/styles to nails. We try to be the one-stop salon. We do some makeup, I typically do it more case by case because we do not have a makeup artist here on staff. I know how to do makeup and have done clients' makeup for photoshoots. I have clients who have gone on the news and needed their makeup done. We also offer hair removal for clients that can range from eyebrow wax to face wax, to lash extensions and hair coloring. I would say that hair coloring is by far one of the most popular services that my clients get. What made you want to pursue community outreach through hair? My heart with what I do is to make it where it is not just about doing hair. There is such a beautiful aspect of being able to service someone to feel better about the way they look, to brighten their day and that goes for a 1-year old to 99-year old, it really doesn't matter. It's something about feeling kept and cared for and being able to sit down and have that service done really means a lot. So I had a number of people who inquired about how to take care of their children's hair, because some of the kids they adopted did not have the same texture hair as the adopted family. This really made me look into that area and to reach out to Dane County foster services to let them know that we love to service those families. But they reached out to us, quicker than we reached out to them, which was really neat, to let us know about a survey they sent to families asking them what hair salon they recommended and we ranked number one. Because of this, we have been offering small classes for over four years where families can come in to the salon to see how what products and tips work best for their children's hair. ___ Distributed by Tribune Content Agency, LLC. Are Women Electable? Even Democratic Voters Who... A Rare Female Boss In The Construction Industry Questions To Help Up-Level Your Life In Online Dating, Remember The Golden Rule Nearly A Year After Layoffs, Former Oreo Plant Workers Struggle To Find Their Way MORE IN LIFE & STYLE Bethenny Frankel Shares Lessons From Her Mistakes In Relationships Beloved Black Cosmetics Brand Fashion Fair Is Planning Its Comeback … And A Showdown With Rihanna Here's How Horses Are Helping Corporate Types And Entrepreneurs To Stay Focused Thousands Turn Out For Women's March Across The Country How Entrepreneurs Are Working Both Sides Of The Sales Street, Online And Brick-And-Mortar Skin Care Subscription Box Combines Korean Culture And Colombian Heritage
{ "redpajama_set_name": "RedPajamaCommonCrawl" }
6,929
Le Moucherolle des pins (Empidonax affinis) est une espèce de passereau de la famille des Tyrannidae. Répartition Cet oiseau se trouve dans les forêts montagneuses tropicales et subtropicales de conifères du Mexique et du Sud-Ouest du Guatemala. Galerie Sous-espèces D'après Alan P. Peterson, 5 sous-espèces ont été décrites : Empidonax affinis affinis (Swainson) 1827 ; Empidonax affinis bairdi Sclater,PL 1858 ; Empidonax affinis pulverius Brewster 1889 ; Empidonax affinis trepidus Nelson 1901 ; Empidonax affinis vigensis Phillips,AR 1942. Liens externes Voir aussi Empidonax Tyrannidae
{ "redpajama_set_name": "RedPajamaWikipedia" }
4,049
Q: SAS: Create Variable from Proc SQL to use in Macro I want to count the number of unique items in a variable (call it "categories") then use that count to set the number of iterations in a SAS macro (i.e., I'd rather not hard code the number of iterations). I can get a count like this: proc sql; select count(*) from (select DISTINCT categories from myData); quit; I can run a macro like this: %macro superFreq; %do i=1 %to &iterationVariable; Proc freq data=myData; table var&i / out=var&i||freq; run; %mend superFreq; %superFreq I want to know how to get the count into the iteration variable so that the macro iterates as many times as there are unique values in the variable "categories". Sorry if this is confusing. Happy to clarify if need be. Thanks in advance. A: You can achieve this by using the into clause in proc sql: proc sql noprint; select max(age), max(height), max(weight) into :max_age, :max_height, :max_weight from sashelp.class; quit; %put &=max_age &=max_height &=max_weight; Result: MAX_AGE= 16 MAX_HEIGHT= 72 MAX_WEIGHT= 150 You can also select a list of results into a macro variable by combining the into clause with the separated by clause: proc sql noprint; select name into :list_of_names separated by ' ' from sashelp.class; quit; %put &=list_of_names; Result: LIST_OF_NAMES=Alfred Alice Barbara Carol Henry James Jane Janet Jeffrey John Joyce Judy Louise Mary Philip Robert Ronald Thomas William
{ "redpajama_set_name": "RedPajamaStackExchange" }
8,390
Q: Convert tibble into a different shape I am using the data found here: https://www.kaggle.com/cdc/behavioral-risk-factor-surveillance-system. In my R studio, I have named the csv file, BRFSS2015. Below is the code I am trying to execute. BRFSS2015%>% filter(!is.na(WTKG3), CHCKIDNY %in% 1:2) %>% mutate(have_kidney_disease = CHCKIDNY == 1, no_kidney_disease = CHCKIDNY == 2, weight = WTKG3) %>% group_by(have_kidney_disease, no_kidney_disease) %>% summarize(mean_weight = mean(weight), sd_weight = sd(weight), .groups = "drop")%>% mutate_if(is.numeric, round, digits=2) This is giving me output that is a 2x4 tibble. The columns are have_kidney_disease, no_kidney disease, mean_weight, and sd_weight. I want to convert this into a 4x4 format, where have_kidney_disease and no_kidney_disease are rows. How can I do this? A: This code will get you there. If you want to rename columns don't do it inside a mutate but use rename(weight = WTKG3) as a statement after you use select or filter. But as you don't actually use this column except as a replacement in the summarize function, there is no need to do this. I used a case_when statement so you can see how it that works. For just 2 values it is simpler to use an if_else. Below I gave 2 different solutions, the first as you had your code, the second without filtering on CHCKIDNY to show you how the case_when deals with other values. BRFSS2015%>% filter(!is.na(WTKG3), CHCKIDNY %in% 1:2) %>% mutate(kidney_disease = case_when(CHCKIDNY == 1 ~ "yes", CHCKIDNY == 2 ~ "no")) %>% group_by(kidney_disease) %>% summarize(mean_weight = mean(WTKG3), sd_weight = sd(WTKG3)) # A tibble: 2 × 3 kidney_disease mean_weight sd_weight <chr> <dbl> <dbl> 1 no 8081. 2152. 2 yes 8392. 2363. No filter on CHCKIDNY BRFSS2015%>% filter(!is.na(WTKG3)) %>% mutate(kidney_disease = case_when(CHCKIDNY == 1 ~ "yes", CHCKIDNY == 2 ~ "no", TRUE ~ "unknown")) %>% group_by(kidney_disease) %>% summarize(mean_weight = mean(WTKG3), sd_weight = sd(WTKG3)) # A tibble: 3 × 3 kidney_disease mean_weight sd_weight <chr> <dbl> <dbl> 1 no 8081. 2152. 2 unknown 8263. 2252. 3 yes 8392. 2363.
{ "redpajama_set_name": "RedPajamaStackExchange" }
4,802
There is a treasure high in the Colorado Rockies waiting for someone to find it. Jake Martin couldn't care less. Since the death of his wife, all Jake wants is to be left alone in his mountain cabin where he and his dog, Fred, can get on with life. But when it becomes known that the location of the treasure is encrypted in a message left by a nineteenth-century miner, people begin to die, and Jake's good friend and neighbor becomes the number one suspect. The old miner used a copy of The Adventures of Tom Sawyer as a key to his coded message, and the real murderer wants that edition so bad, he's willing to kill for it. Can the amateur sleuths decode the message and stop the murderer, or have Jake and Fred finally met their match?
{ "redpajama_set_name": "RedPajamaC4" }
971
Q: Log4j2 Different appender for different level I have a little problem and can´t find a solution. I want to set pattern layout for level info another than for level warn. If I have a log in level INFO everything is OK, but if the log is levelWARN it is written out into console two times (as level info and as level warn). Simply all logs at a specific level is written out us log at that level and the level below. I want to logs in level INFO write out to console like: "%-5level %d{dd-MM-yyyy HH:mm:ss} %msg%n" and level WARN like "%-5level %d{dd-MM-yyyy HH:mm:ss} [%l] %msg%n". <?xml version="1.0" encoding="UTF-8"?> <Configuration status="INFO"> <Appenders> <Console name="ConsoleInfo" target="SYSTEM_OUT"> <ThresholdFilter level="INFO" onMatch="ACCEPT" onMismatch="DENY"/> <PatternLayout pattern="%-5level %d{dd-MM-yyyy HH:mm:ss} %msg%n"/> </Console> <Console name="ConsoleWarning" target="SYSTEM_OUT"> <ThresholdFilter level="WARN" onMatch="ACCEPT" onMismatch="DENY"/> <PatternLayout pattern="%-5level %d{dd-MM-yyyy HH:mm:ss} [%l] %msg%n"/> </Console> <File name="File" fileName="logs/cli.log"> <PatternLayout pattern="%-5level %d{dd-MM-yyyy HH:mm:ss} [%l] %msg%n"/> </File> </Appenders> <Loggers> <Root level="ALL"> <AppenderRef ref="ConsoleInfo"/> <AppenderRef ref="ConsoleWarning"/> <AppenderRef ref="File"/> </Root> </Loggers> </Configuration> A: As I understand it you want to have log events with a level of WARN or higher (WARN,ERROR,FATAL) go to the "ConsoleWarning" appender only rather than going to both "ConsoleWarning" and "ConsoleInfo". The simplest way to do this would be to modify your filter configuration in your "ConsoleInfo" appender to basically do the opposite approach like this: <Console name="ConsoleInfo" target="SYSTEM_OUT"> <ThresholdFilter level="WARN" onMatch="DENY" onMismatch="ACCEPT"/> <PatternLayout pattern="%-5level %d{dd-MM-yyyy HH:mm:ss} %msg%n"/> </Console> This works because as the log4j2 manual states: This filter returns the onMatch result if the level in the LogEvent is the same or more specific than the configured level and the onMismatch value otherwise. For example, if the ThresholdFilter is configured with Level ERROR and the LogEvent contains Level DEBUG then the onMismatch value will be returned since ERROR events are more specific than DEBUG. This will cause the appender to accept only events that have a level less than WARN. Another possible solution would be to use a RoutingAppender to specify the destination for each level. If you do it this way then you don't need the ThresholdFilters at all. Also note that you can ignore events of specific levels by not providing a default route and not providing a route for that level. For example, if you remove <Route ref="ConsoleInfo" key="DEBUG" /> from the configuration below then all DEBUG events will be ignored by the routing appender and will not be printed to console. Here is the configuration: <?xml version="1.0" encoding="UTF-8"?> <Configuration status="INFO"> <Appenders> <Console name="ConsoleInfo" target="SYSTEM_OUT"> <PatternLayout pattern="%-5level %d{dd-MM-yyyy HH:mm:ss} %msg%n"/> </Console> <Console name="ConsoleWarning" target="SYSTEM_OUT"> <PatternLayout pattern="%-5level %d{dd-MM-yyyy HH:mm:ss} [%l] %msg%n"/> </Console> <File name="File" fileName="logs/cli.log"> <PatternLayout pattern="%-5level %d{dd-MM-yyyy HH:mm:ss} [%l] %msg%n"/> </File> <Routing name="Routing"> <Routes> <Script name="RoutingInit" language="JavaScript"><![CDATA[ logEvent.getLevel();]]> </Script> <Route ref="ConsoleInfo" key="TRACE" /> <Route ref="ConsoleInfo" key="DEBUG" /> <Route ref="ConsoleInfo" key="INFO" /> <Route ref="ConsoleWarning" key="WARN" /> <Route ref="ConsoleWarning" key="ERROR" /> <Route ref="ConsoleWarning" key="FATAL" /> </Routes> </Routing> </Appenders> <Loggers> <Root level="ALL"> <AppenderRef ref="Routing"/> <AppenderRef ref="File"/> </Root> </Loggers> </Configuration> Hope this helps!
{ "redpajama_set_name": "RedPajamaStackExchange" }
3,122
{"url":"http:\/\/mathhelpforum.com\/calculus\/102576-local-max-min.html","text":"# Math Help - local max and min\n\n1. ## local max and min\n\nhi\n\nhave function $g(x)=cosxexp(\\frac{2}{3}sinx)$\n\nhas derivative $g'(x)=(2-3sinx-2sin^2x)exp(\\frac{2}{3}sinx)$\n\ni factor out 1\/3 and ignored $exp(\\frac{2}{3}sinx)$ as it always positive and got $(1-2sinx)(2+sinx)$ then set it to zero and got x= $\\frac{\\pi}{6}$ and $\\frac{5\\pi}{6}$ used these to get coordinates, which i am happy with but when i try to find the max and min locals of g(x) i get the oposite result to the graph of g(x). no matter how i try to look at it i get $\\frac{\\pi}{6}$, $\\frac{\\sqrt{3}}{2}$ as minimum point and $\\frac{5\\pi}{6}$, $\\frac{-\\sqrt{3}}{2}$ as a maximum point. have been using first derivitive test as i am unable to work out how to get g''(x)\n\nany help would be great\n\n2. I dont know how you conducted the first devaritive test, but observe that g'(x)>0 for 0<x<(pi\/4)\nand g'(x)<0 for pi\/4<x<5pi\/4. Cos(x) decides the sign because exp(2\/3sin(x)) is always > 0.\n\nThat means that on the interval 0<x<(pi\/4) there is a local maxima, and on the interval pi\/4<x<5pi\/4, there is a local minima.\n\nLooks like you just turned that one around.\n\n3. Originally Posted by smartcar29\nhi\n\nhave function $g(x)=cosxexp(\\frac{2}{3}sinx)$\n\nhas derivative $g'(x)=(2-3sinx-2sin^2x)exp(\\frac{2}{3}sinx)$\n\ni factor out 1\/3 and ignored $exp(\\frac{2}{3}sinx)$ as it always positive and got $(1-2sinx)(2+sinx)$ then set it to zero and got x= $\\frac{\\pi}{6}$ and $\\frac{5\\pi}{6}$ used these to get coordinates, which i am happy with but when i try to find the max and min locals of g(x) i get the oposite result to the graph of g(x). no matter how i try to look at it i get $\\frac{\\pi}{6}$, $\\frac{\\sqrt{3}}{2}$ as minimum point and $\\frac{5\\pi}{6}$, $\\frac{-\\sqrt{3}}{2}$ as a maximum point. have been using first derivitive test as i am unable to work out how to get g''(x)\n\nany help would be great\n1. The roots of the equation\n\n$2-3 \\sin(x)-2\\sin^2(x)=0$\n\ncorresponding to your first factor are $x=\\pi\/6 + 2 \\pi n$ and $x=2 \\pi\/3+2 \\pi m$, for all $n, \\ m \\in \\mathbb{Z}$, and similarly there are multiple roots corresponding to the other factor.\n\n2. How are you determining if a root corresponds to a maximum or minimum?\n\n3. How are you evaluating $g(x)$?\n\nCB","date":"2015-04-19 05:50:25","metadata":"{\"extraction_info\": {\"found_math\": true, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 25, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.6412470936775208, \"perplexity\": 476.8156239304547}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2015-18\/segments\/1429246637445.19\/warc\/CC-MAIN-20150417045717-00214-ip-10-235-10-82.ec2.internal.warc.gz\"}"}
null
null
<!DOCTYPE html> <html> <head> <meta charset="UTF-8"> <title>CI Web Shop</title> <link rel="stylesheet" type="text/css" href="..\node_modules\bootstrap\dist\css\bootstrap.css"> <link rel="stylesheet" type="text/css" href="..\css\layout.css"> <link rel="stylesheet" type="text/css" href="..\css\utils.css"> <script src="..\node_modules\angular\angular.js"></script> <script src="..\node_modules\jquery\dist\jquery.js"></script> <script src="..\node_modules\bootstrap\dist\js\bootstrap.js"></script> <script src="..\scripts\repository.js"></script> <script src="..\scripts\shopping-cart.js"></script> </head> <body ng-app="shopApp"> <nav class="navbar navbar-default navbar-fixed-top"> <div class="container"> <div class="navbar-header"> <button type="button" class="navbar-toggle collapsed" data-toggle="collapse" data-target="#menu"> <span class="icon-bar"></span> <span class="icon-bar"></span> <span class="icon-bar"></span> </button> <a class="navbar-brand" href="..\index.html">CI Web Shop</a> </div> <div class="collapse navbar-collapse" id="menu"> <ul class="nav navbar-nav navbar-right"> <li><a href="#"><span class="glyphicon glyphicon-user"></span> Login</a></li> <li><a href="shopping-cart.html"><span class="glyphicon glyphicon-shopping-cart"></span> Order</a></li> </ul> <div class="navbar-form navbar-right"> <div class="form-group"> <input class="form-control" placeholder="Search" ng-model="query"> </div> <a href="search.html?q={{query}}" class="btn btn-default"> <span class="glyphicon glyphicon-search"></span> </a> </div> </div> </div> </nav> <div class="container" ng-controller="shoppingCartController"> <div class="row" ng-hide="lines.length"> <div class="col-lg-12 text-center"> <p>There are no items in your shopping cart...</p> <p>You should do some shopping!</p> </div> </div> <div ng-repeat="line in lines"> <div class="thumbnail"> <div class="row"> <div class="col-lg-2"> <a href="product.html?id={{line.product.id}}"> <img src="http://placehold.it/150x150" alt="..." /> </a> </div> <div class="col-lg-7"> <div class="caption"> <a href="product.html?id={{line.product.id}}"> <h3>{{line.product.name}}</h3> </a> <p class="label label-default">{{line.product.category}}</p> </div> </div> <div class="col-lg-1 clearfix"> <div class="caption pull-right"> <input ng-model="line.number" /> <button type="button" ng-click="removeLine(line)"> <span class="glyphicon glyphicon-trash"></span> Delete </button> </div> </div> <div class="col-lg-2 clearfix"> <div class="caption pull-right"> <p>{{'&euro; ' + line.subTotal()}}</p> </div> </div> </div> </div> </div> <div class="row"> <div class="col-lg-12"> <p class="pull-right">{{'&euro;' + total()}}</p> </div> </div> <footer class="footer"> <p>Copyright &copy; 2017</p> </footer> </div> </body> </html>
{ "redpajama_set_name": "RedPajamaGithub" }
3,455
Q: Configuration lost in my server after rebooting I have done apt-get update in my server linux - debian and after rebooting the server it get lost all the configuration. I mean it is booting from different partition. Any idea of what happened? A: If your other partition is still there, it should be enough to mount it to any directory : mount /dev/old_partition666 /any_dir then you can try to find your missing etc. To avoid it in future use etckeeper
{ "redpajama_set_name": "RedPajamaStackExchange" }
5,060
\section{Introduction} \label{sec:introduction} {\let\thefootnote\relax\footnotetext{The views expressed in this paper are the personal views of the authors and do not necessarily reflect the views or policies of their current or past employers.}} The highly successful deep learning paradigm~\citep{2015nature} receives a lot of attention in science and engineering, in many different forms and flavors. Within numerical mathematics, the machine learning methodology has successfully entered the field of numerically solving partial differential equations (PDEs)~\citep{data-driven-pde2019,2018SDE-DNN,Han8505,PiNN2019,SIRIGNANO20181339,TempoGAN2018}. The aim with machine learning is to either speed up the solution process or to solve high-dimensional problems that are not easily handled by the traditional numerical methods. There are essentially two types of deep learning approaches, to approximate the solution of PDEs (see~\citep{2018SDE-DNN,Han8505,PiNN2019,SIRIGNANO20181339,TempoGAN2018}), and to design advanced numerical schemes, for example, the authors of \cite{data-driven-pde2019}, derived a data-driven high-order discretization to solve PDEs accurately on a coarse grid. In this paper we will develop a highly accurate numerical discretization scheme for stochastic differential equations (SDEs), which is based on taking possibly large discrete time steps. We ``learn'' to take large time steps, with the help of the Stochastic Collocation Monte Carlo sampler (SCMC) proposed by~\cite{scmc2019}, and by using an artificial neural network (ANN), within the classical supervised learning context. SDEs are widely used to describe uncertain phenomena, in physics, finance, epidemics, amongst others, as a means to model and quantify uncertainty. The corresponding solutions are stochastic processes. Numerical approximation of the solution to an SDE is standard practice, as an analytic solution is typically not available. The most commonly known technique to solve SDEs is based on Monte Carlo (MC) simulation, for which the SDE first needs to be discretized. There are quite a few applications, that could benefit from an accurate and efficient numerical method on the basis of a large time step discretization~\citep{2021largetime-step}, like, in finance, the valuation of path-dependent financial derivatives or financial risk management where counterparty credit risk play a role. Basically, there are two ways to measure the convergence rate of discrete solutions to SDEs, by means of the approximation to the sample path or by approximation to the corresponding distribution. This way, strong and weak convergence of a numerical SDE solution have respectively been defined (see~\cite{intro_sde1999}). Weak convergence, the convergence in distributional sense, is often addressed in the literature. Moment-matching, for example, is a basic technique to improve weak convergence. Strong, path-wise, convergence is particularly challenging, and requires accurate {\it conditional distributions}. There are natural approaches to improve strong convergence properties, i.e. by adding higher order terms or by using finer time grids. However, these are nontrivial and costly, especially when considering multi-dimensional SDEs. We aim to develop highly accurate numerical schemes by means of deep learning, for which the strong error of the discretization does not depend on the size of the simulation time step. For this, we will employ the SCMC method as an efficient approach for approximating (conditional) distribution functions. The distribution function of interest is then expanded as a polynomial in terms of a random variable which is {\it cheap to sample from} at given collocation points, and interpolation takes place between these points. The resulting big time steps discretization, in which the SCMC methodology is combined with deep learning, is called {\em the Seven-League scheme}\footnote{With seven-league boots, we are marching through the time-wise direction, see also \url{https://en.wikipedia.org/wiki/Seven-league_boot} } here, and we abbreviate it by {\em the 7L scheme}. There are different reasons to learn stochastic collocation points instead of the sample paths directly. Stochastic collocation points have a specific physical meaning, which makes the data-driven scheme explainable. Monte Carlo sample paths are random, while collocation points are path independent and deterministic (i.e. representing key features of a probability distribution), which simplifies the learning process when using neural networks. Unlike the SCMC method, which provides accurate Monte Carlo samples given a constant time step and for one specific instance of the SDE parameters, the 7L methodology enables us to generate samples for a wide range of time steps and for many different instances of the model parameters (i.e. for a family of SDEs), by means of a neural network to learn the evolution of the collocation points over time for many different model parameters. The 7L scheme is composed of two separate phases under the framework of supervised learning, i.e., an off-line (training) phase and an on-line (prediction) phase. The training phase, which usually requires heavy computation and many data sets, is done only once and offline. The prediction phase, which is a computationally cheap and highly efficient process, can be performed in an on-line fashion. The remainder of this paper is organized as follows. In Section~\ref{sec:sde}, SDEs, their discretization, stochastic collocation and the connection between SDE discretizations and the SCMC method are introduced. In Section~\ref{sec:method}, the data-driven methodology is explained to address large time step simulation, i.e. the {\em 7L scheme}, for SDEs. ANNs will be used as function approximators to learn the stochastic (conditional) collocation points. A brief description of their details is placed in Section~\ref{sec:ann}. In Section~\ref{sec:cdc}, we introduce a decompression-compression technique to accelerate the computation. This latter efficient variant is named the {\em 7L-CDC scheme} (i.e., seven-league compression-decompression scheme). Section~\ref{sec:results} presents numerical experiments to show the performance of the proposed approach. Furthermore, the corresponding error is analyzed. Section~\ref{cenou} concludes. \section{Stochastic differential equations and stochastic collocation}\label{sec:sde} We first describe the basic, well-known SDE setting, and explain our notation. \subsection{SDE basics}\label{sc:sdeb} We work with a real-valued random variable $Y(t)$, defined on the probability space $(\Omega,\Sigma,\mathbb{P})$ with filtration $\mathcal{F}_{t\in[0,T]}$, sample space $\Omega$, $\sigma$-algebra $\Sigma$ and probability measure $\mathbb{P}$. For the time evolution of $Y(t)$, consider the generic scalar It\^o SDE, \begin{equation}\label{1a} \dY(t) = a(t,Y(t),{\boldsymbol\theta}){\rm d}t + b(t,Y(t),{\boldsymbol\theta}) \dW(t), \;\;\;\;\; 0\leq t \leq T, \end{equation} with the drift term $a(t,Y(t),\boldsymbol\theta)$, the diffusion term $b(t,Y(t),\boldsymbol\theta)$, model parameters $\boldsymbol\theta$, Wiener process $W(t)$, and given initial value $Y_0:=Y(t=0)$. When the drift and diffusion terms satisfy some regularity conditions (e.g., the global Lipschitz continuity~\citep[p.289]{karatzas1988brownian}), existence and uniqueness of the solution of~(\ref{1a}) are guaranteed. The cumulative distribution function of $Y(t)$, $t\in[0,T]$, $F_{Y(t)}(\cdot)$, is available and the corresponding density function, evolving over time, is described by the Fokker-Planck equation~\citep{risken1984fokker}. With a discretization in time interval $[0,T]$, $t_i=i\cdot T/N, \; i=0,\ldots N,$ with equidistant time step $\Delta t = t_{i+1}-t_i$, the discrete random variable at time $t_i$ is denoted by $Y(t_i)$. Traditional numerical schemes have been designed based on It\^o's lemma, in a similar fashion as the Taylor expansion is used to discretize deterministic ODEs and PDEs. The basic discretization, for each Monte Carlo path, is the Euler-Maruyama scheme~\citep{intro_sde1999}, which reads, \begin{equation} \hat{Y}_{i+1} = \hat{Y}_{i} + a(t_i, \hat{Y}_i,\boldsymbol\theta) \Delta t+ b(t_i, \hat{Y}_i,\boldsymbol\theta) \sqrt{\Delta t} \hat{X}_{i+1}, \label{eq:euler} \end{equation} where $\hat{Y}_{i+1}:=\hat{Y}(t_{i+1})$ is a realization (i.e., a number) from random variable $\tilde{Y}(t_{i+1})$, which represents the numerical approximation to exact solution $Y(t_{i+1})$ at time point $t_{i+1}$, and a realization $\hat{X}_{i+1}$ is drawn from the random variable $X$, which here follows the standard normal distribution $\mathcal{N}(0,1)$. Moreover, $\check{Y}(t_i)$ (a number) will be used as the notation for a realization of $Y(t_i)$. In addition, the Milstein discretization~\citep{Milstein1975} reads, \begin{equation} \hat{Y}_{i+1} = \hat{Y}_{i} + a(t_i, \hat{Y}_i,\boldsymbol\theta) \Delta t+ b(t_i, \hat{Y}_i,\boldsymbol\theta) \sqrt{\Delta t} \hat{X}_{i+1} + \frac{1}{2} b'(t_i, \hat{Y}_i, \boldsymbol\theta) b(t_i, \hat{Y}_i,\boldsymbol\theta) \Delta t(\hat{X}_{i+1}^2 -1), \label{eq:milstein} \end{equation} where $b'(t_i, \cdot,\boldsymbol\theta)$ represents the derivative with respect to $\hat{Y}$ of $b(\cdot,\hat{Y},\boldsymbol\theta)$. When the drift and diffusion terms are independent of time $t$, the SDE is called time-invariant. Two error convergence criteria are commonly used to measure the SDE discretization accuracy, that is, the convergence in the weak and strong sense. Strong convergence, which is of our interest here, is defined as follows. \begin{definition} Let ${Y}(t_i)$ be the exact solution of an SDE at time $t_i$, its discrete approximation $\tilde{Y}(t_i)$ with time step $\Delta t \in \mathbb{R}^+$ converges in the strong sense, with order $\beta_s \in \mathbb{R}^+$, if there exists a constant $K$ such that \begin{equation} \E|\tilde{Y}(t_i) - {Y}(t_i)| \leq K(\Delta t)^{\beta_s}. \label{sstr} \end{equation} \end{definition} It is well-known that the Euler-Maruyama scheme~\eqref{eq:euler} has strong convergence $\beta_s=0.5$, while the Milstein scheme~\eqref{eq:milstein} has $\beta_s$ = 1.0. When deriving high order schemes for SDEs, the rules of It\^o calculus must be respected~\citep{intro_sde1999}. As a result, there will be eight terms in a Taylor SDE scheme with $\beta_s$ = 1.5, and twelve with $\beta_s$ = 2.0, and the computational complexity increases. As a consequence, higher order schemes are involved and somewhat expensive. Convergence of the numerical solution for $\Delta t \rightarrow 0$ is guaranteed, but the computational costs increase significantly to achieve accurate solutions. The generic form of the above mentioned numerical schemes to solve the It\^o SDE is as follows, \begin{equation} \label{eq:ddscheme} \hat{Y}_{i+1} | \hat{Y}_i= \sum_{j=0}^{m-1}\alpha_{j} \hat{X}_{i+1}^{j}, \end{equation} where $m$ represents the number of polynomial terms, the coefficients $\alpha_j$ are pre-defined and equation-dependent. For example, for the Euler-Maruyama scheme~\eqref{eq:euler}, with $m=2$, we have \begin{equation} \label{eq:euler_coef} \begin{cases} \alpha_0 = \hat{Y}_{i} + a(t_i,\hat{Y}_i,\boldsymbol\theta) \Delta t, \\ \alpha_1 =b (t_i, \hat{Y}_i,\boldsymbol\theta) \sqrt{\Delta t}, \\ \end{cases} \end{equation} while for the Milstein scheme, with $m=3$, it follows that \begin{equation}\label{eq:milstein_coef} \begin{cases} \alpha_0 = \hat{Y}_i + a(t_i, \hat{Y}_i,\boldsymbol\theta) \Delta t + \frac{1}{2}b'(t_i, \hat{Y}_i, \boldsymbol\theta) b(t_i,\hat{Y}_i,\boldsymbol\theta), \\ \alpha_1 =b (t_i, \hat{Y}_i,\boldsymbol\theta) \sqrt{\Delta t}, \\ \alpha_2 = \frac{1}{2} b'(t_i,\hat{Y}_i,\boldsymbol\theta) b(t_i, \hat{Y}_i,\boldsymbol\theta). \\ \end{cases} \end{equation} With these explicit coefficients we arrive at the probability distribution of the random variable, \begin{equation} \label{eq:distr_Yt} Y(t_{i+1}) | Y(t_i) \approx \tilde{Y}(t_{i+1}) |\tilde{Y} (t_i) \stackrel{d}{=} \sum_{j=0}^{m-1}\alpha_{j} X^{j}. \end{equation} These discrete SDE schemes are based on a series of transformations of the previous realization to approximate the conditional distribution, \begin{equation} \label{eq:condistr} \P \big[Y(t+\Delta t)<y|Y(t) \big]=F_{Y(t + \Delta t)|Y(t)}(y) \approx F_{\tilde{Y}(t+\Delta t)|\tilde{Y}(t)}(y). \end{equation} A numerical scheme is thus essentially based on conditional sampling of $Y(t+\Delta t)|Y(t)$. The Euler-Maruyama scheme draws from a normal distribution, with a specific mean and variance, to approximate the distribution in the next time point, while the Milstein scheme combines a normal and a chi-squared distribution. Similarly, we can derive the stochastic collocation methods. \subsection{ Stochastic collocation method}\label{sc:scmc} Let's assume two random variables, $Y$ and $X$, where the latter one is cheaper to sample from (e.g., $X$ is a Gaussian random variable). These two scalar random variables are connected, via, \begin{equation} F_Y(Y) \stackrel{d}{=} U \stackrel{d}{=} F_X(X), \end{equation} where $U\sim \mathcal{U}([0,1])$ is a uniformly distributed random variable, $F_Y(\bar y):=\P[Y \leq \bar y]$ and $F_X(\bar x):=\P[X \leq \bar x]$ are cumulative distribution functions (CDF). Note that $F_X(X)$ and $F_Y(Y)$ are random variables following the same uniform distribution. $F_Y(\bar y_n)$ and $F_X(\bar x_n)$ are supposed to be strictly increasing functions, so that the following inversion holds true, \begin{equation} \label{eq:xydist} \bar y_n = F^{-1}_Y(F_X(\bar x_n))=:g(\bar x_n) . \end{equation} where $\bar y_n$ and $\bar x_n$ are samples (numbers) from $Y$ and $X$, respectively. The mapping function, $g(\cdot) = F^{-1}_Y(F_X(\cdot))$, connects the two random variables and guarantees that $F_X(\bar x_n)$ equals $F_Y(g(\bar x_n))$, in distributional sense and also element-wise. The mapping function should be approximated, i.e., $g(\bar x_n)\approx g_m(\bar x_n)$, by a function which is cheap. When function $g_m(\cdot)$ is available, we may generate ``expensive'' samples, $\bar y_n$ from $Y$, by using the cheaper random samples $\bar x_n$ from $X$. The Stochastic Collocation Monte Carlo method (SCMC) developed in~\cite{scmc2019} aims to find an accurate mapping function $g(\cdot)$ in an efficient way. The basic idea is to employ Equation~\eqref{eq:xydist} at specific collocation points and approximate the function $g(\cdot)$ by a suitable monotonic interpolation between these points. This procedure, see Algorithm~I, reduces the number of expensive inversions $F^{-1}_Y(\cdot)$ to obtain many samples from $Y(\cdot)$. \zbox{ {\bf Algorithm I: SCMC Method} \\ \label{alg1} Taking an interpolation function of degree $m-1$ (with $m\geq2$, as we need at least two collocation points), as an example, the following steps need to be performed: \begin{itemize} \item[1.] Calculate CDF $F_X(x_j)$ on the points $(x_1,x_2, ..., x_m)$, that are obtained, for example, from Gauss-Hermite quadrature, giving $m$ pairs $({x}_j, F_X({x}_j))$; \item[2.] Invert the target CDF ${y}_j=F^{-1}_Y(F_X({x}_j))$, $j=1,\ldots, m$, and form $m$ pairs $({x}_j, {y}_j)$; \item[3.] Define the interpolation function, $y=g_m(x)$, based on these $m$ point pairs $({x}_j,{y}_j)$; \item[4.] Obtain sample $\hat{Y}$ by applying the mapping function $\hat{Y}=g_m(\hat{X})$, where sample $\hat{X}$ is drawn from $X$. \end{itemize} } The SCMC method parameterizes the distribution function by imposing probability constraints at the given collocation points. Taking the Lagrange interpolation as an example, we can expand function $g_m(\cdot)$ in the form of polynomial chaos, \begin{equation} \label{eq:scmc-lang} Y \approx g_m(X) = \sum_{j=0}^{m-1}\hat\alpha_{j} X^{j} = \hat\alpha_0+\hat\alpha_1X + ... + \hat\alpha_{m-1} X^{m-1}. \end{equation} Monotonicity of interpolation is an important requirement, particularly when dealing with peaked probability distributions. The Cameron-Martin Theorem~\citep{Cameron1947AnnMath} states that any distribution can be approximated by a polynomial chaos approximation based on the normal distribution, but also other random variables may be used for $X$ (see, for example,~\cite{scmc2019}). Clearly, Equation~\eqref{eq:distr_Yt} can be compared to Equation~\eqref{eq:scmc-lang}, as a discretization scheme to approximate the realization in the next time point. \section{Methodology} \label{sec:method} For our purposes, given $Y(t)$, the conditional variable $Y(t+\Delta t)$ can be written as, \begin{equation} \label{eq:condY} Y(t + \Delta t)|Y(t) \approx g_m(X) = \sum_{j=0}^{m-1}\hat\alpha_{j}X^{j}, \end{equation} where the coefficients $\hat\alpha_j \equiv \hat\alpha_j(\hat{Y}_i, t_i, t_{i+1},\Delta t,\boldsymbol\theta)$ are now functions of realization $\hat{Y}_i$. Equation~\eqref{eq:condY}, with large $m$-values, holds for any $\Delta t$, particularly also for large $\Delta t$. As such the scheme can be interpreted as an {\em almost exact simulation scheme} for an SDE under consideration. By the scheme in~\eqref{eq:condY} we can thus take large time steps in a highly accurate discretisation scheme. More specifically, a sample from the known distribution $X$ can be mapped onto a corresponding unique sample of the conditional distribution $Y(t+\Delta t)$ by the coefficient functions. There are essentially two possibilities for using an ANN in the framework of the stochastic collocation method, the first being to directly learn the (time-dependent) polynomial coefficients, $\hat\alpha_{j}$, in~\eqref{eq:condY}, the second to learn the collocation points, ${y}_j$. The two methods are equivalent mathematically, but the latter, our method of choice, appears more stable and flexible. Here, we explain how to learn the collocation points, ${y}_i$, which is then followed by inferring the polynomial coefficients. When the stochastic collocation points at time $t+\Delta t$ are known, the coefficients in~\eqref{eq:condY} can easily be computed. An SDE solution is represented by its cumulative distribution at the collocation points, plus a suitable accurate interpolation $g_m(x)$. In other words, the SCMC method forces the distribution functions (the target and the numerical approximation) to strictly match at the collocation points over time. The collocation points are dynamic and evolve with time. \subsection{Data-driven numerical schemes} \label{sec:CondSCMC} Calculating the conditional distribution function requires generating samples conditionally on previous realizations of the stochastic process. Based on a general polynomial expression, the conditional sample, in discrete form, is defined as follows, \begin{equation} \label{eq:condsam} \hat{Y}_{i+1}|\hat{Y}_i = \sum_{j=0}^{m-1}\hat\alpha_{i+1,j} \left(\hat{Y}_{i}, t_i, t_{i+1}-t_i, \boldsymbol\theta \right) \hat{X}_{i+1}^{j}, \end{equation} where $\Delta t = t_{i+1} -t_i$, and the coefficients $\hat\alpha_{i+1,j}$ are functions of the variables $\hat{Y}_{i}, t_i, t_{i+1}-t_i, \boldsymbol\theta$, for example, see Formulas~\eqref{eq:euler_coef} and~\eqref{eq:milstein_coef}. In the case of a Markov process, the future doesn't dependent on past values. Given $\hat Y(t_i)$, the random variable $\hat{Y} (t_{i+1})$ only depends on the increment $Y (t_{i+1})-Y (t_i)$. The process has independent increments, and the conditional distribution at time $t_{i+1}$ given information up to time $t_i$ only depends on the information at $t_i$. Similar to these coefficient functions, the $m$ {\em conditional} stochastic collocation points at time $t_{i+1}$, $y_j(t_{i+1})|\hat{Y}_i$, with $j=0,\ldots,m-1$, can be written as a functional relation, \begin{equation} {y}_j (t_{i+1}) | \hat{Y}_i = H_j \left(\hat{Y}_i, t_i, t_{i+1}-t_i, \boldsymbol\theta \right). \label{eq:markov-Hfun} \end{equation} A closed-form expression for function $H(\cdot)$ is generally not available. Finding the conditional collocation points can however be formulated as a regression problem. It is well-known that neural networks can be utilized as universal function approximators~\citep{Cybenko1989}. We then generate random data points in the domain of interest and the ANN should ``learn the mapping function $H_j(\cdot)$'', in an off-line ANN training stage. The SCMC method is here used to compute the corresponding collocation points at each time point, which are then stored to train the ANN, in a supervised learning fashion (see, for example,~\cite{Goodfellow-et-al-2016}). \subsection{The Seven-League scheme} Next, we detail the generation of the stochastic collocation points to create the training data. Consider a stochastic process $Y(\tau)$, $\tau\in[0,\tau_{max}]$, where $\tau_{max}$ represents the maximum time horizon for the process that we wish to sample from. When the analytical solution of the SDE is not available (and we cannot use an exact simulation scheme with large time steps), a classical numerical scheme will be employed, based on {\em tiny constant time increments} $\Delta \tau=\tau_{i+1}-\tau_i$, a discretization in the time-wise direction with grid points $0 < \tau_1 <\tau_2 < \ldots <\tau_N \leq \tau_{max}$, to generate a sufficient number of highly accurate samples at each time point $\tau_i$, to approximate the corresponding cumulative functions highly accurately. With the obtained samples, we approximate the collocation points, as follows, \begin{equation} \label{eq:enq1} \hat{y}_j(\tau_i)=F^{-1}_{\tilde{Y}(\tau_i)}(F_X({x}_j)) \approx F^{-1}_{Y(\tau_i)}(F_X({x}_j)) \end{equation} where $\hat{y}_j(\cdot)$ represents the approximate collocation points of $Y(t)$ at time $\tau_i$, and ${x}_j$, $j=1,\dots,M_s$, are collocation points of variable $X$. For simplicity, consider $X\sim\mathcal{N}(0,1)$, so that the points ${x}_j$ are known analytically and do not depend on time point $\tau_i$. In the case of a normal distribution, these points are known quadrature points, and tabulated, for example, in~\citep{scmc2019}. After this first step, we have the set of collocation points, $\hat{y}_j(\tau_i)$, for $i=1,\dots,N$ and $j=1,\dots,M_s$. Subsequently, the $\hat{y}_j(\cdot)$ from~\eqref{eq:enq1} are used as the ground-truth to train the ANN. In the second step, we determine the {\it conditional} collocation points. For each time step $\tau_i$ and collocation point indexed by $j$, a {\em nested Monte Carlo simulation} is then performed to generate the conditional samples. Similar to the first step, we obtain the conditional collocation points from each of these sub-simulations using~\eqref{eq:enq1}. This yields the following set of $M_c$ conditional collocation points, \begin{eqnarray} \label{eqn2} \hat{y}_{k|j}(\tau_{i+1}):=\hat{y}_{k}(\tau_{i+1})|\hat{y}_j(\tau_i)=F^{-1}_{\hat{Y}(\tau_{i+1})|\hat{Y}(\tau_i)=\hat{y}_j(\tau_i)}\biggl(F_X(x_{k|j})\biggr), \end{eqnarray} where $x_{k|j}$ is a conditional collocation point, and $i \in \{0, 1,\dots,N-1\}$, $j \in \{1,\dots, M_s\}$, $k \in \{1,\dots,M_c\}$. Note that, in the case of Markov processes, the above generic procedure can be simplified by just varying the initial value $Y_0$ instead of running a nested Monte Carlo simulation. Specifically, we then set $\hat{Y}_i=\hat{Y}_0$, $\tau_i=\tau_0$ and $\tau_{i+1}=\tau_0+\Delta\tau$ to generate the corresponding conditional collocation points. The inverse, $F^{-1}_{\hat{Y}(\tau_{i+1})|\hat{Y}(\tau_i)}(\cdot)$, is often not known analytically, and needs to be derived numerically. An efficient procedure for this is presented in~\citep{LLech}. Of course, it is well-known that the computation of $F^{-1}(p)$ is equivalent with the computation of the quantile at level $p$. We encounter essentially four types of stochastic collocation (SC) points: ${x}_j$ are called the original SC points, $\hat{x}_j$ are original conditional collocation points, $\tilde{y}_j$ are the marginal SC points, and $\hat{y}_k|\cdot$ are the conditional SC points. For example, $\hat{y}_{k}|\hat{Y_i}$ is conditional on a realization $\hat{Y}_i$. When a previous realization happens to be a collocation point, e.g., $\hat{Y}_i=\hat{y}_j$, we have $\hat{y}_{k|j}:=\hat{y}_{k}|\hat{y}_j$. When the data generation is completed, the ANNs are trained on the generated SC points to approximate the function $H$ in~\eqref{eq:markov-Hfun}, giving us a learned function $\hat{H}$. This is called the training phase. With the trained ANNs, we can approximate new collocation points, and develop a numerical solver for SDEs, which is the Seven-League scheme (7L), see Algorithm~II. Figure~\ref{fig:scheme-ann-scmc} gives a schematic illustration of Monte Carlo sample paths that are generated by the 7L scheme. \zbox{ {\bf Algorithm II: 7L Scheme} \\ \label{alg2} \begin{enumerate} \item Offline stage: Train the ANNs to learn the stochastic collocation points. At this stage, we choose different ${\boldsymbol{\theta}}$ values, simulate corresponding Monte Carlo paths, with small constant time increments $\Delta \tau=\tau_{i+1}-\tau_i$ in $[0,\tau_{max}]$, generate the $\hat y_j$ and $\hat y_{k|j}$ collocation points, and learn the relation between input and output. So, we actually ``learn'' $H_k\approx \hat H_k$. See Section~\ref{sec:ann} for the ANN details. \item Online stage: Partition time interval $[0,T]$, $t_i=i\cdot T/N, \; i=0,\ldots N,$ with equidistant time step $\Delta t = t_{i+1}-t_i$. Given a sample $\hat{Y}_i$ at time $t_i$, compute $m$ collocation points at time $t_{i+1}$ using \begin{equation} \label{eq:funH} \hat{y}_j(t_{i+1}) | \hat{Y}_i = {\hat{H}_j} (\hat{Y}_i, t_i, t_{i+1}-t_i, \boldsymbol\theta ), j=1,2,\ldots,m, \end{equation} and form a vector ${\bf{\hat{y}}}_{i+1}=(\hat{y}_1(t_{i+1})| \hat{Y}_i,\hat{y}_2(t_{i+1})| \hat{Y}_i,\ldots,\hat{y}_m(t_{i+1})| \hat{Y}_i)$. \item Compute the interpolation function $g_m(\cdot)$, or calculate the coefficients $\hat{\boldsymbol{\alpha}}_{i+1}$ (if necessary): \begin{equation} A({x}_{k|i+1}) \hat{\boldsymbol{\alpha}}_{i+1} = {\hat{\bf y}}_{i+1}, \label{eq:gm} \end{equation} see Algorithm~I for details on the computation of original collocation points. We will compare monotonic spline, Chebyshev and the barycentric formulation of Lagrange interpolation for this purpose. See Section~\ref{sec:interp} for a detailed discussion. \item Sample from $X$ and obtain a sample in the next time point, $\hat{Y}_{i+1}$, by $\hat{Y}_{i+1} | \hat{Y}_{i} = g_m(\hat{X}_{i+1})$, or the coefficient form as follows, $$\hat{Y}_{i+1} | \hat{Y}_{i} = \sum_{j=0}^{m-1}\hat{\alpha}_{i+1,j} \hat{X}_{i+1}^{j}.$$ \item Return to Step 2 by $t_{i+1} \xrightarrow{} t_{i}$, iterate until terminal time $T$. \item Repeat this procedure for a number of Monte Carlo paths. \end{enumerate} } \begin{figure}[htb!] \centering \subfloat[Sample paths by 7L]{\label{fig:7L-3D} \includegraphics[width=0.5\textwidth]{7L_scheme_3D}} \subfloat[The 2D projection]{\label{fig:paths-7L-fig}{\includegraphics[width=0.5\textwidth]{7L_scheme}}} \caption{ Schematic diagram of the 7L scheme. Left: Sample paths generated by 7L. Right: The 2D projection of Figure~\ref{fig:7L-3D}. Here conditional SC points, represented by $\blacksquare$, are conditional on a previous realization, denoted by $\bigstar$. ``Conditional PDF'' is the conditional probability density function, defined by these conditional SC points. The density function, which is not required by 7L, is plotted only for illustration purposes.} \label{fig:scheme-ann-scmc} \end{figure} \begin{remark}[Lagrange interpolation issue] In the case of classical Lagrange interpolation, matrix $A(x_{k|i})$ would be the Vandermonde matrix. In that case, it should not get too large, as the matrix would then suffer from ill-conditioning. However, when employing orthogonal polynomials, this drawback is removed. More details can be found in~\cite{scmc2019}. \end{remark} When the approximation errors from ANN and SCMC are sufficiently small, the strong convergence properties of the 7L scheme can be estimated, as follows, \begin{equation} \E|\tilde{Y}(t_i) - {Y}(t_i)| < \epsilon (\Delta \tau) \ll K(\Delta t)^{\beta_s}, \end{equation} where time step $\Delta \tau$ is used to define the ANN training data-set, and the actual time step $\Delta t$ is used for ANN prediction, with $\Delta \tau \ll \Delta t$. Based on the trained 7L scheme, the strong error, $\epsilon(\Delta \tau)$, does thus not grow with the actual time step $\Delta t$. Particularly, let's assume $\Delta \tau= \Delta t/\kappa$, for example $\kappa=100$, when employing the Euler-Maruyama scheme with time step $\Delta \tau$ during the ANN learning phase, we expect a strong convergence of $O(\sqrt{\Delta \tau})$, which then equals $O(\sqrt{\Delta t/\kappa})$, while the use of the Milstein scheme during training would result in $O(\Delta t/\kappa)$ accuracy. When $\kappa=100$, the time step during the learning phase is 100 times smaller than $\Delta t$, which has a corresponding effect on the overall scheme's accuracy in terms of its strong, path-wise convergence. {\em Moreover, the maximum value of the time step $\Delta t$ in the 7L scheme can be set up to $\tau_{max}$ for a Markov process. } With a time step $\Delta t$, we solve the SDE in an iterative way until the actual terminal time $T$, which can be much larger than the training time horizon $\tau_{max}$. The detailed error analysis can be found in Section~\ref{sec:error-analysis}. \subsection{The Artificial Neural Network} \label{sec:ann} The ANN to learn the conditional collocation points is detailed in this subsection. Neural networks can be utilized as powerful functions to approximate a nonlinear relationship. In fact, we will employ a rather basic fully-connected neural network configuration for our learning task. A fully connected neural network, without skip connections, can be described as a composition function, i.e., \begin{equation} \label{eq:fun-dnn} \hat H(\Acute{\mathbf{x}}|\boldsymbol{\tilde{\Theta}}) = h^{(L)}(\ldots h^{(2)}(h^{(1)}(\Acute{\mathbf{x}};\boldsymbol{\tilde{\theta}}_{1});\boldsymbol{\tilde{\theta}}_{2});\ldots\boldsymbol{\tilde{\theta}}_{L_A}), \end{equation} where $\Acute{\mathbf{x}}$ represents the input variables, $\boldsymbol{\tilde{\Theta}}$ being the hidden parameters (i.e. weights and biases), $L_A$ the number of hidden layers. We can expand the hidden parameters as, \begin{equation} \boldsymbol{\tilde{\Theta}} = (\boldsymbol{\tilde{\theta}}_1,\boldsymbol{\tilde{\theta}}_2,\ldots,\boldsymbol{\tilde{\theta}}_L) = (\mathbf{w}_1, \mathbf{b}_1, \mathbf{w}_2, \mathbf{b}_2,\ldots, \mathbf{w}_L,\mathbf{b}_{L_A}), \end{equation} where $\mathbf{w}_\ell$ and $\mathbf{b}_\ell$ represent the weight matrix and the bias vector, respectively, in the $\ell$-th hidden layer. Each hidden-layer function, $h^{(\ell)}(\cdot), \ell=1, 2, \ldots, L_A$, takes input signals from the output of a previous layer, computes an inner product of weights and inputs, and adds a bias. It sends the resulting value in an activation function to generate the output. Let $z_{j}^{(\ell)}$ denote the output of the $j$-th neuron in the $\ell$-th layer. Then, \begin{equation}\label{eq:zz} z_{j}^{(\ell)}=\varphi^{(\ell)} \left(\sum_{i}w_{i,j}^{(\ell)}z^{(\ell-1)}_{i}+b_{j}^{(\ell)} \right), \end{equation} where $w_{i,j}^{(\ell)} \in \mathbf{w}_\ell$, $b_j^{(\ell)} \in \mathbf{b}_\ell$, and $\varphi^{(\ell)}$ is a nonlinear transfer function (i.e. activation function). With a specific configuration, including the architecture, the hidden parameters, activation functions and other specific operations (e.g., drop out), the ANN in~\eqref{eq:fun-dnn} becomes a deterministic, complicated, composite function. Supervised machine learning~\citep{Goodfellow-et-al-2016} is used here to determine the weights and biases, where the ANN should learn the mapping from a given input to a given output, so that for a new input, the corresponding output will be accurately approximated. Such ANN methodology consists of basically two phases. During the (time-consuming, but off-line) training phase the ANN learns the mapping, with many in- and output samples, while in the testing phase, the trained model is used to very rapidly approximate new output values for other parameter sets, in the on-line stage. In a supervised learning context, the loss function measures the distance between the target function and the function implied by the ANN. During the training phase, there are many known data samples available, which are represented by input-output pairs $(\Acute{\mathbf{X}},\Acute{\mathbf{Y}})$. With a user-defined loss function $\mathrm{L}(\boldsymbol{\tilde{\Theta}})$, training neural networks is formulated as \begin{equation} \label{eq:argmin_dnn} \arg \min_{\boldsymbol{\tilde{\Theta}}} \mathrm{L}(\boldsymbol{\tilde{\Theta}} | (\Acute{\mathbf{X}},\Acute{\mathbf{Y}})), \end{equation} where the hidden parameters are estimated to approximate the function of interest in a certain norm. More specifically, in our case, the input, $\Acute{\mathbf{x}}$, equals $\{\hat{Y}_i, t_i, t_{i+1}-t_i, \boldsymbol\theta\}$, and the output, $\Acute{\mathbf{y}}$, represents the collocation points ${\bf{\hat{y}}}_{i+1}$, as in Equation~\eqref{eq:funH}. In the domain of interest $\Acute{\Omega}$, we have a collection of data points $\{\Acute{\mathbf{x}}_k\}$, $k=1,\ldots,M_D$, and their corresponding collocation points $\{ \Acute{\mathbf{y}}_k\}$, which form a vector of input-output pairs $(\Acute{\mathbf{X}},\Acute{\mathbf{Y}})=\{(\Acute{\mathbf{x}}_k,\Acute{\mathbf{y}}_k)\}_{k=1,\ldots,M_D}$. For example, using the $L2$-norm, the discrete form of the loss function reads, \begin{equation} \mathrm{L}(\boldsymbol{\tilde{\Theta}} | (\Acute{\mathbf{X}},\Acute{\mathbf{Y}})) = \sqrt{\frac{1}{M_D} \sum_{k=1}^{M_D} \left(\Acute{\mathbf{y}}_k -\hat{H}(\Acute{\mathbf{x}}_k|\boldsymbol{\hat{\Theta}})\right)^2} . \end{equation} A popular approach for training ANNs is to optimize the hidden parameters via back-propagation, for instance, using stochastic gradient descent~\citep{Goodfellow-et-al-2016}. \section{ An efficient large time step scheme: Compression-Decompression Variant} \label{sec:cdc} The 7L scheme employs the ANNs to generate the conditional collocation points for all samples of a previous time point, see Figure~\ref{fig:paths-7L-fig}. The extensive use of ANNs in the methodology has an impact on the method's computational complexity. In order to speed up the data-driven 7L scheme procedure, we introduce a compression-decompression (CDC) variant, {\em in the on-line validation phase.} Please note that the off-line learning phase is identical for both variants. The so-called 7L-CDC scheme, to be developed in this section, only uses the ANNs to determine the conditional collocation points for the optimal collocation points of a previous time point. All other samples will be computed by means of accurate interpolation. The computational complexity is reduced when the chosen interpolation is computationally cheaper than using ANNs. By the compression-decompression procedure, Monte Carlo sample paths based on SDEs can be recovered from a 3D matrix. We then employ the 7L scheme procedure only to compute the entries of the encoded matrices $C_i$ at time point $t_i$, which leads to a reduction of the computational cost in many cases. Next, we will explain the process of recovering the sample paths from a known matrix $C$ using the decompression method. \subsection{CDC Variant} With a time discretization $\{t_0, t_1,t_2,\dots,t_N\}$, we define a 3D matrix $\hat{C} = \{\hat{C}_0,\hat{C}_1,\dots,\hat{C}_{N-1} \}$, which consists of $N\times(M_s+1)\times(M_c+2)$ entries in total. Recall that $M_s$ represents the number of collocation points and $M_c$ the number of conditional collocation points. $M_s$ and $M_c$ may vary with time points $t_i$ (in case of an adaptive scheme, for example), but we use constant values for $M_s$ and $M_c$. For each time point $t_i$, we construct a 2D matrix $\hat{C}_i$, \begin{eqnarray}\label{eq:bigC} \hat{C}_i = \left( \begin{array}{cccccc} -& -& \hat{x}_1&\hat{x}_2 &\dots&\hat{x}_{M_c}\\ {x}_1& \tilde{y}_1(t_i)&\hat{y}_{1|1}(t_i)&\hat{y}_{2|1}(t_i)&\dots&\hat{y}_{M_c|1}(t_i)\\ {x}_2& \tilde{y}_2(t_i)&\hat{y}_{1|2}(t_i)&\hat{y}_{2|2}(t_i)&\dots&\hat{y}_{M_c|2}(t_i)\\ \vdots&\vdots&\vdots&\vdots&\vdots&\vdots\\ {x}_{M_s}&\tilde{y}_{M_s}(t_i)&\hat{y}_{1|M_s}(t_i)&\hat{y}_{2|M_s}(t_i)&\dots&\hat{y}_{M_c|M_s}(t_i)\\ \end{array} \right)_{(M_s+1)\times(M_c+2)}, \end{eqnarray} with $x_i$, $i=1,\ldots, M_s$, the original SC points, $\hat{x}_k$, $k=1,\ldots,M_c$, the $k$-th original conditional SC points, and the conditional SC points $\hat{y}_{k|j}(t_i) = \hat{y}_{k}(t_{i+1})|\hat{y}_{j}(t_i)$. We thus represent the conditional SC points, $\hat{y}_{k}(t_{i+1})|\hat{y}_j(t_i)$, by matrix elements $c_{i,j,k}$. The two empty cells in~\eqref{eq:bigC} are not addressed in the computation. Moreover, at the last time point, $t_N$, $\hat{C}_{N}$ is not needed. \begin{figure}[!h] \centering \subfloat[Marginal and Cond. SC points]{ \includegraphics[width=0.45\textwidth]{Figure_2_a}} \subfloat[Cond. SC points and densities]{\includegraphics[width=0.45\textwidth]{Figure_2_b}} \caption{Schematic illustration of matrix $C$, with five marginal SC points and five conditional SC points. The conditional SC points are dependent on the realization connected to the corresponding marginal SC point. } \label{fig:collPoints3D} \end{figure} \begin{remark}[Time-dependent elements] As the original collocation points, ${x}_i$ and $\hat{x}_{k}$, do not depend on time, we can remove the first row and the first column of matrix $\hat{C}_i$ to obtain a time-dependent version, $C=\{C_0,C_1,\dots,C_{N-1}\}$, with the following elements, \begin{eqnarray}\label{eq:smallC} C_i=\left( \begin{array}{ccccc} \tilde{y}_1(t_i)&\hat{y}_{1|1}(t_i)&\hat{y}_{2|1}(t_i)&\dots&\hat{y}_{M_c|1}(t_i)\\ \tilde{y}_2(t_i)&\hat{y}_{1|2}(t_i)&\hat{y}_{2|2}(t_i)&\dots&\hat{y}_{M_c|2}(t_i)\\ \vdots&\vdots&\vdots&\vdots&\vdots\\ \tilde{y}_{M_s}(t_i)&\hat{y}_{1|M_s}(t_i)&\hat{y}_{2|M_s}(t_i)&\dots&\hat{y}_{M_c|M_s}(t_i)\\ \end{array} \right)_{M_s\times (M_c+1)}. \end{eqnarray} \end{remark} An entry in matrix $\hat{C}$ can be computed by the trained ANNs, as follows, \begin{equation} \label{eq:cijk} c_{i,j,k}: =\hat{y}_{k|j}(t_i) = \hat{H}_k^{(M_c)}\Big(\hat{y}_{i,j},t_i, t_{i+1}-t_{i}, \boldsymbol\theta\Big), \end{equation} using the marginal SC points, \begin{equation} \label{yij} \tilde{y}_j(t_i) = \hat{H}_j^{(M_s)} \Big(Y_0,t_0, t_i-t_0, \boldsymbol\theta\Big), \end{equation} where $\hat{H}_j^{(\Lambda)} (\cdot), j=1,\dots,m$, $\Lambda=\{M_s,M_c\}$, represents the ANN function which approximates the $j$-th collocation point when $\Lambda=M_s$, and the $j$-th conditional collocation point when $\Lambda=M_c$. When $M_c=M_s$, $\hat{H}_j^{(M_s)}=\hat{H}_j^{(M_c)}$. Figure~\ref{fig:collPoints3D} shows an example of the distribution of the conditional SC points when $M_c=3$ and $M_s=3$. When the matrices have been defined, all sample paths are compressed into a structured matrix. In other words, matrix $\hat{C}$ contains all the information needed to perform the Monte Carlo simulation of the SDEs, apart from the interpolation technique. The resulting matrix $C$ will be {\em decompressed} to generate Monte Carlo sample paths with the help of an interpolation. The process of decompression is straightforward given a matrix $\hat{C}$. In addition to the interpolation process $g_{M_c}(\cdot)$ in SCMC (see Equation~\eqref{eq:gm}), an interpolation $\tilde{g}(\cdot)$ is needed to compute conditional collocation points for previous realizations, based on the matrix $\hat{C}$. \begin{figure}[h!] \subfloat[Paths for marginal SC points]{\includegraphics[width=0.45\textwidth]{Figure_3_a.pdf}} \subfloat[Sample paths by 7L-CDC]{\label{fig:paths-7L-cdc}{\includegraphics[width=0.55\textwidth]{7L_cdc_scheme}}} \caption{ Schematic diagram of the 7L-CDC scheme at time $t_i$. Left: Marginal SC points, corresponding to Equation~\eqref{yij}. Right: Sample paths generated by 7L-CDC. The triple \{2,1,3\}, in the picture, represents the third conditional SC point, dependent on the first marginal SC point at time point $t_2$. The above procedure is also applicable to other time points.} \label{fig:scheme-ann-cdc} \end{figure} Suppose a vector of samples $\hat{\bf{Y}}_i$ at time $t_i$, and we wish to generate samples of $\hat{\bf{Y}}_{i+1}$. For a specific sample $\hat{Y}^*_i$, we need to calculate $M_c$ conditional SC points. To obtain the $k$-th ($1\leq k\leq M_c$) conditional SC point, we take marginal collocation points and their $k$-th conditional collocation points to form $M_s$ pairs $\{ ( \tilde{y}_1 (t_i),\hat{y}_{k|1} (t_i) ), (\tilde{y}_2 (t_i),\hat{y}_{k|2} (t_i) ),\dots,(\tilde{y}_{M_s},\hat{y}_{k|M_s} (t_i)) \}$. This combination gives us the interpolation function $\hat{\bf y}=\hat g(\hat {\bf x})$. Then we can obtain the $k$-th conditional SC point of $\hat{Y}^*_i$, \begin{equation} \label{eq:hatg} \hat{y}^*_k(t_{i+1})|\hat{Y}^*_i = \hat g(\hat{Y}^*_i). \end{equation} As a result, for each sample $\hat{Y}^*_i$, we obtain $M_c$ interpolation nodes, that form a set of pairs, $\big(\hat{x}_1,\hat{y}^*_1(t_{i+1})|\hat{Y}^*_i \big)$,$\big(\hat{x}_2,\hat{y}^*_2(t_{i+1})|\hat{Y}^*_i \big)$, up to $\big(\hat{x}_k,\hat{y}^*_{M_c}(t_{i+1})|\hat{Y}^*_i\big)$, which are used to determine the interpolation function $g_{M_c}(\cdot)$ required by SCMC. Afterwards, to generate a new sample $\hat{Y}^*_{i+1}|\hat{Y}^*_i$, the mapping function $g_{M_c}$ produces a conditional sample by taking in a random sample from $X$, \begin{equation*} \hat{Y}^*_{i+1}|\hat{Y}^*_i= g_{M_c}(\hat{X}_{i+1}). \end{equation*} The choice of the appropriate number of (conditional) collocation points is a trade-off between the computational cost and the required accuracy. When the number of collocation points tends to infinity, the 7L-CDC scheme will resemble the 7L scheme from Section~\ref{sec:CondSCMC}. A schematic picture is presented in Figure~\ref{fig:scheme-ann-cdc}. \begin{remark}[Computation time] During the on-line phase of the method, the total computation time of the large time step schemes consists of essentially two parts, calculation of the conditional SC points, and generating random samples by interpolation (the second part). The difference between the 7L and 7L-CDC schemes is found in the computation of the conditional SC points, the generation of the samples is identical for both schemes. In this first part, for the 7L-CDC scheme, the work consists of setting up matrix $C$ by the ANNs and computing the conditional SC points by the interpolation. In matrix $C$, there are $ M_s \times M_c \times N $ elements that are computed by the ANNs, where $N$ represents the number of time points, $M_s$ the number of collocation points and $M_c$ the number of conditional collocation points. Based on the $M_s$ collocation points, the interpolation is based on $M_c$ conditional collocation points for each path. For the 7L scheme, $M \times M_c \times N$ elements, where $M$ is the total number of paths, are computed by the ANNs. The time ratio between the 7L-CDC and 7L schemes is found as \begin{equation} \label{eq:timeratio} \gamma = \frac{ t_I M + t_AM_s }{t_A M} = \frac{t_I}{t_A} +\frac{M_s}{M}, \end{equation} with $t_A$ the computational time of the ANN (i.e., the function $\hat{H}(\cdot)$), $t_I$ for the interpolation (i.e., the function $\hat g(\cdot)$ in~\eqref{eq:hatg}), which is a polynomial function of $M_s$. Given the fact that the number of sample paths is typically much larger than the number of SC points $M\gg M_s$, $$\gamma \approx \frac{t_I}{t_A}.$$ When the employed interpolation is computationally cheaper than the ANNs, $\gamma<1,$ so that the 7L-CDC scheme needs fewer computations than the 7L scheme. \end{remark} \subsection{Interpolation techniques} \label{sec:interp} To define the function $g_m(x)$ in~\eqref{eq:condY} or $\hat{g}(x)$ in~\eqref{eq:hatg}, we will compare three different interpolation techniques. A bijective mapping function is obtained by the {\em monotonic} Piecewise Cubic Hermite Interpolating Polynomial (PCHIP)~\citep{PCHIP1980}. Assuming there are multiple data points, $(x_k,y_k)$, using, $$ h_k:=x_{k+1}-x_k, \; d_k:= \frac{y_{k+1}-y_k}{x_{k+1}-x_k}, $$ the derivatives $f'_k$ at the points $x_k$ are computed as a weighted average, $$\frac{\hat{w}_1+\hat{w}_2}{f'_k} = \frac{\hat{w}_1}{d_{k-1}} + \frac{\hat{w}_2}{d_k}, \;\; \mbox{ if } d_k \cdot d_{k-1} > 0, $$ where $\hat{w}_1:= 2h_k+h_{k-1}$ and $\hat{w}_2:=h_k +2h_{k-1}$. At each data point the first derivative is guaranteed to be continuous, and a cubic spline is used to interpolate between the data points. If $d_k \cdot d_{k-1} \leq 0$, then $f'_k=0$, PCHIP requires more computations than a Lagrange interpolation, but it results in a monotonic function which is generally advantageous. The convergence of the stochastic collocation method is not really dependent on the monotonicity of the mapping function, so an interpolation based on {\em Lagrange polynomials} is possible in practice. The barycentric version of Lagrange interpolation~\citep{Barycentric2004}, our second interpolation technique, provides a rapid and stable interpolation scheme, which is applied when using Lagrange interpolation in our numerical experiments. With help of the basic Lagrange interpolation expressions, however, we can conveniently perform theoretical analysis. The third technique is based on choosing the interpolation points carefully (e.g., as the Chebyshev zeros) to achieve a stable interpolation. The {\em Chebyshev interpolation}~\citep{rivlin1990chebyshev} is of the form, \begin{equation} g_m(x) =\sum_{j=0}^{m-1}\alpha_j p_{j}(x)= \alpha_0 + \alpha_1p_1(x) + ...+\alpha_{m-1}p_{m-1}(x), \end{equation} where $p_{m-1}(x)$ are interpolation basis functions, here Chebyshev orthogonal polynomials, up to degree $m-1$. The Chebyshev nodes in the interval $[x_a,x_b]$ are computed as, $$\tilde{x}_{k} = x_a + \frac{1+ \cos(\frac{\pi k}{m-1})}{2}(x_b-x_a), k=0,1,\ldots,m-1.$$ When the polynomial degree increases, the Chebyshev interpolation retains uniform convergence. In financial mathematics, Chebyshev interpolation has been successfully used, for example, to compute parametric option prices and implied volatility in~\citep{gass2018chebyshev,glau2019chebyshev,glau2019improved}. When the interpolation points are not Chebyshev nodes (e.g., Gauss quadrature points), the Chebyshev coefficients can be estimated by means of a least squares regression, which is also called the Chebyshev fit. In such case, the coefficients in~\eqref{eq:condsam} can be explicitly computed, in contrast to the barycentric Lagrange interpolation. The selection of a suitable interpolation technique depends on various factors, for instance, speed, monotonicity, availability of coefficients. These three interpolation methods will be compared in the numerical section. \subsection{Path-wise Sensitivity} Often in computations with stochastic variables, we wish to determine the derivatives of the variables of interest, the so-called pathwise sensitivities. This is generally not a trivial exercise in a Monte Carlo setting, see, for example, the discussions in~\citep{fastgreeks,SmokingAF,finbook2019,JAIN201995}. With our new large time step schemes, we determine the pathwise sensitivities of the computed stochastic variables in a natural way, based on the available information in the (conditional) SC points and the interpolation. In this section, we derive the pathwise sensitivity of the state variable $Y(t)$ with respect to model parameters $\theta$. The first derivative with respect to parameter $\theta$ of the conditional distribution in Equation~\eqref{eq:condY} reads, \begin{equation} \label{eq:ywrttheta} \frac{\partial Y(t_{i+1})}{ \partial \theta} = \frac{\partial g(X)}{ \partial \theta} \approx \frac{\partial }{\partial \theta} \left(\sum_{j=1}^m \hat{y}_j(t) p _j(X) \right) = \frac{\partial }{\partial \theta} \left(\sum_{j=1}^m \hat{H}_j p_j(x) \right) = \sum_{j=1}^m \left( \frac{\partial \hat{H}_j}{\partial \theta} p_j(X) \right), \end{equation} where $p_j(X)$ are basis functions, which do not depend on the model parameters. For the derivative $\frac{\partial \hat{H}_j}{\partial \theta}$ in~\eqref{eq:ywrttheta} at time $t_i$, the expression of the ANN~\eqref{eq:fun-dnn}, given the specific activation function, is available. So, the function $\hat{H}$ is analytically differentiable. As a result, $\frac{\partial \hat{H}_j}{\partial \theta}$ can be easily computed, by means of automatic differentiation in the machine learning framework. Thus, we arrive at the sensitivity of a sample path with respect to model parameters, as follows, \begin{equation} \label{eq:ywrtthetapath} \frac{\partial \hat{Y}_{i+1}}{ \partial \theta} = \sum_{j=1}^{m} \frac{\partial \hat{H}_j}{\partial \theta} p_j(\hat{X}_{i+1}). \end{equation} \section{Numerical experiments} \label{sec:results} In this section with numerical experiments we will give evidence of the high quality of our numerical SDE solver, by analyzing first in detail its components. For this purpose, we mainly focus on the Geometric Brownian Motion SDE, which reads, \begin{equation} \label{eq:gbmq} \d Y(t) = \mu Y(t) {\rm d}t+\sigma Y(t) \dW(t), \;\;\;\;\; 0 \leq t \leq T. \end{equation} The model parameters are the constant drift and volatility coefficients, i.e., $\boldsymbol\theta=\{\mu, \sigma\}$, and the initial value is given by $Y_0$. For~\eqref{eq:gbmq} a continuous-time analytic expression for the asset price at time $t$ is available, i.e., \begin{equation} Y(t) = Y_0e^{(\mu-\frac{1}{2} \sigma^2) (t-t_0)+\sigma ( W(t)-W(t_0)) } \stackrel{d}{=} Y_0e^{(\mu-\frac{1}{2} \sigma^2) (t-t_0)+\sigma \sqrt{t-t_0}X}, \label{exa} \end{equation} where $X\sim \mathcal{N}(0,1)$, and $Y(t)$ is governed by the lognormal distribution. The derivative of the stock price with respect to volatility $\sigma$ is available in closed form, and reads, \begin{equation} \label{eq:gbmsense} \frac{\partial Y(t)}{\partial \sigma} \stackrel{d}{=} Y(t)(-\sigma (t-t_0) + \sqrt{t-t_0}X). \end{equation} This expression will be used as the reference value of the sensitivity obtained from the 7L discretization. Furthermore, the Ornstein-Uhlenbeck process is explained and also analyzed, in Subsection~\ref{ouou}. We will employ the large time step discretization, in which the conditional collocation points are computed by the trained ANN, and compare the results of the novel scheme with those obtained by the Milstein SDE discretization. \subsection{ANN Training Details} \label{sec:trainprocess} GBM and the OU process are Markov processes, so the conditional distribution at time $t_{i+1}$ given information up to time $t_i$ only depends on the information at time $t_i$. The ANN~\eqref{eq:markov-Hfun} will therefore be used for the conditional collocation stochastic points, with $\boldsymbol\theta = \{\mu, \sigma\}$, for GBM, and $\boldsymbol\theta = \{\overline{Y},\sigma,\lambda\}$ for the OU process (as will be discussed in Subsection~\ref{ouou}). Regarding the size of the compression-decompression matrix, the more conditional collocation points, the better the accuracy of the 7L-CDC method. A 5x5 matrix size (i.e., five marginal and five conditional SC points) is preferred, taking into account the computing effort and the accuracy. In~\cite{scmc2019} it has been discussed and shown that highly accurate approximations could already be obtained with a small number of collocation points. As the first method component, we evaluate the quality of the ANN which defines the collocation points, for the GBM dynamics. For this purpose, $M_L$ random points (i.e. sets of input parameters) are generated by using Latin Hyper-cube Sampling (LHS) in the domain of interest for the three parameters $(Y_0,\mu, \sigma)$, see Table~\ref{table:ann_train_data}. As the second step, for each point a Monte Carlo method is employed to simulate the discretized SDE based on the tiny time step $\Delta \tau$. We use an Euler-Maruyama time discretization for this purpose, with $N_\tau$ the number of time points and the time horizon $\tau_{max} = N_\tau \cdot \Delta \tau$. At each time step, $j=1,...,N_\tau$, the conditional distribution function $F_{Y(t_j)|Y_0, \mu,\sigma}(\cdot)$ is computed, based on the many generated MC paths. This way, the resulting collocation points for the ``big time step'', $\Delta t = j\cdot\Delta \tau$, are also obtained, to form the required training data set. We set $\tau_{max}=1.6$, $N_\tau=160$, $M_L=500$. The amount of training data used is given by $M_{train} = M_L \cdot N_\tau=80,000$ samples in total, which are divided into an ANN training ($90\%$) and an ANN testing ($10\%$) set. \begin{table}[!h] \begin{center} \caption{ Training data, $\Delta \tau = 0.01$. Here is an example for training on five SC points. } \begin{tabular}{ c|c | c | c } \hline ANN & Parameters & Value range & Method \\ \hline \multirow{3}{*}{input} & drift, $\mu$ & (0.0, 0.10] & LHS \\ &volatility, $\sigma$ & [0.05, 0.60] & LHS \\ &value, $Y_0$ & [0.10, 15.0] & LHS \\ &time, $\tau_{max}$ & (0.0, 1.60] & Equidistant\\ \hline \multirow{1}{*}{$\hat{H}_1(\cdot)$ output} & point, $\hat{y}_1$ & $(0.0, 25.65)$ & SCMC\\ \multirow{1}{*}{$\hat{H}_2(\cdot)$ output} & point, $\hat{y}_2$ & $(0.0, 25.98)$ & SCMC\\ \multirow{1}{*}{$\hat{H}_3(\cdot)$ output} & point, $\hat{y}_3$ & $(0.0, 27.84)$ & SCMC\\ \multirow{1}{*}{$\hat{H}_4(\cdot)$ output} & point, $\hat{y}_4$ & $(0.0, 54.67)$ & SCMC\\ \multirow{1}{*}{$\hat{H}_5(\cdot)$ output} & point, $\hat{y}_5$ & $(0.0, 154.35)$ & SCMC\\ \hline \end{tabular} \label{table:ann_train_data} \end{center} \end{table} The ANN hyper-parameters have an impact on the errors from optimization related to training the ANN, as well as on the model performance. The approximation capacity does not only depend on the number of hidden parameters, but also on the network structure (i.e., on the width and depth of the network). In principle, deep neural networks have more powerful expressiveness than shallow neural networks. The fully connected neural network employed will be composed of one input layer, one output layer and four hidden layers. Each hidden layer consists of 50 neurons, with Softplus, i.e. $\varphi(x)= \ln(1+e^x)$ as the activation function~\citep{activation2018}. Before training the ANN, the hidden parameters are initialized via the Glorot technique~\citep{glorot2010}. Training goes in batches. At each iteration, the stochastic gradient based optimizer, Adam~\citep{Adam2014}, randomly selects a portion of the training samples according to the batch size, to calculate the gradient for updating the hidden parameters. In an epoch, all training samples have been processed by the optimizer. The mean squared error, which measures the distance between the ground-truth and the model values in supervised learning, is used to update the hidden parameters during training. The measure MAE (Mean Absolute Error), i.e., $\text{MAE} = \frac{1}{M_{train}} \sum_j |{y}_j-\hat{y}_j|,$ is also estimated, as the path-wise error of the 7L scheme is related to the maximum absolute difference in the approximated collocation points, $\hat{y}_j$, in Section~\ref{sec:pathwise error}, see the derivation in the next section. The training process starts with a relatively large learning rate (i.e $10^{-3}$) to avoid getting stuck in local optima. After 1000 epochs, the learning rate is reduced to $10^{-4}$, followed by training 500 more epochs, to achieve a steady convergence. Afterwards, the trained ANN is evaluated on the testing data set, with the results presented in Figure~\ref{figure:r2-ANN-performance} (for two of the collocation points) and Table~\ref{table:measures_ann}. Clearly, the predicted values fit very well with the true values of the stochastic collocation points. This implies that the trained ANNs reach a highly satisfactory generalization, and generate accurate and robust approximation results for all five collocation points. \begin{figure}[!h] \centering \subfloat[$\hat{y}_2$]{\label{fig:test_r2_y2} {\includegraphics[width=0.45\textwidth]{test_r2_y2.png}}}% \subfloat[$\hat{y}_4$]{\label{fig:test_r2_y4}{\includegraphics[width=0.45\textwidth]{test_r2_y4.png}}} \caption{The goodness of fit on test data set. Two scatter plots show the relation between the predicted values and the ground truth.} \label{figure:r2-ANN-performance} \end{figure} \begin{table}[htp] \begin{center} \caption{The approximation performance on test data set.} \scalebox{1.0}{ \begin{tabular}{ c | c | c | c | c | c} \hline SC points & $\hat{y}_1$ & $\hat{y}_2$ & $\hat{y}_3$ & $\hat{y}_4$ & $\hat{y}_5$ \\ \hline $R^2$ & 0.999891 & 0.999947 & 0.999980 & 0.999892 & 0.999963 \\ MAE & 0.026 & 0.027 & 0.021 & 0.071 & 0.066 \\ \hline \end{tabular} } \label{table:measures_ann} \end{center} \end{table} \subsection{Error analysis, the Lagrangian case} \label{sec:error-analysis} There are essentially two approximation errors in the 7L scheme, a neural network approximation error when generating the collocation points, and an SCMC error when representing the conditional distribution function. Considering $d$ inputs, the neural network may approximate any function $\zeta_{d,n}$, from the function space $C^{n-1}([0,1]^d)$, where the derivatives up to order $n-1$ are Lipschitz continuous~\cite{ErrorReLuDNN2017}. The input and output variables can be normalized to the unit interval $[0,1]$. With a fixed network architecture during training, the approximation error can be assessed, as follows. \begin{theorem} From~\cite{ErrorReLuDNN2017}. Given any $\hat{\epsilon} \in (0,1)$, there exists a neural network which is capable of approximating any function $\zeta_{d,n}$ with error $\hat{\epsilon}$, based on the following configuration: \begin{itemize} \item at least piece-wise activation functions, \item at least $\tilde{c}(\ln(1/\hat{\epsilon} )+1 )$ hidden layers and $\tilde{c}\hat{\epsilon}^{-d/n}(\ln(1/\hat{\epsilon}) +1 )$ weights and computation units, where $\tilde{c}:=\tilde{c}(d,n)$ depends on the parameters $d$ and $n$. \end{itemize} \end{theorem} When the architecture is dynamic, the error bound can be further reduced, as shown in~\cite{NewErrorBoundsDNN2019} and~\cite{ErrorReLuDNN2017}. One of the assumptions is that the ANNs are sufficiently trained, so that the optimization error is negligible. The error from the SCMC methodology was derived in~\cite{scmc2019}. The optimal collocation points, ${x}_i$, $i=1,\dots,m$, correspond to the zeros of an orthogonal polynomial. In the case of Lagrange interpolation, when the collocation method can be connected to Gauss quadrature, we have \begin{eqnarray} \int_\R \Psi(x)f_X(x) {\rm d}x=\sum_{i=1}^m \Psi({x}_i) \omega_i+ \epsilon_m = \epsilon_m, \end{eqnarray} with $\Psi(x)=\left(g(x) - g_m(x)\right)^2$, the difference between the target and the SC approximated function, $f_X(x)$ the weight function, and $\omega_i$ the quadrature weights. When the Gauss-Hermite quadrature is used with $m$ collocation points. the approximation error of the CDF can be estimated as, \begin{eqnarray}\label{epsilon_scmc} \epsilon_m=\frac{m!\sqrt{\pi}}{2^m}\frac{\Psi^{(2m)}(\xi_1)}{(2m)!}, \end{eqnarray} where $ \xi_1 \in (-\infty, \infty)$ and the distance function $$\Psi(x)=\left(g(x) - g_m(x)\right)^2\approx \left(\frac{1}{m!}\left.\frac{\partial^m g(x)}{\partial x^m}\right|_{x=\xi_2}\prod_{k=1}^m (x-{x}_k)\right)^2,$$ with $\xi_2 \in [{x}_1,{x}_{m-1}]$. In other words, the error of approximating the target CDF converges exponentially to zero when the number of corresponding collocation points increases. At each time point $t_i$, the process $Y(t_i)$ is approximated using the collocation method, by a polynomial $g_m(X)$, i.e., in the case of classical Lagrange interpolation, using $\ell_j(\bar x) = p_j(\bar x)$, \begin{equation} Y(t_i) \approx \tilde{Y}(t_i)=g_m(X)=\sum_{j=1}^m {y}_j(t_i)\ell _j(X), \;\; \ell_j(\bar{x})=\prod_{k=1}^m\frac{X-{x}_k}{{x}_j-{x}_k}, \end{equation} where the collocation points ${y}_j(t_i)=F^{-1}_{Y(t_i)}(F_X({x}_j))$. Because of the use of an ANN, the collocation points are not exact, but they are approximated with ${y}_j(t_i)-\hat{y}_j(t_i)=\epsilon_j^{A}$, where $\hat{y}_j(t_i)$ represents the ANN approximated value. The error associated with $\epsilon_j^{A}$ can be estimated as in~\citep{NewErrorBoundsDNN2019}. The impact of $\epsilon_j^A$ on the obtained output distribution needs to be assessed. Let $\tilde{g}_m$ denote the approximate function based on the predicted ANN collocation points $\hat{y}_j(t_i)$, and $x$ a random sample from the standard normal distribution $X$. The approximation error, in the strong sense, is given by \begin{eqnarray} \E\left[ |g_m(x)-\tilde{g}_m(x) |\right]&=&\E\Big|\sum_{j=1}^m{y}_j(t_i)\ell _j(x)-\sum_{j=1}^m \hat{y}_j (t_i)\ell _j(x)\Big|\nonumber \\ &=&\int_\R\Big|\sum_{j=1}^m {y}_j(t_i)\ell _j(x)-\sum_{j=1}^m\hat{y}_j(t_i)\ell _j(x)\Big|f_X(x){\rm d}x\nonumber \\ &=&\int_\R\Big|\sum_{j=1}^m\epsilon_j^A\ell _j(x)\Big|f_X(x){\rm d}x. \end{eqnarray} Note that the $\ell_j(x)$ interpolation functions are identical as they depend solely on the $x$ values. We arrive at the following error related to the ANNs, \begin{eqnarray} \int_\R\Big|\sum_{j=1}^m\epsilon_j^A\ell _j(x)\Big|f_X(x){\rm d}x &\leq&\int_\R \sum_{j=1}^m\max\{|\epsilon_1^A|,\dots,|\epsilon_m^A|\}\ell _j(x)f_X(x){\rm d}x\nonumber \\ &=&\int_\R \max\{|\epsilon_1^A|,\dots,|\epsilon_m^A|\}f_X(x){\rm d}x\nonumber \\ &=&\max\{|\epsilon_1^A|,\dots,|\epsilon_m^A|\} \end{eqnarray} Considering the error introduced by SCMC in~\eqref{epsilon_scmc}, the total path wise error reads \begin{eqnarray} \label{eq:pwerror} \E\left[ |g(x)-\tilde{g}_m(x) |\right] &\leq& \E\left[ |g(x)-g_m(x) |\right] + \E\left[ |g_m(x)-\tilde{g}_m(x) |\right] \nonumber \\ &\leq& \sqrt{|\epsilon_m|}+ \max\{|\epsilon_1^A|,\dots,|\epsilon_m^A|\}. \end{eqnarray} In other words, the expected pathwise error can be bounded by the approximation CDF error $\sqrt{|\epsilon_m|}$ plus the largest difference in the ANN approximated collocation points. \subsubsection{Kolmogorov-Smirnov Test} The Kolmogorov-Smirnov test, calculating the supremum of a set of distances, is used to measure the nonparametric distance between two empirical cumulative distribution functions. We perform the two-sample Kolmogorov-Smirnov test, as follows, $$KS = \sup_z|F_Y(z) -\hat F_Y(z)|,$$ where $\hat F_Y(\cdot)$ and $F_Y(\cdot)$ are two empirical cumulative distribution functions, one from the 7L-CDC solution and the other one from the reference distribution. We take the analytic solution of the GBM as the reference distribution. \begin{remark}[Time horizon for 7L-CDC ] The information in Table~\ref{table:ann_train_data} is used to train the mapping function between a realization (including marginal SC points) and its conditional SC points, via Equation~\eqref{eq:cijk}. For the marginal SC points in Equation~\eqref{yij}, however, we need training data up to terminal time $T$. So, we generate a second data set in which the time reaches $\tau_{max}$ (the terminal time of interest) and the upper value for $Y_0$ equals 5. These two data sets are merged into one set in order to train the ANNs for the 7L-CDC methodology. \end{remark} Figure~\ref{fig:ks-test} shows the Kolmogorov-Smirnov test at different time points based on 10000 samples. We focus on the CDC methodology here, and compare the accuracy with the different interpolation methods in the figure. Clearly, the KS statistic and also the corresponding $P$-values for the 7L-CDC schemes are much better than those of the Milstein scheme in Figure~\ref{fig:ks-test}. This is an indication that the CDFs that originate from the 7L-CDC schemes resemble the target CDF much better, with high confidence. In addition, unlike the Milstein scheme the 7L-CDC schemes exhibit an almost constant difference between the approximated and target CDFs with increasing time. \begin{figure}[!h] \centering \subfloat[KS statistic]{\label{fig:cdc_kstest_stat} {\includegraphics[width=0.5\textwidth]{7Lcdc_kstest_stat}}}% \subfloat[$P$-value]{\label{fig:cdc_kstest_pval}{\includegraphics[width=0.5\textwidth]{7Lcdc_kstest_pval}}}% \caption{ The Kolmogorov-Smirnov test: $\Delta t =0.5, \mu=0.1, \sigma=0.3, Y_0=1.0$, with 10000 samples. When we have a small KS statistic or a large $P$-value, the hypothesis that the distributions of the two sets of random samples are the same can not be rejected.} \label{fig:ks-test} \end{figure} We will also analyze the costs of the different interpolation methods within 7L-CDC. The two steps which require interpolation are the computation of the conditional collocation points and the generation of conditional samples. The computational speed of the 7L-CDC scheme depends on the employed interpolation method, see Table~\ref{table:computing_time_interp}. In general, to generate a solution with the same strong order in the numerical error, the Milstein scheme will require more computation time, here about 27 seconds, while the 7L scheme needs 13 seconds and 7L-CDC (Barycentric version) 5 seconds when $\Delta t=1.0$. The larger the time step, the more computation time will be saved. \begin{table}[!h] \begin{center} \caption{ The CPU running time (seconds) to reach the same accuracy (CPU: E3-1240, 3.40GHz): simulating 10,000 sample paths until terminal time $T=4.0$, based on $5\times5$ marginal/conditional SC points. Here, for the 7L scheme, PCHIP is used as the interpolant $g_m(\cdot)$ in Step 3 of Algorithm~I. } \begin{tabular}{ c|c c c | c c c} \hline \multirow{2}{*}{Method \slash Time (Sec.)} & \multicolumn{3}{c|}{$\Delta t=1.0$} & \multicolumn{3}{c}{$\Delta t=2.0$} \\ & Create $C$ & Decom. $C$ & Total & Create $C$ & Decom. $C$ & Total \\ \hline 7L-CDC Barycentric & 0.054 & 4.93 & 4.98 & 0.027 & 2.48 & 2.51 \\ 7L-CDC Chebyshev & 0.054 & 9.78 & 9.83 & 0.027 & 4.93 & 4.96 \\ 7L-CDC PCHIP & 0.054 & 11.39 & 11.44 & 0.027 & 5.73 & 5.76 \\ \hline 7L scheme & - & -& 12.80 & - & -& 6.39\\ Milstein & - & - & 27.01 & - & - & 27.70\\ \hline \end{tabular} \label{table:computing_time_interp} \end{center} \end{table} Note that, in order to achieve a similar accuracy in the strong sense, the Euler-Maruyama scheme requires a much finer time grid, by a factor of $\kappa=\Delta t/\Delta \tau$, than the 7L scheme. When $\kappa$ is sufficiently large, the 7L-CDC scheme outperforms the Euler-Maruyama scheme, in terms of both accuracy and speed. For example, in Table~\ref{table:computing_time_interp}, $\kappa=100$ when $\Delta t =1.0$, and $\kappa=200$ when $\Delta t =2.0$. In other words, the ``on-line version'' of the Euler-Maruyama discretization is computationally slower than the on-line phase of the 7L scheme to achieve the same accuracy. Additionally, computational time of the 7L scheme can be further reduced by parallelization, for example, using GPUs . \subsection{Path-wise Error Convergence} \label{sec:pathwise error} In this section, we compare the path-wise errors of our proposed novel discretization with those of the classical discretization schemes. \subsubsection{GBM process} We analyze here the strong convergence properties of the new methodology for the GBM process. For GBM, the exact path is given by the expression~\eqref{exa}. The random number, which is drawn from $X\sim N(0,1)$, is the same for the exact solution ~\eqref{exa}, the novel schemes~\eqref{eq:condsam} and the Milstein scheme~\eqref{eq:milstein}. The path-wise differences between the numerical schemes and the exact simulation are plotted in Figure~\ref{fig:paths_gbm_7Lcdc}. When $\Delta t =0.5$, the 7L-CDC scheme presents superior paths as compared to the Milstein scheme, in terms of its path-wise error comparing to the exact path. As shown in Figure~\ref{figure:convergence_error}, the 7L-CDC scheme gives rise to flat, almost constant, strong and weak error convergence curves for many different $\Delta t$-values, suggesting a small, constant convergence error even with large time steps $\Delta t$. The Milstein scheme has the strong order of convergence $O(\Delta t)$, so that a larger time step gives rise to a larger error. When the time step becomes small, more time points are needed to reach a time $T$, and then the resulting recursive error of the 7L-CDC scheme increases. The number of conditional collocation points, by which the conditional distribution at a next time point is mostly determined, has a significant contribution to the convergence order of the 7L-CDC scheme. As mentioned, we found empirically that five conditional collocation points are preferable in terms of computing effort versus accuracy. CDC matrix $C$ is then of size $N\times5\times5$, that is, at each time point, there are five collocation points and each of these has five conditional collocation points. \begin{figure}[!h] \centering \includegraphics[width=0.6\textwidth]{paths_gbm_7Lcdc} \caption{Paths generated by 7L-CDC: time step $\Delta t=0.5$, GBM with $\sigma=0.3$, $r=0.1$, $Y_0 =1.0$. The paths are with Chebyshev interpolation, which are not plotted, are identical to ones from Lagrange in this case. } \label{fig:paths_gbm_7Lcdc} \end{figure} \begin{figure}[!h] \centering \subfloat[Strong convergence]{\label{fig:cdc_strong_converg}{\includegraphics[width=0.5\textwidth]{7Lcdc_strong_converg}}}% \subfloat[Weak convergence]{\label{fig:cdc_weak_converg} {\includegraphics[width=0.5\textwidth]{7Lcdc_weak_converg}}}% \caption{The strong error is estimated as $\frac{1}{M}\sum |\check{Y}_k(T)-\hat{Y}_k(T)|$, see Equation~\eqref{sstr} and the weak error by $\frac{1}{M}(\sum \check{Y}_k(T)-\sum\hat{Y}_k(T))$, see~\cite[][page 261]{finbook2019} for details on the computation of the convergence rate. There are $M=1000$ sample paths in total.} \label{figure:convergence_error} \end{figure} \subsubsection{Ornstein-Uhlenbeck process} \label{ouou} Any SDE which can be solved by the Euler-Maruyama discretization can be solved by our ANN methodology, with improved strong convergence properties. We also wish to confirm the strong convergence properties for another stochastic process in this section. The mean reverting Ornstein-Uhlenbeck (OU) process~\citep{OUP1930} is defined as, \begin{equation} \label{eq:oup} \d Y(t) = -\lambda (Y(t)-\overline{Y}) {\rm d}t + \sigma \dW(t), \;\;\;\;\; 0 \leq t \leq T, \end{equation} with $\overline{Y}$ the long term mean of $Y(t)$, $\lambda$ the speed of mean reversion, and $\sigma$ the volatility. The initial value is $Y_0$, and the model parameters are $\boldsymbol\theta :=\{\overline{Y},\sigma,\lambda\}$. Its analytical solution is given by, \begin{equation} \label{eq:oup-solu} Y(t) \stackrel{d}{=} Y_0 e^{-\lambda t} + \overline{Y}(1-e^{-\lambda t}) +\sigma \sqrt{ \frac{1-e^{-2\lambda t}}{2\lambda}} X, \end{equation} with $t_0=0$, $X \sim \mathcal{N}(0,1)$. Equation~\eqref{eq:oup-solu} is used to compute the reference value to the path-wise error and the strong convergence. We employ the same data-driven procedure as for GBM to discretize and solve the OU process. In the training phase, the Euler-Maruyama scheme~\eqref{eq:euler} is used to discretize the OU dynamics and generate the data set. Note that the Milstein and Euler schemes are identical in the case of the OU process. As the OU process is a Markov process, we again can vary $Y_0$ to find the relation between the conditional SC points and the marginal SC points (i.e. as in Equation~\eqref{eq:cijk}). Similar to Table~\ref{table:ann_train_data}, we employ five SC points to learn within the ANN, with $\Delta \tau =0.01$, $\tau_{max}=4.1$, $N_\tau=500$, $M_L=410$, see Section~\ref{sec:trainprocess} for the details of the training process. After the training, the obtained ANNs will be applied to solve the OU process with specific parameters and details of our interest. We provide an example in Figure~\ref{fig:oup-test}, which confirms that the sample paths generated by 7L-CDC are as accurate as the exact solution, and the error, in the sense of strong convergence, stays close to zero even with a large time step. \begin{figure}[!h] \centering \subfloat[Path-wise error ($\Delta t=1.0$)]{\label{fig:paths_OUP_cdc} {\includegraphics[width=0.5\textwidth]{paths_OUP_7Lcdc}}} \subfloat[Strong convergence]{\label{fig:anncdc_strong_converg_oup}{\includegraphics[width=0.5\textwidth]{7Lcdc_strong_converg_oup}}} \caption{Paths and strong convergence for the OU process, using $\lambda=0.5$, $ \overline{Y}=1.0$, $\sigma=0.3$, $Y_0=1.0$. The sample paths with barycentric, Chebyshev and PCHIP interpolation overlap for the 7L-CDC scheme. There are five marginal and five conditional SC points at each time point.} \label{fig:oup-test} \end{figure} \subsection{Applications in Finance} The possibility to take large time steps and still get accurate SDE solutions, is certainly interesting in computational finance, as there are several financial products that are updated on a daily basis (think of an over-night interest rate), whereas monitoring of financial contracts and risk management monitoring is typically only done on a weekly, monthly of even yearly basis. In such situations, our novel scheme will be useful. Research into large time step simulations is state-of-the-art in computational finance, see the exact (and almost exact) Monte Carlo simulation papers, like~\cite{KayaBroadie2006,LEITAO2017461} for the SABR and Heston stochastic volatility asset dynamics, respectively. \subsubsection{The Asian option} Moreover, the strong convergence property of an SDE discretization is important in many cases. When valuing so-called path-dependent options, for example, improved strong convergence enhances the convergence of a Monte Carlo simulation. Options are governed by their pay-off function (i.e. the option value at the final time of the contract, $t=T$). Here we consider a path-dependent exotic option, the so-called European-style Asian option, which has a payoff that is based on a time-averaged underlying stock price. For example, the pay-off of a fixed strike Asian option is given by $$ V_A(T) = \max{(A(T)-\tilde{K}, 0)}, $$ where $T$ is the option contract's expiry time, and $\tilde{K}$ is the predetermined strike price. Here $A(T)$ denotes the discrete arithmetic average of the stock prices over $N_b$ monitoring dates $\{t_1,\ldots, t_k\} \in [0,T]$, $$A(T) = \frac{1}{N_b} \sum_{k=1}^{N_b} \hat{Y}(t_k), $$ where $\hat{Y}(t_k)$ is the observed stock price at time $t_k$, $ 1 \leq t_k \leq T$. Averaging thus takes place in the time-wise direction, and we consider pricing financial options based on the discrete arithmetic average of a number of stock prices. We assume here that the underlying stock price follows Geometric Brownian motion, as in Equation~\eqref{eq:gbmq}, under the risk-neutral measure, meaning $\mu\equiv r$, where $r$ is the risk-free interest rate. There is a cash account $M(t)$, governed by $\d M(t) = r M(t){\rm d}t$. The value of European-style Asian option is then given by \begin{equation} V_A(t) = \e^{- r (T-t)} \mathbb{E}^{\mathbb{Q}} \biggl[ \max ( A(T)-\tilde{K} , 0) \bigg| \F(t)\biggr]. \end{equation} Because the pay-off is clearly a path-dependent quantity for such options, it is expected that an improved strong convergence, obtained with the variant 7L-CDC, will result in superior convergence, as compared to classical numerical discretization schemes. \begin{table}[!h] \begin{center} \caption{Pricing Asian European-style option with a fixed strike price, using $Y_0=1.0$, $\tilde{K}=Y_0$, $r=0.1$, $T=\Delta t \times N_b$.} \begin{tabular}{ c | c | c | c } \hline & method & $\Delta t =1.0$, $N_b$=4 & $\Delta t=0.5$, $N_b$=8 \\ \hline \multirow{3}{*}{$\sigma$=0.30} & Analytic MC & 0.24886257 (0.00\%) & 0.22403982 (0.00\%)\\ & Milstein MC & 0.23077000 (7.27\%) & 0.21558276 (3.77\%) \\ & 7L-CDC & 0.24871446 (0.06\%) & 0.22404571 (0.00\%)\\ \hline \multirow{3}{*}{$\sigma$=0.40} & Analytic MC & 0.28515109 (0.00\%) & 0.25723594 (0.00\%) \\ & Milstein MC & 0.26394277 (7.44\%) & 0.24717425 (3.91\%) \\ & 7L-CDC & 0.28482371 (0.11\%) & 0.25647592 (0.30\%)\\ \hline \end{tabular} \label{table:perfomance_asian_pirce} \end{center} \end{table} The relative error is presented, which is defined as $$\epsilon_{rel} = \biggl|\frac{V_A^{ref}(t_0)-V_A(t_0)}{V^{ref}_A(t_0)}\biggr|,$$ where $V^{ref}_A(t_0)$ is based on the exact GBM Monte Carlo simulation. As shown in Table~\ref{table:perfomance_asian_pirce}, the 7L-CDC scheme gives highly accurate Asian option prices, compared to the Milstein scheme. As the accuracy of Asian option prices depends directly on the accuracy of the realized paths, an increasing number of monitoring dates will give rise to higher accuracy by 7L-CDC. Next, we focus on the Asian option's sensitivity. The sensitivity of the option price with respect to volatility $\sigma$ is called vega, which can be computed in a pathwise fashion (see Chapter 7 in~\cite{glasserman2004monte}), as follows, \begin{equation} \label{eq:Vwrtvol} \frac{\partial V}{ \partial \sigma}=e^{-rT}\mathbb{E}^{\mathbb{Q}} \left[ \sum_{i=1}^N \frac{\partial V(T,Y(t_i);\sigma)}{\partial Y(t_i)} \frac{\partial Y(t_i)}{\partial \sigma} \big| Y_0 \right]. \end{equation} The chain rule is employed to derive the sensitivity. First of all, we compute the gradient of the payoff function with respect to the underlying stock price, by \begin{equation} \frac{\partial V(T,Y(t_i))}{ \partial Y(t_i)} = \frac{1}{N} \mathbbm{1}_{A(T)>\tilde{K}}. \end{equation} Then, the derivative of the stock price at time $t_i$ with respect to the model parameter, $\frac{\partial Y(t_i)}{\partial \sigma}$, can be found with the trained ANNs, as given by Equation~\eqref{eq:ywrttheta}. Vega can be estimated by, \begin{equation} \label{eq:Vwrtvolann} \frac{\partial V}{ \partial \sigma} \approx e^{-rT}\frac{1}{N}\mathbb{E}^{\mathbb{Q}} \left[ \sum_{i=1}^N \left( \mathbbm{1}_{A(T)>\tilde{K}} \sum_{j=0}^{m-1} \frac{\partial \hat{H}_j}{\partial \sigma} p_j(X) \right) \big| Y_0 \right]. \end{equation} When there are $M$ sample paths, we have, \begin{equation} \label{eq:Vwrtvolpaths} \frac{\partial V}{ \partial \sigma} \approx e^{-rT} \frac{1}{M} \frac{1}{N} \left[ \sum_{k=1}^{M} \sum_{i=1}^N \left( \mathbbm{1}_{A(T)>\tilde{K}} \sum_{j=0}^{m-1} \frac{\partial \hat{H}_j}{\partial \sigma} p_j(\hat{X}_{k,i+1}) \right) \big| Y_0 \right]. \end{equation} As the realization $\hat{Y}_i$ is a function of the model parameters, at time $t_{i+1}$, the derivative with respect to the volatility in Equation~\eqref{eq:ywrtthetapath} becomes \begin{equation} \label{eq:gbmscmcpathwise} \frac{\partial \hat{H}_{j}(\hat{Y_i},\sigma)}{\partial \sigma} = \frac{\partial \hat{H}_{j}(\sigma;\hat{Y_i})}{\partial \sigma} + \frac{\partial \hat{H}_{j}(\hat{Y_i};\sigma)}{\partial \hat{Y_i}} \frac{\partial \hat{Y}_i}{ \partial \sigma}, \end{equation} where $\frac{\partial \hat{Y}_i}{ \partial \sigma}$ is known at the previous time point. Like simulating the Monte Carlo paths, the calculation of this derivative is done iteratively. Figure~\ref{fig:paths_sense_gbm_scmc} compares the path-wise sensitivities obtained via Equations~\eqref{eq:gbmsense} and~\eqref{eq:gbmscmcpathwise}. Clearly, the path-wise derivative by the 7L scheme is very similar to the analytical solution. Figure~\ref{fig:pathwise-vega} confirms that the ANN methodology computes a highly accurate Asian option vega by means of the above path-wise sensitivity. Summarizing, the sensitivity with respect to model parameters can highly accurately be obtained from the trained ANNs. As the 7L-CDC scheme is composed of marginal and conditional collocation points, the above procedure of computing the path-wise sensitivity is also applicable to the variant 7L-CDC, by using the chain rule. \begin{figure}[!h] \centering \subfloat[Path-wise sensitivity]{\label{fig:paths_sense_gbm_scmc}{\includegraphics[width=0.5\textwidth]{paths_sensitivy_gbm_7L}}} \subfloat[Path-wise vega]{\label{fig:pathwise-vega} {\includegraphics[width=0.5\textwidth]{pathswise_vega_AsianOption_7L}}} \caption{ Path-wise estimator of vega: Exact vega is calculated by means of the central finite difference. The parameters are $Y_0=1.0$, $r=0.05$, $\tilde{K}=Y_0$, $\sigma=0.3$, $\Delta t =1.0$, $N_b=4$, $T=\Delta t \times N_b=4.0$.} \label{figure:Path-wise-sensitivy} \end{figure} \subsubsection{Bermudan option valuation} \label{sec:bermudan} When dealing with so-called Bermudan options, the option contract holder has the right (but not the obligation) to exercise the option contract at a finite number of pre-specified dates up to final time $T$. At an exercise date, when the holder decides to exercise the Bermudan option, she immediately obtains the current payoff value of the contract. Alternatively, she may also wait until the next exercise opportunity. The Bermudan option can be exercised at the following set of exercise dates, $\{t_0,t_1,...,t_{N_b}\}$, with a constant time difference, $\Delta t = t_{i}-t_{i-1}$, for any $0<i\leq N_b$. In this experiment, we compare the performance of the new 7L-CDC discretization scheme with a classical scheme. Valuation of the Bermudan option will take place by means of the well-known Longstaff-Schwartz Monte Carlo method (LSMC)~\citep{Longstaff}, a least squares Monte Carlo method. The Longstaff-Schwartz algorithm is presented, for convenience, in the appendix. The difference between a large time step simulation and a classical simulation, like the Milstein scheme, is that a classical scheme requires additional time steps to be taken between the early-exercise dates of the Bermudan option, while with the 7L-CDC scheme, we can perform one-step Monte Carlo simulation without any intermediate grid points between adjacent early-exercise dates. We also assume here that the underlying stock price follows Geometric Brownian motion, as in Equation~\eqref{eq:gbmq}, under the risk-neutral measure, with $\mu\equiv r$. A Bermudan put option, with risk-free interest rate $r=0.1$, pay-off function $V(t_j) = \max{(\tilde{K}-Y(t_j),0)}$ with strike price $\tilde{K}=1.1$ and initial stock price $Y_0=1.0$, is priced based on $M=100,000$ Monte Carlo paths. The matrix size within the 7L-CDC scheme is set to $N_b\times 5\times5$. The terminal time is $T=\Delta t \times M_B$ with a constant time step $\Delta t$. The random seed is chosen to be zero when drawing random numbers. We compare the relative errors $|\frac{V^{ref}(t_0)-V(t_0)}{V^{ref}(t_0)}|$, where $V^{ref}(t_0)$ is computed with the help of a Monte Carlo method based on the exact simulation of GBM~(\ref{exa}). \begin{table}[!h] \begin{center} \caption{Bermudan put option prices based on large time step Monte Carlo simulations.} \begin{tabular}{ c | c | c | c | c } \hline & method & $\Delta t =1.0$, $N_b$=4 & $\Delta t =0.5$, $N_b$=4 & $\Delta t=0.5$, $N_b$=8 \\ \hline \multirow{3}{*}{$\sigma$=0.30} & Analytic MC & 0.15213858(0.00\%) & 0.14620214(0.00\%) & 0.16161876(0.00\%)\\ & Milstein MC & 0.13872771(8.81\%) & 0.14065252(3.80\%) & 0.15429369(4.53\%) \\ & 7L-CDC & 0.15234901(0.14\%) & 0.14648443(0.19\%) & 0.16196264(0.21\%)\\ \hline \multirow{3}{*}{$\sigma$=0.40} & Analytic MC & 0.21459038(0.00\%) & 0.19552454(0.00\%) & 0.22340304(0.00\%) \\ & Milstein MC & 0.19598488(8.67\%) & 0.18790933(3.89\%) & 0.21297732(4.67\%) \\ & 7L-CDC & 0.21474619(0.07\%) & 0.19590733(0.20\%) & 0.22389360(0.22\%)\\ \hline \end{tabular} \label{table:perfomance_bermudan_pirce} \end{center} \end{table} As shown in Table~\ref{table:perfomance_bermudan_pirce}, the option prices based on the 7L-CDC Monte Carlo simulation are highly satisfactory, and the related error does not increase with larger time steps $\Delta t$. In contrast, a larger time step gives rise to significant pricing errors, in the case of the Milstein discretization. \begin{remark} In principle, a sample value $\hat{Y}_i$ can be any rational number. So, a path value may reach a larger stock price than the prescribed upper bound in Table~\ref{table:ann_train_data}. The stock prices outside the training interval are called {\em outliers}. Outliers did not appear in the experiments of Table~\ref{table:perfomance_bermudan_pirce}. As an alternative method to avoid the appearance of outliers, one may scale the asset price, to remove the dependence on the initial value. For example, GBM can be scaled by $ \bar{Y}(t)= \frac{Y(t)}{Y(t_0) \e^{rt}}.$ Using It\^o's lemma, we have a drift-less process, $ \d \bar{Y}(t) = \bar{Y}(t) \sigma \dW,$ where the initial value $\bar{Y}_0 =1.0$. The following formula returns the original variable, $ Y(t)= \bar{Y}(t) Y(t_0) \e^{rt}.$ In such case, scaling guarantees a fixed initial value, for example, $Y_0=1.0$. \end{remark} \section{Conclusions and Outlook} \label{cenou} We develop a data-driven numerical solver for stochastic differential equations, by which large time step simulations can be carried out accurately in the sense of strong convergence. With a combination of artificial neural networks and the stochastic collocation Monte Carlo method, a small number of stochastic collocation points are learned by the ANN to approximate a nonlinear function which can be used to compute the unknown collocation points. Theoretical analysis indicates that the numerical error is controllable and does not increase when the simulation time step increases. There are several advantages to the proposed approach. The powerful expressive ability of neural networks enables the ANNs to accurately approximate stochastic collocation points. The compression-decompression method reduces the computational costs, so that the numerical method can be applied in practice. In finance, the proposed big time step methodology will be highly beneficial for the generation for path-dependent financial option contracts or in risk management applications. As an outlook, it will be relevant to extend the introduced methodology to solving higher-dimensional or more involved SDE dynamics. We will define multi-dimensional stochastic collocation points (i.e. by means of a tensor) for a multi-dimensional system of SDEs, and choose a Convolutional Neural Network~\citep{2015nature} to efficiently process these collocation points. Of course, this may not trivially generalize to truly high-dimensional systems, but approximation of moderate dimensionality should be possible The computational speed can be further improved by parallel computation, for example, on GPUs. Non-Markovian processes may also be solved with a large time step by the proposed ANN method, where the conditional collocation points are dependent on past realizations. Fractional Brownian motion~\citep{FBM1968} forms a relevant example, which is used for the simulation of rough volatility in finance~\citep{roughvol2018}. In such a context, advanced variants of fully connected neural networks, e.g., recurrent neural networks (RNN) or long short-term memory (LSTM) networks (see a review in~\cite{RNNreview}), are recommended when approximating the nonlinear transition probability function, for example, Equation~\eqref{eq:condY}. As another outlook, Multilevel Monte Carlo (MLMC) methods, as developed by~\cite{mlmc2008,giles_2015}, form another interesting research topic for our large time step accurate discretisation schemes. It is well-known that the strong convergence properties of SDE discretizations impact the efficiency of the MLMC methods. \section{Acknowledgments} S. Liu would like to thank the China Scholarship Council (CSC) for the financial support. \bibliographystyle{plain}
{ "redpajama_set_name": "RedPajamaArXiv" }
8,482
As a blogger, I'm constantly using online resources and apps to create photos/ posters/ banners to enhance my writing and support my posts. I recently ran across one that is TOO GOOD not to share. I simply MUST spread the word to let everyone see what I'm raving about …. FotoJet. FotoJet is an online site that allows you to CREATE, DESIGN, and EDIT. After selecting one of their AWESOME pre-designed templates to start your creation, the sky is the limit! You can rely on their talent, or put your own twist on the highly customizable options they offer. I've done a combination of both, and each time walked away with stunning results. My favorite feature is that FotoJet took the guesswork out of sizing. Looking for something to post to your Pinterest board, specifically? These are only a few of the customizable options you have on FotoJet. Add your pictures, your logo, your words. BAM. Once you choose your template, you make it your own!! I started with the template (on the left) and made it my own design (on the right). Easy peasy…. Below are a few that I use regularly for my reviews and giveaways. FotoJet has helped my blog to have a clean, sleek, professional feel. EVERYONE. No joke. Teachers. Bloggers. Moms. Party Planners. Wedding Planners. PTA presidents. CPA's. Principals. New Mom Creating a Baby Announcement. Friends sending a birthday card. EVERYONE. Many FREE templates that you can customize without paying a dime. Though watch out – The more you play, the more tempted you'll be to buy! NONE. Seriously, I can't think of a single thing I would change about FotoJet. I've yet to want something they can't provide. YES! YES! YES!!! I would happily pay for this service… And am!
{ "redpajama_set_name": "RedPajamaC4" }
8,574
What are they perceived to do and what do they do? Professor Geoffrey C. Fox, Director, Digital Science Center. We consider Big Data Systems such as Hadoop, Spark and TensorFlow and identify what they do well (which is a lot) and where they have omissions. We consider a programming model where "every call" is wrapped by a learning framework that configures execution (auto-tuning) and learns results. We describe our big data framework Twister2 and explain where it can offer improved capabilities over current systems. Fox received a Ph.D. in Theoretical Physics from Cambridge University and is now distinguished professor of Informatics and Computing, and Physics at Indiana University where he is director of the Digital Science Center, Chair of Department of Intelligent Systems Engineering and Director of the Data Science program at the School of Informatics, Computing, and Engineering. He previously held positions at Caltech, Syracuse University and Florida State University after being a postdoc at the Institute of Advanced Study at Princeton, Lawrence Berkeley Laboratory and Peterhouse College Cambridge. He has supervised the Ph.D. of 68 students and published around 1200 papers in physics and computer science with an index of 70 and over 26000 citations. He currently works in applying computer science from infrastructure to analytics in Biology, Pathology, Sensor Clouds, Earthquake and Ice-sheet Science, Image processing, Deep Learning, Manufacturing, Network Science and Particle Physics. The infrastructure work is built around Software Defined Systems on Clouds and Clusters. The analytics focuses on scalable parallelism. He is involved in several projects to enhance the capabilities of Minority Serving Institutions. He has experience in online education and its use in MOOCs for areas like Data and Computational Science. He is a Fellow of APS (Physics) and ACM (Computing).
{ "redpajama_set_name": "RedPajamaC4" }
1,149
Example essay english persuasive examples high school gay marriage thesis expository statement of personal narrative toreto co. For argumentative essays science sample how to write a classification writing exa nuvolexa topics research paper help application synthesis services good co. Address ex samples. An informative about class video and nardellidesign com college after apa format also library form. Sample narrative essay high school thesis support examples for proposal example unique personal. Rated persuasive statement essays and term papers business plan argument autobiography students principal college winning image gallery of admission about family apa format paper english writing simple in also how to an expository romeo juliet highschool on picture a pa gender equality compare ap budget cultural baggage barbara toreto co starting igde write reflective research yellow. Essay narrative example co compare and contrast about high school college general apa format paper english writing examples. Sample application the yellow healthy lifestyle personal biographical of pdf biography for students entrance essays how to write papers from private admission best a twenty hueandi co. Short cover letter topics grade persuasive thesis statement research proposal student position science fiction technology an memories also ap argumentative business studies locavores ideas vs writea perfect catholic society.
{ "redpajama_set_name": "RedPajamaC4" }
2,260
{"url":"https:\/\/brilliant.org\/discussions\/thread\/not-a-big-question\/","text":"\u00d7\n\n# Not a big question...\n\nGuys I have a very obvious query , but still I want to ask you all if this is true or not:\n\nIf a triangle has its circumcenter on one of its sides , then it is a right angled triangle.\n\nPlease tell me if this is true or not , since I am working to construct a problem. (Most probable TONCAS) . Thanks!\n\nNote by Nihar Mahajan\n2\u00a0years ago\n\nSort by:\n\nYes, Let the triangle be ABC and let the circumcentre O lie on BC. Now OB=OC. Thus O is the midpoint of BC and OA=OB=OC. Thus A=90(Simple angle chasing) \u00b7 2\u00a0years ago\n\nYeah I got it too. Also , I got a my own TONCAS problem! Trying to prove it... :D \u00b7 2\u00a0years ago\n\nWhich daily workout are you on? \u00b7 2\u00a0years ago","date":"2017-03-31 00:33:44","metadata":"{\"extraction_info\": {\"found_math\": false, \"script_math_tex\": 0, \"script_math_asciimath\": 0, \"math_annotations\": 0, \"math_alttext\": 0, \"mathml\": 0, \"mathjax_tag\": 0, \"mathjax_inline_tex\": 0, \"mathjax_display_tex\": 0, \"mathjax_asciimath\": 0, \"img_math\": 0, \"codecogs_latex\": 0, \"wp_latex\": 0, \"mimetex.cgi\": 0, \"\/images\/math\/codecogs\": 0, \"mathtex.cgi\": 0, \"katex\": 0, \"math-container\": 0, \"wp-katex-eq\": 0, \"align\": 0, \"equation\": 0, \"x-ck12\": 0, \"texerror\": 0, \"math_score\": 0.8436520099639893, \"perplexity\": 1643.0636483825251}, \"config\": {\"markdown_headings\": true, \"markdown_code\": true, \"boilerplate_config\": {\"ratio_threshold\": 0.18, \"absolute_threshold\": 10, \"end_threshold\": 15, \"enable\": true}, \"remove_buttons\": true, \"remove_image_figures\": true, \"remove_link_clusters\": true, \"table_config\": {\"min_rows\": 2, \"min_cols\": 3, \"format\": \"plain\"}, \"remove_chinese\": true, \"remove_edit_buttons\": true, \"extract_latex\": true}, \"warc_path\": \"s3:\/\/commoncrawl\/crawl-data\/CC-MAIN-2017-13\/segments\/1490218205046.28\/warc\/CC-MAIN-20170322213005-00455-ip-10-233-31-227.ec2.internal.warc.gz\"}"}
null
null
Q: The method of goBack() does not work in Android WebView In my activity which contains a WebView, I override onBackPressed() method, here is my code. @Override public void onBackPressed() { if(mWebView.canGoBack()){ Log.d("DEBUG","canGoBack"); mWebView.goBack(); return; } super.onBackPressed(); } String "canGoBack" is printed if there is a history page, but the activity has no change. Here is my initialize WebView code which is invoked in onCreate() method. private void initWebViewSettings() { mWebView.setWebViewClient(new MyWebViewClient()); WebSettings webSettings = mWebView.getSettings(); webSettings.setCacheMode(WebSettings.LOAD_NO_CACHE); //no cache webSettings.setJavaScriptEnabled(true); webSettings.setCacheMode(WebSettings.LOAD_NO_CACHE); mWebView.addJavascriptInterface(new WebAppInterface(this), "Android"); } Here is my onResume() method, I invoke loadUrl(). @Override protected void onResume() { super.onResume(); mShopId = getIntent().getIntExtra(App.SHOP_ID,-1); if(mShopId != -1){ mWebView.loadUrl(App.SHOP_URL + mShopId + App.INDEX); } } A: Well this is my solution for this issue. I did the same way for many webview project webView.setOnKeyListener(new OnKeyListener() { @Override public boolean onKey(View v, int keyCode, KeyEvent event) { if(event.getAction() == KeyEvent.ACTION_DOWN) { WebView webView = (WebView) v; switch(keyCode) { case KeyEvent.KEYCODE_BACK: if(webView.canGoBack()) { webView.goBack(); return true; } break; } } return false; } }); A: Try out my solution for getting back to the old link on pressing key down public boolean onKeyDown(int keyCode, KeyEvent event) { if (keyCode == event.KEYCODE_BACK && webView.canGoBack()){ webView.goBack(); return false; } return super.onKeyDown(keyCode, event);
{ "redpajama_set_name": "RedPajamaStackExchange" }
8,791
function SetValuestoIndexes(numbers) { let length=Number(numbers[0]); let arr=Array(length).fill(0); for(let i=1;i<numbers.length;i++){ let currentArray=numbers[i].split(' '); let index=Number(currentArray[0]); let value=Number(currentArray[2]); arr[index]=value; } //console.log(arr.join('\n')) for(let i=0;i<arr.length;i++){ console.log(arr[i]); } }
{ "redpajama_set_name": "RedPajamaGithub" }
1,360
Meet The Artists Behind The Third Edition of Dior Lady Art Fashion, News & Trends Elyza Khamil The story goes like this: A bag crafted from the supplest of leathers and stitched to emulate Napoleon III chairs, designer by former Dior creative director Gianfranco Ferré, was gifted to Diana, Princess of Wales, by ex-French first lady Bernadette Chirac in 1995. It wasn't until 1996 that the bag was named Lady Dior, in homage to Princess Diana's irreverent style. Gregg Bleakney This icon has continued to reinvent itself season after season, year after year. However, this 2019 marks the first time 11 female artists of different generations and nationalities have been chosen to reinterpret the Lady Dior bag, and champions unconventional femininity with the most unexpected of savoir-faire. Meet the talented collaborators behind the project: Li Shurui, China Likai The movement and radiance of light is the primary inspiration of the Chinese artist with the elements of nature, water and air are among her essential muses. The use of light and space at different times and in different cultures, reveals the desires, needs and state of mind of individuals. For this collaboration, Li Shurui celebrates that which "can only come from the hand of man and with time", the nobility of savoir-faire specific to craft and the time spent making these unique creations. Her work reflects the contrasting relations between this removed time and that of the digital "flow" of post-industrial society. Exclusive: Phillip Lim on the Importance of Picking Meaningful Collaborations Dior Celebrates Cruise 2022 With a Special Pop-Up Café in Kuala Lumpur Art & About, Lifestyle Carine Roitfeld on Fashion for the Future
{ "redpajama_set_name": "RedPajamaCommonCrawl" }
2,769
Parrésia, also Parrésia Publishers Ltd, is a publishing company in Nigeria founded by Azafi Omoluabi Ogosi and Richard Ali in 2012 with the aim of selling books to the Nigerian reading audience and promote the freedom of the imagination and the free press. It was described in 2017 by The New York Times as one of "a handful of influential new publishing houses" in Africa in the last decade. Character Parrésia publishes its fiction and creative non-fiction works under four imprints: Parrésia Books, Origami Books, Cordite, and Omode Meta. Parrésia Books is a traditional publishing imprint that publishes five books each year — literary fiction, popular fiction and short-story collections. The Parrésia imprint, being a traditional publisher, offers advances on royalties for first-time authors. All responsibilities are borne by the publisher. Origami Books is a specialised publishing imprint which publishes work in all genres. The author takes full control of the book production process. Cordite Books is a new, traditional-style publishing company for genre fiction based in Lagos, Nigeria, and is jointly owned by Parrésia Publishers Ltd and Nigerian writer Helon Habila who is also a professor of Creative Writing at George Mason University in the United States. Habila is the editor for the series. Ọmọde Mẹta is a traditional publishing imprint that publishes children's books for all children ages and in all genres. Authors Helon Habila, winner of the Caine Prize and Commonwealth Writers' Prize for Fiction. His novel Oil on Water was shortlisted for the Orion Environmental Book Award and Pen/Opera Book Award in 2012. Oil on Water was also nominated for the Best Novel Commonwealth Writers' Prize Africa Region in 2011. Chika Unigwe (winner of NLNG Prize 2012) Abubakar Adam Ibrahim (winner of LNG Prize 2016 for Season of Crimson Blossoms, published by Parrésia Books in 2015). Ogaga Ifowodo, whose poetry collection A Good Mourning was shortlisted for NLNG Prize in 2017 Pius Adesanmi, a Nigerian columnist and professor at Carleton University in Canada. References External links Official website 2006 establishments in Nigeria Book publishing companies of Nigeria
{ "redpajama_set_name": "RedPajamaWikipedia" }
3,015
West Brom blog: If James Morrison leaves this summer, who do Albion replace him with? West Brom blogger Brad Poole takes a closer look at Morrison's contractual situation - and potential void-fillers Brad Poole James Morrison celebrates scoring West Bromwich Albion's equaliser against Arsenal James Morrison scored with the final kick of the game to ensure West Bromwich Albion are in the hat in the FA Cup but his future at the club remains in doubt. It has been reported that his contract is up at the end of the season and could be set to leave the side after nine years at the club on a free transfer. The Scot has been a loyal servant to the club during his tenure at the Hawthorns, playing in numerous positions across the midfield with minimal fuss. He has scored sensational goals and on the weekend proved he can still score the odd vital goal for Albion. Despite playing 20 times in all competitions and being Albion's joint top scorer in the league with a staggeringly low three goals, should Albion keep the 29 year old heading into next season? West Brom 2-2 Bristol City - Aftermath International linked Morrison future The midfielder has drifted in and out of games this season, sometimes looking like the best player at the squad, dribbling past players and playing sensational through balls but at other times he can go missing, not making any sort of impact on the game. He fits into Tony Pulis' style very well. He will work hard for the good of the team and isn't afraid to make his feelings heard if he feels someone else is not giving everything for the team. Albion need leaders in the changing room like Morrison to ensure everyone is fighting for the same cause, the club. However, it could be time to go in a new direction, and bring in some young blood who can fit straight into the Albion side and stay there for the best part of a decade like Morrison. Alex Mowatt of Leeds United (L) and James Henry of Wolverhampton Wanderers in action Zach Clough (L) of Bolton Wanderers is tackled by George Thorne West Bromwich Albion's James Morrison is challenged by Newcastle United's Cheick Tiote The first port of call would be at Leeds United and Alex Mowatt. The centre midfielder has broken onto the scene in the last few seasons has scored some truly sensational goals. At just 20 years old he has his whole career ahead of him and has experienced English football at first hand, racking up 87 appearances for the Yorkshire club, scoring 12 goals. Mowatt has all the attributes necessary to make an impact in the Premier League, especially if he is to play alongside the likes of Darren Fletcher, giving the youngster space to let his creativity shine. Other options for Pulis to fill the void which could be left by Morrison in an advanced attacking role could be Zach Clough from Bolton. poll loading Does James Morrison get a new contract in your book? 0+ VOTES SO FAR The 21-year-old would be available on the cheap due to the dismal state of the finances at his club or the club could even resign former Albion player Kemar Roofe who has found his form at Oxford, scoring two to dump Premier League Swansea out of the FA Cup. However, should anyone come in, they need a mentor. Someone who has experience playing in the league in an attacking midfield role. Morrison has to stay for a couple new season to set an example for someone like Mowatt to come in and shine. The season proves that Morrison still has plenty to offer and will only get better with added competition from a younger option. Aston Villa FCAston Villa hero tells Dean Smith the striker 'he'd love' to see him sign in JanuaryAll eyes are on the transfer window with Villa's options looking a little lightweight up top as they face a hectic schedule in the coming weeks which will push the squad to the limit 'Wow' - Fans go nuts as Aston Villa enter 'advanced talks' over transfer Aston Villa FCLatest Aston Villa transfer news from BirminghamLive as Championship side Cardiff City reportedly enter into advanced talks over a move for defender Frederic Guilbert
{ "redpajama_set_name": "RedPajamaCommonCrawl" }
8,045
Vienne-en-Bessin är en kommun i departementet Calvados i regionen Normandie i norra Frankrike. Kommunen ligger i kantonen Ryes som ligger i arrondissementet Bayeux. År hade Vienne-en-Bessin invånare. Befolkningsutveckling Antalet invånare i kommunen Vienne-en-Bessin Referens:INSEE Se även Lista över kommuner i departementet Calvados Källor Kommuner i Calvados
{ "redpajama_set_name": "RedPajamaWikipedia" }
2,913
Q: Restricting number of login attempts by users I'm using bottle's @auth_basic decorator to build my login module. I want to add a feature that if a user inputs the wrong password they will be restricted from trying again for 5 secs. How can this be implemented with bottle's @auth_basic? A: Obviously I don't have the actual login piece. But here is a way leveraging beaker and a decorator to track login attempts. import gevent from gevent import monkey, signal monkey.patch_all() from gevent.pywsgi import WSGIServer from mainapp import mainappRoute import bottle from bottle import request, get, post, template from beaker.middleware import SessionMiddleware from whitenoise import WhiteNoise staticfolder = 'static' beakerconfig = { 'session.type': 'memory', 'session.auto': True, 'session.cookie_path': '/', 'session.key': 'site_id', 'session.secret' : 'lsfkjsdlfhofuhrlifuheroifh', 'session.httponly' : True } class user(object): def __init__(self): self.session = request.environ['beaker.session'] self.login_attempts = 0 def set(self, **kwargs): for k,v in kwargs.items(): self.session[k] = v self.__dict__.update(self.session) def attempt(self): self.session['login_attempts'] = self.session.get('login_attempts', 0) + 1 if self.session['login_attempts'] == 3: #do something like redirect pass def check_login(fn): def check_uid(**kwargs): u = user() u.attempt() return fn(**kwargs) return check_uid def shutdown(): print('Shutting down ...') server.stop(timeout=60) exit(signal.SIGTERM) @get('/login') def login(): return template('login.html') @post('/login') @check_login def process_login(): u = user() #let javascript handle the timeout return template('index.html', attempts=u.login_attempts) botapp = bottle.app() for Route in (mainappRoute,): botapp.merge(Route) botapp = SessionMiddleware(botapp, beakerconfig) botapp = WhiteNoise(botapp) botapp.add_files(staticfolder, prefix='static/') server = WSGIServer(("0.0.0.0", int(80)), botapp) gevent.signal(signal.SIGTERM, shutdown) gevent.signal(signal.SIGINT, shutdown) # CTRL C server.serve_forever()
{ "redpajama_set_name": "RedPajamaStackExchange" }
2,133
Q: Scrapy Python spider unable to find links using LinkExtractor or by manual Request() I am trying to write a Scrapy spider that crawls through all the results pages on the domain:https://www.ghcjobs.apply2jobs.com.... The code should do three things: (1) Crawl through all the pages 1-1000. These pages are identical, save for being differentiated by the final portion of the URL: &CurrentPage=#. (2) Follow each link inside the results table containing job postings where the link's class = SearchResult. These are the only links within the table, so I am not in any trouble here. (3) Store the information shown on the job description page in key:value JSON format. (This part works, in a rudimentary fashion) I have worked with scrapy and CrawlSpiders before, using the 'rule = [Rule(LinkExtractor(allow=' method of recursively parsing a page to find all the links that match a given regex pattern. I am currently stumped on step 1, crawling through the thousand result pages. Below is my spider code: from scrapy.contrib.spiders import CrawlSpider, Rule from scrapy.http.request import Request from scrapy.contrib.linkextractors import LinkExtractor from genesisSpider.items import GenesisJob class genesis_crawl_spider(CrawlSpider): name = "genesis" #allowed_domains = ['http://www.ghcjobs.apply2jobs.com'] start_urls = ['https://www.ghcjobs.apply2jobs.com/ProfExt/index.cfm?fuseaction=mExternal.returnToResults&CurrentPage=1'] #allow &CurrentPage= up to 1000, currently ~ 512 rules = [Rule(LinkExtractor(allow=("^https://www.ghcjobs.apply2jobs.com/ProfExt/ index.cfm\?fuseaction=mExternal.returnToResults&CurrentPage=[1-1000]$")), 'parse_inner_page')] def parse_inner_page(self, response): self.log('===========Entrered Inner Page============') self.log(response.url) item = GenesisJob() item['url'] = response.url yield item Here is the output of the spider, with a bit of the execution code on top cut off: 2014-09-02 16:02:48-0400 [genesis] DEBUG: Crawled (200) <GET https://www.ghcjobs .apply2jobs.com/ProfExt/index.cfm?fuseaction=mExternal.returnToResults&CurrentPa ge=1> (referer: None) ['partial'] 2014-09-02 16:02:48-0400 [genesis] DEBUG: Crawled (200) <GET https://www.ghcjobs .apply2jobs.com/ProfExt/index.cfm?CurrentPage=1&fuseaction=mExternal.returnToRes ults> (referer: https://www.ghcjobs.apply2jobs.com/ProfExt/index.cfm?fuseaction= mExternal.returnToResults&CurrentPage=1) ['partial'] 2014-09-02 16:02:48-0400 [genesis] DEBUG: ===========Entrered Inner Page======== ==== 2014-09-02 16:02:48-0400 [genesis] DEBUG: https://www.ghcjobs.apply2jobs.com/Pro fExt/index.cfm?CurrentPage=1&fuseaction=mExternal.returnToResults 2014-09-02 16:02:48-0400 [genesis] DEBUG: Scraped from <200 https://www.ghcjobs. apply2jobs.com/ProfExt/index.cfm?CurrentPage=1&fuseaction=mExternal.returnToResu lts> {'url': 'https://www.ghcjobs.apply2jobs.com/ProfExt/index.cfm?CurrentPag e=1&fuseaction=mExternal.returnToResults'} 2014-09-02 16:02:48-0400 [genesis] INFO: Closing spider (finished) 2014-09-02 16:02:48-0400 [genesis] INFO: Dumping Scrapy stats: {'downloader/request_bytes': 930, 'downloader/request_count': 2, 'downloader/request_method_count/GET': 2, 'downloader/response_bytes': 92680, 'downloader/response_count': 2, 'downloader/response_status_count/200': 2, 'finish_reason': 'finished', 'finish_time': datetime.datetime(2014, 9, 2, 20, 2, 48, 611000), 'item_scraped_count': 1, 'log_count/DEBUG': 7, 'log_count/INFO': 7, 'request_depth_max': 1, 'response_received_count': 2, 'scheduler/dequeued': 2, 'scheduler/dequeued/memory': 2, 'scheduler/enqueued': 2, 'scheduler/enqueued/memory': 2, 'start_time': datetime.datetime(2014, 9, 2, 20, 2, 48, 67000)} 2014-09-02 16:02:48-0400 [genesis] INFO: Spider closed (finished) Currently, I am stuck on objective (1) of this project. As you can see, my spider only crawls through the start_url page. My regex should be targeting the page navigation buttons correctly as I have tested the regex. My callback function, parse_inner_page, is working, as is shown by the debugging comment I inserted, but only on the first page. Am I using 'Rule' incorrectly? I was thinking that maybe the page being HTTPS was somehow to blame... Jut as a way to tinker a solution, I tried using a manual request for the second page of results; this didn't work. Here is the code for that too. Request("https://www.ghcjobs.apply2jobs.com/ProfExt/index.cfm?fuseaction=mExternal.returnToResults&CurrentPage=2", callback = 'parse_inner_page') Can anyone offer any guidance? Is there maybe a better way to do this? I have been researching this on SO / Scrapy documentation since Friday. Thank you so much. UPDATE: I have resolved the issue. The problem was with the start url I was using. start_urls = ['https://www.ghcjobs.apply2jobs.com/ProfExt/index.cfm?fuseaction=mExternal.returnToResults&CurrentPage=1'] Leads to a post-form-submission page that is the result of clicking the "search" button on This page. This runs javascript on the client side to submit a form to the server, which reports the full job board, pages 1-512. However, there exists another hard-coded URL which apparently calls the server without needing to use any client-side javascript. So now my start url is start_urls = ['https://www.ghcjobs.apply2jobs.com/ProfExt/index.cfm?fuseaction=mExternal.searchJobs'] And everything is back on track! In the future, check and see if there are any javascript independent URLs for calling server resources. A: Are you sure Scrapy sees the web page in the same way as you? Nowadays, more and more sites are built by Javascript, Ajax .. And those dynamic content might need a fully functional browser to be fully populated. However, neither Nutch nor Scrapy will handle those out of box. First of all, you need to make sure the web content you are interested in can be retrieved by scrapy. There are a few ways to do it. I usually use urllib2 and beautifulsoup4 to give it a quick try. And the your start page failed my test. $ python Python 2.7.6 (default, Mar 22 2014, 22:59:56) [GCC 4.8.2] on linux2 Type "help", "copyright", "credits" or "license" for more information. >>> import urllib2 >>> from bs4 import BeautifulSoup >>> url = "https://www.ghcjobs.apply2jobs.com/ProfExt/index.cfm?fuseaction=mExternal.returnToResults&CurrentPage=1" >>> html = urllib2.urlopen(url).read() >>> soup = BeautifulSoup(html) >>> table = soup.find('div', {'id':'VESearchResults'}) >>> table.text u'\n\n\n\r\n\t\t\tJob Title\xa0\r\n\t\t\t\r\n\t\t\n\r\n\t\t\tArea of Interest\xa0\r\n\t\t\t\r\n\t\t\n\r\n\t\t\tLocation\xa0\r\n\t\t\t\r\n\t\t\n\r\n\t\t\tState\xa0\r\n\t\t\t\r\n\t\t\n\r\n\t\t\tCity\xa0\r\n\t\t\t\r\n\t\t\n\n\n\r\n\t\t\t\t\tNo results matching your criteria.\r\n\t\t\t\t\n\n\n' >>> As you can see, "There is no results matching your criteria!" I think you might need to figure out why the content is not populated. Cookies? Post instead of Get? User Agent..etc. Also, you can use scrapy parse command to help you debug. For example, I use this command quite often. scrapy parse http://example.com --rules A few other scrapy commands, maybe Selenium that might be helpful down the road. Here I am using running scrapy shell in iPython to inspect your start url and also the first record that I can see in my browser contains Englewood and it doesn't exist in the html that scrapy grabbed Update: What you are doing is a really trivial scraping work and you really don't need Scrapy, it is a bit overkill. Here are my suggestions: * *Take a look at Selenium (I am assuming you write Python) and make headless Selenium in the end when you try to run it on a server. *You can implement this with PhantomJS which is a much lighter Javascript executor to get your work done. Here is another stackoverflow question that might be helpful. *Several other resources that you can make a career in.
{ "redpajama_set_name": "RedPajamaStackExchange" }
107
Während des knapp 150-jährigen Zeitraums der Dampfschifffahrt auf dem Bodensee änderte sich das Erscheinungsbild der Dampfschiffe mehrfach auf Grund veränderter Bedingungen und Aufgaben – Form follows function: Vom Frachtschiff-Glattdecker zum kombinierten Fracht- und Passagierdampfer bis zum Zweideck-Salondampfschiff. Der technische Fortschritt führte außerdem zu konstruktiven Veränderungen der Dampfmaschine, des Schaufelrads und des Rumpfs der Schiffe. Prägendes Merkmal war bei 70 von insgesamt 75 Dampfschiffen der Seitenradantrieb. Geschichte Der erste Versuch scheiterte Am 29. April 1818 unternahm das erste Dampfschiff auf dem Bodensee, die Stephanie, eine Jungfernfahrt von Konstanz nach Meersburg. Die Rückfahrt schaffte sie wegen der schwachen Maschine nicht mehr und musste von den Passagieren nach Konstanz zurückgerudert werden. Der Bauherr Johann Caspar Bodmer entzog sich seinen Gläubigern durch die Flucht ins Königreich Württemberg. Erste Epoche ab 1824: Die Dampfboote der privaten Gesellschaften Die Bodensee-Dampfschifffahrt begann mit der württembergischen Wilhelm, die 1824 in Konkurrenz zu den zahlreichen Lastsegelschiffen (Lädinen und Segnern) trat. Sie und weitere Dampfboote übernahmen zunächst deren Aufgaben und beförderten in Säcken, Fässern und Kisten verpackte Güter, Holz und Sandsteine, Klein- und Großvieh sowie Personen, meistens Händler. Immer wichtiger wurde die Beförderung von Seepost. Der bauchige, etwa 30 m lange Glattdeckrumpf des neuen Schiffstyps war anfangs aus Eichenholz, die Form orientierte sich an Hochsee-Segelschiffen. Die Dampfmaschine hatte einen hohen, mehrfach verstagten Schornstein. Zwei Radkästen verdeckten die Schaufelräder und vergrößerten die Schiffsbreite auf das Doppelte. Größere Aufbauten gab es keine, stattdessen eine Stoffabdeckung über dem Achterdeck. Einige Dampfboote hatten noch zwei Masten mit Fock- und Gaffelsegel. Der Kapitän stand bei Manövern auf einem der Radkästen und rief dem Maschinisten unter Deck und den Matrosen die Kommandos des für den nautischen Bereich zuständigen Steuermanns auf dem Heck zu. Im Laufe der ersten Epoche wurden die beiden Radkästen zunächst durch eine schmale Brücke verbunden. Durch einen längeren Schiffsrumpf, ab 1838 aus Eisen, wirkten die Boote schmaler. Die Länge (bis 55 m) war durch die engen Häfen und der Tiefgang durch den niedrigen Wasserstand im Winter begrenzt. Um größere Mengen befördern zu können, nahm man bis zu acht Lastkähne in Schlepp. Der Bugspriet war oft mächtig und reichlich verziert. Am Bug befanden sich neben Ankerwinde und Schiffsglocke auch eine Signalkanone. Die Ruderpinne am Heck wurde bald von einem Steuerrad auf dem Achterdeck ersetzt. Eigner der Dampfboote waren drei private Gesellschaften mit den Heimathäfen Konstanz, Friedrichshafen und Lindau. Eines der letzten Schiffe dieser Epoche, die Stadt Constanz, erreichte nach zwei Umbauten die längste Einsatzzeit eines Bodenseeschiffes: 145 Jahre, von 1840 bis 1985. Zweite Epoche ab 1863: Die Dampfschiffe der Ländereisenbahnen Im Laufe der ersten Epoche bauten die drei deutschen Südländer neue Eisenbahnstrecken an den Bodensee, die in den Hafenbereichen von Friedrichshafen, Lindau und Konstanz endeten. Die Schweiz verband die Häfen von Romanshorn, Rorschach und Schaffhausen mit dem Hauptnetz. Zuletzt schloss Österreich den neuen Bregenzer Hafen an. Bis 1901 mit dem Abschnitt Überlingen–Friedrichshafen der Bahnstrecke Stahringen–Friedrichshafen die letzte Lücke geschlossen wurde, war es die neue Aufgabe der Schiffe, die Verbindungen für Zugreisende und Güter herzustellen. Deshalb übernahmen die staatlichen Eisenbahngesellschaften der drei deutschen Länder zwischen 1847 und 1863 die privaten Gesellschaften mit deren Schiffsbestand und erweiterten ihn. Für das Erscheinungsbild, bedeutete dies, dass die Glattdeckdampfer als fast reine Frachtschiffe nach und nach verschwanden und ab 1877 von kombinierten Fracht- und Passagierdampfschiffen in der für den Bodensee typischen Art des Halbsalondampfschiffs abgelöst wurden. Im Gegensatz zu den auf Mittelrhein, Donau und den Schweizer Seen üblichen Decksalonschiffen mit den Passagierräumen auf der Hauptdeckebene, war bei diesem Schiffstyp der Hecksalon bis zur Unterkante der Fenster in den Rumpf abgesenkt. Dadurch wurde die Windangriffsfläche bei Stürmen reduziert und die Sicherheit erhöht. Der Aufbau beschränkte sich auf den Hecksalon, auf dem sich ein offenes Promenadendeck befand. Mittschiffs stand ein erhöhter Ruderstand auf einer Brücke über dem Eingangsbereich des Schiffes. Zwischen den Radkästen war durch eine Öffnung im Hauptdeck die Kurbelwelle der Dampfmaschine sichtbar. Das Vorschiffdeck war der Fracht vorbehalten, der düstere Bugraum den einfachen Fahrgästen. Im Gegensatz zu ihren Vorgängern hatten die Schiffe jetzt alle einen Vormast mit Rah zur Signal- und Lichterführung, teilweise auch mit der Takelage für ein Gaffelsegel. Am Heck und/oder auf den Radkästen befanden sich ein bis zwei Rettungsboote an Davits und Rettungsringe an der Reling. Die Glattdecker auf dem Obersee wurden bald entweder ausgemustert oder mit Aufbauten versehen. Wegen der Brückenpassagen hielten sich die flachen Schiffe mit halbversenkten Radkästen, klappbarem Schornstein und geringem Tiefgang auf Untersee und Rhein am längsten. Schon das erste Schiff der neuen Generation, der Decksalondampfer Kaiser Wilhelm, zog mit seinem repräsentativen Komfort im historistischen Stil Staatsgäste und immer mehr Touristen an. Die Farbgebung der Schiffe war nicht einheitlich und wechselte häufig. Der Schornstein wurde farblich und mit Emblemen ebenso landestypisch gestaltet wie der Klipperbug und das Heckschanzkleid von Schweizer Schiffen. Ein weiterer Trend in dieser Phase waren maritime Adaptionen, wie der negative Vorsteven, Windhutzen und der zweite Mast mit einer Gaffel für die Landesfahne. Eine besondere Rolle spielten in dieser Epoche die drei zwischen 1885 und 1892 gebauten "Kaiserschiffe" der österreichischen Staatsbahnen. Diese komfortablen Schiffe mit elektrischer Beleuchtung und gehobener Bordküche schlossen die Lücke im europäischen Schienennetz Paris/Strassburg – Innsbruck/Wien zwischen Konstanz und Bregenz mit einem Schnellkurs von nur zwei Stunden Dauer. Ganz anders entwickelte sich der Gütertransport über den Bodensee mit Schwerpunkt zwischen Friedrichshafen und Romanshorn. Anstatt die Fracht zweimal umzuladen, wurden bis zu acht Güterwagen ab 1869 komplett auf offenen, zweigleisigen Trajektkähnen befördert, die von Kursschiffen oder Schleppern gezogen wurden. Eine Steigerung stellten die beiden Dampftrajekte mit zwei Schornsteinen dar, die mit 70 m Länge und 18 m Breite die größten Schiffe waren, die den Bodensee je befuhren und die mit zwei Schleppkähnen einen ganzen Güterzug befördern konnten. Die Kursschiffe wurden verstärkt zur Beförderung von Seepost eingesetzt, nach 1882 beförderten sie als Postschiffe auch Schiffspost. Als weiterer Schiffstyp entstand in dieser Epoche die Dampfyacht. Eigner des ersten privaten "Vergnügungsdampfers" auf dem Bodensee war Eugen Zardetti aus Bregenz. Im Vergleich zu den damaligen Raddampfern war seine Dampfjacht Passe Temps eleganter und wendiger. Dritte Epoche ab 1901: Konkurrenz durch die Eisenbahn Die 1901 vollendete Bahnstrecke Stahringen–Friedrichshafen ermöglichte es, fast alle Orte am See per Bahn direkt zu erreichen. Dadurch reduzierte sich einerseits das Fracht- und Fahrgastaufkommen der Schiffe, später noch verstärkt durch den Straßenverkehr, andererseits kamen mit der Bahn immer mehr Touristen aus den Ballungsgebieten an den Bodensee. Dies bewirkte eine weitere Funktionsänderung hin zu Ausflugsschiffen, was wiederum das Aussehen veränderte. Von 1890 bis 1913 wurden zehn ältere Schiffe durch 21 moderne Neubauten ersetzt. Bei allen wurde das Fassungsvermögen und die maximale Geschwindigkeit erhöht. Das Zweideck-Doppelsalondampfschiff Stadt Bregenz blieb bei einer Größe für 1000 Passagiere aber zunächst eine Ausnahme. Die Länderbahnen überboten sich hinsichtlich des Designs der Schiffe und deren Interieur, das teilweise von führenden Jugendstil-Architekten so luxuriös gestaltet wurde, dass manche bei offiziellen Staatsbesuchen eingesetzt wurden. Auffällig waren die großen Wappen aus Bronze, welche die Radkästen schmückten. Selbst die Anordnung und Form der Salonfenster waren oft landesspezifisch. Der Vorsteven war gerade und schmucklos. Die Bodenseedampfschiffe wurden allmählich auch farblich zur "Weißen Flotte". Der Erste Weltkrieg setzte dieser Entwicklung durch einen rapiden Rückgang der Fahrgastzahlen ein Ende. Danach war die Blütezeit der Bodenseedampfschiffe vorbei. Vierte Epoche ab 1918: Die Dampfschiffe der Deutschen Reichsbahn Nach dem Ende der Monarchie wurden 1919 außer in Württemberg die monarchistischen Namen meist gegen Städtenamen ausgetauscht. Am 1. April 1920 gingen die 18 Schiffe der Länderbahnen in den Besitz der Deutschen Reichsbahn über, Landesflaggen und Schornsteinmarkierungen wurden durch nationale ersetzt. Die Verwaltung der Bodenseeschifffahrt blieb auch nach der Übernahme bei den Direktionen in Karlsruhe, Stuttgart und Augsburg. Bald machte der sprunghafte Anstieg des Tourismus eine Flottenvergrößerung notwendig, doch wurde neben zahlreichen Motorschiffen nur noch ein Dampfschiff 1929 neu gebaut. Fast alle anderen wurden einheitlich umgebaut. Um die Kapazität zu erhöhen, entstand ein neuer geräumiger Salonaufbau auf dem jetzt frachtfreien Vorschiff, darüber ein kleinerer Salon mit verlängertem Promenadendeck. Das Steuerhaus wurde entsprechend höher gesetzt. Dadurch veränderte sich das Aussehen erheblich: Aus zierlich anmutenden Dampfbooten wurden stattliche Zweideck-Dampfschiffe. Im Inneren wurden die Jugendstilverzierungen als "altmodisch" entfernt. Alle Schiffe waren weiß, nur die Farbe des Schornsteins variierte anfangs noch. 1933 beheimateten die Häfen Lindau, Friedrichshafen, Konstanz und Radolfzell insgesamt 31 Schiffe. Mit dem Anschluss Österreichs im Jahre 1938 wurden die Schiffe des Nachbarlandes ebenfalls der Reichsbahn unterstellt und mit schwarz-weiß-roten Schornsteinringen versehen. Im Zweiten Weltkrieg wurden die schmückenden Wappen aus Bronze von den Radkästen entfernt und eingeschmolzen. Ein grau-blauer Tarnanstrich sollte ab 1943 die Dampfschiffe der Reichsbahn vor Fliegerangriffen schützen. Nach Kriegsende trugen die Schiffe über der Ankerklüse am Bug eine Registriernummer der französischen Besatzungsmacht und führten zuerst die französische "Tricolore", ab 1949 die Flagge "Cäsar". Sie entspricht der Flagge "Charlie" im aktuellen Flaggenalphabet. Der Tarnanstrich wurde bis zum Ende der 1940er Jahre wieder entfernt. Fünfte Epoche ab 1952: Das Ende der Bodenseedampfschifffahrt 1952 übernahm die Deutsche Bundesbahn die Rechtsnachfolge der Deutschen Reichsbahn und somit auch die deutschen Bodenseedampfschiffe, jetzt wieder ohne Registriernummer aber ab 1957 zusätzlich mit dem DB-Reederei-Logo in schwarz-rot-gold. Im Jahre 1962 zentralisierte die Deutsche Bundesbahn die Verwaltung der Bodenseeschifffahrt in Karlsruhe. In der letzten Epoche der Bodenseedampfschiffe änderte sich deren Aussehen kaum noch. Spektakuläre Auftritte hatten die Raddampfer bei der Wettfahrt Blaues Band des Bodensees und als über die Toppen beleuchtete Zuschauerschiffe bei den Seefeuerwerken. In den Jahren von 1958 bis 1967 wurden alle Bodenseedampfschiffe – sieben deutsche, drei schweizerische und ein österreichisches – ausgemustert. Nur die aufwändig restaurierte Hohentwiel mit dem neuen Heimathafen Hard erlaubt es als schwimmendes Museum, dass man sich heute noch ein reales Bild vom Aussehen der Bodenseedampfschiffe vor hundert Jahren machen kann, da die Umbauten der 1930er Jahre wieder entfernt wurden. Technische Entwicklung der Dampfschiffe Da die Entwicklung auf den Bodensee durch den damals schon globalen technischen Fortschritt stark beeinflusst wurde, zeigt die folgende Übersicht nur die für den Bodensee typischen Daten der Seitenraddampfer. Bauwerften Fünf Dampfschiffe der ersten Epoche wurden von englischen Werften bezogen. Danach baute Escher, Wyss & Cie. in Zürich den größten Teil der Schiffe für alle Schiffsbetriebe. Erst gegen Ende des 19. Jahrhunderts zogen einige Länderbahnen Hersteller ihrer Region vor: Württemberg die Maschinenfabrik Esslingen und G. Kuhn in Stuttgart, Bayern die Maschinenfabrik J. A. Maffei in München, Schweiz die Maschinenfabrik Gebrüder Sulzer in Winterthur und Österreich die Österreichische Schiffswerften AG in Linz. Die Schiffe wurden von ihnen vorgefertigt, wieder demontiert und mit Pferdefuhrwerken oder der Eisenbahn zur Werft des Heimathafens transportiert. Dort fand die Endmontage statt. Schiffsrumpf Nur die ersten vier Schiffe hatten "Schalen" aus (Eichen-)Holz, die nach wenigen Jahren infolge der starken Erschütterungen durch die Dampfmaschine unbrauchbar wurden. Deshalb hatten alle folgenden Schiffe eine Eisenschale. Anfang der 1930er Jahre erhielten die meisten der damaligen Dampfschiffe seitliche Stabilitätswulste zur Verbesserung des Auftriebs. Die Länge über Alles erhöhte sich von 30,20 m beim ersten Schiff auf 60,60 m beim letzten. Dampfmaschine und Dampfkessel Dampfmaschinen in der chronologischen Reihenfolge ihrer Verwendung ab 1824: Einzylindrige Niederdruck-Balanciermaschine, Kofferkessel mit Holzfeuerung. Zweizylindrige Niederdruck-Balanciermaschine, ein oder zwei Kofferkessel mit Holzfeuerung. Niederdruckmaschine mit oszillierenden Zylindern, Kofferkessel mit Holzfeuerung. Niederdruckmaschine mit oszillierenden Zylindern, ein oder zwei Rohrkessel mit Holzfeuerung. Niederdruckmaschine mit oszillierenden Zylindern, Kofferkessel mit Kohlefeuerung. Woolfsche Zwillingsmaschine mit zwei schrägliegenden Zylindern, zwei Flammrohrkessel mit Umkehrkammer. Verbunddampfmaschine mit zwei schrägliegenden Zylindern. Schrägliegende Zweizylinder-Nassdampfverbundmaschine. Schrägliegende Zweizylinder-Expansionsmaschine, Flammrohrkessel mit Umkehrkammer. Vertikale Zweizylinder-Nassdampfverbundmaschine, Flammrohrkessel mit Umkehrkammer. Schrägliegende Dreizylinder-Nassdampfexpansionsmaschine. Schrägliegende Dreizylinder-Heißdampfexpansionsmaschine. Zweizylinder-Heißdampf-Verbundanlage mit Ventilsteuerung. 1952 wurden drei Schiffe von Kohle- auf Schwerölfeuerung umgestellt. Die Hohentwiel wird heute mit schwefelarmen extraleichtem Heizöl befeuert. Schaufelrad der Dampfschiffe mit Seitenradantrieb Die Wilhelm hatte noch zwei Schaufelräder mit zwölf flachen Holzschaufeln ohne Exzenter. Die späteren Schiffe hatten ebenfalls Holzschaufeln unterschiedlicher Anzahl, aber mit Exzenter zur optimalen Verstellung des Neigungswinkels der Schaufeln beim Eintauchen ins Wasser. Erst spät verwendete man Eisenschaufeln. Leistung und Geschwindigkeit Die Leistung der Dampfmaschine erhöhte sich von 20 PSn der Wilhelm auf 960 PSi der Stadt Überlingen. Entsprechend erhöhte sich die maximale Geschwindigkeit von 5,7 kn (10,5 km/h) auf 16,4 kn (30,3 km/h). Schraubendampfer Mit dem Antrieb über eine Schiffsschraube fährt das als Passagierschiff für 14 Personen zugelassene Dampfboot Felicitas vom Heimathafen Wasserburg im Ausflugsverkehr überwiegend im nordöstlichen Seeanteil (Bj. 1991). Siehe auch Bodenseeschifffahrt Liste der Bodenseedampfschiffe Weiße Flotte (Bodensee) Schraubendampfer auf dem Bodensee Schiffsglocken alter Bodenseeschiffe Bodenseehäfen Weblinks Historische Bilder von Bodenseedampfschiffen auf bodenseeschifffahrt.de Geschichte der Bodenseeschifffahrt auf bodenseeschifffahrt.de Geschichte der Schifffahrt in der Schweiz auf schiffe-schweiz.ch Literatur Michael Berg: Die Motorschifffahrt auf dem Bodensee unter der Deutschen Reichsbahn und in der Nachkriegszeit, verlag regionalkultur, Ubstadt-Weiher 2011, ISBN 978-3-89735-614-6. Dietmar Bönke: Schaufelrad und Flügelrad. Die Schiffahrt der Eisenbahn auf dem Bodensee. GeraMond Verlag, München 2013, ISBN 978-3-86245-714-4, Seiten 19–91, 132, 137, 157,173f. Werner Deppert: Mit Dampfmaschine und Schaufelrad. Die Dampfschiffahrt auf dem Bodensee 1817–1967. Verlag Friedr. Stadler, Konstanz 1975, ISBN 3-7977-0015-6. Karl F. Fritz, Werner Häusler, Monika Sonntag: Erlebnis Schiffahrt auf Bodensee/Rhein und Schweizer Seen. Hrsg. Pressebüro Werner Häusler, Konstanz 1995. Emil Krumholz: Die Geschichte des Dampfschiffahrtsbetriebes auf dem Bodensee. Im Kommissionsverlage der Wagner'schen Universitätsbuchhandlung, Innsbruck, 1906. Klaus von Rudloff und Claude Jeanmaire: Schiffahrt auf dem Bodensee, Band 2: Die Blütezeit der Dampfschiffahrt, Beitrag zur Geschichte des Bodensees, Geschichte der einzelnen Schiffe und Register, Verlag Eisenbahn, Villigen (CH) 1981, ISBN 3-85649-071-X. Einzelnachweise und Anmerkungen Bodenseeschifffahrt Geschichte der Binnenschifffahrt Fahrgastschifffahrt
{ "redpajama_set_name": "RedPajamaWikipedia" }
4,271
\section{Introduction} Complex networks is a new emerging branch of random graph theory. For a long time random graphs have been mainly studied by pure mathematics but recently due to the availability of empirical data on real-world networks they have attracted the attention of physics and natural sciences (see for review \cite{ref:ab,ref:dm,ref:n1}). Methods of statistical physics, both empirical and theoretical, have thus begun to play an important role in this research area. The empirical observations of real-networks has had a feedback on theoretical development which now concentrated on the understanding of the observed features. For example fat tails in node degree distribution, small world effect, degree-degree correlations, or high clustering. Two complementary approaches have been developed: diachronic, known as growing networks \cite{ref:ab,ref:dm,ref:n1}, and synchronic being a sort of statistical mechanics of networks \cite{ref:bck,ref:bl1,ref:dms, ref:bjk1,ref:fdpv,ref:pn1}. We will discuss here the latter. This approach is a natural extension of Erd\"os and R\'enyi ideas \cite{ref:er,ref:bb}. It is well suited both for growing (causal) networks for which nodes' labels reflect the causal order of nodes' attachment to the network \cite{ref:kr,ref:bbjk} and for homogeneous networks for which nodes' labels can be permuted freely in an arbitrary way. Here we shall discuss mainly homogeneous networks. We shall shortly comment on causal networks towards the end of the paper. The main aim of the paper is to present a consistent picture of statistical mechanics of networks. Some ideas have already been introduced earlier. They are scattered in many papers and discussed in many different contexts. We put them together, add some new material and introduce a guideline to obtain a self-contained introduction to statistical mechanics of complex networks. The basic concept in the statistical formulation is statistical ensemble. Statistical ensemble of networks is defined by ascribing a statistical weight to every graph in the given set \cite{ref:bck,ref:bjk1}. Physical quantities are measured as weighted averages over all graphs in the ensemble. The probability of the occurrence of a graph in random sampling is proportional to its statistical weight. If the statistical weight changes then also the probability of occurrence of randomly sampled graphs will change and in effect different random graphs will be observed. The concept of statistical weight is crucial, since it defines randomness in the system. Statistical weight is built out of two ingredients: configuration space weight and functional weight. The configuration space weight is proportional to the uniform probability measure on the configuration space which tells us how to uniformly choose graphs in the configuration space. To illustrate the meaning of the uniform measure consider an ensemble of Erd\"os-R\'enyi graphs with $N$ nodes and $L$ links \cite{ref:er,ref:bb}. The configuration space consists of $\binom{\binom{N}{2}}{L}$ graphs with labeled nodes. All those graphs are equiprobable, and therefore the configuration space weight is the same for each graph. It is convenient to choose this weight to be $1/N!$ since then it can be interpreted as a factor which takes care of $N!$ possible permutations of nodes' labels. This factor has the same origin as the corresponding factor in quantum mechanics for indistinguishable particles and it is constant for all graphs in a finite $N$-ensemble. We can calculate the entropy of random graphs as \begin{equation} S = \ln \frac{1}{N!} \binom{\binom{N}{2}}{L} . \label{S0} \end{equation} In the limit of large sparse graphs: $N \rightarrow \infty$ and $\frac{2L}{N} = \alpha = \mbox{const}>2$, the entropy is an over-extensive function of the system size: \begin{equation} S = \frac{\alpha - 2}{2} N \ln N + \dots , \label{S1} \end{equation} unlike in standard thermodynamics. Let us move to weighted graphs. The idea is to modify the Erd\"os-R\'enyi ensemble by introducing a functional weight which explicitly depends on graph's topology. For example, if we choose the functional weight to be a function of the number of loops on the graph, we can suppress of favor loops of typical graphs in the ensemble. In a similar way we can choose statistical weights to control the node degree distribution to produce homogeneous scale-free graphs \cite{ref:bk} or to introduce correlations between degrees of neighboring nodes \cite{ref:b,ref:n3,ref:bp,ref:d2}. Classical thermodynamics describes systems in equilibrium for which the functional weight is given by the Gibbs measure: $\sim \exp (-\beta E)$, where $E$ is the energy of the system. When discussing complex networks it is convenient to abandon the concept of energy and Gibbs measure and consider a more general form of statistical weights because many networks are not in equilibrium. Indeed, many networks emerging as a result of a dynamical process like growth are far from equilibrium \cite{ref:ab,ref:dm,ref:n1}. It does not mean though that one cannot introduce a statistical ensemble of growing networks. On the contrary, one can for example consider an ensemble of networks which result of many independent repetitions of the growth process terminated when the network reaches a certain size. Such a collection of networks does not describe a thermodynamic equilibrium. The functional weight can be deduced from the parameters of the growth process but of course it has nothing to do with the Gibbs measure. In fact, many real-world networks result from a combination of a growth process and some thermalization processes. For example, the Internet grows but at the same time it continuously rearranges. The latter process introduces a sort of thermalization. Today the growth has probably still larger influence on the topology of the underlying network but in the future the growth may slow down due to saturation and then equilibration processes resulting from continuous rewirings will take over. Similarly all evolutionary networks emerge from a growth mixed with a sort of thermalization related to the continuous network rearrangement. Therefore it is convenient to have a formalism which can extrapolate between the two regimes in a flexible way. The approach which we propose here is capable of modeling functional properties of networks by choosing an appropriate functional weight. Let us return to the configuration space weight. As we mentioned this weight is equivalent to the uniform probability measure on the configuration space for which all graphs are equiprobable. It is a very crucial part of the construction of the ensemble to carefully specify what one means by equiprobable graphs. Consider first graphs with $N$ nodes. There are at least two natural candidates for the uniform measure in such a set of graphs. Since one is interested in shape (topology) of graphs one can define all shapes to be equiprobable. Alternatively one can introduce labels for nodes of each graph to obtain a set of labeled graphs and then one can define all labeled graphs to be equiprobable. The two definitions give two different probability measures since the number of ways in which one can label graph nodes depends on graph's topology and thus the probability of occurrence of a given graph will depend on its topology too. It turns out that the latter definition is more natural. As we have seen above this definition leads to Erd\"os-R\'enyi graphs. So we stick to this definition and from here on we shall ascribe to each labeled graph the configuration space weight $1/N!$ which is constant in the set of graphs of size $N$. The situation is more complex if one considers pseudographs that is graphs which have multiple connections (more than one link between two nodes) or self-connections (a link having the same node at its endpoints). In this case one can also label links and ascribe the same statistical weight to each fully labeled graph. For this choice the statistical weight of each graph is equal to the symmetry factor of Feynman diagrams generated in the Gaussian perturbation field theory \cite{ref:bck}. The paper is organized as follows. In the next section we will recall some basic definitions. Then we will discuss Erd\"os-R\'enyi graphs in the context of constructing statistical ensemble and later we will generalize the construction to weighted homogeneous graphs. After this we will describe Monte-Carlo algorithms to generate graphs for canonical, grand-canonical and micro-canonical ensembles and discuss their representation in terms of adjacency matrices. A section will be devoted to pseudographs. In the last section we will shorty summarize the paper. \section{Definitions} Let us first introduce some terminology. Graph is a set of $N$ nodes (vertices) connected by $L$ edges (links). A graph need not be connected. It may have many disconnected components including empty nodes (without any link). If a graph has no multiple or self-connected links we shall call it simple graph or graph. An example is illustrated in Fig. \ref{fig:ex0}. Later we shall also discuss graphs with multiple- and self-connections. To distinguish them from simple graphs we shall call them degenerate graphs or pseudographs. \begin{figure}[ht] \psfrag{1}{$1$} \psfrag{2}{$2$} \psfrag{3}{$3$} \psfrag{4}{$4$} \psfrag{5}{$5$}\psfrag{6}{$6$} \includegraphics[width=2.5cm]{mayer_ex.eps} \caption{An example of simple graph with $N=6,L=5$. Vertices without links (like no. 2) are allowed. Each vertex can have at most $N\!-\!1$ links. Positions of vertices in the picture are meaningless. The only information which matters is connectivity.} \label{fig:ex0} \end{figure} One can consider directed or undirected graphs. Directed graphs are built of directed links while undirected of undirected ones. In this paper we shall discuss undirected graphs but the discussion can easily be generalized to directed ones as well. Sometimes we will find it convenient to represent an undirected link as two oriented links going in opposite directions. A simple graph can be represented by its adjacency matrix which is an $N\times N$ matrix whose entries $A_{ij}$ are equal one if there is a link between vertices $i,j$ or zero otherwise. Since self-connections are forbidden we have $A_{ii}=0$ on the diagonal. The adjacency matrix of an undirected graph is also symmetric because if there is a link $i\rightarrow j$ ($A_{ij}=1$), there must be also the opposite one $j\rightarrow i$ ($A_{ji}=1$). In this paper we want to construct statistical ensembles of homogeneous graphs having desired properties. We discuss three types of ensembles: ensemble of graphs with a fixed number of nodes $N$ and varying number of links, ensemble with a fixed number of nodes $N$ and a fixed number of links $L$, and finally ensemble of graphs with a given node degree sequence $\{q_1,q_2,\dots, q_N\}$, which we shall call grand-canonical, canonical and micro-canonical ensembles, respectively. There are of course many other possibilities like for instance ensembles with varying number of nodes, or with a fixed number of loops etc, but the three mentioned above are encountered most frequently. To construct a statistical ensemble, for the chosen set of graphs, we have to specify statistical weight for each graph in the considered set. In the next section using the Erd\"os-R\'{e}nyi graphs and binomial graphs we will deduce a general logical structure standing behind the construction of ensembles of homogeneous graphs and then we will use this structure to introduce ensembles with an arbitrary functional weight which explicitly depends on the node degrees. \section{Statistical ensemble for Erd\"os-R\'{e}nyi random graphs} For simplicity, we start from a well-known model of Erd\"os-R\'{e}nyi's graphs \cite{ref:er,ref:bb}. In this classical model one considers simple graphs with $N$ labeled nodes and $L$ links \footnote{For simple graphs it is immaterial to label links since each link is uniquely determined by its end points.} chosen at random out of all $\binom{N}{2}$ possibilities. All possibilities are equiprobable and so are the corresponding graphs -- understood as graphs whose vertices are labeled. Usually one is interested in unlabeled graphs that is in their shape or topology and not in their labeled version. To explain what is meant by graph's shape or topology, let us consider a simple graph shown in the upper part of Fig. \ref{fig:shape}. Unlabeled graph (topology) on the left hand side of the figure is represented as labeled graphs on the right hand side. \begin{figure}[h] \includegraphics[width=10cm]{shape.eps} \caption{Top: the graph on the left can be realized as three different labeled graphs. A is equivalent to B, C to D and E to F. They are equivalent because they have the same adjacency matrix. Bottom: triangle-shaped graph has only one realization as labeled graph.} \label{fig:shape} \end{figure} There are six possible realizations, but only three of them: A, C, E are distinct. B is the same as A since it can be obtained from A by a continuous deformation: one can continuously move the vertex $2$, together with the link attached to it, to the position of the vertex $3$, and at the same time the vertex $3$ to the position of the vertex $2$. Such a continuous deformation does not change graph's connectivity. The same holds for pairs: C, D and E, F. This can also be seen if we take into account the adjacency matrix $\mathbf{A}$. Both A and B have identical adjacency matrices which are different from those for C, D and E, F: \begin{equation} \mathbf{A}_{\mbox{\scriptsize A}} = \mathbf{A}_{\mbox{\scriptsize B}} = \left( \begin{array}{ccc} 0 & 1 & 1 \\ 1 & 0 & 0 \\ 1 & 0 & 0 \end{array} \right), \quad \mathbf{A}_{\mbox{\scriptsize C}} = \mathbf{A}_{\mbox{\scriptsize D}} = \left( \begin{array}{ccc} 0 & 1 & 0 \\ 1 & 0 & 1 \\ 0 & 1 & 0 \end{array} \right), \quad \mathbf{A}_{\mbox{\scriptsize E}} = \mathbf{A}_{\mbox{\scriptsize F}} = \left( \begin{array}{ccc} 0 & 0 & 1 \\ 0 & 0 & 1 \\ 1 & 1 & 0 \end{array} \right) \ . \end{equation} Now we remove labels from all graphs in Fig. \ref{fig:shape} to obtain two unlabeled graphs depicted on the left hand side. Although there are three distinct adjacency matrices for the upper shape, all of them lead to the same connections between vertices (unlabeled graph). The graph in the lower line in Fig. \ref{fig:shape}. can be labeled only in one way \footnote{ At first glance one can think that there are six ways because one can put labels in six different ways on a drawing of the triangle. After a while one can see that they all are identical since they can be transformed one into another by a transformation which does not change connectivity. For example, if it is a drawing of equilateral triangle one can change labels 123 into 231 by rotating it by 120$^\circ$.} which is represented by the following adjacency matrix: \begin{equation} \mathbf{A}_{\mbox{\scriptsize G}} = \left( \begin{array}{ccc} 0 & 1 & 1 \\ 1 & 0 & 1 \\ 1 & 1 & 0 \end{array} \right) \ . \end{equation} Thus the triangular shape has only one realization as labeled graph. Furthermore, the upper and lower graphs are obviously distinct because none of the corresponding labeled graphs (adjacency matrices) representing the upper graph is equal to that for the lower graph. In this trivial case the difference is in the number of links. More generally, any two graphs are distinct if the underlying labeled graphs (adjacency matrices) cannot be converted one into another by a permutation of nodes' labels. It is clear that for the graphs in Fig. \ref{fig:shape} there is no such a permutation but in general case the comparison of graphs may be a complex problem. Let us now apply the ideas sketched above to define an ensemble of graphs. As an example we shall consider Erd\"os-R\'{e}nyi graphs with $N=4$, $L=3$. It consists of three distinct graphs A, B, C shown in Fig. \ref{fig:shape2}. Now we want to determine the statistical weight for those graphs. Adjacency matrices of the underlying labeled graphs are essentially different for A, B, C since they cannot be converted one into another by a permutation of node's labels. \begin{figure} \includegraphics[width=7cm]{shape2.eps} \caption{Three possible graphs for $N=4,L=3$. The number of ways of labeling these graphs is: $n_A=12$, $n_B=4$, $n_C=4$.} \label{fig:shape2} \end{figure} Each graph in Fig. \ref{fig:shape2} has a few possible realizations as labeled graph. One can label four vertices in $4!=24$ ways corresponding to permutations of $1-2-3-4$. For the graph A, twelve of them give distinct labeled graphs. For example, the permutation $1-2-3-4$ gives an identical labeled graph (adjacency matrix) as $4-3-2-1$. The same kind of symmetry applies for remaining pairs of permutations. Therefore there are $n_A=12$ labeled graphs for $A$. Similarly one can find that there are $n_B=4$ labeled graphs for $B$ and $n_C=4$ for $C$. Altogether, there are $n_A+n_B+n_C=20$ labeled graphs in accordance with $n=\binom{\binom{N}{2}}{L}=20$. In the Erd\"os-R\'{e}nyi ensemble labeled graphs are equiprobable, so the shapes A, B, C have the following probabilities: \begin{equation} p_A=\frac{n_A}{n} = \frac{3}{5}, \quad p_B=\frac{n_B}{n} = \frac{1}{5}, \quad p_C=\frac{n_C}{n} = \frac{1}{5}. \label{pABC} \end{equation} These probabilities give frequencies of the occurrence of the shapes A, B, C in random sampling. We see that graphs (unlabeled graphs) are not equiprobable in Erd\"os-Renyi's ensembles. Let us denote the statistical weights for A, B, C by $w_A,w_B,w_C$ which are proportional to probabilities of configurations in the ensemble. In our case $w_A:w_B:w_C = p_A:p_B:p_C$. There is a common proportionality constant in the weights. It is convenient to choose this constant in such a way that the weight of each labeled graph be $1/N!$ \footnote{One should however remember that this constant is an irrelevant proportionality factor as long as the number of nodes is fixed.}. For this choice we have \begin{equation} w_A=1/2, \quad w_B=1/6, \quad w_C=1/6 , \end{equation} for the graphs in Fig. \ref{fig:shape2}. This choice compensates for the increasing factor of permutations $N!$, when one considers ensembles with varying $N$, and intuitively removes overcounting coming from summing over permutations of indistinguishable node's labels. However, one should remember that in general the number of distinct labeled graphs of a graph is less than $N!$ and therefore the weight of graph is smaller than $1$. The larger is the symmetry of a graph topology the smaller is the number of underlying labeled graphs and thus the smaller is the statistical weight (see for instance Fig. \ref{fig:shape}). The partition function $Z(N,L)$ for the Erd\"os-R\'enyi ensemble can be written in the form: \begin{equation} Z(N,L) = \sum_{\alpha'\in lg(N,L)} \frac{1}{N!} \;\;= \sum_{\alpha\in g(N,L)} w(\alpha) , \label{eq:er} \end{equation} where $lg(N,L)$ is the set of all labeled graphs with given $N,L$ and $g(N,L)$ is the corresponding set of (unlabeled) graphs. The weight $w(\alpha) = n(\alpha)/N!$, where $n(\alpha)$ is the number of labeled graphs of graph $\alpha$. We are interested in quantities averaged over the ensemble. More precisely, we are interested in quantities which depend on topology of graph and not on node's labels. This means that if $O(\alpha)$ is such an observable then for any two labeled graphs $\alpha'_1$ and $\alpha'_2$ of graph $\alpha$ we have $O(\alpha'_1)=O(\alpha'_2)\equiv O(\alpha)$. The average is defined by \begin{equation} \left\langle O\right\rangle \equiv \frac{1}{Z(N,L)} \sum_{\alpha'\in lg(N,L)} O(\alpha') \frac{1}{N!} \;\;= \frac{1}{Z(N,L)} \sum_{\alpha\in g(N,L)} w(\alpha) O(\alpha) . \end{equation} We shall refer to an ensemble with fixed $N,L$ as to a \emph{canonical ensemble}. The word ''canonical'' is used here to emphasize that the number of links $L$ is conserved like the total number of particles in a container with ideal gas remaining in thermal balance with a source of heat. The partition function $Z(N,L)$ can be calculated by pure combinatorics as we have seen in the introduction. Now for completeness we derive it using the adjacency matrix representation of graphs. The adjacency matrices $\mathbf{A}$ are symmetric, they have zeros on the diagonal and $L$ unities above the diagonal. Thus we have \begin{equation} Z(N,L) = \sum_{A_{12}} \sum_{A_{13}} \dots \sum_{A_{1N}} \sum_{A_{23}} \sum_{A_{24}} \dots \sum_{A_{N-1,N}} \delta\left[L-\sum_{p<r} A_{pr}\right] \; 1/N! , \end{equation} where $\delta\left[x\right]=1$ if $x=0$ and zero elsewhere. The sums are done over all $A_{ij}=0,1$ for all pairs $i,j: \, 1\leq i<j\leq N$. Using an integral representation of $\delta\left[ x \right]$ and exchanging the order of summation and integration we obtain the expected result: \begin{eqnarray} Z(N,L) & = & (1/N!) \frac{1}{2\pi} \int_{-\pi}^{-\pi} dk \; e^{ikL} \left( 1+e^{-i k} \right)^{\binom{N}{2}} = (1/N!) \frac{1}{2\pi} \int_{-\pi}^{-\pi} dk \; e^{ikL} \sum_{m=0}^{\binom{N}{2}} \binom{\binom{N}{2}}{m} e^{-ikm} \nonumber \\ & = & (1/N!) \binom{\binom{N}{2}}{L} . \end{eqnarray} This method can be applied to calculate averages of various quantities. As an example consider the node degree distribution $\pi(q)$ which tells us what is the probability that a randomly chosen vertex on random graph has degree $q$: \begin{equation} \pi(q) = \left\langle \frac{1}{N} \sum_i \delta \left[q-q_i\right] \right\rangle . \label{piq} \end{equation} By random graph we mean that we average over graphs from the given ensemble. We know that for Erd\"os-R\'enyi graphs $\pi(q)$ is a Poissonian distribution in the limit of $N\rightarrow\infty$: \begin{equation} \pi(q) = \frac{\bar{q}^q}{q!} \exp(-\bar{q}) , \label{eq:piq} \end{equation} where $\bar{q}=2L/N$ is the average vertex degree. This result can be rederived using the method described above. Let us look at the degree of a vertex labeled by one. The result does not depend on the vertex label for homogeneous graphs since labels have no physical meaning. One can find that \begin{eqnarray} \pi(q) & = & \frac{1}{Z(N,L)} \frac{1}{N!} \sum_{A_{12}} \sum_{A_{13}} \dots \sum_{A_{23}} \sum_{A_{24}} \dots \sum_{A_{N-1,N}} \delta\left[L-\sum_{p<r} A_{pr}\right] \; \delta\left[ q - \sum_{r=2}^N A_{1r} \right] \nonumber \\ & = & \binom{\binom{N}{2}}{L-q} \binom{N-1}{q} / \binom{\binom{N}{2}}{L} . \label{eq:exact_p} \end{eqnarray} which in the limit $\bar{q}=\mbox{const},\,N\rightarrow\infty$ gives Eq. (\ref{eq:piq}) as expected. So far we have discussed the canonical ensemble of Erd\"os-R\'{e}nyi graphs with $N,L$ fixed. Erd\"{o}s and R\'{e}nyi introduced also a related model called {\it binomial model} where the number of nodes $N$ is fixed but the number of links $L$ is not fixed {\em a priori}. One starts from $N$ empty vertices and connects every pair of vertices with a probability $p$. In this statistical ensemble the probability of obtaining a labeled graph with given $L$ is $P(L)=p^L (1-p)^{\binom{N}{2}-L}$. Thus the partition function is \begin{eqnarray} Z(N,\mu) & = & \sum_L \sum_{\alpha\in lg(N,L)} \frac{1}{N!} P(L(\alpha)) = (1-p)^{\binom{N}{2}} \sum_L \left(\frac{p}{1-p}\right)^L \sum_{\alpha\in lg(N,L)} \frac{1}{N!} \nonumber \\ & \propto & \sum_L \exp (-\mu L) \; Z(N,L) \propto \sum_L \exp (-\mu L + S(N,L)) , \label{grand} \end{eqnarray} where $\frac{p}{1-p}\equiv \exp(-\mu)$ or $\mu = \ln \frac{1-p}{p}$, and the entropy $S(N,L)$ is given by Eq. (\ref{S0}). We skipped an $L$-independent factor in front of the sum in the second line substituting equality by proportionality sign. The weight of labeled graphs is $w(\alpha)=1/N! \; \exp(-\mu L(\alpha))$, where $\mu$ is a constant which can be interpreted as chemical potential for links in the \emph{grand-canonical} ensemble (\ref{grand}). One can calculate the average number of links or its variance as derivatives of the grand-canonical partition function with respect to $\mu$: $\langle L \rangle = -\partial_\mu \ln Z(N,\mu)$ and $\langle L^2 \rangle -\langle L \rangle^2 = \partial^2_\mu \ln Z(N,\mu)$. The sum of states can be done exactly: \begin{equation} Z(N,\mu) = \sum_{L=0}^{\binom{N}{2}} e^{-\mu L} \frac{1}{N!} \binom{\binom{N}{2}}{L} = \frac{1}{N!} (1+e^{-\mu})^{\binom{N}{2}} . \end{equation} It is easy to see that for fixed finite $\mu$ the average number of links behaves as $N^2$ or more precisely as \begin{equation} \langle L \rangle = p \frac{N(N-1)}{2} = \frac{1}{1 + e^\mu} \frac{N(N-1)}{2} . \label{Lp} \end{equation} Thus for $N\rightarrow\infty$ the graphs become dense. The mean value of node degree $\langle \bar{q} \rangle =2\langle L \rangle /N$ increases to infinity. The situation changes when $\mu$ goes to infinity with increasing $N$. This happens in particular if the probability $p$ scales as $p \sim 1/N$ since then $\mu$ behaves as $\mu \sim \ln N$. In this case $L$ is proportional to $N$ (\ref{Lp}) and both the terms $\mu L$ and $S$ in the exponent of Eq. (\ref{grand}) behave as $N\ln N$ and compensate each other. The corresponding graphs become sparse and the mean node degree $\langle \bar{q} \rangle =2\langle L \rangle /N$ is now finite. The situation in which $\mu$ scales as $\ln N$ is very different from the situation known from classical statistical physics, where such quantities like chemical potential $\mu$ are intensive and do not depend on system size $N$ in the thermodynamic limit $N\rightarrow\infty$. The difference between canonical and grand-canonical ensembles gradually disappears in the large $N$ limit \cite{ref:bck,ref:dms}. For canonical ensemble or sparse graphs the node degree $\bar{q} = 2L/N = \alpha$ is kept constant when $N\rightarrow \infty$ while in grand-canonical one it may fluctuate around the average $\langle \bar{q} \rangle = 2\langle L\rangle /N = \alpha$. However, the magnitude of fluctuations around the average disappears in the large $N$ limit since \begin{equation} \langle L^2 \rangle - \langle L\rangle^2 = \binom{N}{2} \frac{e^{-\mu}}{(1+e^{-\mu})^2} , \end{equation} and $\Delta q = \sqrt{\langle L^2 \rangle - \langle L\rangle^2} / \langle L \rangle \sim 1/N \rightarrow 0$, so effectively the system selects graphs with $\bar{q} = \alpha$. Sometimes one also considers an ensemble of graphs with a predefined node degree sequence $\{q_1,q_2,\dots,q_N\}$. We shall call it \emph{micro-canonical}. Again, in the simplest case one assumes that all labeled graphs are equiprobable in this ensemble. Properties of random graphs in such an ensemble strongly depend on the degree sequence. \section{Weighted homogeneous graphs} In the previous section we described ensembles for which all labeled graphs had the same statistical weight. Random graphs in such ensembles have well known properties. It turns out, however, that most of these properties do not correspond to those observed for real world networks. One needs a more general set-up to define an ensemble of complex random networks. Such a set-up can be introduced as follows. One considers the same set of graphs as in Erd\"os-R\'enyi model but one ascribes to each graph a different statistical weight. In other words, one chooses a probability measure on the set of graphs which differs from the uniform measure. In the generalized ensemble, each graph in addition to the configuration space weight $1/N!$ has a functional weight $W(\alpha)$. For homogeneous random graphs this weight is assumed to depend only on graph topology. This means that the weight does not depend on nodes' labels: if $\alpha'_1$ and $\alpha'_2$ are labeled graphs of $\alpha$ then $W(\alpha'_1)=W(\alpha'_2)\equiv W(\alpha)$. The partition function for a weighted canonical ensemble reads \begin{equation} Z(N,L) = \sum_{\alpha'\in lg(N,L)} (1/N!)\, W(\alpha') \;\;= \sum_{\alpha\in g(N,L)} w(\alpha) W(\alpha) , \label{eq:can_g} \end{equation} where $w(\alpha)$ is the same factor $w(\alpha) = n(\alpha)/N!$ as before (\ref{eq:er}), being just the ratio of the number of labeled graphs $\alpha'$ of $\alpha$ (obtained by permutations of nodes' labels giving distinct adjacency matrices) and the number of all nodes' labels permutations $N!$. For Erd\"os-R\'enyi graphs the functional weight is $W(\alpha)=1$. The simplest non-trivial example is a family of product weights $W$: \begin{equation} W(\alpha) = \prod_{i=1}^N p(q_i) , \label{eq:prod} \end{equation} where $p(q)$ is a positive function depending on one node degree $q$. This functional weight does not introduce correlations between node degrees. We shall refer to random graphs generated by this partition function as \emph{uncorrelated networks}. One should however remember that the total weight does not entirely factorize because the configuration space weight $w(\alpha) = n(\alpha)/N!$ written as a function of node degrees $w(q_1,q_2,\dots,q_N)$ does not factorize. There is also another factor which prevents the model from the full factorization and independence of node degrees, namely this is the total number of links $2L = q_1+q_2+\dots + q_N$ which for given $L$ and $N$ introduces correlations between $q_i$'s. For example, if one of $q_i$'s is large, say of order $2L$, then the remaining ones have to be small in order not to violate the constraint on the sum. For a wide class of weights $p(q)$ one can however show that in the large $N$ limit the probability that a randomly chosen graph has degrees $q_1,\dots,q_N$ approximately factorizes: \begin{equation} \pi(q_1,\dots,q_N) \sim \prod_{i=1}^N \pi(q_i) . \label{fact} \end{equation} For large $N$, the node degree distribution $\pi(q)$ (\ref{piq}), that is the probability that a random node on random graph has degree $q$ \cite{ref:bck,ref:bjk1,ref:bl1}, can be approximated by \begin{equation} \pi(q) = \frac{p(q)}{q!} \exp(-Aq-B) , \label{piqAB} \end{equation} where parameters $A,B$ are determined from the conditions for the normalization $\sum_q \pi(q)=1$ and for the average $\sum_q q \pi(q) = \bar{q} \equiv 2L/N$. For example, for $p(q)=1$ which corresponds to Erd\"os-R\'{e}nyi graphs one finds $A = -\ln \bar{q} = \ln 2L/N$ and $B = \bar{q} = 2L/N$, therefore $\pi(q)$ is given by the Poissonian from Eq. (\ref{eq:piq}). Since the node degree distribution $\pi(q)$ for weighted graphs (\ref{eq:prod}) depends on $p(q)$, one can choose the latter to obtain a desired form of the node degree distribution $\pi(q)$. Let $\pi(q)$ be a desired node degree distribution such that \begin{equation} \sum_q \pi(q)=1 \ , \quad \bar{q} = \sum q \pi(q) . \end{equation} If we choose the weight (\ref{eq:prod}) with \begin{equation} p(q) = q! \pi(q) \end{equation} in canonical ensemble with $N$ nodes and $L$ links, in the limit of $N\rightarrow \infty$ and $2L/N = \bar{q}$ we obtain homogeneous random graphs with this node degree distribution. In this case the constants $A$, $B$ from Eq. (\ref{piqAB}) vanish automatically: $A=B=0$. In particular by an appropriate choice of $p(q)$ we can generate scale free graphs with the node degree Barab\'{a}si - Albert distribution \cite{ref:bamodel}: $\pi(q) = \frac{4}{q(q+1)(q+2)}$ for $q=1,2,\dots$ and $\pi(0)=0$ as an ensemble of graphs $L=N$, $\bar{q} = 2$ with $p(q) = q! \frac{4}{q(q+1)(q+2)}$ for $q=1,2,\dots$ and $p(0)=0$. However, for finite $N$ the node degree distribution $\pi(q)$ deviates from the limiting shape due to finite size corrections, which are particularly strong for fat tailed distributions $\pi(q) \sim q^{-\gamma}$. The maximal node degree scales as $q_{max}\sim N^{1/(\gamma-1)}$ for $\gamma\ge 3$ and as $q_{max} \sim N^{1/2}$ for very fat tails: $2 < \gamma < 3$ \cite{ref:bk,ref:bpv} as a result of structural constraints which also lead to the occurrence of correlations between node degrees. One can define more complicated weights than those given by Eq. (\ref{eq:prod}). A natural candidate for networks with degree-degree correlations is the following weight \cite{ref:bl1,ref:b}: \begin{equation} W(\alpha) = \prod_{l=1}^L p(q_{a_l},q_{b_l}) , \end{equation} where the product runs over all links of the graph, and the weight $p(q_a,q_b)$ is a symmetric function of degrees of nodes at the end points of the link. One can choose this function to favor assertive or disassertive behavior \cite{ref:bl1,ref:b,ref:n3,ref:bp,ref:d2}. In a similar way one can introduce probability measures on the set of graphs which mimic some other functional properties of real networks, like for example higher clustering \cite{ref:n2,ref:bjk2,ref:bjk3,ref:d1,ref:pn2}. One can do this in micro-canonical, canonical, grand-canonical or any other ensemble. This is just the most general set-up to handle homogeneous networks. \section{Monte-Carlo generator of homogeneous networks} Erd\"os-R\'enyi graphs are exceptional in the sense that one can calculate for them almost all quantities of interest analytically. This is not the case for weighted networks. Various methods have been proposed for generating random graphs \cite{ref:k}. In this section we will describe a Monte-Carlo method which allows one to study a wide class of random weighted graphs experimentally by a sort of numerical experiments. The basic idea behind this type of experiments is to sample the configuration space of graphs with the probability proportional to the statistical weight or in other words to generate graphs with a desired probability. Again, the Erd\"os-R\'enyi graphs are exceptional because one can generate them one by one independently of each other. This is just because they are equiprobable. For weighted graphs the situation is not that easy since there are no efficient algorithms to pick up an element from a large set with the given probability. The naive algorithm which relies on picking up an element uniformly and then accepting it with the given probability has a very low acceptance rate. Therefore one has to use another idea. We will describe below how to generate graphs using dynamical Monte-Carlo technique. The idea is to run a random walk process in the set of graphs which visits configurations with a frequency proportional to their statistical weight. Mathematically, this means that one has to invent a stationary Markov chain (process) for which the stationary distribution is proportional to the statistical weights of graphs: $\sim W(\alpha)/Z$. The Markov chain is defined by transition probabilities $P(\alpha\rightarrow \beta)$ that the random walker will go in one step from a configuration (graph) $\alpha$ to $\beta$. The probabilities are stored in a transition matrix $\mathbf{P}$: $P_{\alpha\beta} \equiv P(\alpha\rightarrow \beta)$ which is also called Markov's matrix. For a stationary process, the transition matrix $\mathbf{P}$ is constant during the random walk. Random walk is initiated from a certain graph $\alpha_0$ and then elementary steps are repeated producing a sequence (chain) of graphs $\alpha_0 \rightarrow \alpha_1 \rightarrow \alpha_2 \rightarrow \dots \ $. The probability $p_\beta(t+1)$ that a graph $\beta$ will be generated in the $(t+1)$-th step of the Markov process can be calculated as \begin{equation} p_\beta(t+1) = \sum_\alpha p_\alpha(t) P_{\alpha \beta} . \end{equation} The last equation can be written as \begin{equation} \mathbf{p} (t+1) = \mathbf{P}^{\tau} \mathbf{p} (t) , \label{interp} \end{equation} where $\tau$ denotes transposition, and $\mathbf{p}$ is a vector of elements $p_\alpha$. One should note that the stationary state: $\mathbf{p}(t+1) = \mathbf{p}(t)$ corresponds to a left eigenvector of $\mathbf{P}$ to the eigenvalue \footnote{One can show that all eigenvalues of a Markov transition matrix lie inside or on the unit circle $|\lambda| \le 1$.} $\lambda=1$. If the process is ergodic, which means that any configuration can be reached by a sequence of transitions starting from any initial configuration, and if the transition matrix fulfills the detailed balance condition: \begin{equation} W_\alpha P_{\alpha \beta} = W_\beta P_{\beta \alpha} \quad \forall \alpha,\beta , \label{eq:db} \end{equation} then the stationary state can be shown to approach the desired distribution: $p_\alpha(t) \rightarrow W_\alpha/Z$ for $t \rightarrow \infty$. We used a short-hand notation $W_\alpha$ for $W(\alpha)$. In other words, when the length of the Markov chain becomes infinite the probability of occurrence of graphs in the Markov chain becomes proportional to their statistical weights and becomes independent of the initial configuration. Therefore the average over graphs generated in this Markovian random walk is a good estimator of the average over the weighted ensemble. The price to pay for generating graphs in this way is that the consecutive graphs in the Markov chain may be correlated with each other. Therefore one has to find a minimal number of steps for which one can treat measurements on such graphs as independent. One should note that the only characteristics of the Markov process which has a physical meaning from the point of view of the simulated ensemble is the stationary distribution. All other dynamical properties of the random walk which are encoded in the form of transition matrix $P(\alpha\rightarrow\beta)$ are irrelevant. Many different transition matrices $\mathbf{P}$ may have the same stationary distribution. Indeed, many of them fulfill the detailed balance condition for given weights $W_\alpha$ (\ref{eq:db}). The best known choice of $\mathbf{P}$ is \begin{equation} P_{\alpha\beta} = \min\left\{1,\frac{W_\beta}{W_\alpha}\right\} . \label{metrop} \end{equation} This choice is quite general and can be used in many different contexts. It is called Metropolis algorithm. For the current configuration $\alpha$ one proposes a change to a new configuration $\beta$ which differs slightly from $\alpha$ and then one accepts it with the Metropolis probability (\ref{metrop}). One repeats this many times producing a chain of configurations. The proposed modifications should not be too large since then the acceptance rate would be small. Therefore the algorithm makes only small steps (moves) in the configuration space which form a sort of weighted random walk. \section{Monte-Carlo generator of canonical ensemble} Now, we want to apply this method to generate Erd\"os-R\'{e}nyi graphs. Let us begin with the canonical ensemble with $N,L$ fixed. A good candidate for elementary transformation of graph is rewiring of a link as shown in Fig. \ref{fig:tmove}, because it does not change $N$ and $L$. As mentioned before it is convenient to introduce a representation in which each undirected link is represented by two directed links. The rewiring is done in two steps \cite{ref:bck}. First we choose a directed link $ij$ and a vertex $k$ at random. Then we rewire the link $ij$ to $ik$. If there is already a link between $i$ and $k$ or if the vertex $k$ coincides with $i$, we reject the rewiring since it would otherwise lead to a multiple- or self-connection. One should note that the result of rewiring $ij$ is not the same as of rewiring $ji$. \begin{figure}[h] \psfrag{i}{$i$} \psfrag{j}{$j$} \psfrag{k}{$k$} \includegraphics[width=13cm]{tmove.eps} \caption{The idea of rewiring: a random link (dotted line) is rewired from vertex $j$ to a random vertex $k$ (left hand side). Alternatively (right hand side) a random, oriented link (dotted line) is rewired from vertex of its end $j$ to a random vertex $k$. The opposite link $j\rightarrow i$ is simultaneously rewired.} \label{fig:tmove} \end{figure} The move is accepted with the Metropolis probability. For the canonical ensemble of Erd\"os-R\'enyi graph this probability is equal to one since functional weights are $W_\alpha=W_\beta=1$ in Eq. (\ref{metrop}). Let us see how rewiring transformations work in practice. Consider as an example the set of graphs shown in Fig. \ref{fig:shape2}. If we pick up the link $3-2$ in the graph A and rewire it to $3-1$, we will obtain the graph B. If we rewire the link $2-3$ to $2-4$, we will get the graph C. So using the procedure of rewiring showed in Fig. \ref{fig:tmove} we can obtain every graph in the ensemble. The rewiring transformation is ergodic in this set of graphs. To summarize, our procedure of generating graphs in this training ensemble looks as follows. We construct an arbitrary graph having $N=4$ nodes and $L=3$ links to initiate the procedure and then we repeat iteratively rewirings for randomly chosen links and vertices. The only restriction is that the rewirings do not produce self- or multiple-connected links. We keep on repeating until we obtain ''thermalized graphs''. Then we can begin measuring quantities on the generated sequence of random graphs. Let us check that the described Monte-Carlo procedure indeed generates graphs with the expected probabilities (\ref{pABC}). Let us calculate the Markovian matrix $\mathbf{P}$ for the rewiring procedure in this ensemble. First we calculate the transition probability from A to B. The graph A can be converted into B in one step if we rewire the link ''b'' in Fig. \ref{fig:tmove_ex} to the vertex 2, or alternatively the link ''e'' to the vertex 1. We see that for this change we can choose two out of six links and one of four vertices to obtain the desired change. Thus the probability of choosing links is $2/6$ and of choosing correct vertex is $1/4$, so the total probability is $P(A\rightarrow B)=2/6 \cdot 1/4 = 2/24$. Let us now calculate $P(A\rightarrow C)$. To obtain $C$ from $A$ we have to rewire ''a'' to 3 or ''f'' to 4. Thus $P(A\rightarrow C)=2/6 \cdot 1/4 = 2/24$. We can find $P(A\rightarrow A)$ from the condition: $P(A\rightarrow A)+P(A\rightarrow B)+P(A\rightarrow C)=1$. This yields $P(A\rightarrow A)=20/24$. \begin{figure}[ht] \includegraphics[width=3cm]{tmove_ex.eps} \caption{The representation of graph A in Fig. \ref{fig:shape2} as directed graph. } \label{fig:tmove_ex} \end{figure} Repeating the calculations for the remaining cases we find: $P(B\rightarrow A)=6/24$, $P(B\rightarrow B)=18/24$, $P(B\rightarrow C)=0$ and $P(C\rightarrow A)=6/24$, $P(C\rightarrow B)=0$, $P(C\rightarrow C)=18/24$. The results can be collected in a transition matrix: \begin{equation} \mathbf{P} = \frac{1}{24} \left[\begin{array}{ccc} 20 & 2 & 2 \\ 6 & 18 & 0 \\ 6 & 0 & 18 \end{array}\right]. \end{equation} We can now determine the stationary probability distribution of the Markov process as a left eigenvector to the eigenvalue one of the transition matrix $\mathbf{P}$. We obtain $p_A:p_B:p_C = 3:1:1$ in agreement with Eq. (\ref{pABC}). This is not surprising since $\mathbf{P}$ satisfies the detailed balance condition (\ref{eq:db}) and the corresponding changes are ergodic. We have checked above by explicit calculation that the algorithm gives correct weights of Erd\"os-R\'enyi graphs for $N=3,L=4$. One can give a general argument that for any $N,L$ the algorithm generates labeled graphs which are equiprobable. Suppose that we have a certain labeled graph $\alpha$ and want to get $\beta$ by rewiring $ij$ to $ik$. (see Fig. \ref{fig:tmove}). The total probability $P(\alpha\rightarrow\beta)$ can be written as a product of two factors: the probability $P_c$ of choosing a particular candidate for a new configuration and the probability $P_a$ of accepting it. Because we choose a link $i\rightarrow j$ from $2L$ possible directed links and a vertex $k$ from $N$ vertices we have $P_c = 1/(2LN)$. Inserting $P(\alpha'\rightarrow\beta') = 1/(2LN) \, P_a(\alpha'\rightarrow\beta')$ in the Eq. (\ref{eq:db}) and similarly for $\alpha'\leftrightarrow \beta'$ we get \begin{equation} w_{\alpha'} P_a(\alpha'\rightarrow\beta') = w_{\beta'} P_a(\beta'\rightarrow\alpha') . \label{eq:can1} \end{equation} But $w_{\alpha'}=1/N!$ for all labeled graphs, thus $P_a(\alpha'\rightarrow\beta')=P_a(\beta'\rightarrow\alpha')$. This means that every move should be accepted unless it violates the multiple- or self-connections constraints. The rejection does not change the frequency of the occurrence of simple graphs but only restricts the space of sampled graphs to what we need. The weights of (unlabeled) graphs $\alpha$ are in this case $w(\alpha)= n(\alpha)/N!$ where $n(\alpha)$ is the number of distinct graphs of $\alpha$. \begin{table}[h] \begin{center} \begin{tabular}{|c|c|c|c|c|c|c|} \hline Graphs & \multicolumn{6}{|c|}{\includegraphics[width=7cm]{pent_ex.eps}} \\ \hline $w_\alpha$ & 1/2 & 1/2 & 1/12 & 1/2 & 1/8 & 1/24 \\ $p_\alpha$ & 0.286 & 0.286 & 0.048 & 0.286 & 0.071 & 0.024 \\ \hline \multicolumn{7}{|c|}{Simulations} \\ \hline R & \multicolumn{6}{|c|}{} \\ \hline 10 & 0.285(1) & 0.286(1) & 0.048(1) & 0.286(1) & 0.071(1) & 0.024(1) \\ 5 & 0.286(1) & 0.285(1) & 0.047(1) & 0.286(1) & 0.071(1) & 0.024(1) \\ 2 & 0.285(1) & 0.285(1) & 0.047(1) & 0.286(1) & 0.072(1) & 0.024(1) \\ 1 & 0.284(1) & 0.286(1) & 0.048(1) & 0.284(1) & 0.072(1) & 0.025(1) \\ \hline \end{tabular} \caption{Theoretically calculated weights $w_\alpha$ of graphs in the canonical ensemble $N=5,L=4$ are normalized to ensure probabilistic interpretation: $p_\alpha = w_\alpha/\sum_\beta w_\beta$, and compared with the experimental frequencies in the Markov chain using algorithm of rewiring. The results differ by the number $R$ of rewirings between consecutive measurements. } \label{tab1} \end{center} \end{table} Let us numerically test the algorithm. In table \ref{tab1} we compare the weights calculated analytically and computed from Monte-Carlo generated graphs for $N=5,L=4$. There are six different graphs in this ensemble. We generated $10^6$ graphs. Each of them was obtained from the previous one by $R$ rewirings, or more precisely by $R$ attempts of rewiring \footnote{ Even if a rewiring is rejected we count it in since it corresponds to the transition $P(\alpha \rightarrow \alpha)$ from graph to the same graph.}. As we can see in table \ref{tab1}, the frequency of occurrence of each graph is in an excellent agreement with the expected weights. The results do not depend on the separation $R$ between the measurements. In the chain of $10^6$ graphs, each graph in this ensemble is produced many times. For larger ensembles the algorithm would not be able to visit all graphs since the number of graphs is very large (\ref{S1}). In this case the algorithm would choose only those graphs which are most representative. To make sure that the algorithm has reached the stationary distribution one should start a couple of random walks from different corners of the configuration space and run the algorithm so long as the statistical properties of graphs generated in all the random walks become identical. Generalization of the algorithm to a weighted ensemble is straightforward. We insert the statistical weights $W_\alpha$ of this ensemble into the Metropolis formula (\ref{metrop}). Consider in particular a product weight (\ref{eq:prod}). We choose a link $ij$ and a vertex $k$ on the current configuration $\alpha$ at random and attempt to rewire the link to $ik$ to obtain a new configuration $\beta$. The change is accepted with the probability \begin{equation} P_a(\alpha\rightarrow\beta) = \min\left\{1,\frac{W_\beta}{W_\alpha}\right\} = \min\left\{1,\frac{p(q_j-1)p(q_k+1)}{p(q_j)p(q_k)}\right\} . \end{equation} The degrees $q_j,q_k$ are taken from $\alpha$. Clearly the rewiring changes the degrees $q_j\rightarrow q_j-1$, $q_k\rightarrow q_k+1$, and leaves the degrees of remaining nodes intact. The ratio $W_\beta/W_\alpha$ can be calculated for any form of statistical weights, so the algorithm is general. \section{Monte-Carlo generator of grand-canonical ensemble} The rewiring procedure described in the previous section does not change $N$ and $L$. If we want to simulate graphs from a grand-canonical ensemble for which $L$ is variable, we have to supplement the set of elementary transformations in the algorithm by transformations which change the number of links. We can introduce two mutually reciprocal transformations: adding and deleting a link. Both they preserve the number of nodes $N$ but change the number of links: $L \rightarrow L\pm 1$. The two transformations must be carefully balanced. On a given graph $\alpha$ we have to choose one of them. Let the link addition be selected with the probability $p_+$ and the removal with $p_-$. Once the move is selected we have to choose a link-candidate for which the move is to be applied. It is convenient to split the total transition probability into three factors: \begin{equation} P(\alpha\rightarrow \beta) = p_\pm P_c(\alpha\rightarrow \beta) P_a(\alpha\rightarrow\beta) , \end{equation} where $p_\pm$ stands for one of $\{p_-,p_+\}$, $P_c(\alpha\rightarrow \beta)$ for the probability of choosing a candidate configuration for the change and $P_a(\alpha\rightarrow \beta)$ for the probability of accepting the move. Let $\alpha$ and $\beta$ be two graphs which differ by a link which is present on $\beta$ but absent on $\alpha$: $L(\alpha)=L(\beta)-1$. The transition probability for adding a link to $\alpha$ has to be balanced with the probability of removing the link from $\beta$. In order to add a link to $\alpha$ we have to choose two vertices to which the addition of a link is attempted. The probability of choosing a given pair of vertices, if we choose two vertices independently, is $P_c(\alpha\rightarrow\beta)=2/N^2$. Thus the total probability of this move is \begin{equation} P_{\alpha\beta}= P(\alpha\rightarrow\beta) = p_+ \, \frac{2}{N^2} \, P_a(\alpha\rightarrow\beta) . \end{equation} In the reciprocal transformation we have to choose this link among all links. The probability of choosing one among $L$ links is $P_c(\beta\rightarrow\alpha)=1/L_\beta=1/(L_\alpha+1)$. Thus the total probability of this move is \begin{equation} P_{\beta\alpha}= P(\beta\rightarrow\alpha) = p_- \, \frac{1}{L_\beta} \, P_a(\beta\rightarrow\alpha) . \end{equation} Now we have to insert the last two equations to the detailed balance condition which for the grand-canonical ensemble additionally includes the factor $e^{-\mu L}$: \begin{equation} W_\alpha e^{-\mu L_\alpha} P_{\alpha \beta} = W_\beta e^{-\mu L_\beta} P_{\beta \alpha} \ . \label{eq:db_gc} \end{equation} Using this we can calculate the ratio \begin{equation} \frac{P_a(\alpha\rightarrow\beta)}{P_a(\beta\rightarrow\alpha)} = \exp(-\mu) \, \frac{p_-}{p_+} \, \frac{N^2}{2L_\beta} \, \frac{W_\beta}{W_\alpha} . \end{equation} If one chooses the same number of attempts for adding and removing a link: $p_+=p_-$, then the ratio $p_+/p_-=1$ will disappear from the last equation and the acceptance probabilities for adding or removing a link in the Metropolis algorithm will read \begin{equation} P_a(\alpha\rightarrow\beta) = \min\left\{1, \exp(-\mu) \, \frac{N^2}{2(L_\alpha+1)} \, \frac{W_\beta}{W_\alpha} \right\} , \label{eq:gcan_rown} \end{equation} and \begin{equation} P_a(\beta\rightarrow\alpha) = \min\left\{1, \exp(+\mu) \, \frac{2L_\beta}{N^2} \, \frac{W_\alpha}{W_\beta}\right\} , \label{eq:gcan_rown2} \end{equation} respectively. As before if we want to produce only simple graphs we must have an additional condition which eliminates moves leading to self- or multiple connections. The algorithm is complete. One should note that there is no reason to do additional rewirings because a rewiring of a link $ij$ to a link $ik$ is equivalent to removing the link $ij$ and adding $ik$. In principle one could propose other algorithms. For example, one could consider a modified algorithm in which the move removing a link is done in a different way. Instead of picking up a link as a candidate, one could pick up two vertices at random, and then remove a link if there is any between them. The probability of choosing a pair of vertices would be $2/N^2$ and it would cancel with the identical factor for the probability of choosing candidates in the move adding a link. The fractions $N^2/2L$ and $2L/N^2$ would in this case disappear from equations (\ref{eq:gcan_rown}) and (\ref{eq:gcan_rown2}). The two algorithms of course generate the same ensemble. However, the modified algorithm would have much worse acceptance rate for sparse networks since the chance that there is a link between two randomly chosen vertices on a sparse graph is very small. Most of the chosen pairs of vertices are not connected by a link and therefore the algorithm will do nothing since there is no link to remove. This problem is absent for the algorithm which we described previously since in that case only existing links are chosen as candidates for removal. One can easily estimate that the probability of accepting a link removal (\ref{eq:gcan_rown2}) is not very small. Indeed, even for sparse graph the factor $e^{\mu} 2L/N^2$ in Eq. (\ref{eq:gcan_rown2}) is of order unity. In this case both $\exp(\mu)$ and $L$ for large $N$ grow proportionally to $N$ and their product balances the factor $N^2$ in the denominator. The algorithm has a finite acceptance rate which does not vanish when the system size grows. As an exercise, let us consider an example of unweighted ($W_\alpha=1$) graphs with $N=3$. This ensemble consists of four graphs shown in table \ref{tab2}. Their statistical weights can be easily found to be $1/3!, \, 3e^{-\mu}/3!, \, 3e^{-2\mu}/3!, \, e^{-3\mu}/3!$, so we expect that the frequency of occurrence in random sampling should be $1:3e^{-\mu}:3e^{-2\mu}:e^{-3\mu}$. As we see in table \ref{tab2}, the results of Monte-Carlo simulations are in perfect agreement with this expectation. \begin{table}[h] \begin{center} \begin{tabular}{|c|c|c|c|c|c|} \hline & Graphs & \multicolumn{4}{|c|}{\includegraphics[width=5cm]{gr_3.eps}} \\ \hline $\mu = 0$ & Theor. & \ \ 0.125 \ \ & \ \ 0.375 \ \ & \ \ 0.375 \ \ & \ \ 0.125 \ \ \\ & Exp. & 0.125(1) & 0.374(1) & 0.374(1) & 0.126(1) \\ \hline $\mu = 0.2$ & Theor. & 0.166 & 0.408 & 0.334 & 0.091 \\ & Exp. & 0.166(1) & 0.408(1) & 0.334(1) & 0.091(1) \\ \hline $\mu = 0.5$ & Theor. & 0.241 & 0.439 & 0.266 & 0.054 \\ & Exp. & 0.241(1) & 0.439(1) & 0.266(1) & 0.054(1) \\ \hline \end{tabular} \caption{Comparison of the probability distribution of graphs in a grand-canonical ensemble with $N=3$ nodes: calculated theoretically and computed in a run of Monte-Carlo simulation in which $10^6$ graphs were generated.} \label{tab2} \end{center} \end{table} One can easily apply this technique to any form of statistical weights. In particular we can consider the product weights (\ref{eq:prod}). The probability of accepting a new configuration by adding or removing a link between $ij$ reads \begin{eqnarray} \mbox{min} \left\{ 1, \frac{N^2}{2(L+1)} \exp(-\mu) \frac{p(q_i+1)p(q_j+1)}{p(q_i)p(q_j)} \right\} & \mbox{for adding a link,} \nonumber \\ \mbox{min} \left\{ 1, \frac{2L}{N^2} \exp(+\mu) \frac{p(q_i-1)p(q_j-1)}{p(q_i)p(q_j)} \right\} & \mbox{for deleting a link,} \nonumber \end{eqnarray} where $L$ and $q_i,q_j$ refer to the current configuration. \section{Monte-Carlo generator of micro-canonical ensemble} Another frequently encountered ensemble is an ensemble of graphs which have a given node degree sequence $\{q_1,q_2,\dots,q_N\}$. The partition function $Z$ has the form: \begin{equation} Z(N,\left\{ q_i \right\}) = \sum_{\alpha'\in lg(N,L)} \left( \prod_{i=1}^N \delta \left[ q_i(\alpha') - q_i\right] \right) \, 1/N! \, W(\alpha') , \end{equation} where the product of delta functions allows one to include only those graphs which have a prescribed degree distribution $q_i$. As before the factor $1/N!$ is fixed in this ensemble and could in principle be skipped. The canonical partition function $Z(N,L)$ is related to the micro-canonical ones: \begin{equation} Z(N,L) = \sum_{q_1=0}^\infty \dots \sum_{q_N=0}^\infty Z(N,\left\{ q_i \right\}) \, \delta \left[ q_1+q_2+\dots +q_N - 2L \right] . \end{equation} To generate graphs from micro-canonical ensemble one has to have a Markov process preserving node degrees. The main idea is to combine simultaneous rewirings \cite{ref:msz} as shown in Fig. \ref{fig:xmove}. We shall call this combination ''X-move''. \begin{figure}[h] \psfrag{i}{$i$} \psfrag{j}{$j$} \psfrag{k}{$k$} \psfrag{l}{$l$} \includegraphics[width=5cm]{xmove.eps} \caption{The idea of ''X-move'': two oriented links (dotted lines) $ij$ and $kl$ chosen in a random way are rewired, exchanging their endpoints. Then the opposite links (solid lines) are also rewired. } \label{fig:xmove} \end{figure} At each step one picks up two random links: $ij$ and $kl$, and rewires them to $il$ and $kj$. This procedure is ergodic, i.e. it explores the whole configuration space. Such a transformation was discussed in \cite{ref:msz} where it was used to randomize graphs with a given nodes' degree sequence. In that case the functional weight was $W_\alpha=1$ and rewirings were done with probability equal to one. In general case if one considers non-trivial $W_\alpha$, one has to accept the change with a corresponding Metropolis probability (\ref{metrop}). In this way one can for example generate graphs whose statistical weights depend on the number of triangles. In a sense one can perform a weighted randomization of networks with the given node degree sequence. Introducing a weight into randomization may be very important in the construction of scoring functions in problems of motif searching \cite{ref:m,ref:bl2,ref:ia}. If one tries to determine relations between structural motifs and the functionality of network, it is very important to properly construct scoring function which may clearly account for the existence of a particular subgraph on a network and its function. Scoring functions are usually measured as a sort of distance between a network which displays a certain function and a random network which does not. An important problem in such studies is how to construct those networks which should serve as the background reference. The simplest idea is to use networks obtained by uniform randomization. This may however introduce some bias and may be misleading. Imagine for example that a motif which is responsible for a certain network function is built out of a couple of triangular loops and that triangular loops alone have no function. It is clear that one would like to control the abundance of triangular loops to distinguish between specific motifs and motifs which are more frequent by pure chance just because of higher abundance of triangles. Therefore it may be important to control the number of triangles in the randomized reference networks used in the scoring function. It was just an example, but in general case it might be useful to perform a weighted randomization taking into account some desired features of reference networks. \section{Graph generator and adjacency matrices} All the elementary transformations: rewiring, adding or removing a link, and the X-move have a simple representation in terms of adjacency matrices. Rewiring relies on picking up at random an element $A_{ij} = 1$ of the adjacency matrix and flipping it with an element $A_{ik}=0$ so that after the move $A_{ij} = 0$ and $A_{ik}=1$. For undirected links adjacency matrices are symmetric and therefore at the same time one has to flip $A_{ji} = 1$ and $A_{ki}=0$. To add a vertex one chooses at random $A_{ij}$ and if $A_{ij}=0$ and $i\ne j$, one changes it into $A_{ij}=1$ (and for $A_{ji}$ correspondingly). To remove a link one picks up a non-vanishing element $A_{ij}=1$ and substitutes it with $A_{ij}=0$. To perform X-move one picks up two non-vanishing elements of $\mathbf{A}$ at random, say $A_{ij}=1$ and $A_{kl}=1$, and flips them simultaneously with $A_{il}=0$ and $A_{kj}=0$ to: $A_{ij}=0$, $A_{kl}=0$, $A_{il}=1$ and $A_{kj}=1$. Of course one also flips their four symmetric counterparts. In practice, when simulating sparse graphs one does not use the matrix representation since it would require $N^2$ storage capacity. For sparse matrices the number of non-vanishing matrix elements is proportional to $N$ and one can use a linear storage structure. It directly corresponds to the underlying graph structure. Using linear storage one can code graphs having of order $10^6$ nodes or even more on a PC. \section{Degenerated graphs (pseudographs)} In previous sections we described ensembles of simple graphs. Let us now discuss pseudographs that is graphs which may have multiple- and self-connections. A degenerate undirected pseudograph can be represented by a symmetric adjacency matrix $\mathbf{A}$ whose off diagonal entries $A_{ij}$ count the number of links between vertices $i$ and $j$, and the diagonal ones $A_{ii}$ count twice the number of self-connecting links attached to vertex $i$. For example, the graph depicted in Fig. \ref{fig:pseudo} has the following adjacency matrix: \begin{equation} \mathbf{A} = \left( \begin{array}{cccccc} 0 & 0 & 1 & 0 & 0 & 1 \\ 0 & 2 & 0 & 0 & 0 & 0 \\ 1 & 0 & 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 & 0 & 0 \\ 0 & 0 & 0 & 0 & 0 & 2 \\ 1 & 0 & 0 & 0 & 2 & 0 \end{array} \right). \label{eq:pseudo} \end{equation} \begin{figure}[h] \psfrag{1}{$1$} \psfrag{2}{$2$} \psfrag{3}{$3$} \psfrag{4}{$4$} \psfrag{5}{$5$} \psfrag{6}{$6$} \includegraphics[width=7cm]{pseudo.eps} \caption{Left hand side: the example of pseudograph with $N=6,L=5$. Right hand side: its representation as directed graph.} \label{fig:pseudo} \end{figure} In the representation where each undirected link is represented as two opposite directed links all matrix elements including the diagonal ones count the number of directed links emerging from the vertex. As before let us first consider labeled pseudographs. However, in order to have a unique representation of a graph one has to label links as well. We did not have to do this for simple graphs since in that case each link was uniquely determined by its endpoints. It is not anymore the case for degenerate graphs since there may be more than one link between two nodes. A pseudograph with $N$ nodes and $L$ links can be fully labeled by $N$ node labels and $2L$ labels of directed links. Each fully labeled graph has thus the configuration space weight equal to $1/(N!(2L)!)$. Let us work out the consequences of this choice. Denote $\alpha$ a graph, $\alpha'$ a labeled graph of $\alpha$ with labeled nodes only, and $\alpha''$ a fully labeled graph of $\alpha$ with labeled nodes and labeled links. From here on labeled graph means a graph which has only labels on nodes while \emph{fully labeled graph} a graph which has additionally labels on directed links. The configuration space weight of $\alpha$ can be calculated as a sum over all fully labeled graphs $\alpha''$ as follows: \begin{equation} w_\alpha = \sum_{\alpha'' \in flg(\alpha)} \frac{1}{N!(2L)!} = \sum_{\alpha' \in lg(\alpha)} \frac{1}{N!} \left( \prod_{i} \frac{1}{2^{A_{ii}/2} \left( A_{ii}/2\right)!} \right) \prod_{i>j} \frac{1}{A_{ij}!} , \label{wp} \end{equation} where $flg(\alpha)$ denotes the set of fully labeled graphs of graph $\alpha$, $lg(\alpha)$ the set of labeled graphs (labeled nodes only) of graph $\alpha$ . The expression $A_{ii}/2$ counts the number of self-connecting links attached to vertex $i$, and $A_{ij}$ is the multiplicity of links connecting $i$ and $j$. If there are no self-connections ($A_{ii}=0$) and no multiple connections ($A_{ij}\le 1$), the configuration space weight (\ref{wp}) reproduces the weight of simple graphs. One can easily understand the appearance of the combinatorial factors in general case. Suppose that we permute links' labels of a fully labeled graph leaving nodes' labels intact. Among all $(2L)!$ permutations not all are distinct. If we have $A_{ij}$ links between vertex $i$ and $j$ and we will permute their labels, then all $A_{ij}!$ permutations will give the same labeled graph (if we simultaneously permute labels of the directed partners). Similarly if we have vertex with a self-connection and we exchange labels of the two directed links emerging from this vertex, the fully labeled graph will not change. Thus for each self-connection two permutations lead to the same fully labeled graph. To summarize, the number of distinct permutations of link labels is reduced from $(2L)!$ by dividing out the factor $2$ for each self-connection and $k!$ for each $k$-link multiple connection which just gives Eq. (\ref{wp}). It turns out that these weights are identical to the combinatorial factors of Feynman diagrams which appear in perturbative series of a mini-field theory \cite{ref:bck}. One can thus interpret random pseudographs as Feynman diagrams and use perturbation theory to enumerate them. Let us consider as an example a canonical ensemble of pseudographs with $N=3$ and $L=3$. There are $14$ graphs in this ensemble. They are shown in Fig. \ref{fig:pseudo_3}. \begin{figure}[h] \includegraphics[width=12cm]{pseudo_3.eps} \caption{All pseudographs in the canonical ensemble with $N=3,L=3$.} \label{fig:pseudo_3} \end{figure} In table \ref{tab3} we compare the theoretically calculated probability distribution of graphs: \begin{equation} p_\alpha = \frac{w_\alpha}{\sum_\beta w_\beta} , \end{equation} using the weights calculated by the formula (\ref{wp}) with the probability distribution obtained experimentally from the frequency histogram of graphs produced by the Monte-Carlo generator. Now the generator works exactly as before except that it does not reject moves leading to a self- or multiple-connections. The results are in perfect accordance. \begin{table}[h] \begin{center} \begin{tabular}{|c|c|c|c|c|c|c|c|} \hline Graphs & A & B & C & D & E & F & G \\ \hline Weights & 1/6 & 1/2 & 1/12 & 1/4 & 1/2 & 1/8 & 1/4 \\ \hline Theor. & 0.066 & 0.197 & 0.033 & 0.099 & 0.197 & 0.049 & 0.099 \\ \hline Exp. & 0.066(1) & 0.197(1) & 0.033(1) & 0.099(1) & 0.197(1) & 0.049(1) & 0.099(1) \\ \hline \hline Graphs & H & I & J & K & L & M & N \\ \hline Weights & 1/16 & 1/4 & 1/8 & 1/8 & 1/48 & 1/16 & 1/96 \\ \hline Theor. & 0.0247 & 0.099 & 0.049 & 0.049 & 0.0082 & 0.0247 & 0.0041 \\ \hline Exp. & 0.0246(1) & 0.099(1) & 0.049(1) & 0.049(1) & 0.0082(1) & 0.0247(1) & 0.0041(1) \\ \hline \end{tabular} \caption{Comparison of theoretical and experimental (Monte-Carlo) computations of frequencies of graphs' occurrence in the ensemble with $N=3,L=3$ \label{tab3}.} \end{center} \end{table} As an example let us calculate the weight of graph M in Fig. \ref{fig:pseudo_3}. The weight of each labeled graph of graph M, according to the formula (\ref{wp}), is equal \begin{equation} w_{M'} = \frac{1}{3!} \cdot \frac{1}{2^3} \frac{1}{2!} = \frac{1}{96} , \end{equation} where the first factor comes from $1/N!$, the second from the three self-connections, and the third from the fact that the two self-connections are attached to the same vertex and thus can be permuted without changing graph's connectivity. There are six distinct labeled graphs M' of graph M and thus \begin{equation} w_{M} = \sum_{M'} \frac{1}{96} = \frac{6}{96} = \frac{1}{16} . \end{equation} One should note that the number of distinct labeled graphs varies from graph to graph. For example, for graph L there is only one labeled graph. In this case $w_{L'} = 1/3! \cdot 1/2^3 = 1/48$ and $w_L = 1/48$. The calculation can be easily repeated for each graph in Fig. \ref{fig:pseudo_3} yielding the weights $w_\alpha$ listed in table \ref{tab3}. As follows from Eq. (\ref{wp}), the partition function for the canonical ensemble of pseudographs can be written in three different ways: \begin{equation} Z(N,L) = \sum_{\alpha'' \in flg(N,L)} \frac{1}{N! (2L)!} \ = \sum_{\alpha' \in lg(N,L)} \frac{1}{N!} \left( \prod_{i} \frac{1}{2^{A_{ii}/2} \left( A_{ii}/2\right)!} \right) \prod_{i>j} \frac{1}{A_{ij}!} \ = \sum_{\alpha \in g(N,L)} w_\alpha . \label{zw} \end{equation} The first sum runs over fully labeled graphs and has the simplest form since all fully labeled graphs have the same weight. The weight of labeled graphs in the second sum varies. We note that labeled graphs are isomorphic with adjacency matrices that is each labeled graph is uniquely represented by a generalized adjacency matrix \footnote{The elements $A_{ij}$ are indexed by nodes' labels, but the information about links' labels is lost in the adjacency matrix representation.} like Eq. (\ref{eq:pseudo}). The sum over $\alpha' \in lg(N,L)$ can be thus interpreted as a sum over all generalized adjacency $N \times N$ symmetric matrices $\mathbf{A}$ such that $\sum_{ij} A_{ij} = 2L$. We see that not all adjacency matrices have the same statistical weight unlike for simple graphs since the weights depend on the number of self- and multiple-connections. A weighted ensemble of pseudographs is constructed as before by introducing an additional functional weight $W(\alpha)$ under the sum defining the partition function (\ref{zw}): \begin{equation} Z(N,L) = \sum_{\alpha'' \in flg(N,L)} \frac{1}{N! (2L)!} W(\alpha'') \ = \sum_{\alpha \in g(N,L)} w_\alpha W(\alpha) . \end{equation} As before the functional weight $W(\alpha'')$ does not depend on graph's labeling but only on graph's topology. In other words if $\alpha''_1$ and $\alpha''_2$ are two different fully labeled graphs of graph $\alpha$ then $W(\alpha''_1)=W(\alpha''_2) \equiv W(\alpha)$. We can now consider various weights: for example a product weight as in Eq. (\ref{eq:prod}) to mimic graphs with uncorrelated node degrees. But even in this case the total weight does not factorize since the configuration space weight $w(\alpha)$ written as a function of node degrees $w(q_1,q_2,\dots, q_N)$ does not factorize. Due to the absence of the structural constraints the approximation given by equations (\ref{fact}) and (\ref{piqAB}) has now much weaker finite size corrections. A grand-canonical ensemble for pseudographs with arbitrary product weights (\ref{eq:prod}) has the following partition function: \begin{equation} Z(N,\mu) = \sum_L \exp (-\mu L) \sum_{\alpha \in g(N,L)} w_\alpha \prod_{i=1}^N p(q_i(\alpha)) . \label{eq:grps} \end{equation} This means that all pseudographs with fixed nodes' degrees $\{q_i\}$ have the same functional weight $\sim p(q_1)\cdots p(q_N)$, which seems to be similar to that generated by the Molloy-Reed construction of pseudographs \cite{ref:mr}. Let us comment on this. In the Molloy-Reed construction one generates a sequence of non-negative integers $\{q_1,q_2,\dots,q_N\}$ for example as independent identically distributed numbers with the distribution $p(q)$. One interprets $q_i$'s as node degrees. The only requirement is that the sum $q_1+q_2+\dots +q_N=2L$ is even. In the first step of the construction each integer $q_i$ is represented as a hub built out of a vertex and $q_i$ outgoing branches which can be viewed as directed links emerging from this vertex. In the second step the directed links are paired randomly in couples of links in opposite direction to form undirected links connecting vertices. This procedure generates the same subset of pseudographs as the partition function $Z(N,\mu)$ (\ref{eq:grps}). However, statistical weights are different. To see this, let us consider a subset of Molloy-Reed graphs obtained for a given set $\{q_i\}$. There are $N$ labeled vertices and $2L=\sum_i q_i$ labeled directed links. All permutations of labels of links and nodes are equiprobable exactly as it was before for fully labeled pseudographs (\ref{wp}). If one calculates corresponding symmetry factors for node-labeled graphs the same combinatorial factors arise as in Eq. (\ref{wp}): if one pairs two directed links $a$ and $b$ which belong to the same vertex one obtains a self-connecting link. The pair $ab$ is identical as $ba$ since both the links begin and end at the same vertex. This reduces the number of distinct permutations by factor of $2$ as in Eq. (\ref{wp}). Similarly for $k$ pairs of directed links between two vertices one can exchange the order of pairing in $k!$ ways each time obtaining the same multiple connection, so the corresponding factor is $1/k!$ again as in Eq. (\ref{wp}). The conditional probability of choosing a particular graph $\alpha$ under the condition that in the first step of the construction the set of $\{q_1,q_2,\dots,q_N\}$ has been selected, is \begin{equation} w_{M-R}(\{q_i\}|\alpha) = \frac{w_\alpha}{\sum_{\beta \in g\{q_1,\dots ,q_N\}} w_{\beta}} , \end{equation} where the sum is done over all (unlabeled) pseudographs $\beta$ from the micro-canonical set of fixed degrees $\{q_1,\dots ,q_N\}$. The total probability is thus \begin{equation} w_{M-R}(\alpha) = P(\{q_i\}) \, w_{M-R}(\{q_i\}|\alpha) , \end{equation} where $P(\{q_i\})$ is the probability that in the first step of the construction the set $\{q_1,\dots ,q_N\}$ is selected. This probability is proportional to the product of $p(q_i)$'s multiplied by the number of permutations of $\{q_1,\dots ,q_N\}$ which give the same set. We denote this number by $\mbox{Perm}(q_1,\dots, q_N)$. The order of $q_i$'s does not matter since we consider unlabeled graphs. For example, the following permutations (sequences): $(q_1,q_2,q_3)=(3,3,2)$, $(3,2,3)$ and $(2,3,3)$ give the same set $\{3,3,2\}$, so in this case we have $\mbox{Perm}(3,3,2)=3$. In general, the number is given by \begin{equation} \mbox{Perm}(q_1,\dots, q_N) = \frac{N!}{n_0! n_1!\cdots} , \end{equation} where $n_0,n_1,\dots$ are degree's multiplicities: $n_q=\sum_i \delta\left[ q_i-q\right]$. Thus \begin{equation} P(\{q_i\}) \propto \left( \prod_{i=1}^N p(q_i(\alpha)) \right) \mbox{Perm}(\{q_i\}) . \end{equation} \begin{figure}[h] \includegraphics[width=12cm]{grps.eps} \caption{A set of 10 pseudographs for $N=3$ and $p(q)=1/3$ for $q=0,1,2$. Top: three hubs for $\{q_i\}=\{2,2,2\}$ are generated and then directed links are paired randomly giving three pseudographs A1,A2,A3. Bottom: the rest of pseudographs from this ensemble.} \label{fig:grps} \end{figure} \begin{table}[h] \renewcommand*{\arraystretch}{1.3} \begin{center} \begin{tabular}{|c|c|c|c|c||c|c|} \hline L & $\{q_i\}$ & $\mbox{Perm}(q_1,q_2,q_3)$ & Graphs & $w_\alpha$ weights (\ref{wp}) & $w_{M-R}$ weights (\ref{eq:wmr}) & G-C weights from Eq. (\ref{eq:grps}) \\ \hline \hline 3 & 2,2,2 & 1 & A1,A2,A3 & $(\frac{1}{6}:\frac{1}{8}:\frac{1}{48}) = \frac{1}{15} (8:6:1)$ & $\frac{1}{14} (\frac{8}{15},\frac{6}{15},\frac{1}{15})$ & $e^{-3\mu}(\frac{1}{6},\frac{1}{8},\frac{1}{48})$ \\ \hline \hline 2 & 2,2,0 & 3 & B1,B2 & $\frac{1}{8}:\frac{1}{4}=\frac{1}{3}(1:2)$ & $\frac{3}{14} (\frac{1}{3},\frac{2}{3})$ & $e^{-2\mu}(\frac{1}{8},\frac{1}{4})$ \\ \hline & 2,1,1 & 3 & C1,C2 & $\frac{1}{2}:\frac{1}{4}=\frac{1}{3}(2:1)$ & $\frac{3}{14} (\frac{2}{3},\frac{1}{3})$ & $e^{-2\mu}(\frac{1}{2},\frac{1}{4})$ \\ \hline \hline 1 & 2,0,0 & 3 & D & $\frac{1}{4}$ & $\frac{3}{14}$ & $e^{-\mu}\frac{1}{4}$ \\ \hline & 1,1,0 & 3 & E & $\frac{1}{2}$ & $\frac{3}{14}$ & $e^{-\mu}\frac{1}{2}$ \\ \hline \hline 0 & 0,0,0 & 1 & F & $\frac{1}{6}$ & $\frac{1}{14}$ & $\frac{1}{6}$ \\ \hline \end{tabular} \caption{Weights calculated for the Molloy-Reed construction and for the corresponding grand-canonical ensemble with $N=3$. For each sequence $q_1,q_2,q_3$ we give the combinatorial number $\mbox{Perm}(q_1,q_2,q_3)$ of permutations leading to the same graph. Altogether, there are $14$ different sequences of length three, with $q_i=0,1$ or $2$ as can be seen in the third column of the table.} \label{tab:grps} \end{center} \end{table} Collecting all the factors together and normalizing to have the probabilistic interpretation we obtain the following expression for the total weight (probability) for Molloy-Reed's pseudographs: \begin{equation} w_{M-R}(\alpha) = \left( \prod_{i=1}^N p(q_i(\alpha)) \right) \frac{\mbox{Perm}(q_1(\alpha),\dots, q_N(\alpha))} {\sum_{k_1,\dots, k_N} \mbox{Perm}(k_1,\dots, k_N)} \, \frac{w_\alpha}{\sum_{\beta \in g\{q_1,\dots ,q_N\}} w_{\beta}} . \label{eq:wmr} \end{equation} The first factor comes from picking $N$ numbers $q_i$ at random, the second counts permutations and the third includes the weight generated by pairing directed links. As we see, despite many similarities the Molloy-Reed ensemble and the grand-canonical lead to different weights. As an example, in Fig. \ref{fig:grps} we show an ensemble of 10 pseudographs with $N=3$ generated by Molloy-Reed algorithm for $p(q)=1/3$ for $q=0,1,2$ and zero elsewhere. We compare statistical weights of the generated graphs with the corresponding ones in the grand-canonical ensemble. As we can see in table \ref{tab:grps}, the weights are different in the two ensembles. \section*{Summary} We have discussed a statistical approach to homogeneous random graphs. This framework is a natural extension of the Erd\"os-R\'enyi theory to the case of weighted graphs: one considers the same set of graphs but with modified statistical weights. The statistical weights of homogeneous graphs depend only on graphs' topology. In other words, if one assigns some labels to its nodes, they will have no physical meaning similarly as the numbers of indistinguishable particles in quantum mechanics. One can permute them and the graph and its statistical weight will stay intact. The only information which matters is the number (entropy) of distinct permutations of nodes' labels. All permutations of node's labels are equivalent, unlike for growing networks where those permutations have to preserve the causal order corresponding to the order of node's attachment to the graph. The statistical weight of a homogeneous graph is proportional to the number of all labeled graphs of this graph while of a growing network to the number of causally labeled graphs. This leads to a difference between homogeneous and growing networks. For example, a typical homogeneous graph has a larger diameter than the corresponding growing network with the same node degree distribution (\ref{eq:exact_p}). Generally, geometrical properties of homogeneous graphs are different from those of growing networks for which correlations between the time of node's attachment and its degree induce node-node correlations of a specific type \cite{ref:kr,ref:bbjk}. Such correlations are absent for homogeneous graphs. Various functional properties of homogeneous networks can be modeled by an appropriate choice of functional weight. One can easily produce networks with an assertive mixing, higher clustering or any desired property which can reflect any real-data observation. Homogeneous networks can be simulated numerically. We have also described a Monte-Carlo algorithm to generate canonical, grand-canonical and micro-canonical ensembles which performs a sort of weighted random walk (Markov chain) in the configuration space with a desired stationary distribution. We advocated the importance of the possibility of generating random networks with desired statistical properties for advanced motif searching \cite{ref:m,ref:bl2,ref:ia}. Many real networks have resulted from hybrid processes of growth mixed with some thermalization. The framework discussed in this paper can flexibly extrapolate between the two regimes. It allows one to directly investigate the relation between structural and functional properties of complex networks. \bigskip \noindent {\bf Acknowledgments} \medskip \noindent We thank Piotr Bialas, Jerzy Jurkiewicz and Andrzej Krzywicki for stimulating discussions. This work was partially supported by the Polish State Committee for Scientific Research (KBN) grant 2P03B-08225 (2003-2006) and Marie Curie Host Fellowship HPMD-CT-2001-00108 and by EU IST Center of Excellence "COPIRA".
{ "redpajama_set_name": "RedPajamaArXiv" }
1,155