Datasets:

---

LLMDH

/

OpenScience

like 0

Follow

LLM - Digital Humanities 15

ERROR: type should be string, got "https://doi.org/10.1007/s00265-020-02948-4\nBehavioral Ecology and Sociobiology (2021) 75: 6 https://doi.org/10.1007/s00265-020-02948-4\nBehavioral Ecology and Sociobiology (2021) 75: 6 ORIGINAL ARTICLE * Julia A. Kunz\njuliaandrea.kunz@uzh.ch The cost of associating with males for Bornean and Sumatran\nfemale orangutans: a hidden form of sexual conflict? Julia A. Kunz1\n& Guilhem J. Duvot1 & Maria A. van Noordwijk1 & Erik P. Willems1 & Manuela Townsend1 &\nNeneng Mardianah2 & Sri Suci Utami Atmoko2 & Erin R. Vogel3 & Taufiq Purna Nugraha4,5 & Michael Heistermann6 &\nMuhammad Agil5 & Tony Weingrill1 & Carel P. van Schaik1 Julia A. Kunz1\n& Guilhem J. Duvot1 & Maria A. van Noordwijk1 & Erik P. Willems1 & Manuela Townsend1 &\nNeneng Mardianah2 & Sri Suci Utami Atmoko2 & Erin R. Vogel3 & Taufiq Purna Nugraha4,5 & Michael Heistermann6 &\nMuhammad Agil5 & Tony Weingrill1 & Carel P. van Schaik1 Received: 3 August 2020 /Revised: 2 December 2020 /Accepted: 9 December 2020\n# The Author(s) 2020\n/ Published online: 30 December 2020 Abstract Sexual coercion, in the form of forced copulations, is relatively frequently observed in orangutans and generally attributed to their\nsemi-solitary lifestyle. High ecological costs of association for females may be responsible for this lifestyle and may have\nprevented the evolution of morphological fertility indicators (e.g., sexual swellings), which would attract (male) associates. Therefore, sexual conflict may arise not only about mating per se but also about associations, because males may benefit from\nassociations with females to monitor their reproductive state and attempt to monopolize their sexual activities. Here, we evaluate\nassociation patterns and costs for females when associating with both males and females of two different orangutan species at two\nstudy sites: Suaq, Sumatra (Pongo abelii), and Tuanan, Borneo (Pongo pygmaeus wurmbii). Female association frequency with\nboth males and females was higher in the Sumatran population, living in more productive habitat. Accordingly, we found that the\ncost of association, in terms of reduced feeding to moving ratio and increased time being active, is higher in the less sociable\nBornean population. Males generally initiated and maintained such costly associations with females, and prolonged associations\nwith males led to increased female fecal cortisol metabolite (FCM) levels at Tuanan, the Bornean population. We conclude that\nmale-maintained associations are an expression of sexual conflict in orangutans, at least at Tuanan. For females, this cost of\nassociation may be responsible for the lack of sexual signaling, while needing to confuse paternity. 1\nDepartment of Anthropology, University of Zurich,\nZurich, Switzerland 3\nDepartment of Anthropology, Rutgers The State University of New\nJersey, New Brunswick, NJ, USA\n4\nResearch Center for Biology, Indonesian Institute of Sciences (LIPI),\nCibinong, Indonesia\n5\nFaculty of Veterinary Medicine, Bogor Agriculture University,\nBogor, Indonesia\n6\nEndocrinology Laboratory, German Primate Center, Leibniz-Institute\nfor Primate Research, Göttingen, Germany 2\nFaculty of Biology and Primates Research Center, Universitas\nNasional, Jakarta, Indonesia 3\nDepartment of Anthropology, Rutgers The State University of New\nJersey, New Brunswick, NJ, USA Introduction troglodytes]: Emery Thompson and Georgiev 2014;\nGeorgiev et al. 2014). Therefore, males likely benefit from\nassociations with females, as this presumably facilitates mon-\nitoring their reproductive status and sexual activities and may\nbe attempts to mate guard females, even if females are often\nunlikely to be fertile. As a result, females and males may\nexperience a conflict about associating with each other. In\nfact, females may attempt to reduce time spent in associations\nbecause of potential foraging costs (e.g., Knott et al. 2018),\nwhereas males often attempt to prevent them from leaving\n(van Noordwijk and van Schaik 2009). If this is indeed the\ncase, male-female associations in orangutans can be consid-\nered an expression of sexual conflict, and male-maintained\nassociations could be seen as an indirect form of sexual coer-\ncion (cf. Muller et al. 2009). In most mammals, female reproductive success is limited by\naccess to food resources, while that of males is mainly limited\nby access to females (Darwin 1871; Emlen and Oring 1977). Hence, males and females have different behavioral strategies\nto optimize their lifetime fitness, which may lead to sexual\nconflict (Trivers 1972; Parker 1979). The high male-biased\noperational sex ratios in species with long lactational infertil-\nity and no paternal care may exacerbate sexual conflict\n(Clutton-Brock and Parker 1992; van Schaik 2016). Orangutan (Pongo spp.) females exhibit the longest inter-\nbirth intervals in primates of 6 to 9 years (van Noordwijk\net al. 2018), males do not provide direct paternal care for\ninfants (Rijksen 1978; Utami Atmoko et al. 2009), and males\nare not territorial (Spillmann et al. 2017). Hence, male-male\ncompetition for receptive females is high, which carries a high\npotential for sexual conflict (Trivers 1972; Parker 1979). However, the relative importance of male-male competition,\nfemale choice, and sexual conflict in orangutans remains in-\ncompletely understood (Nadler 1981; Fox 1998, 2002; Knott\n2009; Utami Atmoko et al. 2009; Knott et al. 2010; Spillmann\net al. 2010, 2017). There is evidence, however, for the behav-\nioral expression of sexual conflict, in the form of frequent\nforced copulations (Galdikas 1985a; Mitani 1985;\nSchürmann and van Hooff 1986; Knott et al. 2010). Females\nare vulnerable to this form of sexual coercion because of the\npronounced sexual dimorphism (Smuts and Smuts 1993), the\nsemi-solitary lifestyle (Rijksen 1978; van Schaik 1999), and\nthe absence of morphological fertility advertisements, such as\nsexual swellings (Nunn 1999; Zinner et al. 2004). Introduction Interestingly, apparent physical injuries resulting from forced\ncopulations have not been reported, and males seem to use\nonly as much force as is necessary to achieve intromission\n(Knott 2009). Significance statement Socioecological theory predicts a trade-off between the benefits of sociality and the ecological costs of increased feeding\ncompetition. Orangutans’ semi-solitary lifestyle has been attributed to the combination of high association costs and low\npredation risk. Previous work revealed a positive correlation between association frequencies and habitat productivity, but did\nnot measure the costs of association. In this comparative study, we show that females likely incur costs from involuntary, male-\nmaintained associations, especially when they last for several days and particularly in the population characterized by lower\nassociation frequencies. Association maintenance therefore qualifies as another expression of sexual conflict in orangutans, and\nespecially prolonged, male-maintained associations may qualify as an indirect form of sexual coercion. Keywords Concealed ovulation . Cost-of-sexual-attraction hypothesis . Fecal cortisol . Socioecology . Sexu Keywords Concealed ovulation . Cost-of-sexual-attraction hypothesis . Fecal cortisol . Socioecology . Sexual coercion . Pongo spp. Communicated by K. Langergraber Communicated by K. Langergraber Communicated by K. Langergraber * Julia A. Kunz\njuliaandrea.kunz@uzh.ch\n1\nDepartment of Anthropology, University of Zurich,\nZurich, Switzerland\n2\nFaculty of Biology and Primates Research Center, Universitas\nNasional, Jakarta, Indonesia\n3\nDepartment of Anthropology, Rutgers The State University of New\nJersey, New Brunswick, NJ, USA\n4\nResearch Center for Biology, Indonesian Institute of Sciences (LIPI),\nCibinong, Indonesia\n5\nFaculty of Veterinary Medicine, Bogor Agriculture University,\nBogor, Indonesia\n6\nEndocrinology Laboratory, German Primate Center, Leibniz-Institute\nfor Primate Research, Göttingen, Germany 4\nResearch Center for Biology, Indonesian Institute of Sciences (LIPI),\nCibinong, Indonesia 1\nDepartment of Anthropology, University of Zurich,\nZurich, Switzerland 5\nFaculty of Veterinary Medicine, Bogor Agriculture University,\nBogor, Indonesia 6\nEndocrinology Laboratory, German Primate Center, Leibniz-Institute\nfor Primate Research, Göttingen, Germany 6\nEndocrinology Laboratory, German Primate Center, Leibniz-Institute\nfor Primate Research, Göttingen, Germany 6\nEndocrinology Laboratory, German Primate Center, Leibniz-Institute\nfor Primate Research, Göttingen, Germany 6 Page 2 of 22 Behav Ecol Sociobiol (2021) 75: 6 Absence of fertility advertisement and the cost-of-\nsexual-attraction hypothesis Female orangutans do not exhibit any apparent graded, mor-\nphological signals advertising fertility (e.g., sexual swellings;\nNunn 1999; Zinner et al. 2004). Although rare observations of\nfemale proceptive copulations with dominant males have been\nlinked to the peri-ovulatory period (Fox 1998; Knott et al. 2010), ovulation appears largely concealed, as males initiate\ncopulations independent of the females’ reproductive state\nand during periods of lactational infertility (Nadler 1981;\nKnott et al. 2010; Kunz 2020). Unpredictable ovulation in\nother catarrhine primates has been linked to the need to coun-\nteract male monopolization and serves to confuse paternity\nand so offset the risk of infanticide (Hrdy 1979; van\nNoordwijk and van Schaik 2000; van Schaik et al. 2000,\n2004). However, signaling fertility bears costs for females\n(Matsumoto-Oda 1998; Archie et al. 2014). Particularly, the\nenergetic costs of grouping may have prevented females from\nevolving signals that advertise prolonged fertility and attract\nmale associates (Slater et al. 2008; Emery Thompson et al. 2014), thereby achieving such paternity confusion. Accordingly, Wrangham (2002) developed the “cost-of-sexu-\nal-attraction” hypothesis to explain the variation in morpho-\nlogical fertility advertisement in the genus Pan. Indeed, the\ncost of association for females (i) is negatively correlated with\nthe number of swelling cycles to conception (Deschner et al. 2004; Emery Thompson 2005; Douglas et al. 2016) and (ii) is\npositively correlated with the rate of sexual coercion and in-\nfanticide (Wilson et al. 2014). In a high-quality habitat, both\nthe immediate and delayed benefits to associate with males\nand to signal fertility over an extended period may outweigh\nthe costs for females. Bonobos (Pan paniscus) are at this low\ncost of association end (Furuichi 2011; Clay et al. 2016;\nNurmi et al. 2018) and have very prolonged periods of sexual\nattractivity (Douglas et al. 2016), and sexual coercion by\nmales is virtually absent (male aggression: Hohmann and Sexual conflict over associations Sexual conflict may arise not only about actual mating but\nalso about association maintenance. Female orangutans are\nat the solitary end of the fission-fusion spectrum (i.e., females\nspend on average 50–80% of their time with only their own\ndependent offspring; van Schaik 1999; van Noordwijk et al. 2009). The low association frequency suggests that the costs\nof association, in terms of increased scramble feeding compe-\ntition and hence reduced energy acquisition, are substantial for\nboth males and females (Galdikas 1988; van Schaik and Fox\n1996; Utami Atmoko et al. 1997). Yet, associations occur\nnevertheless, even if only one partner benefits. Specifically,\nbecause of the rare siring opportunities, male association de-\ncisions may be less cost-sensitive (van Schaik 1999), and\nmales may accept foraging costs due to increased copulation\nopportunities (orangutans: Mitani 1989; chimpanzees [Pan Page 3 of 22 Behav Ecol Sociobiol (2021) 75: 6 ge 3 of 22 6 Fruth 2003; Paoli 2009; infanticide: Hohmann et al. 2019). The cost of association for chimpanzee females varies geo-\ngraphically. In a more gregarious West African chimpanzee\npopulation (P. troglodytes verus; Boesch and Boesch-\nAchermann 2000; Riedel et al. 2011), rates of sexual coercion\nare low (Stumpf and Boesch 2010) and females may even\ndirectly profit from signaling prolonged sexual attractiveness\n(“Social Passport Hypothesis”; Deschner and Boesch 2007). In contrast, in an East African chimpanzee population\n(P. troglodytes schweinfurthii), females’ foraging effort is\ncompromised and their energy balance decreases with an in-\ncreasing number of males in association (Emery Thompson\net al. 2014). Consistent with the cost-of-sexual-attraction hy-\npothesis, East African chimpanzees have fewer swelling cy-\ncles per conception (Emery Thompson 2005; Deschner et al. 2004; but see Deschner and Boesch 2007), more male sexual\ncoercion (Muller et al. 2007, 2011), and more infanticide\n(Wilson et al. 2014) than West African chimpanzees. Harrison et al. 2010; Vogel et al. 2017). In previous studies,\nit has been shown that day journey length increases with in-\ncreasing association size, indicative of increased scramble\ncompetition both in Sumatran and Bornean populations (Fox\n1998; van Schaik 1999; Wartmann et al. 2010). In some respects, it appears that the orangutan species have\nadapted to their distinct conditions, and even under similar\nfood availability in captivity they show a different response\nto increased sociability. Male bimaturism and sexual conflict Orangutans exhibit a uniquely pronounced male bimaturism,\nwhich has been associated with alternative male reproductive\nstrategies (MacKinnon 1974; Utami Atmoko and van Hooff\n2004; Pradhan et al. 2012; Dunkel et al. 2013). Unflanged\nmales, who lack secondary sexual characteristics, reportedly\nassociate, copulate, and coerce copulations more frequently\nthan flanged males in the majority of study populations\n(MacKinnon 1974; Galdikas 1985b; Sugardjito et al. 1987;\nKnott 2009; Mitra Setia et al. 2009; Utami Atmoko et al. 2009; JAK et al., unpubl. data). Flanged males, who have fully\ndeveloped secondary sexual characteristics, emit long calls\nand are reported to rely largely on female choice around con-\nception (Fox 2002; Mitra Setia and van Schaik 2007;\nSpillmann et al. 2010). Although there is evidence for varia-\ntion among populations and species in the reproductive strat-\negies of the male morphs (Delgado and van Schaik 2000;\nKnott and Kahlenberg 2007; Mitra Setia and van Schaik\n2007; Utami Atmoko et al. 2009; Spillmann et al. 2017), sex-\nual conflict over associations is likely more pronounced with\nunflanged males than with flanged males, because the former\nassociate more frequently with females and cannot rely on\nfemale choice. Sexual conflict over associations In zoos, Bornean orangutans perma-\nnently housed with up to 5 adults exhibited overall higher\nfecal cortisol metabolite (FCM) levels than the more gregari-\nous Sumatran orangutans living in groups with up to 8 adults,\nwhich was attributed to species differences in the sensitivity to\nsocial stress (Weingrill et al. 2011). Moreover, captive\nBornean orangutans that were kept in fission-fusion like hous-\ning systems exhibited lower FCM levels than those kept in a\nstable group (Amrein et al. 2014). Taken together, these re-\nsults indicate that social factors, especially extended sociality,\nlead to a stronger physiological stress response in the less\nsociable Bornean orangutans, suggesting that they will also\nprefer lower association rates in the wild. Following the idea of the cost-of-sexual-attraction hypoth-\nesis, the absence of morphological fertility advertisement in\nthe genus Pongo may reflect the prohibitively high costs of\nassociation to repeatedly signal fertility, while still needing to\navoid male monopolization and thus confuse paternity. Here,\nwe evaluate the costs of association in orangutans; future stud-\nies will investigate how those relate to the frequency of sexual\ncoercion. Geographic variation in sociability in orangutans In the absence of high predation pressure due to their arboreal\nlife style (van Schaik and van Hooff 1983), food abundance is\nthe major constraint to population density and sociality in\norangutans (van Schaik 1999; Hardus et al. 2013; Vogel\net al. 2015). Fruit availability is not only thought to be respon-\nsible for the higher association frequency in West Sumatran\n(Pongo abelii) (average daily party size ranging from 1.6 to\n1.9 individuals) compared to both Eastern Sumatran\n(P. abelii) as well as Bornean orangutans (Pongo pygmaeus)\n(average daily party size ranging from 1.05 to 1.3 individuals)\n(van Schaik 1999; Mitra Setia et al. 2009; Wich et al. 2011;\nRoth et al. 2020), it also likely constrains associations within\npopulations over time (van Schaik and Fox 1996; Fox 1998;\nWich et al. 2006; Roth et al. 2020; J.Meric de Bellefon et al.,\nunpublished data). A high degree of scramble competition has\nbeen held responsible for the low female sociability, and direct\nfemale-female contest competition has also been reported\n(Utami Atmoko et al. 1997; Knott et al. 2008; van\nNoordwijk et al. 2012; Marzec et al. 2016). The ecological\neffect on association frequency, however, would be expected\nto be most prominent in associations between males and fe-\nmales because of the sex differences in ranging patterns and\nactivity budgets connected to their distinct energetic demands\n(Morrogh-Bernard et al. 2009; van Schaik et al. 2009; Aim of the study Associations and their maintenance may present another,\nmore subtle, context of sexual conflict in addition to forced\ncopulations in orangutans. Following the “cost-of-sexual- Behav Ecol Sociobiol (2021) 75: 6 6 Page 4 of 22 attraction” hypothesis (Wrangham 2002), high costs of asso-\nciation may be responsible for the absence of morphological\nfertility advertisements in female orangutans. Here, we evalu-\nate the costs of association for female orangutans with both\nfemales and males at two study sites, Suaq (P. abelii, Sumatra)\nand Tuanan (P. pygmaeus, Borneo) using behavioral and en-\ndocrine data. Because of large within-species variation in\nterms of their socioecology (e.g., Vogel et al. 2015; Roth\net al. 2020) and little evidence for life history differences be-\ntween species (van Noordwijk et al. 2018), we refer to study\nsite rather than species differences, as we evaluate only one\nstudy site per species. We included female-female associa-\ntions as a comparative category with the assumption that fe-\nmales have similar incentives to associate with each other (van\nNoordwijk et al. 2012) as opposed to male-female associa-\ntions and that these therefore are likely cost-sensitive (sensu\nvan Schaik 1999). We measured behavioral changes (i.e.,\nchanges in the daily activity budget) and variation in fecal\ncortisol metabolite (FCM) levels of parous females in relation\nto different types of association and social interactions. We\nhypothesized that (1) if males benefit from monitoring a fe-\nmale’s reproductive state and potentially attempt to monopo-\nlize a female’s sexual activities, they likely initiate and main-\ntain such associations with females; (2) social interactions\nbetween males and females are rare and therefore females\nlikely do not gain direct social benefits from associations with\nmales (for an exception see Marzec et al. 2016), whereas as-\nsociations with other females may provide social benefits for\ntheir infants (e.g., play opportunities; van Noordwijk et al. 2012); (3) associations with both males and females lead to\nhigher foraging costs, i.e., increased moving and reduced\nfeeding time, for females of the less sociable population,\nTuanan, than females of the more sociable population, Suaq;\nand (4) besides scramble competition, costs of grouping fe-\nmales may bear additional costs from agonistic interactions,\nespecially the occurrence of forced copulations, during asso-\nciations with males. In Table 1, we provide a detailed over-\nview of hypotheses and predictions. multiple years (van Noordwijk et al. Aim of the study 2013), both asso-\nciation patterns and the cost-benefit balances incurred\nby sociability are likely different from nulliparous\n(adolescent) females (van Schaik et al. 2009; Ashbury\net al. 2020). Therefore, only data on parous females\nwith a dependent offspring were included in this study\n(N = 20 females [Suaq: 6; Tuanan: 14]). Similarly, fe-\nmales who had lost their infants due to unknown rea-\nsons (Marzec et al. 2016; MAvN et al. unpubl. data)\nwere excluded from the analyses after the loss of their\ninfants, until they had given birth to a new infant. Infant loss is an extremely rare event (van Noordwijk\net al. 2018), and insufficient data were available to add\nthem as a separate category. The age of the dependent\noffspring of females was taken as a proxy for their\nreproductive state and included in all the analyses\n(Table 2). Infant ages were either known because the\nbirth was directly observed or estimated from the first\ntime an infant was observed (Table 2; cf. van\nNoordwijk et al. 2018). Activity budget Behavioral data were collected according to an established,\nstandardized protocol (https://www.aim.uzh.ch/de/\norangutannetwork/sfm.html). We collected 2-min instanta-\nneous data during full-day female focal follows on their activ-\nities (feeding, moving, resting, and social interactions). We\nrecorded all occurrences of any individual in association\n(within 50-m distance) per 2-min interval and ad libitum social\ninteractions with the focal individual. Social partners included\nthe female’s own dependent and independent offspring, ado-\nlescent individuals, and adult females and males (unflanged\nand flanged). We subdivided social interactions into sexual,\naffiliative, and aggressive interactions. Sexual interactions\ncomprised genital investigations by males, copulations, and\ncopulation attempts. Copulations were labelled as either\nforced, if the female showed any resistance behavior\n(e.g., repeated attempt to move away, struggling against\nthe males attempt to intromit), or unforced (following the\ndefinition of Fox 1998). We grouped aggression between\nfemales and both males and other females (excluding\nforced copulations) into non-physical (displays, short\nchases) and physical aggression (fights or coercive hand\nholding by males (van Schaik et al. 2006)). Affiliative\ninteractions comprised allo-grooming, touching another\nindividual, sitting in body contact, and begging for and\nsharing food. Because focal animals were individually\nrecognized, we could not collect blinded data. Only data\nfrom well-trained observers with high inter-observer reli-\nability were included in the analyses. Study sites and study subjects Behavioral focal data on individually recognized adult\nfemales were collected at the long-term field sites of\nTuanan, Mawas Reserve, Central Kalimantan,\nIndonesia (02° 15′ S; 114° 44′ E) and Suaq, Gunung\nLeuser National Park, South Aceh, Indonesia (03° 02′\nN; 97° 25′ E) between July 2003–July 2018 and June\n2007–March 2018, respectively. Because parous females\nare in continued association with their dependent off-\nspring (van Noordwijk et al. 2009) and lactate over Page 5 of 22 6 Behav Ecol Sociobiol (2021) 75: 6 2 6 Table 1\nOverview of the two main hypotheses evaluated in this study with the corresponding predictions\nHypotheses and predictions\nAdditional variables\nTested in\nSite†\nMale morph‡\nTab.$\nFig.$\nMain hypothesis: Associations are a context of sexual conflict in orangutans and may therefore\nqualify as a form of indirect sexual coercion. T > S\nUFM > FLM\nPart 1: Male direct benefits to associate with females exceeds those of females\nAssociation and social\ninteractions (benefits)\nPrediction 1.1: Male-female associations are more frequent than\nfemale-female associations.E\nT > SE\nUFM > FLME\nS 1\nS 1\nPrediction 1.2: Associations are male-initiated and\nmale-maintained.E\nT > SNE\nUFM > FLMNE\nS 2\n1–2\nPrediction 1.3: Male-female associations last longer than\nfemale-female associations, indicating that males benefit from\nprolonged associations.E Tuanan\nT > SE\nUFM > FLMNE\n3\n3\nPrediction 1.4: Affiliative social interactions are rare during\nmale-female associations, indicating near-absence of direct\nsocial benefits to females from associations with males.E\nT < SE\nUFM > FLME\nBuilding up\nrelationship? S 5\nS 2\nPrediction 1.5: Agonistic social interactions outside of the sexual\ncontext are rare during male-female associations, indicating\nnear-absence of direct coercive mate guarding by males.E\nNA\nUFM > FLMO\nS 6\nS 2\nPrediction 1.6: Sexual interactions (particularly forced\ncopulations) and genital investigations by males are frequently\nobserved in associations with females, indicating reproductive\nbenefits for males from associations.E\nT < SE,O\nUFM > FLME\nS 7–8\nS 3\nPart 2: Females incur costs from associations with males as a result of both increased feeding competition and the social interactions meanwhile. Cost of association to females\nPrediction 2.1: Females’ activity budget changes reflect increased\nscramble competition when in association with males. Study sites and study subjects Active\ntime increases and the F:M ratio decreases when in association\nwith males.E\nT > SE\nNot tested (but P1.1:\nUFM > FLME)\n4\nS 9–12\n4\nS 4\nPrediction 2.1.1: Females’ activity budget changes reflect\nincreased scramble competition when in association with any\nadult individual: Active time increases and the F:M ratio\ndecreases when in association with females.E\nT > SE\nNA\n4\nS 9–12\n4\nS 4\nPrediction 2.2: Aggression by males imposes further costs on\nfemales. Female F:M ratio decreases on days when they\nexperience male aggression (either sexual or non-sexual\naggression).NE (E copulation occurrence at Tuanan)\nT = SNE\nNot tested\n(UFM > FLM)\n4\nS 9–12\n5\nS 5\nPrediction 2.2.1: Aggression by other females also imposes costs\non females and therefore, female F:M ratio decreases also on\ndays when they experience aggression from other females,\nindicating that social stressors affect females’ activity budget.NE\nT = SNE\nNA\n4\nS 9–12\nNA\nPrediction 2.3: Female FCM levels are higher on days when in\nassociation with males than on days when females are alone\nwith their dependent offspring or in association with other\nfemales. Male associates therefore are a social stressor to\nfemales.NE\nT = SUnk\nNot tested (but P1.1)\n5\nNA\nPrediction 2.3.1: Female FCM levels are also higher on days\nwhen females are in association with other adult females than\nwhen alone with their dependent offspring. Any adult\nassociation partner may qualify as a social stressor to\nfemales.NE\nT = SUnk\nNA\n5\nNA\nPrediction 2.4: Female FCM levels increase with an increasing\nnumber of days spent in association with any association\npartner as a result of the accumulating negative energy balance,\nindicating energetic stress.E\nT > SUnk\nNA\n5\n6\nPrediction 2.5: Female FCM levels are higher on days when they\nexperience forced copulations, indicating either social (Soc.) or\nenergetic (Eco.) costs.NE\nSoc. T = S\nEco. Study sites and study subjects Observed cost-benefit balances and thus, predictions may vary with both study site († ) and male morph (‡ ). The\ntwo right columns indicate where the corresponding test (table with model output and figure) can be found † Association frequency is reportedly higher at Suaq (S) than at Tuanan (T), indicating generally lower costs of association at Suaq than at Tuanan. Accordingly, each prediction may vary with the socioecological background of the study site (“T > S”: prediction is more pronounced at Tuanan than\nSuaq; “T < S”: prediction more pronounced at Suaq than at Tuanan; “T = S”: no site difference expected) ‡ The two male morphs follow different reproductive tactics. While flanged males (FLM) are reportedly preferred by females, associate, and copulate\nmore selectively, unflanged males (UFM) associate, copulate, and coerce more frequently at Suaq and Tuanan (Kunz 2020). Both the extent of social\nbenefit and the cost of association inflicted on females may therefore vary with male morph and their reproductive tactic ‡ The two male morphs follow different reproductive tactics. While flanged males (FLM) are reportedly preferred by females, associate, and copulate\nmore selectively, unflanged males (UFM) associate, copulate, and coerce more frequently at Suaq and Tuanan (Kunz 2020). Study sites and study subjects T > S\nNot tested\n(UFM > FLM)\n5\n7\nMain hypothesis 2: Female orangutans do not exhibit any apparent morphological fertility advertisements because of the prohibitively high costs of Overview of the two main hypotheses evaluated in this study with the corresponding predictions Hypotheses and predictions Association and social\ninteractions (benefits) Prediction 2.2.1: Aggression by other females also imposes costs\non females and therefore, female F:M ratio decreases also on\ndays when they experience aggression from other females,\nindicating that social stressors affect females’ activity budget.NE\nT = SNE\nNA\n4\nS 9–12\nNA\ndi i\nl\nl\nl\nhi h\nd\nh\ni\nUnk\nd b Prediction 2.3: Female FCM levels are higher on days when in\nassociation with males than on days when females are alone\nwith their dependent offspring or in association with other\nfemales. Male associates therefore are a social stressor to\nfemales.NE\nT = SUnk\nNot tested (but P1.1)\n5\nNA Prediction 2.3.1: Female FCM levels are also higher on days\nwhen females are in association with other adult females than\nwhen alone with their dependent offspring. Any adult\nassociation partner may qualify as a social stressor to\nfemales.NE\nT = SUnk\nNA\n5\nNA Prediction 2.4: Female FCM levels increase with an increasing\nnumber of days spent in association with any association\npartner as a result of the accumulating negative energy balance,\nindicating energetic stress.E\nT > SUnk\nNA\n5\n6 Prediction 2.5: Female FCM levels are higher on days when they\nexperience forced copulations, indicating either social (Soc.) or\nenergetic (Eco.) costs.NE\nSoc. T = S\nEco. T > S\nNot tested\n(UFM > FLM)\n5\n7 g\nMain hypothesis 2: Female orangutans do not exhibit any apparent morphological fertility advertisements, because of the prohibitively high costs of\nassociation (cost-of-sexual-attraction hypothesis prediction 1), while needing to confuse paternity. Study sites and study subjects Observed cost-benefit balances and thus, predictions may vary with both study site († ) and male morph (‡ ). The\ntwo right columns indicate where the corresponding test (table with model output and figure) can be found\n† Association frequency is reportedly higher at Suaq (S) than at Tuanan (T), indicating generally lower costs of association at Suaq than at Tuanan. Accordingly, each prediction may vary with the socioecological background of the study site (“T > S”: prediction is more pronounced at Tuanan than\nSuaq; “T < S”: prediction more pronounced at Suaq than at Tuanan; “T = S”: no site difference expected)\n‡ The two male morphs follow different reproductive tactics. While flanged males (FLM) are reportedly preferred by females, associate, and copulate\nmore selectively, unflanged males (UFM) associate, copulate, and coerce more frequently at Suaq and Tuanan (Kunz 2020). Both the extent of social\nbenefit and the cost of association inflicted on females may therefore vary with male morph and their reproductive tactic\n$ The “S” in front of the number indicates that the table or figure, respectively, can be found in the supplementary materials\nE, evidence for this prediction was found in the current study; NE, no evidence was found for this prediction in the current study; Unk, not enough\nevidence to either support or disapprove the prediction, often because of limited data sets; O, evidence for the opposite pattern (for study site and male\nmorph comparison) (\n)\nHypotheses and predictions\nAdditional variables\nTested in\nSite†\nMale morph‡\nTab.$\nFig.$\nPrediction 3.2: The frequency of male-female associations in-\ncreases with the age of the dependent offspring indicating some\nreproductive benefits (conception and paternity manipulation),\nwhile female-female association remains constant over the age\nof the dependent offspring (indicating socializing benefits).E\nNA\nNA\nS 1\nS 1 Predictions include the evaluation of social benefits (P1.1–1.6) and costs resulting from both associations and social interactions (P2.1–2.5) during\nassociations between males and females. Observed cost-benefit balances and thus, predictions may vary with both study site († ) and male morph (‡ ). The\ntwo right columns indicate where the corresponding test (table with model output and figure) can be found Predictions include the evaluation of social benefits (P1.1–1.6) and costs resulting from both associations and social interactions (P2.1–2.5) during\nassociations between males and females. Table 2 Overview of the data\navailable to assess the cost of\nassociation for parous females at\nTuanan and Suaq § Number of days included in the analyses *Number of samples included in analyses\n§ Number of days included in the analyses Study sites and study subjects Prediction 3.1: Females incur costs from associations with males,\nwhich are consistent with scramble competition of grouping\n(predictions 2.1, 2.4).E\nT > S E\nNA\n4, 5\n4, 6 Table 1 (continued)\nHypotheses and predictions\nAdditional variables\nTested in\nSite†\nMale morph‡\nTab.$\nFig.$\nPrediction 3.2: The frequency of male-female associations in-\ncreases with the age of the dependent offspring indicating some\nreproductive benefits (conception and paternity manipulation),\nwhile female-female association remains constant over the age\nof the dependent offspring (indicating socializing benefits).E\nNA\nNA\nS 1\nS 1\n6 Page 6 of 22\nBehav Ecol Sociobiol (2021) 75: 6 Behav Ecol Sociobiol (2021) 75: 6 Table 1 (continued)\nHypotheses and predictions\nAdditional variables\nTested in\nSite†\nMale morph‡\nTab.$\nFig.$\nPrediction 3.2: The frequency of male-female associations in-\ncreases with the age of the dependent offspring indicating some\nreproductive benefits (conception and paternity manipulation),\nwhile female-female association remains constant over the age\nof the dependent offspring (indicating socializing benefits).E\nNA\nNA\nS 1\nS 1\nPredictions include the evaluation of social benefits (P1.1–1.6) and costs resulting from both associations and social interactions (P2.1–2.5) during\nassociations between males and females. Observed cost-benefit balances and thus, predictions may vary with both study site († ) and male morph (‡ ). The\ntwo right columns indicate where the corresponding test (table with model output and figure) can be found\n† Association frequency is reportedly higher at Suaq (S) than at Tuanan (T), indicating generally lower costs of association at Suaq than at Tuanan. Accordingly, each prediction may vary with the socioecological background of the study site (“T > S”: prediction is more pronounced at Tuanan than\nSuaq; “T < S”: prediction more pronounced at Suaq than at Tuanan; “T = S”: no site difference expected)\n‡ The two male morphs follow different reproductive tactics. While flanged males (FLM) are reportedly preferred by females, associate, and copulate\nmore selectively, unflanged males (UFM) associate, copulate, and coerce more frequently at Suaq and Tuanan (Kunz 2020). Study sites and study subjects Both the extent of social\nbenefit and the cost of association inflicted on females may therefore vary with male morph and their reproductive tactic\n$ The “S” in front of the number indicates that the table or figure, respectively, can be found in the supplementary materials\nE, evidence for this prediction was found in the current study; NE, no evidence was found for this prediction in the current study; Unk, not enough\nevidence to either support or disapprove the prediction, often because of limited data sets; O, evidence for the opposite pattern (for study site and male\nmorph comparison)\nTable 2 Overview of the data\navailable to assess the cost of\nassociation for parous females at\nTuanan and Suaq\nStudy\nsite\nName of\nparous\nfemale\nDependent infant\nActivity budget\nHormone samples\nMinimum\nage of\ninfant\n(years)\nMaximum\nage of infant\n(years)\nNumber\nof follow\nperiods\nNumber of\nfull-day focal\nfollows§\nTotal\nsamples\navailable\nWith\nbehavioral\nreference*\nSuaq\nCissy\n1.6\n5.2\n4\n35\n42\n13\nSuaq\nEllie\n0.3\n2.7\n4\n37\n32\n14\nSuaq\nFriska\n0.9\n4.9\n8\n51\n33\n10\nSuaq\nLisa\n0.5\n7.7\n11\n77\n60\n15\nSuaq\nRaffi\n0.8\n1.7\n1\n6\n12\n0\nSuaq\nSarabi\n0.6\n0.9\n2\n15\n4\n0\nTuanan\nCikipos\n3.2\n3.2\n1\n5\n6\n0\nTuanan\nCinta\n0.9\n2.7\n3\n17\n12\n5\nTuanan\nDesy\n0.0\n5.8\n13\n105\n54\n37\nTuanan\nInul\n0.1\n3.3\n12\n77\n44\n24\nTuanan\nJinak\n0.3\n7.1\n45\n365\n66\n39\nTuanan\nJuni\n0.0\n6.6\n33\n255\n88\n51\nTuanan\nKerry\n0.0\n7.5\n40\n280\n65\n38\nTuanan\nKondor\n0.0\n1.9\n10\n64\n23\n12\nTuanan\nMilo\n0.1\n1.1\n3\n20\n9\n6\nTuanan\nMindy\n0.1\n6.9\n52\n390\n95\n65\nTuanan\nPinky\n0.0\n7.5\n6\n39\n25\n7\nTuanan\nSidony\n0.0\n6.1\n6\n58\n45\n25 Table 1 (continued)\nHypotheses and predictions\nAdditional variables\nTested in\nSite†\nMale morph‡\nTab.$\nFig.$\nPrediction 3.2: The frequency of male-female associations in-\ncreases with the age of the dependent offspring indicating some\nreproductive benefits (conception and paternity manipulation),\nwhile female-female association remains constant over the age\nof the dependent offspring (indicating socializing benefits).E\nNA\nNA\nS 1\nS 1\nPredictions include the evaluation of social benefits (P1.1–1.6) and costs resulting from both associations and social interactions (P2.1–2.5) during\nassociations between males and females. *Number of samples included in analyses Collection, preservation, and extraction of fecal samples We measured fecal cortisol metabolite (FCM) levels for fe-\nmales during association and non-association days. Fecal ma-\nterial was collected non-invasively, when individuals defecat-\ned naturally. Because there is an excretion lag time of 24–72 h\nfor fecal cortisol metabolites (Weingrill et al. 2011), samples\nwere collected on at least 5 consecutive days, once a day,\npreferably in the morning. Due to individual ranging patterns\nin orangutans, samples could only be taken during focal fol-\nlows lasting 5–10 days with at least 5 weeks between succes-\nsive sample periods, because individuals were not followed\nduring this time. The methods to preserve and extract fecal\nsamples from orangutans for hormone analyses have been\nestablished and validated (Weingrill et al. 2011; Amrein\net al. 2014; Marty et al. 2015; Nugraha et al. 2016). Because\nof logistic constraints and varying infrastructures at the two\nfield sites, different preservation and extraction methods had\nto be used. Generally, the fresh feces were homogenized using\na stick and a 2–5-g aliquot was collected for analysis. Only\nsamples not contaminated with urine were taken. When elec-\ntricity from solar power was available, the fresh feces were\ncollected into a polypropylene tube and frozen at −18 °C\nupon return to the field station in the evening. All samples\nremained frozen until transported to the endocrinology labo-\nratory at Bogor Agricultural University where samples were\nlyophilized, pulverized and subsequently extracted with 80%\nmethanol in water as described in detail elsewhere (Weingrill\net al. 2011; Nugraha et al. 2016). At Suaq and when electricity\nsupply was not guaranteed at Tuanan, fecal samples were\nplaced in a tube containing 5 ml of 80% ethanol in water for\npreservation upon collection. Samples were extracted upon\nreturn to the field station using a field-friendly, previously\nvalidated extraction method (Nugraha et al. 2016). Although\nthese extraction methods have been shown to produce results\nwhich are strongly correlated (Nugraha et al. 2016), we con-\ntrolled for potential extraction method differences by normal-\nizing all FCM measurements within individual and method\nusing z-transformations in the statistical analyses (van de Pol\nand Wright 2009; for details see data analysis section). Study sites and study subjects Both the extent of social\nbenefit and the cost of association inflicted on females may therefore vary with male morph and their reproductive tactic The S in front of the number indicates that the table or figure, respectively, can be found in the supplementary materials\nE, evidence for this prediction was found in the current study; NE, no evidence was found for this prediction in the current study; Unk, not enough\nevidence to either support or disapprove the prediction, often because of limited data sets; O, evidence for the opposite pattern (for study site and male\nmorph comparison) E, evidence for this prediction was found in the current study; NE, no evidence was found for this prediction in the current study; Unk, not enough\nevidence to either support or disapprove the prediction, often because of limited data sets; O, evidence for the opposite pattern (for study site and male\nmorph comparison) Study\nsite\nName of\nparous\nfemale\nDependent infant\nActivity budget\nHormone samples\nMinimum\nage of\ninfant\n(years)\nMaximum\nage of infant\n(years)\nNumber\nof follow\nperiods\nNumber of\nfull-day focal\nfollows§\nTotal\nsamples\navailable\nWith\nbehavioral\nreference*\nSuaq\nCissy\n1.6\n5.2\n4\n35\n42\n13\nSuaq\nEllie\n0.3\n2.7\n4\n37\n32\n14\nSuaq\nFriska\n0.9\n4.9\n8\n51\n33\n10\nSuaq\nLisa\n0.5\n7.7\n11\n77\n60\n15\nSuaq\nRaffi\n0.8\n1.7\n1\n6\n12\n0\nSuaq\nSarabi\n0.6\n0.9\n2\n15\n4\n0\nTuanan\nCikipos\n3.2\n3.2\n1\n5\n6\n0\nTuanan\nCinta\n0.9\n2.7\n3\n17\n12\n5\nTuanan\nDesy\n0.0\n5.8\n13\n105\n54\n37\nTuanan\nInul\n0.1\n3.3\n12\n77\n44\n24\nTuanan\nJinak\n0.3\n7.1\n45\n365\n66\n39\nTuanan\nJuni\n0.0\n6.6\n33\n255\n88\n51\nTuanan\nKerry\n0.0\n7.5\n40\n280\n65\n38\nTuanan\nKondor\n0.0\n1.9\n10\n64\n23\n12\nTuanan\nMilo\n0.1\n1.1\n3\n20\n9\n6\nTuanan\nMindy\n0.1\n6.9\n52\n390\n95\n65\nTuanan\nPinky\n0.0\n7.5\n6\n39\n25\n7\nTuanan\nSidony\n0.0\n6.1\n6\n58\n45\n25\nTuanan\nSumi\n0.4\n3.4\n20\n154\n0\n0\nTuanan\nTina\n3.7\n5.5\n5\n36\n27\n9\n*Number of samples included in analyses\n§ Number of days included in the analyses Page 7 of 22 6 Behav Ecol Sociobiol (2021) 75: 6 (Vogel et al. 2017), and can thus be taken as a proxy for forest\nproductivity (Vogel et al. 2015). Association patterns and social interactions Male focal follows collected with the same methods as the\nfemale focal follows were used to enhance the data set on\nthe duration of associations and a fuller record of male-\nfemale dyadic interactions resulting in 960 male-female asso-\nciations (Suaq: 292; Tuanan: 668) with known start and end\ntimes. An association between two individuals could last for\nmultiple days and contain breaks, i.e., the association partners\nwere at a distance of more than 50 m. Orangutans most likely\nperceive the presence of other individuals at distances of more\nthan 50 m better than humans on the ground (van Noordwijk\net al. 2009, 2012), and therefore, brief “separations” (> 50-m\ndistance) likely are not relevant to them. If breaks lasted for\nlonger than one full-day focal follow, we considered it as two\nseparate association units. We recorded the individual respon-\nsible for any distance changes (in distance classes: contact, no\ncontact < 2 m, 2–5 m, 5–10 m, 10–50 m) during the associa-\ntion, as well as the initiator (first approach to < 50 m) and\nterminator (who left to > 50 m) of associations. We calculated\nthe Female Hinde Index (FHI) for female-male associations\nbased on these approaches and leaves as follows: Collection, preservation, and extraction of fecal samples Female Hinde Index FHI\nð\nÞ\n¼\napproaches by female\napproaches by female þ approaches by male\n\u0001\n\u0003\n−\nleaves by female\nleaves by female þ leaves by male\n\u0001\n\u0003 Female Hinde Index FHI\nð\nÞ A positive FHI indicates that the female was on average\nresponsible for the maintenance of the association, while a\nnegative FHI stands for a male-maintained association\n(Hinde and Atkinson 1970). The FHIs were calculated over\nall known approach and leave events and distance classes per\nassociation. Detailed approach and leave data throughout the\nassociation for the FHIs is available for 665 male-female as-\nsociations (Suaq: 223; Tuanan: 442). Ecological data The monthly Fruit Availability Index (FAI; percentage of\ntrees with fruits over all surveyed trees) was obtained from\nmonthly phenology surveys of ~ 1500 trees at Tuanan and ~\n1000 trees at Suaq (Harrison et al. 2010; Vogel et al. 2015)\n(Table 3). Because the FAI is generally higher at Suaq than at\nTuanan (Wich et al. 2011), we z-transformed all the FAIs\nwithin study site prior to the analyses (zFAI) to assess local\nFAI effects rather than between site comparisons. Although\nthe FAI does not include fruits from lianas, which are compo-\nnents of orangutan diet, it corresponds well to the total pro-\nportion of fruits, i.e., high-quality food items, in their diet Hormone measurement Fecal cortisol metabolite levels were measured in a total of\n745 samples (Table 2) using a microtiter plate enzyme immu-\nnoassay (EIA) for 11ß-hydroxyetiocholanolone (Ganswindt\net al. 2003), a major metabolite of cortisol in primate feces\n(Heistermann et al. 2006). The assay has been previously Behav Ecol Sociobiol (2021) 75: 6 6 Page 8 of 22 Table 3\nDefinition of fixed factors included in the activity budget analyses based on female full-day focal follows (z, continuous variables were z-\ntransformed prior to analyses)\nType\nFactor\nDefinition\nSocial\nz Cumulative male\nassociation hours\nThe sum of all association time spent with males during a full-day focal follow, e.g., if a female was 1 h in\nassociation with male A and 5 h with male B on a full-day focal follow, it would result in 6 cumulative\nmale association hours$\nz Cumulative female\nassociation hours\nThe sum of all association time spent with females during a full-day focal follow$\nNumber of consecutive days\nwith males\nThe number of (known) days a female focal animal was in association with males, independent of the\nmales’ identity, e.g., the female may have been in association with male A for 2 days and with male B\nthe next day, which would be 3 consecutive days with males\nNumber of consecutive days\nwith females\nThe number of (known) days a female focal animal was in association with females, independent of the\nfemale partner’s identity\nNumber of copulations\nNumber of observed copulations during the female full-day focal follow\nMale-female cumulative\naggression index\nWe combined the occurrence of forced copulations and other male aggression to a daily male-female\ncumulative aggression index, coded for severity (0 = no aggression; 1 = aggression not directly in a\nsexual context and no physical contact, such as displays, chases and displacements; 2 = forced sexual\ninteractions)\nFemale-female agonistic\ninteractions\nDays with female-female aggression were rare (N = 2 [Suaq], 16 [Tuanan]), and could only be included as\npresence/absence data, and not coded for the severity\nSocial interaction time (h)\nThe total time spent in social interactions with any partner (including own dependent infant and\nassociation partners) during a full-day focal follow (social interactions account for ~ 0.5% of the total\nactive time [Tuanan: 0.4%; Suaq: 1.1%])\nSite (Suaq vs. Tuanan)\nPopulation differences may be the result of either species or ecological differences between the two study\npopulations. did not exceed 15%. All FCM concentrations are expressed\nin ng/g dry fecal weight. did not exceed 15%. All FCM concentrations are expressed\nin ng/g dry fecal weight. did not exceed 15%. All FCM concentrations are expressed\nin ng/g dry fecal weight. validated and successfully applied for assessing adrenocorti-\ncal activity in numerous primate species (e.g., Heistermann\net al. 2006) including captive and wild orangutans\n(Weingrill et al. 2011; Amrein et al. 2014; Marty et al. 2015). Samples used for this study were analyzed in different\ncohorts at two different laboratories (German Primate Center,\nDPZ, by A. Heistermann and Bogor Agricultural University,\nIPB, by JAK), with the locality of analysis for each sample\nincluded as a fixed effect in the statistical analyses (results\nremain the same if we standardize by both laboratory, method\nand individual, and are not shown below). EIAs were per-\nformed as previously described (Heistermann et al. 2004). Samples from the same individual were analyzed on the same\nmicrotiter plate, whenever possible. Each sample was ana-\nlyzed in duplicate. We remeasured samples with a coefficient\nof variation (CV) > 7% between duplicates. Moreover, we\nreran any microtiter plate for which the intra-assay CV of\nthe internal high- and low-value quality controls exceeded\n10%. For the samples analyzed at both IPB and DPZ, the\nintra-assay CVs were below 10%, and the inter-assay CVs Hormone measurement Forest productivity is generally higher at Suaq (Sumatra, P. abelii) than at Tuanan\n(Borneo, P. pygmaeus wurmbii) (Wich et al. 2011). Ecological\nz Fruit Availability Index\nMonthly percentage of trees with fruits over all surveyed trees based on monthly surveys (~ 1500 trees at\nTuanan and ~ 1000 trees at Suaq)\nPhysiological z Age of dependent offspring\n(years)\nInfant ages were either known because the birth was directly observed, or estimated from the first time an\ninfant was observed (Table 2) Table 3\nDefinition of fixed factors included in the activity budget analyses based on female full-day focal follows (z, continuous variables were z-\ntransformed prior to analyses) The total time spent in social interactions with any partner (including own dependent infant and\nassociation partners) during a full-day focal follow (social interactions account for ~ 0.5% of the total\nactive time [Tuanan: 0.4%; Suaq: 1.1%]) Population differences may be the result of either species or ecological differences between the two study\npopulations. Forest productivity is generally higher at Suaq (Sumatra, P. abelii) than at Tuanan\n(Borneo, P. pygmaeus wurmbii) (Wich et al. 2011). Population differences may be the result of either species or ecological differences between the two study\npopulations. Forest productivity is generally higher at Suaq (Sumatra, P. abelii) than at Tuanan\n(Borneo, P. pygmaeus wurmbii) (Wich et al. 2011). Index\nMonthly percentage of trees with fruits over all surveyed trees based on monthly surveys (~ 1500 trees at\nTuanan and ~ 1000 trees at Suaq) Physiological z Age of dependent offspring\n(years)\nInfant ages were either known because the birth was directly observed, or estimated from the first time an\ninfant was observed (Table 2) $ We chose to include the daily cumulative hours spent with either adult females or males to account for multiple individuals in association and the\nduration spent with each of them (including both association time and the number of individuals as separate variables would have led to multi-collinearity\nissues) Behavioral data—association patterns and maintenance We evaluated the time (average daily hours) females spent in\nassociation with either other females or unflanged and flanged\nmales during follow periods in separate analyses (LMM) and\nwith the study site, zFAI, and the age of the dependent off-\nspring as fixed effects. Individual identity was added as a\nrandom intercept. We assessed when associations were male-maintained by\nsetting up a binomial GLMM based on the FHI values (male-\nmaintained when FHI < 0). We added study site, male morph,\nthe age of the dependent offspring of the female (years), local\nfruit availability (zFAI), the occurrence of copulations (both\nunforced and forced), and association duration as fixed ef-\nfects. To account for having the same individuals in several\nassociation dyads, both female and male identities were added\nas crossed random intercepts. p\nWe formulated a Cox proportional hazard mixed\nmodel (survival analysis) using the package “coxme”\n(Therneau 2018) to evaluate if male-female associations\nlasted over more consecutive days than female-female\nassociations based on the female focal follow data. We\nused right-censored data to account for unknown asso-\nciation endings, because females were no longer follow-\ned despite still being in association (N = 625 associa-\ntions [Suaq: 61 associations with females, 53 with\nflanged males and 173 with unflanged males; Tuanan:\n81 with females, 96 with flanged males, 161 with\nunflanged males] during 167 female FPs and of 21 fe-\nmales). We included both associations with known and\nunknown start times, because excluding associations\nwith unknown start times (~ 40% of association dyads)\nwould have introduced a bias against long associations\nin the analysis (for further details on this issue and for\nthe results excluding associations with unknown start\ntime, see supplementary mat). Besides the type of adult\nassociation partners (female, unflanged and flanged\nmale), we added study site, zFAI, and the age of the\ndependent offspring (years) as fixed effects in the mod-\nel. We set contrasts for the association partner type to\nfirst compare association maintenance between male and\nfemale association partners and then the two male\nmorphs. Further, we included the follow period nested\nin the female identity as a random intercept to avoid\npseudo-replication. pp\ny\n(\n)\nAs social factors, we included the total cumulative time\nspent with either males or females and any agonistic and sex-\nual interaction recorded as fixed effects (details in Table 3). Behavioral data—activity budget changes Behavioral data—activity budget changes ecological, and physiological factors) and their possible 2nd-\norder interactions (if applicable) were set up, and com-\npared to the control model, containing all the random\nand control (ecological and physiological) factors, using\nlikelihood ratio tests. All figures were generated using\nthe “ggplot2” (Wickham 2016) and “cowplot” (Wilke\n2019) packages. The daily activity budget was calculated from the 2-min in-\nstantaneous data taken during full-day female focal follows\n(N = 2086; Suaq: 221; Tuanan: 1865), and thus, (1) it includes\nthe association record over the entire day, and (2) it accounts\nfor activity budget variation over daytime. To account for\nvariation in activity budgets (van Noordwijk et al. 2012), we\nonly included female follow periods (FP) that contained at\nleast 5 full-day focal follows (mean = 8.3 ± SE 0.1) within\n40 days (on average within 9 days) in the analyses. First, we\nevaluated the variability of the total active time, which com-\nprised the total hours from leaving the morning nest to enter-\ning the evening nest. Second, we evaluated if female foraging\nbehavior changed on days with associations and social inter-\nactions by analyzing variation in daily feeding time while\ncontrolling for moving time (offset term) (henceforth referred\nto as F:M ratio). Daily moving hours correlate strongly with\nday journey length (DJL) (Pearson correlation for available\nTuanan data: R2 = 0.76, t769 = 32.66, P < 0.0001, N = 771 fe-\nmale full-day follows). We analyzed daily moving hours rath-\ner than DJL, because it is a proxy for daily travel, but also\nincludes moves within feeding patches, with no net displace-\nment in space, which are likely not accurately reflected in\nDJL. We tested for the effects of social, physiological and\necological factors on active time and F:M ratio in linear mixed\nmodels. We built in the female follow period (FP) nested in\nfemale identity as random intercepts to avoid pseudo-replica-\ntion. The separate analyses on the changes of all activity bud-\nget components (feeding, resting, and moving hours) includ-\ning separate analyses for each study site are reported in the\nsupplementary materials (STable 11–14). Statistical analyses All the statistical analyses were conducted in R version 3.5.2\n(R Core Team 2018). We ran (generalized) linear mixed effect\nmodels ([G]LMM) using the “lme4” and “lmerTest” packages\n(Bates et al. 2015; Kuznetsova et al. 2017). Model assump-\ntions (normality [for LMMs], homoscedasticity) were\nchecked by the visual inspection of residual plots. Variance\nInflation Factors (VIF) were calculated to examine potential\nmulti-collinearity issues using the “car” package (VIF < 2, for\nthe full model without interaction terms included and VIF < 4\nfor the full model with interaction terms) (Fox and Weisberg\n2018). Further, we checked all the models for influential cases\nand outliers (Cook’s distance from the package\n“influence.ME” by Nieuwenhuis et al. 2012). The P value of\n0.05 was used as a cutoff value for significance. For all statis-\ntical analyses, full models including all variables (social, Page 9 of 22 Behav Ecol Sociobiol (2021) 75: 6 9 of 22 6 6 Initiation and maintenance of associations The behavioral reference day corresponding to the measured\nFCM level was obtained by backdating 3 days from the date\nof collection of morning fecal samples and 2 days for samples\ncollected after 2 pm (Cadilek 2009; Weingrill et al. 2011;\nAmrein et al. 2014; Nugraha et al. 2016). If there were several\nfecal samples for one behavioral reference day, only the morn-\ning sample was included in the final analysis. FCM levels with\nthe same behavioral reference day were strongly correlated\n(r = 0.73, CI = [0.60, 0.83], P < 0.0001, N = 78). To control\nfor any sample hour bias, the time of sample collection (i.e.,\ntime of defecation) was included in the statistical analyses as a\ncontrol factor although previous data on captive-housed ani-\nmals showed no time-of-day effect (Weingrill et al. 2011). FCM levels were ln-transformed to normalize their distribu-\ntion. Subsequently, the values were standardized within indi-\nvidual and extraction method used, to be able to assess FCM\nlevel changes caused by social and ecological stressors within\nindividuals rather than between individuals (method described\nin van de Pol and Wright 2009). Because such z-transforma-\ntions may be sample size dependent, we only included those\nindividuals in the analyses for which more than 10 samples\nand at least 5 known behavioral reference days for a given\nextraction method were available. The within-individual\ntransformations were done including all available samples,\nincluding the samples without behavioral reference (N =\n745). The analysis included only the samples with a known\nbehavioral reference day (N = 370). The number of total sam-\nples available (per method and individual) was included in the\nanalysis as a control factor. A linear mixed effect model\n(LMM) was set up to test for the effect of social factors on\nfemale FCM levels. The same social factor categories as de-\nscribed in the activity budget analyses were tested. The time to\nsample extraction (days), the total number of days an individ-\nual was followed, the age of the dependent infant (years), an\nactivity budget parameter (feeding proportion), and the Fruit\nAvailability Index (FAI) were included in the full model to\ncontrol for possible confounding factors leading to FCM\nchanges. The female follow period was added as a random\nintercept to avoid pseudo-replication. Initiation and maintenance of associations Because the FCM levels\nwere standardized within individual and method, these two\nfactors were not included as random intercepts in the analysis\nto keep the models as parsimonious as possible. The analyses\nwithout the standardization procedure and including individ-\nual identity and extraction method as random intercepts\nyielded the same patterns and are reported in the supplemen-\ntary materials. The control model with all the potential con-\nfounding factors did not improve the model fit of the null Behavioral data—association patterns and maintenance Because consecutive association days are likely inter-\ndependent and there may be compensatory effects, we also\nincluded the total number of (known) consecutive days in\nassociation with either males or females in the full model. Further, we controlled for potential confounding physiologi-\ncal and ecological factors, overarching site differences\n(Tuanan, Suaq), and the total time spent in social interactions\nwith any partner (Table 3). We tested for interaction terms\nbetween study site and social factors to check for population\ndifferences. Interaction terms were only included in the final\nmodel if they improved the model fit based on likelihood ratio\ntests. Both control models – including study site, zFAI, age of\nthe dependent offspring and total social interaction time –\nsignificantly improved the null models, containing only the\nrandom intercepts (and the offset term) (active time: χ2\n4,8 =\n51.75, P < 0.001; F:M ratio: χ2\n4,8 = 53.08, P < 0.001). We ex-\ncluded 16 days when females fed less than 1 h and their active\ntime was below 6 h because of serious health issues or lack of 6 Page 10 of 22 Behav Ecol Sociobiol (2021) 75: 6 model containing the random intercept term only (χ2\n3,12 =\n6.96, P = 0.64, ΔAIC = 11.04). habituation, as these days revealed to be influential cases and\nthe model assumptions were violated (for one context of these\noutliers see Marzec et al. 2016). habituation, as these days revealed to be influential cases and\nthe model assumptions were violated (for one context of these\noutliers see Marzec et al. 2016). Association frequency Despite substantial day-to-day variation, females spent on av-\nerage (mean) 30.0 ± SE 0.1 min per day in association with\nother females (Suaq: 82.6 ± SE 13.6 min; Tuanan: 23.7 ± SE\n2.4 min), 53 ± SE 0.1 min with unflanged males (Suaq: 200.1\n± SE 20.0 min; Tuanan: 35.1 ± SE 3.1 min), and 20 ± SE\n0.0 min with flanged males (Suaq: 29.1 ± SE 6.8 min;\nTuanan: 19.2 ± SE 2.3 min) (SFig. 1). The time females spent\nin association with both flanged and unflanged males in-\ncreased as the age of their dependent offspring increased\n(flanged: β = 0.448 ± 0.090, t = 5.001, P < 0.001; unflanged:\nβ = 0.606 ± 0.126, t = 4.799, P < 0.001; STable 1; SFig. 1),\nwhile the association time with other parous females remained\nconstant with offspring age (β = 0.067 ± 0.063, t = 1.066, P =\n0.29; STable 1; SFig. 1). Time spent in association with other\nparous females and unflanged males was generally higher at\nSuaq than at Tuanan (females: β = −1.058 ± 0.360, t = −\n2.939, P = 0.008; unflanged: β = −1.821 ± 0.556, t = −\n3.273, P = 0.003; STable 1; SFig. 1), but not with flanged\nmales (β = 0.228 ± 0.332, t = 0.688, P = 0.5). To sum up, fe-\nmales were more frequently in association with unflanged\nmales than with adult females or flanged males, partly\nsupporting prediction 1.1 (Table 1), and time in association\nwith both male morphs increased with the age of the depen-\ndent offspring of females (Table 1: prediction 3.2). Association initiation and maintenance Both flanged males (Tuanan: 82.1%; Suaq: 73.9%) and\nunflanged males (Tuanan: 84.0%; Suaq: 80.7%) initiated as-\nsociations with females more frequently than the females\nthemselves (Fig. 1). Moreover, both flanged males (Suaq:\nmean[FHI] = −0.25 ± SE 0.06 [N = 63 associations];\nTuanan: −0.38 ± SE 0.03 [N = 205]) and unflanged males\n(Suaq: −0.16 ± SE 0.03 [N = 160]; Tuanan: −0.12 ± SE 0.03\n[N = 237]) maintained these associations (Fig. 2). The full\nmodel for the probability that associations were male-\nmaintained explained significantly more variability than the\nnull model (χ2\n3,9 = 72.53, P < 0.0001, N = 665 of 30 female\nand 140 male identities; STable 2). First, especially long as-\nsociations were more likely male-maintained (β = 1.226 ±\n0.255, OR = 3.40, z = 4.804, P < 0.001). Second, flanged\nmales were more likely to maintain associations with females\nthan unflanged males (β = 0.653 ± 0.201, OR = 1.92, z = Fig. 1 Proportion of association\ninitiations by either both, female\nor male by study site (a Suaq, b\nTuanan) and male morph. Only\nassociations with a known\ninitiator are included (N = 957\n[Suaq: 279; Tuanan: 678])\nFig. 2 Female Hinde Index of\nassociations with unflanged males\n(left) and flanged males (right) by\nstudy site (Suaq: top; Tuanan:\nbottom) by the age of the depen-\ndent offspring (year), as a proxy\nfor female reproductive status. The black crosses indicate the\nweighted mean FHI (by the num-\nber of known approaches and\nleaves) and their transparency is\nrelative to the number of associa-\ntions included. Data points (Suaq:\norange; Tuanan: blue) represent\nindividual association units and\nonly include known approaches\nand leaves (N = 665). The data\npoint size is relative to the asso-\nciation duration and a horizontal\njitter function was applied to the\ndata points to make overlapping\ndata points more visible\nPage 11 of 22 6\nBehav Ecol Sociobiol (2021) 75: 6 Page 11 of 22 6 Behav Ecol Sociobiol (2021) 75: 6 Fig. 1 Proportion of association\ninitiations by either both, female\nor male by study site (a Suaq, b\nTuanan) and male morph. Only\nassociations with a known\ninitiator are included (N = 957\n[Suaq: 279; Tuanan: 678]) Fig. Fig. 1 Proportion of association\ninitiations by either both, female\nor male by study site (a Suaq, b\nTuanan) and male morph. Only\nassociations with a known\ninitiator are included (N = 957\n[Suaq: 279; Tuanan: 678]) Association maintenance over multiple days Association maintenance over multiple days\nMale-female associations were maintained over more co\nutive days at Tuanan (maximum 8 days) than female-f\nassociations (maximum 4 days), whereas at Suaq this d\nence between the maintenance of male-female (max\n11 days) and that of female-female (maximum 7 days)\nciations was less pronounced (Fig. 3; Table 4). Accord\nthe survival analysis on the probability of ending an as\ntion was significantly better when including the intera\nbetween study site and partner type (β = −0.234 ± 0\nHR = 0.79, P = 0.009): Female-female associations e\nsooner at Tuanan than male-female associations; at Sua\ndifference was less pronounced (Fig. 3). We could no\nany difference in association maintenance probability be\nthe two male morphs (unflanged vs. flanged) (Tab\nAssociations were maintained over more consecutive\nwith the increasing age of the dependent offspring of a f\n(β = −0.169 ± 0.066, HR = 0.84, P = 0.01). The intera\nbetween the age of the dependent infant and partner typ\nnot improve the model fit (χ2\n2 = 0.83, P = 0.66). Local\ndid not have an effect on the association mainten\n(Table 4). All in all, both unflanged and flanged males\ntained associations with females over more consecutive\nTable 4\nProbability of ending an association: output of th\nproportional hazard mixed model for the total number of (known\nin association by the type of association partner, study site, age\nβ\nSite (Suaq vs. Tuanan)\n0.5\nAssociation partner\nSex (male vs. female)\n0.2\nMale morph (unflanged vs. flanged)\n−0.0\nAge of dependent offspring (years)\n−0. z Fruit Availability Index\n−0.0\nSite: association partner sex (male vs. female)\n−0.2\nSite: partner male morph (unflanged vs. flanged)\n−0.0 Male-female associations were maintained over more consec-\nutive days at Tuanan (maximum 8 days) than female-female\nassociations (maximum 4 days), whereas at Suaq this differ-\nence between the maintenance of male-female (maximum\n11 days) and that of female-female (maximum 7 days) asso-\nciations was less pronounced (Fig. 3; Table 4). Accordingly,\nthe survival analysis on the probability of ending an associa-\ntion was significantly better when including the interaction\nbetween study site and partner type (β = −0.234 ± 0.089,\nHR = 0.79, P = 0.009): Female-female associations ended\nsooner at Tuanan than male-female associations; at Suaq this\ndifference was less pronounced (Fig. 3). Association initiation and maintenance 2 Female Hinde Index of\nassociations with unflanged males\n(left) and flanged males (right) by\nstudy site (Suaq: top; Tuanan:\nbottom) by the age of the depen-\ndent offspring (year), as a proxy\nfor female reproductive status. The black crosses indicate the\nweighted mean FHI (by the num-\nber of known approaches and\nleaves) and their transparency is\nrelative to the number of associa-\ntions included. Data points (Suaq:\norange; Tuanan: blue) represent\nindividual association units and\nonly include known approaches\nand leaves (N = 665). The data\npoint size is relative to the asso-\nciation duration and a horizontal\njitter function was applied to the\ndata points to make overlapping\ndata points more visible 6 Page 12 of 22 Behav Ecol Sociobiol (2021) 75: 6 3.244, P = 0.001), whereas this difference between male\nmorphs was more pronounced at Tuanan than at Suaq, as\nthe model fit marginally improved when adding this interac-\ntion term (χ2\n9,10 = 4.12, P = 0.04). Association maintenance\nby males was independent of the female’s dependent off-\nspring’s age (β = 0.089 ± 0.100, OR = 1.09, z = 0.891, P =\n0.37), the local zFAI (β = −0.019 ± 0.089, OR = 0.98, z = −\n0.215, P = 0.83), and the occurrence of sexual interactions\n(β = 0.556 ± 0.298, OR = 1.74, z = 1.866, P = 0.06). In sum,\nprediction 1.2 (Table 1) was supported, as associations were\nmale-initiated and male-maintained independent of the female\nreproductive state, whereas this was the case at both study\nsites and by both male morphs. than females did with other females at Tuanan, the less socia-\nble population, whereas we find no such difference at Suaq,\nthe more sociable population, supporting prediction 1.3\n(Table 1) and its site difference but not the male morph\ncomponent. Social interactions between males and females Affiliative social interactions occurred in 6.0 ± SE 0.8% of all\nmale-female associations (STable 5), always once or twice\n(mean 1.45 ± SE 0.11 occurrences) during the entire associa-\ntion (female-unflanged: 0.022−h [interactions per association\nhour] [Suaq]; 0.025−h [Tuanan]; female-flanged: 0.016−h\n[Suaq]; 0.011−h [Tuanan]). Male aggression towards females\noutside of the sexual context was observed in 7.4 ± SE 0.9%\nof all dyadic male-female associations. Physical aggression by\nmales directed at females was rare (Suaq: in 1 out of 393\nassociations; Tuanan: 9 of 521 associations) and consisted\nexclusively of coercive hand holding (van Schaik et al. 2006). Flanged males were significantly more likely to direct\nnon-physical aggression in the form of displays, displace-\nments, or short chases towards females both at Tuanan\n(14.0 ± SE 2.2% [0.091−h]) and at Suaq (9.0 ± SE 2.6%\n[0.025−h]) than unflanged males (Tuanan: 6.6 ± SE 1.5%\n[0.030−h]; Suaq: 6.6 ± SE 1.5% [0.013−h]) (STable 6). At\nSuaq and Tuanan both forced and unforced copulations were\nmore frequent during associations involving unflanged males\n(Suaq: 23.5 ± SE 2.6% [0.075−h] [forced: 17.3 ± SE 2.3%\n(0.052−h)]; Tuanan: 15.0 ± SE 2.1% [0.041−h] [forced: 7.3 ±\nSE 1.5% (0.021−h)]) than flanged males (Suaq: 5.7 ± SE 2.1%\n[0.014−h] [forced: 4.1 ± SE 1.8% (0.011−h)]; Tuanan: 6.2 ± SE\n1.5% [0.018−h] [forced: 1.2 ± SE 0.7% (0.005−h)]) (for more\ndetails: Kunz 2020). Moreover, especially unflanged males at\nTuanan frequently investigated the genitals of females during\nassociations (female-unflanged associations: 27.4 ± SE 2.4%\n[Tuanan], 7.7 ± SE 1.9% [Suaq]; female-flanged associations:\n1.3 ± SE 0.7% [Tuanan], 4.2 ± SE 2.1% [Suaq]). These genital\ninvestigations occurred independent of the female’s offspring Active time Female active time on days without any association partners\nexcept for her dependent offspring was on average 10.8 ± SD\n1.0 h (min 6.1 and max 13.1), whereas on days with female\nassociates it increased to 11.4 ± SD 1.0 h (Suaq: 11.6 ± SD\n0.8; Tuanan: 11.3 ± SD 1.0) and on days with males in asso-\nciation (independent of association duration) to 11.4 ± SD\n1.0 h (Suaq: 11.5 ± SD 0.8; Tuanan: 11.3 ± SD 1.0) (suppl. mat. STable 9). Accordingly, in both study populations, fe-\nmale active time increased significantly with increased time in\nassociation with females (β = 0.084 ± 0.025, t = 3.428, P =\n0.001) and with males (β = 0.068 ± 0.034, t = 2.004, P =\n0.045) (Fig. 4a+d). Moreover, at Tuanan a female’s active\ntime increased significantly on days with copulations whereas Association maintenance over multiple days 3 Kaplan-Meier survival curve for the maintenance of associations\nover consecutive days at Suaq (a) and Tuanan (b) by the association\npartner type (color). The survival curve is based on the female focal\ndata from follow periods also including the non-full-day focal follows (e.g., days when an individual was found) (N = 625 [Suaq: 287; Tuanan:\n338] associations of 21 females and 168 different FPs). The left-censored\ndata is indicated in crosses it did not at Suaq (suppl. mat. STable 10), as the significant\ninteraction between study site and days with copulations indi-\ncates (β = 0.387 ± 0.158, t = 2.444, P = 0.02) (Fig. 5a). Active\ntime further increased with increased number of consecutive\ndays with males (β = 0.056 ± 0.024, t = 2.385, P = 0.02), the\ntotal time spent in social interactions with any social partner\n(β = 0.556 ± 0.155, t = 3.587, P < 0.001), and the local fruit\navailability (β = 0.095 ± 0.039, t = 2.427, P = 0.02). Interaction terms between site and any other social factors,\nexcept copulation occurrence, did not significantly improve\nthe model fit. In sum, daily active time increased in both\npopulations for females in associations, and at Tuanan on days\nwith copulations, and accordingly, the model fit significantly\nimproved when including social factors (χ2\n8,16 = 56.27,\nP < 0.001, ΔAIC = 40.27; N = 2086 of 20 females and 279\nFP; for the full model suppl. mat. STable 10). age (for details see suppl. mat. STable 7; SFig. 3). In summa-\nry, both affiliative and agonistic social interactions were rare\nduring male-female associations (predictions 1.4 + 1.5,\nTable 1), indicating that costs of association to females likely\nresult from increased feeding competition rather than the ac-\ncompanying social interactions. However, sexual interactions\nwere on average the most frequent social interactions during\nmale-female associations indicating male mating effort, thus\nsupporting prediction 1.6 and 3.2 (Table 1). Association maintenance over multiple days We could not find\nany difference in association maintenance probability between\nthe two male morphs (unflanged vs. flanged) (Table 4). Associations were maintained over more consecutive days\nwith the increasing age of the dependent offspring of a female\n(β = −0.169 ± 0.066, HR = 0.84, P = 0.01). The interaction\nbetween the age of the dependent infant and partner type did\nnot improve the model fit (χ2\n2 = 0.83, P = 0.66). Local zFAI\ndid not have an effect on the association maintenance\n(Table 4). All in all, both unflanged and flanged males main-\ntained associations with females over more consecutive days Table 4\nProbability of ending an association: output of the Cox\nproportional hazard mixed model for the total number of (known) days\nin association by the type of association partner, study site, age of the dependent offspring (years), and zFAI (χ2\n7 = 28.80, P = 0.0002, N = 625\nassociations of which 426 with known end, of 21 female identities and\n167 FPs). All fixed effects with P < 0.05 are indicated in bold β\nSE\nHazard ratio\nz\nP\nSite (Suaq vs. Tuanan)\n0.551\n0.146\n1.74\n-\n-\nAssociation partner\nSex (male vs. female)\n0.233\n0.071\n1.26\n-\n-\nMale morph (unflanged vs. flanged)\n−0.029\n0.103\n0.97\n-\n-\nAge of dependent offspring (years)\n−0.169\n0.066\n0.84\n−2.560\n0.010\nz Fruit Availability Index\n−0.023\n0.063\n0.98\n−0.360\n0.720\nSite: association partner sex (male vs. female)\n−0.234\n0.089\n0.79\n−2.620\n0.009\nSite: partner male morph (unflanged vs. flanged)\n−0.019\n0.130\n0.98\n−0.150\n0.880 Page 13 of 22 6 Behav Ecol Sociobiol (2021) 75: 6 Fig. 3 Kaplan-Meier survival curve for the maintenance of associations\nover consecutive days at Suaq (a) and Tuanan (b) by the association\npartner type (color). The survival curve is based on the female focal\ndata from follow periods also including the non-full-day focal follows\n(e.g., days when an individual was found) (N = 625 [Suaq: 287; Tuanan:\n338] associations of 21 females and 168 different FPs). The left-censored\ndata is indicated in crosses Fig. 3 Kaplan-Meier survival curve for the maintenance of associations\nover consecutive days at Suaq (a) and Tuanan (b) by the association\npartner type (color). The survival curve is based on the female focal\ndata from follow periods also including the non-full-day focal follows Fig. Foraging behavior Overall daily feeding time (F) decreased with both males and\nfemales in association, whereas moving (M) and resting time\nincreased (Fig. 4; for detailed analyses see suppl. mat. STable 9, 11, 12). At both study sites, the F:M ratio (time\nspent feeding per unit moving time) of females decreased with\nincreased association time with males (β = −0.250 ± 0.053,\nt = −4.698, P < 0.001), whereas it decreased significantly\nmore at Tuanan with increased time with females in Behav Ecol Sociobiol (2021) 75: 6 6 Page 14 of 22 Fig. 4 Daily female activity budget changes (from left to right: active\ntime (a, d), feeding (b, e), and moving (c, f) hours) depending on\ncumulative hours spent with males (a–c) and females (d–f) and by\nstudy site (round, orange: Suaq; triangles, blue: Tuanan). Each data\npoint represents one full-day focal follow (N = 2086), the regression lines\nare the correlations between hours spent with males/females and activity\nhours and do not show model predictions. The shaded areas display 95%\nconfidence intervals point represents one full-day focal follow (N = 2086), the regression lines\nare the correlations between hours spent with males/females and activity\nhours and do not show model predictions. The shaded areas display 95%\nconfidence intervals Fig. 4 Daily female activity budget changes (from left to right: active\ntime (a, d), feeding (b, e), and moving (c, f) hours) depending on\ncumulative hours spent with males (a–c) and females (d–f) and by\nstudy site (round, orange: Suaq; triangles, blue: Tuanan). Each data association compared to Suaq (β = −0.333 ± 0.070, t = −\n4.763, P < 0.001). Only consecutive days with females, but\nnot with males, led to a further decrease in a female’s daily\nF:M ratio (Table 5). Furthermore, on days with copulations,\nthe F:M ratio decreased significantly more at Tuanan than at\nSuaq (β = −0.630 ± 0.246, t = −2.557, P = 0.01). The full\nmodel for the F:M ratio including social factors was signifi-\ncantly better than the control model including ecological and physiological factors only (χ2\n8,17 = 80.47, P < 0.001,\nΔAIC = 62.47; N = 2086 of 20 females and 279 FP)\n(Table 5). Foraging behavior In sum, female foraging behavior was negatively\naffected by associations with both females and males, with the\neffects being more pronounced for the Tuanan population,\nsupporting predictions 2.1 and 2.1.1 (Table 1) that costs of\nassociation arise from scramble competition of grouping. Less support was found for predictions 2.2 and 2.2.1 Fig. 5 Daily female activity budget changes (from left to right: active\ntime (a), feeding (b), and moving (c) hours) depending on the\noccurrence of copulations and by study site (orange Suaq, blue\nTuanan). The boxplots are based on median values of full-day focal\nfollows (N = 2086) and do not show model predictions (the hinges extend\nto the first and third quantiles and the whiskers to the largest value, and\nlowest, respectively, at most 1.5*inter-quartile range. Data points beyond\nthe end of whiskers are plotted individually [Wickham 2016]) follows (N = 2086) and do not show model predictions (the hinges extend\nto the first and third quantiles and the whiskers to the largest value, and\nlowest, respectively, at most 1.5*inter-quartile range. Data points beyond\nthe end of whiskers are plotted individually [Wickham 2016]) Fig. 5 Daily female activity budget changes (from left to right: active\ntime (a), feeding (b), and moving (c) hours) depending on the\noccurrence of copulations and by study site (orange Suaq, blue\nTuanan). The boxplots are based on median values of full-day focal Page 15 of 22 6 Behav Ecol Sociobiol (2021) 75: 6 22 6 Table 5 LMM output of the full\nmodel for daily feeding hours\n(F:M ratio) (N = 2086 full-day\nfollows of 20 parous females and\n279 follow periods; χ2\n8,17 =\n80.47, P < 0.001, ΔAIC = 62.47). All fixed and control effects with\nP < 0.05 are indicated in bold. z,\nfixed effect variable was z-trans-\nformed prior to analysis; O, offset\nterm; C, control factor; F, fixed\neffect Estimate\nSE\nt\nP\nIntercept\n4.908\n0.240\n-\n-\nMoving time (h)\nOffset\nSite (Suaq vs. Tuanan)\nC\n−0.142\n0.262\n-\n-\nz Cumulative female association hours\nF\n0.181\n0.063\n-\n-\nz Cumulative male association hours\nF\n−0.250\n0.053\n−4.698\n< 0.001\nNumber of consecutive days with females\nF\n−0.167\n0.070\n−2.376\n0.018\nNumber of consecutive days with males\nF\n0.037\n0.037\n0.993\n0.321\nNumber of copulations\nF\n0.193\n0.206\n-\n-\nMale-female cumulative aggression index\nF\n−0.052\n0.140\n−0.371\n0.710\nFemale-female agonistic interactions (no vs. Foraging behavior yes)\nF\n−0.549\n0.312\n−1.762\n0.078\nz Fruit Availability Index\nC\n−0.150\n0.063\n−2.387\n0.018\nz Age of dependent offspring (years)\nC\n−0.047\n0.069\n−0.678\n0.498\nSocial interaction time (h)\nC\n−1.045\n0.241\n−4.330\n< 0.001\nSite (Suaq vs. Tuanan): z cumulative female\nassociation hours\nF\n−0.333\n0.070\n−4.763\n< 0.001\nSite (Suaq vs. Tuanan): number of copulations\nF\n−0.630\n0.246\n−2.557\n0.011 when adding the number of consecutive days with males\n(χ2\n12,13 = 7.30, P = 0.007, ΔAIC = 5.30). Although one par-\nticular female (Desy), who had male associations over a\ncourse of 9 days, appeared to be the main driver for this result,\nthere still was a trend for consecutive days with males leading\nto elevated FCM levels when this female was excluded (β =\n0.106 ± 0.056, P = 0.06, N = 333 of 89 follow periods; com-\nparison to control model: χ2\n12,13 = 3.47, P = 0.06). proposing additional costs caused by (agonistic) social inter-\nactions (Table 1). Table 6 LMM output for female\nFCM level changes [z-ln (FCM\nconcentration (ng/g))] in response\nto various ecological,\nphysiological and social factors\n(comparison to control model\n[containing only ecological and\nphysiological factors]: χ2\n12,16 =\n9.13, P = 0.06, ΔAIC = 1.13, N =\n370 of 96 FPs; comparison to null\nmodel: χ2\n3,16 = 16.09, P = 0.24,\nΔAIC = 9.91). All fixed and\ncontrol effects with P < 0.05 are\nindicated in bold. z, fixed effect\nvariable was z-transformed prior\nto analysis; C, control factor; F,\nfixed effect Table 5 LMM output of the full\nmodel for daily feeding hours\n(F:M ratio) (N = 2086 full-day\nfollows of 20 parous females and\n279 follow periods; χ2\n8,17 =\n80.47, P < 0.001, ΔAIC = 62.47).\nAll fixed and control effects with\nP < 0.05 are indicated in bold. z,\nfixed effect variable was z-trans-\nformed prior to analysis; O, offset\nterm; C, control factor; F, fixed\neffect FCM levels 6 Standardized FCM levels (z-ln [FCM concentration (ng/g)]) (y-\naxis) of females in response to consecutive association days with adult\nmales (a) and to consecutive association days with adult females (b). A\njitter function was added to the plot to visualize the overlapping data\npoints (consecutive days are only integers). The black diamond-shaped\npoints indicate the mean FCM levels with the error bar (SE) in black. Study sites indicated by round, orange: Suaq; triangles, blue: Tuanan points (consecutive days are only integers). The black diamond-shaped\npoints indicate the mean FCM levels with the error bar (SE) in black. Study sites indicated by round, orange: Suaq; triangles, blue: Tuanan Fig. 6 Standardized FCM levels (z-ln [FCM concentration (ng/g)]) (y-\naxis) of females in response to consecutive association days with adult\nmales (a) and to consecutive association days with adult females (b). A\njitter function was added to the plot to visualize the overlapping data points (consecutive days are only integers). The black diamond-shaped\npoints indicate the mean FCM levels with the error bar (SE) in black. Study sites indicated by round, orange: Suaq; triangles, blue: Tuanan levels was no longer significant (β = 0.083 ± 0.061, P = 0.18,\nN = 361 of 96 follow periods). In sum, it appears that only\nprolonged male-female associations over more than four con-\nsecutive days lead to increased female FCM levels. directly assess the energetic costs of sociality in orangutans,\nour measure of F:M ratio should be complemented by more\naccurate measures of actual energy intake and measures of\nenergy balance, such as analysis of urinary C-peptide concen-\ntrations (e.g., Emery Thompson and Knott 2008). Orangutan females likely do not gain direct benefits from\nassociations with males, whereas males need associations with\nfemales to monitor their reproductive status. First, genital in-\nvestigations by males and male-initiated sexual interactions\nwere the most frequent social interactions observed during\nmale-female associations, whereas affiliative social interac-\ntions were extremely rare. However, benefits for females by\nassociating with certain (flanged) males, such as protection\nfrom harassing males, cannot conclusively be ruled out\n(Mesnick 1997; Fox 2002). Second, most associations were\nboth male-initiated and male-maintained, regardless of female\nreproductive state, i.e., the age of the female’s dependent off-\nspring, and females likely incurred costs from those involun-\ntary associations as discussed above. FCM levels Female FCM levels increased with the number of consecutive\ndays in association with a male (β = 0.136 ± 0.047, t = 2.870,\nP = 0.004; Fig. 6; Table 6), but not with females (β = −0.082\n± 0.085, t = −0.960, P = 0.34). None of the other social fac-\ntors, including daily association time with either females or\nmales and the occurrence of aggression, further improved the\nmodel fit. Accordingly, the control model containing all phys-\niological and ecological factors was improved significantly Because our FCM data set only contained data for at most 4\nconsecutive days of female-female associations, we restricted\nthe data set to sample days of at most 4 consecutive male-\nfemale association days in a further analysis. Then, the effect\nof consecutive days in association with males on female FCM Type\nEstimate\nSE\nt\nP\nIntercept\n−0.193\n0.390\nSite (Suaq vs. Tuanan)\nC\n−0.021\n0.281\n−0.077\n0.939\nConsecutive days in association with female(s)\nF\n−0.082\n0.085\n−0.960\n0.338\nConsecutive days in association with male(s)\nF\n0.136\n0.047\n2.870\n0.004\nMale-female cumulative aggression index\nF\n−0.110\n0.118\n−0.932\n0.352\nFemale-female agonistic interactions (no vs. yes)\nF\n0.063\n0.443\n0.142\n0.887\nNumber of days followed\nC\n−0.002\n0.018\n−0.141\n0.888\nz Daily feeding proportion\nC\n−0.072\n0.054\n−1.325\n0.186\nz Age of dependent infant (years)\nC\n0.007\n0.121\n0.059\n0.953\nz Fruit Availability Index\nC\n0.057\n0.076\n0.739\n0.462\nHour of sample collection\nC\n0.004\n0.025\n0.151\n0.880\nz Days to sample extraction\nC\n0.067\n0.082\n0.812\n0.419\nLaboratory (DPZ vs. IPB)\nC\n0.027\n0.202\n0.133\n0.894\nTotal samples available with behavioral reference\nC\n0.005\n0.008\n0.560\n0.577 Table 6 LMM output for female\nFCM level changes [z-ln (FCM\nconcentration (ng/g))] in response\nto various ecological,\nphysiological and social factors\n(comparison to control model\n[containing only ecological and\nphysiological factors]: χ2\n12,16 =\n9.13, P = 0.06, ΔAIC = 1.13, N =\n370 of 96 FPs; comparison to null\nmodel: χ2\n3,16 = 16.09, P = 0.24,\nΔAIC = 9.91). All fixed and\ncontrol effects with P < 0.05 are\nindicated in bold. z, fixed effect\nvariable was z-transformed prior\nto analysis; C, control factor; F,\nfixed effect 6 Page 16 of 22 Behav Ecol Sociobiol (2021) 75: 6 Fig. FCM levels When females are ready\nto conceive, however, they may actively seek the association\nwith (dominant) flanged males (Fox 1998, 2002; Spillmann\net al. 2010). With our analyses, we did not capture this short\nwindow around conception. We conclude that females and\nmales are likely at odds about association maintenance. Accordingly, orangutan females have been reported to active-\nly avoid male associates or try to end associations as rapidly as\npossible (Fox 2002; Mitra Setia and van Schaik 2007; Utami\nAtmoko et al. 2009; van Noordwijk and van Schaik 2009;\nSpillmann et al. 2010; Knott et al. 2018). Further investiga-\ntions to understand how and if females attempt to avoid male\nassociates have to be conducted, including the analysis of\nsimultaneous ranging data. In sum, our study indicates that\nfemales incur costs from male-maintained associations, but\nno clear immediate benefits (albeit perhaps indirect ones:\nKunz 2020), especially during the period of lactational Stress and association Female FCM levels increased as they spent more days in\nassociation with males, but not with females. This social factor\nwas the best and only predictor for FCM level changes. Thus,\nrepeated days of increased active time and reduced F:M ratio\nled to a physiological stress response. Interestingly, this was\nnot the case when in association with other females, because\nfemales can avoid lengthy associations with other females\nbefore associations become too costly. Conversely, males ap-\npear to profit from associations with females and they main-\ntain associations over a longer time period than a female\nwould. The behavioral data available support this conclusion:\nFemale-female associations never lasted more than 4 consec-\nutive days at Tuanan, where the F:M ratio decreased signifi-\ncantly more when in association with other females than at\nSuaq, while male-female associations could last up to 8 days. The elevated FCM levels of captive orangutan females when\nartificially confined to permanent association with males\n(Amrein et al. 2014) further support our hypothesis that in-\ncreased sociality over an extended time period leads to a phys-\niological stress response, especially in Bornean orangutans. The findings in captivity suggest that Bornean females show\nstress reactions to extended sociality even in the absence of\nreduced net energy intake, suggesting that in captivity in-\ncreased FCM levels in females associated with males more\nlikely reflect social rather than energetic stress. Although our\nendocrine data set is very limited, especially for the extended\nconsecutive association days with males, we propose that only\nextended association periods with males lead to increased\nFCM levels as seen in captivity. However, whether these\nFCM elevations observed in our wild females are a response\nto the association itself or, alternatively, reflect energetic con-\nstraints due to the association-related decrease in feeding time\nand increase in active time is unclear. Future studies should\ngenerally aim at obtaining a more conclusive endocrine data\nset including larger sample sizes linked to consecutive associ-\nation days. We found no evidence for differences in female FCM\nlevels on days with any agonistic interaction with either males\nor females in the two populations. Even though days with\ncopulations were characterized by increased active time and\nreduced F:M ratio at Tuanan, we found no evidence that male\naggression, in particular sexual coercion (SFig. 7), imposed\nany additional costs, either as reduced feeding time or in ele-\nvated FCM levels. Foraging costs of association Because female reproductive success is generally directly\nlinked to access to resources (chimpanzees: Emery\nThompson et al. 2007; apes: Emery Thompson et al. 2008;\nStumpf et al. 2008; orangutans: Knott et al. 2009), the ener-\ngetic costs of association with conspecifics have been held\nresponsible for the varying degrees of gregariousness across\nthe orangutan geographic distribution (van Schaik 1999). The\nfemales in our study likely suffer energetically from associa-\ntions (with both males and adult females): In both study pop-\nulations, females increased the length of their active day, but\ntheir feeding time decreased, both absolutely (suppl. mat. STable 12) and relative to moving time. This reduction is\nnot only a trade-off directly resulting from increased time\nspent in social interactions, because (1) we controlled for time\nspent in social interactions, and (2) in the more sociable\nSumatran population with higher forest productivity, feeding\ntime was less affected by time spent in association with fe-\nmales. Hence, the reduced F:M ratio and the increased active\ntime can be taken as direct evidence for elevated scramble\ncompetition, indicating that associations incur energetic costs\nto females, whereas we only found limited evidence for costs\nresulting from (agonistic) social interactions. We can conclude\nthat females modify their activity budgets when in association\nwith both males and females, in patterns that are congruent\nwith increased scramble competition. However, to more Page 17 of 22 Behav Ecol Sociobiol (2021) 75: 6 6 the future. A difference in the physiological response to social\nstressors, including energy balances, may be expected in the\nlight of the socioecological theory, because the degree of so-\nciability between the two populations differs (this study; van\nSchaik 1999). Since our activity and feeding data indicate that\nboth associations (with females) and social interactions are\ncostlier to Tuanan females than to Suaq females, where fruit\navailability is generally higher (Wich et al. 2011), a stronger\nphysiological stress response would be expected at Tuanan. Future studies are, however, needed to test this hypothesis and\nthus to evaluate whether females of the more sociable\nSumatran orangutan may be more “stress-resistant” which\ncould explain why there is less need for either behavioral or\nphysiological mechanisms to avoid associations. infertility (~ 6.5 years [van Noordwijk et al. 2018]). These\ncosts of involuntary associations may be relevant, because\norangutan females’ reproductive success highly depends on\nthe availability of resources (Knott et al. Foraging costs of association 2009), particularly\nin a less productive habitat (Wich et al. 2011). The male perspective male-maintained associations, especially if those associations\nlast multiple days. The costs include reduced feeding time and\nincreased moving and resting time, which adds up to longer\nactivity per day and thus shorter night rest. Furthermore,\nprolonged associations with males were associated with ele-\nvated FCM levels, whereas this was not the case for female-\nfemale associations which were usually much shorter. We\nsuggest that the absence of morphological fertility advertise-\nment in female orangutans may be explained by these costs of\nassociation, thus supporting the first prediction of the “cost-of-\nsexual-attraction” hypothesis (Wrangham 2002) for orangu-\ntans. The length of sexual attractivity negatively correlates\nwith the cost of association for females in the genus Pan\n(Wrangham 2002). Orangutans fit into this fission-fusion con-\ntinuum at the solitary end: They do not exhibit any morpho-\nlogical signal of fertility, arguably because this would attract\ntoo many competing males at once leading to a prolonged\nperiod of unacceptably high energetic costs for the females,\nin addition to the mere physiological costs associated with the\nswelling itself (for a review: Nunn 1999). On the contrary,\nfemale orangutans advertise non-availability with small labial\nswellings during pregnancy (Schultz 1938; Galdikas 1981),\nlikely to reduce the costs of association as males refrain from\nmaintaining associations and copulating with pregnant fe-\nmales exhibiting the labial swelling (only 2 out of 34 pregnan-\ncy matings were observed when females exhibited a pregnan-\ncy swelling, JAK et al. unpubl. data). Both unflanged and flanged males are responsible for maintain-\ning associations, independent of the females’ dependent off-\nspring age (as a proxy for reproductive state), which supports\nthe hypothesis that the males’ interest to associate exceeds that\nof the females (Table 1). Besides mating opportunities, these\nassociations may be an attempt to monitor a female’s reproduc-\ntive state and sexual activities. In the absence of any apparent\nsignal of fertility (Nunn 1999), it remains uncertain how males\nassess female reproductive state, if at all. The genital investiga-\ntions reported here may provide some olfactory information to\nmales (cf. chimpanzee: Matsumoto-Oda et al. 2003; review:\nDrea 2015), but data are insufficient to know how and if these\nrelate to sexual interactions (suppl. mat. STable 8). It is likely\nthat males also incur energetic costs from associations and in-\nteractions with females (cf. East African chimpanzees\n[P. troglodytes schweinfurthii]; Emery Thompson and\nGeorgiev 2014; Georgiev et al. The male perspective 2014), and our unpublished data\nsuggest this, too, for orangutan males. Thus, males may have a\nset of decision rules when and for how long to associate with\ncertain females. Accordingly, the time in association with males\nincreases with the increasing age of the dependent offspring of\nfemales (this study; van Schaik 1999), suggesting some type of\nreproductive benefits for males. More detailed analyses on the\nsocial context of associations will provide further insight into\nhow males benefit from sociality with females. Yet, females of both Pan spp. and Pongo spp. exhibit un-\npredictable ovulation, albeit to varying extent (Nadler 1981;\nDeschner et al. 2004; Douglas et al. 2016), which has been\nlinked to paternity confusion serving infanticide avoidance\nstrategies (Hrdy 1979; Hrdy and Whitten 1987; van Schaik\net al. 2004). The concealed ovulation in orangutans (Nadler\n1981) may therefore also serve to reduce the risk of infanticide\nas it does in most other primates (Hrdy 1979; van Schaik et al. 2004). Female orangutans seem to vary their mate preferences\nwith their reproductive status accordingly (Knott et al. 2010). However, evidence for infanticidal attacks by males remains\nindirect (Beaudrot et al. 2009; Knott et al. 2019; Scott et al. 2019) and infant mortality is generally extremely low (van\nNoordwijk et al. 2018), suggesting that male infanticide in\norangutans is extremely rare compared to chimpanzees and\nthat females employ efficient counterstrategies. It remains to be investigated if prolonged male-maintained\nassociations should be labelled as a separate indirect form of\nsexual coercion or may even function as coercive mate guarding,\ni.e., “to constrain female promiscuity” (Muller et al. 2009). First,\ndirect non-sexual aggression towards females by males is rare in\norangutans (STable 6; SFig. 2) providing little evidence for any\nherding, punishment or sequestration (apart from the ten cases of\ncoercive hand holding). However, anecdotal data suggest subtle\nsequestration attempts, in that males may try to influence fe-\nmales’ travel direction away from other males during associa-\ntions (MAvN et al., unpubl. data). Second, copulations regularly\noccur in the presence of other, even more dominant, males (Fox\n2002). (Coercive) mate guarding by males therefore appears to\nbe an inefficient strategy, especially for subordinate, unflanged\nmales. Third, although we found evidence for direct costs for\nfemales resulting from male-maintained associations, which in-\ndicates male coercion, we cannot rule out that females ultimately\nbenefit indirectly from paternity confusion through those male-\ndriven association patterns. The male perspective Future studies are needed to evaluate\nthe social contexts of associations. In a dispersed mating system with high costs of association,\nand where males generally drive association patterns as found\nhere for orangutans, the lack of morphological fertility adver-\ntisement can be explained by the selection on the total conceal-\nment of ovulation. Given a risk of infanticide (Knott et al. 2010,\n2019), females must achieve an optimum distribution of pater-\nnity assessments (van Schaik and Janson 2000; van Schaik et al. 2004) by removing as much information on female fertility sta-\ntus as possible. Accordingly, the absence of morphological fer-\ntility advertisement combined with the concealed ovulation in Stress and association Thus, the com-\nparison should be repeated with a more extensive data set in 6 Page 18 of 22 Behav Ecol Sociobiol (2021) 75: 6 Stress and association If these forced copulations are cost insen-\nsitive, they would not qualify as sexual coercion by the defi-\nnition of Smuts and Smuts (1993) (“use by a male of force, or\nthreat of force, that functions to increase the chances that a\nfemale will mate with him at a time when she is likely to be\nfertile, and to decrease the chances that she will mate with\nother males, at some cost to the female”), while prolonged,\nmale-maintained associations would. However, the absence of\na stress response does not exclude other costs of forced cop-\nulations, such as the limitation to the expression of female\nmating preferences. Indeed, the consistent attempts by females\nto escape from involuntary mating initiations (Fox 2002;\nKnott et al. 2010) suggest that females perceive resisted cop-\nulations as undesired rather than as a way to assess mate qual-\nity. For now, therefore, interpreting forced copulations as sex-\nual coercion remains the most plausible explanation. Since fecal cortisol metabolite levels represent an integra-\ntive measure of pooled endocrine activity over several hours\nor days (Hodges and Heistermann 2011), it is likely unsuited\nto detect short-term stress responses to a specific behavioral\nevent. Forced copulations lasted on average 8.8 ±SD 7.2 min\n(Kunz 2020), and any stress response associated with this\nbehavior is likely to be too short to be detected by our FCM\nmeasure. Urinary cortisol levels may thus be a more appropri-\nate measure to assess whether particular social interactions\ninduce more immediate elevations in cortisol production\n(e.g., Silk et al. 2013) as has been shown for chimpanzees\n(Muller et al. 2007; Emery Thompson et al. 2010). Further\ndetailed studies, with a larger sample size and more immediate\nmeasures of cortisol levels from urine, are needed to examine\nwhether female orangutans do indeed not show stress re-\nsponses to forced copulations. Following the same line of argument, one would expect to\nfind more pronounced FCM level changes in the less sociable\nBornean orangutans in response to involuntary associations\ncompared to Sumatran orangutans. Although we could not\nfind any evidence for differences in FCM level changes be-\ntween Suaq and Tuanan, our data set was very small for the\nSuaq population (N = 52 samples, a maximum of 6 [known]\nconsecutive days in male-female association). Conclusion Here, we report evidence for sexual conflict over associations\nin orangutans. We conclude that females incur costs from Page 19 of 22 6 Behav Ecol Sociobiol (2021) 75: 6 Informed consent\nNot applicable orangutans appears to be the result of a trade-off between the\ncosts of association and the necessity for paternity confusion\n(van Schaik et al. 2004; Knott et al. 2010, 2019). Future work\nwill have to elaborate on the details of this hypothesis. Open Access This article is licensed under a Creative Commons\nAttribution 4.0 International License, which permits use, sharing, adap-\ntation, distribution and reproduction in any medium or format, as long as\nyou give appropriate credit to the original author(s) and the source, pro-\nvide a link to the Creative Commons licence, and indicate if changes were\nmade. The images or other third party material in this article are included\nin the article's Creative Commons licence, unless indicated otherwise in a\ncredit line to the material. If material is not included in the article's\nCreative Commons licence and your intended use is not permitted by\nstatutory regulation or exceeds the permitted use, you will need to obtain\npermission directly from the copyright holder. To view a copy of this\nlicence, visit http://creativecommons.org/licenses/by/4.0/. Supplementary Information The online version contains supplementary\nmaterial available at https://doi.org/10.1007/s00265-020-02948-4. Acknowledgments We thank our local field teams at Suaq and Tuanan\nand all the local and foreign students and researchers for their contribution\nin the long-term data collection. Particularly, we express our gratitude to Dr. Alison Ashbury, Rebecca Brittain, Manon Bodin, Lynda P. Dunkel,\nCaroline Fryns, Dr. Anna Marzec, Dr. Caroline Schuppli, Dr. Brigitte\nSpillmann, and Sofia Vileila. We thank Universitas Nasional (UNAS) for\nsupport and collaboration and particularly Dr. Tatang Mitra Setia, Astri\nZulfa, Misdi bin Abdullah, and Kristana P. Makur. We thank the\nIndonesian State Ministry for Research and Technology (RISTEK), the\nIndonesian Institute of Sciences (LIPI), the Ministry of Environment and\nForestry (KLHK), the Ministry of Internal Affairs, Indonesia, the local\ngovernment in Central Kalimantan, the BKSDA Palangkaraya, the\nBornean Orangutan Survival Foundation (BOSF), MAWAS in\nPalangkaraya, the Sumatran orangutan conservation Program (SOCP),\nand Balai Besar Taman Nasional Gunung Leuser (BBTNGL) in Medan\nfor their permission and support to conduct this study. We are grateful to\nAndrea Heistermann, Dr. Gholib, and Joshua Reukauf for their support\nduring fecal sample analyses. Fecal sample collection was conducted ac-\ncording to Indonesian and international regulations and with the kind per-\nmission of the Indonesian Ministry of Environment and Forestry (KLHK)\n(permit numbers: 17256/IV/SA TS-LN/2012, SK.49/KSDAE/SET/KSA.2/\n1/2017). We thank Kevin Langergraber and two anonymous reviewers for\nhelpful comments. References Amrein M, Heistermann M, Weingrill T (2014) The effect of fission–\nfusion zoo housing on hormonal and behavioral indicators of stress\nin Bornean orangutans (Pongo pygmaeus). Int J Primatol 35:509–\n528 Archie EA, Altmann J, Alberts SC (2014) Costs of reproduction in a\nlong-lived female primate: injury risk and wound healing. Behav\nEcol Sociobiol 68:1183–1193 Ashbury AM, Willems EP, Utami Atmoko SS, Saputra F, van Schaik CP,\nvan Noordwijk MA (2020) Home range establishment and the\nmechanisms of philopatry among female Bornean orangutans\n(Pongo pygmaeus wurmbii) at Tuanan. Behav Ecol Sociobiol 74:42 go pygmaeus wurmbii) at Tuanan. Behav Ecol Sociobiol 74:4 Bates D, Maechler M, Bolker B, Walker S (2015) Fitting linear mixed-\neffects models using lme4. J Stat Softw 67:1–48. https://doi.org/10. 18637/jss.v067.i01 Beaudrot LH, Kahlenberg SM, Marshall AJ (2009) Why male orangutans\ndo not kill infants. Behav Ecol Sociobiol 63:1549–1562. https://doi. org/10.1007/s00265-009-0827-1 Funding Open Access funding provided by University of Zurich. This\nwork was supported by the University of Zurich, the A. H. Schultz\nFoundation, Janggen Pöhn Foundation and the Swiss National Science\nFoundation (grant no. 310030B_160363/1). Boesch C, Boesch-Achermann H (2000) The chimpanzees of the Taï\nForest: behavioural ecology and evolution. Oxford University\nPress, New York Data availability The main data sets generated and analyzed during the\ncurrent study are available in the Harvard Dataverse repository, https://\ndoi.org/10.7910/DVN/WXDVF6. Cadilek MJ (2009) Female-male association patterns, mating conflict and\nhormonal stress response in Bornean orangutan females (Pongo\npygmaeus wurmbii). MSc Thesis, University of Zurich Clay Z, Furuichi T, de Waal FBM (2016) Obstacles and catalysts to\npeaceful coexistence in chimpanzees and bonobos. Behaviour 153:\n1293–1330. https://doi.org/10.1163/1568539X-00003335 Compliance with ethical standards Conflict of interest\nThe authors declare that they have no conflict of\ninterest. Conflict of interest\nThe authors declare that they have no conflict of\ninterest. Clutton-Brock TH, Parker GA (1992) Potential reproductive rates and the\noperation of sexual selection. Q Rev Biol 67:437–456 Darwin C (1871) The descent of man, and selection in relation to sex. Murray, London Ethical approval\nThis is an observational study and fecal samples were\ncollected non-invasively. Observers did not interact with the wild focal\nindividuals in any way and kept a minimum distance of 10 m in order to\nminimize any effect on their natural behavior. The data collection proto-\ncol for this study adheres to legal requirements of Indonesia and was\napproved by the Indonesian State Ministry for Research and\nTechnology (RISTEK) (permit number 66/SIP/FRP/E5/Dit.KI/III/\n2016), the Directorate General of Natural Resources and Ecosystem\nConservation- Ministry of Environment and Forestry of Indonesia\n(KSDAE-KLHK), the Ministry of Internal Affairs, Indonesia, the\nNature Conservation Agency of Central Kalimantan (BKSDA) and\nBalai Besar Taman Nasional Gunung Leuser (BBTNGL). Fecal sample\ncollection was conducted according to Indonesian and international reg-\nulations and with the permission of the Indonesian Ministry of\nEnvironment and Forestry (KLHK) (permit numbers: 17256/IV/SA TS-\nLN/2012, SK.49/KSDAE/SET/KSA.2/1/2017). Delgado RA, van Schaik CP (2000) The behavioral ecology and conser-\nvation of the orangutan (Pongo pygmaeus): a tale of two islands. Evol Anthropol 9:201–218. https://doi.org/10.1002/1520-\n6505(2000)9:5<201::AID-EVAN2>3.0.CO;2-Y Deschner T, Boesch C (2007) Can the patterns of sexual swelling cycles\nin female Taï chimpanzees be explained by the cost-of-sexual-\nattraction hypothesis? Int J Primatol 28:389–406. https://doi.org/\n10.1007/s10764-007-9120-1 yp\n10.1007/s10764-007-9120-1 Deschner T, Heistermann M, Hodges K, Boesch C (2004) Female sexual\nswelling size, timing of ovulation, and male behavior in wild West\nAfrican chimpanzees. Horm Behav 46:204–215. https://doi.org/10. 1016/j.yhbeh.2004.03.013 Douglas PH, Hohmann G, Murtagh R, Thiessen-Bock R, Deschner T\n(2016) Mixed messages: wild female bonobos show high variability\nin the timing of ovulation in relation to sexual swelling patterns. BMC\nEvol Biol 16:140. https://doi.org/10.1186/s12862-016-0691-3 6 Page 20 of 22 Behav Ecol Sociobiol (2021) 75: 6 Drea CM (2015) D’scent of man: a comparative survey of primate\nchemosignaling in relation to sex. Horm Behav 68:117–133. https://doi.org/10.1016/j.yhbeh.2014.08.001 Harrison ME, Morrogh-Bernard HC, Chivers DJ (2010) Orangutan ener-\ngetics and the influence of fruit availability in the nonmasting peat-\nswamp forest of Sabangau, Indonesian Borneo. Int J Primatol 31:\n585–607. https://doi.org/10.1007/s10764-010-9415-5 Dunkel LP, Arora N, van Noordwijk MA, Utami Atmoko SS, Putra AP,\nKrützen M, van Schaik CP (2013) Variation in developmental arrest\namong male orangutans: a comparison between a Sumatran and a\nBornean population. Front Zool 10:12. https://doi.org/10.1186/\n1742-9994-10-12 Heistermann M, Ademmer C, Kaumanns W (2004) Ovarian cycle and\neffect of social changes on adrenal and ovarian function in Pygathrix\nnemaeus. Int J Primatol 25:689–708 Heistermann M, Palme R, Ganswindt A (2006) Comparison of different\nenzymeimmunoassays for assessment of adrenocortical activity in\nprimates based on fecal analysis. Am J Primatol 68:257–273 Emery Thompson M (2005) Reproductive endocrinology of wild female\nchimpanzees (Pan troglodytes schweinfurthii): methodological con-\nsiderations and the role of hormones in sex and conception. Am J\nPrimatol 67:137–158. https://doi.org/10.1002/ajp.20174 Hinde RA, Atkinson S (1970) Assessing the roles of social partners in\nmaintaining mutual proximity, as exemplified by mother-infant re-\nlations in rhesus monkeys. Anim Behav 18:169–176 Primatol 67:137–158. https://doi.org/10.1002/ajp.20174 Emery Thompson M, Georgiev AV (2014) The high price of success:\ncosts of mating effort in male primates. Int J Primatol 35:609–627. https://doi.org/10.1007/s10764-014-9790-4 Hodges KJ, Heistermann M (2011) Field endocrinology: monitoring hor-\nmonal changes in free-ranging primates. In: Setchell JM, Currtis DJ\n(eds) Field and laboratory methods in primatology: a practical guide. Cambridge University Press, Cambridge, pp 353–370 Emery Thompson M, Kahlenberg SM, Gilby IC, Wrangham RW (2007)\nCore area quality is associated with variance in reproductive success\namong female chimpanzees at Kibale National Park. Anim Behav\n73:501–512. https://doi.org/10.1016/j.anbehav.2006.09.007 Hohmann G, Fruth B (2003) Intra- and inter-sexual aggression by bono-\nbos in the context of mating. yp\n10.1007/s10764-007-9120-1 Behaviour 140:1389–1413 Emery Thompson M, Knott CD (2008) Urinary C-peptide of insulin as a\nnon-invasive marker of energy balance in wild orangutans. Horm\nBehav 53:526–535. https://doi.org/10.1016/j.yhbeh.2007.12.005 Hohmann G, Vigilant L, Mundry R, Behringer V, Surbeck M (2019)\nAggression by male bonobos against immature individuals does\nnot fit with predictions of infanticide. Aggressive Behav 45:300–\n309. https://doi.org/10.1002/ab.21819 Emery Thompson M, Muller MN, Kahlenberg SM, Wrangham RW\n(2010) Dynamics of social and energetic stress in wild female chim-\npanzees. Horm Behav 58:440–449. https://doi.org/10.1016/j.yhbeh. 2010.05.009 Hrdy SB (1979) Infanticide among animals: a review, classification, and\nexamination of the implications for the reproductive strategies of\nfemales. Ethol Sociobiol 1:13–40 Hrdy SB, Whitten PL (1987) Patterning of sexual behavior. In: Smuts\nBB, Cheney DL, Seyfarth RM, Wranghym RW, Struhsaker TT\n(eds) Primate Societies. University of Chicago Press, Chicago, pp\n370–384 Emery Thompson M, Muller MN, Wrangham RW (2014) Male chim-\npanzees compromise the foraging success of their mates in Kibale\nNational Park, Uganda. Behav Ecol Sociobiol 68:1973–1983. https://doi.org/10.1007/s00265-014-1803-y Emery Thompson M, Stumpf RM, Pusey AE (2008) Female reproductive\nstrategies and competition in apes: an introduction. Int J Primatol 29:\n815–821. https://doi.org/10.1007/s10764-008-9273-6 Knott CD (2009) Orangutans : sexual coercion without sexual violence. In: Muller MN, Wrangham RW (eds) Sexual coercion in primates\nand humans. Harvard University Press, Cambridge, pp 81–111 Emlen ST, Oring LW (1977) Ecology, sexual selection, and the evolution\nof mating systems. Science 197:215–223 Knott CD, Beaudrot LH, Snaith T, White S, Tschauner H, Planansky G\n(2008) Female-female competition in Bornean orangutans. Int J\nPrimatol 29:975–997. https://doi.org/10.1007/s10764-008-9278-1 Fox EA (1998) The function of female mate choice in Sumatran orang-\nutans (Pongo abelii). PhD thesis, Duke University Knott CD, Emery Thompson M, Stumpf RM, McIntyre MH (2010)\nFemale reproductive strategies in orangutans, evidence for female\nchoice and counterstrategies to infanticide in a species with frequent\nsexual coercion. Proc R Soc Lond B 277:105–113. https://doi.org/\n10.1098/rspb.2009.1552 Fox EA (2002) Female tactics to reduce sexual harassment in the\nSumatran orangutan (Pongo pygmaeus abelii). Behav Ecol\nSociobiol 52:93–101. https://doi.org/10.1007/s00265-002-0495-x Fox J, Weisberg S (2018) An R companion to applied regression. Sage\nPublications, Thousand Oakes, CA Knott CD, Emery Thompson M, Wich SA (2009) The ecology of female\nreproduction in wild orangutans. In: Wich SA, Utami Atmoko SS,\nMitra Setia T, van Schaik CP (eds) Orangutans: geographic varia-\ntion in behavioral ecology and conservation. yp\n10.1007/s10764-007-9120-1 Oxford University\nPress, New York, pp 171–188 Furuichi T (2011) Female contributions to the peaceful nature of bonobo\nsociety. Evol Anthropol 20:131–142. https://doi.org/10.1002/evan. 20308 Galdikas BMF (1981) Orangutan reproduction in the wild. In: Graham\n(ed) Reproductive biology of the great apes: comparative and bio-\nmedical perspectives. Academic Press, London, pp 281–300 Knott CD, Kahlenberg S (2007) Orangutans in perspective: forced cop-\nulations and female mating resistance. In: Bearder S, Campbell CJ,\nFuentes A, MacKinnon KC, Panger M (eds) Primates in perspective. Oxford University Press, Oxford, pp 290–305 Galdikas BMF (1985a) Adult male sociality and reproductive tactics\namong orangutans at Tanjung Puting. Folia Primatol 45:9–24 Galdikas BMF (1985b) Subadult male orangutan sociality and reproduc-\ntive behavior at Tanjung Puting. Am J Primatol 8:87–99 Knott CD, Scott AM, DiGiorgio A, Kane EE, Susanto TW, Riyandi R\n(2018) Are female orangutans less efficient foragers because of the\nrisk of sexual coercion? In: 27th International Primatological\nSociety Congress, Nairobi, Kenya (abstract) Galdikas BMF (1988) Orangutan diet, range, and activity at Tanjung\nPuting, Central Borneo. Int J Primatol 9:1–35 Society Congress, Nairobi, Kenya (abstract) Knott CD, Scott AM, O’Connell CA, Scott KS, Laman TG, Susanto TW\n(2019) Possible male infanticide in wild orangutans and a re-\nevaluation of infanticide risk. Sci Rep 9:7806. https://doi.org/10. 1038/s41598-019-42856-w Ganswindt A, Palme R, Heistermann M, Borragan S, Hodges JK (2003)\nNon-invasive assessment of adrenocortical function in the male\nAfrican elephant (Loxodonta africana) and its relation to musth. Gen Comp Endocr 134:156–166 Kunz JA (2020) Sexual conflict in orang-utans. PhD thesis, University of\nZurich Georgiev AV, Russell AF, Emery Thompson M, Otali E, Muller MN,\nWrangham RW (2014) The foraging costs of mating effort in male\nchimpanzees (Pan troglodytes schweinfurthii). Int J Primatol 35:\n725–745. https://doi.org/10.1007/s10764-014-9788-y Kuznetsova A, Brockhoff PB, Christensen RHB (2017) lmerTest pack-\nage: tests in linear mixed effects models. J Stat Softw 82:1–26 MacKinnon J (1974) The behaviour and ecology of wild orang-utans\n(Pongo pygmaeus). Anim Behav 22:3–74 Hardus ME, de Vries H, Dellatore DF, Lameira AR, Menken SBJ, Wich\nSA (2013) Socioecological correlates of inter-individual variation in\norangutan diets at Ketambe, Sumatra. Behav Ecol Sociobiol 67:\n429–437 Marty PR, van Noordwijk MA, Heistermann M, Willems EP, Dunkel LP,\nCadilek M, Agil M, Weingrill T (2015) Endocrinological correlates Behav Ecol Sociobiol (2021) 75: 6 Page 21 of 22 of male bimaturism in wild Bornean orangutans. Am J Primatol 77:\n1170–1178. https://doi.org/10.1002/ajp.22453 bonobos. Behav Ecol 118:1–15. yp\n10.1007/s10764-007-9120-1 Oxford University Press, New York, pp 245–253 Schultz AH (1938) Genital swelling in the female orang-utan. J Mammal\n19:363–366 Schürmann CL, van Hooff JARAM (1986) Reproductive strategies of the\norang-utan: new data and a reconsideration of existing sociosexual\nmodels. Int J Primatol 7:265–287 Mitra Setia T, van Schaik CP (2007) The response of adult orang-utans to\nflanged male long calls: inferences about their function. Folia\nPrimatol 78:215–226. https://doi.org/10.1159/000102317 Scott AM, Knott CD, Susanto TW (2019) Are male orangutans a threat to\ninfants? Evidence of mother–offspring counterstrategies to infanti-\ncide in Bornean orangutans (Pongo pygmaeus wurmbii). Int J\nPrimatol 40:435–455 Morrogh-Bernard HC, Husson SJ, Knott CD et al (2009) Orangutan\nactivity budgets and diet. In: Wich SA, Utami Atmoko SS, Mitra\nSetia T, van Schaik CP (eds) Orangutans: geographic variation in\nbehavioral ecology and conservation. Oxford University Press, New\nYork, pp 119–134 Silk J, Cheney D, Seyfarth R (2013) A practical guide to the study of\nsocial relationships. Evol Anthropol 22:213–225. https://doi.org/10. 1002/evan.21367 Muller MN, Emery Thompson M, Kahlenberg SM, Wrangham RW\n(2011) Sexual coercion by male chimpanzees shows that female\nchoice may be more apparent than real. Behav Ecol Sociobiol 65:\n921–933. https://doi.org/10.1007/s00265-010-1093-y Slater KY, Schaffner CM, Aureli F (2008) Female-directed male aggres-\nsion in wild Ateles geoffroyi yucatanensis. Int J Primatol 29:1657–\n1669. https://doi.org/10.1007/s10764-008-9311-4 Smuts BB, Smuts RW (1993) Male aggression and sexual coercion of\nfemales in nonhuman primates and other mammals: evidence and\ntheoretical implications. Adv Stud Behav 22:1–63 Muller MN, Kahlenberg SM, Emery Thompson M, Wrangham RW\n(2007) Male coercion and the costs of promiscuous mating for fe-\nmale chimpanzees. Proc R Soc Lond B 274:1009–1014. https://doi. org/10.1098/rspb.2006.0206 Spillmann B, Dunkel LP, van Noordwijk MA, Amda RNA, Lameira AR,\nWich SA, van Schaik CP (2010) Acoustic properties of long calls\ngiven by flanged male orang-utans (Pongo pygmaeus wurmbii) re-\nflect both individual identity and context. Ethology 116:385–395. https://doi.org/10.1111/j.1439-0310.2010.01744.x Muller MN, Kahlenberg SM, Wrangham RW (2009) Male aggression\nand sexual coercion of females in primates. In: Muller MN,\nWrangham RW (eds) Sexual coercion in primates and humans. Harvard University Press, Cambridge, pp 3–22 Spillmann B, Willems EP, van Noordwijk MA, Mitra Setia T, van Schaik\nCP (2017) Confrontational assessment in the roving male promis-\ncuity mating system of the Bornean orangutan. Behav Ecol\nSociobiol 71:20. https://doi.org/10.1007/s00265-016-2252-6 Nadler RD (1981) Laboratory research on sexual behaviour of the great\napes. yp\n10.1007/s10764-007-9120-1 https://doi.org/10.1093/beheco/\nary118 Marzec AM, Kunz JA, Falkner S, Utami Atmoko SS, Alavi SE,\nMoldawer AM, Vogel ER, Schuppli C, van Schaik CP, van\nNoordwijk MA (2016) The dark side of the red ape: male-\nmediated lethal female competition in Bornean orangutans. Behav\nEcol Sociobiol 70:459–466. https://doi.org/10.1007/s00265-015-\n2053-3 Paoli T (2009) The absence of sexual coercion in bonobos. In: Muller\nMN, Wrangham RW (eds) Sexual coercion in primates and humans. Harvard University Press, Cambridge, pp 410–423 Parker GA (1979) Sexual selection and sexual conflict. In: Blum MS,\nBlum NA (eds) Sexual selection and reproductive competition in\ninsects. Academic Press, London, pp 123–166 Matsumoto-Oda A (1998) Injuries to the sexual skin of female chimpan-\nzees at Mahale and their effect on behaviour. Folia Primatol 69:400–\n404 Pradhan GR, van Noordwijk MA, van Schaik CP (2012) A model for the\nevolution of developmental arrest in male orangutans. Am J Phys\nAnthropol 149:18–25. https://doi.org/10.1002/ajpa.22079 Matsumoto-Oda A, Oda R, Hayashi Y, Murakami H, Maeda N,\nKumazaki K, Shimizu K, Matsuzawa T (2003) Vaginal fatty acids\nproduced by chimpanzees during menstrual cycles. Folia Primatol\n74:75–79 R Core Team (2018) R: a language and environment for statistical com-\nputing. R Foundation for Statistical Computing, Vienna http://www. R-project.org Riedel J, Franz M, Boesch C (2011) How feeding competition determines\nfemale chimpanzee gregariousness and ranging in the Taï National\nPark, Côte d’Ivoire. Am J Primatol 73:305–313 Mesnick SL (1997) Sexual alliances: evidence and evolutionary implica-\ntions. In: Gowaty PA (ed) Feminism and evolutionary biology:\nboundaries, intersections, and frontiers. Springer US, Boston, pp\n207–260 Rijksen HD (1978) A field study on Sumatran orangutans (Pongo\npygmaeus abelii): ecology, behavior and conservation. H.Veenman and Zonen, Wageningen Mitani JC (1985) Mating behaviour of male orangutans in the Kutai\nGame Reserve, Indonesia. Anim Behav 33:392–402 Roth TS, Rianti P, Fredriksson GM, Wich SA, Nowak MG (2020)\nGrouping behavior of Sumatran orangutans (Pongo abelii) and\nTapanuli orangutans (Pongo tapanuliensis) living in forest with\nlow fruit abundance. Am J Primatol 82:e23123. https://doi.org/10. 1002/ajp.23123 Mitani JC (1989) Orangutan activity budgets: monthly variations and the\neffects of body size, parturition, and sociality. Am J Primatol 18:87–\n100 Mitra Setia T, Delgado RA, Utami Atmoko SS, Singleton I, van Schaik\nCP (2009) Social organization and male-female relationships. In:\nWich SA, Utami Atmoko SS, Mitra Setia T, van Schaik CP (eds)\nOrangutans: geographic variation in behavioral ecology and conser-\nvation. yp\n10.1007/s10764-007-9120-1 1016/j.jhevol.2018.09.004 Wich SA, Vogel ER, Larsen MD, Fredriksson G, Leighton M, Yeager\nCP, Brearley FQ, van Schaik CP, Marshall AJ (2011) Forest fruit\nproduction is higher on Sumatra than on Borneo. PLoS One 6:\ne21278. https://doi.org/10.1371/journal.pone.0021278 van Noordwijk MA, van Schaik CP (2000) Reproductive patterns in\neutherian mammals: adaptations against infanticide? In: van\nSchaik CP, Janson CH (eds) Infanticide by males and its implica-\ntions. Cambridge University Press, Cambridge, pp 322–360 Wickham H (2016) ggplot2: elegant graphics for data analysis. Springer-\nVerlag, New York van Noordwijk MA, van Schaik CP (2009) Intersexual food transfer\namong orangutans: do females test males for coercive tendency? Behav Ecol Sociobiol 63:883–890. https://doi.org/10.1007/\ns00265-009-0728-3 Wilke CO (2019) cowplot: streamlined plot theme and plot annotations\nfor “ggplot2.” http://bioconductor.statistik.tu-dortmund.de/cran/\nweb/packages/cowplot/cowplot.pdf van Noordwijk MA, Willems EP, Utami Atmoko SS, Kuzawa CW, van\nSchaik CP (2013) Multi-year lactation and its consequences in\nBornean orangutans (Pongo pygmaeus wurmbii). Behav Ecol\nSociobiol 67:805–814. https://doi.org/10.1007/s00265-013-1504-y Wilson ML, Boesch C, Fruth B, Furuichi T, Gilby IC, Hashimoto C,\nHobaiter CL, Hohmann G, Itoh N, Koops K, Lloyd JN,\nMatsuzawa T, Mitani JC, Mjungu DC, Morgan D, Muller MN,\nMundry R, Nakamura M, Pruetz J, Pusey AE, Riedel J, Sanz C,\nSchel AM, Simmons N, Waller M, Watts DP, White F, Wittig\nRM, Zuberbühler K, Wrangham RW (2014) Lethal aggression in\nPan is better explained by adaptive strategies than human impacts. Nature 513:414–417. https://doi.org/10.1038/nature13727 van Schaik CP (2016) The primate origins of human nature. Wiley,\nHoboken van Schaik CP (1999) The socioecology of fission-fusion sociality in\norangutans. Primates 40:69–86. https://doi.org/10.1007/\nBF02557703 van Schaik CP, Fox EA (1996) Temporal variability on orangutan gre-\ngariousness revisited. In: Proceedings from the Sixteenth Congress\nof the International Primatological Society held in Madison,\nWisconsin (abstract) Wrangham RW (2002) The cost of sexual attraction: is there a trade-off in\nfemale Pan between sex appeal and received coercion? In: Boesch\nC, Hohmann G, Marchand LF (eds) Behavioural diversity in chim-\npanzees and bonobos. Cambridge University Press, Cambridge, pp\n204–215 van Schaik CP, Hodges JK, Nunn CL (2000) Paternity confusion and the\novarian cycles of female primates. In: van Schaik CP, Janson CH\n(eds) Infanticide by males its implications. Cambridge University\nPress, Cambridge, pp 361–387 Zinner DP, Nunn CL, van Schaik CP, Kappeler PM (2004) Sexual selec-\ntion and exaggerated sexual swellings of female primates. In:\nKappeler PM, van Schaik CP (eds) Sexual selection in primates:\nnew and comparative perspectives. yp\n10.1007/s10764-007-9120-1 In: Kappeler PM, van\nSchaik CP (eds) Sexual selection in primates: new and comparative\nperspectives. Cambridge University Press, Cambridge, pp 196–207 Vogel ER, Alavi SE, Utami Atmoko SS, van Noordwijk MA, Bransford\nTD, Erb WM, Zulfa A, Sulistyo F, Farida WR, Rothman JM (2017)\nNutritional ecology of wild Bornean orangutans (Pongo pygmaeus\nwurmbii) in a peat swamp habitat: effects of age, sex, and season. Am J Primatol 79:1–20. https://doi.org/10.1002/ajp.22618 Utami Atmoko SS, Wich SA, Sterck EHM, van Hooff JA (1997) Food\ncompetition between wild orangutans in large fig trees. Int J\nPrimatol 18:909–927 Vogel ER, Harrison ME, Zulfa A, Bransford TD, Alavi SE, Husson S,\nMorrogh-Bernard HC, Firtsman T, Utami Atmoko SS, van\nNoordwijk MA (2015) Nutritional differences between two orang-\nutan habitats: implications for population density. PLoS One 10:\ne0138612. https://doi.org/10.5061/dryad.sk616 van de Pol M, Wright J (2009) A simple method for distinguishing\nwithin-versus between-subject effects using mixed models. Anim\nBehav 77:753–758 van Noordwijk MA, Arora N, Willems EP, Dunkel LP, Amda RNA,\nMardianah N, Ackermann C, Krützen M, van Schaik CP (2012)\nFemale philopatry and its social benefits among Bornean orangu-\ntans. Behav Ecol Sociobiol 66:823–834. https://doi.org/10.1007/\ns00265-012-1330-7 Wartmann FM, Purves RS, van Schaik CP (2010) Modelling ranging\nbehaviour of female orang-utans: a case study in Tuanan, Central\nKalimantan, Indonesia. Primates 51:119–130. https://doi.org/10. 1007/s10329-009-0186-6 van Noordwijk MA, Sauren SEB, Nuzuar AA, Morrogh-Bernard HC,\nUtami Atmoko SS, van Schaik CP (2009) Development of indepen-\ndence. In: Wich SA, Utami Atmoko SS, Mitra-Setia T, van Schaik\nCP (eds) Orangutans: geographic variation in behavioral ecology\nand conservation. Oxford University Press, New York, pp 189–203 Weingrill T, Willems EP, Zimmermann N, Steinmetz H, Heistermann M\n(2011) Species-specific patterns in fecal glucocorticoid and andro-\ngen levels in zoo-living orangutans (Pongo spp.). Gen Comp Endocr\n172:446–457. https://doi.org/10.1016/j.ygcen.2011.04.008 Wich SA, Geurts ML, Mitra Setia T (2006) Influence of fruit availability\non Sumatran orangutan sociality and reproduction. In: Hohmann G,\nRobbins MM, Boesch C (eds) Feeding ecology in apes and other\nprimates. Ecological, physical and behavioral aspects. Cambridge\nUniversity Press, New York, pp 335–356 van Noordwijk MA, Utami Atmoko SS, Knott CD, Kuze N, Morrogh-\nBernard HC, Oram F, Schuppli C, van Schaik CP, Willems EP\n(2018) The slow ape: high infant survival and long interbirth inter-\nvals in wild orangutans. J Hum Evol 125:38–49. https://doi.org/10. yp\n10.1007/s10764-007-9120-1 In: Graham C (ed) Reproductive biology of the great apes:\ncomparative and biomedical perspectives. Academic Press, New\nYork, pp 191–238 Stumpf RM, Boesch C (2010) Male aggression and sexual coercion in\nwild West African chimpanzees, Pan troglodytes verus. Anim\nBehav 79:333–342. https://doi.org/10.1016/j.anbehav.2009.11.008 Nieuwenhuis R, te Grotenhuis M, Pelzer B (2012) influence.ME: tools\nfor detecting influential data in mixed effects models. R J 4:38–47 Nugraha TP, Heistermann M, Agil M, Purwantara B, Supriatna I, Gholib\nG, van Schaik CP, Weingrill T (2016) Validation of a field-friendly\nextraction and storage method to monitor fecal steroid metabolites in\nwild orangutans. Primates 58:285–294. https://doi.org/10.1007/\ns10329-016-0583-6 Stumpf RM, Emery Thompson M, Knott CD (2008) A comparison of\nfemale mating strategies in Pan troglodytes and Pongo spp. Int J\nPrimatol 29:865–884. https://doi.org/10.1007/s10764-008-9284-3 Sugardjito J, te Boekhorst IJA, van Hooff JA (1987) Ecological con-\nstraints on the grouping of wild orang-utans (Pongo pygmaeus) in\nthe Gunung Leuser National Park, Sumatra, Indonesia. Int J\nPrimatol 8:17–41 Nunn CL (1999) The evolution of exaggerated sexual swellings in pri-\nmates and the graded-signal hypothesis. Anim Behav 58:229–246. https://doi.org/10.1006/anbe.1999.1159 Therneau T (2018) coxme: mixed effects Cox models, https://cran.r-\nproject.org/package=coxme Nurmi NO, Hohmann G, Goldstone LG, Deschner T (2018) The “tolerant\nchimpanzee”— towards the costs and benefits of sociality in female 6 Page 22 of 22 Behav Ecol Sociobiol (2021) 75: 6 van Schaik CP, van Hooff J (1983) On the ultimate causes of primate\nsocial systems. Behaviour 85:91–117 Trivers RL (1972) Parental investment and sexual selection. In: Campbell\nB (ed) Sexual selection and the descent of man, 1871–1971. Aldine,\nChicago, pp 136–179 van Schaik CP, van Noordwijk MA, Vogel ER (2009) Ecological sex\ndifferences in wild orangutans. In: Wich SA, Utami Atmoko SS,\nMitra Setia T, van Schaik CP (eds) Orangutans: geographic varia-\ntion in behavioral ecology and conservation. Oxford University\nPress, New York, pp 255–268 Utami Atmoko SS, Mitra Setia T, Goossens B, James SS, Knott CD,\nMorrogh-Bernard HC, van Schaik CP, van Noordwijk MA (2009)\nOrangutan mating behavior and strategies. In: Wich SA, Utami\nAtmoko SS, Mitra Setia T, van Schaik CP (eds) Orangutans: geo-\ngraphic variation in behavioral ecology and conservation. Oxford\nUniversity Press, New York, pp 235–244 van Schaik CP, van Noordwijk MA, Wich SA (2006) Innovation in wild\nBornean orangutans (Pongo pygmaeus wurmbii). Behaviour 143:\n839–876. https://doi.org/10.1163/156853906778017944 Utami Atmoko SS, van Hooff JA (2004) Alternative male reproductive\nstrategies: male bimaturism in orangutans. Publisher’s note Springer Nature remains neutral with regard to jurisdic-\ntional claims in published maps and institutional affiliations. yp\n10.1007/s10764-007-9120-1 Cambridge University Press,\nNew York, pp 71–89 van Schaik CP, Janson CH (eds) (2000) Infanticide by males and its\nimplications. Cambridge University Press, Cambridge van Schaik CP, Pradhan GR, van Noordwijk MA (2004) Mating conflict\nin primates: infanticide, sexual harassment and female sexuality. In:\nKappeler PM, van Schaik CP (eds) Sexual selection in primates:\nnew and comparative perspectives. Cambridge University Press,\nCambridge, pp 141–163 Publisher’s note Springer Nature remains neutral with regard to jurisdic-\ntional claims in published maps and institutional affiliations."