Q
stringlengths
4
3.96k
A
stringlengths
1
3k
Result
stringclasses
4 values
Corollary 6. A set \( A \subseteq \left( {S,\rho }\right) \) clusters at \( p \) iff each globe \( {G}_{p} \) (about \( p \) ) contains at least one point of \( A \) other than \( p{.}^{5} \)
Indeed, assume the latter. Then, in particular, each globe\n\n\[ \n{G}_{p}\left( \frac{1}{n}\right) ,\;n = 1,2,\ldots ,\n\]\n\ncontains some point of \( A \) other than \( p \) ; call it \( {x}_{n} \) . We can make the \( {x}_{n} \) distinct by choosing each time \( {x}_{n + 1} \) closer to \( p \) than \( {x}_{n} \) is. It easily follows that each \( {G}_{p}\left( \varepsilon \right) \) contains infinitely many points of \( A \) (the details are left to the reader), as required. The converse is obvious.
No
Theorem 1. Let \( {x}_{m} \rightarrow q,{y}_{m} \rightarrow r \), and \( {a}_{m} \rightarrow a \) in \( {E}^{1} \) or \( C \) (the complex field). Then\n\n(i) \( {x}_{m} \pm {y}_{m} \rightarrow q \pm r \) ;
Proof. (i) By formula (2) of \( §{14} \), we must show that\n\n\[ \left( {\forall \varepsilon > 0}\right) \left( {\exists k}\right) \left( {\forall m > k}\right) \;\left| {{x}_{m} \pm {y}_{m} - \left( {q \pm r}\right) }\right| < \varepsilon . \]\n\nThus we fix an arbitrary \( \varepsilon > 0 \) and look for a suitable \( k \) . Since \( {x}_{m} \rightarrow q \) and \( {y}_{m} \rightarrow r \), there are \( {k}^{\prime } \) and \( {k}^{\prime \prime } \) such that\n\n\[ \left( {\forall m > {k}^{\prime }}\right) \;\left| {{x}_{m} - q}\right| < \frac{\varepsilon }{2} \]\n\nand\n\n\[ \left( {\forall m > {k}^{\prime \prime }}\right) \;\left| {{y}_{m} - r}\right| < \frac{\varepsilon }{2} \]\n\n(as \( \varepsilon \) is arbitrary, we may as well replace it by \( \frac{1}{2}\varepsilon \) ). Then both inequalities hold for \( m > k, k = \max \left( {{k}^{\prime },{k}^{\prime \prime }}\right) \) . Adding them, we obtain\n\n\[ \left( {\forall m > k}\right) \;\left| {{x}_{m} - q}\right| + \left| {{y}_{m} - r}\right| < \varepsilon . \]\n\nHence by the triangle law,\n\n\[ \left| {{x}_{m} - q \pm \left( {{y}_{m} - r}\right) }\right| < \varepsilon \text{, i.e.,}\left| {{x}_{m} \pm {y}_{m} - \left( {q \pm r}\right) }\right| < \varepsilon \text{for}m > k\text{,} \]\n\nas required.\n\nThis proof of (i) applies to sequences of vectors as well, without any change.
Yes
Corollary 1. Suppose \( \lim {x}_{m} = p \) and \( \lim {y}_{m} = q \) exist in \( {E}^{ * } \). (a) If \( p > q \), then \( {x}_{m} > {y}_{m} \) for all but finitely many \( m \). (b) If \( {x}_{m} \leq {y}_{m} \) for infinitely many \( m \), then \( p \leq q \) ; i.e., \( \lim {x}_{m} \leq \lim {y}_{m} \).
This is known as passage to the limit in inequalities. Caution: The strict inequalities \( {x}_{m} < {y}_{m} \) do not imply \( p < q \) but only \( p \leq q \). For example, let \[ {x}_{m} = \frac{1}{m}\text{ and }{y}_{m} = 0 \] Then \[ \left( {\forall m}\right) \;{x}_{m} > {y}_{m} \] yet \( \lim {x}_{m} = \lim {y}_{m} = 0 \) .
No
Corollary 2. Let \( {x}_{m} \rightarrow p \) in \( {E}^{ * } \), and let \( c \in {E}^{ * } \) (finite or not). Then the following are true:\n\n(a) If \( p > c \) (respectively, \( p < c \) ), we have \( {x}_{m} > c\left( {{x}_{m} < c}\right) \) for all but finitely many \( m \) .\n\n(b) If \( {x}_{m} \leq c \) (respectively, \( {x}_{m} \geq c \) ) for infinitely many \( m \), then \( p \leq c\left( {p \geq c}\right) \) .
One can prove this from Corollary 1, with \( {y}_{m} = c \) (or \( {x}_{m} = c \) ) for all \( m \) .
No
A sequence \( \left\{ {x}_{m}\right\} \subseteq \left( {S,\rho }\right) \) clusters at a point \( p \in S \) iff it has a subsequence \( \left\{ {x}_{{m}_{n}}\right\} \) converging to \( p{.}^{1} \)
If \( p = \mathop{\lim }\limits_{{n \rightarrow \infty }}{x}_{{m}_{n}} \), then by definition each globe about \( p \) contains all but finitely many \( {x}_{{m}_{n}} \), hence infinitely many \( {x}_{m} \) . Thus \( p \) is a cluster point.\n\nConversely, if so, consider in particular the globes\n\n\[ \n{G}_{p}\left( \frac{1}{n}\right) ,\;n = 1,2,\ldots \n\]\n\nBy assumption, \( {G}_{p}\left( 1\right) \) contains some \( {x}_{m} \) . Thus fix\n\n\[ \n{x}_{{m}_{1}} \in {G}_{p}\left( 1\right) \n\]\n\nNext, choose a term\n\n\[ \n{x}_{{m}_{2}} \in {G}_{p}\left( \frac{1}{2}\right) \text{ with }{m}_{2} > {m}_{1}. \n\]\n\n(Such terms exist since \( {G}_{p}\left( \frac{1}{2}\right) \) contains infinitely many \( {x}_{m} \) .) Next, fix\n\n\[ \n{x}_{{m}_{3}} \in {G}_{p}\left( \frac{1}{3}\right) \text{, with }{m}_{3} > {m}_{2} > {m}_{1}, \n\]\n\nand so on.\n\nThus, step by step (inductively), select a sequence of subscripts\n\n\[ \n{m}_{1} < {m}_{2} < \cdots < {m}_{n} < \cdots \n\]\n\nthat determines a subsequence (see Chapter 1, §8) such that\n\n\[ \n\left( {\forall n}\right) \;{x}_{{m}_{n}} \in {G}_{p}\left( \frac{1}{n}\right) \text{, i.e.,}\rho \left( {{x}_{{m}_{n}}, p}\right) < \frac{1}{n} \rightarrow 0\text{,} \n\]\n\n--- \n\n\( {}^{1} \) Therefore, cluster points of \( \left\{ {x}_{m}\right\} \) are also called subsequential limits.\n\n--- \n\nwhence \( \rho \left( {{x}_{{m}_{n}}, p}\right) \rightarrow 0 \), or \( {x}_{{m}_{n}} \rightarrow p \) . (Why?) Thus we have found a subsequence \( {x}_{{m}_{n}} \rightarrow p \), and assertion (i) is proved.
Yes
(i) Each bounded infinite set or sequence \( A \) in \( {E}^{n}\left( {{}^{ * }\text{or}\left. {C}^{n}\right) }\right. \) has at least one cluster point \( \bar{p} \) there (possibly outside \( A \) ).
Proof. Take first a bounded sequence \( \left\{ {z}_{m}\right\} \subseteq \left\lbrack {a, b}\right\rbrack \) in \( {E}^{1} \) . Let\n\n\[ p = \overline{\lim }{z}_{m} \]\n\nBy Theorem 2(i) of Chapter 2, \( §{13},\left\{ {z}_{m}\right\} \) clusters at \( p \) . Moreover, as\n\n\[ a \leq {z}_{m} \leq b \]\n\nwe have\n\n\[ a \leq \inf {z}_{m} \leq p \leq \sup {z}_{m} \leq b \]\n\nby Corollary 1 of Chapter 2, §13. Thus\n\n\[ p \in \left\lbrack {a, b}\right\rbrack \subseteq {E}^{1} \]\n\nand so \( \left\{ {z}_{m}\right\} \) clusters in \( {E}^{1} \) .\n\nAssertion (ii) now follows—for \( {E}^{1} \) —by Theorem 1(i) above.
Yes
Theorem 3. We have \( p \in \bar{A} \) in \( \left( {S,\rho }\right) \) iff each globe \( {G}_{p}\left( \delta \right) \) about \( p \) meets \( A \) , i.e.,\n\n\[ \left( {\forall \delta > 0}\right) \;A \cap {G}_{p}\left( \delta \right) \neq \varnothing . \]\n\nEquivalently, \( p \in \bar{A} \) iff\n\n\[ p = \mathop{\lim }\limits_{{n \rightarrow \infty }}{x}_{n}\text{ for some }\left\{ {x}_{n}\right\} \subseteq A. \]
The proof is as in Corollary 6 of \( §{14} \) and Theorem 1. (Here, however, the \( {x}_{n} \) need not be distinct or different from \( p \) .) The details are left to the reader.
No
Theorem 4. A set \( A \subseteq \left( {S,\rho }\right) \) is closed iff one of the following conditions holds.\n\n(i) \( A \) contains all its cluster points (or has none); i.e., \( A \supseteq {A}^{\prime } \) .\n\n(ii) \( A = \bar{A} \) .\n\n(iii) A contains the limit of each convergent sequence \( \left\{ {x}_{n}\right\} \subseteq A \) (if any). \( {}^{2} \)
Proof. Parts (i) and (ii) are equivalent since\n\n\[ A \supseteq {A}^{\prime } \Leftrightarrow A = A \cup {A}^{\prime } = \bar{A}.\;\text{ (Explain!) } \]\n\nNow let \( A \) be closed. If \( p \notin A \), then \( p \in - A \) ; therefore, by Definition 3 in \( §{12} \), some \( {G}_{p} \) fails to meet \( A\left( {{G}_{p} \cap A = \varnothing }\right) \) . Hence no \( p \in - A \) is a cluster point, or the limit of a sequence \( \left\{ {x}_{n}\right\} \subseteq A \) . (This would contradict Definitions 1 and 2 of \( §{14} \) .) Consequently, all such cluster points and limits must be in \( A \), as claimed.\n\nConversely, suppose \( A \) is not closed, so \( - A \) is not open. Then \( - A \) has a noninterior point \( p \) ; i.e., \( p \in - A \) but no \( {G}_{p} \) is entirely in \( - A \) . This means that each \( {G}_{p} \) meets \( A \) . Thus\n\n\[ p \in \bar{A}\text{(by Theorem 3),} \]\n\nand\n\n\[ p = \mathop{\lim }\limits_{{n \rightarrow \infty }}{x}_{n}\text{ for some }\left\{ {x}_{n}\right\} \subseteq A\text{ (by the same theorem),} \]\n\neven though \( p \notin A \) (for \( p \in - A \) ).\n\nWe see that (iii) and (ii), hence also (i), fail if \( A \) is not closed and hold if \( A \) is closed. (See the first part of the proof.) Thus the theorem is proved.
No
Theorem 1. Every convergent sequence \( \left\{ {x}_{m}\right\} \subseteq \left( {S,\rho }\right) \) is Cauchy.
Proof. Let \( {x}_{m} \rightarrow p \) . Then given \( \varepsilon > 0 \), there is a \( k \) such that\n\n\[ \left( {\forall m > k}\right) \;\rho \left( {{x}_{m}, p}\right) < \frac{\varepsilon }{2}. \]\n\nAs this holds for any \( m > k \), it also holds for any other term \( {x}_{n} \) with \( n > k \) . Thus\n\n\[ \left( {\forall m, n > k}\right) \;\rho \left( {{x}_{m}, p}\right) < \frac{\varepsilon }{2}\text{ and }\rho \left( {p,{x}_{n}}\right) < \frac{\varepsilon }{2}. \]\n\nAdding and using the triangle inequality, we get\n\n\[ \rho \left( {{x}_{m},{x}_{n}}\right) \leq \rho \left( {{x}_{m}, p}\right) + \rho \left( {p,{x}_{n}}\right) < \varepsilon , \]\n\nand (1) is proved.
Yes
Theorem 2. Every Cauchy sequence \( \left\{ {x}_{m}\right\} \subseteq \left( {S,\rho }\right) \) is bounded.
Proof. We must show that all \( {x}_{m} \) are in some globe. First we try an arbitrary radius \( \varepsilon \) . Then by (1), there is \( k \) such that \( \rho \left( {{x}_{m},{x}_{n}}\right) < \varepsilon \) for \( m, n > k \) . Fix some \( n > k \) . Then\n\n\[ \left( {\forall m > k}\right) \rho \left( {{x}_{m},{x}_{n}}\right) < \varepsilon \text{, i.e.,}{x}_{m} \in {G}_{{x}_{n}}\left( \varepsilon \right) \text{.} \]\n\nThus the globe \( {G}_{{x}_{n}}\left( \varepsilon \right) \) contains all \( {x}_{m} \) except possibly the \( k \) terms \( {x}_{1},\ldots ,{x}_{k} \) . To include them as well, we only have to take a larger radius \( r \), greater than \( \rho \left( {{x}_{m},{x}_{n}}\right), m = 1,\ldots, k \) . Then all \( {x}_{m} \) are in the enlarged globe \( {G}_{{x}_{n}}\left( r\right) \) .
Yes
Theorem 3. If a Cauchy sequence \( \left\{ {x}_{m}\right\} \) clusters at a point \( p \), then \( {x}_{m} \rightarrow p \) .
Proof. We want to show that \( {x}_{m} \rightarrow p \), i.e., that\n\n\[ \left( {\forall \varepsilon > 0}\right) \left( {\exists k}\right) \left( {\forall m > k}\right) \;\rho \left( {{x}_{m}, p}\right) < \varepsilon . \]\n\nThus we fix \( \varepsilon > 0 \) and look for a suitable \( k \) . Now as \( \left\{ {x}_{m}\right\} \) is Cauchy, there is\na \( k \) such that\n\n\[ \left( {\forall m, n > k}\right) \;\rho \left( {{x}_{m},{x}_{n}}\right) < \frac{\varepsilon }{2}. \]\n\nAlso, as \( p \) is a cluster point, the globe \( {G}_{p}\left( \frac{\varepsilon }{2}\right) \) contains infinitely many \( {x}_{n} \), so we can fix one with \( n > k \) ( \( k \) as above). Then \( \rho \left( {{x}_{n}, p}\right) < \frac{\varepsilon }{2} \) and, as noted above, also \( \rho \left( {{x}_{m},{x}_{n}}\right) < \frac{\varepsilon }{2} \) for \( m > k \) . Hence\n\n\[ \left( {\forall m > k}\right) \;\rho \left( {{x}_{m},{x}_{n}}\right) + \rho \left( {{x}_{n}, p}\right) < \varepsilon , \]\n\nimplying \( \rho \left( {{x}_{m}, p}\right) \leq \rho \left( {{x}_{m},{x}_{n}}\right) + \rho \left( {{x}_{n}, p}\right) < \varepsilon \), as required.
Yes
Theorem 4 (Cauchy’s convergence criterion). A sequence \( \left\{ {\bar{x}}_{m}\right\} \) in \( {E}^{n} \) (*or \( \left. {C}^{n}\right) \) converges if and only if it is a Cauchy sequence.
Proof. If \( \left\{ {x}_{m}\right\} \) converges, it is Cauchy by Theorem 1.\n\nConversely, let \( \left\{ {x}_{m}\right\} \) be a Cauchy sequence. Then by Theorem 2, it is bounded. Hence by the Bolzano-Weierstrass theorem (Theorem 2 of \( §{16} \) ), it has a cluster point \( \bar{p} \) . Thus by Theorem 3 above, it converges to \( \bar{p} \), and all is proved.
Yes
Corollary 1. If \( A \) clusters at \( p \) in \( \left( {S,\rho }\right) \), then a function \( f : A \rightarrow \left( {T,{p}^{\prime }}\right) \) can have at most one limit at \( p \) ; i.e.,
Proof. Suppose \( f \) has two limits, \( q \) and \( r \), at \( p \) . By the Hausdorff property, \[ {G}_{q}\left( \varepsilon \right) \cap {G}_{r}\left( \varepsilon \right) = \varnothing \;\text{ for some }\varepsilon > 0. \] Also, by (2), there are \( {\delta }^{\prime },{\delta }^{\prime \prime } > 0 \) such that \[ \left( {\forall x \in A \cap {G}_{\neg p}\left( {\delta }^{\prime }\right) }\right) \;f\left( x\right) \in {G}_{q}\left( \varepsilon \right) \text{ and } \] \[ \left( {\forall x \in A \cap {G}_{\neg p}\left( {\delta }^{\prime \prime }\right) }\right) \;f\left( x\right) \in {G}_{r}\left( \varepsilon \right) . \] Let \( \delta = \min \left( {{\delta }^{\prime },{\delta }^{\prime \prime }}\right) \) . Then for \( x \in A \cap {G}_{\neg p}\left( \delta \right), f\left( x\right) \) is in both \( {G}_{q}\left( \varepsilon \right) \) and \( {G}_{r}\left( \varepsilon \right) \) , and such an \( x \) exists since \( A \cap {G}_{\neg p}\left( \delta \right) \neq \varnothing \) by assumption. But this is impossible since \( {G}_{q}\left( \varepsilon \right) \cap {G}_{r}\left( \varepsilon \right) = \varnothing \) (a contradiction!).
Yes
Corollary 2. \( f \) is continuous at \( p\left( {p \in {D}_{f}}\right) \) iff \( f\left( x\right) \rightarrow f\left( p\right) \) as \( x \rightarrow p \) .
The straightforward proof from definitions is left to the reader.
No
Theorem 1 (sequential criterion of continuity). (i) A function \[ f : A \rightarrow \left( {T,{\rho }^{\prime }}\right) ,\text{ with }A \subseteq \left( {S,\rho }\right) ,\] is continuous at a point \( p \in A \) iff for every sequence \( \left\{ {x}_{m}\right\} \subseteq A \) such that \( {x}_{m} \rightarrow p \) in \( \left( {S,\rho }\right) \), we have \( f\left( {x}_{m}\right) \rightarrow f\left( p\right) \) in \( \left( {T,{\rho }^{\prime }}\right) \). In symbols, \[ \left( {\forall \left\{ {x}_{m}\right\} \subseteq A \mid {x}_{m} \rightarrow p}\right) \;f\left( {x}_{m}\right) \rightarrow f\left( p\right) . \] \( \left( {1}^{\prime }\right) \)
Proof. We first prove (ii). Suppose \( q \) is a limit of \( f \) at \( p \), i.e. (see \( §1 \) ), \[ \left( {\forall \varepsilon > 0}\right) \left( {\exists \delta > 0}\right) \left( {\forall x \in A \cap {G}_{\neg p}\left( \delta \right) }\right) \;f\left( x\right) \in {G}_{q}\left( \varepsilon \right) . \] (2) Thus, given \( \varepsilon > 0 \), there is \( \delta > 0 \) (henceforth fixed) such that \[ f\left( x\right) \in {G}_{q}\left( \varepsilon \right) \text{ whenever }x \in A, x \neq p,\text{ and }x \in {G}_{p}\left( \delta \right) . \] (3) We want to deduce \( \left( {2}^{\prime }\right) \). Thus we fix any sequence \[ \left\{ {x}_{m}\right\} \subseteq A - \{ p\} ,{x}_{m} \rightarrow p.{}^{1} \] Then \[ \left( {\forall m}\right) \;{x}_{m} \in A\text{ and }{x}_{m} \neq p, \] and \( {G}_{p}\left( \delta \right) \) contains all but finitely many \( {x}_{m} \). Then these \( {x}_{m} \) satisfy the con- ditions stated in (3). Hence \( f\left( {x}_{m}\right) \in {G}_{q}\left( \varepsilon \right) \) for all but finitely many \( m \). As \( \varepsilon \) is arbitrary, this implies \( f\left( {x}_{m}\right) \rightarrow q \) (by the definition of \( \left. {\mathop{\lim }\limits_{{m \rightarrow \infty }}f\left( {x}_{m}\right) }\right) \), as is required in \( \left( {2}^{\prime }\right) \). Thus \( \left( 2\right) \Rightarrow \left( {2}^{\prime }\right) \). Conversely, suppose (2) fails, i.e., its negation holds. (See the rules for forming negations of such formulas in Chapter 1, §§1-3.) Thus \[ \left( {\exists \varepsilon > 0}\right) \left( {\forall \delta > 0}\right) \left( {\exists x \in A \cap {G}_{\neg p}\left( \delta \right) }\right) \;f\left( x\right) \notin {G}_{q}\left( \varepsilon \right) \] (4) --- \( {}^{1} \) If no such sequence exists, then \( \left( {2}^{\prime }\right) \) is vacuously true and there is nothing to prove. --- by the rules for quantifiers. We fix an \( \varepsilon \) satisfying (4), and let \[ {\delta }_{m} = \frac{1}{m},\;m = 1,2,\ldots \] By (4), for each \( {\delta }_{m} \) there is \( {x}_{m} \) (depending on \( {\delta }_{m} \) ) such that \[ {x}_{m} \in A \cap {G}_{\neg p}\left( \frac{1}{m}\right) \] (5) and \[ f\left( {x}_{m}\right) \notin {G}_{q}\left( \varepsilon \right) ,\;m = 1,2,3,\ldots \] (6) We fix these \( {x}_{m} \). As \( {x}_{m} \in A \) and \( {x}_{m} \neq p \), we obtain a sequence \[ \left\{ {x}_{m}\right\} \subseteq A - \{ p\} \] Also, as \( {x}_{m} \in {G}_{p}\left( \frac{1}{m}\right) \), we have \( \rho \left( {{x}_{m}, p}\right) < 1/m \rightarrow 0 \), and hence \( {x}_{m} \rightarrow p \). On the other hand, by (6), the image sequence \( \left\{ {f\left( {x}_{m}\right) }\right\} \) cannot converge to \( q \) (why?), i.e., \( \left( {2}^{\prime }\right) \) fails. Thus we see that \( \left( {2}^{\prime }\right) \) fails or holds accordingly as (2) does. This proves assertion (ii). Now, by setting \( q = f\left
Yes
Corollary 1. Let \( \\left( {T,{\\rho }^{\\prime }}\\right) \) be complete, such as \( {E}^{n} \) . Let a map \( f : A \\rightarrow T \) with \( A \\subseteq \\left( {S,\\rho }\\right) \) and a point \( p \\in S \) be given. Then for \( f \) to have a limit at \( p \) , it suffices that \( \\left\\{ {f\\left( {x}_{m}\\right) }\\right\\} \) be Cauchy in \( \\left( {T,{\\rho }^{\\prime }}\\right) \) whenever \( \\left\\{ {x}_{m}\\right\\} \\subseteq A - \\{ p\\} \) and \( {x}_{m} \\rightarrow p \) in \( \\left( {S,\\rho }\\right) \) .
Indeed, as noted above, all such \( \\left\\{ {f\\left( {x}_{m}\\right) }\\right\\} \) converge. Thus it only remains to show that they tend to one and the same limit \( q \), as is required in part (ii) of Theorem 1. We leave this as an exercise (Problem 1 below).
No
Theorem 3. Let \( \left( {S,\rho }\right) ,\left( {T,{\rho }^{\prime }}\right) \), and \( \left( {U,{\rho }^{\prime \prime }}\right) \) be metric spaces. If a function \( f : S \rightarrow T \) is continuous at a point \( p \in S \), and if \( g : T \rightarrow U \) is continuous at the point \( q = f\left( p\right) \), then the composite function \( g \circ f \) is continuous at \( p \) .
Proof. The domain of \( g \circ f \) is \( S \) . So take any sequence\n\n\[ \left\{ {x}_{m}\right\} \subseteq S\text{with}{x}_{m} \rightarrow p\text{.} \]\n\nAs \( f \) is continuous at \( p \), formula \( \left( {1}^{\prime }\right) \) yields \( f\left( {x}_{m}\right) \rightarrow f\left( p\right) \), where \( f\left( {x}_{m}\right) \) is in \( T = {D}_{g} \) . Hence, as \( g \) is continuous at \( f\left( p\right) \), we have\n\n\[ g\left( {f\left( {x}_{m}\right) }\right) \rightarrow g\left( {f\left( p\right) }\right) \text{, i.e.,}\left( {g \circ f}\right) \left( {x}_{m}\right) \rightarrow \left( {g \circ f}\right) \left( p\right) \text{,} \]\n\nand this holds for any \( \left\{ {x}_{m}\right\} \subseteq S \) with \( {x}_{m} \rightarrow p \) . Thus \( g \circ f \) satisfies condition \( \left( {1}^{\prime }\right) \) and is continuous at \( p \) .
Yes
Corollary 2. With the notation of Theorem 3, suppose\n\n\\[ \nf\\left( x\\right) \\rightarrow q\\text{as}x \\rightarrow p\\text{, and}g\\left( y\\right) \\rightarrow r\\text{as}y \\rightarrow q\\text{.}\n\\]\n\nThen\n\n\\[ \ng\\left( {f\\left( x\\right) }\\right) \\rightarrow r\\text{ as }x \\rightarrow p,\n\\]\n\nprovided, however, that\n\n(i) \\( g \\) is continuous at \\( q \\), or\n\n(ii) \\( f\\left( x\\right) \\neq q \\) for \\( x \\) in some deleted globe about \\( p \\), or\n\n(iii) \\( f \\) is one to one, at least when restricted to some \\( {G}_{\\neg p}\\left( \\delta \\right) \\) .
Indeed, (i) and (ii) suffice, as was explained above. Thus assume (iii). Then \\( f \\) can take the value \\( q \\) at most once, say, at some point\n\n\\[ \n{x}_{0} \\in {G}_{\\neg p}\\left( \\delta \\right)\n\\]\n\nAs \\( {x}_{0} \\neq p \\), let\n\n\\[ \n{\\delta }^{\\prime } = \\rho \\left( {{x}_{0}, p}\\right) > 0.\n\\]\n\nThen \\( {x}_{0} \\notin {G}_{\\neg p}\\left( {\\delta }^{\\prime }\\right) \\), so \\( f\\left( x\\right) \\neq q \\) on \\( {G}_{\\neg p}\\left( {\\delta }^{\\prime }\\right) \\), and case (iii) reduces to (ii).
Yes
Theorem 1. For any functions \( f, g, h : A \rightarrow {E}^{1}\left( C\right), A \subseteq \left( {S,\rho }\right) \), we have the following:\n\n(i) If \( f, g, h \) are continuous at \( p\left( {p \in A}\right) \), so are \( f \pm g \) and \( {fh} \). So also is \( f/h \), provided \( h\left( p\right) \neq 0 \); similarly for relative continuity over \( B \subseteq A \).\n\n(ii) If \( f\left( x\right) \rightarrow q, g\left( x\right) \rightarrow r \), and \( h\left( x\right) \rightarrow a \) (all, as \( x \rightarrow p \) over \( B \subseteq A \) ), then\n\n(a) \( f\left( x\right) \pm g\left( x\right) \rightarrow q \pm r \);\n\n(b) \( f\left( x\right) h\left( x\right) \rightarrow {qa} \);\n\n(c) \( \frac{f\left( x\right) }{h\left( x\right) } \rightarrow \frac{q}{a} \), provided \( a \neq 0 \).\n\nAll this holds also if \( f \) and \( g \) are vector valued and \( h \) is scalar valued.
For a simple proof, one can use Theorem 1 of Chapter 3, §15. (An independent proof is sketched in Problems 1-7 below.)\n\nWe can also use the sequential criterion (Theorem 1 in \( §2 \) ). To prove (ii), take any sequence\n\n\[ \left\{ {x}_{m}\right\} \subseteq B - \{ p\} ,{x}_{m} \rightarrow p. \]\n\nThen by the assumptions made,\n\n\[ f\left( {x}_{m}\right) \rightarrow q, g\left( {x}_{m}\right) \rightarrow r\text{, and}h\left( {x}_{m}\right) \rightarrow a\text{.} \]\n\nThus by Theorem 1 of Chapter 3, §15,\n\n\[ f\left( {x}_{m}\right) \pm g\left( {x}_{m}\right) \rightarrow q \pm r, f\left( {x}_{m}\right) g\left( {x}_{m}\right) \rightarrow {qa},\text{ and }\frac{f\left( {x}_{m}\right) }{g\left( {x}_{m}\right) } \rightarrow \frac{q}{a}. \]\n\nAs this holds for any sequence \( \left\{ {x}_{m}\right\} \subseteq B - \{ p\} \) with \( {x}_{m} \rightarrow p \), our assertion (ii) follows by the sequential criterion; similarly for (i).
No
Theorem 2 (componentwise continuity and limits). For any function \( f : A \rightarrow \) \( {E}^{n}\left( {{}^{ * }{C}^{n}}\right) \), with \( A \subseteq \left( {S,\rho }\right) \) and with \( f = \left( {{f}_{1},\ldots ,{f}_{n}}\right) \), we have that\n\n(i) \( f \) is continuous at \( p\left( {p \in A}\right) \) iff all its components \( {f}_{k} \) are, and\n\n(ii) \( f\left( x\right) \rightarrow \bar{q} \) as \( x \rightarrow p\left( {p \in S}\right) \) iff\n\n\[{f}_{k}\left( x\right) \rightarrow {q}_{k}\text{ as }x \rightarrow p\;\left( {k = 1,2,\ldots, n}\right) ,\]\ni.e., iff each \( {f}_{k} \) has, as its limit at \( p \), the corresponding component of \( \bar{q} \) .
We prove (ii). If \( f\left( x\right) \rightarrow \bar{q} \) as \( x \rightarrow p \) then, by definition,\n\n\[ \left( {\forall \varepsilon > 0}\right) \left( {\exists \delta > 0}\right) \left( {\forall x \in A \cap {G}_{\neg p}\left( \delta \right) }\right) \;\varepsilon > \left| {f\left( x\right) - \bar{q}}\right| = \sqrt{\mathop{\sum }\limits_{{k = 1}}^{n}{\left| {f}_{k}\left( x\right) - {q}_{k}\right| }^{2}}; \]\n\nin turn, the right-hand side of the inequality given above is no less than each\n\n\[ \left| {{f}_{k}\left( x\right) - {q}_{k}}\right| ,\;k = 1,2,\ldots, n. \]\n\nThus\n\n\[ \left( {\forall \varepsilon > 0}\right) \left( {\exists \delta > 0}\right) \left( {\forall x \in A \cap {G}_{\neg p}\left( \delta \right) }\right) \;\left| {{f}_{k}\left( x\right) - {q}_{k}}\right| < \varepsilon \]\n\ni.e., \( {f}_{k}\left( x\right) \rightarrow {q}_{k}, k = 1,\ldots, n \) .\n\nConversely, if each \( {f}_{k}\left( x\right) \rightarrow {q}_{k} \), then Theorem 1(ii) yields\n\n\[ \mathop{\sum }\limits_{{k = 1}}^{n}{\bar{e}}_{k}{f}_{k}\left( x\right) \rightarrow \mathop{\sum }\limits_{{k = 1}}^{n}{\bar{e}}_{k}{q}_{k}.{}^{1} \]\n\nBy formula (1), then, \( f\left( x\right) \rightarrow \bar{q} \) (for \( \mathop{\sum }\limits_{{k = 1}}^{n}{\bar{e}}_{k}{q}_{k} = \bar{q} \) ). Thus (ii) is proved;\n\nsimilarly for (i) and for relative limits and continuity.
Yes
Theorem 3. Any rational function (in particular, every polynomial) in one or several variables is continuous on all of its domain.
Proof. Consider first a monomial of the form\n\n\[ f\left( \bar{x}\right) = {x}_{k}\;\left( {k\text{ fixed }}\right) \]\n\nit is called the kth projection map because it \
No
Theorem 2. If a function \( f : A \rightarrow {E}^{ * }\left( {A \subseteq {E}^{ * }}\right) \) is nondecreasing on a finite or infinite interval \( B = \left( {a, b}\right) \subseteq A \) and if \( p \in \left( {a, b}\right) \), then\n\n\[ f\left( {a}^{ + }\right) \leq f\left( {p}^{ - }\right) \leq f\left( p\right) \leq f\left( {p}^{ + }\right) \leq f\left( {b}^{ - }\right) ,\]\n\nand for no \( x \in \left( {a, b}\right) \) do we have\n\n\[ f\left( {p}^{ - }\right) < f\left( x\right) < f\left( p\right) \text{ or }f\left( p\right) < f\left( x\right) < f\left( {p}^{ + }\right) {;}^{1} \]\n\nsimilarly in case \( f \downarrow \) (with all inequalities reversed).
Proof. By Theorem \( 1, f \uparrow \) on \( \left( {a, p}\right) \) implies\n\n\[ f\left( {a}^{ + }\right) = \mathop{\inf }\limits_{{a < x < p}}f\left( x\right) \text{ and }f\left( {p}^{ - }\right) = \mathop{\sup }\limits_{{a < x < p}}f\left( x\right) ; \]\n\nthus certainly \( f\left( {a}^{ + }\right) \leq f\left( {p}^{ - }\right) \) . As \( f \uparrow \), we also have \( f\left( p\right) \geq f\left( x\right) \) for all \( x \in \) \( \left( {a, p}\right) \) ; hence\n\n\[ f\left( p\right) \geq \mathop{\sup }\limits_{{a < x < p}}f\left( x\right) = f\left( {p}^{ - }\right) . \]\n\nThus\n\n\[ f\left( {a}^{ + }\right) \leq f\left( {p}^{ - }\right) \leq f\left( p\right) \]\n\nsimilarly for the rest of (1).\n\nMoreover, if \( a < x < p \), then \( f\left( x\right) \leq f\left( {p}^{ - }\right) \) since\n\n\[ f\left( {p}^{ - }\right) = \mathop{\sup }\limits_{{a < x < p}}f\left( x\right) \]\n\nIf, however, \( p \leq x < b \), then \( f\left( p\right) \leq f\left( x\right) \) since \( f \uparrow \) . Thus we never have \( f\left( {p}^{ - }\right) < f\left( x\right) < f\left( p\right) \) . Similarly, one excludes \( f\left( p\right) < f\left( x\right) < f\left( {p}^{ + }\right) \) . This completes the proof.
Yes
Theorem 1. If a set \( B \subseteq \left( {S,\rho }\right) \) is compact, so is any closed subset \( A \subseteq B \) .
Proof. We must show that each sequence \( \left\{ {x}_{m}\right\} \subseteq A \) clusters at some \( p \in A \) . However, as \( A \subseteq B,\left\{ {x}_{m}\right\} \) is also in \( B \), so by the compactness of \( B \), it clusters at some \( p \in B \) . Thus it remains to show that \( p \in A \) as well.\n\nNow by Theorem 1 of Chapter \( 3,§{16},\left\{ {x}_{m}\right\} \) has a subsequence \( {x}_{{m}_{k}} \rightarrow p \) . As \( \left\{ {x}_{{m}_{k}}\right\} \subseteq A \) and \( A \) is closed, this implies \( p \in A \) (Theorem 4 in Chapter 3, §16). \( ▱ \)
Yes
Theorem 2. Every compact set \( A \subseteq \left( {S,\rho }\right) \) is closed.
Proof. Given that \( A \) is compact, we must show (by Theorem 4 in Chapter 3, §16) that \( A \) contains the limit of each convergent sequence \( \left\{ {x}_{m}\right\} \subseteq A \) . Thus let \( {x}_{m} \rightarrow p,\left\{ {x}_{m}\right\} \subseteq A \) . As \( A \) is compact, the sequence \( \left\{ {x}_{m}\right\} \) clusters at some \( q \in A \), i.e., has a subsequence \( {x}_{{m}_{k}} \rightarrow q \in A \) . However, the limit of the subsequence must be the same as that of the entire sequence. Thus \( p = q \in A \) ; i.e., \( p \) is in \( A \), as required.
Yes
Theorem 3. Every compact set \( A \subseteq \left( {S,\rho }\right) \) is bounded.
Proof. By Problem 3 in Chapter 3, \( §{13} \), it suffices to show that \( A \) is contained in some finite union of globes. Thus we fix some arbitrary radius \( \varepsilon > 0 \) and, seeking a contradiction, assume that \( A \) cannot be covered by any finite number of globes of that radius.\n\nThen if \( {x}_{1} \in A \), the globe \( {G}_{{x}_{1}}\left( \varepsilon \right) \) does not cover \( A \), so there is a point \( {x}_{2} \in A \) such that\n\n\[{x}_{2} \notin {G}_{{x}_{1}}\left( \varepsilon \right) \text{, i.e.,}\rho \left( {{x}_{1},{x}_{2}}\right) \geq \varepsilon \text{.}\]\n\nBy our assumption, \( A \) is not even covered by \( {G}_{{x}_{1}}\left( \varepsilon \right) \cup {G}_{{x}_{2}}\left( \varepsilon \right) \) . Thus there is a point \( {x}_{3} \in A \) with\n\n\[{x}_{3} \notin {G}_{{x}_{1}}\left( \varepsilon \right) \text{and}{x}_{3} \notin {G}_{{x}_{2}}\left( \varepsilon \right) \text{, i.e.,}\rho \left( {{x}_{3},{x}_{1}}\right) \geq \varepsilon \text{and}\rho \left( {{x}_{3},{x}_{2}}\right) \geq \varepsilon \text{.}\]\n\nAgain, \( A \) is not covered by \( \mathop{\bigcup }\limits_{{i = 1}}^{3}{G}_{{x}_{i}}\left( \varepsilon \right) \), so there is a point \( {x}_{4} \in A \) not in that union; its distances from \( {x}_{1},{x}_{2} \), and \( {x}_{3} \) must therefore be \( \geq \varepsilon \) .\n\nSince \( A \) is never covered by any finite number of \( \varepsilon \) -globes, we can continue this process indefinitely (by induction) and thus select an infinite sequence \( \left\{ {x}_{m}\right\} \subseteq A \), with all its terms at least \( \varepsilon \) -apart from each other.\n\nNow as \( A \) is compact, this sequence must have a convergent subsequence \( \left\{ {x}_{{m}_{k}}\right\} \), which is then certainly Cauchy (by Theorem 1 of Chapter 3,§17). This is impossible, however, since its terms are at distances \( \geq \varepsilon \) from each other, contrary to Definition 1 in Chapter 3, §17. This contradiction completes the proof.
Yes
Theorem 4. In \( {E}^{n}\left( {{}^{ * }\text{and}\left. {C}^{n}\right) }\right. \) a set is compact iff it is closed and bounded.
Proof. In fact, if a set \( A \subseteq {E}^{n}\left( {{}^{ * }{C}^{n}}\right) \) is bounded, then by the Bolzano-Weierstrass theorem, each sequence \( \left\{ {x}_{m}\right\} \subseteq A \) has a convergent subsequence \( {x}_{{m}_{k}} \rightarrow p \) . If \( A \) is also closed, the limit point \( p \) must belong to \( A \) itself.\n\nThus each sequence \( \left\{ {x}_{m}\right\} \subseteq A \) clusters at some \( p \) in \( A \), so \( A \) is compact.\n\nThe converse is obvious.
Yes
Theorem 5 (Cantor's principle of nested closed sets). Every contracting sequence of nonvoid compact sets,\n\n\[ \n{F}_{1} \supseteq {F}_{2} \supseteq \cdots \supseteq {F}_{m} \supseteq \cdots ,\n\]\n\nin a metric space \( \left( {S,\rho }\right) \) has a nonvoid intersection; i.e., some \( p \) belongs to all \( {F}_{m} \) .
Proof. We prove the theorem for complete sets first.\n\nAs \( {F}_{m} \neq \varnothing \), we can pick a point \( {x}_{m} \) from each \( {F}_{m} \) to obtain a sequence \( \left\{ {x}_{m}\right\} ,{x}_{m} \in {F}_{m} \) . Since \( d{F}_{m} \rightarrow 0 \), it is easy to see that \( \left\{ {x}_{m}\right\} \) is a Cauchy sequence. (The details are left to the reader.) Moreover,\n\n\[ \n\left( {\forall m}\right) \;{x}_{m} \in {F}_{m} \subseteq {F}_{1}\n\]\n\nThus \( \left\{ {x}_{m}\right\} \) is a Cauchy sequence in \( {F}_{1} \), a complete set (by assumption).\n\nTherefore, by the definition of completeness (Chapter \( 3,§{17} \) ), \( \left\{ {x}_{m}\right\} \) has a limit \( p \in {F}_{1} \) . This limit remains the same if we drop a finite number of terms, say, the first \( m - 1 \) of them. Then we are left with the sequence \( {x}_{m},{x}_{m + 1},\ldots \) , which, by construction, is entirely contained in \( {F}_{m} \) (why?), with the same limit \( p \) . Then, however, the completeness of \( {F}_{m} \) implies that \( p \in {F}_{m} \) as well. As \( m \) is arbitrary here, it follows that \( \left( {\forall m}\right) p \in {F}_{m} \), i.e.,\n\n\[ \np \in \mathop{\bigcap }\limits_{{m = 1}}^{\infty }{F}_{m}\text{, as claimed.}\n\]\n\nThe proof for compact sets is analogous and even simpler. Here \( \left\{ {x}_{m}\right\} \) need not be a Cauchy sequence. Instead, using the compactness of \( {F}_{1} \), we select from \( \left\{ {x}_{m}\right\} \) a subsequence \( {x}_{{m}_{k}} \rightarrow p \in {F}_{1} \) and then proceed as above.
No
Theorem 1 (Lebesgue). Every open covering \( \left\{ {G}_{j}\right\} \) of a sequentially compact set \( F \subseteq \left( {S,\rho }\right) \) has at least one Lebesgue number \( \varepsilon \) . In symbols,\n\n\[ \left( {\exists \varepsilon > 0}\right) \left( {\forall x \in F}\right) \left( {\exists i}\right) \;{G}_{x}\left( \varepsilon \right) \subseteq {G}_{i}. \]
Proof. Seeking a contradiction, assume that (1) fails, i.e., its negation holds. As was explained in Chapter 1, §§1-3, this negation is\n\n\[ \left( {\forall \varepsilon > 0}\right) \left( {\exists {x}_{\varepsilon } \in F}\right) \left( {\forall i}\right) \;{G}_{{x}_{\varepsilon }}\left( \varepsilon \right) \nsubseteq {G}_{i} \]\n\n(where we write \( {x}_{\varepsilon } \) for \( x \) since here \( x \) may depend on \( \varepsilon \) ). As this is supposed to hold for all \( \varepsilon > 0 \), we take successively\n\n\[ \varepsilon = 1,\frac{1}{2},\ldots ,\frac{1}{n},\ldots \]\n\nThen, replacing \
Yes
Theorem 2 (generalized Heine-Borel theorem). A set \( F \subseteq \left( {S,\rho }\right) \) is compact iff every open covering of \( F \) has a finite subcovering.
Proof. Let \( F \) be sequentially compact, and let \( F \subseteq \bigcup {G}_{i} \), all \( {G}_{i} \) open. We have to show that \( \left\{ {G}_{i}\right\} \) reduces to a finite subcovering.\n\nBy Theorem \( 1,\left\{ {G}_{i}\right\} \) has a Lebesgue number \( \varepsilon \) satisfying (1). We fix this \( \varepsilon > 0 \) . Now by Note 1 in \( §6 \), we can cover \( F \) by a finite number of \( \varepsilon \) -globes,\n\n\[ F \subseteq \mathop{\bigcup }\limits_{{k = 1}}^{n}{G}_{{x}_{k}}\left( \varepsilon \right) ,\;{x}_{k} \in F.\]\n\nAlso by (1), each \( {G}_{{x}_{k}}\left( \varepsilon \right) \) is contained in some \( {G}_{i} \) ; call it \( {G}_{{i}_{k}} \) . With the \( {G}_{{i}_{k}} \) so fixed, we have\n\n\[ F \subseteq \mathop{\bigcup }\limits_{{k = 1}}^{n}{G}_{{x}_{k}}\left( \varepsilon \right) \subseteq \mathop{\bigcup }\limits_{{k = 1}}^{n}{G}_{{i}_{k}}\]\n\nThus the sets \( {G}_{{i}_{k}} \) constitute the desired finite subcovering, and the \
Yes
Theorem 1. If a function \( f : A \rightarrow \left( {T,{\rho }^{\prime }}\right), A \subseteq \left( {S,\rho }\right) \), is relatively continuous on a compact set \( B \subseteq A \), then \( f\left\lbrack B\right\rbrack \) is a compact set in \( \left( {T,{\rho }^{\prime }}\right) \) . Briefly, the continuous image of a compact set is compact.
Proof. To show that \( f\left\lbrack B\right\rbrack \) is compact, we take any sequence \( \left\{ {y}_{m}\right\} \subseteq f\left\lbrack B\right\rbrack \) and prove that it clusters at some \( q \in f\left\lbrack B\right\rbrack \) . As \( {y}_{m} \in f\left\lbrack B\right\rbrack ,{y}_{m} = f\left( {x}_{m}\right) \) for some \( {x}_{m} \) in \( B \) . We pick such an \( {x}_{m} \in B \) for each \( {y}_{m} \), thus obtaining a sequence \( \left\{ {x}_{m}\right\} \subseteq B \) with \[ f\left( {x}_{m}\right) = {y}_{m},\;m = 1,2,\ldots \] Now by the assumed compactness of \( B \), the sequence \( \left\{ {x}_{m}\right\} \) must cluster at some \( p \in B \) . Thus it has a subsequence \( {x}_{{m}_{k}} \rightarrow p \) . As \( p \in B \), the function \( f \) is relatively continuous at \( p \) over \( B \) (by assumption). Hence by the sequential criterion (§2), \( {x}_{{m}_{k}} \rightarrow p \) implies \( f\left( {x}_{{m}_{k}}\right) \rightarrow f\left( p\right) \) ; i.e., \[ {y}_{{m}_{k}} \rightarrow f\left( p\right) \in f\left\lbrack B\right\rbrack \] Thus \( q = f\left( p\right) \) is the desired cluster point of \( \left\{ {y}_{m}\right\} \) .
Yes
Lemma 1. Every nonvoid compact set \( F \subseteq {E}^{1} \) has a maximum and a minimum.
Proof. By Theorems 2 and 3 of \( §6, F \) is closed and bounded. Thus \( F \) has an infimum and a supremum in \( {E}^{1} \) (by the completeness axiom), say, \( p = \inf F \) and \( q = \sup F \) . It remains to show that \( p, q \in F \) . Assume the opposite, say, \( q \notin F \) . Then by properties of suprema, each globe \( {G}_{q}\left( \delta \right) = \left( {q - \delta, q + \delta }\right) \) contains some \( x \in B \) (specifically, \( q - \delta < x < q \) ) other than \( q \) (for \( q \notin B \), while \( x \in B \) ). Thus \[ \left( {\forall \delta > 0}\right) \;F \cap {G}_{\neg q}\left( \delta \right) \neq \varnothing \] i.e., \( F \) clusters at \( q \) and hence must contain \( q \) (being closed). However, since \( q \notin F \), this is the desired contradiction, and the lemma is proved.
Yes
Theorem 2 (Weierstrass).\n\n(i) If a function \( f : A \rightarrow \left( {T,{\rho }^{\prime }}\right) \) is relatively continuous on a compact set \( B \subseteq A \), then \( f \) is bounded on \( B \) ; i.e., \( f\left\lbrack B\right\rbrack \) is bounded.\n\n(ii) If, in addition, \( B \neq \varnothing \) and \( f \) is real \( \left( {f : A \rightarrow {E}^{1}}\right) \), then \( f\left\lbrack B\right\rbrack \) has a maximum and a minimum; i.e., \( f \) attains a largest and a least value at some points of \( B \) .
Proof. Indeed, by Theorem \( 1, f\left\lbrack B\right\rbrack \) is compact, so it is bounded, as claimed in (i).\n\nIf further \( B \neq \varnothing \) and \( f \) is real, then \( f\left\lbrack B\right\rbrack \) is a nonvoid compact set in \( {E}^{1} \), so by Lemma 1, it has a maximum and a minimum in \( {E}^{1} \) . Thus all is proved.
Yes
Theorem 3. If a function \( f : A \rightarrow \left( {T,{\rho }^{\prime }}\right), A \subseteq \left( {S,\rho }\right) \), is relatively continuous on a compact set \( B \subseteq A \) and is one to one on \( B \) (i.e., when restricted to \( B \) ), then its inverse, \( {f}^{-1} \), is continuous on \( f\left\lbrack B\right\rbrack {\text{.}}^{1} \)
Proof. To show that \( {f}^{-1} \) is continuous at each point \( q \in f\left\lbrack B\right\rbrack \), we apply the sequential criterion (Theorem 1 in \( §2 \) ). Thus we fix a sequence \( \left\{ {y}_{m}\right\} \subseteq f\left\lbrack B\right\rbrack \) , \( {y}_{m} \rightarrow q \in f\left\lbrack B\right\rbrack \), and prove that \( {f}^{-1}\left( {y}_{m}\right) \rightarrow {f}^{-1}\left( q\right) \) .\n\n\( {}^{1} \) Note that \( f \) need not be one to one on \( {all} \) of its domain \( A \), only on \( B \) . Thus \( {f}^{-1} \) need not be a mapping on \( f\left\lbrack A\right\rbrack \), but it is one on \( f\left\lbrack B\right\rbrack \) . (We use \
Yes
Theorem 4. If a function \( f : A \rightarrow \left( {T,{\rho }^{\prime }}\right), A \subseteq \left( {S,\rho }\right) \), is relatively continuous on a compact set \( B \subset A \), then \( f \) is also uniformly continuous on \( B \) .
Proof (by contradiction). Suppose \( f \) is relatively continuous on \( B \), but (4) fails. Then there is an \( \varepsilon > 0 \) such that\n\n\[ \left( {\forall \delta > 0}\right) \left( {\exists p, x \in B}\right) \;\rho \left( {x, p}\right) < \delta ,\text{ and yet }{\rho }^{\prime }\left( {f\left( x\right), f\left( p\right) }\right) \geq \varepsilon ; \]\n\nhere \( p \) and \( x \) depend on \( \delta \) . We fix such an \( \varepsilon \) and let\n\n\[ \delta = 1,\frac{1}{2},\ldots ,\frac{1}{m},\ldots \]\n\nThen for each \( \delta \) (i.e., each \( m \) ), we get two points \( {x}_{m},{p}_{m} \in B \) with\n\n\[ \rho \left( {{x}_{m},{p}_{m}}\right) < \frac{1}{m} \]\n\n(5)\n\nand\n\n\[ {\rho }^{\prime }\left( {f\left( {x}_{m}\right), f\left( {p}_{m}\right) }\right) \geq \varepsilon ,\;m = 1,2,\ldots \]\n\n(6)\n\nThus we obtain two sequences, \( \left\{ {x}_{m}\right\} \) and \( \left\{ {p}_{m}\right\} \), in \( B \) . As \( B \) is compact, \( \left\{ {x}_{m}\right\} \) has a subsequence \( {x}_{{m}_{k}} \rightarrow q\left( {q \in B}\right) \) . For simplicity, let it be \( \left\{ {x}_{m}\right\} \) itself; thus\n\n\[ {x}_{m} \rightarrow q,\;q \in B. \]\n\nHence by (5), it easily follows that also \( {p}_{m} \rightarrow q \) (because \( \rho \left( {{x}_{m},{p}_{m}}\right) \rightarrow 0 \) ; see Problem 4 in Chapter 3,§17). By the assumed relative continuity of \( f \) on \( B \) , it follows that\n\n\[ f\left( {x}_{m}\right) \rightarrow f\left( q\right) \text{ and }f\left( {p}_{m}\right) \rightarrow f\left( q\right) \text{ in }\left( {T,{\rho }^{\prime }}\right) . \]\n\nThis, in turn, implies that \( {\rho }^{\prime }\left( {f\left( {x}_{m}\right), f\left( {p}_{m}\right) }\right) \rightarrow 0 \), which is impossible, in view of (6). This contradiction completes the proof.
Yes
Lemma 1 (principle of nested line segments). Every contracting sequence of closed line segments \( L\left\lbrack {{\bar{p}}_{m},{\bar{q}}_{m}}\right\rbrack \) in \( {E}^{n} \) (*or in any other normed space) has a nonvoid intersection; i.e., there is a point\n\n\[ \bar{p} \in \mathop{\bigcap }\limits_{{m = 1}}^{\infty }L\left\lbrack {{\bar{p}}_{m},{\bar{q}}_{m}}\right\rbrack \]
Proof. Use Cantor’s theorem (Theorem 5 of \( §6 \) ) and Example (1) in \( §8 \) .
No
Theorem 2. If a function \( f : A \rightarrow {E}^{1} \) is monotone and has the Darboux property on a finite or infinite interval \( \left( {a, b}\right) \subseteq A \subseteq {E}^{1} \), then it is continuous on \( \left( {a, b}\right) \) .
Proof. Seeking a contradiction, suppose \( f \) is discontinuous at some \( p \in \left( {a, b}\right) \) . For definiteness, let \( f \uparrow \) on \( \left( {a, b}\right) \) . Then by Theorems 2 and 3 in \( §5 \), we have either \( f\left( {p}^{ - }\right) < f\left( p\right) \) or \( f\left( p\right) < f\left( {p}^{ + }\right) \) or both, with no function values in between. On the other hand, since \( f \) has the Darboux property, the function values \( f\left( x\right) \) for \( x \) in \( \left( {a, b}\right) \) fill an entire interval (see Note 1). Thus it is impossible for \( f\left( p\right) \) to be the only function value between \( f\left( {p}^{ - }\right) \) and \( f\left( {p}^{ + }\right) \) unless \( f \) is constant near \( p \), but then it is also continuous at \( p \), which we excluded. This contradiction completes the proof. \( {}^{3} \)
Yes
Theorem 3. If \( f : A \rightarrow {E}^{1} \) is strictly monotone and continuous when restricted to a finite or infinite interval \( B \subseteq A \subseteq {E}^{1} \), then its inverse \( {f}^{-1} \) has the same properties on the set \( f\left\lbrack B\right\rbrack \) (itself an interval, by Note 1 and Theorem 1). \( {}^{4} \)
Proof. It is easy to see that \( {f}^{-1} \) is increasing (decreasing) if \( f \) is; the proof is left as an exercise. Thus \( {f}^{-1} \) is monotone on \( f\left\lbrack B\right\rbrack \) if \( f \) is so on \( B \) . To prove the relative continuity of \( {f}^{-1} \), we use Theorem 2, i.e., show that \( {f}^{-1} \) has the Darboux property on \( f\left\lbrack B\right\rbrack \) . Thus let \( {f}^{-1}\left( p\right) < c < {f}^{-1}\left( q\right) \) for some \( p, q \in f\left\lbrack B\right\rbrack \) . We look for an \( r \in f\left\lbrack B\right\rbrack \) such that \( {f}^{-1}\left( r\right) = c \), i.e., \( r = f\left( c\right) \) . Now since \( p, q \in f\left\lbrack B\right\rbrack \), the numbers \( {f}^{-1}\left( p\right) \) and \( {f}^{-1}\left( q\right) \) are in \( B \), an interval. Hence also the intermediate value \( c \) is in \( B \) ; thus it belongs to the domain of \( f \), and so the function value \( f\left( c\right) \) exists. It thus suffices to put \( r = f\left( c\right) \) to get the result.
No
Theorem 1. A function \( f : \left( {A,\rho }\right) \rightarrow \left( {T,{\rho }^{\prime }}\right) \) is continuous on \( A \) iff \( {f}^{-1}\left\lbrack B\right\rbrack \) is closed in \( \left( {A,\rho }\right) \) for each closed set \( B \subseteq \left( {T,{\rho }^{\prime }}\right) \) ; similarly for open sets.
Indeed, this is part of Problem 15 in \( §2 \) with \( \left( {S,\rho }\right) \) replaced by \( \left( {A,\rho }\right) \).
No
Theorem 2. The only connected sets in \( {E}^{1} \) are exactly all convex sets, i.e., finite and infinite intervals, including \( {E}^{1} \) itself.
Proof. The proof that such intervals are exactly all convex sets in \( {E}^{1} \) is left as an exercise.\n\nWe now show that each connected set \( A \subseteq {E}^{1} \) is convex, i.e., that \( a, b \in A \) implies \( \left( {a, b}\right) \subseteq A \) .\n\nSeeking a contradiction, suppose \( p \notin A \) for some \( p \in \left( {a, b}\right), a, b \in A \) . Let\n\n\[ P = A \cap \left( {-\infty, p}\right) \text{ and }Q = A \cap \left( {p, + \infty }\right) . \]\n\nThen \( A = P \cup Q, a \in P, b \in Q \), and \( P \cap Q = \varnothing \) . Moreover, \( \left( {-\infty, p}\right) \) and \( \left( {p, + \infty }\right) \) are open sets in \( {E}^{1} \) . (Why?) Hence \( P \) and \( Q \) are open in \( A \), each being the intersection of \( A \) with a set open in \( {E}^{1} \) (see Note 1 above). As \( A = P \cup Q \), with \( P \cap Q = \varnothing \), it follows that \( A \) is disconnected. This shows that if \( A \) is connected in \( {E}^{1} \), it must be convex.
No
Theorem 3. If a function \( f : A \rightarrow \left( {T,{\rho }^{\prime }}\right) \) with \( A \subseteq \left( {S,\rho }\right) \) is relatively continuous on a connected set \( B \subseteq A \), then \( f\left\lbrack B\right\rbrack \) is a connected set in \( \left( {T,{\rho }^{\prime }}\right) {}^{4} \) .
Proof. By definition (§1), relative continuity on \( B \) becomes ordinary continuity when \( f \) is restricted to \( B \) . Thus we may treat \( f \) as a mapping of \( B \) into \( f\left\lbrack B\right\rbrack \), replacing \( S \) and \( T \) by their subspaces \( B \) and \( f\left\lbrack B\right\rbrack \) .\n\nSeeking a contradiction, suppose \( f\left\lbrack B\right\rbrack \) is disconnected, i.e.,\n\n\[ f\left\lbrack B\right\rbrack = P \cup Q \]\n\nfor some disjoint sets \( P, Q \neq \varnothing \) closed in \( \left( {f\left\lbrack B\right\rbrack ,{\rho }^{\prime }}\right) \) . Then by Theorem 1, with \( T \) replaced by \( f\left\lbrack B\right\rbrack \), the sets \( {f}^{-1}\left\lbrack P\right\rbrack \) and \( {f}^{-1}\left\lbrack Q\right\rbrack \) are closed in \( \left( {B,\rho }\right) \) . They also are nonvoid and disjoint (as are \( P \) and \( Q \) ) and satisfy\n\n\[ B = {f}^{-1}\left\lbrack {P \cup Q}\right\rbrack = {f}^{-1}\left\lbrack P\right\rbrack \cup {f}^{-1}\left\lbrack Q\right\rbrack \]\n\n(see Chapter 1, \( §§4 - 7 \), Problem 6). Thus \( B \) is disconnected, contrary to assumption.
Yes
Lemma 1. A set \( A \subseteq \left( {S,\rho }\right) \) is connected iff any two points \( p, q \in A \) are in some connected subset \( B \subseteq A \) . Hence any arcwise connected set is connected.
Proof. Seeking a contradiction, suppose the condition stated in Lemma 1 holds but \( A \) is disconnected, so \( A = P \cup Q \) for some disjoint sets \( P \neq \varnothing, Q \neq \varnothing \) , both closed in \( \left( {A,\rho }\right) \) .\n\nPick any \( p \in P \) and \( q \in Q \) . By assumption, \( p \) and \( q \) are in some connected set \( B \subseteq A \) . Treat \( \left( {B,\rho }\right) \) as a subspace of \( \left( {A,\rho }\right) \), and let\n\n\[ \n{P}^{\prime } = B \cap P\text{ and }{Q}^{\prime } = B \cap Q.\n\]\n\nThen by Theorem 4 of Chapter \( 3,§{12},{P}^{\prime } \) and \( {Q}^{\prime } \) are closed in \( B \) . Also, they are disjoint (for \( P \) and \( Q \) are) and nonvoid (for \( p \in {P}^{\prime }, q \in {Q}^{\prime } \) ), and\n\n\[ \nB = B \cap A = B \cap \left( {P \cup Q}\right) = \left( {B \cap P}\right) \cup \left( {B \cap Q}\right) = {P}^{\prime } \cup {Q}^{\prime }.\n\]\n\nThus \( B \) is disconnected, contrary to assumption. This contradiction proves the lemma (the converse proof is trivial).\n\nIn particular, if \( A \) is arcwise connected, then any points \( p, q \) in \( A \) are in some arc \( B \subseteq A \), a connected set by Corollary 2 . Thus all is proved.
Yes
Corollary 3. Any convex or polygon-connected set (e.g., a globe) in \( {E}^{n} \) (or in any other normed space) is arcwise connected, hence connected.
Proof. Use Lemma 1 and Example (c) in part I of this section.
No
Theorem 4. Every open connected set \( A \) in \( {E}^{n} \) (*or in another normed space) is also arcwise connected and even polygon connected.
Proof. If \( A = \varnothing \), this is \
No
Theorem 5. If a function \( f : A \rightarrow {E}^{1} \) is relatively continuous on a connected set \( B \subseteq A \subseteq \left( {S,\rho }\right) \), then \( f \) has the Darboux property on \( B \) .
In fact, by Theorems 3 and \( 2, f\left\lbrack B\right\rbrack \) is a connected set in \( {E}^{1} \), i.e., an interval. This, however, implies the Darboux property.
Yes
Theorem 3. In every metric space \( \\left( {S,\\rho }\\right) \), the metric \( \\rho : \\left( {S \\times S}\\right) \\rightarrow {E}^{1} \) is a continuous function on the product space \( S \\times S \) .
Proof. Fix any \( \\left( {p, q}\\right) \\in S \\times S \) . By Theorem 1 of \( §2,\\rho \) is continuous at \( \\left( {p, q}\\right) \) iff\n\n\[ \n\\rho \\left( {{x}_{m},{y}_{m}}\\right) \\rightarrow \\rho \\left( {p, q}\\right) \\text{ whenever }\\left( {{x}_{m},{y}_{m}}\\right) \\rightarrow \\left( {p, q}\\right) ,\n\]\n\ni.e., whenever \( {x}_{m} \\rightarrow p \) and \( {y}_{m} \\rightarrow q \) . However, this follows by Theorem 4 in Chapter 3, §15. Thus continuity is proved.
Yes
Theorem 1. Given a sequence of functions \( {f}_{m} : A \rightarrow \left( {T,{\rho }^{\prime }}\right) \), let \( B \subseteq A \) and\n\n\[ \n{Q}_{m} = \mathop{\sup }\limits_{{x \in B}}{\rho }^{\prime }\left( {{f}_{m}\left( x\right), f\left( x\right) }\right) .\n\]\n\nThen \( {f}_{m} \rightarrow f \) (uniformly on \( B \) ) iff \( {Q}_{m} \rightarrow 0 \) .
Proof. If \( {Q}_{m} \rightarrow 0 \), then by definition\n\n\[ \n\left( {\forall \varepsilon > 0}\right) \left( {\exists k}\right) \left( {\forall m > k}\right) \;{Q}_{m} < \varepsilon .\n\]\n\nHowever, \( {Q}_{m} \) is an upper bound of all distances \( {\rho }^{\prime }\left( {{f}_{m}\left( x\right), f\left( x\right) }\right), x \in B \) . Hence (2) follows.\n\nConversely, if\n\n\[ \n\left( {\forall x \in B}\right) \;{\rho }^{\prime }\left( {{f}_{m}\left( x\right), f\left( x\right) }\right) < \varepsilon ,\n\]\n\nthen\n\n\[ \n\varepsilon \geq \mathop{\sup }\limits_{{x \in B}}{\rho }^{\prime }\left( {{f}_{m}\left( x\right), f\left( x\right) }\right)\n\]\n\ni.e., \( {Q}_{m} \leq \varepsilon \) . Thus (2) implies\n\n\[ \n\left( {\forall \varepsilon > 0}\right) \left( {\exists k}\right) \left( {\forall m \geq k}\right) \;{Q}_{m} \leq \varepsilon \n\]\n\nand \( {Q}_{m} \rightarrow 0 \) .
Yes
Theorem 2. Let \( {f}_{m} : A \rightarrow \left( {T,{\rho }^{\prime }}\right) \) be a sequence of functions on \( A \subseteq \left( {S,\rho }\right) \) . If \( {f}_{m} \rightarrow f \) (uniformly) on a set \( B \subseteq A \), and if the \( {f}_{m} \) are relatively (or uniformly) continuous on \( B \), then the limit function \( f \) has the same property.
Proof. Fix \( \varepsilon > 0 \) . As \( {f}_{m} \rightarrow f \) (uniformly) on \( B \), there is a \( k \) such that\n\n\[ \left( {\forall x \in B}\right) \left( {\forall m \geq k}\right) \;{\rho }^{\prime }\left( {{f}_{m}\left( x\right), f\left( x\right) }\right) < \frac{\varepsilon }{4}. \]\n\nTake any \( {f}_{m} \) with \( m > k \), and take any \( p \in B \) . By continuity, there is \( \delta > 0 \), with\n\n\[ \left( {\forall x \in B \cap {G}_{p}\left( \delta \right) }\right) \;{\rho }^{\prime }\left( {{f}_{m}\left( x\right) ,{f}_{m}\left( p\right) }\right) < \frac{\varepsilon }{4}. \]\n\nSetting \( x = p \) in (3) gives \( {\rho }^{\prime }\left( {{f}_{m}\left( p\right), f\left( p\right) }\right) < \frac{\varepsilon }{4} \) . Combining this with (4) and (3), we obtain \( \left( {\forall x \in B \cap {G}_{p}\left( \delta \right) }\right) \)\n\n\[ {\rho }^{\prime }\left( {f\left( x\right), f\left( p\right) }\right) \leq {\rho }^{\prime }\left( {f\left( x\right) ,{f}_{m}\left( x\right) }\right) + {\rho }^{\prime }\left( {{f}_{m}\left( x\right) ,{f}_{m}\left( p\right) }\right) + {\rho }^{\prime }\left( {{f}_{m}\left( p\right), f\left( p\right) }\right) \]\n\n\[ < \frac{\varepsilon }{4} + \frac{\varepsilon }{4} + \frac{\varepsilon }{4} < \varepsilon \]\n\nWe thus see that for \( p \in B \), \n\n\[ \left( {\forall \varepsilon > 0}\right) \left( {\exists \delta > 0}\right) \left( {\forall x \in B \cap {G}_{p}\left( \delta \right) }\right) \;{\rho }^{\prime }\left( {f\left( x\right), f\left( p\right) }\right) < \varepsilon ,\]\n\ni.e., \( f \) is relatively continuous at \( p \) (over \( B \) ), as claimed.\n\nQuite similarly, the reader will show that \( f \) is uniformly continuous if the \( {f}_{n} \) are.
Yes
Theorem 3 (Cauchy criterion for uniform convergence). Let \( \left( {T,{\rho }^{\prime }}\right) \) be complete. Then a sequence \( {f}_{m} : A \rightarrow T, A \subseteq \left( {S,\rho }\right) \), converges uniformly on a set \( B \subseteq A \) iff\n\n\[ \left( {\forall \varepsilon > 0}\right) \left( {\exists k}\right) \left( {\forall x \in B}\right) \left( {\forall m, n > k}\right) \;{\rho }^{\prime }\left( {{f}_{m}\left( x\right) ,{f}_{n}\left( x\right) }\right) < \varepsilon . \]
Proof. If (5) holds then, for any (fixed) \( x \in B,\left\{ {{f}_{m}\left( x\right) }\right\} \) is a Cauchy sequence of points in \( T \), so by the assumed completeness of \( T \), it has a limit \( f\left( x\right) \) . Thus we can define a function \( f : B \rightarrow T \) with\n\n\[ f\left( x\right) = \mathop{\lim }\limits_{{m \rightarrow \infty }}{f}_{m}\left( x\right) \text{ on }B. \]\n\nTo show that \( {f}_{m} \rightarrow f \) (uniformly) on \( B \), we use (5) again. Keeping \( \varepsilon, k \) , \( x \), and \( m \) temporarily fixed, we let \( n \rightarrow \infty \) so that \( {f}_{n}\left( x\right) \rightarrow f\left( x\right) \) . Then by Theorem 4 of Chapter \( 3,§{15},{\rho }^{\prime }\left( {{f}_{m}\left( x\right) ,{f}_{n}\left( x\right) }\right) \rightarrow {p}^{\prime }\left( {f\left( x\right) ,{f}_{m}\left( x\right) }\right) \) . Passing to the limit in (5), we thus obtain (2).\n\nThe easy proof of the converse is left to the reader (cf. Chapter 3, §17, Theorem 1).
No
Theorem 4. Let\n\n\\[ \nf = \\mathop{\\sum }\\limits_{{m = 1}}^{\\infty }{f}_{m}\\text{ (pointwise) on }B{.}^{3} \n\\]\n\nLet \\( {m}_{1} < {m}_{2} < \\cdots < {m}_{n} < \\cdots \\) in \\( N \\), and define\n\n\\[ \n{g}_{1} = {s}_{{m}_{1}},{g}_{n} = {s}_{{m}_{n}} - {s}_{{m}_{n - 1}},\\;n > 1.\n\\]\n\n(Thus \\( {g}_{n + 1} = {f}_{{m}_{n} + 1} + \\cdots + {f}_{{m}_{n + 1}} \\) .) Then\n\n\\[ \nf = \\mathop{\\sum }\\limits_{{n = 1}}^{\\infty }{g}_{n}\\text{ (pointwise) on }B\\text{ as well; }\n\\]\n\nsimilarly for uniform convergence.
Proof. Let\n\n\\[ \n{s}_{n}^{\\prime } = \\mathop{\\sum }\\limits_{{k = 1}}^{n}{g}_{k},\\;n = 1,2,\\ldots\n\\]\n\nThen \\( {s}_{n}^{\\prime } = {s}_{{m}_{n}} \\) (verify!), so \\( \\left\{ {s}_{n}^{\\prime }\\right\} \\) is a subsequence, \\( \\left\{ {s}_{{m}_{n}}\\right\} \\), of \\( \\left\{ {s}_{m}\\right\} \\) . Hence \\( {s}_{m} \\rightarrow f \\) (pointwise) implies \\( {s}_{n}^{\\prime } \\rightarrow f \\) (pointwise); i.e.,\n\n\\[ \nf = \\mathop{\\sum }\\limits_{{n = 1}}^{\\infty }{g}_{n}\\text{ (pointwise). }\n\\]\n\nFor uniform convergence, see Problem 13 (cf. also Problem 19).
No
Theorem 1. Let the range space of the functions \( {f}_{m} \) (all defined on \( A \) ) be \( {E}^{1} \) , \( C \), or \( {E}^{n} \) (*or another complete normed space). Then for \( B \subseteq A \), we have the following:\n\n(i) If \( \sum \left| {f}_{m}\right| \) converges on \( B \) (pointwise or uniformly), so does \( \sum {f}_{m} \) itself. Moreover,\n\n\[ \left| {\mathop{\sum }\limits_{{m = 1}}^{\infty }{f}_{m}}\right| \leq \mathop{\sum }\limits_{{m = 1}}^{\infty }\left| {f}_{m}\right| \;\text{ on }B \]
Proof.\n\n(i) If \( \sum \left| {f}_{m}\right| \) converges uniformly on \( B \), then by Theorem \( {3}^{\prime } \) of \( §{12} \),\n\n\[ \left( {\forall \varepsilon > 0}\right) \left( {\exists k}\right) \left( {\forall n > m > k}\right) \left( {\forall x \in B}\right) \]\n\n\[ \varepsilon > \mathop{\sum }\limits_{{i = m}}^{n}\left| {{f}_{i}\left( x\right) }\right| \geq \left| {\mathop{\sum }\limits_{{i = m}}^{n}{f}_{i}\left( x\right) }\right| \;\text{ (triangle law). } \]\n\n(1)\n\nHowever, this shows that \( \sum {f}_{n} \) satisfies Cauchy’s criterion (6) of \( §{12} \), so it converges uniformly on \( \bar{B} \).\n\nMoreover, letting \( n \rightarrow \infty \) in the inequality\n\n\[ \left| {\mathop{\sum }\limits_{{m = 1}}^{n}{f}_{m}}\right| \leq \mathop{\sum }\limits_{{m = 1}}^{n}\left| {f}_{m}\right| \]\n\nwe get\n\n\[ \left| {\mathop{\sum }\limits_{{m = 1}}^{\infty }{f}_{m}}\right| \leq \mathop{\sum }\limits_{{m = 1}}^{\infty }\left| {f}_{m}\right| < + \infty \text{ on }B\text{, as claimed. } \]\n\nBy Note 1, this also proves the theorem for pointwise convergence.
Yes
Theorem 2 (comparison test). Suppose\n\n\[ \n\\left( {\\forall m}\\right) \\;\\left| {f}_{m}\\right| \\leq \\left| {g}_{m}\\right| \\text{ on }B.\n\]\n\nThen\n\n(i) \( \\mathop{\\sum }\\limits_{{m = 1}}^{\\infty }\\left| {f}_{m}\\right| \\leq \\mathop{\\sum }\\limits_{{m = 1}}^{\\infty }\\left| {g}_{m}\\right| \) on \( B \) ;\n\n(ii) \( \\mathop{\\sum }\\limits_{{m = 1}}^{\\infty }\\left| {f}_{m}\\right| = + \\infty \) implies \( \\mathop{\\sum }\\limits_{{m = 1}}^{\\infty }\\left| {g}_{m}\\right| = + \\infty \) on \( B \) ; and\n\n(iii) If \( \\sum \\left| {g}_{m}\\right| \) converges (pointwise or uniformly) on \( B \), so does \( \\sum \\left| {f}_{m}\\right| \) .
Proof. Conclusion (i) follows by letting \( n \\rightarrow \\infty \) in\n\n\[ \n\\mathop{\\sum }\\limits_{{m = 1}}^{n}\\left| {f}_{m}\\right| \\leq \\mathop{\\sum }\\limits_{{m = 1}}^{n}\\left| {g}_{m}\\right|\n\]\n\nIn turn, (ii) is a direct consequence of (i).\n\nAlso, by (i),\n\n\[ \n\\mathop{\\sum }\\limits_{{m = 1}}^{\\infty }\\left| {g}_{m}\\right| < + \\infty \\text{ implies }\\mathop{\\sum }\\limits_{{m = 1}}^{\\infty }\\left| {f}_{m}\\right| < + \\infty .\n\]\n\nThis proves (iii) for the pointwise case (see Note 1). The uniform case follows exactly as in Theorem 1(i) on noting that\n\n\[ \n\\mathop{\\sum }\\limits_{{k = m}}^{n}\\left| {f}_{k}\\right| \\leq \\mathop{\\sum }\\limits_{{k = m}}^{n}\\left| {g}_{k}\\right|\n\]\n\nand that the functions \( \\left| {f}_{k}\\right| \) and \( \\left| {g}_{k}\\right| \) are real (so Theorem \( {3}^{\\prime } \) in \( §{12} \) does apply).
Yes
Theorem 4 (necessary condition of convergence). If \( \sum {f}_{m} \) or \( \sum \left| {f}_{m}\right| \) converges on \( B \) (pointwise or uniformly), then \( \left| {f}_{m}\right| \rightarrow 0 \) on \( \bar{B} \) (in the same sense).
Proof. If \( \sum {f}_{m} = f \), say, then \( {s}_{m} \rightarrow f \) and also \( {s}_{m - 1} \rightarrow f \) . Hence\n\n\[ \n{s}_{m} - {s}_{m - 1} \rightarrow f - f = \overline{0}.\n\]\n\nHowever, \( {s}_{m} - {s}_{m - 1} = {f}_{m} \) . Thus \( {f}_{m} \rightarrow \overline{0} \), and \( \left| {f}_{m}\right| \rightarrow 0 \), as claimed.\n\nThis holds for pointwise and uniform convergence alike (see Problem 14 in §12).
Yes
Theorem 5 (root and ratio tests). A series of constants \( \sum {a}_{n}\left( {\left| {a}_{n}\right| \neq 0}\right) \) converges absolutely if\n\n\[ \overline{\lim }\sqrt[n]{\left| {a}_{n}\right| } < 1\text{ or }\overline{\lim }\left( \frac{\left| {a}_{n + 1}\right| }{\left| {a}_{n}\right| }\right) < 1. \]\n\nIt diverges if\n\n\[ \overline{\lim }\sqrt[n]{\left| {a}_{n}\right| } > 1\text{ or }\underline{\lim }\left( \frac{\left| {a}_{n + 1}\right| }{\left| {a}_{n}\right| }\right) > 1. \]\n\nIt may converge or diverge if\n\n\[ \overline{\lim }\sqrt[n]{\left| {a}_{n}\right| } = 1 \]\n\nor if\n\n\[ \underline{\lim }\left( \frac{\left| {a}_{n + 1}\right| }{\left| {a}_{n}\right| }\right) \leq 1 \leq \overline{\lim }\left( \frac{\left| {a}_{n + 1}\right| }{\left| {a}_{n}\right| }\right) . \]\n\n(The \( {a}_{n} \) may be scalars or vectors.)
Proof. If \( \overline{\lim }\sqrt[n]{\left| {a}_{n}\right| } < 1 \), choose \( r > 0 \) such that\n\n\[ \overline{\lim }\sqrt[n]{\left| {a}_{n}\right| } < r < 1 \]\n\nThen by Corollary 2 of Chapter 2, \( §{13},\sqrt[n]{\left| {a}_{n}\right| } < r \) for all but finitely many \( n \) . Thus, dropping a finite number of terms (§12, Problem 17), we may assume that\n\n\[ \left| {a}_{n}\right| < {r}^{n}\text{for all}n\text{.} \]\n\nAs \( 0 < r < 1 \), the geometric series \( \sum {r}^{n} \) converges. Hence so does \( \sum \left| {a}_{n}\right| \) by Theorem 2.\n\nIn the case\n\n\[ \overline{\lim }\left( \frac{\left| {a}_{n + 1}\right| }{\left| {a}_{n}\right| }\right) < 1 \]\n\nwe similarly obtain \( \left( {\exists m}\right) \left( {\forall n \geq m}\right) \left| {a}_{n + 1}\right| < \left| {a}_{n}\right| r \) ; hence by induction,\n\n\[ \left( {\forall n \geq m}\right) \;\left| {a}_{n}\right| \leq \left| {a}_{m}\right| {r}^{n - m}.\;\text{ (Verify!) } \]\n\nThus \( \sum \left| {a}_{n}\right| \) converges, as before.\n\nIf \( \overline{\lim }\sqrt[n]{\left| {a}_{n}\right| } > 1 \), then by Corollary 2 of Chapter 2, \( §{13},\left| {a}_{n}\right| > 1 \) for infinitely many \( n \) . Hence \( \left| {a}_{n}\right| \) cannot tend to 0, and so \( \sum {a}_{n} \) diverges by Theorem 4 .\n\nSimilarly, if\n\n\[ \underline{\lim }\left( \frac{\left| {a}_{n + 1}\right| }{\left| {a}_{n}\right| }\right) > 1 \]\n\nthen \( \left| {a}_{n + 1}\right| > \left| {a}_{n}\right| \) for all but finitely many \( n \), so \( \left| {a}_{n}\right| \) cannot tend to 0 again.
Yes
Theorem 6. For any power series \( \sum {a}_{n}{\left( x - p\right) }^{n} \), there is a unique \( r \in {E}^{ * } \) \( \left( {0 \leq r \leq + \infty }\right) \), called its convergence radius, such that the series converges absolutely for each \( x \) with \( \left| {x - p}\right| < r \) and does not converge (even conditionally) if \( \left| {x - p}\right| > r{.}^{8} \)
Proof. Fix any \( x = {x}_{0} \) . By Theorem 5, the series \( \sum {a}_{n}{\left( {x}_{0} - p\right) }^{n} \) converges absolutely if \( \overline{\lim }\sqrt[n]{\left| {a}_{n}\right| }\left| {{x}_{0} - p}\right| < 1 \), i.e., if\n\n\[ \left| {{x}_{0} - p}\right| < r\;\left( {r = \frac{1}{\overline{\lim }\sqrt[n]{\left| {a}_{n}\right| }} = \frac{1}{d}}\right) ,\]\n\nand diverges if \( \left| {{x}_{0} - p}\right| > r \) . (Here we assumed \( d \neq 0 \) ; but if \( d = 0 \), the condition \( d\left| {{x}_{0} - p}\right| < 1 \) is trivial for any \( {x}_{0} \), so \( r = + \infty \) in this case.) Thus \( r \) is the required radius, and clearly there can be only one such \( r \) . (Why?)
No
Theorem 7. If a power series \( \sum {a}_{n}{\left( x - p\right) }^{n} \) converges absolutely for some \( x = {x}_{0} \neq p \), then \( \sum \left| {{a}_{n}{\left( x - p\right) }^{n}}\right| \) converges uniformly on the closed globe \( {\bar{G}}_{p}\left( \delta \right) \) , \( \delta = \left| {{x}_{0} - p}\right| \) . So does \( \sum {a}_{n}{\left( x - p\right) }^{n} \) if the range space is complete (Theorem 1).
Proof. Suppose \( \sum \left| {{a}_{n}{\left( {x}_{0} - p\right) }^{n}}\right| \) converges. Let\n\n\[ \delta = \left| {{x}_{0} - p}\right| \text{ and }{M}_{n} = \left| {a}_{n}\right| {\delta }^{n} \]\n\nthus \( \sum {M}_{n} \) converges.\n\nNow if \( x \in {\bar{G}}_{p}\left( \delta \right) \), then \( \left| {x - p}\right| \leq \delta \), so\n\n\[ \left| {{a}_{n}{\left( x - p\right) }^{n}}\right| \leq \left| {a}_{n}\right| {\delta }^{n} = {M}_{n} \]\n\nHence by Theorem 3, \( \sum \left| {{a}_{n}{\left( x - p\right) }^{n}}\right| \) converges uniformly on \( {\bar{G}}_{p}\left( \delta \right) \) .
Yes
Theorem 1. If a function \( f : {E}^{1} \rightarrow E \) is differentiable at a point \( p \in {E}^{1} \), it is continuous at \( p \), and \( f\left( p\right) \) is finite (even if \( E = {E}^{ * } \) ).
Proof. Setting \( {\Delta x} = x - p \) and \( {\Delta f} = f\left( x\right) - f\left( p\right) \), we have the identity\n\n\[ \left| {f\left( x\right) - f\left( p\right) }\right| = \left| {\frac{\Delta f}{\Delta x} \cdot \left( {x - p}\right) }\right| \;\text{ for }x \neq p. \]\n\n(2)\n\nBy assumption,\n\n\[ {f}^{\prime }\left( p\right) = \mathop{\lim }\limits_{{x \rightarrow p}}\frac{\Delta f}{\Delta x} \]\n\nexists and is finite. Thus as \( x \rightarrow p \), the right side of (2) (hence the left side as well) tends to 0 , so\n\n\[ \mathop{\lim }\limits_{{x \rightarrow p}}\left| {f\left( x\right) - f\left( p\right) }\right| = 0\text{, or }\mathop{\lim }\limits_{{x \rightarrow p}}f\left( x\right) = f\left( p\right) ,\]\n\nproving continuity at \( p \) .\n\nAlso, \( f\left( p\right) \neq \pm \infty \), for otherwise \( \left| {f\left( x\right) - f\left( p\right) }\right| = + \infty \) for all \( x \), and so \( \left| {f\left( x\right) - f\left( p\right) }\right| \) cannot tend to 0 .\n\nNote 1. Similarly, the existence of a finite left (right) derivative at \( p \) implies left (right) continuity at \( p \) . The proof is the same.\n\nNote 2. The existence of an infinite derivative does not imply continuity, nor does it exclude it. For example, consider the two cases\n\n(i) \( f\left( x\right) = \frac{1}{x} \), with \( f\left( 0\right) = 0 \), and\n\n(ii) \( f\left( x\right) = \sqrt[3]{x} \).\n\nGive your comments for \( p = 0 \) .
Yes
Theorem 2. A function \( f : {E}^{1} \rightarrow E \) is differentiable at \( p \), and \( {f}^{\prime }\left( p\right) = c \), iff there is a finite \( c \in E \) and a function \( \delta : {E}^{1} \rightarrow E \) such that \( \mathop{\lim }\limits_{{x \rightarrow p}}\delta \left( x\right) = \delta \left( p\right) = 0 \) , and such that\n\n\[{\Delta f} = \left\lbrack {c + \delta \left( x\right) }\right\rbrack {\Delta x}\;\text{ for all }x \in {E}^{1}.\]
Proof. If \( f \) is differentiable at \( p \), put \( c = {f}^{\prime }\left( p\right) \) . Define \( \delta \left( p\right) = 0 \) and\n\n\[ \delta \left( x\right) = \frac{\Delta f}{\Delta x} - {f}^{\prime }\left( p\right) \text{ for }x \neq p.\]\n\nThen \( \mathop{\lim }\limits_{{x \rightarrow p}}\delta \left( x\right) = {f}^{\prime }\left( p\right) - {f}^{\prime }\left( p\right) = 0 = \delta \left( p\right) \) . Also,(3) follows.\n\nConversely, if (3) holds, then\n\n\[ \frac{\Delta f}{\Delta x} = c + \delta \left( x\right) \rightarrow c\text{ as }x \rightarrow p\text{ (since }\delta \left( x\right) \rightarrow 0\text{ ). }\]\n\nThus by definition,\n\n\[ c = \mathop{\lim }\limits_{{x \rightarrow p}}\frac{\Delta f}{\Delta x} = {f}^{\prime }\left( p\right) \text{ and }{f}^{\prime }\left( p\right) = c\text{ is finite. }\]
Yes
Theorem 3 (chain rule). Let the functions \( g : {E}^{1} \rightarrow {E}^{1} \) (real) and \( f : {E}^{1} \rightarrow E \) (real or not) be differentiable at \( p \) and \( q \), respectively, where \( q = g\left( p\right) \). Then the composite function \( h = f \circ g \) is differentiable at \( p \), and \[ {h}^{\prime }\left( p\right) = {f}^{\prime }\left( q\right) {g}^{\prime }\left( p\right) \]
Proof. Setting \[ {\Delta h} = h\left( x\right) - h\left( p\right) = f\left( {g\left( x\right) }\right) - f\left( {g\left( p\right) }\right) = f\left( {g\left( x\right) }\right) - f\left( q\right) , \] we must show that \[ \mathop{\lim }\limits_{{x \rightarrow p}}\frac{\Delta h}{\Delta x} = {f}^{\prime }\left( q\right) {g}^{\prime }\left( p\right) \neq \pm \infty . \] Now as \( f \) is differentiable at \( q \), Theorem 2 yields a function \( \delta : {E}^{1} \rightarrow E \) such that \( \mathop{\lim }\limits_{{x \rightarrow q}}\delta \left( x\right) = \delta \left( q\right) = 0 \) and such that \[ \left( {\forall y \in {E}^{1}}\right) \;f\left( y\right) - f\left( q\right) = \left\lbrack {{f}^{\prime }\left( q\right) + \delta \left( y\right) }\right\rbrack {\Delta y},{\Delta y} = y - q. \] Taking \( y = g\left( x\right) \), we get \[ \left( {\forall x \in {E}^{1}}\right) \;f\left( {g\left( x\right) }\right) - f\left( q\right) = \left\lbrack {{f}^{\prime }\left( q\right) + \delta \left( {g\left( x\right) }\right) }\right\rbrack \left\lbrack {g\left( x\right) - g\left( p\right) }\right\rbrack , \] where \[ g\left( x\right) - g\left( p\right) = y - q = {\Delta y}\text{ and }f\left( {g\left( x\right) }\right) - f\left( q\right) = {\Delta h}, \] as noted above. Hence \[ \frac{\Delta h}{\Delta x} = \left\lbrack {{f}^{\prime }\left( q\right) + \delta \left( {g\left( x\right) }\right) }\right\rbrack \cdot \frac{g\left( x\right) - g\left( p\right) }{x - p}\;\text{ for all }x \neq p. \] Let \( x \rightarrow p \) . Then we obtain \( {h}^{\prime }\left( p\right) = {f}^{\prime }\left( q\right) {g}^{\prime }\left( p\right) \), for, by the continuity of \( \delta \circ g \) at \( p \) (Chapter 4, \( §2 \), Theorem 3), \[ \mathop{\lim }\limits_{{x \rightarrow p}}\delta \left( {g\left( x\right) }\right) = \delta \left( {g\left( p\right) }\right) = \delta \left( q\right) = 0. \]
Yes
Theorem 4. If \( f, g \), and \( h \) are real or complex and are differentiable at \( p \), so are \[ f \pm g,{hf},\text{ and }\frac{f}{h} \] (the latter if \( h\left( p\right) \neq 0 \) ), and at the point \( p \) we have (i) \( {\left( f \pm g\right) }^{\prime } = {f}^{\prime } \pm {g}^{\prime } \) ; (ii) \( {\left( hf\right) }^{\prime } = h{f}^{\prime } + {h}^{\prime }f \) ; and (iii) \( {\left( \frac{f}{h}\right) }^{\prime } = \frac{h{f}^{\prime } - {h}^{\prime }f}{{h}^{2}} \) .
(i) \( {\left( f \pm g\right) }^{\prime } = {f}^{\prime } \pm {g}^{\prime } \) ; (ii) \( {\left( hf\right) }^{\prime } = h{f}^{\prime } + {h}^{\prime }f \) ; and (iii) \( {\left( \frac{f}{h}\right) }^{\prime } = \frac{h{f}^{\prime } - {h}^{\prime }f}{{h}^{2}} \)
Yes
Theorem 5 (componentwise differentiation). A function \( f : {E}^{1} \rightarrow {E}^{n}\left( {{}^{ * }{C}^{n}}\right) \) is differentiable at \( p \) iff each of its \( n \) components \( \left( {{f}_{1},\ldots ,{f}_{n}}\right) \) is, and then
\[ {f}^{\prime }\left( p\right) = \left( {{f}_{1}^{\prime }\left( p\right) ,\ldots ,{f}_{n}^{\prime }\left( p\right) }\right) = \mathop{\sum }\limits_{{k = 1}}^{n}{f}_{k}^{\prime }\left( p\right) {\bar{e}}_{k}, \] with \( {\bar{e}}_{k} \) as in Theorem 2 of Chapter 3, \( §§1 - 3 \) . In particular, a complex function \( f : {E}^{1} \rightarrow C \) is differentiable iff its real and imaginary parts are, and \( {f}^{\prime } = {f}_{\mathrm{{re}}}^{\prime } + i \cdot {f}_{\mathrm{{im}}}^{\prime } \) (Chapter \( 4,§3 \), Note 5).
Yes
Lemma 1. If \( {f}^{\prime }\left( p\right) > 0 \) at some \( p \in {E}^{1} \), then\n\n\[ x < p < y \]\n\nimplies\n\n\[ f\left( x\right) < f\left( p\right) < f\left( y\right) \]\n\nfor all \( x, y \) in a sufficiently small globe \( {G}_{p}\left( \delta \right) = \left( {p - \delta, p + \delta }\right) {}^{1} \)\n\nSimilarly, if \( {f}^{\prime }\left( p\right) < 0 \), then \( x < p < y \) implies \( f\left( x\right) > f\left( p\right) > f\left( y\right) \) for \( x, y \) in some \( {G}_{p}\left( \delta \right) \) .
Proof. If \( {f}^{\prime }\left( p\right) > 0 \), the \
No
Corollary 1. If \( f\left( p\right) \) is the maximum or minimum value of \( f\left( x\right) \) for \( x \) in some \( {G}_{p}\left( \delta \right) \), then \( {f}^{\prime }\left( p\right) = 0 \) ; i.e., \( f \) has a zero derivative, or none at all, at \( p \) .
For, by Lemma \( 1,{f}^{\prime }\left( p\right) \neq 0 \) excludes a maximum or minimum at \( p \) . (Why?)
No
Theorem 1. Let \( f : {E}^{1} \rightarrow {E}^{ * } \) be relatively continuous on an interval \( \left\lbrack {a, b}\right\rbrack \) , with \( {f}^{\prime } \neq 0 \) on \( \left( {a, b}\right) \) . Then \( f \) is strictly monotone on \( \left\lbrack {a, b}\right\rbrack \), and \( {f}^{\prime } \) is sign-constant there (possibly 0 at a and b), with \( {f}^{\prime } \geq 0 \) if \( f \uparrow \), and \( {f}^{\prime } \leq 0 \) if \( f \downarrow \) .
Proof. By Theorem 2 of Chapter \( 4,§8, f \) attains a least value \( m \), and a largest value \( M \), at some points of \( \left\lbrack {a, b}\right\rbrack \) . However, neither can occur at an interior point \( p \in \left( {a, b}\right) \), for, by Corollary 1, this would imply \( {f}^{\prime }\left( p\right) = 0 \), contrary to our assumption.\n\n\( {}^{1} \) This does not mean that \( f \) is monotone on any \( {G}_{p} \) (see Problem 6). We shall only say in such cases that \( f \) increases at the point \( p \) .\n\n![d5cb5ba6-f920-47a9-b7ff-044c628ad635_273_0.jpg](images/d5cb5ba6-f920-47a9-b7ff-044c628ad635_273_0.jpg)\n\nFIGURE 22\n\nThus \( M = f\left( a\right) \) or \( M = f\left( b\right) \) ; for the moment we assume \( M = f\left( b\right) \) and \( m = f\left( a\right) \) . We must have \( m < M \), for \( m = M \) would make \( f \) constant on \( \left\lbrack {a, b}\right\rbrack \) , implying \( {f}^{\prime } = 0 \) . Thus \( m = f\left( a\right) < f\left( b\right) = M \).\n\nNow let \( a \leq x < y \leq b \) . Applying the previous argument to each of the intervals \( \left\lbrack {a, x}\right\rbrack ,\left\lbrack {a, y}\right\rbrack ,\left\lbrack {x, y}\right\rbrack \), and \( \left\lbrack {x, b}\right\rbrack \) (now using that \( m = f\left( a\right) < f\left( b\right) = M \) ), we find that\n\n\[ f\left( a\right) \leq f\left( x\right) < f\left( y\right) \leq f\left( b\right) .\;\text{ (Why?) } \]\n\nThus \( a \leq x < y \leq b \) implies \( f\left( x\right) < f\left( y\right) \) ; i.e., \( f \) increases on \( \left\lbrack {a, b}\right\rbrack \) . Hence \( {f}^{\prime } \) cannot be negative at any \( p \in \left\lbrack {a, b}\right\rbrack \), for, otherwise, by Lemma \( 1, f \) would decrease at \( p \) . Thus \( {f}^{\prime } \geq 0 \) on \( \left\lbrack {a, b}\right\rbrack \).\n\nIn the case \( M = f\left( a\right) > f\left( b\right) = m \), we would obtain \( {f}^{\prime } \leq 0 \) .\n\nCaution: The function \( f \) may increase or decrease at \( p \) even if \( {f}^{\prime }\left( p\right) = 0 \) . See Note 1.
Yes
Corollary 2 (Rolle’s theorem). If \( f : {E}^{1} \rightarrow {E}^{ * } \) is relatively continuous on \( \left\lbrack {a, b}\right\rbrack \) and if \( f\left( a\right) = f\left( b\right) \), then \( {f}^{\prime }\left( p\right) = 0 \) for at least one interior point \( p \in \left( {a, b}\right) \) .
For, if \( {f}^{\prime } \neq 0 \) on all of \( \left( {a, b}\right) \), then by Theorem \( 1, f \) would be strictly monotone on \( \left\lbrack {a, b}\right\rbrack \), so the equality \( f\left( a\right) = f\left( b\right) \) would be impossible.
Yes
Theorem 2 (Cauchy’s law of the mean). Let the functions \( f, g : {E}^{1} \rightarrow {E}^{ * } \) be relatively continuous and finite on \( \left\lbrack {a, b}\right\rbrack \) and have derivatives on \( \left( {a, b}\right) \), with \( {f}^{\prime } \) and \( {g}^{\prime } \) never both infinite at the same point \( p \in \left( {a, b}\right) \) . Then\n\n\[ \n{g}^{\prime }\left( q\right) \left\lbrack {f\left( b\right) - f\left( a\right) }\right\rbrack = {f}^{\prime }\left( q\right) \left\lbrack {g\left( b\right) - g\left( a\right) }\right\rbrack \text{for at least one}q \in \left( {a, b}\right) \text{.} \n\]
Proof. Let \( A = f\left( b\right) - f\left( a\right) \) and \( B = g\left( b\right) - g\left( a\right) \) . We must show that \( A{g}^{\prime }\left( q\right) = \) \( B{f}^{\prime }\left( q\right) \) for some \( q \in \left( {a, b}\right) \) . For this purpose, consider the function \( h = {Ag} - {Bf} \) . It is relatively continuous and finite on \( \left\lbrack {a, b}\right\rbrack \), as are \( g \) and \( f \) . Also,\n\n\[ \nh\left( a\right) = f\left( b\right) g\left( a\right) - g\left( b\right) f\left( a\right) = h\left( b\right) .\;\text{ (Verify!) } \n\]\n\nThus by Corollary \( 2,{h}^{\prime }\left( q\right) = 0 \) for some \( q \in \left( {a, b}\right) \) . Here, by Theorem 4 of \( §1,{h}^{\prime } = {\left( Ag - Bf\right) }^{\prime } = A{g}^{\prime } - B{f}^{\prime } \) . (This is legitimate, for, by assumption, \( {f}^{\prime } \) and \( {g}^{\prime } \) never both become infinite, so no indeterminate limits occur.) Thus \( {h}^{\prime }\left( q\right) = A{g}^{\prime }\left( q\right) - B{f}^{\prime }\left( q\right) = 0 \), and (1) follows.
Yes
Corollary 3 (Lagrange’s law of the mean). If \( f : {E}^{1} \rightarrow {E}^{1} \) is relatively continuous on \( \left\lbrack {a, b}\right\rbrack \) with a derivative on \( \left( {a, b}\right) \), then\n\n\[ f\left( b\right) - f\left( a\right) = {f}^{\prime }\left( q\right) \left( {b - a}\right) \text{ for at least one }q \in \left( {a, b}\right) . \]
Proof. Take \( g\left( x\right) = x \) in Theorem 2, so \( {g}^{\prime } = 1 \) on \( {E}^{1} \) .
No
Corollary 4. Let \( f \) be as in Corollary 3. Then\n\n(i) \( f \) is constant on \( \left\lbrack {a, b}\right\rbrack \) iff \( {f}^{\prime } = 0 \) on \( \left( {a, b}\right) \) ;
Proof. Let \( {f}^{\prime } = 0 \) on \( \left( {a, b}\right) \) . If \( a \leq x \leq y \leq b \), apply Corollary 3 to the interval \( \left\lbrack {x, y}\right\rbrack \) to obtain\n\n\[ f\left( y\right) - f\left( x\right) = {f}^{\prime }\left( q\right) \left( {y - x}\right) \text{ for some }q \in \left( {a, b}\right) \text{ and }{f}^{\prime }\left( q\right) = 0. \]\n\nThus \( f\left( y\right) - f\left( x\right) = 0 \) for \( x, y \in \left\lbrack {a, b}\right\rbrack \), so \( f \) is constant.
Yes
Theorem 3 (inverse functions). Let \( f : {E}^{1} \rightarrow {E}^{1} \) be relatively continuous and strictly monotone on an interval \( I \subseteq {E}^{1} \). Let \( {f}^{\prime }\left( p\right) \neq 0 \) at some interior point \( p \in I \). Then the inverse function \( g = {f}^{-1} \) (with \( f \) restricted to \( I \)) has a derivative at \( q = f\left( p\right) \), and\n\n\[ \n{g}^{\prime }\left( q\right) = \frac{1}{{f}^{\prime }\left( p\right) }.\n\]\n\n(If \( {f}^{\prime }\left( p\right) = \pm \infty \), then \( {g}^{\prime }\left( q\right) = 0 \).)
Proof. By Theorem 3 of Chapter \( 4,§9, g = {f}^{-1} \) is strictly monotone and relatively continuous on \( f\left\lbrack I\right\rbrack \), itself an interval. If \( p \) is interior to \( I \), then \( q = f\left( p\right) \) is interior to \( f\left\lbrack I\right\rbrack \) . (Why?)\n\nNow if \( y \in f\left\lbrack I\right\rbrack \), we set\n\n\[ \n{\Delta g} = g\left( y\right) - g\left( q\right) ,{\Delta y} = y - q, x = {f}^{-1}\left( y\right) = g\left( y\right) \text{, and}f\left( x\right) = y\n\]\n\nand obtain\n\n\[ \n\frac{\Delta g}{\Delta y} = \frac{g\left( y\right) - g\left( q\right) }{y - q} = \frac{x - p}{f\left( x\right) - f\left( p\right) } = \frac{\Delta x}{\Delta f}\text{ for }x \neq p.\n\]\n\nNow if \( y \rightarrow q \), the continuity of \( g \) at \( q \) yields \( g\left( y\right) \rightarrow g\left( q\right) \) ; i.e., \( x \rightarrow p \). Also, \( x \neq p \) iff \( y \neq q \), for \( f \) and \( g \) are one-to-one functions. Thus we may substitute \( y = f\left( x\right) \) or \( x = g\left( y\right) \) to get\n\n\[ \n{g}^{\prime }\left( q\right) = \mathop{\lim }\limits_{{y \rightarrow q}}\frac{\Delta g}{\Delta y} = \mathop{\lim }\limits_{{x \rightarrow p}}\frac{\Delta x}{\Delta f} = \frac{1}{\mathop{\lim }\limits_{{x \rightarrow p}}\left( {{\Delta f}/{\Delta x}}\right) } = \frac{1}{{f}^{\prime }\left( p\right) }\n\]\n\n\( \left( {2}^{\prime }\right) \)\n\nwhere we use the convention \( \frac{1}{\infty } = 0 \) if \( {f}^{\prime }\left( p\right) = \infty \) .
No
Theorem 4 (Darboux). If \( f : {E}^{1} \rightarrow {E}^{ * } \) is relatively continuous and has a derivative on an interval \( I \), then \( {f}^{\prime } \) has the Darboux property (Chapter 4,§9) on \( I \) .
Proof. Let \( p, q \in I \) and \( {f}^{\prime }\left( p\right) < c < {f}^{\prime }\left( q\right) \) . Put \( g\left( x\right) = f\left( x\right) - {cx} \) . Assume \( {g}^{\prime } \neq 0 \) on \( \left( {p, q}\right) \) and find a contradiction to Theorem 1. Details are left to the reader.
No
Theorem 1 (finite increments law). Let \( f : {E}^{1} \rightarrow E \) and \( g : {E}^{1} \rightarrow {E}^{ * } \) be relatively continuous and finite on a closed interval \( I = \left\lbrack {a, b}\right\rbrack \subseteq {E}^{1} \), and have derivatives \( {}^{2} \) with \( \left| {f}^{\prime }\right| \leq {g}^{\prime } \), on \( I - Q \) where \( Q \subseteq \left\{ {{p}_{1},{p}_{2},\ldots ,{p}_{m},\ldots }\right\} \) . Then \[ \left| {f\left( b\right) - f\left( a\right) }\right| \leq g\left( b\right) - g\left( a\right) . \]
The proof is somewhat laborious, but worthwhile. (At a first reading, one may omit it, however.) We outline some preliminary ideas. Given any \( x \in I \), suppose first that \( x > {p}_{m} \) for at least one \( {p}_{m} \in Q \) . In this case, we put \[ Q\left( x\right) = \mathop{\sum }\limits_{{{p}_{m} < x}}{2}^{-m} \] here the summation is only over those \( m \) for which \( {p}_{m} < x \) . If, however, there are no \( {p}_{m} \in Q \) with \( {p}_{m} < x \), we put \( Q\left( x\right) = 0 \) . Thus \( Q\left( x\right) \) is defined for all \( x \in I \) . It gives an idea as to \
No
If \( f : {E}^{1} \rightarrow E \) is relatively continuous and finite on \( I = \left\lbrack {a, b}\right\rbrack \subseteq \) \( {E}^{1} \), and has a derivative on \( I - Q \), then there is a real \( M \) such that \n\n\[ \n\left| {f\left( b\right) - f\left( a\right) }\right| \leq M\left( {b - a}\right) \text{ and }M \leq \mathop{\sup }\limits_{{t \in I - Q}}\left| {{f}^{\prime }\left( t\right) }\right| .\n\]
Proof. Let \n\n\[ \n{M}_{0} = \mathop{\sup }\limits_{{t \in I - Q}}\left| {{f}^{\prime }\left( t\right) }\right| \n\] \n\nIf \( {M}_{0} < + \infty \), put \( M = {M}_{0} \geq \left| {f}^{\prime }\right| \) on \( I - Q \), and take \( g\left( x\right) = {Mx} \) in Theorem 1 . Then \( {g}^{\prime } = M \geq \left| {f}^{\prime }\right| \) on \( I - Q \), so formula (1) yields (5) since \n\n\[ \ng\left( b\right) - g\left( a\right) = {Mb} - {Ma} = M\left( {b - a}\right) .\n\] \n\nIf, however, \( {M}_{0} = + \infty \), let \n\n\[ \nM = \left| \frac{f\left( b\right) - f\left( a\right) }{b - a}\right| < {M}_{0}.\n\] \n\nThen (5) clearly is true. Thus the required \( M \) exists always. \( {}^{3} \)
Yes
Corollary 2. Let \( f \) be as in Corollary 1. Then \( f \) is constant on \( I \) iff \( {f}^{\prime } = 0 \) on \( I - Q \) .
Proof. If \( {f}^{\prime } = 0 \) on \( I - Q \), then \( M = 0 \) in Corollary 1, so Corollary 1 yields, for any subinterval \( \left\lbrack {a, x}\right\rbrack \left( {x \in I}\right) ,\left| {f\left( x\right) - f\left( a\right) }\right| \leq 0 \) ; i.e., \( f\left( x\right) = f\left( a\right) \) for all \( x \in I \) . Thus \( f \) is constant on \( I \) .\n\nConversely, if so, then \( {f}^{\prime } = 0 \), even on all of \( I \) .
Yes
Corollary 3. Let \( f, g : {E}^{1} \rightarrow E \) be relatively continuous and finite on \( I = \) \( \left\lbrack {a, b}\right\rbrack \), and differentiable on \( I - Q \) . Then \( f - g \) is constant on \( I \) iff \( {f}^{\prime } = {g}^{\prime } \) on \( I - Q \) .
Proof. Apply Corollary 2 to the function \( f - g \) .
No
Theorem 2. Let \( f \) be real and have the properties stated in Corollary 1. Then\n\n(i) \( f \uparrow \) on \( I = \left\lbrack {a, b}\right\rbrack \) iff \( {f}^{\prime } \geq 0 \) on \( I - Q \) ; and\n\n(ii) \( f \downarrow \) on \( I \) iff \( {f}^{\prime } \leq 0 \) on \( I - Q \) .
Proof. Let \( {f}^{\prime } \geq 0 \) on \( I - Q \) . Fix any \( x, y \in I\left( {x < y}\right) \) and define \( g\left( t\right) = 0 \) on \( {E}^{1} \) . Then \( \left| {g}^{\prime }\right| = 0 \leq {f}^{\prime } \) on \( I - Q \) . Thus \( g \) and \( f \) satisfy Theorem 1 (with their roles reversed) on \( I \), and certainly on the subinterval \( \left\lbrack {x, y}\right\rbrack \) . Thus we have\n\n\[ f\left( y\right) - f\left( x\right) \geq \left| {g\left( y\right) - g\left( x\right) }\right| = 0\text{, i.e.,}f\left( y\right) \geq f\left( x\right) \text{whenever}y > x\text{in}I\text{,}\]\n\nso \( f \uparrow \) on \( I \) .\n\nConversely, if \( f \uparrow \) on \( I \), then for every \( p \in I \), we must have \( {f}^{\prime }\left( p\right) \geq 0 \), for otherwise, by Lemma 1 of \( §2, f \) would decrease at \( p \) . Thus \( {f}^{\prime } \geq 0 \), even on all of \( I \), and (i) is proved. Assertion (ii) is proved similarly.
Yes
Theorem 1. If \( F \) and \( G \) are primitive to \( f \) on \( I \), then \( G - F \) is constant on \( I \).
Proof. By assumption, \( F \) and \( G \) are relatively continuous and finite on \( I \) ; hence so is \( G - F \) . Also, \( {F}^{\prime } = f \) on \( I - Q \) and \( {G}^{\prime } = f \) on \( I - P \) . ( \( Q \) and \( P \) are countable, but possibly \( Q \neq P \) .)\n\nHence both \( {F}^{\prime } \) and \( {G}^{\prime } \) equal \( f \) on \( I - S \), where \( S = P \cup Q \), and \( S \) is countable itself by Theorem 2 of Chapter 1, §9.\n\nThus by Corollary 3 in \( §4,{F}^{\prime } = {G}^{\prime } \) on \( I - S \) implies \( G - F = c \) (constant) on each \( \left\lbrack {x, y}\right\rbrack \subseteq I \) ; hence \( G - F = c \) (or \( G = F + c \) ) on \( I \).
Yes
Corollary 1 (linearity). If \( \int f \) and \( \int g \) exist on \( I \), so does \( \int \left( {{pf} + {qg}}\right) \) for any scalars \( p, q \) (in the scalar field of \( E{)}^{.3} \) Moreover, for any \( a, b \in I \), we obtain\n\n(i) \( {\int }_{a}^{b}\left( {{pf} + {qg}}\right) = p{\int }_{a}^{b}f + q{\int }_{a}^{b}g \) ;\n\n(ii) \( {\int }_{a}^{b}\left( {f \pm g}\right) = {\int }_{a}^{b}f \pm {\int }_{a}^{b}g \) ; and\n\n(iii) \( {\int }_{a}^{b}{pf} = p{\int }_{a}^{b}f \) .
Proof. By assumption, there are \( F \) and \( G \) such that\n\n\[ \n{F}^{\prime } = f\text{ on }I - Q\text{ and }{G}^{\prime } = g\text{ on }I - P.\n\]\n\nThus, setting \( S = P \cup Q \) and \( H = {pF} + {qG} \), we have\n\n\[ \n{H}^{\prime } = p{F}^{\prime } + q{G}^{\prime } = {pf} + {qg}\text{ on }I - S,\n\]\n\nwith \( P, Q \), and \( S \) countable. Also, \( H = {pF} + {qG} \) is relatively continuous and finite on \( I \), as are \( F \) and \( G \) .\n\nThus by definition, \( H = \int \left( {{pf} + {qg}}\right) \) exists on \( I \), and by (1),\n\n\[ \n{\int }_{a}^{b}\left( {{pf} + {qg}}\right) = H\left( b\right) - H\left( a\right) = {pF}\left( b\right) + {qG}\left( b\right) - {pF}\left( a\right) - {qG}\left( a\right) = p{\int }_{a}^{b}f + q{\int }_{a}^{b}g,\n\]\n\nproving \( \left( {\mathrm{i}}^{ * }\right) \) .\n\nWith \( p = 1 \) and \( q = \pm 1 \), we obtain \( \left( {\mathrm{{ii}}}^{ * }\right) \) .\n\nTaking \( q = 0 \), we get \( \left( {\mathrm{{iii}}}^{ * }\right) \) .
Yes
Corollary 2. If both \( \int f \) and \( \int \left| f\right| \) exist on \( I = \left\lbrack {a, b}\right\rbrack \), then\n\n\[\n\left| {{\int }_{a}^{b}f}\right| \leq {\int }_{a}^{b}\left| f\right|\n\]
Proof. As before, let\n\n\[{F}^{\prime } = f\text{and}{G}^{\prime } = \left| f\right| \text{on}I - S\left( {S = Q \cup P\text{, all countable}}\right) \text{,}\]\n\nwhere \( F \) and \( G \) are relatively continuous and finite on \( I \) and \( G = \int \left| f\right| \) is real. Also, \( \left| {F}^{\prime }\right| = \left| f\right| = {G}^{\prime } \) on \( I - S \) . Thus by Theorem 1 of \( §4 \) ,\n\n\[ \left| {F\left( b\right) - F\left( a\right) }\right| \leq G\left( b\right) - G\left( a\right) = {\int }_{a}^{b}\left| f\right| .\]
Yes
Corollary 3. If \( \int f \) exists on \( I = \left\lbrack {a, b}\right\rbrack \), exact on \( I - Q \), then\n\n\[ \left| {{\int }_{a}^{b}f}\right| \leq M\left( {b - a}\right) \]\n\nfor some real\n\n\[ M \leq \mathop{\sup }\limits_{{t \in I - Q}}\left| {f\left( t\right) }\right| \]
This is simply Corollary 1 of \( §4 \), when applied to a primitive, \( F = \int f \) .
No
Corollary 4. If \( F = \int f \) on \( I \) and \( f = g \) on \( I - Q \), then \( F \) is also a primitive of \( g \), and\n\n\[{\int }_{a}^{b}f = {\int }_{a}^{b}g\;\text{ for }a, b \in I\]\n\n(Thus we may arbitrarily redefine \( f \) on a countable \( Q \) .)
Proof. Let \( {F}^{\prime } = f \) on \( I - P \) . Then \( {F}^{\prime } = g \) on \( I - \left( {P \cup Q}\right) \) . The rest is clear.
No
Corollary 5 (integration by parts). Let \( f \) and \( g \) be real or complex (or let \( f \) be scalar valued and \( g \) vector valued), both relatively continuous on \( I \) and differentiable on \( I - Q \) . Then if \( \int {f}^{\prime }g \) exists on \( I \), so does \( \int f{g}^{\prime } \), and we have\n\n\[{\int }_{a}^{b}f{g}^{\prime } = f\left( b\right) g\left( b\right) - f\left( a\right) g\left( a\right) - {\int }_{a}^{b}{f}^{\prime }g\;\text{ for any }a, b \in I.\]
Proof. By assumption, \( {fg} \) is relatively continuous and finite on \( I \), and\n\n\[{\left( fg\right) }^{\prime } = f{g}^{\prime } + {f}^{\prime }g\text{ on }I - Q.\]\n\nThus, setting \( H = {fg} \), we have \( H = \int \left( {f{g}^{\prime } + {f}^{\prime }g}\right) \) on \( I \) . Hence by Corollary 1, if \( \int {f}^{\prime }g \) exists on \( I \), so does \( \int \left( {\left( {f{g}^{\prime } + {f}^{\prime }g}\right) - {f}^{\prime }g}\right) = \int f{g}^{\prime } \), and\n\n\[{\int }_{a}^{b}f{g}^{\prime } + {\int }_{a}^{b}{f}^{\prime }g = {\int }_{a}^{b}\left( {f{g}^{\prime } + {f}^{\prime }g}\right) = H\left( b\right) - H\left( a\right) = f\left( b\right) g\left( b\right) - f\left( a\right) g\left( a\right) .\]\n\nThus (2) follows.
Yes
A function \( f : {E}^{1} \rightarrow {E}^{n}\left( {{}^{ * }{C}^{n}}\right) \) is integrable on \( I \) iff all its components \( \left( {{f}_{1},{f}_{2},\ldots ,{f}_{n}}\right) \) are, and then (by Theorem 5 in \( §1 \) )
\[ {\int }_{a}^{b}f = \left( {{\int }_{a}^{b}{f}_{1},\ldots ,{\int }_{a}^{b}{f}_{n}}\right) = \mathop{\sum }\limits_{{k = 1}}^{n}{\overrightarrow{e}}_{k}{\int }_{a}^{b}{f}_{k}\;\text{ for any }a, b \in I. \] Hence if \( f \) is complex, \[ {\int }_{a}^{b}f = {\int }_{a}^{b}{f}_{\mathrm{{re}}} + i \cdot {\int }_{a}^{b}{f}_{\mathrm{{im}}} \] (see Chapter 4, §3, Note 5).
Yes
Corollary 8. If \( f = 0 \) on \( I - Q \), then \( \int f \) exists on \( I \), and
\[ \left| {{\int }_{a}^{b}f}\right| = {\int }_{a}^{b}\left| f\right| = 0\;\text{ for }a, b \in I. \]
No
Theorem 2 (change of variables). Suppose \( g : {E}^{1} \rightarrow {E}^{1} \) (real) is differentiable on \( I \), while \( f : {E}^{1} \rightarrow E \) has a primitive on \( g\left\lbrack I\right\rbrack ,{}^{4} \) exact on \( g\left\lbrack {I - Q}\right\rbrack \) . Then \[ \int f\left( {g\left( x\right) }\right) {g}^{\prime }\left( x\right) {dx}\;\left( {\text{ i.e.,}\int \left( {f \circ g}\right) {g}^{\prime }}\right) \] exists on \( I \), and for any \( a, b \in I \), we have \[ {\int }_{a}^{b}f\left( {g\left( x\right) }\right) {g}^{\prime }\left( x\right) {dx} = {\int }_{p}^{q}f\left( y\right) {dy},\text{ where }p = g\left( a\right) \text{ and }q = g\left( b\right) . \]
Proof. Let \( F = \int f \) on \( g\left\lbrack I\right\rbrack \), and \( {F}^{\prime } = f \) on \( g\left\lbrack {I - Q}\right\rbrack \) . Then the composite function \( H = F \circ g \) is relatively continuous and finite on \( I \) . (Why?) By Theorem 3 of \( §1 \) , \[ {H}^{\prime }\left( x\right) = {F}^{\prime }\left( {g\left( x\right) }\right) {g}^{\prime }\left( x\right) \text{ for }x \in I - Q; \] i.e., \[ {H}^{\prime } = \left( {{F}^{\prime } \circ g}\right) {g}^{\prime }\text{ on }I - Q. \] Thus \( H = \int \left( {f \circ g}\right) {g}^{\prime } \) exists on \( I \), and \[ {\int }_{a}^{b}\left( {f \circ g}\right) {g}^{\prime } = H\left( b\right) - H\left( a\right) = F\left( {g\left( b\right) }\right) - F\left( {g\left( a\right) }\right) = F\left( q\right) - F\left( p\right) = {\int }_{p}^{q}f. \]
Yes
Theorem 3. If \( f, g : {E}^{1} \rightarrow {E}^{1} \) are integrable on \( I = \left\lbrack {a, b}\right\rbrack \), then we have the following:\n\n(i) \( f \geq 0 \) on \( I - Q \) implies \( {\int }_{a}^{b}f \geq 0 \) .
Proof. By Corollary 4, we may redefine \( f \) on \( Q \) so that our assumptions in (i)-(iv) hold on all of \( I \) . Thus we write \
No
Corollary 9 (first law of the mean). If \( f \) is real and \( \int f \) exists on \( \left\lbrack {a, b}\right\rbrack \), exact on \( \left( {a, b}\right) \), then\n\n\[{\int }_{a}^{b}f = f\left( q\right) \left( {b - a}\right) \text{ for some }q \in \left( {a, b}\right) .
Proof. Apply Corollary 3 in \( §2 \) to the function \( F = \int f \) .
No
Theorem 2. Let \( f : {E}^{1} \rightarrow {E}^{ * } \) be of class \( {\mathrm{{CD}}}^{n} \) on \( {G}_{p}\left( \delta \right) \) for an even number \( n \geq 2 \), and let\n\n\[ \n{f}^{\left( k\right) }\left( p\right) = 0\text{ for }k = 1,2,\ldots, n - 1, \n\] \n\nwhile \n\n\[ \n{f}^{\left( n\right) }\left( p\right) < 0\text{ (respectively,}{f}^{\left( n\right) }\left( p\right) > 0\text{). } \n\] \n\nThen \( f\left( p\right) \) is the maximum (respectively, minimum) value of \( f \) on some \( {G}_{p}\left( \varepsilon \right) \) , \( \varepsilon \leq \delta \) .\n\nIf, however, these conditions hold for some odd \( n \geq 1 \) (i.e., the first nonvanishing derivative at \( p \) is of odd order), \( f \) has no maximum or minimum at \( p \) .
Proof. As \n\n\[ \n{f}^{\left( k\right) }\left( p\right) = 0,\;k = 1,2,\ldots, n - 1, \n\] \n\nTheorem \( {1}^{\prime } \) (with \( n \) replaced by \( n - 1 \) ) yields \n\n\[ \nf\left( x\right) = f\left( p\right) + {f}^{\left( n\right) }\left( {q}_{n}\right) \frac{{\left( x - p\right) }^{n}}{n!}\;\text{ for all }x \in {G}_{p}\left( \delta \right) , \n\] \n\nwith \( {q}_{n} \) between \( x \) and \( p \) .\n\nAlso, as \( f \in {\mathrm{{CD}}}^{n},{f}^{\left( n\right) } \) is continuous at \( p \) . Thus if \( {f}^{\left( n\right) }\left( p\right) < 0 \), then \( {f}^{\left( n\right) } < 0 \) on some \( {G}_{p}\left( \varepsilon \right) ,0 < \varepsilon \leq \delta \) . However, \( x \in {G}_{p}\left( \varepsilon \right) \) implies \( {q}_{n} \in {G}_{p}\left( \varepsilon \right) \), so \n\n\[ \n{f}^{\left( n\right) }\left( {q}_{n}\right) < 0 \n\] \n\nwhile \n\n\[ \n{\left( x - p\right) }^{n} \geq 0\text{ if }n\text{ is even. \n\] \n\nIt follows that \n\n\[ \n{f}^{\left( n\right) }\left( {q}_{n}\right) \frac{{\left( x - p\right) }^{n}}{n!} \leq 0 \n\] \n\nand so \n\n\[ \nf\left( x\right) = f\left( p\right) + {f}^{\left( n\right) }\left( {q}_{n}\right) \frac{{\left( x - p\right) }^{n}}{n!} \leq f\left( p\right) \;\text{ for }x \in {G}_{p}\left( \varepsilon \right) , \n\] \n\ni.e., \( f\left( p\right) \) is the maximum value of \( f \) on \( {G}_{p}\left( \varepsilon \right) \), as claimed.\n\nSimilarly, in the case \( {f}^{\left( n\right) }\left( p\right) > 0 \), a minimum would result.\n\nIf, however, \( n \) is odd, then \( {\left( x - p\right) }^{n} \) is negative for \( x < p \) but positive for \( x > p \) . The same argument then shows that \( f\left( x\right) < f\left( p\right) \) on one side of \( p \) and \( f\left( x\right) > f\left( p\right) \) on the other side; thus no local maximum or minimum can exist at \( p \) . This completes the proof.
Yes
Corollary 1. The sum \( S\left( {f, P}\right) = \ell W \) cannot decrease when \( P \) is refined.
Thus when new partition points are added, \( S\left( {f, P}\right) \) grows in general; i.e., it approaches some supremum value (finite or not). Roughly speaking, the inscribed polygon \( W \) gets \
No
Corollary 2 (monotonicity of \( {V}_{f} \) ). If \( a \leq c \leq d \leq b \), then\n\n\[ \n{V}_{f}\left\lbrack {c, d}\right\rbrack \leq {V}_{f}\left\lbrack {a, b}\right\rbrack \n\]
Proof. By Theorem 1,\n\n\[ \n{V}_{f}\left\lbrack {a, b}\right\rbrack = {V}_{f}\left\lbrack {a, c}\right\rbrack + {V}_{f}\left\lbrack {c, d}\right\rbrack + {V}_{f}\left\lbrack {d, b}\right\rbrack \geq {V}_{f}\left\lbrack {c, d}\right\rbrack . \n\]
Yes
For each \( t \in \left\lbrack {a, b}\right\rbrack \) , \n\n\[ \left| {f\left( t\right) - f\left( a\right) }\right| \leq {V}_{f}\left\lbrack {a, b}\right\rbrack \] \n\nHence if \( f \) is of bounded variation on \( \left\lbrack {a, b}\right\rbrack \), it is bounded on \( \left\lbrack {a, b}\right\rbrack \) .
Proof. If \( t \in \left\lbrack {a, b}\right\rbrack \), let \( P = \{ a, t, b\} \), so \n\n\[ \left| {f\left( t\right) - f\left( a\right) }\right| \leq \left| {f\left( t\right) - f\left( a\right) }\right| + \left| {f\left( b\right) - f\left( t\right) }\right| = S\left( {f, P}\right) \leq {V}_{f}\left\lbrack {a, b}\right\rbrack , \] \n\nproving our first assertion. \( {}^{3} \) Hence \n\n\[ \left( {\forall t \in \left\lbrack {a, b}\right\rbrack }\right) \;\left| {f\left( t\right) }\right| \leq \left| {f\left( t\right) - f\left( a\right) }\right| + \left| {f\left( a\right) }\right| \leq {V}_{f}\left\lbrack {a, b}\right\rbrack + \left| {f\left( a\right) }\right| . \] \n\nThis proves the second assertion.
Yes
Corollary 4. A function \( f \) is finite and constant on \( \left\lbrack {a, b}\right\rbrack \) iff \( {V}_{f}\left\lbrack {a, b}\right\rbrack = 0 \) .
The proof is left to the reader. (Use Corollary 3 and the definitions.)
No
Theorem 2. Let \( f, g, h \) be real or complex (or let \( f \) and \( g \) be vector valued and \( h \) scalar valued). Then on any interval \( I = \left\lbrack {a, b}\right\rbrack \), we have\n\n(i) \( {V}_{\left| f\right| } \leq {V}_{f} \)\n\n(ii) \( {V}_{f \pm g} \leq {V}_{f} + {V}_{g} \) ; and\n\n(iii) \( {V}_{hf} \leq s{V}_{f} + r{V}_{h} \), with \( r = \mathop{\sup }\limits_{{t \in I}}\left| {f\left( t\right) }\right| \) and \( s = \mathop{\sup }\limits_{{t \in I}}\left| {h\left( t\right) }\right| \) .\n\nHence if \( f, g \), and \( h \) are of bounded variation on \( I \), so are \( f \pm g,{hf} \), and \( \left| f\right| \) .
Proof. We first prove (iii).\n\nTake any partition \( P = \left\{ {{t}_{0},\ldots ,{t}_{m}}\right\} \) of \( I \) . Then\n\n\[ \left| {{\Delta }_{i}{hf}}\right| = \left| {h\left( {t}_{i}\right) f\left( {t}_{i}\right) - h\left( {t}_{i - 1}\right) f\left( {t}_{i - 1}\right) }\right| \]\n\n\[ \leq \left| {h\left( {t}_{i}\right) f\left( {t}_{i}\right) - h\left( {t}_{i - 1}\right) f\left( {t}_{i}\right) }\right| + \left| {h\left( {t}_{i - 1}\right) f\left( {t}_{i}\right) - h\left( {t}_{i - 1}\right) f\left( {t}_{i - 1}\right) }\right| \]\n\n\[ = \left| {f\left( {t}_{i}\right) }\right| \left| {{\Delta }_{i}h}\right| + \left| {h\left( {t}_{i - 1}\right) }\right| \left| {{\Delta }_{i}f}\right| \]\n\n\[ \leq r\left| {{\Delta }_{i}h}\right| + s\left| {{\Delta }_{i}f}\right| \]\n\nAdding these inequalities, we obtain\n\n\[ S\left( {{hf}, P}\right) \leq r \cdot S\left( {h, P}\right) + s \cdot S\left( {f, P}\right) \leq r{V}_{h} + s{V}_{f}. \]\n\n\( {}^{3} \) By our conventions, it also follows that \( \left| {f\left( a\right) }\right| \) is a finite constant, and so is \( {V}_{f}\left\lbrack {a, b}\right\rbrack + \left| {f\left( a\right) }\right| \) if \( {V}_{f}\left\lbrack {a, b}\right\rbrack < + \infty \) . As this holds for all sums \( S\left( {{hf}, P}\right) \), it holds for their supremum, so\n\n\[ {V}_{hf} = \sup S\left( {{hf}, P}\right) \leq r{V}_{h} + s{V}_{f} \]\n\n as claimed.\n\nSimilarly, (i) follows from\n\n\[ \left| \right| f\left( {t}_{i}\right) \left| -\right| f\left( {t}_{i - 1}\right) \left| \right| \leq \left| {f\left( {t}_{i}\right) - f\left( {t}_{i - 1}\right) }\right| . \]\n\nThe analogous proof of (ii) is left to the reader.\n\nFinally,(i)-(iii) imply that \( {V}_{f},{V}_{f \pm g} \), and \( {V}_{hf} \) are finite if \( {V}_{f},{V}_{g} \), and \( {V}_{h} \) are. This proves our last assertion.
No
(ii) If \( f \) is real and monotone on \( I \), it is of bounded variation there.
Proof. We prove (ii) first.\n\nLet \( f \uparrow \) on \( I \) . If \( P = \left\{ {{t}_{0},\ldots ,{t}_{m}}\right\} \), then\n\n\[ \n{t}_{i} \geq {t}_{i - 1}\text{ implies }f\left( {t}_{i}\right) \geq f\left( {t}_{i - 1}\right) .\n\]\n\nHence \( {\Delta }_{i}f \geq 0 \) ; i.e., \( \left| {{\Delta }_{i}f}\right| = {\Delta }_{i}f \) . Thus\n\n\[ \nS\left( {f, P}\right) = \mathop{\sum }\limits_{{i = 1}}^{m}\left| {{\Delta }_{i}f}\right| = \mathop{\sum }\limits_{{i = 1}}^{m}{\Delta }_{i}f = \mathop{\sum }\limits_{{i = 1}}^{m}\left\lbrack {f\left( {t}_{i}\right) - f\left( {t}_{i - 1}\right) }\right\rbrack \n\]\n\n\[ \n= f\left( {t}_{m}\right) - f\left( {t}_{0}\right) = f\left( b\right) - f\left( a\right) \n\]\n\nfor any \( P \) . (Verify!) This implies that also\n\n\[ \n{V}_{f}\left\lbrack I\right\rbrack = \sup S\left( {f, P}\right) = f\left( b\right) - f\left( a\right) < + \infty .\n\]\n\nThus (ii) is proved.
Yes
Theorem 4. (i) A function \( f : {E}^{1} \rightarrow {E}^{n}\left( {{}^{ * }{C}^{n}}\right) \) is of bounded variation on \( I = \left\lbrack {a, b}\right\rbrack \) iff all of its components \( \left( {{f}_{1},{f}_{2},\ldots ,{f}_{n}}\right) \) are.
## Proof. (i) Take any partition \( P = \left\{ {{t}_{0},\ldots ,{t}_{m}}\right\} \) of \( I \) . Then \[ {\left| {f}_{k}\left( {t}_{i}\right) - {f}_{k}\left( {t}_{i - 1}\right) \right| }^{2} \leq \mathop{\sum }\limits_{{j = 1}}^{n}{\left| {f}_{j}\left( {t}_{i}\right) - {f}_{j}\left( {t}_{i - 1}\right) \right| }^{2} = {\left| f\left( {t}_{i}\right) - f\left( {t}_{i - 1}\right) \right| }^{2}; \] i.e., \( \left| {{\Delta }_{i}{f}_{k}}\right| \leq \left| {{\Delta }_{i}f}\right|, i = 1,2,\ldots, m \) . Thus \[ \left( {\forall P}\right) \;S\left( {{f}_{k}, P}\right) \leq S\left( {f, P}\right) \leq {V}_{f}, \] and \( {V}_{{f}_{k}} \leq {V}_{f} \) follows. Thus \[ {V}_{f} < + \infty \text{ implies }{V}_{{f}_{k}} < + \infty ,\;k = 1,2,\ldots, n. \] The converse follows by Theorem 2 since \( f = \mathop{\sum }\limits_{{k = 1}}^{n}{f}_{k}{\overrightarrow{e}}_{k} \) . (Explain!)
No
Theorem 1. The following are equivalent:\n\n(i) \( f \) is (weakly) absolutely continuous on \( I = \\left\\lbrack {a, b}\\right\\rbrack \) ;\n\n(ii) \( {v}_{f} \) is finite and relatively continuous on \( I \) ; and\n\n(iii) \( \\left( {\\forall \\varepsilon > 0}\\right) \\left( {\\exists \\delta > 0}\\right) \\left( {\\forall x, y \\in I \\mid 0 \\leq y - x < \\delta }\\right) {V}_{f}\\left\\lbrack {x, y}\\right\\rbrack < \\varepsilon \) .
Proof. We shall show that (ii) \( \\Rightarrow \) (iii) \( \\Rightarrow \) (i) \( \\Rightarrow \) (ii).\n\n(ii) \( \\Rightarrow \) (iii). As \( I = \\left\\lbrack {a, b}\\right\\rbrack \) is compact,(ii) implies that \( {v}_{f} \) is uniformly continuous on \( I \) (Theorem 4 of Chapter 4,§8). Thus\n\n\[\\left( {\\forall \\varepsilon > 0}\\right) \\left( {\\exists \\delta > 0}\\right) \\left( {\\forall x, y \\in I \\mid 0 \\leq y - x < \\delta }\\right) \\;{v}_{f}\\left( y\\right) - {v}_{f}\\left( x\\right) < \\varepsilon .\]\n\nHowever,\n\n\[{v}_{f}\\left( y\\right) - {v}_{f}\\left( x\\right) = {V}_{f}\\left\\lbrack {a, y}\\right\\rbrack - {V}_{f}\\left\\lbrack {a, x}\\right\\rbrack = {V}_{f}\\left\\lbrack {x, y}\\right\\rbrack\]\n\nby additivity (Theorem 1 in \( §7 \) ). Thus (iii) follows.\n\n(iii) \( \\Rightarrow \) (i). By Corollary 3 of \( §7,\\left| {f\\left( x\\right) - f\\left( y\\right) }\\right| \\leq {V}_{f}\\left\\lbrack {x, y}\\right\\rbrack \) . Therefore,(iii) implies that\n\n\[\\left( {\\forall \\varepsilon > 0}\\right) \\left( {\\exists \\delta > 0}\\right) \\left( {\\forall x, y \\in I\\left| \\right| x - y \\mid < \\delta }\\right) \\;\\left| {f\\left( x\\right) - f\\left( y\\right) }\\right| < \\varepsilon ,\]\n\nand so \( f \) is relatively (even uniformly) continuous on \( I \) .\n\nNow with \( \\varepsilon \) and \( \\delta \) as in (iii), take a partition \( P = \\left\{ {{t}_{0},\\ldots ,{t}_{m}}\\right\} \) of \( I \) so fine that\n\n\[{t}_{i} - {t}_{i - 1} < \\delta ,\\;i = 1,2,\\ldots, m.\]\n\nThen \( \\left( {\\forall i}\\right) {V}_{f}\\left\\lbrack {{t}_{i - 1},{t}_{i}}\\right\\rbrack < \\varepsilon \) . Adding up these \( m \) inequalities and using the additivity of \( {V}_{f} \), we obtain\n\n\[{V}_{f}\\left\\lbrack I\\right\\rbrack = \\mathop{\\sum }\\limits_{{i = 1}}^{m}{V}_{f}\\left\\lbrack {{t}_{i - 1},{t}_{i}}\\right\\rbrack < {m\\varepsilon } < + \\infty .\]\n\nThus (i) follows, by definition.\n\nThat (i) \( \\Rightarrow \) (ii) is given as the next theorem.
Yes
Theorem 2. If \( {V}_{f}\left\lbrack I\right\rbrack < + \infty \) and if \( f \) is relatively continuous at some \( p \in I \) (over \( I = \left\lbrack {a, b}\right\rbrack \) ), then the same applies to the length function \( {v}_{f} \) .
Proof. We consider left continuity first, with \( a < p \leq b \) .\n\nLet \( \varepsilon > 0 \) . By assumption, there is \( \delta > 0 \) such that\n\n\[ \left| {f\left( x\right) - f\left( p\right) }\right| < \frac{\varepsilon }{2}\text{ when }\left| {x - p}\right| < \delta \text{ and }x \in \left\lbrack {a, p}\right\rbrack . \]\n\nFix any such \( x \) . Also, \( {V}_{f}\left\lbrack {a, p}\right\rbrack = \mathop{\sup }\limits_{P}S\left( {f, P}\right) \) over \( \left\lbrack {a, p}\right\rbrack \) . Thus\n\n\[ {V}_{f}\left\lbrack {a, p}\right\rbrack - \frac{\varepsilon }{2} < \mathop{\sum }\limits_{{i = 1}}^{k}\left| {{\Delta }_{i}f}\right| \]\n\nfor some partition\n\n\[ P = \left\{ {{t}_{0} = a,\ldots ,{t}_{k - 1},{t}_{k} = p}\right\} \text{of}\left\lbrack {a, p}\right\rbrack \text{. (Why?)} \]\n\nWe may assume \( {t}_{k - 1} = x, x \) as above. (If \( {t}_{k - 1} \neq x \), add \( x \) to \( P \) .) Then\n\n\[ \left| {{\Delta }_{k}f}\right| = \left| {f\left( p\right) - f\left( x\right) }\right| < \frac{\varepsilon }{2} \]\n\nand hence\n\n\[ {V}_{f}\left\lbrack {a, p}\right\rbrack - \frac{\varepsilon }{2} < \mathop{\sum }\limits_{{i = 1}}^{{k - 1}}\left| {{\Delta }_{i}f}\right| + \left| {{\Delta }_{k}f}\right| < \mathop{\sum }\limits_{{i = 1}}^{{k - 1}}\left| {{\Delta }_{i}f}\right| + \frac{\varepsilon }{2} \leq {V}_{f}\left\lbrack {a,{t}_{k - 1}}\right\rbrack + \frac{\varepsilon }{2}. \]\n\n(1)\n\nHowever,\n\n\[ {V}_{f}\left\lbrack {a, p}\right\rbrack = {v}_{f}\left( p\right) \]\n\nand\n\n\[ {V}_{f}\left\lbrack {a,{t}_{k - 1}}\right\rbrack = {V}_{f}\left\lbrack {a, x}\right\rbrack = {v}_{f}\left( x\right) . \]\n\nThus (1) yields\n\n\[ \left| {{v}_{f}\left( p\right) - {v}_{f}\left( x\right) }\right| = {V}_{f}\left\lbrack {a, p}\right\rbrack - {V}_{f}\left\lbrack {a, x}\right\rbrack < \varepsilon \text{for}x \in \left\lbrack {a, p}\right\rbrack \text{with}\left| {x - p}\right| < \delta \text{.} \]\n\nThis shows that \( {v}_{f} \) is left continuous at \( p \) .\n\nRight continuity is proved similarly on noting that\n\n\[ {v}_{f}\left( x\right) - {v}_{f}\left( p\right) = {V}_{f}\left\lbrack {p, b}\right\rbrack - {V}_{f}\left\lbrack {x, b}\right\rbrack \text{ for }p \leq x < b.\text{ (Why?) } \]\n\nThus \( {v}_{f} \) is, indeed, relatively continuous at \( p \) . Observe that \( {v}_{f} \) is also of bounded variation on \( I \), being monotone and finite (see Theorem 3(ii) of \( §7 \) ).\n\nThis completes the proof of both Theorem 2 and Theorem 1.
No
Corollary 1. If \( f \) is real and absolutely continuous on \( I = \left\lbrack {a, b}\right\rbrack \) (weakly), so are the nondecreasing functions \( g \) and \( h\left( {f = g - h}\right) \) defined in Theorem 3 of \( §7 \) .
Indeed, the function \( g \) as defined there is simply \( {v}_{f} \) . Thus it is relatively continuous and finite on \( I \) by Theorem 1. Hence so also is \( h = f - g \) . Both are of bounded variation (monotone!) and hence absolutely continuous (weakly).
Yes
Corollary 2. If \( F = \int f \) on \( I = \left\lbrack {a, b}\right\rbrack \) and if \( f \) is bounded \( \left( {\left| f\right| \leq K \in {E}^{1}}\right) \) on \( I - Q \) ( \( Q \) countable), then \( F \) is weakly absolutely continuous on \( I \) .
Proof. By definition, \( F = \int f \) is finite and relatively continuous on \( I \), so we only have to show that \( {V}_{F}\left\lbrack I\right\rbrack < + \infty \) . This, however, easily follows by Problem 3 of \( §7 \) on noting that \( {F}^{\prime } = f \) on \( I - S \) ( \( S \) countable). Details are left to the reader.
No
Theorem 3. If \( f : {E}^{1} \rightarrow E \) is continuously differentiable on \( I = \left\lbrack {a, b}\right\rbrack \left( {§6}\right) \) , then \( {v}_{f} = \int \left| {f}^{\prime }\right| \) on \( I \) and\n\n\[ \n{V}_{f}\left\lbrack {a, b}\right\rbrack = {\int }_{a}^{b}\left| {f}^{\prime }\right| .\n\]
Proof. Let \( a < p < x \leq b,{\Delta x} = x - p \), and\n\n\[ \n\Delta {v}_{f} = {v}_{f}\left( x\right) - {v}_{f}\left( p\right) = {V}_{f}\left\lbrack {p, x}\right\rbrack .\;\text{ (Why?) }\n\]\n\nAs a first step, we shall show that\n\n\[ \n\frac{\Delta {v}_{f}}{\Delta x} \leq \mathop{\sup }\limits_{\left\lbrack p, x\right\rbrack }\left| {f}^{\prime }\right| \n\]\n\n(2)\n\nFor any partition \( P = \left\{ {p = {t}_{0},\ldots ,{t}_{m} = x}\right\} \) of \( \left\lbrack {p, x}\right\rbrack \), we have\n\n\[ \nS\left( {f, P}\right) = \mathop{\sum }\limits_{{i = 1}}^{m}\left| {{\Delta }_{i}f}\right| \leq \mathop{\sum }\limits_{{i = 1}}^{m}\mathop{\sup }\limits_{\left\lbrack {t}_{i - 1},{t}_{i}\right\rbrack }\left| {f}^{\prime }\right| \left( {{t}_{i} - {t}_{i - 1}}\right) \leq \mathop{\sup }\limits_{\left\lbrack p, x\right\rbrack }\left| {f}^{\prime }\right| {\Delta x}.\n\]\n\nSince this holds for any partition \( P \), we have\n\n\[ \n{V}_{f}\left\lbrack {p, x}\right\rbrack \leq \mathop{\sup }\limits_{\left\lbrack p, x\right\rbrack }\left| {f}^{\prime }\right| {\Delta x} \n\]\n\nwhich implies (2).\n\nOn the other hand,\n\n\[ \n\Delta {v}_{f} = {V}_{f}\left\lbrack {p, x}\right\rbrack \geq \left| {f\left( x\right) - f\left( p\right) }\right| = \left| {\Delta f}\right| .\n\]\n\nCombining, we get\n\n\[ \n\left| \frac{\Delta f}{\Delta x}\right| \leq \frac{\Delta {v}_{f}}{\Delta x} \leq \mathop{\sup }\limits_{\left\lbrack p, x\right\rbrack }\left| {f}^{\prime }\right| < + \infty \n\]\n\n(3)\n\nsince \( {f}^{\prime } \) is relatively continuous on \( \left\lbrack {a, b}\right\rbrack \), hence also uniformly continuous and bounded. (Here we assumed \( a < p < x \leq b \) . However,(3) holds also if \( a \leq x < p < b \), with \( \Delta {v}_{f} = - V\left\lbrack {x, p}\right\rbrack \) and \( {\Delta x} < 0 \) . Verify!)\n\nNow\n\n\[ \n\left| \right| {f}^{\prime }\left( p\right) \left| -\right| {f}^{\prime }\left( x\right) \left| \right| \leq \left| {{f}^{\prime }\left( p\right) - {f}^{\prime }\left( x\right) }\right| \rightarrow 0\;\text{ as }x \rightarrow p,\n\]\n\nso, taking limits as \( x \rightarrow p \), we obtain\n\n\[ \n\mathop{\lim }\limits_{{x \rightarrow p}}\frac{\Delta {v}_{f}}{\Delta x} = \left| {{f}^{\prime }\left( p\right) }\right| \n\]\n\nThus \( {v}_{f} \) is differentiable at each \( p \) in \( \left( {a, b}\right) \), with \( {v}_{f}^{\prime }\left( p\right) = \left| {{f}^{\prime }\left( p\right) }\right| \) . Also, \( {v}_{f} \) is relatively continuous and finite on \( \left\lbrack {a, b}\right\rbrack \) (by Theorem 1). \( {}^{2} \) Hence \( {v}_{f} = \int \left| {f}^{\prime }\right| \) on \( \left\lbrack {a, b}\right\rbrack \), and we obtain\n\n\[ \n{\int }_{a}^{b}\left| {f}^{\prime }\right| = {v}_{f}\left( b\right) - {v}_{f}\left( a\right) = {V}_{f}\left\lbrack {a, b}\right\rbrack ,\text{ as asserted. }\n\]\n\n(4)
Yes
Theorem 1. Let \( {F}_{n} : {E}^{1} \rightarrow E\left( {n = 1,2,\ldots }\right) \) be finite and relatively continuous on \( I \) and differentiable on \( I - Q \) . Suppose that\n\n(a) \( \mathop{\lim }\limits_{{n \rightarrow \infty }}{F}_{n}\left( p\right) \) exists for some \( p \in I \) ;\n\n(b) \( {F}_{n}^{\prime } \rightarrow f \neq \pm \infty \) (uniformly) on \( J - Q \) for each finite subinterval \( J \subseteq I \) ;\n\n(c) \( E \) is complete.\n\nThen\n\n(i) \( \mathop{\lim }\limits_{{n \rightarrow \infty }}{F}_{n} = F \) exists uniformly on each finite subinterval \( J \subseteq I \) ;\n\n(ii) \( F = \int f \) on \( I \) ; and\n\n(iii) \( {\left( \lim {F}_{n}\right) }^{\prime } = {F}^{\prime } = f = \mathop{\lim }\limits_{{n \rightarrow \infty }}{F}_{n}^{\prime } \) on \( I - Q \) .
Proof. Fix \( \varepsilon > 0 \) and any subinterval \( J \subseteq I \) of length \( \delta < \infty \), with \( p \in J(p \) as in (a)). By (b), \( {F}_{n}^{\prime } \rightarrow f \) (uniformly) on \( J - Q \), so there is a \( k \) such that for \( m, n > k \) ,\n\n\[ \left| {{F}_{n}^{\prime }\left( t\right) - f\left( t\right) }\right| < \frac{\varepsilon }{2},\;t \in J - Q; \]\n\n(1)\n\nhence\n\n\[ \mathop{\sup }\limits_{{t \in J - Q}}\left| {{F}_{m}^{\prime }\left( t\right) - {F}_{n}^{\prime }\left( t\right) }\right| \leq \varepsilon .\;\text{ (Why?) } \]\n\n(2)\n\nNow apply Corollary 1 in \( §4 \) to the function \( h = {F}_{m} - {F}_{n} \) on \( J \) . Then for each \( x \in J,\left| {h\left( x\right) - h\left( p\right) }\right| \leq M\left| {x - p}\right| \), where\n\n\[ M \leq \mathop{\sup }\limits_{{t \in J - Q}}\left| {{h}^{\prime }\left( t\right) }\right| \leq \varepsilon \]\n\nby (2). Hence for \( m, n > k, x \in J \) and\n\n\[ \left| {{F}_{m}\left( x\right) - {F}_{n}\left( x\right) - {F}_{m}\left( p\right) + {F}_{n}\left( p\right) }\right| \leq \varepsilon \left| {x - p}\right| \leq {\varepsilon \delta }. \]\n\n(3)\n\nAs \( \varepsilon \) is arbitrary, this shows that the sequence\n\n\[ \left\{ {{F}_{n} - {F}_{n}\left( p\right) }\right\} \]\n\nsatisfies the uniform Cauchy criterion (Chapter 4, \( §{12} \), Theorem 3). Thus as \( E \) is complete, \( \left\{ {{F}_{n} - {F}_{n}\left( p\right) }\right\} \) converges uniformly on \( J \) . So does \( \left\{ {F}_{n}\right\} \), for \( \left\{ {{F}_{n}\left( p\right) }\right\} \) converges, by (a). Thus we may set\n\n\[ F = \lim {F}_{n}\text{ (uniformly) on }J, \]\nproving assertion (i). \( {}^{1} \)
No
Theorem 2. Let the functions \( {f}_{n} : {E}^{1} \rightarrow E, n = 1,2,\ldots \), have antiderivatives, \( {F}_{n} = \int {f}_{n} \), on \( I \). Suppose \( E \) is complete and \( {f}_{n} \rightarrow f \) (uniformly) on each finite subinterval \( J \subseteq I \), with \( f \) finite there. Then \( \int f \) exists on \( I \), and \[ {\int }_{p}^{x}f = {\int }_{p}^{x}\mathop{\lim }\limits_{{n \rightarrow \infty }}{f}_{n} = \mathop{\lim }\limits_{{n \rightarrow \infty }}{\int }_{p}^{x}{f}_{n}\text{ for any }p, x \in I. \]
Proof. Fix any \( p \in I \). By Note 2 in \( §5 \), we may choose \[ {F}_{n}\left( x\right) = {\int }_{p}^{x}{f}_{n}\text{ for }x \in I. \] Then \( {F}_{n}\left( p\right) = {\int }_{p}^{p}{f}_{n} = 0 \), and so \( \mathop{\lim }\limits_{{n \rightarrow \infty }}{F}_{n}\left( p\right) = 0 \) exists, as required in Theorem 1(a). Also, by definition, each \( {F}_{n} \) is relatively continuous and finite on \( I \) and differentiable, with \( {F}_{n}^{\prime } = {f}_{n} \), on \( I - {Q}_{n} \). The countable sets \( {Q}_{n} \) need not be the same, so we replace them by \[ Q = \mathop{\bigcup }\limits_{{n = 1}}^{\infty }{Q}_{n} \] (including in \( Q \) also the endpoints of \( I \), if any). Then \( Q \) is countable (see Chapter 1,§9, Theorem 2), and \( I - Q \subseteq I - {Q}_{n} \), so all \( {F}_{n} \) are differentiable on \( I - Q \), with \( {F}_{n}^{\prime } = {f}_{n} \) there. Additionally, by assumption, \[ {f}_{n} \rightarrow f\text{ (uniformly) } \] on finite subintervals \( J \subseteq I \). Hence \[ {F}_{n}^{\prime } \rightarrow f\text{ (uniformly) on }J - Q \] for all such \( J \), and so the conditions of Theorem 1 are satisfied. By that theorem, then, \[ F = \int f = \lim {F}_{n}\text{ exists on }I \] and, recalling that \[ {F}_{n}\left( x\right) = {\int }_{p}^{x}{f}_{n} \] we obtain for \( x \in I \) \[ {\int }_{p}^{x}f = F\left( x\right) - F\left( p\right) = \lim {F}_{n}\left( x\right) - \lim {F}_{n}\left( p\right) = \lim {F}_{n}\left( x\right) - 0 = \lim {\int }_{p}^{x}{f}_{n}. \] As \( p \in I \) was arbitrary, and \( f = \lim {f}_{n} \) (by assumption), all is proved.
Yes