Suppose that $g: (a,b] \rightarrow \mathbb{R}$ is uniformly continuous. Suppose that both $(x_n), (y_n)$ are sequence in $(a,b]$ which converge to $a$. Show that if $g(x_n)) \rightarrow l$ and $g((y_n)) \rightarrow m$, then $l=m$.

By definition, since $(x_n) \rightarrow a$, $\forall \varepsilon > 0$, $\exists N_0 \in \mathbb{N}$ $\forall n\geq N_0$


A parallel statement can be made for the sequence $(y_n)$. This implies that

$|x_n-y_n|=|x_n-a+a-y_n| \leq |x_n-a|+|a-y_n| \equiv |x_n-a|+|y_n-a| < \frac{\varepsilon}{2}+\frac{\varepsilon}{2}=\varepsilon$

Since $g$ is uniformly continuous, this informs us that $\forall \varepsilon>0$, $\exists \delta>0$ s.t. $\forall x_n \in (x_n)$ and $\forall y_n \in (y_n)$,

$|g(x_n)-g(y_n)|<\varepsilon$ (we exclude the delta condition since we showed that it is always satisfied)

Similar conditions can be made using each sequence exclusively



Struggling to figure out the last few steps.

  • $\begingroup$ Your proof needs work. Too many "$\forall \varepsilon > 0$, $\exists N_0 \in \mathbb{N}$ $\forall n\geq N_0$" statements. Suggestion: Start with "let $\epsilon > 0.$" Now, it seems you have the basic idea for proving $x_n-y_n \to 0.$ But this also converges to ___ . $\endgroup$
    – zhw.
    Apr 20, 2015 at 19:12

1 Answer 1


Here is a proof in the framework of IST, assuming $g$, $a$, $l$ and $m$ are standard. Let $n$ be infinitely large. Then $x_n$ and $y_n$ are both infinitely close to $a$, so $x_n$ is infinitely close to $y_n$. By uniform continuity, $g(x_n)$ is infinitely close to $g(y_n)$. Because $g(x_n)\to l$, $g(x_n)$ is infinitely close to $l$ and similarly $g(y_n)$ is infinitely close to $m$. Thus $l$ and $m$ are infinitely close, hence (being standard) equal.

  • 1
    $\begingroup$ Translating this idea into a non-non-standard proof is left as an exercise to the reader. $\endgroup$ Apr 20, 2015 at 18:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.