Let $f$ be continuous on $X$ and $(x_n)$ be a Cauchy sequence on $X$. Show that $(f(x_n))$ doesnt have to be a Cauchy sequence. And show that $(f(x_n))$ always has to be a Cauchy sequence when $f$ is uniformly continuous.

My attempt:

We take the Cauchy sequence $(x_n)=\frac{1}{n}$ and choose $f:\mathbb{R}^+\longrightarrow\mathbb{R}:x\mapsto\frac{1}{x}$ which is a continuous function on given domain.

Since $x_n \longrightarrow0\,\,\Longrightarrow (f(x_n))\longrightarrow+\infty$

So eventhough $(x_n)$ is a Cauchy sequence, $(f(x_n))$ on the continuous $f$ isnt!

$\exists f\in C^0(\mathbb{R}^+)$ where the requirement does not hold!

Now we show that from the fact that $f$ is uniformly continuous we can always conclude:

Is $(x_n)$ a Cauchy sequence $\Longrightarrow$ $(f(x_n))$ is a Cauchy sequence

A function $f$ is uniformly continuous on $X$ when the Heine-Cantor theorem holds:

$\forall \epsilon>0 \,\,\, \exists \delta>0 \,\,\,\forall x,y\in X:x\in \mathcal{U}_{\delta}(y)\Longrightarrow f(x)\in \mathcal{U}_{\epsilon}(f(y))$

So for a $N\in \mathbb{N}$ we know that for all $n>N:x_n\in \mathcal{U}_{\delta}(\xi)$

Where $\xi$ is the point $(x_n)$ is converging to.

Following the Heine-Cantor theorem we know that if $n>N:x_n\in \mathcal{U}_{\delta}(\xi) \Longrightarrow f(x_n)\in \mathcal{U}_{\epsilon}(f(\xi))$

So when $(x_n)$ is a Cauchy sequence, $(f(x_n))$ has to be one aswell!

Would be great if someone could check my reasoning and give me advice for improving :)

  • 2
    $\begingroup$ I'm convince by the first part. But not by the second one. Can you restate the definition of a Cauchy sequence? Also, I don't know why you mention Heine-Cantor theorem. Heine-Cantor theorem gives conditions implying that a map is uniform continuous. But here we're already supposing that it is uniform continuous. $\endgroup$ – mathcounterexamples.net May 30 '20 at 8:39
  • $\begingroup$ The second part follows from definition of uniform continuity and definition of a Cauchy sequence. I don't understand why you are quoting a theorem for this. $\endgroup$ – Kavi Rama Murthy May 30 '20 at 8:51
  • $\begingroup$ I thought that the cauchy sequence definition gives us, that a infinit amount of values lie in every epsilon region of our point of convergence. so there is this $\xi$ for which every $x_n$ with $N>n$ lies in any epsilon region around $\xi$. For any epsilon we might choose. And from the Heine Cantor theorem it follows that since the function is uniformly continuous our $(f(x_n))$ also has to lie in the close region of $f(\xi)$.. $\endgroup$ – CoffeeArabica May 30 '20 at 8:58
  • $\begingroup$ Cauchy sequence does not need to converge. $\endgroup$ – N.Quy May 30 '20 at 9:05
  • $\begingroup$ Ohh.. could you give me an example for that case? :) so I can ponder with that? $\endgroup$ – CoffeeArabica May 30 '20 at 9:13

Consider $A = \{1, 1/2, 1/3, · · ·\}$ and $f(1/n) = \begin{cases}1, &if ~~n ~is~ odd\\−1,& if~~ n~ is ~even\end{cases}$.

Then $f$ is continuous but not uniformly continuous. The sequence $(x_n)= \frac{1}{n}$ in $A$ is Cauchy but the sequence $f(x_n) = (1, −1, 1, −1, · · ·)$ is not Cauchy.

So the Uniformely continuous is necessary condition for the statement you want to prove.


If $f$ is uniformly continuous on $A$, then given $ε > 0$ there is $δ > 0$ such that if $x, y\in A$ and $|x−y| < δ$, then $|f(x) − f(y)| < ε$. Let $(x_n)$ be a Cauchy sequence in $A$. Then for given $δ > 0$ there is $M$ such that if $p, q > M$, then $|x_p − x_q| < δ$, and thus $|f(x_p) − f(x_q)| < ε$, implying that $(f(x_n))$ is a Cauchy sequence.

  • $\begingroup$ Sorry if i miss the point.. but wasnt i stating (especially) the second part in my post? $\endgroup$ – CoffeeArabica May 30 '20 at 9:01
  • $\begingroup$ @Coffee Arabic: Sorry,I don't see really your comment. I simply proved it by the definition of uniform continuity. I don't get that comment. $\endgroup$ – Aman Pandey May 30 '20 at 9:12

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.