Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Let $f(x),f_n(x):[a,b]\to[c,d]$ s.t $f_n\to f$ uniformly on [a,b]. Let $g:[c,d]\to\mathbb R$ continious function. Prove that $g(f_n(x))\to g(f(x))$

Let $\epsilon>0$. If $f_n\to f$, $\exists n_0\in\mathbb N$ such that for all $n>n_0$ $|f_n(x)-f(x)|<\epsilon$. Since g is continious on [c,d], exists $\delta>0$ s.t if $|x-x_0|<\delta$ then $|g(x)-g(x_0)|<\epsilon$. In fact $\forall x\in [a,b]: f_n(x),f(x)\in[c,d]$ so if $n>n_0$ and $|f_n(x)-f(x)|<\delta$ then $|g(f_n(x))-g(f(x))|<\epsilon$. I'm sure something is really wrong with this proof (except it seems I proved it only for specific environment of $x_0$). What is it and how can my proof can be improved?

share|improve this question
What makes you sure that something is wrong? Would you mind splitting up the logic a little to make it more readable? I don't think the $|x-x_0| < \delta$ part is worded correctly; you should say that "for every $\epsilon > 0$ $\exists \delta > 0$ such that...". –  abiessu Aug 24 '13 at 17:55
Your title says that you want $g\circ f_n$ to converge uniformly, but you omitted this from the text. Note that your proof uses the fact that a continuous function on a closed interval is uniformly continuous. It is not specific to any $x_0$ when you realize that. –  Ted Shifrin Aug 24 '13 at 17:58

1 Answer 1

up vote 3 down vote accepted

Since $f_n\to f$ uniformly, there exists, given $\eta >0$; $N=N_\eta$ such that $$|f_n(x)-f(x)|<\eta $$ whenever $n\geqslant N$, $x\in [a,b]$.

Since $g:[c,d]\to\Bbb R$ is continuous, it is uniformly continuous, whence given $\varepsilon >0$ there exists $\eta=\eta_\varepsilon >0$ such that for any $x,x'\in [c,d]$ with $|x-x'|<\eta$ we have $$|g(x)-g(x')|<\varepsilon$$

Combine the above to find, given $\varepsilon >0$; an $N=N_\varepsilon$ such that $$|g\circ f_n(x)-g\circ f(x)|<\varepsilon$$ whenever $n\geqslant N$.

share|improve this answer
I'm assuming you meant $x, x' \in [c,d]$. –  abiessu Aug 24 '13 at 18:00
why the composition is less then epsilon? –  Danis Fischer Aug 24 '13 at 18:14
@CoarguAliquis That is for you to explain! Try to write out the proof. –  Pedro Tamaroff Aug 24 '13 at 18:16

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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