Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

f is twice differentiable on $(0,\infty)$ and $f''$ bounded on $(0,\infty)$, $f(x)\rightarrow 0$ as $x\rightarrow\infty$. show $f'(x)\rightarrow 0$ as $x\rightarrow\infty$.

We know $\exists M >0 $ s.t. $|f''(x)|<M\space\space\forall x\in (0,\infty)$. By MVT,

$$|f'(x+h)-f(x)|=|h||f''(z)|\le \frac{\delta}{2M}M=\delta/2$$ $\forall |x+h-x|\le\frac{\delta}{2M}$

$$\left|f'(x+h)-\lim_{c\rightarrow\infty}\frac{f(x+c)-f(x)}{c}\right|\le \frac{\delta}{2}$$

Since $f(x)\rightarrow 0$ as $x\rightarrow\infty$, then if we fix $\delta >0$, we have

$$|f'(x+h)|\le \frac{\delta}{2}$$

So $f'(x)\rightarrow 0$ as $x\rightarrow\infty$. Is this right?

share|cite|improve this question
In your argument, it seems that you used $\lim_{c\to \infty} \frac{f(x+ c) - f(x)}{c} = f'(x)$, which is not correct (should be $c\to 0)$. – Arctic Char Nov 19 '13 at 6:47
Thanks, yes that's what I meant. would the proof be correct then? – lightfish Nov 19 '13 at 12:32
up vote 2 down vote accepted

By Taylor Theorem, we have $$ f(x+h)-f(x)-f'(x)h=f''(z)h^2/2. $$ for $z\in(x,x+h)$. The right hand side is bounded independently of $x$, thus, the bound for $f'$ for large $x$ only depends on $h$, which is an arbitrary number and we conclude.

share|cite|improve this answer
Your idea is correct, but might be you can elaborate more as you need to choose $h$ a bit more carefully depending on $x$. – Arctic Char Nov 19 '13 at 6:54
$$ ||f(x+h)-f(x)|-|f'(x)|h|\leq \frac{\max |f''|}{2}h^2, $$ Thus $$ \lim_{x\rightarrow\infty}|f'(x)|h\leq \frac{\max |f''|}{2}h^2, $$ and we conclude. I don't see why $h$ depends on $x$. Maybe I'm doing something wrong. – guacho Nov 19 '13 at 18:16
You are absolute correct, but I will choose $M$ such that $|f(x|\leq \epsilon^2$ for all $x>M$ and get something like $$|f'(x)| \epsilon \leq C\epsilon^2\ .$$ somehow in your prove its seems that you assume limit $f'(x)$ exists already. – Arctic Char Nov 19 '13 at 20:37
Oooooooook, you are right. Change $\lim$ by $\limsup$ and everything works, right? – guacho Nov 19 '13 at 21:06

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.