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

If $F(t)$ is twice differentiable at $x$ and $$G(h)=\max_{t\in(0,h)}\left[\frac{F'(x+t)-F'(x-t)}{2t}-F''(x)\right],$$ where $x$ is fixed; then how can we show that $\displaystyle\lim_{h\to 0}G(h)=0$.

share|cite|improve this question
There are abstract duplicates of this question here and here. – Zev Chonoles Jun 9 '12 at 15:48
up vote 1 down vote accepted

Hint: $$\frac{F'(x+t)-F'(x-t)}{t}=\frac{F'(x+t)-F'(x)}{t}+\frac{F'(x-t)-F'(x)}{-t}.$$

share|cite|improve this answer
@Thomas: Indeed very useful! I started with the obvious one, then tried to give a little more. – André Nicolas Jun 9 '12 at 15:57
You removed the $t$, I removed my comment. A pity, this equation would have allowed me to show (no, I won't tell ;-) – user20266 Jun 9 '12 at 15:59
@Kns: I was being too indirect. Have changed the hint, added primes. Note that each part is related to the derivative of $F'$. – André Nicolas Jun 9 '12 at 16:05
Thanks a lot Andre! – Kns Jun 9 '12 at 16:17

Hint: By definition

$$ F''(x) = \lim_{h\to 0} \frac{F'(x+h)-F'(x)}{h}$$

Couple this with Andre's comment

share|cite|improve this answer

Since $\,t\in(,h)\,$ , we have that $\,h\to 0\Longrightarrow t\to 0\,$ , so: $$\lim_{t\to 0}\frac{F'(x+t)-F'(x-t)}{2t}=\lim_{t\to 0}\frac{1}{2}\left[\frac{F'(x+t)-F'(x)}{t}+\frac{F'(x-t)-F'(x)}{-t}\right]$$ and you get what you want since we know $\,F''(x)\,$ exists, so the limit defining this second derivative exists.

share|cite|improve this answer

Since $h\to0$ means $t\to0$, so $$\displaystyle\lim_{h\to 0}G(h) = \lim_{t\to 0}\left[\frac{F'(x+t)-F'(x-t)}{2t}-F''(x)\right]$$ apply l'Hôpital's rule,we can get $$ \lim_{t\to 0}\left[\frac{F'(x+t)-F'(x-t)}{2t}-F''(x)\right] = \lim_{t \to 0}\left[\frac{F''(x+t)+F''(x-t)}{2}-F''(x)\right] = 0$$

ps: your tags include real analysis, I assume $x\in \mathbf{R^{n}}$, although real analysis isn't only about real numbers. I dou't know if l'Hôpital's rule can apply under other situation.

share|cite|improve this answer
(This was my first intent, too: L'Hospital) I've a doubt here: in the RHS the $\,t's\,$ disappeared while you still are taking the limit when $\,t\to 0\,$ , but this is equivalent to plug $\,t=0\,$ after applying L'Hospital, which is justified if $\,F''\,$ is continuous in $\,x\,$, something we can't know... – DonAntonio Jun 9 '12 at 16:27
If I recall correctly, L'Hospital does not require the derivative to be continuous(namely $F''$ here), the existent of the derivative in the area is enough – haohaolee Jun 9 '12 at 17:05
Indeed it doesn't, but you require it to put $$\lim_{t\to 0}\frac{F''(x+t)+F''(x-t)}{2}=\frac{F''(x)+F''(x)}{2}$$which is what in fact you did, or, of course, justify otherwise this equality. – DonAntonio Jun 9 '12 at 17:13
oh, you are right, my fault, thanks. edit it. And then I get what you meant.... thinking now – haohaolee Jun 9 '12 at 17:18
seems it is not appropriate here. Thanks for the reminding – haohaolee Jun 9 '12 at 17:28

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.