3

According to Mathematica,

$$\displaystyle\lim_{x \rightarrow 0} \frac{(1-\cos x)^2}{\log (1 + \sin^4x)} = \frac{1}{4}$$

For my purposes it is sufficient to know this limit exists and is finite, but I may not use L'Hopital's rule (or anything relying on differentiability). However, I am stuck on how to show this limit holds. I've tried giving an epsilon-delta proof, but it seems needlessly complicated. Ideally I should be able to show the limit exists using only trig rules, algebra, and eventually reducing this out of being an indeterminate form so that I can simply plug in $0$.

However the problem with this is that I can't seem to break the log term. It seems like I would need to factor its argument, but to do so I need to manipulate $1 + \sin^4 x$ so that I can factor it and separate out the terms. But nothing has worked thus far.

StubbornAtom
  • 17,052
Zach Halle
  • 163
  • 4

3 Answers3

4

$$ \lim_{x \rightarrow 0} \frac{(1-\cos x)^2}{\log (1 + \sin^4x)} = \lim_{x \rightarrow 0} \frac{(2\sin^2( x/2))^2)}{\log (1 + \sin^4x)}=1\cdot\lim_{x \rightarrow 0} \frac{4\sin^4(x/2))}{\sin^4x} $$ and this behaves as $$ \frac{4x^4}{2^4x^4}, $$ which indeed tends to $1/4$.

We used $1-\cos x=(\sin^2(x/2)+\cos^2(x/2))-(\cos^2(x/2)-\sin^2(x/2))$ and $\displaystyle\lim_{y\to0}\dfrac{\ln(1+y)}{y}=1$.

Przemysław Scherwentke
  • 13,668
  • 5
  • 35
  • 56
  • Oh jeez, this was a typo on my part. The square term should be applied to the whole numerator. I'll edit my question to reflect this. – Zach Halle Nov 05 '14 at 03:40
  • Who downvoted an old version of answer (to old version of question)?! Now it is corrected to be an answer to new version of question. – Przemysław Scherwentke Nov 05 '14 at 03:50
  • Some people really rushed to downvote... @Prz , would you mind to explain your second equality in the first line? – Timbuc Nov 05 '14 at 03:50
  • @Timbuc Enlarged. – Przemysław Scherwentke Nov 05 '14 at 03:57
  • @PrzemysławScherwentke, so in fact you're using an approximation: $;\log(1+\sin^4x)\sim\sin^4x;$ for values close to zero, right? But, even if this approx. is done without Taylor polynomials/series (which require derivatives), how do you prove $;\frac{\log(1+y)}y\xrightarrow[y\to 0^+]{}1;$ without l'Hospital or something similar? – Timbuc Nov 05 '14 at 04:03
  • @Timbuc It is one of basic limits, taught before the beginning of differential calculus. It is for example a corollary from the limit $\lim_{x\to0}\frac{e^x-1}{x}=1$, which is obtained by elementary methods. – Przemysław Scherwentke Nov 05 '14 at 04:06
  • I'm just asking how that can be deduced without derivatives, @PrzemysławScherwentke. I'm not saying it can't be done. And for me l'Hospital and Taylor series is elementary methods, yet not for the OP. – Timbuc Nov 05 '14 at 04:08
  • @Timbuc A comment may be too short to answer. Please see http://math.stackexchange.com/questions/152605/proving-that-lim-limits-x-to-0-fracex-1x-1 (jon jones's [http://math.stackexchange.com/users/26706/jon-jones] answer). – Przemysław Scherwentke Nov 05 '14 at 04:16
2

HINT:

$$\frac{(1-\cos x)^2}{\ln(1+\sin^4x)}$$

$$=\frac{(1-\cos x)^2\cdot(1+\cos x)^2}{(1+\cos x)^2\cdot\ln(1+\sin^4x)}$$

$$=\frac1{\dfrac{\ln(1+\sin^4x)}{\sin^4x}}\cdot\frac1{(1+\cos x)^2}$$

Now $$\lim_{h\to0}\frac{\ln(1+h)}h=1$$

lab bhattacharjee
  • 274,582
2

I always rember that $1-\cos x \sim \frac{x^2}{2}$ and $\ln(1+x) \sim x$ as $x\to 0$. When I need to compute a limit, I always use it.

Paul
  • 20,553