# A problem on limit involving various functions

Find the value of $$\lim \limits_{x\to 0} \frac{\tan\sqrtx\ln(1+3x)}{(\tan^{-1}\sqrt{x})^2(e^{5\large \sqrtx}-1)}$$ Applying L'Hospital's rule does not seem to simplify the expression.

• Have you tried Taylor expansion? – Git Gud Nov 2 '13 at 15:19
• Apply power series for each term in the expression. I got a similar problem in my maths olympiad. – GTX OC Nov 2 '13 at 15:34

Note that $$\lim \limits_{y\to 0}\frac{\tan y}{y}=\lim_{y\to 0}\frac{\tan^{-1} y}{y}=\lim_{y\to 0}\frac{\ln (1+y)}{y}=\lim_{y\to 0}\frac{e^y-1}{y}=1.$$ Therefore,

$$\lim \limits_{x\to 0^+} \frac{\tan\sqrtx\cdot \ln(1+3x)}{\left(\tan^{-1}\sqrt{x}\right)^2\cdot(e^{5 \sqrtx}-1)}=\lim_{x\to 0^+}\frac{\sqrtx\cdot 3x}{(\sqrt{x})^2\cdot 5 \sqrt{x}}=\frac{3}{5}.$$

Edit: Some comments below suggest that the answer above may not be detailed enough. Let me explain a little more about the gap between the two lines of equations above. The limit we are concerned about can be written as

$$\lim \limits_{x\to 0^+}\left(\frac{\tan\sqrtx}{\sqrtx}\cdot \frac{\ln(1+3x)}{3x}\cdot \left(\frac{\sqrt{x}}{\tan^{-1}\sqrt{x}}\right)^2\cdot\frac{5 \sqrtx}{e^{5 \sqrtx}-1}\cdot \frac{\sqrtx\cdot 3x}{(\sqrt{x})^2\cdot 5 \sqrt{x}}\right).$$ On the one hand, as $x\to 0^+$, $\sqrtx$, $3x$ $\sqrt x$ and $5\sqrtx$ all approach to $0$, so from the first displayed line we know that each limit of first four terms in the product above exists and equals $1$. On the other hand, the limit of the last term exists and equals $\dfrac{3}{5}$. As a result, the original limit exists and equals $\dfrac{3}{5}$.

• I don't see how you get the first equality in the second line from the first line. – Git Gud Nov 2 '13 at 15:36
• @GitGud: Substitute $y=\sqrt{x}$, $y=3x$, $y=\sqrt{x}$ and $y=5\sqrt{x}$ respectively. – user104254 Nov 2 '13 at 15:39
• I thought you were using taylor series implicitly, I just realised you aren't. (+1) – Git Gud Nov 2 '13 at 15:43
• It seems that behaviour of functions of arguments of $x$ is similar to the behaviour of arguments of $x$, as $x$ tends to zero, after observing the second line. Is it rigorous? – Tejas Nov 2 '13 at 15:48
• @TBrendle I think you're incurring on the falacy of not looking at $=$ as symmetric relation, but only in one direction. For instance $\lim \limits_{x\to 0}\left(\dfrac{\sin(x)}{x}\dfrac{1}{\cos(x)}\right)=\lim \limits_{x\to 0}\left(\dfrac{\sin (x)}{x}\right)\lim \limits_{x\to 0}\left(\dfrac 1{\cos(x)}\right)=1\cdot 1=1$ is perfectly rigorous simply because you can read the equalities backwards and symmetry of $=$ takes care it. – Git Gud Nov 2 '13 at 16:24

If you change variable (x = y^3) and expand as a Taylor series at y=0, the first terms are 3/5 - 3 y / 2 + 29 y^2 / 20 +...