$\lim_{x\to \infty}e^{-x^2}\int_x^{x+1/x}e^{t^2}dt$
This is problem from Spivak calculus 1994, chapter 18, 31.
The book gives following solution: simply apply Lhopitatal rule, but what puzzles me is the way numerator is differentiated.
So they got $e^{(x+1/x)^2}-e^{x^2}\over 2xe^{x^2}$
But I think it should be $(1- 1/x^2) e^{(x+1/x)^2}-e^{x^2}\over 2xe^{x^2}$ . I think they are missing derivative of $x+1/x$. Am I wright or wrong?
P.S Could anyone please help me editing this question?
The function $\;e^{x^2}\;$ is continuous everywhere so it has a primitive function, say $\;F\;$ , and then
$$\int_x^{x+\frac1x}e^{t^2}dt=F\left(x+\frac1x\right)-F(x)\implies$$
$$\lim_{x\to\infty}e^{-x^2}\int_x^{x+\frac1x}e^{t^2}dt=\lim_{x\to\infty}\frac{F\left(x+\frac1x\right)-F(x)}{e^{x^2}}\stackrel{L'H}=\lim_{x\to\infty}\frac{\left(1-\frac1{x^2}\right)e^{\left(x+1/x\right)^2}-e^{x^2}}{2xe^{x^2}}=$$
$$=\lim_{x\to\infty}\frac{\left(\left(1-\frac1{x^2}\right)\left(e^2+e^{1/x^2}\right)\right)-1}{2x}=0\cdot\left((e^2+1)-1\right)\cdot1=0$$