Each of the three following definite integrals are well known to have the same value of $\pi$: $$\int_{-1}^1\frac{1}{\sqrt{1-x^2}}\,\mathrm{d}x=2\int_{-1}^1\sqrt{1-x^2}\,\mathrm{d}x=\int_{-\infty}^{\infty}\frac{1}{1+x^2}\,\mathrm{d}x=\pi.$$

I like taking the first definite integral as the definition of $\pi$, since it represents half the circumference of the unit circle. The second integral obviously represents the area of the unit circle, but as an exercise I wanted to prove its equality with integral #1 using just elementary integration rules. I was successful once I tried integrating by parts:

$$\begin{align} \int\sqrt{1-x^2}\,\mathrm{d}x &=x\sqrt{1-x^2}-\int\frac{-x^2}{\sqrt{1-x^2}}\,\mathrm{d}x\\ &=x\sqrt{1-x^2}+\int\frac{1}{\sqrt{1-x^2}}\,\mathrm{d}x-\int\frac{1-x^2}{\sqrt{1-x^2}}\,\mathrm{d}x\\ &=x\sqrt{1-x^2}+\int\frac{1}{\sqrt{1-x^2}}\,\mathrm{d}x-\int\sqrt{1-x^2}\,\mathrm{d}x\\ \implies2\int\sqrt{1-x^2}\,\mathrm{d}x&=x\sqrt{1-x^2}+\int\frac{1}{\sqrt{1-x^2}}\,\mathrm{d}x\\ \implies2\int_{-1}^1\sqrt{1-x^2}\,\mathrm{d}x&=\int_{-1}^1\frac{1}{\sqrt{1-x^2}}\,\mathrm{d}x. \end{align}$$

Having accomplished this much, I decided I'd like to demonstrate the equality of these two integrals to $\int_{-\infty}^{\infty}\frac{1}{1+x^2}\mathrm{d}x$ as well, in a similarly elementary manner, but I'm stumped as to what to try. Can anyone suggest a substitution or transformation that demonstrates their equality?

  • $\begingroup$ I am not sure you can equate the last one to either of the first two since the bounds of integration are different. $\endgroup$ – Laars Helenius Dec 18 '13 at 20:56
  • $\begingroup$ On your note about complex analysis: this integral, if computed rigorously (of course, $\text{atan}(x) \rightarrow \pi/2$ as $x \rightarrow \infty$), is usually done with contour integration. So this would suggest complex analysis as a common approach. $\endgroup$ – Chris K Dec 18 '13 at 20:57
  • $\begingroup$ see Chebyshev polynomials:Orthogonality... $\endgroup$ – draks ... Dec 18 '13 at 21:15
  • $\begingroup$ @LaarsHelenius That can be avoided if the third one is replaced by $$\int_{-1}^1 \frac{2dx}{1+x^2} = \pi$$ $\endgroup$ – Jaume Oliver Lafont Apr 19 '16 at 8:42

These are both immediate using two subsitutions. Everything is even, so split all of them at $0$.

$$\begin{aligned} t=\sqrt{1-x^2}:\quad\int_0^1 \frac{dx}{\sqrt{1-x^2}}= \int_0^1 2\sqrt{1-t^2}\,dt\end{aligned}$$


$$\begin{aligned} t=\frac{x}{\sqrt{1-x^2}}:\quad\int_0^{1} \frac{dx}{\sqrt{1-x^2}}= \int_0^{\infty}\frac{dt}{1+t^2}\end{aligned}$$

  • 1
    $\begingroup$ Whoa! Way better than my answer. Thanks! $\endgroup$ – John Hughes Dec 18 '13 at 21:25
  • $\begingroup$ Awesome. This is precisely the kind of answer I was looking for! Much appreciated. $\endgroup$ – David H Dec 18 '13 at 21:40
  • $\begingroup$ I can't see how these two substitutions would satisfy these two equations. Could someone clarify it a bit further? $\endgroup$ – Albert Netymk Mar 14 '18 at 23:24
  • $\begingroup$ @AlbertNetymk What do you mean? Applying those substitutions takes you from one integral to the other. Try it! $\endgroup$ – L. F. Jul 20 '18 at 15:54

Let $x = \sin \theta, \int_{-1}^1 \frac{1}{\sqrt{1-x^2}} dx = 2\int_{0}^1 \frac{1}{\sqrt{1-x^2}} dx = 2 \int_0^{\pi/2} \frac{1}{\cos\theta} d\sin\theta =\pi$

$\int_{-\infty}^{\infty}\frac{1}{1+x^2}dx = \int_{-\infty}^{\infty} d \left(\tan^{-1}x \right) =\int_{-\pi/2}^{\pi/2} dy = \pi, where \quad y=\tan^{-1}x$

  • 4
    $\begingroup$ I think the OP's goal was to equate the two integrals without evaluating them explicitly. $\endgroup$ – Chris K Dec 18 '13 at 21:08
  • $\begingroup$ @ChrisK Indeed, as I would tend to take $\int_0^x\frac{1}{1+x^2}dx$ as the definition of $\arctan$, which would render this answer circular. Perhaps I should have made my question more explicit. $\endgroup$ – David H Dec 18 '13 at 21:12

If in the third integral you let $x=\tan(t)$, you get

$$ I_3 = \int_{-\pi/2}^{\pi/2} 1 dt. $$

If, in the second integral, you do the substitution $x=\sin(u)$, you get $$ I_2 = 2\int_{-\pi/2}^{\pi/2} \cos^2 t ~ dt. $$

Alternatively, substitute $x = \cos(u)$ to get $$ I_2 = 2\int_{-\pi/2}^{\pi/2} \sin^2 t ~ dt. $$

So combining these last two,

$$ I_2 = \frac{1}{2} \left (2\int_{-\pi/2}^{\pi/2} \sin^2 t ~ dt + 2\int_{-\pi/2}^{\pi/2} \sin^2 t ~ dt \right) \\ = \frac{1}{2} \left (2\int_{-\pi/2}^{\pi/2} \sin^2 t + \cos^2 t~ dt \right) \\ = \int_{-\pi/2}^{\pi/2} 1~ dt $$

hence the second and third integrals are equal, as you wanted.

Not completely satisfactory, but a shift from integrating from $-1$ to $1$ to integrating from $-\infty$ to $\infty$ is likely to involve something like a tangent substitution in general, so it's not too surprising to see this route work out.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.