Question: Compute $$\int_0^1 \frac{\sqrt{x-x^2}}{x+2}dx.$$

Attempt: I've tried various substitutions with no success. Looked for a possible contour integration by converting this into a rational function of $\sin\theta$ and $\cos \theta$.

  • $\begingroup$ Through a number of substitutions I've reduced the problem to computing $\int_0^1 \frac{\cos^2(\theta)}{\sin(\theta) + 5} d\theta$, but I'm not quite sure where to go from there. $\endgroup$ – GiantTortoise1729 Jun 26 '15 at 21:39
  • $\begingroup$ @GiantTortoise1729: From there, the usual tangent half angle substitution and partial fractions will work. There might be a better way. Also, your bounds for $\theta$ are probably wrong. $\endgroup$ – JimmyK4542 Jun 26 '15 at 21:41
  • $\begingroup$ Oh yes, sorry, I meant to write $x=1$, $x=0$ $\endgroup$ – GiantTortoise1729 Jun 26 '15 at 21:46

We certainly can use contour integration here. First, through a sub of $x \mapsto x^2$ and a little algebra, we can express the integral as

$$2 \int_0^1 dx \, \sqrt{1-x^2} - 4 \int_0^1 dx \frac{\sqrt{1-x^2}}{2+x^2} = \frac{\pi}{2} - 2 \int_{-1}^1 dx \frac{\sqrt{1-x^2}}{2+x^2}$$

Now consider

$$\oint_C dz \frac{\sqrt{z^2-1}}{z^2+2} $$

where $C$ is (1) the circle $z=R e^{i \theta}$, $\theta \in [-\pi,\pi)$, (2) a line extending from the circle at $\theta=\pi$ to the dogbone contour, (3) the dogbone, and (4) a line back to the circle at $\theta=-\pi$. Note that the integral vanishes along the lines to the dogbone and along the small circles of the dogbone.

Thus, the contour integral is equal to

$$i R \int_{-\pi}^{\pi} d\theta \, e^{i \theta} \frac{\sqrt{R^2 e^{i 2 \theta}-1}}{R^2 e^{i 2 \theta}+2} + i 2 \int_{-1}^1 dx \frac{\sqrt{1-x^2}}{2+x^2}$$

Note that I have ignored the small semicircular and circular contours, as the integrals around those vanish as their radii vanish. Also, I have skipped over the step of assigning a phase to each branch of the square root along $[-1,1]$.

The contour integral is also equal to $i 2 \pi$ times the residue of the poles inside $C$, which are at $z=\sqrt{2} e^{\pm i \pi/2}$. Thus, taking the limit as $R \to \infty$, we get

$$i 2 \pi + i 2 \int_{-1}^1 dx \frac{\sqrt{1-x^2}}{2+x^2} = i 2 \pi \left (\frac{i \sqrt{3}}{i \sqrt{2}} + \frac{-i \sqrt{3}}{-i \sqrt{2}} \right ) = i \sqrt{6} \pi $$

Putting this altogether, we get

$$\int_{-1}^1 dx \frac{\sqrt{1-x^2}}{2+x^2} = \left (\sqrt{6}-2 \right ) \frac{\pi}{2} $$

and therefore, the original integral is equal to

$$\int_0^1 dx \frac{\sqrt{x-x^2}}{x+2} = \left (\frac{5}{2}-\sqrt{6} \right ) \pi $$

  • $\begingroup$ Good answer! Out of curiosity how did you get the inspiration for the substitution that moved the singularity to the complex plane? $\endgroup$ – Zach466920 Jun 26 '15 at 22:44
  • 3
    $\begingroup$ @Zach466920: anytime I see $\sqrt{x}$ in an integrand, I try to sub $x \mapsto x^2$. $\endgroup$ – Ron Gordon Jun 26 '15 at 23:16
  • $\begingroup$ Given this result, seems like a reasonable thing to do. Thanks for the response :) $\endgroup$ – Zach466920 Jun 26 '15 at 23:25

Letting $x=u^2, dx=2udu$ gives $\displaystyle2\int_0^1\frac{u^2\sqrt{1-u^2}}{u^2+2}du$; then letting $u=\sin\theta, du=\cos\theta d\theta$ gives

$\displaystyle2\int_0^{\pi/2}\frac{\sin^2\theta\cos^2\theta}{\sin^2\theta+2}d\theta=2\int_0^{\pi/2}\frac{\tan^2\theta\sec^2\theta}{(\sec^4\theta)(3\tan^2\theta+2)}d\theta=2\int_0^{\infty}\frac{t^2}{(t^2+1)^2(3t^2+2)}dt$ [$t=\tan\theta$]




Substituting $x=\frac{1+\cos(\theta)}2$ and $z=e^{i\theta}$, we can compute this integral as a contour integral around the unit circle: $$ \begin{align} \int_0^1\frac{\sqrt{x-x^2}}{x+2}\,\mathrm{d}x &=\frac12\int_0^\pi\frac{\sin^2(\theta)}{5+\cos(\theta)}\,\mathrm{d}\theta\\ &=\frac14\int_0^{2\pi}\frac{\sin^2(\theta)}{5+\cos(\theta)}\,\mathrm{d}\theta\\ &=-\frac1{8i}\oint\frac{z^2-2+\frac1{z^2}}{z+10+\frac1z}\frac{\mathrm{d}z}z\\ &=-\frac1{8i}\oint\frac{z^4-2z^2+1}{z^2+10z+1}\frac{\mathrm{d}z}{z^2}\\[3pt] &=-\frac1{8i}\oint\left(\color{#0000F0}{1+\frac1{z^2}}\color{#00A000}{-\frac{10}z+\frac{4\sqrt6}{z+5-\sqrt{24}}}\color{#C00000}{-\frac{4\sqrt6}{z+5+\sqrt{24}}}\right)\,\mathrm{d}z\\[6pt] &=-\frac\pi4\left(\color{#00A000}{-10+4\sqrt6}\right)\\[6pt] &=\pi\,\frac{5-2\sqrt6}2 \end{align} $$ The terms in blue have no residue and the pole of the term in red is outside the unit circle; therefore, we need only consider the terms in green.

  • $\begingroup$ That's the approach i was initially hoping for- any trick for coming up with the right substitution? $\endgroup$ – Ashley Jun 27 '15 at 19:34
  • 1
    $\begingroup$ @Hamton: I completed the square with $\sqrt{x-x^2}=\frac12\sqrt{1-(2x-1)^2}$, which lead to $\cos(\theta)=2x-1$. Then, $z=e^{i\theta}$ is a standard substitution when integrating along the unit circle in $\mathbb{C}$. $\endgroup$ – robjohn Jun 27 '15 at 23:52

Let us make the problem a bit more general considering $$I=\int \frac{\sqrt{x-x^2}}{x+a}dx$$ and let us apply the same approach Mario G proposed. We then arrive to $$\begin{align*} I=\int\frac{\sqrt{x-x^2}}{x+2}dx &=\frac{1}{2}\int\frac{\cos^2( t)}{\sin t+(2a+1)}\,dt\\ \end{align*}$$ Now, let use the tangent half-angle substitution $y=\tan(\frac t2)$

$$I=\int\frac{ \left(y^2-1\right)^2}{\left(y^2+1\right)^2 \left((2 a+1) y^2+2 y+(2 a+1)\right)}\,dy$$ Now, partial fraction decomposition which gives for the integrand $$\frac{2 a+1}{y^2+1}-\frac{2 y}{\left(y^2+1\right)^2}-\frac{4 a\left(a+1\right)}{(2 a+1)y^2+2y+(2 a+1) }$$

Integrating the different species, we then arrive to $$I=(2 a+1) \tan ^{-1}(y)+\frac{1}{y^2+1}-2 \sqrt{a(a+1)} \tan ^{-1}\left(\frac{(2 a+1) y+1}{2 \sqrt{a(a+1)}}\right)$$


Let $x=\frac{1}{2}\sin t+\frac{1}{2}$, then we have \begin{align*} \int_0^1\frac{\sqrt{x-x^2}}{x+2}dx&=\int_0^1\frac{\sqrt{\left(\frac{1}{2}\right)^2-\left(x-\frac{1}{2}\right)^2}}{x+2}dx\\ &=\int_{-\frac{\pi}{2}}^{\frac{\pi}{2}}\frac{\sqrt{\left(\frac{1}{2}\right)^2-\left(\frac{1}{2}\sin t\right)^2}}{\frac{1}{2}\sin t+\frac{1}{2}+2}\frac{1}{2}\cos t\,dt\\ &=\frac{1}{2}\int_{-\frac{\pi}{2}}^{\frac{\pi}{2}}\frac{\cos t}{\sin t+5}\cos t\,dt\\ \end{align*}

  • $\begingroup$ Is that last integral simpler? $\endgroup$ – Ashley Jun 26 '15 at 22:52
  • $\begingroup$ @Hamton tangent half angle substitution, yields and evaluable form. $\endgroup$ – Zach466920 Jun 26 '15 at 22:58
  • $\begingroup$ ah okay, thanks $\endgroup$ – Ashley Jun 27 '15 at 19:34




Using the generating function for the Catalan numbers, this equals



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.