Can you help me evaluating this integral? (no symbolic software please).

\begin{equation} \int \frac{C \; d \theta}{ \sin \theta \sqrt{\sin^2 \theta - C }} \end{equation}

It can be reduced to an algebraic form \begin{equation} \int \frac{dx}{ x \sqrt{ a + b x^2 + c x^4}} \end{equation} How about evaluating the second integral?


Actually I like very much the substitutions suggested by @juanheron. However he rushed a bit at the end and introduced errors on the result. In addition he left the work on half.

Let me start where Juan introduced the error:

Let us make the substitution $v^2 = k u^2 -1 $, so $ v dv = k \, u \, du$, and $u^2 -1 = (v^2 +1)/k -1 = (v^2 +1 - k)/k$, so \begin{equation} I = C \int \frac{v dv/k}{(v^2+1-k)/k) v} = C \int \frac{dv}{(v^2+1-k)} = C \int \frac{dv}{(v^2+C)} \end{equation} That is, \begin{equation} I = \int \frac{dv}{(v/\sqrt{C})^2 + 1}. \end{equation} A final substitution of $v/\sqrt{C}=z$, $dv = \sqrt{C} dz$, yields \begin{equation} I = \int \frac{\sqrt{C} dz}{ z^2 + 1} = \sqrt{C} \tan^{-1} z + C' \end{equation}

Now comes some algebra work. We need to do back-substitution to get the expression in terms of the original variable $\theta$.

Here is the back substitution:

\begin{eqnarray} I &=& \int \frac{\sqrt{C} dz}{ z^2 + 1} \\ &=& \sqrt{C} \tan^{-1} z + C' \\ &=& \sqrt{C} \tan^{-1} (v/\sqrt{C}) + C' \\ &=& \sqrt{C} \tan^{-1} (\sqrt{(k u^2 -1)/C}) + C' \\ &=& \sqrt{C} \tan^{-1} (\sqrt{k/t^2 -1/C}) + C' \\ &=& \sqrt{C} \tan^{-1} (\sqrt{[(1-C)/\cos^2 \theta -1]/C}) + C' \\ &=& \sqrt{C} \tan^{-1} \left ( \sqrt{ (\sin^2 \theta - C)/\cos^2 \theta /C} \right ) + C' \\ &=& \sqrt{C} \tan^{-1} \left ( \frac{\sqrt{\sin^2 \theta - C}}{\sqrt{C} \cos \theta} \right ) + C' \end{eqnarray}


Hint One can handle the second form of the integral by writing it as $$\int \frac{x \,dx}{x^2 \sqrt{a + b x^2 + c x^4}}$$ and substituting $$u = x^2, \quad du = 2 x \,dx ,$$ to write the integral as $$\int \frac{du}{u \sqrt{a + b u + c u^2}} .$$ This latter integral is a standard sort of exercise and can be handled with trigonometric (or hyperbolic trigonometric) substitution. The appropriate substitution, as well as the qualitative behavior of the antiderivative, depends on the sign of the discriminant $b^2 - 4 c a$ of the quadratic expression in the radical. (See this integrable table.)


Let $\displaystyle I = \int\frac{Cd\theta}{\sin \theta \sqrt{\sin^2 \theta - C}} = \int\frac{C \sin \theta}{\sin^2 \theta \sqrt{\sin^2 \theta -C}}d\theta$

So we get $$I = \int\frac{C\sin \theta}{(1-\cos^2 \theta)\sqrt{1-\cos^2 \theta -C}}d\theta\;,$$

Now Put $\cos \theta = t\;,$ Then $\sin \theta d\theta = -dt$

so we get $$I = -\int\frac{C}{(1-t^2)\sqrt{\underbrace{(1-C)}_{k}-t^2}}dt = -C\int\frac{1}{(1-t^2)\sqrt{k-t^2}}dt$$

Now Put $\displaystyle t = \frac{1}{u}\;,$ Then $\displaystyle dt = -\frac{1}{u^2}du$

So we get $$I = C\int\frac{u}{(u^2-1)\sqrt{ku^2-1}}du$$

Now at last Put $(ku^2-1)=v^2\;,$ Then $\displaystyle udu = \frac{1}{k}vdv$

So we get $$I = \frac{C}{k}\int\frac{k}{v^2+1}dv = C\int\frac{1}{v^2+1}dv= C\tan^{-1}(v)+\mathcal{C'}$$


$$\int \frac{C \; d \theta}{ \sin \theta \sqrt{\sin^2 \theta - C }}$$ $$=\int \frac{C \; \csc ^2 \theta d \theta}{ \csc ^2 \theta \cdot \sin \theta \sqrt{\sin^2 \theta - C }}$$ $$=\int \frac{C \; \csc ^2 \theta d \theta}{ \sqrt{1 - C \csc^2 \theta}}$$ $$=\int \frac{C \; \csc ^2 \theta d \theta}{ \sqrt{1 - C (1+\cot^2 \theta)}}$$ Use $\sqrt{C} \cot \theta = z$ Then $-\sqrt{C} \; \csc ^2 \theta d \theta = dz$ $$=\int \frac{-\sqrt{C} dz}{ \sqrt{1 - C - z^2}}$$ Say $1-C=a^2$ $$=\int \frac{-\sqrt{C} dz}{ \sqrt{a^2 - z^2}}$$ $$=-\sqrt{C}\sin^{-1}\left(\frac{z}{a}\right)+k$$ $$=-\sqrt{C}\sin^{-1}\left(\frac{\sqrt{C} \cot \theta}{a}\right)+k$$

  • $\begingroup$ Why the downvote??? $\endgroup$ – SchrodingersCat Nov 4 '15 at 14:04
  • $\begingroup$ I find your reasoning spotless. I would only change your TeX commands so that trigonometric functions will not show in italics. This is not the place for discussion but I find that some people are mean on these forums. They shoot and do not show their faces :( $\endgroup$ – Herman Jaramillo Nov 4 '15 at 17:39
  • $\begingroup$ Thanks for your comment... :-) $\endgroup$ – SchrodingersCat Nov 4 '15 at 17:47
  • $\begingroup$ I believe you have an error in your equation. Please look at the expression $\int -\sqrt{C} dz/\sqrt{1 - C + z^2}$, just before the line with "Say $1-C=a^2$. The variable $z^2$ should have a $C$ in front. Right? $\endgroup$ – Herman Jaramillo Nov 5 '15 at 22:00
  • 2
    $\begingroup$ Yes you were right. Thanks for pointing out the mistake. I have updated my answer correctly. $\endgroup$ – SchrodingersCat Nov 6 '15 at 16:44

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.