For what values of $s\in\mathbb{R}$ does the following identity hold for all $\theta\in\mathbb{R}$:


for some $a,b\in\mathbb{R}$? In other words, for what values of $s$ can the expression $1-s\cos^2\theta\sin^2\theta$ be simplified to $a\sin^6\theta+b\cos^6\theta$ for some $a$ and $b$ not depending on $s$ or $\theta$? I know that $s=3$ holds, since we have the identity:


but I would like to know (purely out of interest) if there are other values of $s$ for which it holds. I would also like to know for what if any values of s the following identity holds:


for some $a,b\in\mathbb{R}$; and I would also like to know for what if any values of $s$ we have the following identity:


I am not aware of any theory that would help me solve trigonometric equations like $(1)$, $(2)$ and $(3)$ except by trial and error, but I would love to know if there is a method. I assume that there are only particular values of $s$ for which equations like these can hold, but I can not see any way of finding them.

The context of this question relates to a question I recently asked here, where I was attempting to evaluate integrals of the form $\int_0^\infty\left(a+\sin^2{\theta}\right)^{-\frac{1}{3}}\;d\theta$; there are complicated reasons as to which values of $a$ give nice values, but I noticed that the simple method I was using could possibly be extended for certain values of $a$ if I knew the solutions to equations $(1)$, $(2)$ and $(3)$ above. Thus I am wondering if anybody knows how to solve these equations


Since the equality should hold for all $\theta$, try first with $\theta=0$, that gives $b=1$. Then for $\theta=\pi/2$ we get $a=1$.

Let's try now with $\theta=\pi/4$: $$ 1-s\frac{1}{2}\frac{1}{2}=\frac{1}{8}+\frac{1}{8} $$ so $s=3$.

It should be easy to apply a similar method for showing the other two are not identities for any value of $s,a,b$.

The identity $1-3\cos^2\theta\sin^2\theta=\sin^6\theta+\cos^6\theta$ is indeed true, but the argument above doesn't prove it: we only got necessary conditions. However, it's easy to derive it as a consequence of $$ x^3+y^3=(x+y)^3-3xy(x+y) $$ with $x=\sin^2\theta$ and $y=\cos^2\theta$.

  • $\begingroup$ Oh I see. +1. I hadn't even thought of that. But could this be extended to actually proving the purported identity was correct? $\endgroup$ – Anon Jan 22 '17 at 23:57
  • 1
    $\begingroup$ @Anon The first identity can only hold for $s=3$, $a=1$ and $b=1$; but it has to be proved for these values. $\endgroup$ – egreg Jan 22 '17 at 23:59

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.