Reading Landau and Lifshitz "Quantum Mechanics. Non-relativistic theory", I've come across an identity, which after being a bit simplified, reads


and apparently holds for pure imaginary $a$ and $b$. The book says the left hand side can be calculated to equal right hand side using the reflection formula:

$$\Gamma(x)\Gamma(1-x)=\frac{\pi}{\sin\pi x}.\tag2$$

But I fail to see how I could actually do this calculation. Plotting the difference I see that this actually is false for real $a$ and $b$, so the reflection formula, which AFAIK works for all $x\in\mathbb C$ seems to be not enough to get this result.

So, how do I prove $(1)$, or even better, derive RHS from LHS?


The identity holds for pure imaginary $a$ and $b$. So, let $a=i u$ and $b=i v$ where $u$ and $v$ are real. The identity to be proved is : $$\left|\frac{(i(u+v)\Gamma(2iu)\Gamma(-i(u+v))^2}{i(u-v)\Gamma(-2iu)\Gamma(i(u-v))^2}\right|=\left|\frac{\sin(\pi i(u-v))}{\sin(\pi i(u+v))}\right|$$

At first look, I don't see any interest to use the reflexion relationship. It is much easier to use the formula giving the modulus : $$|\Gamma{(iy)}|=\sqrt{\frac{\pi}{y \sinh(\pi y)}}$$ With this formula, remplace $y$ by $(u+v)$ or $(u-v)$ and put it back into the left term above. The simplification leads to $\frac{\sinh(\pi(u-v))}{\sinh(\pi(u+v))}$

With $\sinh(y)=-i\sin(i y)$ we have: $\frac{\sinh(\pi(u-v))}{\sinh(\pi(u+v))}=\frac{\sin(\pi i(u-v))}{\sin(\pi i(u+v))}$ which is the right term of the equation and so, it is proved.

  • $\begingroup$ Wow! In fact the original variables were indeed sum and difference multiplied by $i$, it was me who changed them into $a$ and $b$ trying to work out the result. $\endgroup$ – Ruslan Nov 18 '14 at 19:19
  • 1
    $\begingroup$ You seem to have lost a minus in $\sinh(y)=-i\sin(iy)$. $\endgroup$ – Ruslan Nov 18 '14 at 19:47
  • $\begingroup$ Of course, you are right. I corrected the typing mistake in my answer. Thanks for the point. $\endgroup$ – JJacquelin Nov 18 '14 at 20:03
  • $\begingroup$ @Ruslan: you should probably mention this change in the question. People attempting to answer the question without reading your comment to this answer might spend a lot of time trying to prove something that is not true. $\endgroup$ – robjohn Nov 18 '14 at 20:23
  • $\begingroup$ @robjohn no, the statement doesn't become false — it's just change of variables. $\endgroup$ – Ruslan Nov 19 '14 at 4:02

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.