MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

The following identity arose while I was working on a recent MO question:


I have no doubt that the identity is true, but I am not able to prove it. Can anyone help?

It is easy to prove by Taylor expansion that the left-hand-side of the identity can equivalently be written as $\sum_{n=1}^{\infty}\ln(1+x^n)$, which is the logarithm of the q-Pochhammer symbol $(-x,x)_{\infty}$, so an alternative way to pose my question is to ask for a proof of the series expansion


share|cite|improve this question
up vote 6 down vote accepted

First notice that $$\sum _{n=1} ^{\infty} \frac{x^n}{n(1-x^{2n})} = \sum _{r=0} ^{\infty} \sum _{m=1} ^{\infty}\left(\frac{1}{2^r}\sum _{k|2m-1} \frac{1}{k}\right)x^{2^r(2m-1)}.$$ And similarly $$-\sum _{n=1}^{\infty}\frac{(-x)^n}{n(1-x^n)} = \sum _{s=1}^{\infty} \left(\sum _{k|s}\frac{(-1)^{k+1}}{k}\right)x^s.$$ So we need to show that the respective coefficients match, i.e.: $$\frac{1}{2^r}\sum _{k|2m-1} \frac{1}{k}=\sum _{k|s}\frac{(-1)^{k+1}}{k},$$ for $s=2^r(2m-1)$. But this is a simple corollary of $\frac{1}{2^r}=1-(\frac{1}{2}+\cdots+\frac{1}{2^r})$.

share|cite|improve this answer
answered within 33 minutes --- MO rules! – Carlo Beenakker Nov 6 '12 at 15:22
@Gjergji Zaimi Thanks for the help. I could not derive any of the 2 equalities that you stated in the beginning. Can you kindly give some proof or reference for them? Especially in the second one I don't see any occurrence of the summing variable k in the summand. Can you kindly add some more lines of explanation! – Anirbit Nov 19 '12 at 23:56
Anirbit, thanks for catching the typo. I fixed it. – Gjergji Zaimi Nov 20 '12 at 1:51
I may try to add some more explanations a bit later, but if you expand the fractions $\frac{1}{1-x^n}=1+x^n+x^{2n}+\cdots$, it shouldn't be very hard to see the equations that I wrote. – Gjergji Zaimi Nov 20 '12 at 1:52

I would make a mere comment since Gjergji has already answered, but I am not allowed to make comments.

... so an alternative way to pose my question is to ask for a proof of the series expansion $\ln(-x,x)_{\infty}=\sum_{n=1}^{\infty}\frac{1}{n}\frac{x^{n}}{1-x^{2n}}$.

This is a corollary of Euler's theorem that the number of partitions of $n$ into distinct parts is equal to the number of partitions of $n$ into odd parts. In terms of generating functions, Euler's theorem is just $(-x,x)_{\infty}=\frac{1}{(x,x^2)_\infty}$, which can be easily proved by replacing the term $(1+x^i)$ in $(-x,x)_\infty$ by $\frac{1-x^{2i}}{1-x^i}$ and cancelling all the terms in the numerator against the corresponding terms in the denominator. By Euler's theorem, $\ln \left( (-x,x)_{\infty}\right) =\ln\left( \frac{1}{(x,x^2)_\infty}\right) =\sum_{i=1}^\infty \ln \left( \frac{1}{1-x^{2i-1}} \right) =\sum_{i=1}^\infty \sum_{n=1}^\infty \frac 1 n x^{n (2i-1)}$

$=\sum_{n=1}^\infty \sum_{i=1}^\infty \frac 1 n x^{n (2i-1)}$ $ =\sum_{n=1}^\infty \frac 1 n \frac{x^n}{1-x^{2n}} $.

share|cite|improve this answer
thanks a lot, very helpful indeed. – Carlo Beenakker Nov 6 '12 at 19:33
@Garth Payne So in the light of your answer does this function $\prod _{n=1} ^{\infty} (1+x^n)$ have any modular properties? Also can you give a more specific reference to this Euler's theorem that you refer to whose special this is you say. – Anirbit Nov 21 '12 at 3:20
@anirbit. Euler's theorem is in George Andrew's book The Theory of Partitions (p. 5). The proof I refer to above is given there in more detail. In case you don't have access to the book, someone has dupiicated the proof here:… . – Garth Payne Nov 27 '12 at 15:22

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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