Take the 2-minute tour ×
MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

MacMahon in the paper Divisors of Numbers and their Continuations in the Theory of Partitions defines several generalized notions of the sum-of-divisors function; for example, if we write $a_{n,k}$ for the sum $$ \sum s_1 \cdots s_k $$ where this sum is taken over all ways of writing $n = s_1m_1 + \cdots s_km_k$ with $m_1 < \cdots < m_k$ (note the asymmetry in $s_k, m_k$), he then studies the generating functions $$ A_k(q) = \sum_{n=1}^\infty a_{n,k}q^n $$ for fixed $k$, as well as a number of other variants. Note that for $k=1$, this is nothing but the generating function for the ordinary sum-of-divisors function.

These functions (Specifically, from his paper, the functions $A_k$ and $C_k$) have arisen in my research and I would like to know what literature there is on them. In particular, I would like to know if there are any well-known identities that hold between them. MacMahon himself lists the identity $$ A_2(q) = \tfrac{1}{8}\sum_{n=1}^\infty\big(\sigma_3(n) - (2n-1)\sigma_1(n)\big)q^n $$ as well as similar ones for $A_3$ and $A_4$, but the identities that I am looking for are more in line with an attempt to write these as quasi-modular forms, if possible. For example, it turns out that the following is true: $$ A_2(q) = \tfrac{1}{10}\Big(3A_1(q)^2 + A_1(q) - q\frac{d}{dq}A_1(q)\Big) $$ and I conjecture that you can always write $A_k(q)$ (and similarly $C_k(q)$) recursively in terms of previous such functions.

So my question is: What literature is there on these functions? Is the relation above a well known one?

share|improve this question
add comment

2 Answers

up vote 2 down vote accepted

For anyone who comes by this later, it turns out that the following relationship is true for the functions $A_k(q)$: $$ A_k(q) = \frac{1}{(2k+1)2k}\Big(\big(6A_1(q) + k(k-1)\big)A_{k-1}(q) - 2q\frac{d}{dq}A_{k-1}(q)\Big) $$ After contacting George Andrews (as suggested), he and I wrote a joint paper proving this result (as well as a few other related ones), which can be found at


share|improve this answer
add comment

I don't know, but if you type $$\rm macmahon\ divisors\ continuations$$ into Google you'll find a number of papers by George Andrews, and one by Jim Huard, citing the MacMahon paper.

share|improve this answer
Andrews might be a good person to ask directly; from reading some of his papers it looks like he's really interested in MacMahon's work. –  Qiaochu Yuan Oct 7 '10 at 22:27
I've already downloaded and skimmed over two of those, "stacked lattice boxes" and "partitions with designated summands" which don't seem to really help; they're mostly about the case where you simply count the number of such 'divisor partitions', instead of summing them. I could try email him though. –  Simon Rose Oct 7 '10 at 22:37
add comment

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.