Distribution of the partition function modulo composite integers \(M\) (Q1592361)

Ramanujan's partition congruences are \[ p(5n+4)\equiv 0\pmod 5, \quad p(7n+5)\equiv 0\pmod 7, \quad p(11n+6)\equiv 0\pmod {11}. \] These have extensions modulo powers of 5, 7, 11. In general, congruences of the form \[ p(An+B)\equiv 0\pmod M \] are very rare. However, \textit{K. Ono} [Ann. Math. (2) 151, 293-307 (2000; Zbl 0984.11050)] proved there are infinitely many of these congruences for any prime value of \(M\), and subsequently \textit{R. Weaver} [Ramanujan J. 5, 53-63 (2001; Zbl 0986.11071)] gave over 70 000 explicit examples of such congruences for \(M\in \{13, 17, 19, 23, 29, 31\}\). The cases \(M=2\) and \(M=3\) are more difficult. In the paper under review, composite values of \(M\) are considered. It is shown that if \(M\) is a positive integer coprime to 6, then there are infinitely many arithmetic progressions \(An+B\) (none of which is contained in any other) with the property that \(p(An+B)\equiv 0\pmod M\) for every integer \(n\). The proof uses \(M\)-adic modular forms, the Shimura correspondence and a theorem of Serre. It is an existence proof and no explicit examples of congruences are given. Using the same methods, Newman's conjecture is shown to be true for good positive integers \(M\) which are coprime to~6.