On a function related to Ramanujan's tau function (Q1069965): Difference between revisions

For each positive divisor \(\alpha\) of 24, the arithmetical function \(\psi_{\alpha}(n)\) is defined by \[ \sum^{\infty}_{n=1}\psi_{\alpha}(n) x^ n=x\prod^{\infty}_{n=1}(1-x^{24n/\alpha} )^{\alpha}. \] For \(\psi =\psi_{12}\) it is shown that \(\psi (2m)=0\) for \(m\geq 1\), and that \[ \psi (2m+1)=\sum^{m}_{i=0}(-1)^ i r_ 8(i) \sigma (2m-2i+1), \] where \(r_ 8(n)\) is the number of representations of n as a sum of 8 squares, and \(\sigma\) (n) is the sum of the positive divisors of n. Jacobi's formula for \(r_ 8(n)\) leads to an alternate expression for \(\psi (2m+1)\) entirely in terms of \(\sigma\).