Conjugacy classes of \(p\)-torsion in \(SP_{p-1}(\mathbb{Z})\) (Q1372667)

Let \(S\) be a connected compact Riemann surface of genus \(g\)(\(\geq 2\)) without boundary and let \(\Aut(S)\) be its group of analytic automorphisms. It is known that there exists a monomorphism \(\theta\colon\Aut(S)\to SP_{2g}(\mathbb{Z})\), where \(SP_{2g}(\mathbb{Z})\) is the symplectic group (of degree \(2g\)) over \(\mathbb{Z}\). In this paper the authors are concerned with the question of determining which matrices of order \(p\) in \(SP_{2g}(\mathbb{Z})\), where \(p\) is a prime, lie in the image of \(\theta\). Such matrices are said to be realizable. With the exceptions of the 2-sphere and the torus, the least genus of a surface \(S\) in which the cyclic group \(\mathbb{Z}_p\) acts is \((p-1)/2\) and \(p\geq 5\). Using algebraic number theory applied to the number field \(\mathbb{Q}(\zeta)\) and its maximal order \(\mathbb{Z}[\zeta]\), where \(\zeta=\exp(2\pi i/p)\), the authors demonstrate in a precise way that not all matrices of order \(p\) are realizable. In the process they determine the number of conjugacy classes among the order \(p\) elements in \(SP_{p-1}(\mathbb{Z})\).