On hyperelliptic involutions of hyperbolic 3-manifolds (Q5950056)

It is known by classical results that an orientable closed hyperbolic \(2\)-manifold (compact Riemann surface) has at most one hyperelliptic involution. The paper deals with the question in the 3-dimensional situation. Let \(M\) be a closed orientable hyperbolic 3-manifold (we will always assume that \(M\) denotes such a manifold). How many hyperelliptic involutions, that is, isometric involutions whose quotient is the \(3\)-sphere, does \(M\) admit. The main theorem in the paper is that the number of conjugacy classes of hyperelliptic involutions in \(\text{Isom}_{+}(M)\) is bounded by a universal constant \(C\) which is independent of \(M\). Subsequent work by Mecchia and Reni shows that \(C\) can be taken to be \(9\). It should be remarked that there remains a small gap between the bound \(C=9\) and the number of different hyperelliptic involutions which the authors can actually construct although they believe that \(9\) is probably the best bound. The paper also considers the case of hyperelliptic isometries of \(M\) of prime order. In the case of an odd prime \(p\), they show that for any such \(M\), up to conjugacy, there are at most three cyclic subgroups of \(\text{Isom}_+(M)\) generated by a hyperelliptic isometry of order \(p\), and that this bound is sharp in the sense that for any odd prime \(p\), there exists an \(M\) having exactly three conjugacy classes of cyclic subgroups in \(\text{Isom}_+(M)\) which are generated by a hyperelliptic isometry of order \(p\). Finally, they also show that for any triple of integers \(p\), \(q\) and \(r\) greater than \(2\), there exists an \(M\) admitting hyperelliptic involutions of orders \(p\), \(q\) and \(r\). Furthermore, there is no universal bound for the number of orders of hyperelliptic isometries for closed hyperbolic \(3\)-manifolds.