Maximal volume representations are Fuchsian (Q2494099)

Let \(M\) be a complete hyperbolic \(k\)-manifold with finite volume (\(k\geq 3\)). Then the universal covering space \(\widetilde M\) is the hyperbolic \(k\)-space \({\mathbf H}^k\) and there is a representation \(\rho:\pi_1(M)\to \text{Isom}({\mathbf H}^k)\) such that \(\rho\) is faithful and discrete, and the volume of \(\rho\) is equal to the volume of \(M\). (The definition of the volume of a representation is given in the section 2). Inversely, is for a given representation \(\rho:\pi_1(M)\to \text{Isom}({\mathbf H}^n)\) (\(k\leq n\)), the volume of \(\rho\) is related to \(\text{Vol}(M)\)? Dunfield showed that \(\text{Vol}(\rho) \leq \text{Vol}(M)\) for \(k=n=3\) closed \(M\). (The equality holds only when \(\rho\) is the above elementary example.) In this paper the authors give the full answer for any \(k\) and \(n\) (\(3\leq k\leq n\)). We consider a pseudo-developing map \(D : {\widetilde M}\to {\mathbf H}^n\), that is, \(D\) is a piecewise smooth \(\rho\)-equivariant map. Such a map can be constructed for any \(\rho\) (not unique). \(D\) has a volume and \(\text{Vol}(\rho)\) is defined by the minimum of \(\text{Vol}(D)\) for any pseudo-developing map \(D\) of \(\rho\). So if \(M\) is closed, mainly it is good to show the existence of \(D\) such that \(\text{Vol}(D) \leq \text{Vol}(M)\), and it is shown in the former paper of the first author. If \(M\) is not closed but \(M\) has a finite volume, the problem is very difficult. The main route is similar and it is good to find out a `good' map \(D : {\widetilde M}\to {\mathbf H}^n\) with good properties around its ends. The authors succeed to find such a map within short pages.