Hearing quaternions in hyperbolic surfaces
It is well known that the Laplace spectrum of a closed hyperbolic surface determines the lengths of its closed geodesics and hence the trace set of the corresponding Fuchsian group. Long-standing conjectures of Sarnak (bounded clustering and positive trace gap) and Schmutz (linear growth) relate the distribution of these traces to the arithmetic structure of the group.
In 2008, Geninska and Leuzinger proved Sarnak’s bounded-clustering conjecture for non-cocompact Fuchsian lattices. Hao subsequently gave an algebraic proof of the positive trace gap conjecture and, more recently, proved Schmutz’s linear-growth conjecture in the same setting. These arguments rely essentially on parabolic elements in non-cocompact lattices.
I will introduce these conjectures and the relevant notions of arithmeticity, and then sketch a proof of the positive trace gap conjecture in the cocompact case: a cocompact Fuchsian lattice with uniformly separated traces is derived from a quaternion algebra, a condition even stronger than arithmeticity. The proof combines a consequence of Ratner’s orbit closure theorem, an elementary matrix construction based on the Cayley--Hamilton theorem, the Margulis commensurator rigidity theorem, and Vinberg’s results on invariant trace fields and quaternion algebras. This talk is based on my recent paper https://arxiv.org/abs/2609.29033

