The limit set of non-orientable mapping class groups (Q6167215)

The article under review explores Teichmüller theory for nonorientable closed surfaces. In the case of classical Teichmüller theory, there is a heuristic analogy where Teichmüller space is compared to hyperbolic space~\(\mathbb{H}^n\), the mapping class group is compared to a Kleinian group and the Thurston boundary (the set of projective measured foliations) is compared to the boundary of~\(\mathbb{H}^n\). While there are certain strong similarities, things turn out to be surprisingly different in the nonorientable case. For example, in the unpublished preprint [``What's wrong with the growth of simple closed geodesics on nonorientable hyperbolic surfaces'', Preprint, \url{arXiv:1706.08798}], \textit{M. Gendulphe} showed that the number of curves of length~\(\le L\) does \emph{not} grow like~\(L^d\), where~\(d\) is the dimension of Teichmüller space. Rather, Gendulphe conjectures that the exponent should be the Hausdorff dimension of the limit set of the action of the mapping class group, which in the nonorientable case does not contain all projective measured foliations. In the heuristic analogy to hyperbolic geometry, Gendulphe suggests that the mapping class group of a nonorientable surface might be thought of more like an infinite covolume geometrically finite Kleinian group. The first main result of the article under review concerns the limit set of the mapping class group in the nonorientable case and supports the analogy to hyperbolic geometry: it is shown that the complement of the limit set is open and dense, and more precisely, that the limit set is contained in the set of projective measured foliations containing no closed one-sided leaf (whose complement is known to be open and dense and of full measure in the set of all projective measured foliations). A stronger version of this result was independently obtained by \textit{V. Erlandsson} et al. [Geom. Funct. Anal. 33, No. 3, 637--693 (2023; Zbl 1528.57012)], who proved, using different methods, that the limit set is exactly the set of projective measured foliations containing no closed one-sided leaf. The second main result is negative: it concerns a certain retract~\(T^-_\varepsilon\) of Teichmüller space, namely the set of points in Teichmüller space that have no one-sided curves shorter than~\(\varepsilon\), which was shown by Gendulphe to be of finite covolume. Furthermore, Gendulphe asked whether this retract is quasi-convex. A negative answer to this question is given in the article under review.