Completely periodic directions and orbit closures of many pseudo-Anosov Teichmüller discs in \(\mathrm{Q}(1,1,1,1)\). (Q422381)

It is no exaggeration to say the problem of classifying \(\mathrm{GL}^+(2, \mathbb R)\) and unipotent orbit closures on moduli spaces of abelian differentials is one of the most important questions in the field of dynamical systems, given the wealth of applications to billiards in rational polygons and more generally dynamics on translation surfaces. The hope is to prove a version of Ratner's theorem for unipotent orbit closures and invariant measures on homogeneous spaces. Important progress toward this was made by \textit{C. T. McMullen} [Ann. Math. (2) 165, No. 2, 397--456 (2007; Zbl 1131.14027)], who classified \(\mathrm{GL}^+(2, \mathbb R)\)-orbit closures in the setting of abelian differentials in genus \(2\), by `splitting' the genus \(2\) surfaces into the connected sum of two tori, and bootstrapping Ratner's theorem on unipotent orbit closures on the moduli space of (affine) tori, which is a homogenous space. In this paper, the authors use a similar technique to show that for a certain class of abelian differentials on genus \(3\) surfaces (which correspond to a more natural class of quadratic differentials on genus \(2\) surfaces) has orbit closure `as large as possible'. They study the locus \(\mathcal L\), the hyperelliptic locus of the non-hyperelliptic component of the stratum \(\mathcal H (2, 2)\) of genus \(3\) abelian differentials with \(2\) double zeros. This is \(\mathrm{GL}^+(2,\mathbb R)\)-equivariantly isomorphic to the principal stratum \(Q(1, 1, 1, 1)\) of quadratic differentials with simple zeros on genus \(2\) surfaces. The main result of this paper states that if \((X, \omega) \in \mathcal L\) is a surface obtained by the Thurston-Veech construction of pseudo-Anosov diffeomorphisms, which has trace field of degree \(3\), and a completely periodic non-parabolic direction, then \[ \overline{ \mathrm{GL}^+(2, \mathbb R) (X,\omega)} = \mathcal L. \]