Quasi-isometric rigidity of Fuchsian buildings (Q2570064)

Quasi-isometric rigidity theorems are known for a variety of spaces. In the paper under review the author shows that if \(g:\Delta_1\to\Delta_2\) is a quasi-isometry between Fuchsian buildings \(\Delta_1\) and \(\Delta_2\) that admit cocompact lattices, then \(g\) lies at a finite distance from an isomorphism. The proof of this rigidity result involves the following Theorem 1.5, whose verification takes up much of the paper and is its main focus. Let \(h:\partial\Delta_1\to\partial\Delta_2\) be a homeomorphism between the ideal boundaries of two Fuchsian buildings \(\Delta_1\) and \(\Delta_2\) and assume that \(h\) preserves the combinatorial cross ratio almost everywhere. Then \(h\) extends to an isomorphism \(f:\Delta_1\to\Delta_2\). The quasi-isometric rigidity of Fuchsian buildings then follows using arguments due to \textit{M. Bourdon} and \textit{H. Pajot} [Comment. Math. Helv. 75, 701--736 (2000; Zbl 0976.30011)] and yet unpublished results by B. Kleiner which assert that the cotangent bundle of the ideal boundary of a Fuchsian building, which is an Ahlfors regular Loewner space, is 1-dimensional and that in this situation the full quasiconformal group is uniformly quasiconformal. For the proof of Theorem 1.5 the author first develops a property that allows him to decide by means of the combinatorial cross ratio whether \(\xi\eta\) where \(\xi,\eta\in\Delta\) is contained in the 1-skeleton \(\Delta^{(1)}\) of the Fuchsian building \(\Delta\). A further key step is to show that the images of all geodesics in the 1-skeleton of \(\Delta_1\) through a vertex \(v\) intersect in a unique vertex \(w\) in \(\Delta_2\); the map \(v\to w\) then is a bijection between the vertex sets that extends to an isomorphism from \(\Delta_1\) to \(\Delta_2\). To this end the author considers triangles and qudrilaterals contained in the 1-skeleton each of which bounds a topological disk which is the union of a finite number of chambers. Gauss-Bonnet implies that there are only a few possibilities for such triangles and quadrilaterals and the author obtains a complete list of all such triangles \((m_1,m_2,m_3)\) (unique up to isometry) in \({\mathbb H}^2\) with angles \(\pi/m_1,\pi/m_2,\pi/m_3\). To finally prove Theorem 1.5 three main cases are considered: (1) chambers are not right triangles; (2) both chambers are right triangles but none is \((2,3,8)\); (3) both chambers are \((2,3,8)\). It then is a case by case analysis. Usually a triangle \(T\subset\Delta^{(1)}\) is present that has a certain property. A check is then made against the list of triangles found before to see whether any of them or which ones of them have the stated property.