On Fuchsian groups of genus zero (Q1907460)

To each Fuchsian group \(\Gamma\) of signature \((0; n_1, \dots, n_r)\) is associated in a canonical way a complex reflection group \(\widetilde \Gamma\) acting discontinuously on the domain \(\widetilde B = \{z \in \mathbb{C}^2 \mid |z_0 |^2 - |z_1 |^2 < 0\}\) in \(\mathbb{C}^2\): the image \(B\) of \(\widetilde B\) in the projective space \(P \mathbb{C}^2 = P^1 (\mathbb{C})\) serves as a model for the real hyperbolic plane on which \(\Gamma\) acts (with the 2-sphere as quotient), and the elliptic generators of \(\Gamma\) (whose orders are \(n_1, \dots, n_r)\) correspond to reflections of \(\widetilde B\) which generate \(\widetilde \Gamma\). The main result of the paper says that exactly for the signatures satisfying the Diophantine equation \(2/m + \sum 1/n_i = r - 2\), for some positive integer \(m\), the quotient spaces \(\widetilde B// \widetilde \Gamma\) (in the sense of the theory of invariants) is isomorphic to \(\mathbb{C}^2\); also, there are exactly 288 signatures of such type. This can be considered an analogue for Fuchsian groups of genus 0 of Chevalley's classical theorem on the invariants of finite linear reflection groups. It also means that exactly for Fuchsian groups of such signatures the algebra of \(\Gamma\)-automorphic forms, with respect to some cocycle, is free.