Nonequivalent families of group actions on Riemann surfaces (Q392229)

Let \(S\) denote a compact Riemann surface and let \(G\) denote its full group of conformal automorphisms. Two subgroups \(H_{0}, H_{1} \leq G\) are said to be conformally equivalent if there exists a conformal automorphism \(t\in G\) such that \(tH_{0} t^{-1} =H_1\). They are said to be directly topologically equivalent if there exists an orientation preserving homeomorphism \(t: S\rightarrow S\) such that \(tH_{0} t^{-1} =H_1\). Clearly conformally equivalent subgroups are directly topologically equivalent. The converse however is not true, see for example [\textit{G. González-Diez} and \textit{R. A. Hidalgo}, Bull. Lond. Math. Soc. 29, No. 3, 280--284 (1997; Zbl 0865.14011)] where an infinite family of Riemann surfaces of genus \(9\) is constructed for which there exists two cyclic subgroups which define directly topologically equivalent actions, but not conformally equivalent actions. Building on these results, in the article under review, for each \(n \in {\mathbb N}\) the author describes an infinite family of Riemann surfaces of genus \(g=3(2n -1)\) for which there exists two cyclic subgroups which define directly topologically equivalent actions, but not conformally equivalent actions. The construction of these surfaces and the corresponding automorphisms by the author is explicit. For a given \(a,n\in {\mathbb N}\) with \(a\) odd and \(\lambda \in {\mathbb C}\) with \(\lambda^4\neq 1\), let \(S_{a,\lambda}\) denote the compact Riemann surface with defining equation \(f_{a,\lambda} (a,y)=y^{2^n} -x^a(x^2-1)^a(x^2-\lambda^2) (x^2-\lambda^{-2})\). For a given \(n\), such a surface has genus \(g=3(2n-1)\). Moreover, except for finitely many cases, the automorphisms \(\tau_1\) and \(\tau_2\) defined by \[ \tau_1 (x,y)=(-x,e^{\frac{2\pi i}{2^{n+1}}} y) \text{ and } \tau_2 (x,y)=(\frac{1}{x},\frac{e^{\frac{2\pi i}{2^{n+1}}} y}{x^c}) \] with \(c\) satisfying \(c2^{n-1}=2a+2\) generate the full group of conformal automorphisms. By employing a very useful criterion developed by the author earlier in the paper based on the local behavior of the corresponding automorphisms (Theorem 7), it is shown that the groups \(\langle \tau_1 \rangle\) and \(\langle \tau_2\rangle\) are directly topological equivalent. Since for all but finitely many cases \(G= \langle \tau_1,\tau_2\rangle\) and we have either \(G={\mathbb Z}/2^{n+1} {\mathbb Z} \times {\mathbb Z}/2 {\mathbb Z}\) or \(G={\mathbb Z}/2^{n+1} {\mathbb Z} \rtimes {\mathbb Z}/2 {\mathbb Z}\) depending on the parity of \(c\), it is easy to see that the groups \(\langle \tau_1 \rangle\) and \(\langle \tau_2\rangle\) cannot possibly be conformally equivalent, hence the result. The paper is very nicely written with a good balance of preliminaries. It should also be noted that the aforementioned criterion for when two cyclic groups define directly topological actions (Theorem 7) is itself a very useful result that might be applied to find other families of surfaces for which there are directly topologically equivalent actions which are not conformally equivalent.