\(p\)-groups of automorphisms of compact non-orientable Riemann surfaces (Q2050197)

Classical Riemann surfaces are orientable and unbordered, and both conditions can be relaxed in order to obtain the more general concept of Klein surfaces. These surfaces, and their groups of automorphisms, are uniformized by non-Euclidean crystallographic groups, which are a generalization of Fuchsian groups, allowing orientation-reversing elements. The automorphism groups of classical Riemann surfaces have been studied since very long time. For instance, the \(p\)-groups of automorphisms were studied by \textit{R. Zomorrodian} who obtained upper bounds for their order in [Glasg. Math. J. 29, 237--244 (1987; Zbl 0626.20035)]. In this article, the authors consider non-orientable unbordered surfaces, which they call non-orientable Riemann surfaces. Their goal is to obtain an upper bound for the order of a \(p\)-group which acts as an automorphism group of a non-orientable Riemann surface of topological genus \(g \geq 3\). Using that result, the same problem is considered for the full automorphism group of the surface. Firstly, the authors prove that if \(p\) does not divide \(g-2\), then the \(p\)-group \(G\) acting on a surface of genus \(g\) is cyclic or dihedral. Then, they remove that arithmetic restriction and study cyclic and dihedral \(p\)-groups, and obtain the searched upper bound, and show that it is sharp, that is to say, it is attained for infinitely many values of \(g\). Finally, they consider general non-cyclic groups \(G\). In this last section, the result depends not only on \(p\) and \(g\), but on the minimal number of generators of \(G\). In all cases, the bounds are proved to be sharp, the structure of the group \(G\) and its action on the surface are determined, and the bounds are seen to hold for the full automorphism group. Just a single case remains open, concerning 2-generator 2-groups, in which the bound is not proved to be sharp. It is worth to detail a little this open question. According to Theorem 6.1 in the paper, if \(G\) is a non-cyclic \(2\)-group of automorphisms of a compact non-orientable Riemann surface of topological genus \(g\), and \(G\) is generated by 2 elements, then the order of \(G\) is at most \(8(g-2)\). This bound is attained for \(g=3\) by the dihedral group \(D_4\). However, this is the unique example known to attain the bound. In fact, using the lists by Marston Conder that can be found on his webpage, one sees that there exists only a handful of groups of order \(2^n\) with \(n \leq 8\) that satisfy the bound, but all of them require 3 generators, except just \(D_4\). Thus, it looks very probable to the reviewer that this bound is not sharp, but really difficult to prove. If this is the case, then the next candidate bound is \(4(g-2)\). This bound should be sharp, because for all \(n \geq 4\) the dihedral group \(D_{2^{n-1}}\) of order \(2^n\) acts with order \(4(g-2)\) on a surface of genus \(g\), as an image of an NEC group with signature \((0,+,[-],\{(2,2,2,2,2)\})\). Returning to the article, it is really well written and all the constructions are clear and detailed.