On Ovals of Riemann Surfaces of Even Genera

Abstract

Let X be a Riemann surface of genus g≥2. A symmetry of σ of X is an antiholomorphic involution acting of X. A classical theorem of Harnack states that the set Fix (σ) of fixed points of σ is either emplty or it consists of ‖σ‖≤ g+1 disjoint simple closed curves called, following Hilbert′s terminology, the ovals of σ. A Riemann surface admitting a symmetry corresponds to a real algebraic curve and nonconjugate symmetries correspond to different real models of the curve. The number of ovals of the symmetry equals the number of connected components of the corresponding real model. It is well known that two symmetries of a Riemann surface of genus g have at most 2g+2 ovals, and the bound is attained for every genus and just for commuting symmetries. Natanzon showed that three and four nonconjugate symmetries of a Riemann surface of genus g have at most 2g+4 and 2g+8 ovals respectively, and these bounds are attained for every odd genus and for commuting symmetries. Natanzon found that a Riemann surface of genus g has at most 2($\sqrt g$+1) nonconjugate symmetries and, again, this bound is attained for infinitely many of g. Recently we have showed that a Riemann surface of even genus g admits at most four symmetries. Our aim here is to show, using NEC groups and combinatorial methods, that three nonconjugate symmetries of a surface of even genus g has at most 2g+3 ovals and, surprisingly, if such a surface admits four nonconjugate symmetries then its total number of ovals does not exceed 2g+2. Furthermore, we show that this last bound is sharp for every even genus g and for surfaces with automorphism group D _n × Z₂, for each n dividing 2g.