Spheres not admitting smooth odd-fixed-point actions of \(S_5\) and \(SL(2, 5)\) (Q2176711): Difference between revisions

This nice paper investigates smooth actions of finite groups on manifolds such that the number of fixed-points is odd including the case when there is exactly one fixed-point. One can check easily that if an action of a finite group on a sphere of dimension 0, 1, or 2 has at least one isolated fixed-point, then there are exactly two fixed-points. The situation is different for higher dimensional spheres. It is known that there are one fixed-point actions of \(\mathrm{SL}(2, 5)\) on \(\mathbb{S}^7\). Further, in his earlier works, the first author has constructed one fixed-point actions of \(A_5\) on \(\mathbb{S}^n\) for \(n\ge 6\). On the other hand, works of Furuta, Demichelis, Buchdahl, Kwasik and Schultz show that there are no one fixed-point actions of any finite group on \(\mathbb{S}^n\) for \(n = 3, 4, 5\). \par For homotopy spheres, it is known due to \textit{E. Laitinen} and \textit{P. Traczyk} [Proc. Am. Math. Soc. 97, 151--157 (1986; Zbl 0593.57020)] that for \(n = 5\) or \(n\ge 7\), a group \(G\neq A_5\) does not admit any one fixed-point action on an \(n\)-dimensional homotopy sphere \(\Sigma\) such that \(\dim \Sigma^H \le 2\) for all non-trivial subgroups \(H\) of \(G\). The present paper considers the case of homology spheres and the main result is the following theorem. \par Theorem 1.1. Let \(G\) be the group \(S_5\) (resp. \(\mathrm{SL}(2, 5)\)) and \(\Sigma\) a \(\mathbb{Z}\)-homology sphere. If \(\dim \Sigma \in \{0, 1, 2, 3, 4, 5, 7, 8, 9, 13\}\) (resp. \(\{0, 1, 2, 3, 4, 5, 6, 8, 9\}\)), then there is no effective odd fixed-point action of \(G\) on \(\Sigma\). This extends a result of \textit{A. Borowiecka} [Bull. Pol. Acad. Sci., Math. 64, No. 1, 85--94 (2016; Zbl 1352.57037)] who proved that there do not exist one fixed-point actions of \(\mathrm{SL}(2, 5)\) on a \(\mathbb{Z}\)-homology sphere of dimension 8.