Even sets of nodes on sextic surfaces (Q2478596)

Let \(F\) be a nodal complex surface in \(\mathbb P^3\) of degree \(d\), i.e., \(F\) has only nodes (ordinary double points) as singularities. A classical problem, still open for \(d>6\), is to find the maximum number of nodes \(\mu(d)\) that such a surface \(F\) can have. In order to show that \(\mu(5)\) is achieved by Togliatti quintic and that it is \(31\), \textit{A. Beauville} [Algebraic Geometry (Angers, 1979), Sijthoff and Noordhoff, 207--215 (1980; Zbl 0445.14016)], introduced the notion of an even set of nodes. This notion can be derived from the coding theory framework, which was later used by \textit{D. B. Jaffe} and \textit{D. Ruberman} [J. Algebraic Geom. 6, 151--168 (1997; Zbl 0884.14015)] to show that \(\mu(6)\) is achieved by Barth sextic and that it is \(65\). The main results of the paper under review are the proofs of two theorems: A and B. Theorem A states: Let \(F\) be a nodal surface of degree \(d=6\) in \(\mathbb P^3\) with an even set of \(t\) nodes. Then \(t \in \{24,32,40,56\}\). These four possibilities occur and can be explicitly described. Partial results were known, e.g. an even set of \(48\) nodes was ruled out by Jaffe and Ruberman in the previously-quoted paper. Considering \(d<6\) as well, cf. e.g. \textit{G. Casnati} and \textit{F. Catanese} [J. Differ. Geom. 47, 237--256 (1997; Zbl 0896.14017)], the situation for \(d \leq 6\) is as follows: \[ d=3, \;t=4; \quad d=4, \;t \in \{8,16\}; \quad d=5, \;t \in \{16,20\};\quad d=6, \;t \in \{24,32,40,56\}. \] Theorem B concerns results obtained for a family of nodal sextic surfaces with \(56\) nodes forming an even set.