Two new sporadic semibiplanes related to \(M_{22}\) (Q1332355): Difference between revisions

Let \(\Gamma\) be a rank 3 Buekenhout geometry with point, lines and circles. Suppose that the residue of any point is a complete graph on \(n + 2\) vertices (the vertices correspond to circles in \(\Gamma\); the edges to lines), that the residue of a line is a thin digon and that the residue of every circle is again a complete graph on \(n + 2\) vertices (where the vertices correspond to points of \(\Gamma\) and the edges to lines), then \(\Gamma\) is a \(c.c^*\)-geometry with parameter \(n\). The connection between these geometries and semibiplanes is that any \(c.c^*\)-geometry defines a semibiplane (by truncating to points and circles, i.e. by ``forgetting'' the lines) if the intersection property holds. Conversely, a semibiplane defines a \(c.c^*\)-geometry by defining lines as sets of two points and two circles forming a digon in the incidence graph. The paper under review classifies all flag-transitive \(c.c^*\)- geometries with parameter 13. There are essentially three examples. One arising from (covered by) a `two-coloured hypercube' (and this belongs to an infinite class), and two new examples, related to the Mathieu group \(M_{22}\). One is a 2-cover of the other. Let us briefly define the latter, as a semibiplane. The points are the octads of the Steiner system \(S(5, 8, 24)\) related to \(M_{24}\) through some fixed point \(x_ 1\) and not containing some fixed point \(x_{24}\); the circles are the octads through \(x_{24}\) not containing \(x_ 1\). It turns out that all three of the examples have the intersection property, hence define semibiplanes. The eventual classification uses coset enumeration by a computer.