On finite generalized quadrangles of even order (Q6175950)

In this paper, the author proves two longstanding conjectures on automorphism groups of finite generalised quadrangles. The first one concerns the conjecture of \textit{S. E. Payne} [Conf. Comb. Graph Theor. Comput. 485--504 (1975; Zbl 0336.05016)] that every skew translation generalized quadrangle of even order is a(n ordinary) translation generalized quadrangle. To explain this result it is good to recall that in the classical quadrangles of order \(s\), either every point \(p\) is a point of symmetry (that is, the group of collineations fixing all points collinear to \(p\) has order \(s\)), or each line is an axis of symmetry (this is the dual of the previous). But if the characteristic of the underlying field is equal to \(2\), then the quadrangle is self dual and hence all points are centres of symmetry and all lines are axes of symmetry. Now it is an easy consequence of the classification of so-called Moufang quadrangles that, if all points of a quadrangle of order \(s\) are centres of symmetry, and \(s\) is even, then all lines are axes of symmetry (and the quadrangle is classical). The local version, however, is much more difficult to prove. It says that, if a point \(p\) of a generalised quadrangle of even order \(s\) is a point of symmetry, and the \(s\) collineations that fix all points collinear to \(p\) belong to a collineation group that acts simply-transitively on the set of points not collinear to \(p\), then all lines through \(p\) are axes of symmetry. That is exactly what the author proves. The second conjecture of \textit{D. Ghinelli} [Geom. Dedicata 41, No. 2, 165--174 (1992; Zbl 0746.51011)] is easier to explain: the author shows that no generalised quadrangle of even order \(s\) admits a collineation group acting simply-transitively on the point set. In particular, there is no such thing as a Singer cycle for generalised quadrangles of even order. The proofs of both results use new ideas (e.g. character theory, bounds for bipartite graphs using spectral theory) compared to the classical approaches.