Sets in a finite plane with few intersection numbers and a distinguished point (Q1183988): Difference between revisions

This paper is an extremely rich source of deep, difficult results concerning sets of conics or Hermitian curves in \(PG(2,q)\), the projective plane of order \(q\). To quote the authors: ``there has been much study of algebraic curves such as conics and Hermitian curves as well as subsets with various intersection properties which characterize such curves. For example, for \(q\) odd, a set \(K\) of \(q+1\) points such that every line meets \(K\) in 0, 1 or 2 points is a conic''; ``for \(q\) square, a set \(K\) of \(q\sqrt{q}+1\) points such that (a) every line meets \(K\) in 1 or \(\sqrt{q}+1\) points, and (b) every such \(\sqrt{q}+1\) points is a subline \(PG(1,\sqrt{q})\), is a Hermitian curve'' [\textit{G. Faina} and \textit{G. Korchmàros}, Ann. Discrete Math. 18, 335-342 (1983; Zbl 0513.05020) and \textit{C. Lefèvre-Percsy}, Arch. Math. 39, 476-480 (1982; Zbl 0494.51003)]. ``A more difficult question is to characterize sets of conics or Hermitian curves in the plane. Here we make some advances in this direction by looking at pencils of conics or Hermitian curves with one base point. This is done by generalizing the notion of a set with two intersection numbers to a set with a special point \(P\) such that all lines through \(P\) apart from a tangent meet the set in a fixed number of points and all other lines have one of two intersection numbers with the set. A key result that emerges from the investigation is contained in Corollary 5.6 and Theorem 5.7 which together show that, for \(q\) odd, there exists a Hermitian arc (unital) in \(PG(2,q)\) that is the union of \(\sqrt{q}\) conics and that this unital is nonclassical of Buekenhout-Metz type''. For a description of these unitals, see the first author, ``Finite Projective Spaces of Three Dimensions'' (1985; Zbl 0574.51001). Independently, \textit{R. D. Baker} and \textit{E. L. Ebert} [Congr. Numerantium 70, 87-94 (1990; Zbl 0695.51009)] have shown the existence of a unital as the union of conics. The details are too involved to be described here; the reader may find them in the original paper.