On non-singular Hermitian varieties of \(\mathrm{PG}(4, q^2)\) (Q2279280): Difference between revisions

Several authors have been interested in giving purely combinatorial characterizations of objects classically defined in an algebraic way. A non-singular Hermitian variety of \(\operatorname{PG}(r,q^2)\) is projectively equivalent to the hypersurface defined by \(\sum_{i=0}^r X_i^{q+1} = 0\). In [Combinatorica 30, No. 1, 25--45 (2010; Zbl 1224.05074)], \textit{S. De Winter} and \textit{J. Schillewaert} prove that a subset of points of \(\operatorname{PG}(r,q^2),\ r\ge 3\), having the same intersection numbers with respect to hyperplanes and subspaces of codimension \(2\) as a non-singular Hermitian variety, is a non-singular Hermitian variety. In an earlier paper [J. Aust. Math. Soc. 107, No. 1, 1--18 (2019; Zbl 1420.51007)], the first author shows that a hypersurface \(\mathcal{H}\) of \(\operatorname{PG}(4,q^2)\) of degree \(q+1\) having the same intersection numbers with respect to hyperplanes of a non-singular Hermitian variety is a non-singular Hermitian variety. The main result of this paper is as follows: Let \(\mathcal{H}\) be a hypersurface of \(\operatorname{PG}(4,q^2),\ q>3\), defined over \(\operatorname{GF}(q^2)\), and not containing planes. If the degree of \(\mathcal{H}\) is \(q+1\) and the number of its rational points is \(q^7+q^5+q^2+1\), then every plane of \(\operatorname{PG}(4,q^2)\) meets \(\mathcal{H}\) in at least \(q^2+1\) rational points. If there is at least one plane that intersects \(\mathcal{H}\) in exactly \(q^2+1\) points defined over \(\operatorname{GF}(q^2)\), then \(\mathcal{H}\) is a non-singular Hermitian variety of \(\operatorname{PG}(4,q^2)\). The main steps of the argument are to show that \(\mathcal{H}\) intersects each line of \(\operatorname{PG}(4,q^2)\) in at least one point, and then to study the set of planes of \(\operatorname{PG}(4,q^2)\) meeting \(\mathcal{H}\) in exactly \(q^2+1\) rational points. The result of De Winter and Schillewaert is used to conclude the proof.