Maximal partial spreads of \(T_2(\mathcal O)\) and \(T_3(\mathcal O)\) (Q1869032)

For \(n=2,3\) let \(\mathcal O\) be an oval or an ovoid, respectively, of PG\((n,q)\), where the latter geometry is assumed to be some hyperplane of \(\text{ PG}(n+1,q)\). The points of the generalized quadrangles \(T_n({\mathcal O})\) are defined to be the points of \(\text{ PG}(n+1,q)\setminus \text{ PG}(n,q)\) plus the hyperplanes of \(\text{ PG}(n+1,q)\), which intersect \(\mathcal O\) in exactly one point, plus one additional point \((\infty)\), whereas the lines of \(T_n({\mathcal O})\) are just the lines of \(\text{ PG}(n+1,q)\) not contained in the hyperplane \(\text{ PG}(n,q)\) meeting \(\mathcal O\) plus all points of \(\mathcal O\). Incidence within \(T_n({\mathcal O})\) is the one inherited by \(\text{ PG}(n+1,q)\) and the point \((\infty)\) is incident with just the points of \(\mathcal O\). Let \(S\) be a maximal partial spread of \(T_n({\mathcal O})\) of size \(q^n+1-\delta\), where \(\delta \geq 0\). The authors prove the following two theorems. (1) Let \(n=2\) and \(\delta > 0\). If \((\infty\)) is not a hole (i.e. a point which is not incident with any line of \(S\)) and if \(q\) is not square, then \(S\) can be extended to a spread if \(\delta < \varepsilon\), where \(q+\varepsilon\) is the size of the smallest non-trivial blocking set in PG\((2,q)\). If \(q\) is square and if \((\infty)\) is not a hole, then \(S\) can be extended to a spread if \(\delta < \varepsilon\) such that any blocking set of size \(q+\varepsilon\) contains either a line or a Baer subplane. If \(q\) is even and if \((\infty)\) is a hole, then \(\delta \geq q\). (2) Let \(n=3\). If \((\infty\)) is not a hole, then \(\delta\geq\varepsilon\), where \(q+\varepsilon\) is the size of the smallest non-trivial blocking set in PG\((2,q)\). If \((\infty\)) is a hole and if \(q > 2\), then \(\delta \geq 3\) for some maximal partial spread of \(T_3({\mathcal O})\). For \(T_2({\mathcal O}) \cong Q(4,q)\) this improves a theorem of \textit{G. Tallini} [Mitt. Math. Semin. Gießen 201, 141-147 (1991; Zbl 0738.51008)]. For the case of Tallini's result, the result of the paper under review turns out to represent a sharp bound.