Partial unitals and related structures in Desarguesian planes (Q1281407)

A \((k,n)\)-arc in a projective plane \(\Pi\) is a set of \(k\) points, at most \(n\) on every line. If the order of the plane is \(q\), where \(q\) is a square, a unital \(\mathcal U\) is defined to be a \((q\sqrt q+1,\sqrt q+1)\)-arc meeting all lines of \(\Pi\) in 1 or \(\sqrt q+1\) points. Lines meeting \(\mathcal U\) in exactly one point are called tangent lines. A partial unital is a \((k,\sqrt q+1)\)-arc \(\mathcal X\) such that each point of \(\mathcal X\) lies on a tangent. A unital is a minimal blocking set; that is it contains at least one point on every line and is minimal since each point lies on a tangent. For a minimal blocking set \(\mathcal B\), it is well-known that (a) \(q+\sqrt q+1\leq | {\mathcal B}| \leq q\sqrt q+1\), (b) \(| {\mathcal B}| =q+\sqrt q+1\) if and only if \(q\) is a square and \({\mathcal B}\) is a Baer subplane, (c) \(| {\mathcal B}| =q\sqrt q+1\) if and only if \(q\) is a square and \({\mathcal B}\) is a unital. The question arises of what are the integers \(k\), \(q+\sqrt q+1< k< q\sqrt q+1\), such that a minimal blocking set of size \(k\) exists -- such values form the so-called spectrum of \(\Pi\). Then it could be interesting to decide whether in particular the integers \(k\), near \(q\sqrt q+1\), belong to the spectrum. \textit{A. Blokhuis} and \textit{K. Metsch} [Lond. Math. Soc. Lect. Note Ser. 191, 37-52 (1993; Zbl 0797.51012)] give a negative answer for \(q\geq 25\), \(q\) a square, \(k=q\sqrt q\). \textit{S. Innamorati} extends this result to \(q=9\) [J. Discrete Math. Sci. Cryptography 1, 1-26 (1998; Zbl 0911.51009)]. In the paper under review it is shown that in a Desarguesian finite plane of order \(q\), a partial unital \(\mathcal X\) with \(q\sqrt q+1-\sqrt q<| {\mathcal X}| <q\sqrt q+1\) can be extended to a unital. As a consequence, \(\mathcal X\) cannot be a blocking set. The techniques used are based on those used in the papers by Ball, Blokhuis and C. O'Keefe (to appear) concerning unitals, and \textit{S. Ball} and \textit{A. Blokhuis} [Geom. Dedicata 74, 325-332 (1999)] concerning the incompleteness of maximal arcs.