Sets with a large number of nuclei on a conic (Q1207775)

Let \(B\) be a set of \(q+1\) points in \(PG(2,q)\), the desarguesian projective plane of order \(q\), for some odd prime power \(q\). A point of \(PG(2,q)\) with the property that every line through it intersects \(B\) in exactly one points is called a nucleus of \(B\). We denote by \(N(B)\) the set of all nuclei of \(B\). Clearly, if \(B\) is a line, then \(N(B)=PG(2,q)\backslash B\). We therefore assume \(B\) is not a line for the remainder of this review. A result of \textit{B. Segre} and \textit{G. Korchmáros} [Atti Accad. Naz. Lincei, VIII. Ser., Rend., Cl. Sci. Fis. Math. Nat. 62, No. 5, 613-619 (1977; Zbl 0383.51006)] says that in this case \(N(B)\) cannot contain a conic. However, certain examples of J. C. Fisher (unpublished but described in this paper) show that \(N(B)\) can contain a surprisingly large portion of a conic. In the paper under review the authors use a very clever counting argument to show that if \(q>10^ 7\) \((q\) odd) and \(B\) as more than \(0.326q\) nuclei on a given conic \(C\), then \(B\) is of Fisher type. The constants \(10^ 7\) and 0.326 obviously arise from the particular proof given. The authors conjective that for fixed \(c\) and \(q\) large enough the only noncollinear \((q+1)\)- sets \(B\) with more than \(cq\) nuclei in a conic are those of Fisher type. The main ideas include the ``lifting'' technique from \(AG(2,q)\) to \(AG(3,q)\) of the first author and \textit{F. Mazzocca} [Advances in Finite Geometries and Designs, 27-35 (1991; Zbl 0736.51006)] and some fundamental results on nuclei found in the paper of Segre and Korchmáros mentioned above.