Projective geometries in dense matroids (Q2519009)

If \(M\) is a matroid, let \(\varepsilon(M)\) denote its number of \textit{points}. The main result of the paper under review is the following theorem: {\parindent=6mm \begin{itemize} \item{}Let \(\mathcal M\) be a minor--closed class of matroid. Then either \begin{itemize} \item[(1)]\(\varepsilon(M) \leq r(M){}^{c}{}_{\mathcal M}\) for each \(M\in \mathcal M\), \item[(2)]there is a prime--power \(q\) such that \(\varepsilon(M)\leq c_{\mathcal M} q^{r(M)}\) for each \(M\in \mathcal M\), and \(\mathcal M\) contains all \({\mathrm GF}(q)\)--representble matroids, or \item[(3)]\(\mathcal M\) contains arbitrarily long lines. \end{itemize} \item{}Where \(c_{\mathcal M}\) is an integer constant depending on \(\mathcal M\). \end{itemize}} This result, is motivated by the following conjecture [\textit{J. P. S. Kung}, Extremal matroid theory. N. Robertson, Graph structure theory. Proceedings of the AMS-IMS-SIAM joint summer research conference on graph minors held June 22 to July 5, 1991 at the University of Washington, Seattle, WA (USA). Providence, RI: American Mathematical Society. Contemp. Math. 147, 21-61 (1993; Zbl 0791.05018)] which reads as follows. {\parindent=6mm \begin{itemize} \item{}Conjecture (\textit{Kung's Growth rate Conjecture}) Let \(\mathcal M\) be a minor--closed class of matroid. Then either \begin{itemize} \item[(1)]\(\varepsilon(M) \leq {c_{\mathcal M}} r(M)\) for each \(M\in \mathcal M\), \item[(2)]\(\varepsilon(M) \leq {c_{\mathcal M}} r(M)^2\) for each \(M\in \mathcal M\) and \(\mathcal M\) contains all graphic matroids, \item[(3)]there is a prime--power \(q\) such that \(\varepsilon(M)\leq \frac{q^{r(M)-1}}{q - 1}\) for each \(M\in \mathcal M\) with sufficiently high rank, and \(\mathcal M\) contains all \({\mathrm GF}(q)\)--representble matroids, or \item[(4)]\(\mathcal M\) contains arbitrarily long lines. \end{itemize} \end{itemize}}