On circuit exchange properties for matroids (Q1823955)

The starting point for this paper was the following result proved by \textit{J.-C. Fournier} in [J. Comb. Theory, Ser. B 31, 249-250 (1981; Zbl 0469.05024)] stating that a matroid M is binary if and only if, whenever \(C_ 1\) and \(C_ 2\) are distinct circuits of M and x and y are elements of \(C_ 1\cap C_ 2\), then the set \((C_ 1\cup C_ 2)-\{x,y\}\) contains a circuit of M. In this paper under review, this result is extended in several directions, and the following circuit exchange property is defined and studied. Given two positive integers n and k with \(n\geq 2\), a matroid M has the (n,k)-exchange property for circuits, if, whenever \(C_ 1,C_ 2,...,C_ n\) are distinct circuits of M and \(x_ 1,x_ 2,...,x_ k\) are distinct elements of \(C_ 1\cap C_ 2\cap...\cap C_ n,\) then there is a circuit contained in \((C_ 1\cup C_ 2\cup...\cup C_ n)-\{x_ 1,x_ 2,...,x_ k\}.\) The author's main aim was to give excluded-minor characterizations of matroids having these properties. More precisely, in Section 2 these characterizations are given for the matroids having (2,k)-exchange properties for \(k=3,4\) and 5. Similar characterization is given for matroids with the (n,2)- exchange property in Section 4, and weakening the (3,3)-exchange property a new characterization of when a 3-connected matroid is binary is given in Section 5. The conjecture stating that, if \(k\geq 2\), a matroid with the (2,k)-exchange property has the \((2,k+t)\)-exchange property for all \(t\geq 0\) is discussed in Section 3. From the results given in Section 2 it follows, that the conjecture is true for \(k=2,3\) and 4, and the theorem establishing this conjecture for \(k=5\) is stated without proof, as well. For details the reader should consult the paper itself.