The center of the generic division algebra (Q755856)

The ring \(G_{m,n}\) of m generic \(n\times n\) matrices over some algebraically closed field k of characteristic 0 is well-known to possess a classical division algebra of fractions \(D_{m,n}\). Since the center \(K_{m,n}\) of \(D_{m,n}\) is known to be purely transcendental over \(K_{2,n}\), the problem of (dis)proving that \(K_{m,n}\) is purely transcendental over k reduces to the case \(m=2.\) The author presents the following geometric interpretation of \(K_{2,n}\). For any pair of positive integers (n,d), denote by \(Q_{n,d}\) the variety parametrizing the couples (C,D), where \(C\subseteq {\mathbb{P}}^ 2_ k\) is a curve of degree n and where D is a divisor of degree d of C \((Q_{n,d}\) is essentially the Picard scheme over a generic plane curve). The main result of the paper then says that \(K_{2,n}\) is the function field of \(Q_{n,n(n-1)/2}\). As an application, it is then shown that \(K_{2,3}\) is rational, by proving that the variety \(Q_{3,3}\) is rational. (The rationality of \(K_{2,n}\) has been proved by E. Formanek for \(n=2,3,4\).)