The trisecant conjecture (Q1914735)

Let \((X, \Theta)\) be an indecomposable principally polarized abelian variety of dimension \(g\) over an algebraically closed field \(k\) of characteristic different from 2. Let \(K(X) \subset |2\Theta |\simeq \mathbb{P}^N\) be the Kummer variety associated with \((X, \Theta)\). \textit{G. E. Welters} conjectured [``A criterion for Jacobi varieties'', Ann. Math., II. Ser. 120, 497-504 (1984; Zbl 0574.14027)]\ that the existence of a trisecant line to \(K(X)\), i.e., a line in the projective space \(\mathbb{P}^N\) that meets \(K(X)\) in at least three distinct points, characterizes the Jacobians of algebraic smooth curves among all indecomposable principally polarized abelian varieties (known as the trisecant conjecture). We prove that there exist trisecant lines to the Kummer variety of every Prym variety associated with a double étale cover of smooth connected complete curves. Then, we deduce that the trisecant conjecture is false.