The generalized Franchetta conjecture for some hyperkähler fourfolds (Q2279518)

The author continues his investigations on the properties of the Chow ring. His motivation here are Voisin's conjectures for hyperkähler varieties [\textit{C. Voisin}, Prog. Math. 315, 365--399 (2016; Zbl 1352.32010)]. The paper deals with the case of fourfolds, specifically with the Hilbert squares of \(K3\) surfaces of genus \(g\), with \( 2 \le g \le 8 , g \neq 7 \). Two main results are proved. The first says that, given a cycle \(\Gamma \) of codimension 2 in the Chow group of the universal family \( \mathcal X \to B \) of Hilbert squares, its restriction \(\Gamma_b \) belongs to \(A^2_{(0)}(X_b)\) for every fibre \(X_b\). Here \(\bigoplus_i \bigoplus_j A^i_{(j)}(X_b)\) is the splitting of the Chow ring of \textit{M. Shen} and \textit{C. Vial} [The Fourier transform for certain hyperkähler fourfolds. Providence, RI: American Mathematical Society (AMS) (2016; Zbl 1386.14025)]. The second result states that given a codimension 2 subvariety \(\Gamma \) on the universal family, if the restriction to a very general fibre is Lagrangian ( meaning that multiplication by the class of \(\Gamma_b \) kills \(H^{2,0} (X_b)\, \)), then, for every fibre, multiplication by \(\Gamma_b\) kills those elements in the Chow group of codimension \(2\) which are homologous to zero. The paper has a sequel [\textit{L. Fu} et al., J. Math. Pures Appl. (9) 130, 1--35 (2019; Zbl 1423.14033)], there the results above are extended to more general cases.