Remarks on the Segre cubic (Q1423772)

A set \(s\) of six distinct points on \(\mathbb{P}^1\) determines a curve \(C\) of genus 2 and the Kummer surface \(J(C)/ \langle\pm 1\rangle\). The bi-canonical map maps \(C\) onto a conic and the linear system of plane cubics passing through the six branch points maps \(\mathbb{P}^2\) into a nodal cubic surface \(V_s\subset\mathbb{P}^3\). The Segre cubic \({\mathcal S}_3\) is a \(S_6\)-symmetric compactification of the configuration space of six points on \(\mathbb{P}^1\) [\textit{I. Dolgachev} and \textit{D. Ortland}, Point sets in projective spaces and theta functions, Asterisque 165 (1985; Zbl 0685.14029)]. In the paper under review the author proves the following theorem: Let \(s\) be a general point on \({\mathcal S}_3\) (i.e. the corresponding six points on \(\mathbb{P}^1\) are distinct) and let \(F(t_0,\dots,t_4)\) be the defining equation of \({\mathcal S}_3\) and \(Q_s{\mathcal S}_3\) be the quadratic hypersurface defined by \((t_0,\dots,t_4) H^t(t_0,\dots,t_4)\) where \(H=(\frac{\partial^2 F} {\partial t_i\partial t_j})_s\), then the following holds: 1) the tangent space at \(s\) cuts out \(V_s\) 2) the intersection \(W_s={\mathcal S}_3\cap Q_s{\mathcal S}_3\) is birationally equivalent to the Kummer surface corresponding to \(s\), 3) there is a 5-dimensional variety \(\Gamma\subset{\mathcal S}_3 \times {\mathcal S}_3\) such that \(\tau_1^{-1}(s)=V_s\) and \(\pi_2^{-1}(s) =W_s\) where \(\pi_1\) (resp. \(\pi_2): \Gamma\to {\mathcal S}_3\) is the first (resp. the second) projection.