Two-descent on some genus two curves (Q2040156)

Consider the hyperelliptic curves of genus two, varying \(p\) a prime, \[C_p:y^2=x(x^2-p^2)(x^2-4p^2),\] which are a quadratic twist of \(y^2=x(x^2-1)(x^2-4)\) over \(\mathbb{Q}(\sqrt{p})\) and \(\mathbb{Q}(\sqrt{-p})\). The authors of the paper under review investigate the arithmetic of \(C_p\) over \(\mathbb{Q}\) in terms of \(p\), in particular the rank of the Jacobian \(J_p\) of \(C_p\). The results of the paper under review are mainly the Master thesis of two of the authors supervised by the third author of the paper, corresponding to [\textit{T. Evink}, Two descent on hyperelliptic curves of genus 2. Groningen: University of Groningen (Master Thesis) (2020), \url{http://fse.studenttheses.ub.rug.nl/21636}; \textit{G.-J. van der Heiden}, Computing the 2-descent over \(\mathbb{Q}\) for curves of genus 2. Groningen: University of Groningen (Master Thesis) (1998), \url{http://fse.studenttheses.ub.rug.nl/8657}]. In the paper under review, first, the authors compute the \(\mathbb{F}_2\)-dimension of the \(2\)-Selmer group of \(J_p\) over \(\mathbb{Q}\), denoted by \(S^2(J_p/\mathbb{Q})\), by explicit 2-descent, method described in [\textit{E. F. Schaefer}, J. Number Theory 51, No. 2, 219--232 (1995; Zbl 0832.14016)]. The result is presented in terms as \(p\) congruent to 1,5,7,11,13,17,19,23 modulo 24, and next they study the family of \(C_p\) with \(p\) congruent to a fix of such odd primes modulo 24. From the usual sequence: \[0\rightarrow J_p(\mathbb{Q})/2J_p(\mathbb{Q})\rightarrow S^2(J_p/\mathbb{Q})\rightarrow \text{Ш}(J_p/\mathbb{Q})[2]\rightarrow 0,\] where \(\text{Ш}(J_p/\mathbb{Q})[2]\) is the 2-torsion of the Tate-Shafarevich group associated to \(J_p\) over \(\mathbb{Q}\), (and because all Weierstrass points on \(C_p\) are \(\mathbb{Q}\)-rational), they prove that the torsion group of \(J_p(\mathbb{Q})\) corresponds to the group generated by the Weierstrass points. And deduce that \[\mathrm{rank}(J_p(\mathbb{Q}))+\mathrm{dim}_{\mathbb{F}_2}\text{Ш}(J_p/\mathbb{Q})[2]=\mathrm{dim}_{\mathbb{F}_2}S^2(J_p/\mathbb{Q})-4.\] Next, the authors are interested to obtain the rank of \(J_p(\mathbb{Q})\), and the dimension of the two torsion of the Tate-Shafarevich group. To attack such questions, because the curves \(C_p\) are twists of the hyperelliptic curve \(C:y^2=x(x^2-1)(x^2-4)\), the authors observe the relation \[\mathrm{rank} J_p(\mathbb{Q})+\mathrm{rank}\, \mathrm{Jac}(C)(\mathbb{Q})=\mathrm{rank} \,\mathrm{Jac}(C)(\mathbb{Q}(\sqrt{\pm p})),\] for both possibilities of the sign \(\pm\). By a computation of the authors in the paper under review they obtain that \(\mathrm{rank}\,(\mathrm{Jac}(C)(\mathbb{Q}))=0\), and it follows that \[\mathrm{rank}J_p(\mathbb{Q})\leq \mathrm{dim}_{\mathbb{F}_2} S^2(\mathrm{Jac}(C)/\mathbb{Q}(\sqrt{\pm p}))-4.\] Next, the authors compute the dimension of \(S^2(\mathrm{Jac}(C)/\mathbb{Q}(\sqrt{\pm p}))\), computation inspired by [Evink, loc. cit.] but using Rédei symbols (recalled in the paper) instead of different reciprocity arguments in [Evink, loc. cit.]. Finally, the \(\mathbb{Q}\)-rationals points of \(C_p\) are briefly discussed, in particular the authors prove in the paper that the set \(C_{241}(\mathbb{Q})\) is the expected set, (where \(\mathrm{rank}\,\mathrm{Jac}(C_{241})(\mathbb{Q})=2\), so the standard Chabauty method does not apply to verify the computational expectation).