A consequence of the relative Bogomolov conjecture (Q2235726)

In the paper under review, the authors give a precise formulation of the relative Bogomolov conjecture and show that such formulation completely answers a question posed by B. Mazur. More precisely, in [Bull. Am. Math. Soc., New Ser. 14, 207--259 (1986; Zbl 0593.14021)], \textit{B. Mazur} introduces the following conjecture. Conjecture 1 (Mazur). Let \(g\geq 2\) be an integer. Then there exists a constant \(c(g)\) with the following property. Let \(C\) be a smooth curve of genus \(g\) defined over a field \(F\) of characteristic \(0\); let \(P_0\) be a point in \(C(F)\) and \(\Gamma\) a subgroup of \(\mathrm{Jac}(C)(F)\) of finite rank \(r\). Using the Abel-Jacobi map based at \(P_0\), view \(C-P_0\) as a curve in \(\mathrm{Jac}(C)\). Then \[ \# (C(F)-P_0)\cap \Gamma \leq c(g)^{1+r}. \] In [Ann. of Math. (2) 194, No. 1, 237--298 (2021; \url{doi:10.4007/annals.2021.194.1.4})], the authors of the paper under review prove Conjecture 1 under the assumption that \(C\) has a modular height that is larger than a constant \(\delta(g)\) depending on the genus \(g=g(C)\). In this paper, they give the following precise formulation of the relative Bogomolov conjecture and then they show that it implies Conjecture 1. Let \(\pi:\mathcal{A}\to S\) be an abelian scheme dominating a quasi-projective variety \(S\) defined over \(\overline{\mathbb{Q}}\); let \(\mathcal{L}\) be a symmetric relatively ample line bunde on \(\mathcal{A}/S\), also defined over \(\overline{\mathbb{Q}}\). Then one can define the Néron-Tate height \(h\) associated to \(\mathcal{L}\) on \(\mathcal{A}(\overline{\mathbb{Q}})\). Let \(\eta\) denote the generic point of \(S\). Conjecture 2 (Relative Bogomolov conjecture). Let \(X\) be an irreducible subvariety of \(\mathcal{A}\) defined over \(\overline{\mathbb{Q}}\) that dominates \(S\). Assume that the fiber \(X_\eta\) above \(\eta\) is irreducible and not contained in any proper algebraic subgroup of \(\mathcal{A_\eta}\). If \(\textrm{codim}_\mathcal{A} X > \dim S\), then there exists a \(\epsilon>0\) such that the set \[ \{\, x\in X(\overline{\mathbb{Q}})\, : \, h(x)\leq \varepsilon \, \} \] is not Zariski dense in \(X\). Then they prove the following result. Main Theorem. Conjecture 2 implies Conjecture 1. The proof essentially relies on two steps. First, the authors reduce the case with \(F\) any field of characteristic \(0\) to the case with \(F=\overline{\mathbb{Q}}\). Then, they apply and replicate the arguments used in their previous work to prove the result over \(\overline{\mathbb{Q}}\). In addition, using \textit{S.-W. Zhang} [Ann. Math. (2) 147, No. 1, 159--165 (1998, Zbl 0991.11034)], they prove that any isotrivial abelian scheme satisfies Conjecture 2 (and hence Conjecture 1). \textit{L. Kühne} [``Equidistribution in families of abelian varieties and uniformity'', Preprint, \url{arXiv: 2101.10272}] shows that Conjecture 2 holds for any curve whose modular height is smaller than \(\delta(g)\). These are exactly the curves left out by the result in the authors' previous paper. Hence Kühne's result, the Main theorem and Theorem 1.2 in [the authors, loc. cit.] provide a complete proof of Conjecture~1.