Potential density of rational points on the variety of lines of a cubic fourfold (Q955159)

The main object of the paper under review is the variety \(F\) of lines on a cubic fourfold \(X\) defined over a number field \(k\). The main result (Theorem 0.1) states that in many cases \(F\) satisfies the potential density property, i.e., the set of rational points \(F(k')\) is dense in \(F\) for some finite extension \(k'\) of \(k\). The authors give precise conditions under which the conclusion of the main theorem holds. The first one is that the geometric Picard group of \(F\) is cyclic. The second one is formulated in terms of a two-dimensional family of (singular) surfaces \((\Sigma_b)_{b\in B}\) covering \(F\) [cf. \textit{C.~Voisin}, Current developments in mathematics, 2003. Somerville, MA: International Press. 53--125 (2003; Zbl 1215.32014)]: for some closed point \(b\) in the moduli space \(B\) which is defined over a number field, the abelian surface \(A_b\) defined as the Albanese variety of some desingularization \(\Sigma'_b\) of \(\Sigma_b\) is assumed to have geometric Néron--Severi group equal to \(\mathbb Z\). The authors show that the above two conditions are indeed satisfied for many cubic fourfolds \(X\). The proof is based on dynamic properties of the rational self-map \(\varphi \colon F \dashrightarrow F\) constructed by the second author in [The Fano conference. Papers of the conference organized to commemorate the 50th anniversary of the death of Gino Fano (1871--1952), Torino, Italy, September 29--October 5, 2002. Torino: Università di Torino, Dipartimento di Matematica. 761--792 (2004; Zbl 1177.14040)]. The argument of \textit{E. Amerik} and \textit{F.~Campana} [Pure Appl. Math. Q. 4, 509--545 (2008; Zbl 1143.14035)], showing that the orbit of a general complex point of \(F\) is dense in \(F\), is not applicable because it requires throwing away too many ``bad'' countable unions of proper subvarieties making it impossible to find a point in \(F(\overline{\mathbb Q})\) with Zariski dense orbit in such a way. The authors circumvent this difficulty with the help of ingenious arguments proving that there exists a surface \(\Sigma_b\) as above whose orbit is Zariski dense in \(F\). They first prove that there are many \(\Sigma_b\)'s with infinite orbit (i.e., not preperiodic) by showing, using subtle geometric considerations, that the orbit contains a divisor. Second, they exclude the case where the Zariski closure of the union of surfaces in the orbit is of dimension three. This step also requires involved techniques from the geometry of threefolds.