The topology of smooth divisors and the arithmetic of abelian varieties. (Q5954580)

The author investigates the following question of Fulton, motivated by the study of degeneracy loci: Can one determine the Betti (or, more generally, the Hodge) numbers of a smooth divisor \(D\) on a smooth projective variety \(X\) in terms of its cohomology class in \(H^2(X,\mathbb{Z})\)? Though this can be done in the case of ample divisors or \(b_1(X)=0\) by using Lefschetz principle and Riemann-Roch theorem (as the author shows in section four), the answer turns out to be ``no'' in general. However, the study of this problem leads to surprising insights into the topology of smooth divisors. The counterexample constructed by the author is a ruled surface with a smooth connected divisor which is homologous to a smooth disconnected divisor. However, it turns out that such examples are quite rare and can be characterized in the following way: They occur only if there is a surjective morphism from \(X\) to a curve with some multiple fibers, and the two divisors are both unions of fibers. In the course of proving this result, the author obtains that any smooth projective variety containing at least three pairwise disjoint effective divisors whose rational cohomology classes lie on a line in \(H^2(X,\mathbb{Q})\) has a unique fibration over a smooth curve such that the divisors are multiples of fibers. This result, which fails in the case of just two disjoint divisors, is certainly of its own interest. In the last part, the author gives an example of two connected smooth divisors which are homologous but have different Betti numbers. Motivated by this example, he conjectures that any smooth connected homologous divisors have cyclic étale coverings that are deformation equivalent, and proves a slightly weaker form if the Picard variety of \(X\) is isogenous to a product of elliptic curves. He shows that the general conjecure would follow from a well-known open problem in the arithmetic theory of abelian varieties, which states that if \(A\) is an abelian variety defined over a number field \(F\), then there are infinitely many primes \(\mathfrak p\) of the ring of integers \({\mathcal O}_F\) such that the finite group \(A({\mathcal O}_F/\mathfrak p)\) has order prime to the characteristic of the field \({\mathcal O}_F/\mathfrak p\).