Title: On vanishing of Abelian integrals and contraction of curves

Abstract: In this article we consider a smooth curve $C$ of genus at least two. We prove that if the integral of a differential 1-form of the first kind $\omega$ on a topological cycle $\gamma$ of $C$ is zero and both $C$ and $\omega$ are defined over a number field, then there must be a non-trivial contraction of $C$ to another curve $D$ of smaller positive genus such that the topological cycle $\gamma$ is mapped to zero under this contraction and $\omega$ is the pull-back of some differential form on $D$. We give an upper bound for the degree of the contraction and for the degree of the field of definition of $D$ in terms of some numerical invariants. The basic tool of the proof is the abelian Subvariety Theorem. This ensures the existence of an abelian subvariety $B$ of a given abelian variety $A$ and of bounded degree, under the condition that there exists a proper subspace of the tangent space of $A$ at zero which contains a period of $A$. The fact that a curve can be embedded in its Jacobian, provides the link between the two situations.