"Arcs are expected to be the fundamental brick of the Continental Crust (CC). The main problem with this hypothesis is that the bulk CC has an andesitic composition while most arcs are formed by successive inputs of basaltic magmas.If island arcs accretion is indeed the main mechanism of crustal growth, one must explain how arcs can reach an intermediate composition and if these processes are effective in the nature.Intra-arc differentiation takes place in their lower crustal section. Removal of magmatic cumulates is not a sufficient process to drive the bulk composition of island arcs towards intermediate ones. Because melting and dehydration lead to the production of dense iron-rich melting and/or dehydration residues, it could causes delamination of mafic-ultramafic rocks into the mantle due to gravitational instability in the lowermost arc crust. Melting at the basis of arcs is most probably a necessary conditions to evolve towards andesitic bulk composition.The mechanisms of arc differentiation and their consequences are still poorly understood; this is due to the scarcity of exhumed section of arcs that have preserved evidences for lower crustal melting and/or dehydration. This project will focus on selected samples from well-constrained Precambrian and Phanerozoic exhumed arc sections. The main objectives are to: (i) quantify the geochemical differentiation induced during partial melting by directly analysing and modelling the composition of melts and residues sampled in exhumed arc roots; (ii) model the P-T-X conditions of intra-arc differentiation and to establish the mechanisms responsible for such a process; (iii) evaluate the causes and consequences of melting on the bulk structure, composition and stability of arcs by numerical modelling.The results of this study will be used to evaluate if delamination and foundering of dense residues at the base of arcs is a viable mechanism to change their bulk composition towards the one of continental crust."