Final Report Summary - MELTARC (Modelling and geochemical implications of intra-arc melting: consequences for the composition of the continental crust) The continental crust (CC) on which humans are living, is a major reservoir of mineral and organic resources necessary for social development. How continental crust grew and reached its present day composition is a question that still puzzle geologists. Volcanic arcs forming in subduction zones are thought to be the main building blocks of CC. Recent studies however pointed out that volcanic arcs have not the same composition that the CC. We are able to access the superficial section of volcanic arcs, those who are exposed at surface, and these are relatively well-known. The structure of arc roots, lying down to 40 km below sea level is not accessible except indirectly with from seismic studies, but they are a potential key to link the arcs to the continental crust. In the MELARC project, we proposed to study the deep roots of arcs where the crust to mantle transition is exposed. Some complexes representing arc sections now sandwiched in orogenic belts (Kohistan arc in Himalaya; Amalaoulaou complex in West Africa) have been accreted on the continental crust and we can sample them and observe their structure.We analysed the composition of rock forming the arcs roots and measured their physical properties such as the density and the seismic velocities. Our findings allow a better understanding of deep arc rocks and allow the possibility to link the results of remote geophysical prospection on active arcs and the nature of rocks forming the crust to mantle transition. We found that several petrogenetic processes are involved in the formation of arc roots, each process leading to a modification of arc roots. Percolation of basaltic magma in the mantle below arcs produce pyroxenites that are stable compared to their host rocks. In the crust, however, repeated intrusions of magma lead to the production of dense pyroxenites and leads to melting in the deep arc roots. These processes are occurring at 30 to 40 km depth and they induced changes in the bulk density of the arc root. Despite an increase in the density, the distribution of the dense pyroxenites does not favour their gravitational instability and their recycling in the upper most mantle. This mean that arcs can be considered as system where magmatic inputs are high but where outputs from crust to mantle are probably low. Active arcs cannot be considered as an equivalent of bulk continental crust which was created and evolved over more than 3.5 billon years. Numerical models will confirm if arc roots are gravitationally stable and if no, what are the output fluxes associated with gravitational instabilities.