The main objective is to contribute to the understanding of the dynamics of differentiation in large magma chambers. We will evaluate the importance of crystal settling in the formation of a chemically stratified solid. A major aim is to provide a physical model of crystal nucleation and growth that relates the thermal and chemical history of the chamber to the final crystal size. The novelty will be to test the results of numerical models with laboratory experiments which will consist of cooling a solution from above to generate crystals that can settle in a well-controlled thermal regime. We will evaluate the variation of the crystal size in different regimes as a function of the temperature and the liquid composition. We hope to provide a flexible numerical model applicable to geologically realistic boundary conditions, and capable of relating the thermal history of magma chambers to field data on crystal sizes.