Volcanic eruptions may have serious human and economic consequences (e.g., ash clouds, large cities close to active volcanic systems). The eruptive styles and mechanisms are mainly dependent on degassing processes in magmas. The main dissolved gases are H2O and CO2 and understanding the solubility of mixed volatiles, such as H2O and CO2, in silicate melts as well as the partitioning of volatiles between melts and fluids is of particular importance to model magmatic and volcanic processes. The available models recently developed fail in reproducing H2O and CO2 solubilites for a variety of natural compositions, mainly because there is no data to model accurately compositional parameters (e.g., role of alkalis, alkaline Earth, NBO/T).The first goal of this project is the determination of the role of compositional parameters on the distribution of volatiles, H2O and CO2, between silicate melts and coexisting gas phase at highpressure and high temperature. Starting from a base composition corresponding to a basalt, parameters such as K2O /(Na2O+ MgO+CaO), CaO /(Na2O+ K2O+ MgO) ratios, and NBO/T will be modified systematically in the investigated melt composition. The H2O and CO2 solubility and partitioning data will be used to improve the existing H2O and CO2 solubility models (e.g., Papale et al. 2006). The results will also have a particular impact for the interpretation of volatile concentration in natural glass inclusions.In a second part, complementary to the volatile solubility investigations, the kinetics of degassing will be studied in isobaric isothermal experiments as well as in experiments with fixed decompression rates. Particular attention will be given to the nucleation and growth of bubbles. Innovative techniqes (automatic processing of microtomographic data) will be used to analyse the experimental products. The experiments will be designed to model degassing processes during eruptions from the magma chamber to the eruptive column.