To assess the alteration rate of silicate glasses in natural hydrothermal conditions, experimental kinetic measurements were performed at 150, 200, and 300-degrees-C in both batch and mixed-flow reactors. A Li-spiked synthetic basalt glass was reacted with silica solutions of various concentrations. Initial dissolution rates were calculated from Li release rates and compared with those obtained from long-term dissolution experiments using silica-free solutions. Some experiments were conducted using Al or Na doped solutions. Under neutral pH and for all temperatures, the presence of silica in solution lowered the initial dissolution rate as much as a factor 20. Under acidic conditions (0.1 M and 0.01 M HCl) the dependence of the initial dissolution rate on the aqueous silica concentration changed with the temperature. A weak silica dependence was observed at 300-degrees-C, whereas at 150-degrees-C, the presence of silica in solution had no effect on the measured rates, which became dependent on the dissolved Al concentration. These data suggest that the glass dissolution is dominated by the speciation and charge density of the glass surface: = SiOH at neutral pH and charged species (= AlOH2+) at low pH. At 300-degrees-C and under acidic conditions, the initial dissolution rate does not reflect the bulk pH of the solution; the solution pH at the glass-water interface is buffered at nearly neutral values by the hydrolysis of Al and Fe oxides. Experiments conducted in silica-free solutions over long durations exhibit a decrease of dissolution rates with time. Eventually, these rates approach the initial rate measured at equilibrium with amorphous silica. This reflects the amorphous state of the residual leached layer which develops at the glass surface, which itself buffers the aqueous silica concentration near the reaction interface to values close to the amorphous silica solubility. Thus, although the hydrothermal alteration rate of basalt glasses is controlled by the diffusion of aqueous species (silica, protons) through the alteration products, the measurement of initial dissolution rates in solution saturated with amorphous silica can provide accurate data for ''long-term'' modelling. As an application, the timing of sealing of hydrothermal fractures, resulting from wall-rock alteration, is calculated as a function of temperature and thickness of the fractures.
