Granite rheology: magma flow and melt migration

Granite magma rheology has long been studied by rock physicists, petrologists and experimentalists. New data on viscosity of dry and H(2)O-bearing melts and glasses in a wide range of temperatures and H(2)O content changed drastically the general view on the Arrhenian behaviour of granitic melts. Recent studies on the deformation of partially molten granite rocks at up to 50 vol% of melting show the failure of the RCMP (rheologically critical melt percentage) concept for granites. We present new direct measurements of melt migration from a combination of centrifuge experiments, falling sphere viscometry of partially molten granites assisted by a centrifugal field and viscoelastic behaviour of partially molten granites in torsion oscillation apparatus. The results suggest that the extrapolation of the Einstein-Roscoe equation for the viscosity of partially molten rocks does not give adequate results for granites with 50 vol% of melt phase. Rheology of granite magmas is very sensitive to chemical composition, and viscosity of the melt phase. Adding H(2)O to partially molten granites results in a decrease in the viscosity of the melt phase, as well as dissolution of quartz crystals. In turn, this produces more felsic and viscous melt. Thus, these factors smear out the contrast between partially and completely molten granites, and between anhydrous and H(2)O-bearing magmas at moderate H(2)O contents (1 to 2 wt%).