Abstract:
Experimental and natural investigations of partially molten granite are compared and reviewed. Experiments suggest that deformation of partially molten granite with low-viscosity melt (104 Pa s) exhibits a rheological critical melt percentage (RCMP). In case of high viscosity melts (108 Pa s), however, the relationship between melt fraction and log viscosity of the partially molten granite may be linear. Considerations about viscosity, rheological thresholds, and segregation of natural melts suggest that low-viscosity melt experiments simulate natural conditions more realistically. Therefore, an RCMP is to be expected under natural conditions. Both diffusion creep and dislocation creep may occur under natural conditions, whereas cataclastic flow is only observed under experimental conditions. A melt-induced transition from dislocation creep to diffusion creep occurs under experimental and natural conditions. Melt topology is controlled by the magnitude of differential stress under experimental conditions. If differential stress is higher than ~100-150 MPa, melt pockets are elongate and oriented at a low angle to the maximum compressive stress. In contrast, in nature, melt pockets tend to be oriented subparallel to the foliation plane, i.e., presumably at a high angle to the maximum compressive stress.