Numerical models (sleep, 1996,1997) of mantle plumes that consider variations in lithospheric thicknesss suggest that deep cratonic roots influence the flow of hot, bouyant plume material. This process may explain the longlevity of cold, thick cratonic roots and the generation of kimberlites in crtons. We will use these methods to: a) predict the temporal and spatial distribution of adiabatic decompression melting and b) track the diamond/graphite stability field beneath and along the margins of the supeior craton, the world's largest, and the small tanzania craton, working in collaboration with n. Sleep. Initial lithospheric rhickness will be constrained by new and existing seismic, as well as zenolith, gravity, and heat flow data. These results will improve models for mineral exploration, as well as plume processes.