A general solution for the slope of a fusion curve is presented. It is not the Clausius-Clapeyron equation which is a special case for one-component systems. The connection between fusion curve slope and crystal/liquid buoyancy is obscure in multicomponent systems even though the slopes of fusion curves for any given composition are always well defined. The olivine flotation hypothesis at high pressure is poorly constrained by information derived from fusion curve analysis. Physical property measurements provide the best constraints on buoyancy questions. ¿ American Geophysical Union 1988