The rate of mass transfer was measured for solid metal shapes dissolving into mercury at room temperature. Sherwood numbers for horizontal tin, cadmium, zinc, and lead cylinders dissolving by natural convection agreed with Nusselt numbers for heat transfer in nonmetallic liquids at the same Rayleigh (Grashof X Randtl) numbers. Dissolving of zinc tubes by mercury flowing turbulently within them qgreed with heat transfer to nonmetals in tubes. Dissolving of random beds of lead spheres by mercury flowing through the bed agreed with similar nonmetal systems. It is concluded that mass transfer processes in liquid metals follow substantially the correlations for other fluids in heat or mass transfer, which with moderate safety factors may thus be used for at least preliminary design purposes.Cases of mass transfer in liquid metals are becoming more frequent with the increasing popularity of liquid metals as heat transfer media and as solvents. The pyrometallurgical refining of metals has long einployed mass transfer between the melted metal and a molten flux, another molten metal, a solid, or a gas. Corrosion of pipes and vessels by contaihed liquid metals due to temperature differentials plus a variation of solubility with temperature has long been encountered. The purification of molten sodium by the freezing out of sodium oxide is a similar case. E'inally, liquid-metal fuel reactors (L.M.F.R.) have been proposed (20) which feature mass transfer to accomplish principal objectives. Replenishment of the fissionable element would involve mass transfer, which, however, would be applied principally in the continuous removal of radioactive and neutron-consuming fission products.In the design of such processes, or in predicting the effect on performance of changes in operating conditions, it is desirable to have available the appropriate dimensionless correlations of mass transfer for each proposed geometry. A few such correlations from studies on nonmetallic liquids have been published. However, before one could confidently employ such correlations for liquid metals it would seem desirable to establish their validity for liquid metals in at least several key geometries. This paper presents the results of such a study employing four binary systems and three rigid geometries. The solids were lead, tin, cadmium, and zinc, and the liquid phase was mercury at room temperature. Natural convection was studied by dissolving horizontal cyliiiders of all four metals in a stationary pool of mercury, the Grashof X Schmidt number product ranging from 107 to 109. Forced convection was studied by passing mercury through zinc pipes at Reynolds numbers of 1,200 to 12,000 and through fixed beds of randomly packed lead spheres at void Reynolds numbers of 20 to 1,000. In all cases the Schmidt number was of the order of magnitude of 100.The ditrerential equations for forcedconvection mass transfer with constant physical properties (dilute solutions) are identical with those for heat transfer, provided that the Nusselt and Prandtl numbers in...