Current theories of surface segregation are based on either the difference in the sizes or the heats of sublimation of the alloy components. We show that neither of these theories gives adequate agreement with experiment. A simple rule, based on the melting curve of the alloy, gives good agreement with all available experimental data.

• Received 6 August 1976

DOI:https://doi.org/10.1103/PhysRevLett.37.1433