Some simple exploratory theoretical models of the thermal effects of core segragation have been investigated, assuming an initially homogeneous earth and including convective heat transport through a 'parametrized convection' approximation. The results indicate that either (1) mantle temperatures 30% or more above present values may have resulted from the gravitational energy released during core segregation, (2) the earth retained very little of its accretional energy, (3) core segregation lasted for 1 b.y. or more, or (4) the earth accreted heterogeneously. Option 3 seems to be precluded by terrestrial lead isotope data, and the alternatives each raise substantial questions concerning the mechanics, chemistry, and petrology of the earth's early history. There is no recognized evidence for the early hot phase of option 1, and option 4 implies, among other things, an analogous early hot phase. Although it has not been favored, option 2 may be viable.