A long-standing question is whether old oceanic lithosphere continues cooling as the boundary layer of a halfspace or approaches thermal equilibrium, as modeled by a finite thickness plate. Although the latter is the most direct inference from seafloor depths and heat flow, other explanations have been proposed. We investigate this issue using published results for the derivative of the oceanic geoid with age estimated from geoid offsets across fracture zones. Such data have not been used extensively in analyses of the thermal evolution of the lithosphere, primarily because they are inconsistent with two commonly used thermal models: a halfspace or a 125-km-thick plate. Recent studies, however, find that depth and heat flow data are better fit by a thinner (95 km) plate model. We thus compile published geoid slope results, and find that these data, though scattered, can discriminate between the models. Geoid slope changes with age, rather than being constant as predicted for a cooling halfspace. This variation is greater than predicted for a thick plate and is better fit by a thin plate. Geoid data should thus be useful for improving thermal models of the lithosphere. ¿ American Geophysical Union 1995