Lithospheric buoyancy, the product of lithospheric density and thickness, is an important physical property that influences both the long-term stability of continents and their state of stress. We have determined lithospheric buoyancy by applying the simple isostatic model of Lachenbruch and Morgan (1990). We determine the crustal portion of lithospheric buoyancy using the USGS global database of more than 1700 crustal structure determinations (Mooney et al., 2002), which demonstrates that a simple relationship between crustal thickness and surface elevation does not exist. In fact, major regions of the crust at or near sea level (0-200 m elevation) have crustal thicknesses that vary between 25 and 55 km. Predicted elevations due to the crustal component of buoyancy in the model exceed observed elevations in nearly all cases (97% of the data), consistent with the existence of a cool lithospheric mantle lid that is denser than the asthenosphere on which it floats. The difference between the observed and predicted crustal elevation is assumed to be equal to the decrease in elevation produced by the negative buoyancy of the mantle lid. Mantle lid thickness was first estimated from the mantle buoyancy and a mean lid density computed using a basal crust temperature determined from extrapolation of surface heat flow, assuming a linear thermal gradient in the mantle lid. The resulting values of total lithosphere thickness are in good agreement with thicknesses estimated from seismic data, except beneath cratonic regions where they are only 40-60% of the typical estimates (200-350 km) derived from seismic data. This inconsistency is compatible with petrologic data and tomography and geoid analyses that have suggested that cratonic mantle lids are ??? 1% less dense than mantle lids elsewhere. By lowering the thermally determined mean mantle lid density in cratons by 1%, our model reproduces the observed 200-350+ km cratonic lithospheric thickness. We then computed gravitational potential energy by taking a vertical integral over the computed lithosphere density. Our computed values suggest that the thick roots beneath cratons lead to strong negative potential energy differences relative to surrounding regions, and hence exert compressive stresses superimposed on the intraplate stresses derived from plate boundary forces. Forces related to this lithosphere structure thus may explain the dominance of reverse-faulting earthquakes in cratons. Areas of high elevation and a thin mantle lid (e.g., western U.S. Basin and Range, East African rift, and Baikal rift) are predicted to be in extension, consistent with the observed stress regime in these areas.