The Airy hypothesis of isostatic compensation is very useful in accounting for structural differences between oceans and the continental margins. However, within the continents the compensation mechanism becomes more complicated. The thickness of the crust under much of the plains in the United States and Canada is between 40 and 55 kilometers. Determinations of crustal thickness under the Rocky Mountains gives results between 30 and 50 kilometers. Although local mountain ranges may have small roots, the Cordilleran region does not have a crustal root when compared to the plains. It follows that a modified form of Pratt's hypothesis of isostasy must be applied to continental regions. The density of the upper mantle is then different under the plains from what it is under the Rocky Mountains. In the plains it appears that there is a broad conformity of the Precambrian basement surface and the Mohorovicic discontinuity. Therefore the cause of epeirogenesis must lie within the upper mantle, possibly at the level of Gutenberg's low-velocity layer. The crustal studies in the plains and mountains indicate that more consideration should be given to gravitational gliding tectonics in the development of the Rocky Mountain system, since it is possible that there was substantial vertical motion of large crustal blocks.