Gravitational collapse has been suggested as the major cause of Cenozoic extension in the North American Cordillera and many other orogenic belts. Although both crustal thickening and mantle upwelling may have contributed to the Cordilleran extension, previous models of gravitational collapse have focused on the former; the cause of mantle upwelling and its relationship to crustal collapse remain obscure. Here we attempt to address the question of whether gravitational collapse of an overthickened crust could induce major mantle upwelling and whole-lithosphere extension. Thermal-rheological calculations indicate that crustal collapse may decouple from the mantle lithosphere, because the extensional forces arising from an overthickened crust are limited to the crust, while the rheology of continental lithosphere is intrinsically stratified. Even when the mantle lithosphere is mechanically coupled to the crust, thermomechanical modeling indicates that strain is localized in the weak lower crust during crustal collapse, and no significant (<10 km) thinning of the mantle lithosphere may be induced at the absence of extensional forces from plate boundaries. Crustal collapse of the Sevier-Laramide orogen seems adequate to account for much of the mid-Tertiary extension in the Cordillera, including formation of many core complexes, but it is unlikely to have been the major cause of the more recent basin-and-range extension. We suggest that a strong pulse of mantle upwelling in the mid-Tertiary, as indicated by the ignimbrite flare-up, may have triggered basin-and-range extension by weakening the lithosphere and providing excess gravitational potential energy. The cause of mantle upwelling remains uncertain, but the continued extension and volcanism since mid-Miocene in the northern Basin and Range province favor an active mantle upwelling with internal convective heating. ¿ 1998 American Geophysical Union