After examining 15 reliable estimates of the elastic thickness of the lithosphere where it underthrusts mountain belts in Europe, Asia, and the Americas, we conclude that the strength of continental lithosphere as a function of depth must be very different from that of oceanic lithosphere. Depending on the state of stress, continental lithosphere may be either very much stronger or very much weaker than would be predicted for oceanic lithosphere of the same thermal age.

We explain this observation as the result of two competing effects: a thicker thermal plate beneath the continents which allows a deeper elastic/ductile transition for plates more than 100 million years old, and a low activation energy for ductile creep in the continental crust which leads to massive failure of the upper elastic lithosphere at high bending stresses. The 100+-km values for elastic plate thickness of billion-year-old lithosphere require a thermal plate thickness of at least 250 km. The magnitude of apparent thinning of the elastic plate at high stress is consistent with that calculated from a simple yield envelope which includes a ductile zone in the lower crust. The low yield strength of continental materials leads to correlations between effective elastic plate thickness and all of the following: dip of the underthrust plate; radius of curvature of the thrust front in map view; and total length of the mountain belt. Using insight derived from laboratory experiments on the buckling of spherical shells and drawing analogies with the geophysical characteristics of oceanic subduction zones, we arrive at the following sequence of causes and effects. The primary variable is the dip of the downgoing plate, presumably prescribed by its own excess mass and pressure exerted by dynamic flow in the mantle. Where this dip is steep, the rigidity of the plate is weakened due to inelastic yielding. The weakened plate can then buckle at small wavelengths into short, arcuate mountain belts. Thus we believe that the mantle ultimately controls the pattern of surface tectonics, but that the nonelastic rheology of the lithosphere is the mechanism by which that control is imposed. ¿ American Geophysical Union 1988