The origin of low-angle normal faults or detachment faults mantling metamorphic core complexes in the southwestern United States remains controversial. If σ1 is vertical during extension, the formation of, or even slip along, such low-angle normal faults is mechanically implausible. No records exist of earthquakes on low-angle normal faults in areas currently undergoing continental extension, except from an area of actively forming core complexes in the Solomon Sea, Papua New Guinea. In light of such geophysical and mechanical arguments, W. R. Buck and B. Wernicke and G. J. Axen proposed models in which detachment faults originate as high-angle normal faults, but rotate to low angles and become inactive as extension proceeds. These models are inconsistent with critical field relations in several core complexes. The Rawhide fault, an areally extensive detachment fault in western Arizona, propagated at close to its present subhorizontal orientation late in the Tertiary extension of the region. Neither the Wernicke and Axen nor Buck models predict such behavior; in fact, both models preclude the operation of low-angle normal faults. We must seek alternative explanations or modify existing models to explain the evidence that detachment faults form and operate with gentle dips.