Some geologists believe that exfoliation results when rocks formed at depth are exposed at the ground surface; the previous compressional forces would decrease and thus allow the rock to expand by fracturing parallel to the surface. Quite often, however, the fractures are not parallel to the ground surface, and this circumstance is taken as an indication of some other method of formation. Large daily variations in temperature, especially pronounced in deserts, were also credited with producing exfoliation—expansion from heating during the day followed by contraction from rapid cooling at night was thought to cause the separation of thin slabs from large blocks of rock at the surface. This view has been discredited by careful experiments using an electric heating and cooling device; thousands of alternations between temperatures considerably higher and lower than those measured in deserts have failed to produce in samples of rock any fractures detectable even under high magnification.

Study of thin shells that separate from rock exposed to the weather reveals as a common cause of the separation the slow development of clay minerals, which involves an increase in volume. The outer surface of exposed rock dries rapidly after wetting; but moisture that penetrates into minor crevices stays until some decay is started, and the resultant swelling causes flaking roughly parallel to the outer rock surface.

A small-scale form of exfoliation, called spheroidal weathering, is restricted to boulder-sized rock material and may occur at some depth within the Earth. In this case, rounded boulders are found surrounded by layers of disintegrated material.