Subsidence and sink formation resulting from brining operations in the Windsor-Detroit area include the 1954 sink at the Canadian Salt Company brine field near Windsor, Ontario, and the 1971 sinks at the BASF Wyandotte Corporation brine field at Grosse Ile, Mich. Earlier investigations into both occurrences concluded that the mechanism of sink development consisted of the gradual stoping of poorly supported brine-gallery roof rock to the near surface with subsequent surface collapse. A more recent study attempted to describe the mechanism of sink development in terms of the geometry of a cylindrical chimney formed by stoping of roof rock, the height of a cavity at depth, the depth of overlying rock, and the bulking ratio of the rubble formed during stoping. Persons with extensive experience in solution mining in the Windsor-Detroit area have expressed doubt that the stoping mechanism could fully explain the development of these sinks. Further, they have proposed that the relatively shallow (300-ft-deep) Sylvania Sandstone, in this case, may be responsible for the sinks by a secondary undermining mechanism to be examined in this paper. The mechanism involves downwarping of the beds overlying the salt cavity and development of a shallower cavity in the Sylvania Sandstone by downward migration of cohesionless sand grains from the Sylvania through openings in the disturbed rock to the lower cavity. This study indicates that under natural conditions the Sylvania will not migrate, even in the presence of large underground water flows because the sandstone possesses some cohesion throughout its depth. However, further investigation has formulated a mechanism that could allow the Sylvania Sandstone to loose its cohesion in response to high horizontal stresses. These stresses could be the result of deformation that accompanies general subsidence and (or) of past geologic processes. Included in this study were experimental and analytical investigations. As determined by uniaxial and triaxial testing, the Sylvania Sandstone in the Detroit area has been shown to have low compressive strength. In addition, it exhibits an explosive type failure whereby over 50 percent of the sample is reduced to loose granular sand. As a result of these characteristics, the Sylvania Sandstone can loose its cohesion when subjected to high horizontal stresses. Efforts at mechanically modeling the Sylvania were made to account for the measurements and observations. Linear arch theory was used for an elastic analysis. Linear arch theory predicts two modes of failure: (1) arch crushing, a compressive failure of the upper portion of the arch due to compressive stresses exceeding the compressive strength of the material, and (2) arch collapse, a sagging of the beds due to compressive strains which reduce the arch line to a length less than the original arch length. The arch crushing mode of failure would then yield the loose granular sand as observed in laboratory testing. Arch collapse would simply result in bed sagging without granulation of the sandstone. Arch collapse is favored by thin-bedded material while arch crushing is favored by thick-bedded material. Arch crushing seems to be a likely mode of failure for the Windsor-Detroit sinks. It is believed that after a crushing failure the sand-water slurry (specific gravity 1.2) which exceeds the density of the cavity brine will migrate downward through cracks and open joints eventually reaching the practically limitless open spaces of the rubble column and salt cavity. As the extent of the cavity within the Sylvania increases in depth and width because of sand migration, a critical span will be reached where the immediately overlying upper Sylvania and the overlying Detroit River Dolomite will fail. The collapse will allow a path for the approximately 100 ft of clay to collapse, resulting in a sink as the surface manifestation.