Some published data indicate that several minerals, some of which are usually considered relatively stable species, become corroded within fresh and unweathered sandstone strata. This solution or decomposition is evidently due to the action of solutions within the buried strata, and the surficial alteration due to weathering is not included in this discussion. This corrosion of minerals has suggested that some of the less stable ones may have completely disappeared from ancient strata. Examples are presented which show that hornblende and epidote have disappeared from some Miocene sandstones. The mere absence of some relatively unstable mineral, especially some of the ferro-magnesian group, therefore seems doubtful evidence for conclusions regarding the source rocks or the stratigraphic relations of sandstones. These considerations seem especially important for older formations, as the small assemblage of stable minerals, that is usual in early Paleozoic strata, may represent only a survival of the fittest.