Sulfur isotope analyses were carried out on ore minerals from Mississippi Valley-type deposits throughout the southern Appalachians, as well as on evaporite sulfate minerals that might have supplied this sulfur, in an effort to determine the origin(s) of sulfur in the deposits. Mississippi Valley-type deposits are found largely in two major paleoaquifers in the region, the Lower Cambrian Shady Dolomite and the Lower Ordovician Knox Group karst zone. In general, Mississippi Valley-type sulfide delta 34 S values are different in these two paleoaquifers, but appear to be relatively homogeneous within each paleoaquifer over areas of thousands of square kilometers. In the Knox paleoaquifer, the most common sulfide sulfur has delta 34 S values of 31 to 36 per mil and appears to have been derived by complete reduction of seawater sulfate of Lower Ordovician age, with little or no fractionation. A second population of sulfide sulfur in the Knox paleoaquifer has delta 34 S values of 26 to 29 per mil and appears to have been derived by similar reduction of Lower Cambrian seawater sulfate. Barite deposits in the Knox paleoaquifer have delta 34 S values that also fall in the range of Lower Ordovician seawater but extend to values that could reflect mixing with sulfate from underlying Upper Cambrian evaporites. The isotopic composition of sulfate sulfur in the Lower Cambrian paleoaquifer is similar to that of local evaporites, but sulfide minerals in the northern part of the paleoaquifer may have been derived from Mississippian evaporites.These results indicate that sulfur isotope compositions are not specific to individual Mississippi Valley-type deposits but, rather, have similar compositions over large regions. Compositions are most similar within specific stratigraphic units, with local evidence for cross-formational flow. The close similarity of sulfide and sulfate delta 34 S values suggests that large volumes of sulfate were reduced quantitatively with little or no fractionation. The lack of isotopic zonation within individual crystals indicates that sulfate reduction took place largely before sulfide deposition. Although thermochemical sulfate reduction appears to be the most likely mechanism to produce this reduced sulfur, some might have been produced by bacterial reduction in bottom waters of the overlying anoxic Sevier Shale basin, which recharged the Knox paleoaquifer.