Bisulfite reductase was purified from extracts of Desulfovibrio vulgaris. By colorimetric analyses trithionate was found to be the major product, being formed in quantities 5 to 10 times more than two other detectable products, thiosulfate and sulfide. When [35S]bisulfite was used as the substrate, all three products were radioactively labeled. Degradation of [35S]trithionate showed that all of its sulfur atoms were equally labeled. In contrast, [35S]thiosulfate contained virtually all of the radioactivity in the sulfonate atom while the sulfane atom was unlabeled. These results, in conjunction with the funding that the sulfide was radioactive, led to the conclusion that bisulfite reductase reduced bisulfite to trithionate as the major product and sulfide as the minor product; the reason for the unusual labeling pattern found in the thiosulfate molecule was not apparent at this time. When bisulfite reductase was incubated with [35S]bisulfite in the presence of another protein fraction, FII, the thiosulfate formed from this reaction contained both sulfur atoms having equal radioactivity. This discovery, plus the fact that trithionate was not reduced to thiosulfate under identical conditions, led to the speculation that bisulfite could be reduced to thiosulfate by another pathway not involving trithionate.