The reductive dechlorination of Aroclor 1254 by microorganisms eluted from Hudson River sediments was investigated. The rate, extent, and pathway of dechlorination were compared in biphenyl-enriched and non-enriched sediments over a 24-month period. The greatest dechlorination activity took place within the first 5 months of incubation, where approximately 30% of the total chlorines was removed in both treatments. Dechlorination activity slowed substantially so that by the end of 24 months only an additional 8 and 2% of the total chlorines were removed in the nonenriched and biphenyl-enriched sediments, respectively. Aroclor 1254 was dechlorinated primarily with the loss of chlorines meta to the biphenyl link. It ranged from 64 to 66% in the biphenyl-enriched and nonenriched sediments, respectively. The more extensive overall dechlorination in the nonenriched sediments was characterized by the loss of 40% of the para chlorines compared with only a 12% loss in the biphenyl-enriched sediments. However, dechlorination was congener-specific as evidenced by the recalcitrance of two peaks 2,4,2',4'- + 2,4,5,2'-chlorobiphenyl (CB) and 2,4,2',5'-CB. The difference in para dechlorination was also in fact due primarily to the inability to dechlorinate 2,4,4'-CB with biphenyl enrichment. These results suggest that environmental factors may alter the congener specificity of dechlorination. In both treatments, several parent congeners persisted at low levels after 24 months, suggesting the existence of a threshold concentration below which no dechlorination took place.