Marked differences were observed in intermediate sulphur metabolism between non-pathogenic strains of Bacteroides melaninogenicus var melaninogenicus (CP-) and pathogenic Bacteroides melaninogenicus asaccharolyticus (CP+). The CP+ strains, which produced collagenase and protease and caused formation of abscesses when injected subcutaneously into groins of guinea pigs, produced copious amounts of volatile sulphur compounds (VSC) which consisted predominantly of CH3SH and (CH3S)2. Hydrogen sulphide occurred in considerably lesser amounts. CP+ cultures yielded 8-fold more total volatile S, 15-fold more CH3SH and 260-fold more (CH3S)2 during 24 h of incubation in trypticase-yeast extract medium. Whereas H2S accounted for 60 per cent of the total volatile S content of the head-space of CP- cultures, it represented only 8 per cent of the volatile S in CP + systems. Although the CP-organisms did not grow as well as CP +, the differences in concentration of VSC may be only partly related to the disparity in growth rates. When the VSC concentrations were calculated on the basis of equivalent optical density of 1.0, the CP + strains still produced over 3-fold more total volatile S, 6-fold more CH3SH and 100-fold more (CH3S)2. A similar allowance for growth rate suggests that CP-strains may possess a greater potential to produce H2S. Both groups metabolized S-containing amino acids and serine, resulting in appreciable increases in H2S production by CP-. However, the two groups appeared to metabolize the carbon moiety of cystine an cysteine by different pathways. The addition of glucose to the medium depressed total volatile S production by both CP+ and CP-strains, attributable mostly to lower H2S levels. Whereas the omission of yeast extract and charcoal treatment of trypticase did not adversely effect the activity of CP+, it further markedly reduced the capacity of CP-cultures to produce VSC. These results suggest that VSC analysis offers a convenient means of assessing strain differences and pathogenic potential of B. melaninogenicus.