Enantiomers of the diastereomeric pair of bay-region benz(a)anthracene 3,4-diol-1,2-epoxides in which the benzylic 4-hydroxyl group and epoxide oxygen are either cis (isomer 1) or trans (isomer 2) were evaluated for mutagenic activity in two histidine-dependent strains of Salmonella typhimurium, as well as in an 8-azaguanine-sensitive Chinese hamster cell line. In strain TA 98 of S. typhimurium, the diol-epoxide with (1S,2R,3R,4S) absolute configuration [(-)-diol-epoxide 2] was the most active isomer, although there was less than a 3-fold difference in the mutagenicity of the four diol-epoxides. However, in strain TA 100 of S. typhimurium, the enantiomeric diol-epoxide with (1R,2S,3S,4R) absolute configuration [(+)-diol-epoxide 2] was the most active diol-epoxide, and the two isomers with (3S,4R) absolute configuration [(-)-diol-epoxide 1 and (+)-diol-epoxide 2] were three to eight times more active than were the two isomers with (3R,4S) configuration. The highest degree of sensitivity to absolute configuration was observed in Chinese hamster V79 cells, in which the (1R,2S,3S,4R) isomer [(+)-diol-epoxide 2] was from three to 20 times more mutagenic than were the other three isomers. This metabolically predominant (+)-diol-epoxide 2 isomer, which has high activity in strain TA 100 of S. typhimurium and the Chinese hamster V79 cells, has the same absolute configuration as do the bay-region diol-epoxide isomers of benzo(a)pyrene and chrysene that have been shown previously to be exceptionally mutagenic to mammalian cells and highly tumorigenic in mice. Analysis of the mutagenic activity of the (+)- and (-)-isomers of the 1,2- and 3,4-tetrahydroepoxides of benz(a)anthracene revealed only small enantiomeric differences in strain TA 98 of S. typhimurium (2.5 fold) and little, if any, differences (less than 1.5-fold) in the other two mutagenicity systems. However, the extent to which the four tetrahydroepoxides were converted to nonmutagenic products by homogeneous microsomal epoxide hydrolase (EC 3.3.2.3) indicated marked differences in the stereoselectivity of the enzyme. (-)-(3R,4S)-Epoxy-1,2,3,4-tetrahydrobenz(a)anthracene appears to be an exceptionally good substrate for epoxide hydrolase.