The spontaneous and induced revertant frequency there was little difference among the responses of of four Salmonella typhimurium strains (TA1535, the Salmonella strains on any of the agars. How-TA1537, TA98, and TA100) and Escherichia coli ever, with E. coli the responses include either a re-[WP2 uvrA (pKM101)] was evaluated using Vogel duction or an increase in spontaneous revertants Bonner minimal plates prepared with ten different numbers as well as a reduction in absolute and agars. In addition to the Difco Bacto agar originally relative induced revertant frequency. Difco Bacto recommended by Ames, Difco Noble, granulated agar appears to be the most consistent agar for and Bitek agars; BD grade A, BBL granulated and use with these strains. As an alternative, only BBL purified agars; Oxoid purified and No. 1 agars; purified agar resulted in consistent results for all of and GIBCO select agar were tested. Several of these strains under all testing conditions. These rethese agars have been reported as acceptable alter-sults emphasize the need to evaluate the componatives for these Salmonella strains, but comparable nents of the standard mutation assay when incorpostudies with E. coli have not been done. The bacte-rating additional bacterial strains. Suboptimal reria were treated with DMSO or an appropriate posi-sponses related to the agar or other components tive control in the presence or absence of an Aroclor could compromise the detection of weak muta-1254-induced rat liver activation system.