An interlaboratory study of an EPA/Ames/Salmonella test protocol.

7 laboratories participated in a collaborative study to evaluate an EPA standard protocol for the Ames test. The study utilized Salmonella typhimurium (strains TA98 and TA100) and 3 metabolic activation levels (0%, 2%, and 10% S9 in the S9 mix). 6 pure chemicals and 2 complex mixtures were tested as coded unknowns. Ability to obtain qualitative results in agreement with published data was less (% agreement) than that reported in an earlier study (% agreement) by de Serres and Ashby (1981) in which each laboratory used its own protocol. The conclusion from analysis of the quantitative data from this interlaboratory Ames study was that both intralaboratory and interlaboratory variations were substantial. Results for the same substance varied by an order of magnitude or more (CV of 115%) when the mutagenic response was measured as the slope of the dose response in revertants/microgram. Taking interlaboratory variation into account, one chemical must be more than an order of magnitude more mutagenic than another (ratio of slopes greater than 10) to have only an even chance of finding a statistically significant difference between the two chemicals at the 5% level. Such large variations must be taken into account when evaluating Ames/Salmonella data.