Both the spontaneous and the induced mutation rates in Salmonella tester strains vary among different laboratories, and also within the same laboratory over time. If there is an association between spontaneous and induced mutagenesis, a measure of mutagenic potency that incorporates the background may be more consistent than the simple measure of the induced slope. We have used the statistical procedures recently described by Bernstein et al. (1982), and a large data-base of Salmonella test results to examine the association between spontaneous and induced mutation and to compare several alternative measures of mutagenic potency. A correlation analysis indicated an association between spontaneous and induced mutation for TA98, TA1537 and TA1535; TA1538 was close to being significant. This was observed over a wide range of chemicals. In addition, for TA98, for which we observed the strongest association, we obtained a rough estimate of the relationship between slope and intercept by using least squares to fit K and p in the power curve beta = k alpha p. We then chose 3 simple potency measures: the slope, the ratio of slope to spontaneous background, and the ratio of slope to the square-root of spontaneous background. These corresponded to the range of p's estimated from the least-squares fit procedure. The reproducibility of these measures was compared and no significant differences were found. Though there were some differences in the relative potency ranking of chemicals using the different measures, they were highly correlated.