Several automated approaches to Salmonella mutagenicity testing have been proposed in recent years but have failed to gain acceptance in the scientific community due to poor performance or lack of demonstrated usefulness. The paper reports on an automated system that successfully generates dose-response data and, moreover, reduces the labor, materials, and sample mass required to obtain such information. In the standard Salmonella plate-incorporation assay, dose-response relationships are defined by testing discrete doses of the test agent on a series of agar plates. In contrast, the spiral Salmonella assay generates dose-response data from a continuous concentration gradient on a single agar plate. Upon analysis, each spiral plate yields a dose-response curve consisting of 13 data points that span a concentration range of about 15:1. The performance of the spiral Salmonella assay was compared to that of the conventional plate-incorporation assay using 13 mutagens and 7 nonmutagens selected from a variety of chemical classes.