Ozone is a highly reactive gas that has been tested for genotoxicity in a number of systems. Induced genetic damage resulting from ozone treatment may not be readily observed because of the high toxicity of the chemical and difficulties in generating and administering controlled concentrations. The mutagenicity of ozone was investigated in Salmonella typhimurium using a plate test protocol designed for reactive vapours and gases. Ozone, at two to three consecutive doses, induced weak, albeit statistically significant, mutagenic responses in tester strain TA102 with and without Aroclor-induced rat liver s9 (lowest effective mean concentration of 0.019 ppm; 35 min total exposure). However, dose-related responses were not always obtained. No mutagenicity was detected in strains TA98, TA100, or TA1535, with or without S9. In strain TA104, ozone induced a weak response only at a single dose with S9; this response was not reproducible. Mutagenicity was dependent on the ozone flow rate and total exposure time, with variations in the optimum dase-time regimen leading to toxicity or complete inactivity. The data show that ozone is a very weak bacterial mutagen and only when tested under narrowly prescribed, subtoxic dosing conditions. 0 1992 Wiley-Liss, Inc.