Tobacco smoke aerosols with fewer mutagens in the particulate fraction may present reduced risk to the smoker. The objective of this study was to test the hypothesis that the temperature at which tobacco is pyrolyzed or combusted can affect the mutagenicity of the particulate fraction of the smoke aerosol. Tobacco smoke aerosol was generated under precisely controlled temperature conditions from 250 to 550 degrees C by heating compressed tobacco tablets in air. The tobacco aerosols generated had a cigarette smoke-like appearance and aroma. The tobacco smoke aerosol was passed through a Cambridge filter pad to collect the particulate fraction, termed the smoke condensate. Although condensates of tobacco smoke and whole cigarette mainstream smoke share many of the same chemical components, there are physical and chemical differences between the two complex mixtures. The condensates from smoke aerosols prepared at different temperatures were assayed in the Ames Salmonella microsome test with metabolic activation by rat liver S9 using tester strains TA98 and TA100. Tobacco smoke condensates were not detectably mutagenic in strain TA98 when the tobacco smoke aerosol was generated at temperatures below 400 degrees C. Above 400 degrees C, condensates were mutagenic in strain TA98. Similarly, condensates prepared from tobacco smoke aerosols generated at temperatures below 475 degrees C were not detectably mutagenic in strain TA100. In contrast, tobacco tablets heated to temperatures of 475 degrees C or greater generated smoke aerosol that was detectably mutagenic as measured in TA100. Therefore, heating and pyrolyzing tobacco at temperatures below those found in tobacco burning cigarettes reduces the mutagenicity of the smoke condensate. Highly mutagenic heterocyclic amines derived from the pyrolysis of tobacco leaf protein may be important contributors to the high temperature production of tobacco smoke Ames Salmonella mutagens. The relevance of these findings regarding cancer risk in humans is difficult to assess because of the lack of a direct correlation between mutagenicity in the Ames Salmonella test and carcinogenicity.