A comparative potency method for cancer risk assessment has been developed based on a constant relative potency hypothesis. This method was developed and tested using data from a battery of short-term mutagenesis bioassays, animal tumorigenicity data and human lung cancer risk estimations. This data base was developed for a series of complex mixtures including emissions from coke ovens, roofing tar pots, cigarette smoke and automotive engines. The series of automobiles used in this study included both diesel- and gasoline-powered vehicles. The assumptions inherent in this method are discussed, together with the methods and data base used to test these assumptions.