In the classical two-stage models of carcinogenesis, initiation has been usually related to a DNA-damage/gene-mutation event, while promotion has been related to the non-genotoxic effects of clonal expansion of preneoplastic cells and/or modulation of cell differentiation. It is now clear that the process of carcinogenesis is linked to more than one irreversible alteration in the genome. Likewise, we can envisage that non-genotoxic events can take place after perhaps 0, 1, 2 or more irreversible alterations in the genome. Initiating and promoting activities of a chemical can be considered clearly separated in theory, but in practice, the chemicals we work with only rarely will be purely of the genotoxic or non-genotoxic type. We will discuss an empirical approach to classify genotoxic or prevalently non-genotoxic chemical carcinogens. For prevalently non-genotoxic carcinogens we will analyze what fraction of them can be detected as promoters of in vivo rat liver carcinogenesis. We will analyze carcinogenic potency of genotoxic and non-genotoxic carcinogens.