The carcinogenic process is usually multifactor in its causation and multistep in its evolution. It is likely that entirely different molecular mechanisms underlie the many steps in this process. In contrast to initiating carcinogens, the action of the tumor-promoting phorbol esters does not appear to involve covalent binding to cellular DNA and they are not mutagenic. Recent studies in cell culture have revealed two interesting biologic effects of the phorbol esters and related macrocyclic plant diterpenes. The first is that at nanomolar concentrations they induce several changes that resemble those seen in cells transformed by chemical carcinogens or tumor viruses. These include altered morphology and increased saturation density, altered cell surface fucose-glycopeptides, decrease in the LETS protein, increased transport of deoxyglucose, and increased levels of plasminogen activator and ornithine decarboxylase. In transformed cells exposed to phorbol esters the expression of these features is further accentuated. Phorbol esters do not induce normal cells to grow in agar but they do enhance the growth in agar of certain transformed cells. The second effect of the phorbol esters is inhibition of terminal differentiation. This effect extends to a variety of programs of differentiation and is reversible when the agent is removed. With certain cell culture systems induction of differentiation, rather than inhibition, is observed. Both the transformation mimetic and the differentiation effects are exerted by plant diterpenes that have tumor-promoting activity but not by congeners that lack such activity. The primary target of phorbol esters appears to be the cell membrane. Early membrane-related effects include enhanced uptake of 2-deoxyglucose and other nutrients, altered cell adhesion, induction of arachidonic acid release and prostaglandin synthesis, inhibition of the binding of epidermal growth factor to cell surface receptors, altered lipid metabolism, and modifications in the activities of other cell surface receptors. A model of "two stage" carcinogenesis encompassing the known molecular and cellular effects of initiating carcinogens and tumor promoters is presented. According to this model, initiating carcinogens induce stable alterations in the cellular genome but these are not manifested until tumor promoters modulate programs of gene expression and induce the clonal outgrowth of the initiated cell.