Historically our knowledge about the direct carcinogenic activity of cigarette smoke and its constituents grew from painting experiments on the skin of mice to produce papillomas and carcinomas. The neutral fraction of cigarette smoke condensate had most of the carcinogenic activity in this test and was rich in carcinogenic polycyclic aromatic hydrocarbons (PAHs), the most abundant by far being BP. However, the concentration of BP in the condensate was only about 2% the amount of pure BP required to cause skin tumors. In other fractions there were non-carcinogenic constituents that promoted tumor formation when applied repeatedly to mouse skin that had been initiated by a single subcarcinogenic application of BP. There were also constituents of cigarette smoke that acted as cocarcinogens when applied simultaneously with repeated applications of BP. BP was effective as an initiator at lower concentrations than as a complete carcinogen, and some non-carcinogenic PAHs in the condensate were also active initiators. It was concluded from these studies that cigarette smoke condensate is primarily a tumorpromoting and co-carcinogenic agent with weak activity as a complete carcinogen. A major effect of promoters, and possibly of co-carcinogens, is a diffuse hyperplasia which includes selective expansion of clones carrying endogenous mutations and/or mutations induced by PAHs and other carcinogens such as NNK. The induced mutations as well as damaged cells would occur throughout the exposed region and, along with the hyperplasia, increase the permissiveness of the cellular microenvironment for neoplastic expression of any potential tumor cell in its midst. Since neither the promoters nor co-carcinogens in tobacco smoke are known to interact directly with DNA, their effects can be considered epigenetic processes that act upon genetically altered cells. Examples are cited from studies of experimental skin carcinogenesis, smoking-induced histopathological changes in human lung and spontaneous transformation in cell culture to illustrate the genetic and epigenetic interactions of neoplastic development in general and their significance for smoking-induced lung cancer in particular. Certain dietary modifications that appear to be effective in moderating the promotional phase of animal and human carcinogenesis are suggested for trial in managing lung cancer.