The intake of known dietary carcinogens was compiled and the cancer risk was estimated on the basis of carcinogenic potencies in animals as derived from the Carcinogenic Potency Database by Gold and co-workers. The total cancer risk was compared with the number of cancer cases attributed by epidemiologists to dietary factors (one-third of all cancer cases, i.e. approximately 80,000 per one million lives). Except for alcohol, the known dietary carcinogens could not account for more than a few hundred cancer cases. This was seen both with the DNA-reactive carcinogens (heterocyclic aromatic amines, polycyclic aromatic hydrocarbons, N-nitroso compounds, estragole, aflatoxin B1, ethyl carbamate, to name the most important factors) as well as with those carcinogens which have not been shown to react with DNA (e.g. caffeic acid and the carcinogenic metals arsenic and cadmium). Residues and contaminants turned out to be negligible. Among the various possibilities to explain the discrepancy we investigated the role of overnutrition. Dietary restriction in animals is well known for its strong reducing effect on spontaneous tumor formation. These data can be used to derive a carcinogenic potency for excess macronutrients: the tumor incidence seen with the restricted animals is taken as a control value and the increased tumor incidence in the animals fed ad libitum is attributed to the additional feed intake. For excess standard diet in rats, a carcinogenic potency TD50 of 16 g/kg/day was deduced from a recent study. Overnutrition in Switzerland, estimated to be 5.5 kcal/kg/day, was converted to excess food (1.9 g/kg/day) and the cancer incidence was calculated. The result, 60,000 cancer cases per one million lives, is provocatively close to the number of cases not explained by the known dietary chemical carcinogens. Mechanistic studies will be required to test our hypothesis and investigate the role of different types of macronutrients in overnutrition.