The association of cigarette smoking and atherosclerorosis was investigated in 1320 autopsied men, 25--64 years of age. Aortic and coronary lesions were evaluated visually in coded specimens and objectively by analysis of radiographs. Using schedules that had been tested on pairs of living persons, interviewers obtained estimates of cigarette smoking habits of the deceased men from surviving relatives. Data were analysed for black and white men in the total sample of cases and also in groups according to the presence (selected disease group) or absence (basal group) of diseases thought to be associated with smoking (emphysema, lung cancer, etc.) or with coronary heart disease (myocardial infarction, hypertension, diabetes, stroke, etc.). Atherosclerotic involvement of aorta and coronary arteries was greatest in heavy smokers and least in nonsmokers for both races in the total sample of cases, the basal group and the selected disease group.