Information on several variables, including occupational history and various coronary risk factors, was collected from the wives of 568 married men who died of coronary heart-disease (CHD) and an equal number of matched control subjects. The crude matched-pair relative risk of fatal CHD among men who retired compared with non-retirees was 2.9 (95% confidence limits from 1.9 to 4.9). After adjustment for age and history of hospitalisation for myocardial infarction by means of a paired multiple-logistic regression analysis the relative risk was reduced to 1.8 (range 1.0 to 3.3). These data suggest that retirement and subsequent coronary mortality may be linked.