For a series of 568 married white men aged 30-70 years who died from coronary heart diseases, (CHD), and a matched sample of living neighbourhood controls, information was collected on a large number of variables, including physical activity, (PA), classified according to the Health Insurance Plan Study criteria. Increased leisure PA was associated with a decreased risk of coronary death (p less than .001) but there was no association between job activity and coronary deaths. Step-up multiple regression, based on within-pair differences, was used to control for the possible confounding effects of all variables studied. Even when these variables were controlled for there was still a significant association between increased leisure PA and decreased risk of death due to CHD (P less than .001). These data are consistent with the hypothesis that increased leisure PA can contribute to the prevention of death from CHD.