Few studies of physical activity and coronary heart disease (CHD) have included women or blacks. We examined this association in a biracial cohort of 45- to 64-yr-old adults. We related the sports, leisure, and work indices developed by J. A. H. Baecke et al. to CHD incident events (N = 97 in women, N = 223 in men) over 4-7 yr in the Atherosclerosis Risk in Communities study. The age-, race-, and field center-adjusted relative risk of CHD was 0.73 in women and 0.82 in men per each standard deviation increment in the sports index (P < 0.05). For the leisure index, these relative risks were 0.78 for both sexes (P < 0.05). The work index was not associated with CHD. These inverse associations held for non-blacks, but there was no association between the sport or leisure indices and CHD among blacks. Vigorous sports participation was strongly inversely associated with CHD, but an independent contribution of nonvigorous activity (e.g., walking) could not be demonstrated conclusively. Adjustment for other risk factors attenuated the relative risks, as one might expect if these risk factors mediated any protective effect of physical activity. Our findings reinforce evidence that regular physical activity should protect women, as well as men, from CHD. Explanations for no association among blacks, if real, are needed.