Use of the category of race in epidemiologic research presupposes scientific validity for a system that divides man into subspecies. Although the significance of race may be clear-cut in many practical situations, an adequate theoretical construct based on biologic principles does not exist. Anthropologists have in large measure abandoned the biologic concept of race, and its persistent widespread use in epidemiology is a scientific anachronism. The assumption that race designates important genetic factors in a population is in most cases false. Racial definitions should be seen as primarily social in origin and should be clues to environmental-rather than genetic-causes of disease. An understanding of the social forces leading to racial differentials in health will give further direction to preventive campaigns.