Epidemiologic methods quantitatively assess the relations between body weight and general or cause-specific morbidity and mortality. This research is especially difficult because of the complex interrelations between weight and diet, physical activity, cigarette smoking, blood pressure, diabetes, ischemic heart disease, and other conditions. The interactions are not easily summarized with available mathematical models. Weight may be either a dependent or an independent variable, according to the analysis. Epidemiologic studies of body weight are subject not only to biases of sampling and selection, but also marked difficulties in definition and measurement. There are variations of interpretation, sometimes even of the same set of data. However, epidemiology continues to uncover important information that is consistent across studies, and that may be used to formulate programs for disease prevention.