Research on the epidemiology of bulimia nervosa has focused largely on the prevalence of the disorder. As methods have improved, consensus has increased regarding the prevalence rate among adolescent and young adult women--about 1%. However, the accuracy of this figure and its clinical significance must be questioned. In this synthesis of the epidemiological work to date, the authors review the literature from a clinical and research perspective. They recommend a shift in emphasis away from studies of the distribution of the disorder toward studies of the determinants of the whole spectrum of the disturbance that exists in the community.