Epidemiologic inferences about the epidemic of acquired immunodeficiency syndrome (AIDS) are developed from three different sources: (1) case-control studies, (2) cohort studies, and (3) national data on prevalence and incidence. Together these data implythat reducing the number of sexual partners may in itself provide little protection. Shifting to a “lower-risk” behavior will only be helpful if the cumulative risk of transmitting the viruscan be reduced nearly to zero.However, if all membersof a minimally affected population made such a shift, the benefit to that population could be dramatic. The emerging realization that latency periods are longer than was first thought implies that the epidemic could be much more severe than has been anticipated. Regional trends giveno assurance that it has reached a peak, evenin areas of highest prevalence. Much more work needs to be done to establish the actual role of saliva in transmitting HTLV-III because oral contagion has the greatest potential to spread the disease. Efforts to determine the latency period are crucial to predicting the future of the epidemic. Preventive measures behavior modification and education - hold some hope of retarding spread of the epidemic pending development of effective therapy or a vaccine.