Persons coming into contact with patients who have active pulmonary tuberculosis are at risk of tuberculous infection. There is a need for a simple way to identify contacts who have the greatest risk, so that they can be investigated with least delay. In addition, if resources and manpower limit investigation of contacts, then the lower-risk contacts can be omitted. By using 2 characteristics of the exposure, local health department personnel in Birmingham and New Orleans categorized 1,590 contacts into 4 groups with differing risks of infection. The 2 characteristics used were the relative concentration of tubercle bacilli in the sputum of the source case (evaluated by Ziehl-Neelsen smear; high=smear-positive, low=smear-negative) and the intimacy of exposure of the contact to the source case (household or nonhousehold). Household contacts of smear-positive patients were at highest risk of infection (infection rate of 46%). At lower risk were nonhousehold contacts of smear-positive patients (34%), followed by household contacts of smear-negative patients (28%) and nonhousehold contacts of smear-negative patients (24%). The authors conclude that local health departments can assign priority to contacts simply and reliably on the basis of the sputum smear of the source case and the intimacy of exposure of the contact to the source case.