Several studies have confirmed that the contacts of sputum-smear-positive patients are a high risk group. We made a prospective survey to investigate the contacts of infectious cases of pulmonary tuberculosis in order to evaluate whether different grades of sputum-smear positivity have any consequence in the emergence of new cases. The number of contacts reported by 134 index cases was 609. These included 136 (22%) who had been in close contact to the index case, 69 of them to patients heavily positive by sputum smear. Tracing of 609 contacts over 2 yr revealed four (0.7%) new cases of active tuberculosis. All of them were close contacts, and, moreover, all four (5.8%) belonged to the group of 69 whose index case had a heavily positive sputum smear. This was significantly more than in other contacts, p = 0.0002. To be productive, tracing should be limited to close contacts of heavily smear-positive patients. This seems also to be the group in which chemoprophylaxis could be cost-effectively focused.