SETTING: A rural district in Southern Ethiopia. OBJECTIVE: To estimate the prevalence of smear-positive pulmonary tuberculosis (TB). DESIGN: In this cross-sectional study, adults aged >14 years were surveyed by home-to-home visit, and asked about cough of =2 weeks with or without sputum, chest pain or difficulty in breathing. Symptomatic suspects submitted three sputum samples for standard smear microscopy. RESULTS: Of 16697 adults surveyed, 436 (2.6%) were symptomatic and submitted sputum samples. Thirteen (3%) were positive for acid-fast bacilli, and the prevalence of smear-positive TB was 78 per 100000 population (95%CI 36–120). Twenty-four smear-positive cases identified through the existing health care delivery were on anti-tuberculosis medication at the time of the survey. The ratio of smear-positive cases on treatment to those newly detected by the survey was 2:1. CONCLUSION: The prevalence of TB in this rural setting was unexpectedly low. For every two cases of smear-positive TB on treatment, there was one undetected infectious case in the community. However, as our screening technique did not allow detection of cases who did not report symptoms, the true prevalence may have been underestimated.