The ability of any screening test to correctly identify diseased patients is directly related to the prevalence of the disease in question. The continuing use of smears for the detection of acid-fast bacilli when the prevalence of tuberculosis is declining will produce increasing numbers of false-positive results. To determine the usefulness of smear techniques, we carried out a retrospective study of approximately 10 000 consecutive smears and cultures. Positive smears were classified as true- or false-positive on the basis of whether the corresponding cultures grew mycobacteria. The overall incidence of false-positive smears was 0.7%; however, the ratio of the number of false-positive to total positive smear was about 0.5. Thus, for each true-positive result, one false-positive result was obtained. Patients in the true-positive group had clinical and radiological evidence to support the diagnosis of tuberculosis, while those in the false-positive group had new findings of the disease. These data suggest that the smear is a poor screening technique in a population where the prevalence of tuberculosis is low.