During a single week in April 1982, cultures for Mycobacterium tuberculosis were reported positive from nine patients who did not appear clinically to have active infection. Each of the patients had only one positive culture out of multiple specimens cultured. At the time of investigation, five specimens were available and were found to be all of the same phage type which strongly suggested cross-contamination. Four patients received antituberculosis chemotherapy. In one year of follow-up of the five who did not receive chemotherapy, none developed clinical disease. The contamination was probably due to faulty laboratory technique, but the source of the contaminant is uncertain. This investigation suggests that patients without clinical evidence of active infection and with isolated positive cultures for Mycobacterium tuberculosis should be carefully evaluated before they are subjected to a prolonged, potentially toxic, and expensive course of chemotherapy.