An unusual series of positive acid-fast smears, including positive stains of a commercial saline solution, has been studied. Short of waiting 2--6 weeks for culture results, a laboratory investigation of these findings appeared hopeless: repeat staining failed to confirm the origin smear results, and a review of laboratory techniques and supplies failed to pinpoint any source of specimen or smear contamination. An alternative epidemiological approach was adopted. A review of the laboratory records indicated that, during the time the unusual series occurred, there had been a distinct change in the distribution of positive smears by type of specimen submitted and by ward of origin. The change in distribution indicated that the unusual series was probably part of a larger episode of false-positive acid-fast smears caused by random specimen or smear contamination in the hospital laboratory. Culture results eventually confirmed this and showed that the period of random laboratory contamination had ended just before the present investigation (thus explaining the failure of the initial approach). This experience suggests that ongoing analysis of specimen and ward distribution of positive acid-fast smear results will enable hospitals to detect episodes of false-positive smears earlier, thus reducing erroneous diagnoses and permitting prompt evaluation of sources of specimen and smear contamination.