At Yale-New Haven Hospital, 46 specimens submitted for mycobacterial culture during an 8-week period in 1989 were positive for Mycobacterium gordonae, a nontuberculous acid-fast bacterium (AFB) of low pathogenicity. The specimens were submitted from 34 patients who came from various inpatient and outpatient services. Four patients were begun on antimycobacterial therapy on the basis of an AFB isolate which was later identified as M. gordonae. Isolation of M. gordonae was associated with use of the BACTEC TB system (BACTEC TB; Becton Dickinson Diagnostic Instrument Systems, Towson, Md.) and an antimicrobial solution, BACTEC PANTA PLUS (PANTA; Becton Dickinson Diagnostic Instrument Systems). The manufacturer reported that two lots (B9K1 and C9K1) of PANTA kits containing a single production lot (N8C1) of PANTA, which had been shipped to 173 laboratories, had been contaminated with M. gordonae. A survey of mycobacteriology laboratories in the United States revealed that, during April to July 1989, the M. gordonae isolation rate was 5.8/1,000 AFB specimens processed at laboratories that did not use BACTEC TB, 11.4/1,000 AFB specimens at laboratories that used BACTEC TB but not the implicated lot of PANTA, and 23.5/1,000 AFB specimens at laboratories that used BACTEC TB and the lot of implicated PANTA. Intrinsic contamination of PANTA was attributed to ineffective sterilization of water used in the manufacturing process and was not detected prior to product shipment because cultures for AFB were not part of the quality control regimen. This episode emphasizes that clinical laboratories can detect pseudoepidemics promptly if they are alert to abrupt increases in isolation rates, especially of unusual or generally nonpathogenic organisms.