In 1979 an outbreak of tuberculosis occurred in medical students at the University of Sydney. Eight of 35 Mantoux-negative students who attended the autopsy of an immunosuppressed patient with unsuspected active tuberculosis became infected and one developed clinical disease. A report of the incident was prepared for publication because it supported the then controversial University policy of recommending BCG vaccination to medical and dental students in a country where the reported prevalence of tuberculosis is very low. The report was never published, mainly in order to protect the privacy of the individual students involved, but also because it was felt by the administration of the time that it might undermine confidence in infection control procedures in the autopsy room. The original report, updated and reproduced here, suggested that tuberculosis might be an emerging nosocomial problem. This has been all too clearly realised since its re-emergence as an opportunistic infection in AIDS patients. Worldwide, the problem of antibiotic resistance in Mycobacterium tuberculosis provides an added risk of a return to the situation which prevailed early this century when tuberculosis was a major occupational risk for young health care workers. Infection often restricted career choices, even in those whose disease was relatively benign. Our purpose in bringing this incident to light after so many years is to point out the relevance of the extensive studies of the problem which were conducted in the 1930s and 1940s to the current situation and to suggest that health care students are vulnerable to airborne infections as well as those spread by inoculation injuries. In retrospect, our 1979 conclusions about prospects for preventing nosocomial tuberculosis appear optimistic.