AIMS: To determine factors that may influence the efficacy of an oral pelleted vaccine containing Mycobacterium bovis bacille Calmette-Guérin (BCG) to induce protection of brushtail possums against tuberculosis. To determine the duration of protective immunity following oral administration of BCG. METHODS: In Study 1, a group of possums (n=7) was immunised by feeding 10 pellets containing dead Pasteur BCG, followed 15 weeks later with a single pellet of live Pasteur BCG. At that time, four other groups of possums (n=7 per group) were given a single pellet of live Pasteur BCG orally, a single pellet of live Danish BCG orally, 10 pellets of live Pasteur BCG orally, or a subcutaneous injection of live Pasteur BCG. For the oral pelleted vaccines, BCG was formulated into a lipid matrix, and each pellet contained approximately 107 colony forming units (cfu) of BCG, while the vaccine injected subcutaneously contained 106 cfu of BCG. A sixth, non-vaccinated, group (n=7) served as a control. All possums were challenged by the aerosol route with a low dose of virulent M. bovis 7 weeks after vaccination, and killed 7–8 weeks after challenge. Protection against challenge with M. bovis was assessed from pathological and bacteriological findings. In Study 2, lipid-formulated live Danish BCG was administered orally to three groups of possums (10–11 per group), and these possums were challenged with virulent M. bovis 8, 29 or 54 weeks later. The possums were killed 7 weeks after challenge, to assess protection in comparison to a non-vaccinated group. RESULTS: The results from Study 1 showed that vaccine efficacy was not adversely affected by feeding dead BCG prior to live BCG. Feeding 10 vaccine pellets induced a level of protection similar to feeding a single pellet. Protection was similar when feeding possums a single pellet containing the Pasteur or Danish strains of BCG. All vaccinated groups had significantly reduced pathological changes or bacterial counts when compared to the non-vaccinated group. In Study 2, oral administration of Danish BCG induced protection against challenge with M. bovis, which persisted for at least 54 weeks after vaccination. Some protection was observed in possums challenged 54 weeks after vaccination, but this protection was significantly less than that observed in groups vaccinated 29 or 8 weeks prior to challenge. There was a strong relationship between the proportion of animals producing positive lymphocyte proliferation responses to M. bovis antigens and protection against challenge with M. bovis. CONCLUSIONS: Factors considered potentially capable of interfering with vaccination, including feeding dead BCG to possums prior to feeding live BCG, feeding multiple doses of BCG at one time, and changing strains of BCG, were shown not to interfere with the acquisition of protective immune responses in possums. Protection against tuberculosis was undiminished up to 29 weeks after vaccination with BCG administered orally. It is concluded that vaccination of possums by feeding pellets containing BCG is a robust and efficient approach to enhance the resistance of these animals to tuberculosis.