In several countries, bovine tuberculosis (caused by infection with Mycobacterium bovis) is a major economic problem with the potential to be a significant public health risk. Where traditional test-and-slaughter policies based on skin testing with tuberculin have not been fully successful, new tools including additional diagnostic tests and improved vaccines are required urgently. This paper considers how recent developments in knowledge of immune responses and mycobacterial antigens can be used in the logical development of more efficient strategies for the identification of infected cattle.