Despite the use of combinations of drugs that permit a substantial shortening of treatment and prophylaxis, as well as intermittent directly observed therapy, current regimens for treating patients with tuberculosis still require many months to be effective, and sustaining adherence remains a challenge. Indeed, the worldwide prevalence of tuberculosis is steady and each year millions of persons throughout the world die of the disease. The World Health Organization projects that the combination of HIV and tuberculosis is likely to result in a substantial increase in deaths in the coming years. Moreover, outbreaks of drug-resistant (and often multidrug-resistant) tuberculosis (1) and outbreaks clearly linked to nonadherent patients with drug-sensitive disease (2) have underscored the limitations of currently available treatments and control programs. In this context, it is easy to understand why some experts think that real control of tuberculosis, particularly on a worldwide basis, is unlikely to occur until an effective vaccine becomes available. This objective is certainly not new. The tuberculin skin test, which is used to identify tuberculous infection, was developed years ago and originated in an effort to develop an effective therapeutic vaccine for tuberculosis. The antigenic complexity of mycobacteria in general and Mycobacterium tuberculosis in particular has, to date, created an insurmountable challenge. However, the remarkable developments in recent years in our understanding of the immunobiology of tuberculosis, as well as the availability of the powerful new tools of molecular biology, make this goal seem more possible today.