Tuberculosis (TB) kills more people in the world today than any other infectious disease, and the number of drug-resistant Mycobacterium tuberculosis isolates is increasing. Vaccines, better than most of the currently available strains of bacille Calmette–Guérin (BCG), are urgently needed to control this disease. TB in rabbits resembles human TB more closely than TB in any other common laboratory animal and a most pertinent method of assessing vaccine efficacy is Lurie's tubercle count method in this species. Vaccinated and control rabbits were infected by aerosol with virulent human-type tubercle bacilli (H37Rv). At necropsy 5 weeks thereafter, the grossly visible primary tubercles in the entire lung were counted. A decrease in the number of such tubercles is a quantitative measure of vaccine efficacy: An effective vaccine prevents microscopic tubercles from growing to grossly visible (clinically apparent) size. The Pasteur substrain of BCG and two substrains of Mycobacterium microti (the vole bacillus) reduced the number of visible primary tubercles an average of 75%, whereas three other substrains of BCG and three other substrains of vole bacilli only reduced the number an average of 40%. These initial studies indicate that Lurie's tubercle-count method in rabbits is a precise way to choose the best available tuberculosis vaccines.