PIP: A controlled trial of Bacille Calmette Guerin (BCG) vaccination was conducted in 1950 among 64,136 residents of Muscogee County, Georgia, and Russell County, Alabama. Only 11% of the total study population is now estimated to have been infected with Mycobacterium tuberculosis. About half of the initial nonreactors to tuberculin were vaccinated. Throughout 14 years of observation, tuberculosis continued to develop more frequently among initial reactors than among those who were nonreactors at the start of the trial. Among the nonreactors, vaccination accounted for a reduction in tuberculosis of only 14%, meaning that the net reduction in the entire study population was less than 5%. The slight contribution made by vaccination was manifested almost entirely during the 1st 4 years of the trial. Moreover, BCG vaccination was least effective among groups most in need of protection, especially Blacks. Although tuberculosis appeared to be most common among persons with small amounts of subcutaneous fat and those in inadequate housing, there was no indication that either of these factors could account for the lack of effectiveness of BCG vaccination. It is concluded that the risk of tuberculosis among nonreactors is too low to justify the use of BCG in the US. To achieve a dramatic decline in tuberculosis morbidity, it is necessary to find ways to stop the continued development of tuberculosis among persons with old established infections.