In 1977, a comprehensive tuberculosis management program was introduced at a medical facility which then served approximately 65,000 black gold miners. The effectiveness of this program was evaluated from the relapse rate after treatment in 2,776 men. Post-treatment follow-up averaged 23.5 months. Limited information on potential determinants of relapse was available for all subjects and more detailed information for subsets. The drug regimen was the major determinant of relapse. Age was not a determinant of relapse nor was primary resistance of Mycobacterium tuberculosis to antituberculosis drugs, which was infrequent. The initial extent and severity of disease had a direct relation with the risk of relapse. The risk of relapse was greatest during the first 3 yr after treatment but persisted for the duration of the observation period. Neither continued exposure to mine dust nor the presence of pre-existing silicosis were shown to be determinants of relapse. A 4.5-month, four-drug weekday regimen was identified as being particularly effective in this working population and may be suitable for use in other working populations in the developing world.