Anecdotal historical evidence suggests that tuberculosis was uncommon at high altitude, but whether transmission is affected by high altitude is not known. To test whether high altitude lowers infection by Mycobacterium tuberculosis, the prevalence of tuberculin skin test (TST) positivity was compared between two high altitude villages (3,340 meters [10,960 feet] and 3,500 meters [11,480 feet]) and three sea-level sites in Peru. High altitude villages had lower TST-positive prevalence rates (5.7% and 6.8%) than sea level areas (25-33%), and the difference remained significant (odds ratio = 4.5-6.0) after adjusting for age, education, bacille Calmette-Guérin vaccination, and contact with tuberculosis patients. The TST-positive individuals clustered within highland families more than within sea level families. These data suggest that prevention and control efforts targeted to families may be more effective at high altitude. The mechanism by which TST-positivity prevalence is decreased at high altitude is unknown, but may reflect relative hypoxia, low humidity, or an increased ultraviolet effect.