Homosexual/bisexual men from Amsterdam, The Netherlands, New York, New York, and San Francisco, California, were entered into trials of the efficacy of hepatitis B vaccine shortly before the acquired immunodeficiency syndrome (AIDS) epidemic was recognized (1978-1980). The authors analyzed data, including serial blood samples tested for antibody to human immunodeficiency virus type 1 (HIV-1) as well as demographic and behavioral information, to characterize the spread of HIV-1 infection within the cohorts. By the end of 1982, the cumulative incidence of HIV-1 infection within the cohorts. By the end of 1982, the cumulative incidence of HIV-1 infection was 7.5% in Amsterdam, 26.8% in New York City, and 42.6% in San Francisco. Covariate analysis showed that differences in sexual activity (number of male sexual partners) and correlates of sexual activity (age and hepatitis B incidence) accounted for the differences in incidence of HIV-1 infection between the New York City and San Francisco cohorts. These covariates did not explain the lower incidence in the Amsterdam cohort. In conclusion, significant differences were found in the spread of HIV-1 in cohorts of homosexual men in Amsterdam, New York City, and San Francisco. These dissimilarities were probably due to a combination of differences in sexual activity at the time the epidemic began and a later introduction of HIV-1 in Amsterdam.