The authors studied the historical spread of human immunodeficiency virus type 1 (HIV-1) infection in homosexual/bisexual men and projected its future spread in these men using data from an AIDS-free cohort recruited during late 1984 in Baltimore, Maryland; Chicago, Illinois; Los Angeles, California; and Pittsburgh, Pennsylvania. Dates of preentry seroconversion in HIV-1 seroprevalent men were estimated using study entry values of hematologic variables influenced by HIV-1 infection. The authors used survival methods incorporating truncation to determine numbers/dates of seroconversion for men with a pre-1984 AIDS diagnosis who were selectively excluded by design from the 1984 AIDS-free cohort. Overall, the annual seroconversion hazard rose progressively from 0.4% in 1978 to 13.8% in 1983, dropped to 4.6% in 1985, and remained relatively stable at 1.1-2.2% from 1986 to 1990. By January 1990, almost 46% of men who were seronegative in 1978 had seroconverted. The authors estimated historical rates of spread by city, age, education, and ethnicity to examine the effects of these factors in the early and continuing stages of the HIV-1 epidemic. There were striking differences among cities with respect to pre-1985 seroconversion rates but not with respect to post-1985 seroconversion rates. Age, education, and ethnicity were all associated with 1978-1990 seroconversion rates. Future seroconversion among homosexual men was predicted assuming that the "stabilized" 1986-1990 hazards (stratified by age) observed here will be representative of future rates. Truncated Kaplan-Meier methods gave the probability of a seronegative 20-year-old man's remaining seronegative in subsequent years. Such a man has a 20.2% chance of seroconverting before reaching the age of 25 years (a 4.4% yearly hazard). The annual hazard drops to 2.5% between 25 and 30 years, to about 1.5% between 30 and 45 years, and to 1.0% between 45 and 55 years. The overall probability of seroconversion prior to age 55 years is about 50%, with seroconversion still continuing at and after age 55. Given that this cohort consists of volunteers receiving extensive anti-HIV-1 transmission education, the future seroconversion rates of the general homosexual population may be even higher than those observed here.