The interaction of herpes zoster and the human immunodeficiencyvirus (HIV) was evaluated in a cohort study of 287 homosexual men with well-defined dates of HIV seroconversion and 499 HIV-seronegative homosexual men. The incidence of herpes zoster was significantly higher among HIV-seropositive men (29.4 cases/1000 person-years) than among HIV-seronegative men (2.0 cases/1000 person-years); the overall age-adjusted relative risk (RR) was 16.9 (95% confidence interval [CI], 8.7–32.6). When compared with that of age-matched population controls from 1945 to 1959, the incidence of zoster was significantlyhigher among seropositive men (RR, 26.7; 95% CI, 19.3–37.1) and slightly higher among seronegative men (RR, 1.85; 95% CI, 1.0–3.3); the latter may reflect increasing background rates over several decades. The risk of herpes zoster was not associated with duration of HIV infection and was not predictive of faster progression to AIDS.