We examined sexual behavior as a risk factor for Kaposi's sarcoma-associated herpesvirus (KSHV) infection and examined the relation between KSHV seropositivity and development of KS in cross-sectional and cohort studies of 130 homosexual men diagnosed with AIDS in Sydney, Australia during the period from 1991 to 1993. KSHV serology was measured using antibody tests to latency-associated nuclear antigen (LANA) and lytically expressed open reading frame (ORF) 65.2. In the cross-sectional analysis, 52% (68) of study subjects were KSHV-seropositive by either assay. KSHV-seropositive men were significantly more likely to be seropositive to both herpes simplex type 2 (odds ratio [OR] 3.0; 95% confidence interval [CI], 1.2-7.5 for LANA and OR 2.8; 95% CI, 1.3-6.0 for ORF 65) and hepatitis A virus (OR 2.2; 95% CI, 1.1-4.5 for ORF 65). KSHV-seropositive men reported nonsignificantly more casual sexual partners and were nonsignificantly more likely to report insertive oroanal contact with casual partners. These data suggest that KSHV might be sexually transmitted among homosexual men. Men were observed until October 1996 for development of KS. Those seropositive to either KSHV assay at baseline were more likely than the seronegative to develop KS during follow-up (rate ratio [RR] 4.4; 95% CI, 1.9-10.2). Of those seropositive for KSHV, 53% developed KS.