To examine the extent of hepatitis B virus infection (HBV) in an inner-city community, we determined the prevalence, incidence, and correlates of HBV seroreactivity in a representative sample of unmarried whites, African-Americans, and Hispanics living in San Francisco during 1988 to 1989 and again 1 year later in 1989 to 1990. Unmarried men and women aged 20 to 44 years were surveyed in a random household sample drawn from three neighborhoods of varying geographic and cdtural characteristics. Hepatitis B infection was determined by testing specimens for antibody to hepatitis B core antigen (anti-HBc). Of blood samples available from 1,108 participants from the initial survey, 159 (14%) were anti-HBc positive. There was a strong positive association between anti-HBc positivity and positive serological tests for human immunodeficiency virus, herpes virus type 2, and syphilis. In women and heterosexual men, after controlling for other variables, anti-HBc positivity was significantly associated with older age (P < .001), nonwhite ethnicity (P < .Ol), less education (P < .05), injection drug use (P < .001), being paid for sex (P < .05), and lifetime number of sexual partners (P < .05). Among homosexually active men, after controlling for other variables, anti-HBc positivity was significantly associated with nonwhite ethnicity (P < .001), injection drug use in a sexual partner (P < .05), and number of lifetime sexual partners (P < .05). There were 19 (3.2%) incident HBV infections. Participants who used injection drugs (relative risk [RR], 8.2; 95% confidence interval [GI] 3.9 to 17.41, crack cocaine (RR, 3.5; 95% CI, 1.2 to 9.6), or who were paid for sex (RR, 7.6; 95% GI, 1.4 to 41.1) were more Abbreviations: HBV, hepatitis B virus; STD, sexually transmitted disease; HIV, human immunodeficiency virus; anti-HBc, antibody to hepatitis B core antigen; HSV-2, herpes simplex virus type 2; PR, prevalence ratio; CI, confidence interval; OR, odds ratio; RR, relative risk.