PURPOSE: Acquisition of human papillomavirus (HPV) in women occurs predominantly through vaginal sex. However, HPV has been detected in girls reporting no previous sex. We aimed to determine incidence and risk factors for HPV acquisition in girls who report no previous sex in Tanzania, a country with high HPV prevalence and cervical cancer incidence. METHODS: We followed 503 adolescent girls aged 15-16 years in Mwanza, Tanzania, with face-to-face interviews and self-administered vaginal swabs every 3 months for 18 months; 397 girls reported no sex before enrollment or during follow-up; of whom, 120 were randomly selected. Samples from enrollment, 6-, 12-, and 18-month visits were tested for 37 HPV genotypes. Incidence, clearance, point prevalence, and duration of any HPV and genotype-specific infections were calculated and associated factors were evaluated. RESULTS: Of 120 girls who reported no previous sex, 119 were included, contributing 438 samples. HPV was detected in 51 (11.6%) samples. The overall incidence of new HPV infections was 29.4/100 person-years (95% confidence interval: 15.9-54.2). The point prevalence of vaccine types HPV-6,-11,-16, and -18 was .9%, .9%, 2.0%, and 0%, respectively. Spending a night away from home and using the Internet were associated with incident HPV, and reporting having seen a pornographic movie was inversely associated with HPV incidence. CONCLUSIONS: Incident HPV infections were detected frequently in adolescent girls who reported no previous sex over 18 months. This is likely to reflect under-reporting of sex. A low-point prevalence of HPV genotypes in licensed vaccines was seen, indicating that vaccination of these girls might still be effective.