OBJECTIVES: To quantify the association between prevalent or incident Herpes simplex virus type-2 (HSV2) infection and the incidence of HIV seroconversion among adults in the general population in rural Tanzania. STUDY POPULATION: Adults aged 15-54 years sampled randomly from 12 rural communities in Mwanza Region, Tanzania and recruited to a randomized trial of improved treatment of sexually transmitted diseases. STUDY DESIGN: Unmatched case-control study nested within trial cohort. METHODS: Participants included 127 cases who seroconverted to HIV during the 2-year follow-up period and 636 randomly selected controls who remained HIV negative. Subjects were tested for HSV2 serology at baseline and follow-up, and associations between HIV and HSV2 were analysed with adjustment for socio-demographic and behavioural factors. RESULTS: After adjusting for confounding factors, a strong association between HSV2 infection and HIV seroconversion was observed in men (test for trend: P < 0.001), with adjusted odds ratios (OR) of 6.12 [95% confidence interval (CI), 2.52-14.9] in those HSV2 positive at baseline, and 16.8 (95% CI, 6.06-46.3) in those acquiring HSV2 infection during follow-up. A weaker association was observed in women (tests for trend: P = 0.14), with adjusted OR of 1.32 (95% CI, 0.62-2.78) and 2.36 (95% CI, 0.81-6.84), respectively. Population attributable fractions of incident HIV infection due to HSV2 were estimated as 74% in men and 22% in women. CONCLUSIONS: The results suggest that HSV2 plays an important role in the transmission of HIV infection in this population. There is an urgent need to identify effective HSV2 control measures in order to reduce HIV incidence in Africa.