Recent studies in sub-Saharan Africa have shown that fertility is reduced among HIV-infected women compared with uninfected women. The size and pattern of this fertility reduction has important implications for antenatal clinic-based surveillance of the epidemic and also for estimates and projections of the demographic impact of the epidemic. This paper examines the association between HIV and fertility in Kisesa, a rural area in Tanzania, where HIV prevalence among adults is about 6% and gradually increasing. The analysis is based on data obtained through a demographic surveillance system in Kisesa during 1994-98 and two large sero-surveys of all residents in 1994-95 and 1996-97. The HIV-associated fertility reduction among women was investigated by estimating fertility rates by HIV status and prevalence rates by fertility status. A substantial reduction (29%) was observed in fertility among HIV-infected women compared with HIV-uninfected women. The fertility reduction was most pronounced during the terminal stages of infection, but no clear association with duration of infection was observed. Use of modern contraception was higher among HIV-infected women. However, both among contracepting and non-contracepting women, a substantial reduction in fertility was seen among HIV-infected women.