Objective To study in depth sexual history and sexual behaviour variables as risk factors for HIV-1 infection in a rural Ugandan population. Methods Following a socioeconomic and serological survey of a rural population in Masaka District, south-west Uganda, 233 randomly selected HIV-1-positive cases and 233 negative controls matched on age and village of residence, were invited in October 1990 to participate in a case-control study. A total of 132 cases and 161 controls attended for in-depth investigation including an interview about sexual behaviour. Results The factor most strongly associated with increased risk of infection was a greater number of lifetime sexual partners, with odds ratios (OR) of 2.1 and 4.9 for those reporting 4–10 and 11 or more partners, respectively, compared with those reporting less than four partners. Having only one sexual partner did not provide complete protection, a total of seven (one male, six female) subjects reporting only one sexual partner were HIV-1-positive. Other significant factors were a history of genital ulcers |OR, 2.9; 95% confidence intervals (CI), 1.0–9.1) and not being a Muslim (OR, 5.4; 95% Cl, 1.8–16.5) suggesting a possible protective effect of circumcision. There was a suggestion that those who married within the last 7 years (OR, 2.4; 95% Cl, 0.9–6.1) and men exposed to menstrual blood (OR, 5.7; 95% Cl, 0.7–49.8) were at an increased risk of HIV-1 infection. Conclusions These results confirm the predominant role of sexual behaviour in the HIV-1 epidemic. Of particular concern is the observation of HIV-1 infection among those reporting only one partner. Where HIV-1 infection is widely distributed in the general population, risk reduction strategies should, in addition to the promotion of partner reduction, place strong emphasis on safe-sex techniques.