The spermicide nonoxynol-9 (N-9) has been used as a contraceptive for over 30 years, but the use of a vaginal spermicide and condoms for the prevention of sexually transmitted infections has not been examined in randomised studies. We report a single-blind randomised field trial to assess the effect of N-9 film on the rate of gonococcal and chlamydial cervical infection in women at high risk of these diseases. 343 women were randomly assigned to use either condoms and N-9 (186 women) or condoms and a placebo (157). Compliance with condom use was much the same in the two groups. Overall, N-9 reduced the rate of cervical infection by 25% (rate ratio [RR] 0.75, 95% confidence interval [Cl] 0.5-1.1); in women who used N-9 for more than 75% of their coital acts the infection rate was reduced by 40% (RR 95% Cl 0.3-1.0). The rate of yeast vulvovaginitis or genital ulcers was not higher in N-9 users than in placebo users, but the rate of symptomatic irritation was increased by 70% (RR 95% Cl 1.1-2.6) among N-9 users. Condom use was more protective against cervical infection than N-9 use. The rate of infection was 50% (RR 95% Cl 0.3-0.7) lower with 75% than with 0-50% condom compliance. The use of a vaginal N-9 spermicide with condoms whenever possible seems to be a better strategy than the use of condoms only for prevention of gonococcal and chlamydial cervical infection.