Using data on incidence and secondary attack rates, we examined the protective efficacy of high-titre Edmonston-Zagreb (EZ) and Schwarz (SW-HT) measles vaccines administered at 5 months. Control children were assigned to placebo at age 5 months and standard Schwarz (SW-std) measles vaccine at 9-10 months of age. A large proportion of measles cases was verified serologically. Though high-titre vaccines seemed to be protective before 10 months of age, a significant reduction in disease could not be demonstrated due to low incidence of measles. After 10 months of age, SW-std given at 10 months gave a vaccine efficacy of 100% and induced better protection than SW-HT (P = 0.030) and EZ-HT (P = 0.128) administered at 5 months. In studies of secondary attack rates in the compound, vaccine efficacy was 91% (75%-97%) for EZ-HT, 85% (40%-96%) for SW-HT, and 100% for SW-std. Attack rates were correlated with intensity of exposure (P = 0.0006), being much higher for children exposed in the same hut than for those living in the same compound but in a different household (relative risk = 3.36 [1.32-8.57]). The attack rate was significantly lower among vaccinated than unvaccinated children with no detectable measles antibody (relative risk = 0.41 [0.18-0.93]). In rural areas with a high coverage in the surrounding community, a single dose at 9-10 months may provide sufficient protection. Since high-titre vaccines have been associated with higher mortality than SW-std, further improvements in measles control before 9 months may require two-dose strategies with standard vaccines.