The epidemiology of measles has been investigated in Niakhar, a rural area of Senegal, during two periods, 1983-1986 and 1987-1990. Following a major increase in immunization coverage beginning in 1987, the case fatality ratio for all ages declined fourfold from the first to the second period (relative risk (RR) = 0.24, 95% confidence interval (CI) 0.13-0.46). The measles incidence for children under 10 years of age declined by 69% (95% CI 65-72) and the risk of dying of measles by 91% (95% CI 82-95). Vaccinated children who contracted measles had significantly lower case fatality ratio than unvaccinated children with measles (p = 0.038). Children infected by an immunized case tended to have lower case fatality ratio than those infected by an unimmunized index case (p = 0.104) and immunized index cases generated fewer secondary cases than unimmunized index cases (p < 0.001). Respiratory complications were more common in secondary cases infected by an index case with respiratory complications than by an index case without such complications (RR = 1.60, 95% CI 1.08-2.37), which suggests that severe cases give rise to further severe cases. As expected, there was a significant increase in the proportion of vaccinated cases in the second period (RR = 1.41, 95% CI 1.00-1.98). Mean age at infection increased from 4 to 7 years between the two periods and the change in age structure accounted for 20% of the decline in case fatality ratio. Measles immunization may contribute to lower mortality directly through reduced incidence and indirectly through increases in age at infection, less severe infection for immunized cases and changes in transmission patterns leading to reduced severity of measles.