In a special project area of rural India, the age-specific mortality rates of a birth cohort 0-59 months old were calculated from 1991 to 1999. The mortality rates were lower than the national average, but the common causes of child deaths were similar. Since 1985, when the Universal Immunization Programme of India introduced routine measles vaccination, there has been a proportionate decline in child deaths unlike other parts of India. The absolute numbers of child deaths from 'all' causes have also declined significantly over the same period of time. The vaccination coverage in the project area is higher than the national average. This suggests a possible 'beneficial' non-specific effect of measles vaccination on child survival over this time period.