Earlier studies have suggested that general measles vaccination programmes should not be made a priority in developing countries because the presumably malnourished children saved from measles are likely to die from something else. Recent community studies indicate, however, that malnutrition is not the cause of high measles mortality. In an urban community in Guinea-Bissau, child mortality has been registered for a period of 3 years; 1 year before and 2 years after the introduction of a general measles vaccination program. In the years following the introduction of measles vaccination, mortality for children aged 6 to 35 months has significantly diminished. Though this is not a controlled study of vaccinated and unvaccinated children, much of the reduced mortality can apparently be attributed to the protective effect of measles vaccination. Children with a history of earlier measles infection had a significantly higher mortality rate than children vaccinated against measles. Rather than being a mechanism of natural selection taking the weakest children, measles apparently aggravates the condition of many children, leading to delayed excess mortality. In areas where the case fatality rate is high, vaccination against measles should be made an indispensable part of primary health care.