A case-control study of risk factors for child diarrhoeal disease was undertaken in a rural area of Nicaragua. Some 1229 children under the age of five were matched with an equal number of children of the same age presenting with other illnesses unrelated to water and sanitation. The main types of water supply were sampled at monthly intervals and tested for the presence of faecal coliforms in order to characterize their microbiological quality. In spite of marked differences in water quality between the different types of water supply, no relationship was found with diarrhoea morbidity. In contrast, there was a statistically significant association between water availability and diarrhoea morbidity. Children from homes with water supplies over 500 meters from the house had incidence rates of diarrhoea 34% higher than those of children from houses with their own water supply. Owning a latrine was not found to be significantly related to diarrhoea morbidity. A mother's level of schooling was inversely correlated with the frequency of diarrhoea in her children. A significant association was also found between the number of children under the age of five living in the house and the incidence of diarrhoea. These effects remained significant after controlling for confounding variables by conditional logistic regression.