Abstract
One hundred and fifty-three water samples from rural Nicaragua were examined for the presence of faecal coliforms during both wet and dry periods. A linear model was fitted by analysis of covariance with the logarithm of the faecal coliform count as the dependent variable. As expected, traditional water sources were grossly contaminated at all times whereas piped water sources were much cleaner. Hand-dug protected wells had significantly higher levels of faecal contamination than unprotected riverside wells and springs during the dry season. The possible reasons for this unexpected finding are discussed. A close association between rainfall and faecal contamination was demonstrated but the effect of rainfall depended on the type of water source. An association between water quality and the size of the community served by the source was also detected. The finding that stored water was usually more contaminated than fresh water samples is consistent with the results from other studies. Since it is unusual for water quality to be inversely correlated with accessibility, this study site would be suitable for investigating the relative importance of water-borne versus water-washed transmission mechanisms in childhood diarrhoea.