To evaluate the importance of public-domain transmission of pathogens in drinking-water, an intervention study was carried out by chlorinating the public water-supply system in a village in Pakistan. The water quality improved and reached a geometric mean of 3 Escherichia coli per 100 mL at the last standpipe of the water-supply system. Drinking-water source used and the occurrence of diarrhoea were monitored on a weekly basis over a six-month period among 144 children aged less than five years in the village. In this group, the children using chlorinated water from the water-supply scheme had a higher risk of diarrhoea than children using groundwater sources, controlled for confounding by season and availability of a toilet and a water-storage facility. The incidence of diarrhoea in the village (7.3 episodes per 10(3) person-days) was not statistically different from that in a neighbouring village where most children used water from a non-chlorinated water-supply system with very poor water quality. In this study area, under non-epidemic conditions, the reduction of faecal bacteria in the public drinking-water supply by chlorination does not seem to be a priority intervention to reduce childhood diarrhoea. However, the study was of limited size and cannot provide conclusive evidence.