In response to a recorded increasing incidence of diarrhoea in Tumpat District, Malaysia, a case-control study was performed to identify modifiable risk factors for the transmission of diarrhoea, in children aged 4-59 months. Ninety-eight pairs of children, matched on age and sex, were recruited prospectively from health centres. Exposure status was determined during a home visit. Interviewers were 'blinded' as to the disease status of each child. Odds ratios were measured through matched pair analysis and conditional logistic regression. Risk factors for diarrhoea identified were: reported--drinking of unboiled water, storage of cooked food before consumption and bottle feeding; and observations--animals inside the house and absence of washing water in latrines. Water quality, source of drinking water, reported hand washing behaviour, indiscriminate defecation by children, cup use and the absence of a functional latrine were not associated with diarrhoea. Nonsignificant associations were found for: accessibility of washing water source, type of water storage container and use of fly covers for food.