An iodinator was fitted to the existing gravity-fed water-supply of a remote village in Sarawak, Malaysia, where goitre was endemic. Within nine months, the prevalence of goitre had been reduced from 61% to 30%, with 79% of goitres showing visible reduction in size. All subjects were clinically euthyroid before and nine months after the start of iodination, although pre-treatment serum thyroid-stimulating hormone (TSH) concentrations varied from normal up to 24 mU/l. Before treatment basal serum triiodothyronine (T3) and thyroxine (T4) concentrations were typical of endemic goitre with a low mean serum T4 (80 +/- 30 [SD] nmol/l) and a slightly raised mean serum T3 (2.3 +/- 0.7 nmol/l). After iodination, circulating TSH concentration was generally undetectable (less than 0.1 mU/l), mean T3 concentration was unchanged, but the mean T4 rose significantly to 109 +/- 41 nmol/l (p less than 0.01). Urinary iodine concentrations fluctuated; this largely reflected intermittent blockage of the iodinator, but concentrations became consistent with a return to the iodine-replete state. There was no evidence of the Jod Basedow effect in the group studied. Iodinated water was more convenient to distribute than iodised salt and is less likely to cause Jod Basedow phenomenon than are injections of iodised oil. Moreover, iodination of water is effective in killing most microorganisms and this additional benefit could contribute significantly to village health.