Ninety four children aged less than 5 years with diarrhoeal dehydration and acidosis were treated randomly with either World Health Organisation (WHO) oral rehydration solution containing sodium chloride, potassium chloride, sodium bicarbonate and glucose or an oral solution with tripotassium citrate monohydrate replacing the sodium bicarbonate and potassium chloride in the WHO solution. Fifty five children (58%) were hypokalaemic (potassium less than 3.5 mmol/l) on admission. All but two in the citrate group were successfully treated. There were no significant differences in rehydration solution intake, stool output, gain in body weight, and fall in plasma specific gravity and haematocrit between the two treatment groups after 48 hours' treatment. Significant improvement in the serum potassium concentration was observed in the hypokalaemic children receiving potassium citrate solution compared with children receiving WHO solution after 24 and 48 hours' treatment. None developed hyperkalaemia. Although children receiving potassium citrate solution corrected their acidosis at a slower rate than the WHO solution group during the first 24 hours, by 48 hours satisfactory correction was observed in all. Tripotassium citrate can safely replace sodium bicarbonate and potassium chloride and may be the most useful and beneficial treatment for diarrhoea and associated hypokalaemia.