A prospective randomized study of 100 well-nourished infants with acute gastroenteritis resulting in dehydration and acidosis was carried out at the Jackson Memorial Hospital, Miami from 1981 to 1983. Patients were randomly assigned to receive either standard intravenous therapy or oral rehydration. Infants in the latter group first received solution A containing 75 mEq/L sodium, 30 mEq/L potassium, 75 mEq/L chloride [corrected], 30 mEq/L bicarbonate, and 2 gm/dL glucose [corrected]. After ad libitum feeding for six hours, solution B containing 50 mEq/L sodium, 30 mEq/L potassium, 50 mEq/L chlorine, 30 mEq/L bicarbonate, and 3 gm/dL [corrected] glucose was given. With three exceptions (6%), oral rehydration was comparable to the intravenous regimen in clinical estimates of improvement, although the oral group had more stools in the first day. The oral group had faster correction of acidosis and a sustained rise in serum potassium concentration, whereas in the intravenous group the potassium concentration showed first a drop with a later increase, but levels were at all times below those in the oral group. Although potassium was given from the beginning of oral rehydration, and at a higher concentration than recommended by the World Health Organization, no hyperkalemia occurred. We concluded that oral therapy is safe, less expensive for patients, and more convenient for the medical and nursing staffs.