To investigate the optimal timing for treatment of small premature infants, we performed a double-blind, controlled trial of indomethacin therapy on the first day of life in 104 infants weighing between 700 and 1300 gm. Infants were given indomethacin or placebo at a mean age of 15 hours. Eleven of the 56 infants given placebo developed large left-to-right shunts through a patent ductus arteriosus. In contrast, only two of the 51 infants given indomethacin developed large shunts (P less than 0.025). There were no significant differences in incidence of surgical ligation, duration of oxygen therapy, duration of endotracheal intubation, days required to regain birth weight, or incidence of complications. However, the power of the tests of significance was low because of the small number of patients. Thus, although the incidence of large left-to-right ductus shunts was decreased in the indomethacin group, morbidity was not otherwise altered for the entire group of patients, possibly because of the relatively low incidence (21%) of large shunts in the placebo group. We conclude that although treatment with indomethacin on the first day of life appears to be safe, there is little advantage to its use in centers where the incidence of large shunts through a patent ductus arteriosus is relatively low.