We conducted a case-control study of antecedents of bronchopulmonary dysplasia (BPD) in 223 infants enrolled in a prospective, randomized clinical trial of phenobarbital prophylaxis for intracranial hemorrhage. The trial took place at three Boston neonatal intensive care units between June 1981 and April 1984. The 76 babies with BPD had radiographic evidence of the condition and required oxygen therapy for 28 days or more. All 147 control babies survived until day 28 of life without meeting either of these criteria for BPD. Compared with control infants, those with BPD received greater quantities of total, crystalloid, and colloid fluids per kilogram per day in the first 4 days of life. In addition, infants with BPD generally had a net weight gain in the first 4 days of life in contrast to the normal pattern of weight loss seen in control infants. Finally, the infants with BPD were more likely to be given a clinical diagnosis of patent ductus arteriosus and to have received furosemide on days 3 and 4 of life. From these observations we infer that early postnatal phenomena such as excessive fluid therapy may be important in the pathogenesis of BPD.