This report describes the Vermont-Oxford Trials Network, a voluntary collaborative research network, and summarizes the outcomes and medical interventions for very low birth weight infants at participating centers in 1990. The Vermont-Oxford Trials Network included 36 centers in 1990 (11% university hospitals, 44% university affiliates, 44% nonaffiliated) with a total of 2961 infants weighing 501 to 1500 g (median 73 infants, range 5 to 172). Eighty percent of the infants were inborn and 65% were white. The overall network frequencies for selected interventions and outcomes were as follows: prenatal care, 90%; a complete course of antenatal corticosteroids, 12%; cesarean section, 56%; surfactant therapy, 49%; postnatal steroids for chronic lung disease, 16%; high-frequency ventilation, 4%; patent ductus arteriosus, 31%; necrotizing enterocolitis, 6%; bacterial sepsis, 16%; and intraventricular hemorrhage, 26%. By 28 days, 15% of the infants had died and 8% had been transferred, whereas by discharge 18% had died and 18% had been transferred. There were marked variations among the centers in the frequencies of different medical interventions and in the frequencies of various clinical outcomes. The Vermont-Oxford Trials Network is a unique collaborative research group composed of a broad range of neonatal intensive care units. During 1990 there were considerable differences among the centers in the interventions used and patient outcomes observed. The investigators plan to devote the resources of the Network to a research program of randomized trials and outcome studies so that effective interventions can be identified and the quality of neonatal intensive care can be continuously improved.