Fifty-six premature infants with a mean gestational age at birth of 30 weeks were randomly assigned to a transfusion group, for whom the hemoglobin level was kept above 10.0 g/dL, and a nontransfusion group, who were transfused only for specific clinical indications. The groups were followed up longitudinally with weekly determinations of reticulocyte count, the partial pressure of oxygen at which 50% of hemoglobin is saturated, and hemoglobin F percentage, as well as weight gain, length of stay, hospital cost, and frequency and severity of apnea. At birth, there was no significant difference in birth weight, gestational age, and hemoglobin level between the two groups. At discharge, laboratory differences were noted between the two groups, but there was no clinical difference. We found no clinical advantage to the use of "booster" RBC transfusions in growing premature infants.