The preceding paper dealt with the occurrence of hyperbilirubinemia with levels comparable to those encountered in hemolytic disease of the newborn, in full-term and premature infants exhibiting no evidence of either hemolytic or hepatic disease. The serum bilirubin in some of the patients in this group of 46 infants was of a magnitude which in hemolytic disease would be associated with a high incidence of kernicterus and is therefore generally regarded as an indication for exchange transfusion. The rationale of an exchange transfusion in the treatment of hemolytic disease may not be generally agreed upon, but it would seem that most investigators perform the procedure with the chief aim of removing vulnerable sensitized red cells, thereby decreasing the potential for bilirubin formation resulting from the accelerated breakdown of hemoglobin. Secondarily an exchange transfusion can be expected to remove a certain amount of bilirubin directly by the exchange of plasma involved