The incidence of kernicterus in the neonate has been dramatically reduced with treatment by exchange transfusion in recent years. The clinical decision to exchange transfuse is at present based on total bilirubin measurements, acid-base status, degree of prematurity, etc. A wealth of experimental and clinical evidence suggests that the unbound (‘free’) bilirubin concentration is probably the most critical parameter in establishing the risk for bilirubin encephalopathy. We have measured free bilirubin in plasma by measuring its initial rate of oxidation by peroxidase in the presence of ethylhydroperoxide. Despite the problems of bilirubin standardization this method has proved to be sensitive and reasonably precise. We have measured free bilirubin in the plasma of all infants in our neonatal ward who had a plasma total bilirubin concentration > 200 μmol/l. A total of 190 patients have been studied over a period of 18 mth. Levels of plasma free bilirubin of > 1 μmole/l were taken to be abnormal; 25 exchange transfusions were carried out on 19 patients using current clinical criteria. The pre-exchange free bilirubin was raised in 7 of these cases. In addition, 6 patients who were not exchanged were found to have plasma free bilirubin >1 μmole/l. Although no clinical evidence of kernicterus has been found in any of these children to date, long-term follow-up is to be attempted to check for any neurological sequelae as a result of high free bilirubin levels in the neonatal period.