1. The hemoglobin concentration in the cord blood of infants affected with hemolytic disease is very closely related to their chance of survival. If the hemoglobin concentration is plotted against the chance of survival, a sigmoid curve is obtained. Probit analysis can be applied and the results of treatment can then be summarized in terms of the severity of the cases treated. For example, in the present series the cord hemoglobin concentration corresponding to a 50 per cent chance of survival was 8.09 Gm./100 ml.
2. Application of probit analysis makes it possible to compare results obtained in different centers, for due allowance is made for any difference in severity in the cases included in the different series.
3. In the present series, despite uniform treatment (exchange transfusion), over-all mortality fell from 37 percent in 1947 and 1948 to 3 per cent in 1950. It is shown that this can be entirely explained by a decrease in the number of severely affected infants.
4. Predictions of survival based on the cord hemoglobin concentration and cord bilirubin concentration combined are not significantly better than predictions based on cord hemoglobin concentration alone.
5. Infants with positive direct Coombs tests whose cord hemoglobin concentrations exceed 15.5 Gm./100 ml. have some risk (approximately 1 in 12) of developing kernicterus, if untreated.