The assertion that an infant has no blood group was tested by the writer some time ago and shown to be false. On several occasions a newly born infant was tested and found to show well-marked agglutination reactions indicating Groups II or III as the case might be. Even in 1905 it had been shown by Martin that reactions could often be demonstrated between an infant’s corpuscles and the maternal serum, and sometimes between the infant’s serum and the maternal corpuscles. More recently (March 1920) the results of a full investigation into the reactions found in infants and children have been published by W. M. Happ in America. These researches began with the testing of blood from the umbilical cord, and this was seldom found to show the blood reactions as given by the adult. So far the statement quoted above was justified. It is even true that the serum of an infant’s blood will usually not give any reaction at birth or during the first month. The percentage in which it does give a reaction increases with the age of the child; after one year it is usually, and after two years always, established. On the other hand, the agglutination reaction in the corpuscles appears before that in the serum, so that the grouping tested in this way may be present immediately after birth, as I found to be the case. It is possible that the grouping which first appears may afterwards be modified, but any change which occurs is always by the addition of factors and not by their subtraction; thus an apparent Group IV may become a Group II or III, or an apparent Group II or III may become a Group I. It is found that when a reaction is present in both the corpuscles and the serum, the group does not afterwards change. Happ’s conclusion, based on his investigations, was that it is unsafe to transfuse an infant with its mother’s blood without first making the usual tests, and the reasons for this will now be evident. In the first place an infant _may_ be possessed of its final blood reactions very shortly after birth, and should therefore be treated in the same way as if it were an adult. In the second place, although its serum may be without agglutinating powers, so that transfused corpuscles will not be attacked, yet its corpuscles may be possessed of pronounced agglutinophilic properties, so that they may be seriously affected by the serum of transfused blood from an incompatible group. In the third place, as will presently be seen, it is by no means the rule that an infant should belong to the same group as its mother, whatever its blood reactions may be.