It was at one time believed that the group reactions were clear-cut and absolute rather than relative. At the present time, however, the view is gaining ground that there may be some “over-lapping” of groups, that is to say, a serum may contain agglutinins which give a gross reaction with the corpuscles of one group and a reaction with another group so slight that it can be detected only with difficulty, or alternatively the recipient’s corpuscles may give a definite and limited group reaction, while his serum may cause some agglutination in the blood of a theoretically compatible group. These properties have recently been termed “major” and “minor agglutinins” by Unger, who claims that the possible presence of minor agglutinins makes it advisable to test the recipient’s blood directly against the donor’s in every case. The term “universal donor” commonly applied to Group IV is, in fact, misleading. The blood of Group IV cannot be used indiscriminately with complete impunity. The groups are determined by the major agglutinins, and by these the ordinary gross reactions may be eliminated. Everyone who has used blood transfusions extensively has observed that slight reactions may occur after transfusion with a compatible blood, irrespective of the methods employed. Usually these reactions are slight, and do not in any way prejudice the benefits conferred by the transfusion, but they may become greatly accentuated in the later transfusions of a series, and it is probable that minor agglutinins may be developed in certain pathological conditions. Further reference to these phenomena will be made elsewhere (p. 93). In addition to this, it has been commonly observed that the intensity of the reaction varies greatly with the sera of different individuals of the same group. It has also been stated by Stansfeld that the agglutinating power of the serum of an individual may vary from time to time. As a rule the corpuscles of a person belonging to Group I are not agglutinated with equal rapidity or intensity by the sera of Groups II and III, but the meaning of this phenomenon has not been fully investigated.