The reaction is, however, found to become hæmolytic in the recipients of transfusions, and it is then selective. It seems, therefore, that the group reactions may not be as clearly defined as was at one time supposed. Probably there are slight incompatibilities of an unknown nature between individuals of the same or compatible groups. These are very seldom of any consequence in a first transfusion, but become accentuated as the result of “sensitization,” and in later transfusions have a pronounced influence. This “over-lapping” of groups has been mentioned on another page. It must not be supposed that any untoward results follow repeated transfusions as a general rule, for usually no such effect is observed. In order, however, to minimize the risk, it may be suggested that the following precautions should be taken: (1) The donor should be actually of the same group as the recipient, and not merely of a theoretically compatible group; a patient, for instance, of Group II should receive blood of Group II rather than of Group IV. (2) The same donor should not be used for the later transfusions of a series, on the grounds that the sensitization appears to be an individual rather than a group phenomenon. (3) In performing the later transfusions, the blood should be given at first very slowly, so that it may be discontinued at the first appearance of any signs of a reaction. CHAPTER VI THE CHOICE OF BLOOD DONOR The physiology of blood groups having been examined, the principles governing the choice of a blood donor can be more readily understood. It is evident that this choice is determined largely by blood groups, and in the present chapter therefore the clinical methods of testing for the groups will be described.