the apparent blood group of the patient. Some of the facts of auto-hæmolysis have been recently investigated by Bond, but it is not necessary to give the details here. He concludes that the development of auto-hæmolysins, which are non-specific and independent of the specific hæmolysins of the blood groups, has a biological significance in the history of the red corpuscle, and is a product of ageing. The biochemistry, however, of the process remains at present entirely unknown. The necessity for careful blood grouping in every case before performing a transfusion has now been sufficiently emphasized, but before proceeding to the description of the methods of choosing a donor and of grouping, a possible danger must be mentioned which may arise even when the blood groups are known. In the preceding chapters references were made to the effects which have been observed to follow repeated transfusions given in the treatment of a condition such as pernicious anæmia. In such cases, although the groups were ascertained, and the bloods were also tested directly against one another without any incompatibility being detected, yet when the third or fourth transfusion was given, symptoms of toxæmia followed, sometimes with hæmolysis. The death of the patient has even been hastened in this way. A very striking instance of this phenomenon, which has been recently reported (278), will serve to bring home the reality of the danger. A boy was transfused by the citrate method with blood from his father, and this was followed only by a mild febrile reaction such as is often observed. Eighteen days later a second transfusion with blood from the same donor was performed, and after 150 cc. had been given, a severe reaction resulted, which was followed later by pronounced hæmoglobinuria. In this case the bloods of donor and recipient had been tested against one another directly, but this was not repeated, and the groups were not ascertained until afterwards. Probably there was some error in the original test, for it afterwards appeared that the boy belonged to Group I and his father to Group III, so that there should have been agglutination of the boy’s corpuscles by his father’s serum outside the body. Nevertheless, Group I individuals have been called the “universal recipients,” and no ill effects are usually observed whatever blood be used for transfusing them. In the other cases already mentioned a reaction followed the later transfusions, even when the donor and recipient belonged to the same group. It appears that by repeated transfusions the recipient becomes as it were sensitized to the blood of another individual even of the same group, and consequently great caution must be used in giving the later transfusions of a series. Some light is thrown on this question by the observations of Ottenberg, already referred to, concerning the artificial production of iso-hæmolysins in cats. In these animals iso-agglutinins are found, but iso-hæmolysins seldom or never.