I possess, on the other hand, evidence that an alteration may take place in some other diseases, such as pernicious anæmia and familial, or acholuric, jaundice. Evidence for the former was provided recently by a patient whose condition was typical, clinically, of the last stages of the disease. Her corpuscles, tested with stock sera, belonged to Group II, but her serum, tested directly with the corpuscles of prospective donors known to belong to Group IV, agglutinated these vigorously, so that a transfusion could not safely be performed. The same phenomenon has been found by other observers. In acholuric jaundice there is a progressive destruction of red corpuscles in the patient’s circulation. This appears to be connected in some way with an abnormal functioning of the greatly enlarged spleen, since the destruction of corpuscles ceases almost at once when this organ is removed. There seems to be, in addition, an alteration in the blood reactions. In a case which I tested recently, the patient’s corpuscles were quickly agglutinated by serum of Group III, and he therefore nominally belonged to Group II. His serum, however, when separated and tested against other bloods of known groups gave, in addition to a rapid agglutination of corpuscles belonging to Group III, a definite, though slower, agglutination of corpuscles belonging to Groups II and IV, showing that it had acquired abnormal properties. It is possible that there are similar alterations of reactions in other pathological conditions. The instances mentioned above suggest that the serum is affected rather than the corpuscles, but further investigations are needed. It is an observed fact that blood outside the body soon develops the property of auto-hæmolysis. If blood is drawn from a vein, put into a test-tube, and allowed to clot, then after twenty-four hours or more the serum which has separated from the clot begins to be tinged with hæmoglobin, even though it has remained absolutely sterile. It appears, therefore, that the serum develops a hæmolysin and the corpuscles the corresponding iso-hæmolysin, the interaction of which results in the breaking up of corpuscles. If this process takes place in normal blood outside the body, it would not be surprising to find that it may also occur abnormally inside the body. This actually happens in the condition known as paroxysmal hæmoglobinuria. The pathology of the disease is obscure, but it seems that a hæmolysin develops in the serum as the result of cooling in the extremities and hæmolysis takes place when the cooled serum is again warmed by being restored to the general circulation. The presence of this hæmolysin in addition to the normal hæmolysins has been demonstrated by Moss. It is possible that a similar though less acute change takes place in acholuric jaundice. Blood transfusion, therefore, is not likely to be efficacious in such conditions, since the transfused corpuscles may be destroyed whatever