75 



2. Tlie liorses SH 1, 147, 160, and 172, showing the litgliest 



liaemolytic indices for horse and mule blood, had also been 

 most frequently injected with the greatest quantities of blood, 

 but a strict proportionality between the amount of blood 

 injected and tlie value of tlie liaemolytic index cannot be 

 demonstrated. The haemolytic index seems rather to a great 

 extent to be independent of the amount of blood injected 

 and to differ individually ; for instance, in liorse 611, after 

 liaving received 4,500 c.c. the serum had an index of 0'166 for 

 mules. Horse 612, treated in tlie same way, even with tlie 

 blood of the same horses, showed an index of 0'204; tlie 

 haemolytic index of serum 624 was only 0'075, though the 

 quantity injected into tliis horse amounted to nearly the 

 double of that injected into horses 611 and 612. On the other 

 hand, althougli horses 615 and 624 had both leceived 8,000 c.c, 

 the haemolytic ijower of serum 615 is three and four times 

 greater respectively than tliat of serum 624. 



3. Tlie haemolytic index varied considerably in the same horse, 

 (a) The haemolytic index increased when further injections 



were made (horse 611 for mules and horses, horse 612 for 

 horses, horse 615 for mules and horses). 



(&) The haemolytic index decreased, notwithstanding the higher 

 liyperimmunisation (horse 160 for horses, and horse 612 

 for mules). 



(c) The haemolytic index increased without further injection of 

 blood (horse 160 for mules). 



4. A serum of which the haemolytic index was 1"3 or lower 



could safely be injected into horses ; it did not produce 

 haemolysis in vivo. 



