852 
Proceedings of the Royal Society 
sent. With this reservation, his statement may he accepted as at 
east corroborative of that of Worm-Muller. 
The result of my own experiments, made on rabbits, go in the 
main to confirm the accuracy of this estimate of Worm-Miiller and 
Quincke. 
The method of transfusion adopted by myself differed somewhat 
from that of previous observers. Instead of being introduced 
directly into the circulation through a vein, the blood was injected 
into the peritoneal cavity, the absorption from which, as is well 
known, is so rapid and continuous, that peritoneal injection has 
claims to he considered as a slow method of intravenous injection. 
In some respects, indeed, for experimental purposes it offers advan- 
tages over the latter, since, apart from the greater simplicity with 
which the operation can he carried out, the absorption into the cir- 
culation (mainly through the lymphatics of the diaphragm and the 
thoracic duct) is so continuous and steady, prolonged as it is over 
a period of some 24 to 48 hours, that time is given during 
its progress for the removal of the fluid part of the blood, and hence 
little or no distension of the vascular system is likely to occur after 
the injection of even the largest quantities. 
The method is naturally open to the objection, that the fact of 
the blood corpuscles having been extravasated before reaching the 
circulation may possibly affect their vitality, and thus shorten the 
duration of their life. That a certain number of them, under such 
circumstances always suffer a local death, is certain ; but this number 
is relatively small, and, as will be seen, the results show that the 
vitality of those absorbed is not in any way affected by their tem- 
porary sojourn in the peritoneal cavity. 
The increase in the number of corpuscles in the blood as ascer- 
tained by actual enumeration, after reaching a maximum on the 
second or third day, became gradually less and less, till at the end 
of a certain time, varying somewhat in the different experiments, 
the number of blood corpuscles had returned to their normal, this 
period of time representing, therefore, the duration of life of the 
injected corpuscles in that particular case. 
The experiments were made both with defibrinated and with 
entire blood. The duration of the increase after the injection of 
entire blood was: — In two experiments 26 days, in one 21 days, 
