258 MRS. ERNEST HART. 



more of the pale corpuscles, and in fig. S, which is drawn faith- 

 fully from the preparation, this departure of the fibrils of fibrine 

 from the pale corpuscles is, I think, demonstrated. 



In the well-known and generally accepted theory of A. 

 Schmidt as to the formation of fibrine, it is assumed that the 

 red corpuscles part with fibrino-plastin before the formation of 

 fibrine can be accomplished. In the course of my research this 

 theory seemed to me to be often capable of physical demonstra- 

 tion by the appearances the red corpuscles sometimes present, for 

 they occasionally appear to be in the act of discharging part of 

 their contents. This appearance is shown in fig. 9, e. The 

 crescentic corpuscles, which are also figured (d., fig. 9), seem to 

 show a loss of substance. They were very frequently found both 

 in human and in rabbit's blood. Of these crescentic corpuscles 

 I am unable to give any satisfactory explanation. 



It may be objected that the tails of the colourless and pale 

 corpuscles are produced by currents of the serum. This objec- 

 tion was present to my mind the whole time that these experi- 

 ments were being performed, and great care was taken to ascer- 

 tain if the processes pointed in the direction taken by the fluid. 

 In fig. 5 it will be noted, however, that the tails of the corpuscles 

 point in opposite directions, and their position does not seem 

 to me at all to justify the hypothesis that they may be produced 

 by eddies ; also it will be observed that in fig. 5 both ends of 

 the corpuscles are sending out processes. In fig. 3, moreover, 

 it may be seen that the processes do not take the direction of 

 any possible current, but that they regularly divide and ramify 

 like the branches of a tree. In fig. 10 two pale corpuscles (c) 

 also will be seen sending out branches, which cross and lie over 

 one another ; this one fact militates strongly against the explana- 

 tion by currents. 



Again, it may be objected that the processes sent out by the 

 corpuscles and the bands and fibres represented in figs. 6 and 7 

 are not true fibrine. This is, I allow, a weighty objection. 



As the preparations are all fixed by osmic acid vapour 

 they are, therefore, incapable of being tested by chemical re- 

 agents for fibrine. To eliminate this error I may, however, 

 state that I defibrinated fresh rabbit's blood and treated 

 the defibrinated blood in the way already described, with the 

 result of finding many pale corpuscles mixed with the red, 

 (the pale corpuscle being, I imagine, the very first step in 

 the formation of fibrine), but extremely few transparent cor- 

 puscles, a few crescents, and no fibrils. I also treated the serum 

 of blood- clot in the same way, and found neither transparent cor- 

 puscles nor fibrils. To be certain, also, that the transparent 

 branching corpuscles were not accidental productions, minute 



