312 THE AMERICAN MONTHLY [Oct., 



This now passes through and the leucocyte regains its 

 former shape, the granules becoming distributed over its 

 surface. 



In several instances, while the leucocyte was in the 

 act of penetrating the vessel wall, it was swept from it 

 by the blood stream, but the granules still held their po- 

 sition in the part of the cell away from the penetrating 

 process, and more than this, in one case a leucocyte so 

 swept away, adhered again to the wall, while still in the 

 field, and penetrated it with the self-same process, pass- 

 ing through all the stages mentioned. 



These phenomena which I have just mentioned indi- 

 cate to my mind that the leucocyte penetrates the ves- 

 sel wall largely by its own cell activity, and I am con- 

 vinced that such is the fact. 



As the blood stream becomes slower and slower, the 

 process is obscured by the accumulation of leucocytes, 

 and when at last the current stops, the vessels become 

 filled with them, and diapedesis can no longer be 

 watched. To see the process the best, a point in the 

 vessel where the blood stream is quite rapid and the 

 leucocytes few in number, is most favorable. 



To sum up then, the following facts are those which I 

 think are most important and which I have not as yet 

 seen any other account of. 



1. The leucocytes, in penetrating the vessel wall, send 

 off" processes similar in shape and size. 



2. If a leucocyte, while in the act of penetrating the 

 vessel, was swept off" into the blood current it retained its 

 process as long as I could see it. 



3. The granular matter in the leucocytes becomes ar- 

 ranged in the portion of the cell furthest from the pene- 

 trating powers so that the process contains only clear 

 protoplasm. 



4. If the leucocyte be swept away while penetrating, 

 the granules still retain their position, and in one case 



