PEROPHORA ANNECTENS. 5 I 



or more blunt processes are seen projecting from it. As 

 already said, these cells are very frequently situated in 

 spaces, or capsules in the test matrix. Whether this is 

 wholly due to the shrinkage of the cells I am not sure. 

 Even when these cells are found so close to the endothe- 

 lium as to be actually in contact with it, as very frequenth^ 

 happens, they are still of the same form, size and struct- 

 ure. This, of itself, is strong evidence against their 

 havinjT arisen from this endothelium. It is almost con- 

 elusive proof that they are not endothelial cells which 

 have migrated into the test; and if they had arisen from 

 these cells by division, it would still seem improbable that 

 they should at once be so different from their mother 

 cells. Again, it would seem that if they have such an 

 origin, cases in which the division is going on might be 

 found. Although I have searched diligently for such 

 cases my efforts have been in vain. 



My second reason for believing the cells of the test to 

 be derived from the contents of the vessels and the body 

 spaces is that cells are found here which are, so far as I am 

 able to make out, precisely similar to the test cells. Figs. 

 30 and 32, m. d' ., pi. iii, show some of these, or rather their 

 nuclei, in the first figure taken from one of the stolonic 

 vessels, and in the second from the body space in the 

 region of the branchial siphon. 



Fig. 37, pi. iii, represents three of them from one of 

 the vessels as they appear under the y'j o^^ immersion 

 objective. These are so similar to the cells of the test 

 in size, form and behavior toward stains that they need 

 no description. That which has already been given of 

 the one applies in every respect to the other. 



The third point which I present in evidence for my 

 contention is that I have found the cells in the process of 

 migration throutrh the endothelium. The most convincing 



