52 CALIFORNIA ACADEMY OF SCIENCES. 



instance of this is shown in fig. 30, ni. c, pi. iii. Although 

 neither nucleolus nor cell-body could be distinguished in 

 this, still it differs in no way from many nuclei that are 

 found on both sides of the membrane, /. e., within the 

 vessel and in the test. That it is embedded in the pro- 

 toplasmic portion of the cells of the endothelium does 

 not, I think, admit of doubt. I suppose it is passing 

 between the two cells, the nuclei of which lie each side 

 of it. 



The same process of migration is seen also, I believe, 

 in fig, 31. This is from a section which cuts the endo- 

 thelium of a vessel tangentially, or rather at a very oblique 

 angle. On one side of the section test alone appears, 

 while on the other side endothelium appears only. It 

 will be seen that in the region where there is no test three 

 of these nuclei are found. Of course there is grreat 

 danger of error here, since the endothelium is very thin 

 and cells on its surface on either side may easily be mis- 

 taken as being situated within it. With this chance of 

 error in mind, I have studied this and many other similar 

 sections with care, and have convinced myself that several 

 instances have been found where the migrating cells are 

 in the same plane as the nuclei of the endothelial cells. 



Figure 32, pi. iii, represents an interesting condition, 

 which strongly confirms the belief here maintained. The 

 section is through the point at which the branchial siphon, 

 already fully formed, so far as the ectodermal membrane 

 is concerned, is about to break through the test. 



What is taking place will be made apparent by com- 

 paring this figure with fig. 33, which represents a sim- 

 ilar section of an earlier stage in the formation of the 

 siphon. Shortly after the breaking through is com- 

 pleted at the point of fusion of the invaginated ectoder- 

 mal layer with the wall of the branchial sac, a prolific 



