32 J. B. JOHNSTON. 



off from the entoblast separately from the rest of the meso- 

 blast. Finally, single cells wander off from the cephalic surface 

 of the entoblast and go immediately to the formation of head 

 mesenchyme. 



The mode of formation of the heart endothelium from the ven- 

 tral keel of entoblast differs in details in different forms. In the 

 Urodeles studied by Brachet, the keel of entoblast extending from 

 the mouth anlage to the region of the liver splits off as a contin- 

 uous rod, the cells of which later arrange themselves into a tube. 

 In the species studied by the writer the cells of this keel do not 

 remain in a continuous rod but split off singly or in groups of a 

 few cells and form a loose mass which remains connected with 

 the entoblast longest at the end nearest the mouth. At this 

 point there is continued growth and there is probably a migra- 



FIG. 4. 



tion of cells from this point backward, and also upward into the 

 several branchial arches, to form the aortic arches. The splitting 

 off of this keel of entoblast is taking place simultaneously with 

 the spreading ventrally and mesially of the pericardial mesoblast. 

 My preparations leave no doubt whatever that the heart endo- 

 thelium is formed from the most superficial portion of the ento- 

 blast in the mid-ventral region and that the lateral sheets of 

 mesoblast are formed wholly outside of this area. Brachet's de- 

 scription makes it clear that the same thing is true of the Urodeles 

 which he studied, but this important relation seems not to have 

 attracted his attention and the fact is not mentioned by him. 



We are now able to state definitely the nature of the cceno-' 

 genetic modifications connected with the formation of the heart 

 endothelium in amphibia. According to the earlier accounts 



