THE DEVELOPMENT OF LEPIDOSIREN PARADOX A. 37 



in methods of pi-eservation, section cutting, etc. ; so much 

 so that my final description varies in many important respects 

 from the rough draft made on a preliminary study of a few 

 embryos. My experience convinces me of the futility of 

 trying to give a fair description of the embryology of any 

 type unless one has a very large material to go upon. Much 

 of the discussion, involving often flat contradiction of dis- 

 tinguished observers^ statements, which is constantly taking 

 place appears to me to have a very probable cause in the small 

 amount of material which has been made use of. 



Summary of the more Important New Facts. 



1. The segmentation cavity arises in Lepidosiren from 

 intercellular chinks. 



2. The roof of the segmentation cavity early becomes two- 

 layered, and assumes the character of epiblast. 



3. Gastrulation takes place, for the most part, by a true 

 invaginatory process. 



4. Spreading of small cells over large takes place by de- 

 lamination, there being no true epibole. 



5. The disappearance of the segmentation cavity is in- 

 augurated by its penetration by a sponge-work of small 

 blastomeres from its floor and sides. 



6. The notochordal and mesodermal rudiments are at first 

 quite continuous across the middle line. 



7. The notochordal rudiment remains attached to the 

 hypoblast for some time after the mesoderm has separated 

 off on each side. 



8. The enteric roof is formed in situ directly from the 

 archenteric roof. 



9. The mesoderm grows outwards on each side by delami- 

 uation fi-om the large yolk-cells. 



10. The myocoele arises by breaking down of cells in the 

 middle of the myotome. 



11. Later on the myotome wall is composed of a single 

 layer of regular columnar cells. 



