134 s. HATTA. 



appearance can be g'athered from ûgs. 2, o, and 4, PJ. XIII. As al- 

 ready stated, I am inclined to believe that fig. 2 shows the stage in 

 which the invagination is just commencing. The cells where this process 

 beo-ins (pp. fig. 2) assume first of all a columnar shape, while the rest 

 still remain materially unchanged. Once begun, tlie invagination pro- 

 gresses rapidly, the roof of the invaginated tube being formed by the 

 cells coming from the animal pole, while its floor is furnished by the 

 yolk cells. According to Shipley (loc. cit.) and Kuptfer (loc. cit.), 

 the epiblast extends itself over the vegetative half of the egg by the 

 conversion of the outermost row of yolk-cells into small columnar 

 cells. This appears to me very doubtful. I think it more prol^able 

 that the whole yolk-cells are bodily invaginated into the interior, i. e., 

 the point marked x in fig. 2 moves itself constantly towards the 

 l)lastopore Qrp.), so that the whole mass of yolk-cells is really pushed 

 inside. According to this view, what was at first the external surfîice 

 of the yolk-cell mass, becomes the lioor of the invaginated tube. The 

 seo'mentation- cavity is pushed forwards and early obliterated. This 

 process is very much like that described by Hertwig^ in Triton. 



Tio-, 4 represents a median sagittal section through an egg in 

 wliich the invagination is almost completed. Fig. 3 is a transverse 

 section of an egg a little younger. From these two sections some idea of 

 the character and disposition of the cells forming the différent layers can 

 be easily formed. The roof of the mesenteron consists from the first of 

 a single layer of columnar cells appressed against the under surface of 

 the epiblast. I could not find throughout the entire extent of the in- 

 vaoinated tube, either in longitudinal or transverse series, any trace of 

 cells interposed between the epiblast and hypoblast. I can not therefore 

 accept Scott's invaginated mcmhhi^i (loc. cit.), nor can I find any trace 

 of Kupffer's teloblast (loc. cit.). The epiblast cells are also columnar 



1. 0. Hertwig :— Die Entwicklung des mittleren Keimblattes der Wirbelthiere. Jena, 18S3. 



