THE DUVKLOPMyNT OF LEFIDOSIHEN PARADOXA. 11 



sious soon become joined up to form a continuous groove 

 stretching through about one tliircl of the circumference of 

 the egg at this latitude (cf. Text-fig. 1, fig. 10). A section 

 through the whole egg at this stage is given in fig. 6 (PI. 2), 

 and sections through the groove itself under a higher magni- 

 fication in figs. 5 and 7 (PI. 1). 



In the cells lining the groove much of the yolk lias passed 

 into a state of fine subdivision, thus pointing to cell activity. 

 From the open character of the groove during this stage it is 

 obvious that we have to do with a process of true invagina- 

 tion. In some series of sections one can see very well (fig. 7) 

 how the groove, although to the naked eye apparently coin- 

 cident with the boundary between the cells with small and 

 those with large yollc granules, lies really just within the 

 region of the latter. Tiie invagination in Lepidosiren is 

 thus essentially a lower cell phenomenon. 



The groove, as mentioned in my previous paper, gradually 

 becomes more limited in extent by its lateral portions be- 

 coming flattened out. Had it extended at any period of 

 ontogeny completely round the exposed area of large cells, 

 we should have been able to speak of a yolk-plug. As it is 

 probable that the disappearance of a yolk-plug bounded all 

 round by an invagination groove is due to increase in size 

 and richness of yolk in the egg, I had hoped to find it present 

 in Protopterus. In this I have been disappointed, the 

 condition in this respect being just as in Lepidosiren. 



While the lateral parts of the groove flatten out and 

 disappear, the middle part is deepening to form the arch- 

 enteric cavity. 



B. The exact method by which this takes place in its 

 earlier siages forms a problem of considerable general 

 interest, but at the same time one the final solution of which 

 is attended with great difficulties. 



The appearance of sections during this period is illustrated 

 by fig. 8 (PI. 2). The archenteric cavity runs obliquely in- 

 wards from the surface of the egg, and at its inner end turns 

 upwards so as to run roughly parallel to the surface. Th^ 



