Î'ORMATION OF GERMINAL LAYERS IN CHELONlA. 241 



mesoblastic mass becomes continuous on each side with the entoblast, 

 just at the point where the chorda- and the darm-entoblast meet each 

 other. The cells being much smaller in Trionyx than in Triton, it 

 is not possible to distinguish the parietal from the visceral layer of 

 the mesoblast; but if both the chorda- and the darm-entoblast pass 

 into the mesoblastic mass, the relations found here amount to the 

 same thing as found in Triton and in Amphioxus. We think our 

 figs. 25 and 29 ought to convince the most sceptical on this point. 

 It is significant that at one time (fig. 17) the chorda-entoblast occu- 

 pies a recess of the alimentary canal by itself, and from the two sides 

 of this recess the mesoblastic masses stretch out — a relation which 

 recalls vividly the development of Amphioxus. 



Our fig. 19 may prove a stumbling block to some in the way of 

 comparison with Amphibia. But we think this figure is soon 

 reduced to the general rule. We have already pointed out that the 

 cells forming the floor of the blastoporic passage in this figure are 

 difl:erent from those of the roof and the sides. If we consider the 

 chorda-entoblast as extending on each side to the spot marked with 

 the star, and this spot as corresponding with the similarly marked 

 spot in fig. 15, Taf. iv, of Hertwig, which passes through the 

 corresponding part of Triton, the comparison will become easy. The 

 apparent difiiculty is brought about by the cells of the floor being 

 many layered in Trionyx. 



There is another point on which we wish to touch. Although 

 there is no doubt tliat the mesoblastic masses arise as what mor- 

 phologically amount to diverticula of the alimentary canal, the 

 development in Trionyx has so far changed from the primitive 

 method that the mesoblast no longer forms an epithelium as in Am- 

 phioxus or Triton or even compact masses throughout, but at places 

 appear as only loose masses of spindle and stellate cells (figs. 25 



