REVIEW. 89 



the aperture, and are finally forced within it. Counting, there- 

 fore, as endoderm, not only those cells which actually form the 

 lining of the archenteron, but also those which will, at a later 

 period, constitute a portion of that lining, it is perfectly certain 

 that the whole of the mesodermal cells at present existing are 

 derived by a process of division from the endoderra. This im- 

 portant fact was first observed by Bobretsky. 



In the fourth stage (PI. VI, fig. 4), the thoracico-abdominal 

 process has increased considerably, forming now a well-defined, 

 transversely-oval elevation (a), just in front of the blastopore. 

 The latter (g), has decreased to a third of its original dimen- 

 sions, and the endodermal plug, which had already begun to 

 diminish in the third stage, has almost disappeared. 



This last observation corrects in a very important particular the 

 statements of Rathke and LerebouUet. Both these authors observed 

 the blastopore, but thought that the "plug" was converted into 

 the thoracico-abdominal process, the anus appearing in its centre. 



Sections of the embryo at this stage show the great forward ex- 

 tension of the archenteron, the lining cells of which are now, at the 

 anterior boundary of the blastopore, sharply distinguished by their 

 form and size from the ectoderm, while on its posterior boundary 

 they still pa?s insensibly into the latter membrane. 



Many of the endoderm cells, especially those immediately 

 beneath the abdominal process, have increased to twice their 

 former size, and exhibit certain very remarkable characters. 

 They contain, in many cases, two or even three nuclei, the 

 shape of which may be round, oval, biscuit-shaped, or semi- 

 lunar. Besides the finely-granular protoplasm, accumulated 

 mostly round the nuclei, they contain large, often vacuolated 

 spheres of deutoplasm, of precisely similar character to the 

 spheres of which the food-yolk itself is now made up, and upon 

 which these endoderm cells abut. But the most interesting 

 point of all has regard to the shape of the cells. Under a low 

 power they seem to have the ordinary columnar form, but when 

 highly magnified (see fig. 5, woodcut) it is seen that the ends 

 of them, which are turned towards the yolk, are irregularly 

 lobed, and give ofi^ more or less fine threads of protoplasm {j).), 

 which pass between, and in some cases surround, the yolk 

 spheres (y.) These processes have, in fact, all the characteristics 

 of pseudopodia, and it seems perfectly evident that the endoderm 

 cells absorb the nutritive matter of the yolk, not by a passive 

 process of diffusion, but by an active process of ingestion, the 

 food particles being immediately ''plunged into the living pro- 

 toplasm of the cell,'^ and there digested. It is extremely 

 interesting to see so archaic a mode of nutrition as this retained 

 in the embryo of so highly organised an animal as the Crayfish, 



