THE ECTODEEM. 55 



gress of Stage f the nuclei which have hitherto been placed in 

 the deep parts of the layer (Plate VII, figs. 23 a — e, 25) acquire 

 a superficial position, excepting in the hump, where they retain 

 their deep position until after Stage g. Contemporaneously 

 with this change the deeper parts of the ectoderm become 

 filled with very large vacuoles, so that the protoplasm is re- 

 duced to fine cords, passing inwards from the superficial nucleated 

 layer. This vacuolated deeper part of the dorsal ectoderm 

 now becomes much reduced, so that in Stage f the dorsal 

 ectoderm consists mainly, if not entirely, of a thin layer of 

 nucleated protoplasm derived from the superficial layer of the 

 preceding stage. The hump, however, still persists, retaining 

 the characters it had in Stage e. 



With regard to the internal boundary of the ectoderm, in 

 the gastrula stage there was no line of demarcation between 

 it and the eadoderm. In Stage b the mesoderm appears, 

 but causes no break in the continuity (Plate VI, fig. 

 5 a — /). In Stages c and d, however, a definite separa- 

 tion occurs, firstly by the appearance of the cavity in the 

 somites, and secondly — and this happens later, in Stage 

 D — by the dorsal and ventral separation of the endo- 

 derm from the ectoderm. The endoderm is, however, still 

 continuous with the splanchnic layer of mesoderm, and the 

 ectoderm with the somatic. In the subsequent development 

 this continuity seems to be retained and to be extended in 

 consequence of the growth of the mesoderm over the internal 

 surfaces of the at first uncovered parts of the two primary 

 layers. At any rate I have never been able to see any well- 

 defined boundary between the layers in question, even in the 

 best-preserved embryos, if a careful examination was made 

 with a high power. The defined line drawn in my figures has 

 only an existence with a low power ; it is therefore extremely 

 difficult to say whether or no nuclei pass in from the ectoderm 

 to the mesoderm, and often not possible to settle for certain 

 whether a given mass of nuclei belong to the ectoderm or 

 somatic mesoderm. In the later stages (f) this absence of a 

 defined line continues, so that it becomes difficult to decide 



