12 Primitive Streak and Notochordal Canal in CJielonia. 



surface views of the hardened embryo (Plate I. Figs. 3 and 3'). 

 The dorsal view exhibits the three-pronged or trident-like mark- 

 ing which Mitsukuri and Ishikawa (^^6) have already observed 

 and described. The central prong ends abruptly. Cross sections 

 show this to be the point at which the columnar arrangement of 

 the thick dorsal wall of the notochordal cavity ends. From this 

 point a thin layer of loosely arranged cells continues anteriad 

 underneath the remaining portion of the shield. 



The reconstruction (Plate VIII. Fig. 44) from transverse sec- 

 tions is intended to illustrate the condition which would have been 

 found along the axis of the shield if the shield had been cut in the 

 sagittal plane. The notochordal entoderm, which is continuous 

 with the axial ectoderm at the dorsal lip of the blastopore, stretches 

 anteriad for some distance as a thickened layer of columnar cells 

 several cells deep. In this instance the entoderm for a con- 

 siderable distance in front of the blastopore is thicker than the 

 ectoderm which lies above it. About two thirds of the distance 

 towards the anterior end of the shield the entoderm suddenly 

 diminishes in thickness, and continues forward as the thin layer 

 already mentioned. 



The anterior portion only of the notochordal cavity has become 

 continuous with the subgerminal space. Half of the remaining 

 portion of the floor, however, shows indications of being about to 

 disintegrate. Anterior to the region in which only a few scattered 

 cells exist, the floor of the notochordal canal continues for some 

 distance as a jelly-like granulated sheet. Here and there in the 

 anterior portion of the canal small, free clumps of granulated sub- 

 stance also exist. Along the walls of the canal are seen fragments 

 of degenerating cells, such as have been previously described. 

 Between these shaggy margins the lumen of the canal extends 

 posteriad for some distance, to end blindly in a plug of cornified 

 or degenerating cells, which completely fill the lumen of the canal 

 as far as the blastopore itself. The notochordal entoderm above 

 this plug has assumed a perfect columnar arrangement. Below 

 the plug exists a mass of cells which has been identified by many 

 embryologists as the beginning of Rabl's gastral mesoderm. The 

 plug itself is composed mainly of what appear to be empty cell 

 walls, which either refuse to stain at all, or stain but slightly. 

 Figure 35 (Plate VIII.) represents, under higher magnification 

 than Figure 44, a cross section through this region. The lumen 



