FORMATION OF NOTOCHORD IN AMPHIBIA. 293 



derm and endoderm are connected for a considerable distance on 

 either side of the middle line. 



In the posterior part of the embryo the cells forming the dorsal 

 wall of the archenteron do not differ in size, shape, or in power of 

 staining from the mesodermal cells above them, and at the sides of 

 the archenteron they grade into the larger yolk cells forming the 

 ventral and lateral walls. There is a comparatively narrow re- 

 gion in the mid-dorsal wall where the edges of the cells border- 

 ing the archenteric cavity are quite heavily pigmented ; but the 

 other cells of the dorsal wall contain about the same amount of 

 pigment as do the mesoderm cells above them, and there is no 

 definite dorsal plate of small, deeply pigmented, rectangular cells 

 as in the toad embryo. I can find no evidence that any of the 

 cells of the dorsal wall of the archenteron ever form a permanent 

 union with the mesoderm. 



When the medullary folds are beginning to form, the noto- 

 chord has extended into the posterior region of the embryo and 

 appears as in Fig. 8. It is a more rounded structure than is the 

 notochord of the toad embryo at a corresponding stage of de- 

 velopment (Fig. 3), yet it, too, is closely connected with the en- 

 dodermal layer of cells forming the mid-dorsal wall of the archen- 

 teron. As shown in Fig. 8, the lateral qiesoderm and the en- 

 doderm of the archenteric wall are connected for some distance 

 on either side of the notochord. The cells of both of these tis- 

 sues have the same general characteristics, and there is no sharp 

 distinction between them as in the embryo of Bufo. As all of 

 the cells in the dorsal part of the embryo have the same power 

 of staining, it is not easy to follow the changes that take place, 

 even with an abundance of material for study. Not until I had 

 made camera drawings of a complete series of sections through 

 the posterior region of an unusually favorable embryo was I 

 able to tell with certainty how the notochord is formed. Four 

 of these drawings (from the same embryo as Fig. 8) are repro- 

 duced in Figs. 9-12. For the sake of clearness only the dorsal 

 wall of the archenteron and the notochord are shown. In al 

 of the sections the mesoderm is entirely cut off from the noto- 

 chord, and also from the endoderm beneath it. 



A short distance in front of the region shown in Fig. S, almost 



