536 LOCV. [Vol. XL 



4. They are formed Independently of Me so dermic Influence. 



The next point I wish to maintain with regard to these seg- 

 ments is that they are formed independently of mesodermic 

 influence. 



In a preliminary account, I went so far as to state that the 

 segmentation is epiblastic, but that statement should be quali- 

 fied. It is certainly most clearly expressed in the epiblast, but 

 the other germ-layers apparently feel the same influence, and 

 partake of this segmental division in a modified degree. It is 

 least clear in the mesoderm — in sections it may be made out 

 in this layer, and also in the endoderm. The so-called neural 

 segments are developed throughout the entire length of the 

 embryo before any protovertebrae are formed and, therefore, 

 they must be independent of any formative influence of the 

 latter. 



There is also a notable difference in the two forms of 

 metameric division ; the formation of protovertebrae, as is well 

 known, proceeds from a certain point forwards and backwards, 

 and if the segmentation I have described were moulded over 

 the former, we should expect the neural segments to appear 

 gradually, keeping pace with the formation of protovertebrae. 



Nevertheless, the majority of authors who have touched 

 upon the question have taken the position that segmental 

 division of the mesoderm is primary and that the neural seg- 

 ments are moulded over it, but all this time the early history 

 of the neural segments has not been known. Kupffer 

 announced, in 1885, that segments appear upon the median 

 neural plate of Salamandra atra while the groove is widely 

 open and before the appearance of any protovertebrae. . He 

 designated this primary metamerism. Seven years later, 

 Froriep ('92) studied the question of metameric segmentation 

 upon closely related forms (Salamandra maculosa and Triton 

 cristatus), but reached different conclusions as regards the 

 nature of the segments. 



Froriep's conclusions are based on both surface study and 

 sections. He shows in very young Triton embryos a seg- 

 mented condition in the neural plate included between the 



