[47] EMBEYOGRAPHY OF OSSEOUS FISHES. 501 



iug- promiuence of the veutral keel of the blastotlei m, as shown iu Fig. 

 22, another process has been going on by which the tissues of the neu- 

 rula have been more markedly difit'erentiated from those adj acent. A very 

 distinct line of demarkatiou is established on either side of the neurula 

 separating it from the muscular mesoblast on either hand; this is shown 

 in Fig. 23. This demarkatiou was, however, already established when 

 the sensory layer was split otf from that lying- below it, but as the carina 

 of the neurula is developed, the mesoblast is gradually separated off 

 into two lateral masses, the stratum of hypoblast and mesoblast below 

 the neurula being now only a relatively very thin layer which it has 

 apparently pushed down before it. The line of separation between the 

 sensory and muscular layers at first coincided with the plane of their 

 upper and lower surfaces, but as the neurula grew iu depth, this line of 

 separation became more and more nearly vertical at the sides of the lat- 

 ter. In the cod's blastoderm this is accomplished about the tenth day 

 in embryos which hatch in from sixteen to eighteen days, and before the 

 closure of the blastoderm over the yelk. The muscular layer has by this 

 time also been quite sej^arated from the peritoneal layer below it, aiul 

 the muscle plates, as we may designate the three-sided, longitudinal, lat- 

 eral masses of cells on either side of the neurula, extend from behind 

 the optic vesicles up to the tail end of the embryo and are continued 

 into the lower layer of the rim of the blastoderm at r. The substance 

 of this rim seems, in fact, to be entirely incorporated into the building- 

 uj) of the embryo's hody, up to the time the blastoderm has inclosed 

 the yelk. The neurula has meanwhile also been undergoing further 

 differentiation. At its front end in the vicinity of cv, Fig. 23, a distinct 

 median split or cleft has apj)eared in it which extends forward and 

 backward some way, but is not developed in the tail end of this structure. 

 This cleft or cavity in the fore part of the neurula represents the neural 

 canal of other types, and from it in the head region the cerebral vesicles 

 will arise at a later period. At the tail end, on the other hand, the 

 neurula is solid, and on the caudal swelling no trace of the medullary 

 groove is visible. The caudal swelling is a mass of cells in which it is 

 impossible to discover any traces of the differentiation of laminte or 

 layers, except that of the outermost epiblastic or dermal layer. The 

 development of the neurula proceeds therefore from the head towards 

 the tail end of the embryonic axis, where it also grows in length as the 

 rim r of the blastoderm advances to finally close over the yelk. Once 

 the closure is accomplished the neurula becomes more defined at the 

 caudal region; it was greatly depressed here on the eve of closure, but 

 the caudal end of the embryo rapidly thickens as it incorporates the 

 caudal plate derived from the rim of the blastoderm. The now more 

 pronounced development of the caudal extremity of the neurula is due 

 to a process very similar to that concerned in the formation of the keel 

 or carina at the head end of the embryo ; in fact the keel of the neurula 

 develops from before backwards just as does its lumen or cavity. 



