60 EMBEYOLOGY OF THE LOWEE VEETEBEATES ch. 



mesoderm. This latter gradually spreads ventralwards by delamina- 

 tion from the large yolk-cells and eventually the mesoderm sheets on 

 the two sides become continued into one another ventrally. 



As will be noticed there are no coelomic spaces within the 

 mesoderm rudiments at these early stages: they arise secondarily 

 later on. 



If we review the above-described stages in the early development 

 of the mesoderm segment in Lepidosiren, in which, as already indi- 

 cated, the large size of the cell-elements ensures unusual freedom 

 from the danger of errors of observation, we see that the last 

 described stage is clearly in agreement with the hypothesis that it is 

 a repetition of the stage in Amphioxus when the mesoderm existed 

 in the form of a series of enterocoelic pouches on each side. The 

 only conspicuous difference is that, whereas in Amphioxus these 

 were actual pouches, here they are solid blocks of cells in which a 

 cavity only makes its appearance at a later stage of development. 

 That this difference is in no way a serious one will become apparent 

 to the reader as he realizes that it is one of the commonest modifica- 

 tions of developmental phenomena, when yolk is abundant, that 

 primitively hollow organs develop in the embryo from solid rudi- 

 ments and only form their cavity secondarily. 



It may be accepted then with confidence that the solid mesoderm 

 segments of Lepidosiren at the stage indicated, continuous ventrally 

 with the endoderm, represent the enterocoelic pouches of Amphioxus 

 modified in correlation with the abundance of yolk. 



The first stages in the development of the mesoderm of Lepido- 

 siren are obviously very different from what are found in Amphioxus 

 and the differences here also we may justifiably attribute to the 

 immense thickening of the endodermal wall of the archenteron corre- 

 lated with the storing up of a large amount of yolk in its cells. 



In the other groups of holoblastic vertebrates the main features 

 in the early development of the mesoderm agree with those just 

 described for Lepidosiren. In all of them the archenteron is pro- 

 vided with a thick wall of heavily yolked endoderm cells, those 

 forming the roof or dorsal part of the wall being smaller and pro- 

 vided with finer yolk-granules. Out of this smaller-celled mass the 

 mesoderm segments become carved by the development of splits very 

 much in the same way as in Lepidosiren (cf. Fig. 40, ~R—Petromyzon). 



Amongst these groups the Amphibia call for a little further 

 consideration. 



In the frog a split develops on each side which separates the roof 

 of the archenteric cavity into two layers, an inner layer, one cell 

 thick, of definitive endoderm and an outer, two cells thick for the 

 most part, the mesoderm. This split is seen in Fig. 36 which repre- 

 sents a section, transverse to the axis of the archenteron, through an 

 egg with large yolk-plug. The split in this section terminates below 

 at about the level of the floor of the archenteric cavity while above 

 it stops short of the level of the notochord. 



