[113] EMBRYOGEAPHY OF OSSEOUS FISHES. 567 



from Kowale^sky. In fact there is abundant evidence of the truth of 

 the cceloma theory, proposed by the Hertwigs, presented in the mode 

 of development of tlie muscle plates of Teleostei, as lateral outgrowths 

 of the lower layer before the differentiation of the mesoblast from the 

 hypoblast by delamination, as indicated in Fig. H. Ziegler has in fact 

 reached the conclusion that the chorda is of hypoblastic origin, so that 

 embryologists are aluiost unanimous in regard to the origin of the primi- 

 iiive axis of the body of the chordata. Such an origin is indicated in 

 the cod by the position of the caudal end of the chorda below the level 

 of the upper half of the thickness of the caudal end of the body of the 

 embryo shown in Fig. 31. 



The evidence favoring the marginal ingrowth of the lower layer is to 

 me not as strong as that in favor of delamination. My reasons for this 

 opinion ma}^ be stated somewhat more clearly by referring to Figs. 14 

 and G. In the first, and in both optical and actual sections of a similar 

 stage in other forms, the lenticular germ is shown to be composed of a 

 mass of equal sized cells; in fact a morula condition is developed and no 

 differentiation of layers is perceptible except the single outermost and 

 epithelial stratum of the epi blast. In flattening and spreading there 

 is not enough lateral movement to account for the formation of the wide 

 rim in Figs. 15 and 16, by an infolding of the edge of the disk as it spreads 

 so as to bend the separate lower hxj-er inwards all round the margin. In 

 fact, the segmentation cavity is at first smaller and deeper in i)roj)or- 

 tiou to its w^dth than shown in Fig. 15, and the marginal infolding, as it 

 might appear, does not continue with the spreading of the disk, but 

 afterwards takes place only at one point, viz, where the rim is contin- 

 ued into the sides of the embryo. The weight of the evidence is there- 

 fore in favor of delamination as the means by which the primary dip- 

 loblastic condition of the germ is developed and not by gastrulation. 

 The first diploblastic phase of the germ of the fish egg is therefore appar- 

 ently a plannla, and neither its upper nor its lower layers are but one 

 cell deep, but consist in both cases of three or more rows of cells, which 

 can scarcely be said to be arranged in regular layers except those 

 which limit the upper and lower surfaces of two the primary ones. 



lu Fig. Gr the upper or epiblastic stratum in vertical section is left 

 white, while the lower combining the mesoblast and hypoblast is shown 

 black in section with its superficial extent in,the blastoderm indicated 

 by the hatched lines. The embryonic axis of this blastoderm is cut 

 through from a to hi. At first one would suppose that there was clear evi- 

 dence of invagination from behind forwards from the point hi in the 

 earlier stages, but, as it has already been remarked, this does probably 

 not begin much before the stage represented in figure G has been at- 

 tained, and then by the peculiar mode of concrescence previously de- 

 scribed. 



With the growth of the blastoderm over the yelk the lower layer 

 of the rim r does not increase in width, as required by the invagination 



