BULLETIN OP THE UNITED .STATES FISH COMMISSION. 287 



By the tenth hour the segmentation of the disk has advanced very 

 much, and the cleavage of the component cells has proceeded so as to 

 have split them up into superimposed layers lying in the plane of the 

 great diameter of the disk, as shown in Fig. 3. Besides the develop- 

 ment ot superimposed layers of cells by another process, which I do 

 not clearly understand, a portion of the germinal matter of the disk has 

 been segmented off at its margin to form a wreath, tv, of much depressed 

 cells, which seem to be severed from the edge of the disk proper by a 

 slight interval all the way round. These appear to take an important 

 share in the development of the thick rim of cells r, which limits the 

 border of the blastoderm after it has spread out somewhat, as indicated 

 in Figs. 4, 5, and G. Up to the tenth hour of development the disk has 

 expanded but slightly; it now measures about one twenty-lifth of an 

 inch in transverse diameter, exclusive of the wreath of marginal cells, 

 or about the same as in the stage represented in Fig. 1. In Fig. 3 it is, 

 however, lenticular, convex above and below, and it is only during the 

 next twelve hours that it begins to spread, become of almost uniform 

 thickness, convex above and concave below. The singular changes un- 

 dergone by the disk of the cod were not so narrowly observed in this 

 species, although they probably occur. What is alluded to is the change 

 from the biscuit-shape of the morula stage, with a thick. margin and 

 almost flat upper and lower surfaces, to^the lenticular form of Fig. 3, 

 which is viewed somewhat obliquely, to that of the concavo-convex 

 form, which is alre;;dy assumed somewhat earlier. 



With the lateral expansion of the disk, the segmentation cavity sc is 

 developed beneath the upper germinal layers, which constitute its roof. 

 Here, as in other forms studied by the writer, this cavity does not disap- 

 pear, but persists and expands laterally as the growth of the blastoderm 

 proceeds. In Goregonus albus the cavity is principally roofed over by 

 the epiblast, which is composed of flattened, juxtaposed cells, while 

 smaller, rounded cells constitute its imperfect floor. The cells of the 

 floor appear to have been budded off from the mesoblast near the edge 

 of the blastoderm. A similar state of affairs probably exists here, for 

 as yet I can flud no evidence of a positive character to show that we 

 have in Belone an exception to the mode of develoi)ment generally ex- 

 hibited by embryo fishes ; but this structural feature will be further con- 

 sidered, in relation to the genesis of the blood, at another place. 



In Fig. 4 the embryo-swelling, which extends from e to the edge of 

 the blastoderm, is still in a verj' primitive condition. The cells, which 

 are to develop'into the bodj^ of the embryo, have not yet been arranged 

 into tracts, and little more than the upper or epiblast layer, with the 

 mesoblast lying below the latter, and above the hypoblast, can be said 

 to be ditt'erentiated. There is still no indication of a neural or primitive 

 groove; no differentiation of lateral mesoblastic plates, from which the 

 muscular segments or somites are to be dift'erentiated. Whether these 

 are lateral outgrowths, or diverticula from the hypoblast of the primitive 



