494 KEPORT OF COMMISSIONER OF FISH AND FISHERIES. [40] 



from A to B. This figure also shows the very narrow slit-like lumen of 

 the segmentation cavitj' to the right of A, and its extent over the left 

 half of the blastoderm is shown by the line below and to the left of at/. 



I).— TUE DEVELOPMENT OF THE GERMINAL LAYERS. 



This portion of the subject is one upon which I cannot, unfortunately, 

 throw nuK'li light on the basis of observations made upon the develoi*- 

 meut of the cod's e^^g, and L shall therefore place under contribution 

 the labors of O^llacher and others on the trout, and my own observa- 

 tions ui)on those of several other species. It is evident that in respect 

 to the developmental changes which the blastoderm undergoes in dif- 

 ferent species, there is considerable variation. In the trout, for example ; 

 the embryonic shield or area, corresponding in the cod's egg to the space 

 from A to B in Figs. IG, 17, and 18; there is a considerably less i)romi- 

 nent development of the embryonic shield at a correspondingly early 

 stage. In Tylosiirtis, at an early stage, the conditions of the two in re- 

 spect to the size of the embryonic shield is about the same. 



The embryonic shield, as develoi)ment advances, grows farther and 

 farther inwards towards the center of the blastodermic disk, or rather, 

 as it grows in length before and behind, the disk at the same time spreads 

 in consequence of the continued segmentation of its component cells, so 

 that these are spreading themselves over a greater and greater area 

 while they are at the same time undergoing a definite rearrangement 

 into strata or layers, each of which has a definite share in building up 

 the different parts of the embryo's body. This mode of spreading, how- 

 ever, never affects the relation of the embryo to the edge of the disk. 

 Its tail-end lies at the edge, its head at the center for a considerable 

 time (in small or moderate sized ova constantly), with the axis of the 

 body of the future embryo lying in one of the radii of the disk. In un- 

 usually large ova, like tliose of the salmon, Tylosnrm and Arins, the 

 blastoderm spreads so fast after a while that the embryo does not grow 

 in length rapidly enough to maintain the position of the head near the 

 center of the blastoderm. These last facts explain (EUacher's position 

 in regard to this plienomenon in the trout's ovum. 



With the development of the embryonic shield the diflerentiation of 

 the lower layers comnumces. The first to be split off is the sensor^' or 

 epiblastic layer; in the cod's egg this is formed on the seventh day in 

 ova which hatched in sixteen days. The process is truly one of dela mi- 

 nation, and cannot be regarded as produced by a true gastrulation at 

 all, as ITaeckel has tried to sliow in a paper alrea«ly noticed. The ])ro- 

 cess of the differentiation of the layers we saw began with the develop- 

 ment of an epithelial layer of epiblast over the surface of the germinal 

 disk before the appearance of the segmentation cavity. Immediately 

 after the ai)pearance of the latter the embryoinc disk begins to be de- 

 veloped and the layers differentiated. The sensory layer is split oft" 

 first from the underlying stratum of cells at the head end of the embry- 



