198 BULLETIN OF THE UNITED STATES FISH COMMISSION. 



once becomes blastodermic in morphological character. Nor do holo- 

 blastic ova even behave similarly to each other. Even in the progress 

 of development the gastrula stage is sometimes so modified or interfered 

 with by the mechanical accidents of the construction of the ovum as to 

 be almost wholly obscured or suppressed. Such occurs in a few cases 

 like Eucope and Stepkanomia, where a gastric cavity is formed without 

 gastrulation before the oral opening is broken through, so that the fact 

 of gastrulation is not universally a preliminary to the formation of the 

 archenteric cavity; formerly, in its positive form, an embryological 

 doctrine which observed facts have shown to be untenable when uni- 

 versally applied. Again, it is certain that the blastopore which results 

 from the closure of the outermost layers of the ovum over the innermost 

 ones, or over the yelk, does not always correspond either to stomodeum 

 or proctodeum, or to mouth or vent, but that it is sometimes an evanes- 

 cent structure of little or no morphological significance. This is the 

 case with what is probably the true blastopore of the Teleostean ovum. 

 The mechanical relation of germ disk or blastoderm to the yelk in the 

 Teleostean ovum is a peculiar one and has not hitherto received the atten- 

 tion which its importance has deserved. The peculiar mode of devel- 

 opment which has grown out of this relation is in the highest degree 

 interesting as compared with other forms. After the disk has been 

 formed by the ainseboid aggregation of the germinal protoplasm, a series 

 of phenomena present themselves during segmentation which will well 

 repay attention. As soon as the segmentation is fairly under way, so that 

 t lie rudiment of the embryo's axis begins to be apparent at the edge of the 

 blastoderm, it is found that a shallow, crescent-shaped cavity has ap- 

 peared underneath the central portion of the disk not immediately em- 

 braced by the embryo. This space extends beneath the blastoderm and 

 grows laterally with the growth of the disk. It is, in fact, a space filled 

 with a film of fluid over which the disk may spread without friction on the 

 underlying yelk. It remains until the blastoderm finally closes over the 

 yelk entirely, when it may still be seen as a space, in some species sep- 

 arating the membranes of the embryo all round from the true yelk sac 

 within. This arrangement, of course, causes the development to present 

 some peculiar features, since the development of the embryo is perfectly 

 sessile on the yelk. Later on, the heart communicates directly with 

 this cavity, as in Alosa and Pomolob us, and yelk corpuscles are budded 

 off directly into it and sucked up by the heart, to be carried directly into 

 the vascular system. Nor is this the only feature of novelty ; at the tail, 

 the rim of the blastoderm closes and forms what I have denominated 

 tin- caudal plale, which afterwards enters into the development of the 

 tail. The whole of the blastodermic rim (ringwulst) is thus absorbed, 

 and no similarity remains to make the Teleostean embryo comparable in 

 this respect with the Elasmobranch, in which the rim of the blastoderm 

 is not incorporated in this manner. In the region of the caudal plate a 

 solid neural cord or rod is continuous with the notochord and rudiment 



