BULLETIN OF THE UNITED STATES FISH COMMISSION. 297 



ter upon an umbilical stalk, so that the portion of the blastodermic rim 

 which still remains, but is separated liom the embryo, and which will 

 finally coalesce some distance behind the umbilical stalk, is probably 

 not homologous with any i)art of the rim of the blastoderm of the teleos- 

 teau. It is clear, at any rate, that this part of the rim of the blasto- 

 derm of the Elasmobranch takes no share in the formation of the cau- 

 dal plate, and indirectly of the tail end of the body, as happens in the 

 teleost. To urge the example of tlie Elasmobranch blastoderm, as 

 Balfour has done, in refutation of the arguments of His and Eauber in 

 relation to the part taken by the blastodermic rim of the germinal 

 membranes of the teleost in the formation of the body, is therefore 

 hardly fair. 



As already stated in my paper on the development of the Spanish 

 mackerel, the teleostean ovum is remarkable for the way in which the 

 superficial layer or ijellicle of germinal protoplasm, destined to form the 

 germinal disk, migrates towards one pole of the vitellus to aggregate 

 into a biscuit-shaped germ-mass. The process has been studied by the 

 writer in detail in the egg of the cod ( Gadus), where, owing to the low 

 temperature of the water in which the eggs develop, it requires some 

 time for its completion, so that it may be studied very minutely. It 

 appears that the nucleus undergoes disintegration or rearrangement in 

 the fish ovum before it leaves the ovarian follicle in which it grew. 

 The nucleus in young ova is observed to be embedded in the center of 

 the ovum ; as the latter acquires maturity it migrates toward the sur- 

 face and its contents are apparently broken uj) to be involved partly or 

 wholly in the peripheral germinal protoplasm. In some teleostean ova it 

 appears that the germinal disk is formed at the time of oviposition, but 

 this is not the case in any of the species studied by the wi^^ter. In the 

 cod, for examjjle, the germinal disk was not formed until about four hours 

 after impregnation. In this sj^ecies, as well as in BeJone and Ci/hium, the 

 germinal layer of protoplasm from which the germinal disk is developed 

 is a distinct external layer enveloping the true vitelline protoplasm. 

 It appears that in some species this peripheral layer of protoplasm is con- 

 nected with the interior of the vitellus by strands or i^rocesses of itself 

 which pass inwards between the vitelline corpuscles, often forming an 

 intricate investing matrix in which the latter are embedded. Notwith- 

 staudiug all these modifications, however, the portion of the ovum which 

 is directly inlluenced bj" the act of impregnation is the germinal disk alone, 

 which in turn has been derived from the external germinal pellicle. The 

 vitellus is, throughout the whole of development, passive ; as the embryo 

 is developed, the heart, through the intermediation of the segmentation 

 cavity and blood vessels, becomes, in part, the means by which it is ab- 

 sorbed, the process being assisted by the formation of free nuclei in its 

 substance as well as by germination, and, iierhaps, by intussusception 

 or absorption by the overlying hypoblast itself. The theory of the in- 

 termediarij layer proposed by the writer, in the essay on the develop- 



