VIVIPAROUS FISHES OF THE PACIFIC COAST. 
433 
could more forcibly illustrate the manner in which the discontinuity of the lower 
layer of cells was formed by the addition of yolk to the Branchiostoma gastrula and 
the result of subsequent reductions, than the early gastrula of Stolepliorus and the 
condition in Cymatogaster. A comparison of the gastrula of Cymatogaster with that 
of Branchiostoma and of the frog is interesting. The gastrula of Cymatogaster , exclu- 
sive of the yolk, is remarkably like the gastrula of Branchiostoma as figured by 
Kowalewsky. In fact the peculiar conditions in Cymatogaster are just what might be 
expected if the ventral hypoblast cells of Branchiostoma should in some way become 
eliminated. This elimination has probably been brought about by the fusion of 
these cells, through their enormous surcharge with yolk and by the subsequent 
reduction of this yolk without a restoration of its original cellular constitution. In 
those cases (elasmobranchs and teleosts in general) in which the yolk is functional 
it occupies the space originally oceupied by the ventral hypoblast cells ; i.e., the region 
between the tip of the head and the anterior or ventral margin of the embryonic rim. 
Detailed homologies between parts of the Branchiostoma and parts of the Cyma- 
togaster gastrula can not with certainty be pointed out till the first stages in the for- 
mation of the embryo of Cymatogaster have been observed. Both the gastrulas are 
formed by a layer of epiblast and a layer of hypoblast, and the region of the dorsal lip 
of the blastopore of Cymatogaster is in all probability homologous with the dorsal lip 
of the blastopore of Branchiostoma. The ventral portion of the former gastrula has, 
however, been so reduced that the embryonic axis extends over more than half the 
circumference of the gastrula. While the gastrula of Cymatogaster is thus seen to 
approach that of Branchiostoma , it is by no means primitive, but is highly specialized 
by the reduction of the yolk, or, in other words, by viviparity. This is to be expected, 
for, as Balfour long ago expressed it ( 11 , 342) : 
If the descendants of a form with a large amount of food yolk in its ova were to produce ova 
with hut little food yolk, the type of formation of the germinal layers which would thereby result 
would be by no means the same as that of the ancestors of the forms with much food yolk, but would 
probably be something very different, as in the case of Mammalia. 
Cymatogaster appears to me to stand at the very end of the series of teleostean 
eggs which have been derived from large-yolked elasmobranch eggs. 
A glance at the gastrulas of Cymatogaster (fig. 41) and of the frog (Balfour, fig. 71, 
B) shows an interesting similarity. The yolk in the frog bears the same relation to 
the blastopore that it does in Cymatogaster , the yolk plug of the frog being represented 
by the yolk nucleus in Cymatogaster. But here again the embryo occupies less than 
half the circumference of the entire egg. The conditions in this case would be 
nearer those of Cymatogaster were there less yolk. There is, however, a very serious 
objection to deriving the teleost gastrula directly from the amphibian gastrula, for by 
such a supposition the periblast of the teleost must arise de novo from the yolk cells of 
the amphibian and stand in no relation phylogenetically to the yolk nuclei of the elas- 
mobranch, a supposition very absurd on its face. 
In Wilson’s opinion one of the prime objections to deriving the teleost gastrula 
from the Branchiostoma is that “the theory leads us nowhere; it does not admit 
of any exact comparison between the teleostean embryo and those of other vertebrates.” 
It will be well to bear this statement in mind in further considering the homology 
between the amphibian and the teleostean egg. 
According to Wilson’s view the whole course of the fish development becomes 
F. C. B. 1892 — -28 
