432 
BULLETIN OF THE UNITED STATES FISH COMMISSION. 
of Cymatogaster , which contains much less yolk than the amphibian egg, persists in 
segmenting only the germ and not the yolk tends to show that the teleostean egg is not 
immediately derived from an egg similar to the amphibian, as Wilson states, but that 
the teleostean condition is deep-rooted and of very long standing. 
Balfour’s (1885, p. 68) view (which Wilson does not mention) is thoroughly in 
harmony with the conditions found in Cymatogaster , as well as with those of Salmo, 
and I see no valid reason for changing it. His view very tersely expressed is: 
The peculiarities of the development of the teleostean ovum can best be understood by regarding 
it as an elasmobranch ovum, very much reduced in size. 
Agassiz & Whitman (1884) indorse this view, and Miss Clapp’s discovery of the 
closing of the blastapore behind the embyro in Batrachus , a large-yolked teleost, 
brings the final proof of this view. It is probable that conditions will be found in 
Tachysurus which are similar to those in Batrachus. 
There can be no question but that teleostean ova, as all other telolecithal ova, 
have been derived from alecithal ova, such as those of Branchiostoma. The question 
at issue is whether the conditions found in the fishes have been derived immediately 
from some egg resembling the amphibian egg, or whether from some egg like that of 
elasmobranchs. The nonsegmentation of the yolk in small-yolked teleosts, the disco- 
gastrula of Cymatogaster , and the close approach to the discogastrula in the oval egg, 
in which the germ lies at the narrow end of the yolk, as in Stolephorus , all point to the 
fact that the teleostean egg was not directly derived from an amphibian-like egg. 
Ryder’s (1887, p. 493) modification of Haeckel’s theory seems to me much nearer 
the truth. Both suppose the yolk to fill the archenteron of the primitive Branchiostoma 
gastrula — very nearly the condition found in Cymatogaster. Wilson admits “that this 
theory offers an explanation of the early teleost gastrula, but it becomes utterly 
unsatisfactory as soon as what Balfour has called ‘ the asymmetry of the vertebrate 
gastrula’ begins to appear in the fish embryo. For the teleost gastrula of Ryder 
and Henneguy is a symmetrical gastrula, and they are consequently unable to explain 
why it is that (continued) invagination takes place at one pole of the blastoderm, tvhile 
the other pole grows epibolically round the yolk. v The major premise of this argument 
lies in the words I have italicized. Wilson takes it for granted that this premise is 
correct, although he has not proved that this is really the condition in Serr anus — the 
form he studied (see page 436). Certainly it is not the case in Cymatogaster, in 
which the gastrula is a symmetrical gastrula, and the blastopore closes at exactly the 
entodermic pole of the egg. It is not the case in two species of pelagic eggs ( Stole - 
pliorus), in which I have examined the gastrulation, for these also have symmetrical 
gastrulas, and the blastopore closes at the entodermic pole of the egg. The preceding- 
italics, describe exactly what does not take place in Cymatogaster and in Stolephorus, 
in which the anterior and the posterior margins of the blastoderm grow equally over 
the yolk. In Stolephorus, as I have stated elsewhere (1891), the germ lies over the 
narrow end of a very elongate yolk. To all intents and purposes we have, there- 
fore, a small-yolked teleost ovum, and when gastrulation takes place the marginal 
ingrowth exceeds (on account of the narrow yolk) the restricted lateral spreading 
of the blastoderm. The result is that the tip of the embryonic shield comes in 
close proximity to, if not in contact with, the inner margin of the anterior portion 
of the blastodermic ring — an almost compfete diplastic gastrula is formed. Nothing 
