266 BULLETIN OF THE UNITED STATES FISH COMMISSION. 
Amphibia in whicii the posterior end of the blastopore does not persist as the anus, but 
closes), Fig. 65, PI. xcvi, the blastopore remnant being plugged up in both types by 
the yolk. 
The asymmetry which Balfour (7) showed to be a characteristic attribute of verte- 
brate gastrulation, is present in the highest degree in the Teleost gastrula. At the 
posterior pole of the blastoderm (dorsal lip of the blastopore) there is an extensive 
invagination, which gives rise to the roof of the archenteron. The opposite pole of the 
blastoderm (ventral lip of blastopore) incloses the yolk (=yolk cells=floor of archen- 
terou) in an epibolic fashion. The cause of the asymmetry must be looked for in a 
peculiarly localized distribution of the yolk in the egg. The yolk not only lies in the 
hypoblastic iiart of the egg (so to speak), but in that part of it which corresponds to 
the ventral hypoblast of the gastrula. Consequently when the meroblastic blastula 
is perfected (Fig. 30, PI. xci) the so-called “ blastoderm ” contains not only the ecto- 
dermic half of the primitive simple blastula, but also contains one half of the hypo- 
blast. This half of the hypoblast invaginates as it did in the yolkless ancestor, but 
the other half (yolk) must be inclosed epibolically. 
The alimentary canal is formed from the roof of the archenteron exclusively. How 
this was effected is easy to see. The increase in size of the mass of yolk cells (of 
Amphibia) brought it about that the dorsal parts of the embryo were early folded off — 
some time before the alimentary canal was completed ventrally. The division of labor 
already far advanced between the dorsal and ventral hypoblast of the gastrula next 
took the final step: the dorsal hypoblast assumed the entire function of forming 
the gut, while the ventral hypoblast became transformed into pure food material. 
The yolk is consequently to be looked on as an organ of the gastrula which has lost 
its original function, but which in doing so became adapted to another function to 
which it owes its large size. 
Tip to this point the discussion of the extra-embryonic germ ring has been avoided. 
The interpretation of this part of the embryo is a mere corollary of Ziegler’s concep- 
tion of the gastrula, as originally stated (48). A comijarison of the frog’s gastrula 
(Pig. 11) with the fish gastrula (Fig. 10), after the preceding discussion, leads at 
once to the homology of the extra-embryonic germ ring, [e. e. g. r.) with the ventral 
mosoblast of the frog {v. mes.) In my preliminary communication (46) I made the fol- 
lowing statement : “With reference to the meaning of what may be called the non- 
embryonic part of the germ ring, Ziegler is by no means clear, though the interpreta- 
tion seems to me a mere corollary of the foregoing (Ziegler’s) propositions.” Since 
writing this I have received a letter from Professor Ziegler which, read in connection 
with his brief mention of the point in his second paper (47), satisfies me that he has 
Jield for some years the view of the extra-embryonic germ ring to which the present 
piece of work has led me. It gives me pleasure to find that as regards this point as 
well as in the general interpretation of the gastrula, I have been led to the same con- 
clusions as Professor Ziegler. The point is one, however, which deserves a somewhat 
ampler notice than Ziegler gives it. 
The adherents of the concrescence theory in vertebrates regard the extra-embryonic 
germ ring as hypoblastic. Thus Agassiz and Whitman (1) state their opinion as fol- 
lows: “We think that what we have described as the entodermic ring (germ ring) 
corresponds to the chorda-entoblast of Eana ; and it seems plausible that the periblast 
should correspond to the ‘ Darmentoblast.’ On tliis view we should expect the peri- 
blast to take some share in forming the alimentary canal, which can not be admitted if 
