24 J. T. CUNNINGHAM. 
the blastoderm along the median dorsal line. It is thus easy 
to see that the formation of the embryo by concrescence in the 
actual development of the Teleostean is simply the gradual 
closing of the elongated dorsal blastopore from before back- 
wards. The diagram fig. 27 may be considered as representing 
the ancestral vertebrate gastrula with its elongated dorsal blas- 
tepore, and fig. 28 the same gastrula after the blastopore has 
closed anteriorly, leaving a small opening at the posterior end 
of the dorsal surface. The transition from the condition in 
fig. 27 to that in fig. 28 is represented in actual Teleostean de- 
velopment by the changes shown in the diagrams figs. 24, 25, 26. 
But we have further to inquire into the relations between the 
condition of the hypoblast in Teleosteans to that in the ances- 
tral gastrula. Itis evident that the part of the periblast on 
which the edge of the invaginated layer rests at fig. 24, is not 
the part with which it becomes continuous on the formation of 
the intestine. In order to understand the Teleostean gastrula 
we have to perceive that at the period of invagination (fig. 24) — 
a separation takes place between the cells destined to form the 
sides and dorsal wall of the archenteron and the cells destined 
to form its floor. In fact the yolk with its periblast represents 
the part of the hypoblast in fig. 27 between the lines o and p, 
and the continuity between the ventral hypoblast and the dorsal 
is not re-established in the Teleostean until the period of the 
formation of the intestine. 
It is this solution of continuity between ventral hypoblast 
and dorsal in the Teleostean which enables the invagination to 
take place all round the edge of the blastoderm at so early a 
stage. In the Amphibian and Petromyzon no such solution 
of continuity takes place, and the invagination of the part of 
the blastoderm corresponding to the point z (fig. 24), therefore, 
does not occur until the yolk is in the position or near it which 
it occupies at the formation of the intestine. This, it seems to 
me, is the real explanation of the difference between the inva- 
gination in the Teleostean and in the Amphibian. The edge 
of the blastoderm is, according to what we know at present, 
finally invaginated round its whole periphery in the frog, and 
