346 Some Phases of the Gastrulation of the Horned Toad 
blast sae laterally and forward, that the cavity of the pouch is obliterated 
to a mere crevice except in the median plane where it remains as the 
mesoblast sac, in its last condition, the canalis neurentericus, open ven- 
trally where mesoblast and hypoblast come together. 
DISCUSSION OF LITERATURE. 
While the general process of gastrulation in Phrynosoma is similar to 
that in other reptiles, there are some striking differences. In Hatteria 
(Dendy, 99) there is a stage when the embryonic area very much re- 
sembles the cap-like elevation of the Phrynosoma embryo represented in 
Figs. 1-3. The ectoderm is similarly elevated and thickened, beneath is 
a cellular mass of hypoblast and on the posterior margin of the area is 
the blastopore, a depression of considerable breadth. From the vicinity 
of the blastopore, the mesoblast spreads laterally and the head. process 
grows forward. * This early growth of mesoblast, when the mesoblast sac 
is very shallow, exists in Phrynosoma. In Hatteria, however, the head 
fold either appears much earlier than in any of the lizards, or Dendy’s 
figure is of a stage in which the blastopore has become closed with age. 
The lack of information as to further details of the development of the 
blastopore in Hatteria renders the above comparison of little value. Be- 
cause of the work of Mitsukuri our knowledge of the embryology of the 
Chelonia is much more complete. In appearance, the blastoderm of 
Clemmys Japonica (Mitsukuri 94, Fig. 1, Pl. VI) as seen in surface 
view, is much the same as the older stages of Phrynosoma after the blasto- 
derm has flattened on the yolk. None of Mitsukuri’s figures suggest any 
elevation of the blastoderm. From his figures (Mitsukuri, 94, Figs. 9, 
13, 15, Pl. VIII), it is very apparent that in Clemmys there is less cellu- 
lar differentiation than in Phrynosoma at this early stage. The accumu- 
lation of yolk in and about the region of the primitive streak obscures 
the lateral spread of mesoblast. There are certainly no median or lat- 
eral cellular pockets at this or later stages. Mitsukuri describes a lateral 
growth of mesoblast from the primitive streak, but does not call attention 
to any evidence that it is a hollow evagination of the mesoblast sac. 
Similarly the growth of the head process anteriorly is en masse and gives 
no evidence of a ventral wall (Mitsukuri, 94, Figs. 15 and 16, Pl. VIII). 
Only in stages subsequent to the breaking through of the mesoblastic 
canal and the cellular differentiation of all three germ layers does Mit- 
sukuri find evidence that the lateral mesoblast is an evaginated growth 
from between the hypoblast and chorda. There is a somatic and splanch- 
nic mesoblast, the former continuous with the chorda, the latter con- 
