76 A. A. W. HUBRECHT. 
folds meet has the surface of the Pilidium become separated 
from the future body-wall of the embryo and are these two 
separated by a closed cavity which, also in the case of the 
Pilidium larva, has for many long years borne the name of 
amnion cavity. It makes no difference that in the Pilidium 
the process occurs at four different spots, the products of 
which fuse later. A. Willey (98) has speculated upon 
similar relations between arthropod embryos and their larval 
envelope, also designated as amnion in Peripatus, Lepisma, 
Gryllus, Forficula, and others; has rightly interpreted the 
direct comparability with the vertebrate trophoblast, and 
has looked upon it (as I have done [795 8] for vertebrates) as 
an adaptation to a viviparous habit acquired by the terrestial 
descendant of an aquatic ancestor. 
And so not only can we link on the larval trophoblast to 
cases met with amongst the invertebrates, but even for the 
development either of a closed amnion or of an amnion 
arising by circular folds (perhaps in the case of the inverte- 
brates also a secondary modification) do we find examples 
in the invertebrate kingdom. 
It seems to me that in the case of vertebrates we may be 
content to say: (a) that wherever an amnion is met with it is 
the sequence of the separation of the embryonic ectoderm 
from the larval envelope; (b) that this larval envelope (tro- 
phoblast, giving rise to chorion, diplotrophoblast, or serosa) 
is always antecedent to and must be older than the amnion; 
(c) that the actual separation of the embryonic ectoderm from 
the trophoblast can be witnessed in those mammals where the 
amnion is from the first a closed cavity ; and finally (d) that 
in those cases, both among mammals and Sauropsids where 
we do not notice a direct separation between embryonic 
ectoderm and trophoblast in consequence of which an 
amniotic space arises, we see the amniotic cavity appear at 
a later stage, thanks to a folding process, which may be 
entirely restricted to the ectodermal tissues, and for the 
formation of which the presence of mesoblast is in no way 
required, 
