No. 388.] CORRESPONDENCE. 365 
and met in the middle line to assist in forming the new ventral 
surface, and the head fold is completed, we are watching the embry- 
onic representation of the transformation of the Limulus-like animal 
into the scorpion-like ancestor of the vertebrates.” 
It is not quite clear what Mr. Gaskell means by the above state- 
ment, but if I understand him correctly, it is clear that if the embry- 
onic shield of a vertebrate embryo represents an arthropod embryo 
on the old arthropod hzemal surface, then the growing margins of the 
mesodermic area must lie defween the nerve cords, and they should 
grow toward the mid-dorsal line and concresce there. But nothing 
of the kind is to be seen in vertebrates, because the mesoderm lies 
mainly /azera/ to the nerve cords, and grows in the opposite direction 
to what it should according to Gaskell’s theory, namely, toward the 
vertebrates’ true hamal surface. Moreover, the transformation of 
the “ Limulus-like animal into the scorpion-like ancestor” should 
show us not merely the growth of the mesodermic plates, but the 
migration of the canalis centralis and the nerve cords from their first 
position in the embryonic shield of a vertebrate (ż.e., the Limulus 
condition) to the opposite side of the egg, z.2., the vertebrate position. 
But the nerve cords and canalis centralis are already in their perma- 
nent vertebrate position, and for that very reason, according to his own 
supposition, they cannot also be in the Limulus position! The reader 
may perhaps doubt whether Gaskell, in his earlier papers, thought 
that the alimentary canal of arthropods arose from the dorsal or the 
ventral surface of the egg. If he supposed it arose from the ventral 
surface between the nerve cords, then how can it appear in verte- 
brates at the very outset of development at the opposite side of the 
egg? Consultation of the text leads one to suspect that the difficulty 
is to be solved by taking the entire roof off the crab, leaving nothing 
but.the ventral surface behind, for he also claims that in vertebrates 
the infolding of the medullary plate represents the simultaneous 
development of the nerve cords and alimentary canal of the ancestral 
crab, hence they must appear on what was the ancestral ventral side, 
so that we must suppose the medullary plate of vertebrates is seen 
through an imaginary dorsal portion of the ancestral crab, like the 
coat-tail buttons on the back of Marley’s ghost; or, if he wishes to 
avoid this dilemma by an appeal to the tottering theory of concres- 
cence, he will be forced to look on the vertebrate blastopore not as 
the original mouth of a remote ancestor, but as the entire dorsal 
surface and sides of a crab, into which are gradually swept all the 
organs which, in a crustacean, lie dorsal to the nerve cords and 
