366 THE AMERICAN NATURALIST. [Vov. XXXIII. 
alimentary canal. If Mr. Gaskell accepts this alternative, he will 
find it difficult to explain how the alimentary canal of the ancestral 
crab was split into two surface cords of cells that sweep round to 
the opposite side of the ovum to form the canalis centralis of verte- 
brates. 
Gaskell bases his comparison of the pineal eye of vertebrates 
with the ocelli of arthropods mainly on their minute structure, neg- 
lecting their mode of development and the fact that there are two 
distinct types of arthropod ocelli, namely, the paired ocelli, with 
upright retinas, as in Dytiscus and Hydrophilus, and the ocelli with 
inverted retinas, formed by the fusion in the median line of two 
or more ocelli (¢.g., median ocelli of scorpions, Limulus, and pos- 
sibly some crustacea). As we have shown elsewhere, the latter are 
the only ocelli that can be compared with the pineal eye of verte- 
brates, both on account of their position and the fact that they are 
` the only ocelli which lie at the end of tube-like outgrowths from the 
roof of a cerebral vesicle. This condition is brought about in scor- 
pions and in Limulus by the overgrowth of a lateral fold of ectoderm, 
which in the scorpion completely encloses the cephalic lobes or fore- 
brain, leaving the inverted ocelli at the end of a tubular outgrowth 
of the roof of the vesicle. This is one of the most satisfactory indi- 
cations we have of a relation between vertebrates and arthropods, 
for it shows us in detail how the eyes of vertebrates have in all prob- 
ability been inverted and transferred from the lateral edges of the 
cephalic plate to the ends of tubular outgrowths of a cerebral vesicle. 
In Gaskell’s earlier paper (Quart. Journ. p. 50, 1889) he compares the 
pineal eye of vertebrates with the ocelli of Dytiscus and Hydrophilus. 
He then constructs a diagram of the vertebrate pineal eye, which is 
almost an exact copy of the ocelli of Dytiscus, and this diagram is 
printed in three colors, to show that the pineal eye of vertebrates 
“is clearly that of an arthropod, and indeed of an ancient form ” 
(p. 53, 1889). As a matter of fact, however, the ocelli of Dytiscus 
and the pineal eye of vertebrates are not in the least alike, and the 
mode of development in the two cases is entirely different. On the 
other hand, the median ocelli of scorpions and of Limulus do resemble 
in a most striking way, in their position and development, the pineal 
eye of vertebrates. These ocelli a/ready “ie at the end of a tubular out- 
growth from the roof of a forebrain vesicle. These facts perhaps escaped 
Gaskell’s attention, although they were described in the same number 
of the Quarterly Journal as his earlier paper. They are fatal to his 
view that the forebrain vesicle is a remnant of an arthropod alimen- 
