No. 388.] CORRESPONDENCE. 367 
tary canal, because if the arthropod brain is already hollow, why 
should any one introduce the alimentary canal into it in order to 
explain why the brain is hollow in vertebrates ? 
I have described in the same number of the Quarterly Journad in 
which Gaskell’s earlier paper appeared how the cartilaginous endo- 
sternite in scorpion and Limulus is comparable in shape and in its 
relations to the brain and alimentary canal with the primordial cranium 
of vertebrates, and how a complete ring is formed about the posterior 
part of the brain, like the occipital ring seen in the early stages of 
Fic. 5. — Gaskell’s diagrams of the aa ConA oy him to show how much the 
cartilaginous skeleton of Limulus f Ammoceetes, B. 
the cartilaginous cranium of vertebrates. Gaskell makes no refer- 
ence to these facts, but nevertheless he lays great stress on the pres- 
ence of the endosternite, although it lies on the opposite side of both 
the nervous system and the alimentary canal from what his theory 
demands. To meet the requirements of his theory, the endosternite 
must first be split in halves lengthwise, and the two parts transferred 
to the opposite side of the nervous system, and then, after their 
reunion, a new occipital ring must be formed on the opposite side 
of the sternite from the one where it actually is formed in the arach- 
nids. (Compare Figs. 1 and 2.) And before all these changes could 
take place, the long fragments of the sternite would have to plough 
their way through the nervous system, beneath the epithelium of the 
cerebral vesicles and the canalis centralis (Fig. 1). But even then, 
