366 
the eyestalk to the orbital wall would change this relation. To produce 
the urodelian relation of nerve and process by a simple shifting of 
the parts concerned, the eyestalk would have to be carried relatively 
farther backward in the orbit, and the inferior division of the oculomo- 
torius, together with certain of the muscles it innervates, would have 
to slip, in the process, over the outer end of the stalk before the latter 
had acquired an attachment to the palato-quadrate. And the same 
applies to the nervus abducens and its related muscle. But there is 
still another and it would seem a more probable supposition that can 
be made to explain the differing conditions, and that is that the eye- 
stalk, in acquiring the position of the urodelian process near the hind 
end of the orbit, had pushed backward, before it, certain of the eye 
muscles so that they had a curved and disadvantageous course around 
the eyestalk and then forward internal to it. Because of this disad- 
vantageous course, the muscles, in growing from their myotomes of 
origin toward the cranial wall to acquire their points of attachment 
there, would naturally have gradually acquired a shorter and more 
direct course anterior to the eyestalk. The related nerves, growing 
outward from the brain, would then have encountered and joined 
the muscles approximately at the same places as formerly, but lying, 
in their course, always anterior or internal to the eyestalk or process, 
as they do in urodeles. This assumed readjustment of nerves and 
muscles may even give an explanation of the radical and puzzling 
differences found in the innervation of these muscles in different 
members of the gnathostome vertebrate series. For it is to be noted 
that, so far as known, it is only in those gnathostome vertebrates that 
have a proper eyestalk that the superior branch of the oculomotorius 
innervates two of the recti muscles. 
An eyestalk is said to be found in many teleosts, and I have my- 
self described it in Scomber (Aunıs, 1903), but Harman (1899) concludes 
that this stalk in teleosts is probably not homologous with the carti- 
laginous eyestalk of elasmobranchs. And it seems a somewhat note- 
worthy circumstance that in all teleosts so far properly examined 
in this respect, with the single exception of Ameiurus, the internal 
and inferior recti muscles are not innervated as they are in Amia and 
in the higher vertebrates (Auuıs, 1908). 
In an earlier work (Auuıs, 1897a), I gave the innervation of the 
eye muscles in urodeles as similar to that in elasmobranchs, this being 
based on the very unsatisfactory descriptions I could find at the time 
