140 EDWARD PHELPS ALLIS, JUN. 



branch o£ the oculomotorius that innervates the rectus 

 internus would still necessnrily always remain on the same 

 side of the opticus, ventral or dorsal as the case may be. He 

 chooses Bsox to represent one of the two manners of innerva- 

 tion, and Carcharias to represent the other. Take his 

 figure 4, which shows the muscles and nerves of Esox, and 

 imagine the rectus internus to shift, at its origin, forward in 

 front of the opticus and then backward above it. This is 

 simply what Corning proposes, but in the reverse direction, 

 and it is the only way in Avhich the muscle can be brought 

 dorsal to the opticus without assuming that the nerve cuts 

 through the muscle or the muscle through the nerve. The 

 muscle itself, in thus shifting at its origin, would evidently 

 acquire the selachian position, as Corning asserts, but the 

 nerve that innervates the muscle would most certainly not. 

 It would still remain ventral to the opticus. 



In Esox the internal and inferior recti are not innervated, 

 as I assumed, from earlier descriptions, that they were in all 

 teleosts, and as they ai'e in Amia. Corning calls attention to 

 this, and Herrick (32), before him, had called attention to 

 a similar difference in Menidia. I myself had already found 

 this to be true of Scomber and certain other teleosts, and had 

 called attention to it in a work on Scomber sent to press now 

 nearly three years ago. In all these fishes the arrangement 

 seems practically to be that the superior division of the 

 oculomotorius innervates the rectus superior alone. A branch 

 is then given off which separates into two parts, one for the 

 rectus internus and the other for the rectus inferior, both 

 branches lying morphologically ventral to the optic nerve, 

 but the branch to the rectus internus passing dorsal to the 

 rectus inferior, instead of postero-ventral to it as it does 

 in Amia. The remainder of the oculomotorius then runs 

 downward and forward around the hind edge of the rectus 

 inferior, and supplies the obliquus inferior. Herrick con- 

 siders this latter part of the nerve as the first branch of the 

 oculomotorius, the remainder of that nerve later separating 

 into dorsal and ventral portions, the former for the rectus 



