260 



side wall of the skull, and having passed through a foramen between the 

 prootic and parasphenoid, enters the rayodonie. There it unites with its 

 fellow of the opposite side, and the single median artery so formed 

 turns upward and is said, in all descriptions, to perforate the floor 

 of the cranial cavity and enter that cavity. There it separates im- 

 mediately into two parts, one on either side, and from each of the 

 two cerebral arteries so formed a branch is sent out of the cranial 

 cavity again, through the opticus foramen, and accompanying the 

 nervus opticus, enters the eye ball to supply the retina. This, at 

 least, is the arrangement as described by Allen ('05) and myself 

 (Allis, 1 and 2) in the mail cheeked fishes, and as indicated but not 

 definitely given in my (Allis, '03) descriptions of Scomber and in 

 Silvester's ('04) descriptions of Lopholatilus. And it would also 

 seem to be the arrangement in the Anacanthiui, although Cole and 

 Johnstone's ('01) descriptions of Pleurouectes, and Parker's ('84) 

 descriptions of Gadus are not detailed enough to make this certain. 

 The apparently typical arrangement in these fishes thus differs quite 

 markedly from that in both Ameiurus and Polypterus (Allis, 3), in 

 both of which fishes the internal carotids of opposite sides do not 

 coalesce in the median line, and the artery itself does not enter the 

 cranial cavity until after it has given off the retinalis artery. This 

 seeming an important and unaccountable difference, I have re-examined 

 the artery in the mail cheeked fishes, to see if the current descriptions 

 were not in some way incomplete or wrong; using for this purpose 

 sections of a small adult Scorpaena. 



The internal carotid of this small Scorpaena, as it passes inward 

 through its foramen between the prootic and parasphenoid passes over 

 the ventral or ventro-anterior edge of the prootic without perforating 

 either the bone or the related cartilage; thus entering the myodome 

 either through the hypophysial fenestra of the primordial cranium, or 

 through the orbital opening of the myodome, these two openings being 

 confluent. Reaching the myodome, the artery there lies in a cavity 

 in the cranial wall and not in the cranial cavity proper, as I have 

 fully explained in several of my works. While still in this cavity the 

 artery coalesces with its fellow of the opposite side and the single 

 artery so formed penetrates the membrane that forms the roof of the 

 myodome and also the floor of the cranial cavity. But it does not 

 perforate the membrane and reach its inner surface; an apparently 

 important point, wholly overlooked in all earlier descriptions, my own 

 included. The median artery, as it reaches the dorsal end of the short 

 and nearly vertical portion of its intramyodomic course, separates into 



