491 



openings, meet in the median line and there fuse without indication 

 of crossing. In the unclassified head the carotids enter the skull by 

 a single median opening and there definitely cross each other. They 

 then turn towards the median line again, there meet and fuse and later 

 separate, and, superficially examined, the arteries might here be 

 supposed to have each definitely crossed from one side of the head 

 to the other. But the apparent crossing can be fully explained by 

 the assumption of a simple twisting, such as is found in Chlamydo- 

 selachus. And, of course, this must be the condition if the internal 

 carotids are each simply the anterior prolongation of the lateral dorsal 

 aorta of its side. And if there were crossing without anastomosis, such 

 as Parker's description would lead one to assume in Mustelus antarc- 

 ticus, there would have to be fusion and subsequent separation with 

 interchange of the anterior prolongations of the arteries. 



According to Gegenbaur (1872, pp. 74 — 79, and Fig. 1, Taf. IV), 

 the carotid canals, in Heptanchus, begin at a pair of openings lying 

 close together, one on either side, on the ventral surface of the keel- 

 shaped base of the chondrocranium, the two canals soon uniting to 

 form a single median canal. This canal runs upward and slightly 

 forward in the cartilage until it reaches the ventral edge of the canalis 

 transversus where it turns forward and upward, still as a single 

 median canal, and as such opens into the cranial cavity on the bottom 

 of the pituitary fossa (Sattelgrube). The canal is said to be definitely 

 separated from the canalis transversus by membrane (,, durch eine 

 häutige Scheidew^and abgeschlossen"). The canalis transversus is 

 said to be traversed by a vessel which puts the orbital sinuses of 

 opposite sides in communication one with the other, and these sinuses 

 are considered by Gegenbaur as large orbital lymph spaces. The 

 canalis transversus is shown communicating with the cranial cavity 

 by a small, median, vertical canal. 



In each of my two specimens the carotid canals begin inde 

 pendently on either side of the base of the chondrocranium and then 

 immediately fuse to form a single median canal, exactly as given 

 by Gegenbaur, but when this median canal reaches the ventral 

 surface of the canalis transversus it communicates largely with that 

 canal and then itself separates into its two component parts, one 

 on either side, each of which parts has a separate and independent 

 course upward and forward and opens independently on the bottom 

 of the pituitary fossa. Into this dorsal portion of each carotid canal 



