474 Edward Phelps Allis ji%, 



dorsal one of the two has pushed forward beyond its own territory, 

 and occupied the line of least resistance usually reserved, in elasmo- 

 branchs, for the subrostral section of the supraorbital line. Then, when 

 the growing end of the supraorbital line arrived at the point where 

 its own rostral section would normally terminate, in Selachians, it would 

 find the line of its subrostral section already occupied, and would be 

 quite naturally switched off onto the unoccupied prenasal but infra- 

 orbital line of least resistance, and would follow that line and then 

 that of the median and selachian nasal lines until it met that end of 

 the suborbital canal with which it normally fuses, in selachians, that 

 end here being the anterior end of the misnamed angular, and would 

 there end. This being so, the nasal canal of Chimaera, while corre- 

 sponding morphologically to the nasal, median and prenasal sections 

 together of plagiostomes, would not occupy an homologous line of least 

 resistance, and hence would not be the exact homologue of those canals. 

 It would however probably be the exact homologue of the ethmoidal 

 canal of the bony ganoids, and perhaps also of the transverse median 

 canal of those plagiostomes in wliich that canal is found. 



In none of the several elasmobranchs above referred to, with the 

 single exception of Mustelus, and also in none of the many described 

 and figured by Garman, is there any mention of pit lines on the dorsal 

 surface of the head. In Mustelus two pit organs on each side of the 

 top of the head, slightly anterior to the corresponding endolymphatic 

 pore, were considered by me [8, p. 117] as possibly homologous with 

 one of the head lines of Amia, probably with the middle head line. 

 The anterior head line was assumed to be included in the posterior 

 part of the supraorbital canal, but there is no direct proof that it is 

 so included. This anterior head line would seem certainly to be in- 

 cluded in the posterior portion of the supraorbital groove of Chimaera, 

 and also in that of Callorynchus, in both of which fishes it anasto- 

 moses with the supratemporal commissure, and not with the main infra- 

 orbital line. The Holocephalae would seem in this, as also in the 

 apparent possession of an ethmoidal cross-commissure, to approach the 

 teleostean arrangement rather than the plagiostomian; to represent, 

 in fact, to a certain degree, a type intermediate between the other two. 



