148 Edward Phelps Allis jr., 



lip of the fish. The canal, from there, runs backward, upward and 

 mesially and enters the outer end of a short tubular process which 

 projects anteriorly, laterally, and but slig-htly dorsally from the dorsal 

 surface of the anterior end of the skull. The outer end of this pro- 

 cess is fringed with flbro-cartilage similar to that associated , with 

 other parts of the lateral system. Turning downward into and tra- 

 versing this process the canal reaches a relatively large median ca- 

 vity in the bone, and in this cavity there is a median sack-like 

 enlargement of the membranous lateral canal. From this median sack- 

 like enlargement the posterior section of the ethmoidal canal on each 

 side has its origin, this section arising from the sack-like enlargement 

 postero-dorsal to the point where the anterior section of the canal 

 joins the same enlargement. Running forward and laterally this 

 posterior section of the ethmoidal canal issues from the skull by tra- 

 versing a process that is exactly similar to and lies directly and im- 

 mediately posterior to the process traversed by the anterior section 

 of the canal. It then joins and fuses with the supraorbital lateral 

 canal immediately anterior to the first sense organ of that line. At 

 the base of each of these two sections of the ethmoidal canal, or 

 perhaps more properly in the median sensory chamber near the points 

 of entrance or departure of the canals, there is a lateral sense organ, 

 the two sense organs on each side of the head lying close together, 

 one posterior and somewhat dorsal to the other, and both being 

 innervated by long branches of a separate and special branch of the 

 buccalis facialis. These two buccal branches, on either side, run 

 backward in a median canal in this part of the skull, and issue by 

 a special foramen on the lateral surface of the internasal part of the 

 skull. Continuing backward, mesial to the nasal sac, the two branches 

 first unite to form à single nerve and then join that branch or division 

 of the buccalis that innervates organs 1 to 4 infraorbital. Strictly 

 similar conditions must exist in the other Muraenidae examined, but, 

 as already stated, no attempt was made to determine the innervation 

 of the sense organs concerned. 



There is thus here an arrangement of canals totally different 

 from anything that is known in other teleosts. Unable at first to 



