The Lateral Sensory System in the Muraenidae. 135 



deepl,y buried beneath overlapping- portions of the trunk muscles, Muraena 

 (lift'ering- somewhat in this from the other Muraenidae examined. 



The preoperculo- mandibular canal, beginning- at the point of 

 anastomosis -with the main infraorbital canal above described, runs 

 downward through a long tubular structure similar to the shorter 

 ones that enclose the supratemporal commissure, this long tube Ijing 

 as do those of the commissure and the adjoining parts of the main 

 infraorbital, deeply buried beneath overlapping portions of the muscles 

 of the region. The long tube lies parallel to and slightly posterior 

 to the hind edg-e of the external surface of the hyomandibular, and 

 ends approximately at the ventral end of that element. There the 

 lateral sensory canal leaves the tube and enters and traverses a small 

 bone that lies immediately posterior to and against the quadrate part 

 of the fused hyomandibular and quadrate, two blind pouches arising 

 from the canal as it traverses the bone. This little bone must there- 

 fore be the homologue of the preoperculum of Ophicthys and Myrus, 

 that bone of each of those fishes, j^lus the tubular structure that 

 lodges the dorsal organ of the line, naturally together representing 

 the preoperculum of such fishes as Amia and Scomber. The relations 

 of these two parts of the preoperculum of Muraena, one to the 

 hyomandibular and the other to the quadrate, is noteworthy, though 

 just what its significance may be I cannot tell. It is clearly in accord 

 with the supposition that the hj^omandibular and quadrate of the fish, 

 or at least the hyomandibular and a part of the quadrate, belong to 

 the same visceral arch. 



On leaving the ventral end of the preoperculum, the sensory 

 canal traverses the subcutaneous tissues for a short distance, and then 

 enters the mandible, which it traverses to its anterior end, six primary 

 tubes, one of them a terminal one, arising from this part of the canal, 

 and each of them opening on the outer surface by a single pore. 

 There is no indication of a primary tube between the preoperculum 

 and the hind end of the mandible. 



The ethmoidal section of the main infraorbital canal begins at a 

 pore that lies on the anterior end of the snout, ventro-mesial to the 

 anterior nasal tube and close to the ventral edge of the upper lip. 

 From there the canal runs upward and backward, traverses a central 



