672 E. PH. ALLIS, 



are both enclosed or partly enclosed in a bony envelope apparently of 

 dermal origin, dermal bone thus here being found inside the cartilage 

 of the skull. Posterior to the commissure, as far as traced, the 

 sensory canal lies external to but close to the lateral edge of the 

 cartilage of the rostrum, and is partly enclosed in bone. 



In the adult the commissural sensory canal above described tra- 

 verses a relatively well developed canal near the end of the carti- 

 laginous rostrum, and the median portion of this canal has become a 

 fairly large median chamber. This chamber lies anterior to and wholly 

 independent of the large central cavity of the rostrum described by 

 Bridge, and was apparently entirely overlooked by that author. It 

 is a median sensory chamber developed in relation to the lateral 

 sensory system, and is not found in any other fish I know of, ex- 

 cepting in the Muraenidae, in which fishes it is, however, wholly en- 

 closed in dermal bone and not in cartilage. This median chamber 

 in the Muraenidae lodges a part of the ethmoidal section of the main 

 infraorbital canal, as is fully described in another work (5). The 

 cross-commissural canal of Polyodon is thus probably also an ethmoidal 

 section of the main infraorbital line of the fish. 



The supraorbital canal, starting from its point of anastomosis 

 with the main intraorbital canal, lies for a short distance in the 

 antero-mesial arm of bone i, and then enters and traverses the series 

 of three or more ossicles that together constitute bone 5; the canal 

 turning gradually forward and then forward and laterally as it tra- 

 verses the chain of ossicles. The one or two posterior ossicles of the 

 series lie, as already stated, on the dorsal surface of bone i, or on 

 the dorsal surface of that bone and bone B- of Bridge's descriptions. 

 The more anterior ossicles span the space between B- and bone 2. 

 Beyond the anterior ossicle of this series the canal crosses somewhat 

 transversely the dorsal surface of bone ^, there being a dermal tube 

 supported by imperfect bone formation. At the lateral edge of bone 2 

 the canal turns downward and forward and enters a little ossicle that 

 lies in the roof of the nasal cavity, between the two nasal openings. 

 Several tubules arise from the canal as it traverses this little bone, 

 and they would seem to belong to two dendritic systems, though they 

 may all belong to a single system. If they belong to two systems 

 one of them is certainly a terminal system; for the canal is not 

 continued beyond the bone. The bone is thus evidently the nasal 

 bone of the fish, and it was apparently wholly overlooked by both 

 Bridge and Collinge, the former of which authors says that the 



