FISHES. 123 



are grouped in one order, because at certain points they approach closely to each 

 other. On the other hand, the sijecialized extremes of some of these lines, as Hemi- 

 branchii, Lophobranchii, Plectognathi, Heterosomata, are regarded as forming distinct 

 orders, although the evidence of their descent from certain groups of Teleocephali 

 seems clear enough. It is in some regards unnatural to separate these extreme forms 

 as orders, while leaving the main branches, with their diverging relations, in one 

 group. 



As to the facts in the case, there is, however, no essential disagreement, in the 

 views of American ichthyologists at least; and whether the chief subdivisions of the 

 Teleocephali be regarded as sub-orders (as by Dr. Gill) or as orders (as Professor 

 Cope and the writer have preferred to do) is not a question of high importance. It 

 is rather a matter of taxonomlc convenience. 



Admitting then, for the present purpose, the order Teleocephali as defined by Dr. 

 Gill, we may proceed to the consideration of its several sub-orders. Of these, the first 

 four to be mentioned (Gymnonoti, Eventognathi, Isospondyli, and Haplomi) are dis- 

 tinguished by the presence of a slender duct, connecting the air-bladder with the 

 alimentary canal, as is also the case with the N"ematognathi and Scyphophori already 

 mentioned. These groups so characterized are often known collectively as the Phy- 

 sostomi. With few exceptions their ventral fins are abdominal, and, with the dorsal 

 and anal fins, are devoid of sjsines. 



In the remaining sub-orders, as in the Plectognathi, Pediculati, and other aberrant 

 or specialized orders, there is no pneumatic duct. These groups constitute the Physo- 

 clisti, and for the most part the physoclistous fishes have the ventral fins anteVior in 

 position, and inarticulate rays or spines in the dorsal, anal, and ventrals. To this there 

 are, however, many exceptions, as will be seen. Of the sub-orders of Teleocephali, the 

 one least specialized in structure, and therefore the ' lowest,' is the Gymnonoti. 



Sub-Order I. — Gymnonoti. 



This sub-order contains about twenty species of fishes from the fresh waters of 

 South and Central America, chiefly east of the Andes. They are slender and com- 

 pressed, somewhat eel-like in form, and with tapering tail, which is said to be i-epro- 

 duced when injured in any way. The characters assigned to the order by Professor 

 Cope are the following: "Scapular arch suspended to the cranium, parietal bones 

 extensively in contact, distinct ; pterotic normal ; a symplectic ; opercular bones com- 

 plete ; mouth bounded by premaxillaries chiefly ; six or seven basilar ])ectoral rays ; no 

 priEcoracoid or interclavicles ; anterior vei-tebrse united, modified, and with ossicula 

 auditus ; superior pharyngeal bones subequal, continuous." 



More obvious characters, besides the elongate body and the tapering, usually fin- 

 less, tail, are the absence of the dorsal fin and the great length of the anal fin, the 

 vent being at the throat, and the latter fin extending thence to the end of the tail. 

 The head is naked, the body naked, or more usually covered with small scales. The 

 mouth is small, sometimes placed at the end of a long tube-like snout. The gill open- 

 ings are narrow, as in the eels, and as in the latter grouj:" there are no ventral fins. 



The name Glanencheli, given to a portion of this group by Professor Cope, sug- 

 gests his view that it is intermediate between the eel-like siluroids (Clariidse) and the 

 eels. A similar view seems to be held by Professor Gill, who observes of the eels that 

 " their afiinities with the more generalized forms of the order are possibly with the 



