Classification of the Order Ostariophysi. 29 



and its allies, in which the cleitlira are more expanded and 

 tnincated anteriorly; all the American genera seem to be 

 Leuciscines, and Alburnus and Ahramis also pertain to this 

 group, to which Bhodeus is nearly related ; the Barhun 

 group differs in that the cleithra are distinctly emarginate 

 anteriorly. 



These characters are not sufficiently well marked for the 

 definition of subfamilies, and others, such as the pharyngeal 

 dentition, the form of the pharyngeal process, &c., are of use 

 only in defining genera or small groups of genera. 



(jtiinther's Cyprinina seems to be a natural group, after 

 excluding the North-American genera, but to it should be 

 added Eohteichthys and Osteohrama, with the osteological 

 characters of Barbus, and doubtless Leptobarhus and Myslacu- 

 leucus also ; Tinea seems to be nearer to Barbus than to 

 Leuciscus. The Jtiasborina and Danionina should be united 

 and some of the Abramidina should be added to this group, 

 some to the Leuciscina. Xenocypris is a Leuciscine, and the 

 aberrant Semiplotus is, perhaps, nearest to it. Hypophthalm- 

 ickthys is nearly related to the Barilius group. 



Thanks to the kindness of Mile. Dr. C. L. Popta, I have 

 been able to examine a specimen of the remarkable Bornean 

 Cyprinidj Gyrinochilus pustulosus, Vaill. This fish is 

 evidently closely related to Discognathus, which it resembles 

 in form, scaling, structure, and position of the fins, structure 

 of the air-bladder, inferior mouth with the united lips ex- 

 panded and papillose, and even in the groove on the snout 

 and the disposition of the tubercles on the head. Gyrino- 

 chilus differs externally from Discognathus especially in the 

 much broader lips, folded vviien retracted and when expanded 

 recalling the suctorial disc of Petromyzon^ and in the structure 

 of the gill-opening, the upper part of which forms an in- 

 halent orifice, the opercular membrane being curved inwards 

 in front of the pectoral arch in this region. As described 

 and figured by Vaillant, each branchial arch has a double 

 series of gill-rakers developed along the upper edge of the 

 gill, filtering the inhalent current of water. Internally 

 Gyrinochilus is remarkable chiefly for the very long and 

 much convoluted intestine, the slender toothless lower pha- 

 ryngeals, the absence of a horny pad, and the reduction of 

 the pharyngeal processes of the basioccipital to a pair of short 

 blunt projections, much as in some Cobitidse and Homalo- 

 pterida3. 



Extraordinarily aberrant as Gyrinochilus is, its place '\n 

 the system seems to be in the family Cyprinidse next to 

 Crossochilus and Discognathus ; to make it the type of a 



