122 Cockerell and Callaway — The Scales of Fishes. 



origin. In their scaling, however, they show a distinct approxi- 

 mation to the otherwise aberrant Campostominse , and it is proba- 

 bly safe to say that they are much nearer Campostoma than are 

 any of the other herbivorous genera. 



Depending principally on scale characters, the four sub- 

 families of herbivorous Cyprinids may be separated as follows: 



Subfamilies of North America Herbivorous Cyprinid.^. 



Scales small, with small radii all around; breadth usually 

 greater than length, the scales transversely suboval. Small 

 fishes of the Eastern United States, with two dark bands on 

 each side, the upper sometimes faint. 



Chrosomin.e iChrosomiis Raf.). 



Scales small, with strong but not very numerous apical radii, 

 basal radii evanescent but usually more or less indicated. 

 Silvery fishes of the Pacific coast region without evident dark 

 band. 



Chrondrostomin^ iAcrocheilus Agass., Orthodon Girard). 



Scales large and broad, apical radii numerous, no trace of basal 



radii . 

 Nuclear area large and very broad, the radii subparallel. 



Campostomin^ (Campostoma Agass.).* 



Nuclear area small, the radii converging; fishes of small or 

 moderate size with one dark lateral band. 



Pimephalin^ (Pimephales Raf., Hyhognathus Agass.). 



Hybognathus is remarkable for the very thick skin covering 

 the scales; in Pimephales the skin is thin, and the radii are 

 clearly visible. The gill-lamella^ judging from our material, 

 appear to present diagnostic characters. In Orthodon and Acro- 

 cheilus they are disposed in a very regular manner, fitting closely 

 together. In Jlj/bognatJius nuchalis they are regular, and scarcely 

 fimbriate, but in Pimephales amdi they are strongly fimbriate. 

 In P. anuli they are obtuse apically; in P. votatus tapering. 

 ]n Ghrosomus daJcotensis ihey are stout, in C. erythrogaster slender . 

 Possibly these characters would not hold good in a long series 



* The pharyngeal teeth of Campostoma are more or less hooked and really very like 



tliose of Lencigciif!. etc. 'I'hc rchitionsliiii of the ('ainposloniiiKr mid riiiu'|ilnilii];v' (o tho 

 Leuciscine scries is a inattor to be further investigated. 



