4 SMITHSONIAN MISCKLLANEOUS COLLECTIONS VOL. 56 



Tlie basal radii of Elops are also much more regularly spaced, and 

 the laterobasal angles are rounded. 



According to the scales, the four famil''"'^ considered are related 

 thus : 



A. Albulidte. 



B f «. Monnvridse. 



\h. . . '. . . / I. Rlopidte. 



( 2. niodontida?. 



A remarkable thing about the scales of Hiodon is their close re- 

 semblance to those of certain old world Cyprinidse. Comparison 

 may be made, for instance, with Cyprimis carpio. The resemblance 

 is such that, if the scales had come to me nameless, I should certainly 

 have guessed them to be Cyprinoid. The Hiodontidae cannot be 

 directly related to the Cyprinid;e, but I believe that they may stand 

 close to the ancestors of the Characinidie. I do not know a Chara- 

 cinid scale similar to Hiodon, but, of all the South American Chara- 

 cinids, 1 know scales of only two genera. 



