Cockerell — Scales of Panama Fishes. 155 



are all latini;cleate, but they agree with those of B. falcatus in all 

 essential characters; the apical radii are many and parallel, and the 

 apical margin is strongly crenulate or subdentate, though not ctenoid 

 in any proper sense. 



TETRAGONOPTERiN,E. Two species of Astyanax are given in the table 

 above. A. fischeri is related in the characters shown by the scales to 

 ^4. bimaculatus, though the scales are much smaller (5-5.5 mm. broad), 

 the circuli are less dense, and the weak subparallel apical radii are very 

 different. The relationship is perhaps actually closer with the still 

 smaller scales of the A. polylepis group. A. grandis is very close to 

 fischeri. Thus the two Panama species form a little group intermediate 

 between the South American groups of A. bimaculatus (large scales 

 with very coarse spreading apical radii) and of A. polylepis, abramoides , 

 etc. (small scales with very weak apical radii). The J., mucronatus 

 group (with V-like apical radial pattern) stands apart from all these; 

 to it must be added A. fasciatus Cuvier, Rio Primero, Argentine 

 (J. W. Titcomb). The mucronaius-fasciatus group is typical Astyanax. 

 Poecilurichthys, which Eigenmann has recently treated as a distinct 

 genus, is typified by A. bimaculatus ; but according to scale characters 

 we should be inclined to include in it some of the species which 

 Eigenmann, in his catalogue, has left in typical Astyanax. 



GASTEROPELECIN.E. Gasteropclecus maculatus has scales which do not 

 differ in any tangible way from those of G. sternicla L. 



piABUCiNiE. Piabucina panamensis has scales of the same general type 

 as those of Chalceus macrolepidotus , but much smaller (length 7, 

 breadth 8 mm.). The intermediate, minor apical radii are not so evi- 

 dent in the Piabucina as in the Chalceus. In all respects the scales of 

 Piabucina are essentially as in the African genus Alestes. Both show 

 a radial polygonal pattern in the middle, at least in some of the scales. 

 Are the Piabucinae separable as a subfamily from the Alestinae? 



HYDRocYNiN^. Luciochavax insculptus Steind., Rio Abaco, Panama 

 (Meek & Hildebrand). Scales subquadrate, about 6 mm. long and 

 broad; basal margin strongly undulate, or emarginate in middle; 

 nucleus a little above the middle; circuli fine; a few basal, apical and 

 lateral radii (1 or 2 apical, 1 to 3 basal, 1 to 5 lateral, but when more 

 than one lateral, they are only partly developed) , in Alestiform fashion ; 

 apical field without circuli, but thrown into strong parallel pleats or 

 folds (not radii), which end as strong sharp teeth on the margin. The 

 circuli are microscopically beaded. These are ctenoid scales, using 

 that word in a purely descriptive sense. There is no resemblance to 

 Hydrocynus, at least judging from the scales of the latter genus seen 

 by me, which are, however, apparently quite immature. Eigenmann 

 states that adult scales of Hydrocynus have denticulate apical margins. 

 The scale of Luciocharax is singularly like that of Phractolxmus. 



CHARAciN^. The genera of this subfamily now before me may be sepa- 

 rated thus : 



