110 Annals of the Carnegie Museum. 



Serrasalmonin.e. 



Pygopristis denticiilatus (Cuvier). Scales oblong, not very far from 

 circular, the nucleus a short distance below the middle; the whole 

 scale covered with practically uniform circuli, forming complete 

 circles; no radii. The nucleus is very small, not sculptured. Some 

 of the scales differ by having a very broad, strongly reticulated 

 nuclear field, and the circuli (13 or 14 in all) widely spaced. Inter- 

 mediate forms also occur. These scales are very closel}' related in 

 structure to those of Ctenobrycon. 



Pygocefitrus piraya (Cuvier). Small scales with precisely the same 

 characters as those of Pygopristis, and the same variation, except 

 that when the nuclear field is broad and sculptured, the sculpturing 

 is labyrinthine. 



Serrasalmo gymnogenys Giinther. Again the same type as Pygopristis. 



Serrasalmo rhombeus (Linnaeus) (Plate XXVII, figs. 7 and 7a). Again 

 the same, with two forms, as shown in the figures. These scales 

 closely resemble those of Argyrosomus, in the Salmonidae. 



MVLINiE. 



Myleus rubripinnis Mtiller & Troschel. Scales exactly like those 



of the Serrasalmoninae. The circuli very fine and dense in the 



normal form. Plate XXVII, fig. 8, shows a double scale, a 



monstrosity. 



Cynodontin^. 



Hydrolycus scomberoides (Cuvier) (Rockstone, British Guiana, col- 

 lected by Max Ellis). Scale agrees with that of Myleus rubripinnis, 

 but is larger, and the lateral line scale has a curious system of basal 

 grooves, with four broad Y-like prongs pointing basad, the most 

 lateral reaching the prominent laterobasal angles. The ordinary 

 scales are rounded, without angles. The grooves of the lateral line 

 scales are curiously suggestive of the extinct Chirocentrid genus 

 Cladocyclus. 



Characin.e. 



Charax gibbosa (Linnaeus) (Plate XXVUI, fig. i). Very broad scales 

 without angles, as the figure well shows. The structure is like that 

 of the Serrasalmoninae, with however the important difference that 

 the apical region is sharply differentiated, without true circuli, but 

 with fine growth-lines simulating them. x-Xlong the transition line, 

 especially mesad, is a fine reticulation. There are no radii. Al- 



