104 CALIFORNIA ACADEMY OF SCIENCES. 



Francisco; Bon Jardin, R. S. F. ; Rio Doce; Bahia; Santa 

 Cruz; Santa Clara; Rio Mucury; Sao Matlieos; Rio San 

 Antonio; Rio Janeiro; Goyaz; San Gongallo; Jequitinhonha; 

 Rio Puty; Buenos Ayres; Rio Grande; Porto Alegre; Rio 

 Arassuahy, Minas Geraes; Murialie; Guiana; Rio Ipojuco, 

 Province Pernambuco. 



This seems to be one of the commonest food fishes of all 

 the rivers east of the Andes. It is found in the whole 

 course of these streams, occurring in the Amazons between 

 Para and the Rio Huallaga. 



The coloration varies considerably with the localities and 

 with age, the different localities regulating the degree of 

 light or darkness. In specimens .06 m. long there is a con- 

 spicuous dark lateral band from the tip of the snout to the 

 base of the caudal; in larger specimens portions of the band 

 become broadened to form blotches while the remaining 

 portions are margined by a dark line and bordered by a 

 light or silvery area; with age the band, and afterwards the 

 blotches disappear; in the largest specimens the dorsal 

 surface is thickly covered with small spots. The fin rays of 

 the vertical fins are spotted and the spots frequently extend 

 on the membrane to form bands. The pectoral and ventral 

 fins are occasionally dusky without distinct spots. The 

 belly is white, occasionally spotted; the chin frequently 

 crossed by purplish bars; there is usually a round dark 

 spot at the upper posterior angle of the caudal peduncle. 



The scales are uniformly larger in this species than in 

 microlepis. In a specimen .39 m. collected by Mr. Wheat- 

 land at Buenos Ayres there are 12, 11, 11, 9 series of scales 

 from one lateral line to the other in respectively the first, 

 second, third and fourth series behind the dorsal; this is 

 due however to the splitting of the median scales of the 

 back. Several other specimens present similar conditions. 

 The scales of the lateral line vary from 39-43. 



The size of the canine teeth and the lingual patches of 

 teeth vary considerably. We have been unable to detect 



