THE AMERICAN TETRAGONOPTERINAE. 295 



ciennes. It appeared that these specimens were received from another collector 

 a year after the pubUcation of the original description of A . fasciatus. 



Dr. Pellegrin has kindly written me that the date 1820 does not preclude 

 one of the specimens being the type, for it probably only indicates the date of 

 registering. Also that the type of Chalceus opalinus credited in Cuvier's de- 

 scription to Saint-Hilaire, was, in reality collected by Delalande in Rio de 

 Janeiro. There seems to be no doubt, therefore, that Cuvier inadvertently 

 exchanged the names of the collectors of his A. opalinus and A. fasciatus and 

 there is every reason to beUeve that one of the three specimens mentioned by 

 Cuvier and Valenciennes served Cuvier as the type of A. fasciatus. His state- 

 ment, that his specimen was but 5 inches long, can apply only to the largest of 

 the three specimens. 



These specimens are stuffed and lacquered so that none of the characters 

 can be made out clearly but they are evidently the recently described A. rutilus, 

 and not the short analed fasciatus of Steindachner. 



This species, the most widely distributed of the characins, has been and is 

 giving rise to a number of distinct forms by isolation in different rivers. Stein- 

 dachner says, "Almost every river system possesses a peculiar variety of this 

 species; according to age, sex, season; according to abundance or scarcity of 

 food; according to the habitat in cool or clear mountain brooks or deeper stag- 

 nant waters the outlines of the body vary and in part also the number of hori- 

 zontal rows of scales and of the anal rays." 



Some of the forms have differentiated far enough to be universally considered 

 as distinct species. Such are the A. inexicanus reaching the United States, and 

 the A. aeneus of Central America; to these should probably be added A. para- 

 hybae. Of equal value are A. scabripinnis and A. jenynsii of southeastern 

 Brazil. Aside from these there are a number of statistical forms such as can 

 only be differentiated if a comparatively large number of specimens are examined 

 in each locality. The Rio Novo for instance has a well-marked variety of this 

 sort. In an examination of the material in the U. S. N. M. A. nicaraguensis 

 from Lake Nicaragua was thus defined as another one o£ them, but at that time 

 the southern varieties were not distinguished because there was not enough 

 material from the southern localities. Whether we call these forms species, 

 varieties or do not recognize them as worthy of name, the fact remains that 

 different rivers are inhabited by individuals that in the aggregate differ from 

 the individuals of another river — that we have here a series of species in the 

 making as the result of segregation. 



