254 THE AMERICAN CHARACIDAE. 



the series above this, surrounded by a pale area; a spot on the caudal peduncle, 

 fading out forward and continued behind to the tip of the middle caudal rays; 

 caudal markings varying greatly in intensity. Sides silvery or brassy; some- 

 times a dusky cross-shade behind the pale area surrounding the humeral spot. 



Vertebrae 13 + 19. 



Air-bladders very large, the posterior one regularly curved, ending coni- 

 cally near the origin of the anal, twice as long as the anterior, 2^ times as long 

 as wide, its width 21 in the head. 



Alimentary canal about equal to the total length. 



This species varies very much in shape with the sex and with the character 

 of its indi\'idual habitat. It also varies much in more constant characters, 

 i. e. such as are not dependent on age, sex, or nutriment, with the geographical 

 distribution. The specimens from the Rio San Francisco and Bahia seem to 

 occupy the center from which the characters of the specimens to the north as 

 well as those to the south vary. These variations have for the most part 

 received distinct names which may be retained as varietal designations. Still 

 other deviations are so well separated that they are recognized as distinct 

 species. 



The following table (p. 255) indicates the total variation of the species and 

 its immediate relatives in the number of anal rays and scales in the lateral line, 

 as well as the range of variation in each locality. 



16. ASTYANAX (POECILURICHTHYS) BIMACULATUS BOREALIS EigenmaUU. 



Tetraganopterus maculatus Steindachner, Denks. K. akad. Wien, 1878, 39, p. 58 (Rio Magdalena) ; 



1880, 42, p. 73 (Rio Cauca near Caceres, Colombia). 

 Astyanax bimaculafus borealis Eigenmann, Bull. M. C. Z., 1908, 52, p. 96; Rept. Princeton univ. exped. 



Patagonia, 1910, 3, p. 432. 



Habitat. — Magdalena Basin. 



In nine specimens from the Cauca and Magdalena Steindachner found one 

 with 32 anal rays, three with 38, and the rest with between 36-39. This would 

 give the usual number to be 38 and the average 37. The nearest average num- 

 ber of anal rays in A. himaculatus in any other locality is 32.2. This difference 

 is worthy of nominal recognition. I have seen no specimens. 



The species is evidently restricted locally. 



