b PROCEEDINGS OF THE NATIONAL MUSEUM vol. 119 



the origin of the pelvic fin and the anal fin, the fourth above the 

 anal fin origin, and the fifth in the lower half of the caudal fin root. 



Caudal fin rays and body swimming position: The normal principal 

 caudal ray count is 10/9 in characids, and this is true of all Nanno- 

 stomina examined by me. However, in two species, two of the rays 

 normally in the upper lobe of the caudal fin have their distal ends 

 in the lower lobe. The proximal ends of these two rays retain the 

 normal position on the hypural fan and thus originate in the upper 

 lobe. The two species having this arrangement swim at an inclined 

 angle, head up. Steindachner (1876) fu'st noticed that the lower 

 caudal fin lobe is larger in some Nannostomina, and Hoedeman (1950) 

 noted that these unequal lobes are correlated with the swimming 

 position. Hoedeman did not describe accurately the morphological 

 arrangement of the caudal fin rays. He stated that there is a differ- 

 ence in the shape of the swim bladder correlated with the swimming 

 position. This is possibly true, but investigation of this feature has 

 been postponed pending a histological examination of the swim bladder 

 and its musculature. A comparative histological examination of the 

 semicircular canals and associated structures probably would show 

 also interesting morphological differences associated with the oblique 

 swmiming position. Braemer and Braemer (1958) have begun to 

 investigate some of the morphological differences in the arrangement 

 of statoliths in characids, including that in Poecilobrycon eques, one 

 of the oblique swimmers. 



Other characters have been reported to have generic significance 

 in Nannostomina. For example, Eigenmann (1909) based Poecilo- 

 brycon mainly on the presence of an adipose fin in the type species, 

 P. harrisoni. However, additional specimens have shown that 

 members of this species may or may not have an adipose fin, and the 

 character is of no generic significance in this group. Steindachner 

 (1876) noted that the adipose fin is variably present in Poecilobrycon 

 eques. Hoedeman (1950) attempted to use the shape of the teeth 

 as a generic character, but their shape often varies from individual to 

 individual as well as among different geographical populations of the 

 same species. 



The characters most useful in separating species are color pattern, 

 scale counts, gill-raker counts, numbers of vertebrae and teeth, and 

 body proportions. Preliminary observations of living specimens in 

 aquaria indicate that behavior also may be of considerable aid in 

 understanding specific and generic relationships. 



The synonymy used here is selected, and no attempt has been made 

 to include all references to these fishes that have appeared in the 



