328 PROCEEiDINGS OF THE NATIONAL MUSEUM \ol. 94 



black blotch on body on upper base of pelvic fins; one or two black 

 blotches on naked area of body at each side of front of anal iin, 

 sometimes lacking in young specimens; often black pigment areas on 

 under sides of lateral plates; black blotch arovmd bases of last dorsal 

 rays; blackish blotch in middle of first plate in front of dorsal fin; 

 peritoneum dusky. 



Remarks. — This new subspecies differs from L. uracantha uracantha 

 Kner and Steindachner in having a wider body, a shorter caudal 

 peduncle* (see table 24 and key), the plates along midsides with keels 

 widely spaced numbering 15 to 17 (average 16), and the plates with 

 keels approximated numbering 12 to 14 (usually 13 or 14) instead of 

 13 to 17 (average 14 or 15) and 12 to 16 (usually 14 or 15), respectively. 

 The black spots each side of anal fin base on naked area at front of anal 

 are lacking in L. uracantha uracantha. 



Named rupestre in reference to its habitat among rocks and stones 

 on the stream bottom. 



LORICARIA FiLAMENTOSA Steindachner 



Armadillo Machete 



Loricaria filamentosa Steindachner, Denkschr. Akad. Wiss. Wien, vol. 39, p. 45^ 



pi. 9, 1878 (Magdaleiia River). 

 Loricaria fimbriata Eigenmann and Vance, Indiana Univ. Studies, No. 16, p. 12, 



1912 (Boca dc Certegr.i). — Meek and Hildebrand, Publ. Field Mus. Nat. 



Hist., zool. ser., vol. 10, No. 15, p. 260, 1916 (Rio Capeti, tributary to Rfo 



Tiiyra). 

 Loricaria filamentosa latiura Eigenmann and Vance, Indiana Univ. Studies, 



No. 16, p. 13, 1912 (Boca de Certegui, Colombia). — Meek and Hildebrand, 



Publ. Field Mus. Nat. Hist., zool. ser., vol. 10, No. 16, p. 257, 1916. 

 Loricaria tuyrensis Meek and Hildebrand, Publ. Field Mus. Nat. Hist., zool. 



ser., vol. 10, No. 8, p. 81, 1913 (Tuyra Basin). 



After examining a series of para types of the species described by 

 Eigenmann and Vance, Meek and Hildebrand, and the figm-e by 

 Steindachner, then making com.parisons with my material from the 

 Maracaibo Basin, I conclude that L. Jlmbriata is the young of this 

 species and has a naked abdomen until it reaches a length of about 

 100 mm. when the plates begin to fonn; at a length of 114 mm. (see 

 Eigenmann's fig. 1, pi. 15, m Mem. Carnegie Mus., vol. 9, 1922) some 

 plates have formed along midventral region and at sides. Stein- 

 da diner's figure of filamentosa and Eigenmann's figure 3, plate 15 

 (loc. cit.), of latiura represent the abdomen as almost completely 

 plated except an area from the anus to the base of pel vies, this condi- 

 tion being found in specimens about 200 to 250 mm. in standard 

 length. Specimens of larger sizes up to 355 mm. have been described 

 as tuyrensis and latiura. 



The form from the Magdalena system appears to be slenderer than 

 the specimens before me from the Maracaibo Basm. I hesitate to 

 describe these from the Maracaibo Basin as a new subspecies because 



