GILBEUT AND STARKS— FISHES OF PANAMA BAY 29 



posteriorly) is contained hut three times in its length. One specimen is a young 

 male, the other an adult female with very narrow convexly curved snout, and small 

 convex mouth with thick lips. The adult agrees exactly with adults of N. planiceps, 

 except in the characters mentioned. N. osculd may represent an extreme variation 

 in that most variable species, l)ut the two forms are retained until intermediate speci- 

 mens are obtained. 



It is doubtful whether the specimen described by Eigenmann and Eigenmann 

 (1890, p, 74) as I'achisitrus oscnlus is properly referred to this species, as the mouth 

 is wider (two in head), and the intermaxillary band is wider antero-postcriorly, its 

 width one-fourth its length. The vomerine patches seem also much more widely 

 separated than in N. oi^cula or N. planiceps. In our adult female the head is 

 contained of times in total length (without caudal). 



45. Netuma elattura {Jordan <i- Gilbert). 



Known from the type specimen obtained by Dr. Gilbert at Panama in 1881; 

 and from another obtained at Albatross Station 2800, Bay of Panama, at a depth 

 of seven fathoms (Jordan and Bollman, 1889, p. 179). 



46. Tachysurus steindachneri sp. nov. 



Plate V, Fig. 9. 



Arius mrlanopus Steindachner, icSyG /;, p. 29 (Panama); not Arius mclanopus Gunther (Rio 

 Motagua, Atlantic slope of Guatemala). 



' It has been pointed out by Dr. Steindachner, in the article above cited, that 



Panama specimens of Tachjmirus, allied to T. melanopus, differ not a little from 

 Giinther's description of that species. Two specimens in the present collection, 

 21 (3) and 26(9) cm. long, agree with those examined by Steindachner. They 

 differ fi'om melanopus in the shorter head (4^ in melanop)m), the character of 

 the longitudinal groove on top of head ("indistinct, narrow, linear behind, 

 scarcely extending to the base of the occipital process" in melanopus), the shorter 

 maxillary barbels (not quite extending to the middle of the pectoral fin, in melano- 

 pus), and in the small size of the axillary pore ("nearly as wide as a nasal 

 opening" in melanopus). The description of melanopus is so lacking in detail that 

 other differences may well exist. When to these considerations is added the fact that 

 melanopus belongs to the Atlantic fauna, while no species of marine catfish is as yet 

 known to be common to the two oceans, it seems advisable to recognize the Pacific 

 form as distinct. 



Of the Pacific species, 2\ steindachneri is most nearly allied to T. liropus, but 

 the latter has the inner faces of the paired fins light or slightly dusky, instead of 

 black on basal half; the spines are longer and more slender, the snout longer and 

 more rounded at its extremity, the anterior divergent extensions of the granulated 



( 5 I October 6, 1903. 



