FISHES FROM BERING SEA AND KAMCHATKA. 5 1 



Dorsals wholly separate; spines very slender, the fifth spine longest, 2.5 in head; pectoral reaching 

 to fourth anal ra)', its lower 7 or 8 rays thickened, the membranes incised, bat not deeply; ventrals 

 reaching nearly to front of anal, the spine and outer ray closely joined, thickened, bordered with a 

 membranous flap; caudal forked, but not so deeply as in Triglops forficala, the middle rays about 

 four-fifths the outer. Series of spinous scales on basal portion of upper pectoral and upper caudal 

 rays; very fine scales on rays of dorsal fin. 



Color above brownish gray, below white; breast and abdomen silver}-; back crossed with 5 saddle- 

 shaped dark bars which extend down to the lateral line, the first under the first to the ninth dorsal 

 spines, the second under the fifth to the eleventh dorsal rays, the third and fourth equally spaced 

 under the remainder of the soft dorsal, the fifth on the caudal peduncle; an irregular indistinct more 

 or less wavy dark streak below the lateral line, and projecting from its lower margin a series of short 

 dark bars or blotches, one or more of them often double, with their lower ends joined by a horizontal 

 line, thus forming a dusky rectangle with a light central area, the spaces between these blotches 

 silvery; a dark streak runs forward and downward from the base of the upper caudal lobes, a small 

 blotch often at the base of the lower lobe; a dark blotch near the tips of the outer caudal rays (absent 

 in females); a dusky bar extending downward and backward from the eye; a dusky streak on the 

 lower half of the preorbital, extending to the edge of the snout; Ihree blotches on premaxillar>-, the 

 posterior one continued on maxillarj-; lower lip dusky; dorsals faintly crossbarred; distal half of anal 

 rays dusky; ventrals pale; branchiostegals without dark bar. 



This species differs from Triglops forficata in the shorter body, the smaller number of dorsal and 

 anal rays, the less widely forked caudal and in a number of minor characters. From Triglops jordani, 

 from Japan, it differs, among other details, in the presence of a distinct series of dorsal scutes, in the 

 incised branchiostegal membranes (these having the posterior border a straight transverse line on 

 jordani) and in the presence of wider naked strips along base of anal fin and on lower median line of 

 caudal peduncle. 



Seven specimens besides the type were taken at stations 4777 and 4779, Petrel Bank, Bering Sea; 

 depth 52 to 54 fathoms. 



Triglops forficata (Gilbert). 



Two specimens were taken at station 4779, on Petrel Rank, and numerous specimens at stations 

 4788, 4789, and 4792, between Medni and Bering Islands, Bering Sea; depths, 54 to 72 fathoms. 



The genus Elanura, based on the present species, was characterized by the widely forked caudal 

 fin, the elongate body, and the lengthened dorsal and anal fins. Two species more recently discovered, 

 Elanura jordani from Japan and Triglops melopias of this paper, are intermediate between Elanura and 

 typical Triglops, for the body is scarcely more elongate tlian in Triglops beani, the dorsal and anal fins 

 are but little lengthened, and the caudal fin, though usually sharply forked, is less widely cleft than in 

 forficala and may be only deeply emarginate, as is usually the case in T. beani. 



Closely related to the Elanura group is Prionistius (macellus), differing in no important respect 

 save the absence of cross folds on the breast. But as these are reduced in Triglops vieiopias, and may 

 even be wholly absent, Prionistius can not be retained. The row of dorsal scutes is subject to pro- 

 gressive reduction among these species, forficata and melopias having them distinct but very small, 

 jordani having the series so reduced that the scutes are distinguished with difficulty (the species hav- 

 ing been described as without scutes), and macellus lacking any trace of them. Different as are the 

 extremes of the series, it seems im])racticable to draw a line separating them, and the genus Elanura 

 is here withdrawn. 



In Triglops forficata the caudal fin is much more deeply forked in males than in females, the shape 

 of the caudil in females being similar to the more deeply cleft fins in jordani and macellus. The cross 

 folds on breast are subject to variation in forficala, being usually present in small number, frequently 

 reduced to but one or two and occasionally wholly wanting. 



