WESTERN ATLANTIC PERCOPHIDID FISHES — GINSBURG 625 



area near base, scaleless in larger part; caudal scaled on a larger area 

 near base; dorsal and anal scaleless; scales on sides of body moderately 

 ctenoid, others cycloid, at least in large part. (The scalation is incom- 

 plete or nearly all missing in almost all specimens examined and not 

 all details are accurately determinable; consequently, the above 

 statements might be subject to a moderate revision.) Lateral line 

 beginning at upper angle of gill opening, running horizontally over 

 upper margin of opercle, curving directly behind head, except in in- 

 dividual variants of Bembrops anatirostris (p. 636), and dipping rather 

 sharply down, posterior end of curve approximately at end of pectoral, 

 thence running nearly straight backward and nearer to lower than 

 upper profile. First dorsal with six flexible spines (the count constant 

 as determined in 124 specimens of all five species). Second dorsal 

 with 14-18 rays, all segmented; anal similar to second dorsal, with 

 17-19 segmented rays; the rays in the two fins notably well spaced. 

 Pelvic horizontally placed on ventral aspect of abdomen, its base well 

 in advance of pectoral, approximately under preopercular margin, its 

 end falling considerably short of anus; with one flexible spine rather 

 well developed and five segmented rays. Pectoral extending moder- 

 ately'" behind a vertical through anal origin, with 22-28 rays. (Caudal 

 damaged in nearly all specimens examined; seems to be moderately 

 emarginate with the rays near the upper margin somewhat longer 

 than the corresponding lower rays.) 



Methods of study: The dorsal and anal fins are very often dam- 

 aged, many of the rays being broken off, leaving only a short stump. 

 However, the rays are widely separated and readily countable even 

 when damaged and in the small specimens. The last ray is split to 

 its base and the two branches have been counted as one ray. 



The scales are deciduous and most or nearly all of them are missing 

 in almost all specimens examined. However, when the scales are 

 missing their pockets are prominently outlined. The counts of the 

 scales here given generally refer to the number of oblique rows of scale 

 pockets, beginning with the row that is wholly or partly behind the 

 short, rather sharp spine at the upper anterior angle of the gill opening 

 and ending at the caudal base. Some of the counts here recorded may 

 be off by one or two scales, especially for specimens on which part of the 

 count was made of the rows of actual scales and part of the rows of 

 pockets, or in case of specimens having the rows somewhat irregular 

 in places. But, on the whole, the counts made closely approach the 

 true numbers. 



The gill raker standing squarely within the angle of the arch is the 

 longest, and it has one "root" on the upper limb and one on the lower. 

 Following my usual procedure, it has been uniformly included in the 

 count of the lower limb. The tuberclelike outgrowths at both ends 



