PHILIPPINE MACROUROID FISHES GILBERT AND HUBBS. 395. 



the premaxillary band, which is flattened and greatly expanded in 

 width at the middle of the length of the maxillary; the greatest 

 width of the band is contained 1.25 (to 1.65) times in the least 

 width "of the bony suborbital. 



The length and strength of the barbel are too variable to serve 

 as good specific characters, as can be understood by an examination 

 of the tables of measurements following the descriptions of each 

 species. Pseudobranchial filaments are developed, but are exposed" 

 only in our smallest specimens, having become covered in the large 

 specimens by the fold of membrane which precedes the first gill- 

 arch, and lines the deep pit opposite the upper angles of the arches. 

 The gill-rakers are blunt at their tips, rather robust, smooth, rather 

 widely spaced, and short, their length in the larger specimens being 

 less than least width of the bony suborbital; they number 5-J-17 and 

 5+19, on the left and right sides, respectively, of the type, and vary 

 from -A to 5-J-19 to 21 in the nine paratypes counted. The branchial 

 aperture curves forward to below the middle of the eye. A narrow 

 slit is present behind the fourth gill-arch. Seven branchiostegals. 

 As in our specimens of all four Philippine species of this genus, as 

 well as in the specimen of G. colletti of Japan which we examined, 

 the scapular foramen is in contact with the suture separating the 

 hypercoracoid from the hypocoracoid. 



The. scales are thin, deciduous, and cycloid. When examined under 

 the compound microscope, -the numerous striae are seen to be finely 

 denticulate; on the exposed portion of the scale the striae are ar- 

 ranged concentrically, but on the basal portion, which is the larger, 

 they form an extremely irregular zigzag pattern, and are perpen- 

 dicular (or nearly so) to the basal margin of the scale; these striae 

 are connnected by numerous cross ridges. Certain marks or narrow 

 zones may be annular rings, as they bear a striking similarity to 

 certain marks so interpreted in other marine fishes. 1 Several scales 

 from the type specimen, 112 mm. long to anus, show three such rings; 

 those from the paratype 111 mm. long, from Albatross station 5624, 

 show two rings. A region of closely approximated circuli, occurring 

 just outside the focus of the scale, suggests a similar area in sal- 

 monoid and other fishes, which in those forms is regarded as the winter 

 zone of the first full year's growth. A similar area is more clearly 

 marked off on the scales from G. colletti of Japan, and is also indi- 

 cated on the scales of G. longifilis and the other species reported on 

 in this paper. Those species inhabiting deeper water, however, do 

 not show more than traces, if anything, of .those marks which re- 

 semble annular rings. That fact is in harmony with the view that 



iSuch as Cynoscion (Taylor, Bull. U. S. Bur. Fish.. No. 34, 1914 (Sept. 23, 1916), pp. 

 295-330; numerous figures). 



