FISHES OF FAIMILY GKAMMICOLEPIDAE MYERS 147 



tails, and I am inclined to think that there may be more than one 

 family type among them. Neocyttus and Cyttomimus are certainly 

 greatly different from Zeus^ Zen, Zenopsis, and Cyttus. The gram- 

 micolepids are not particularly close to either of these groups in form 

 and a number of minor details, and their scales are so vastly different 

 that, for the present at least, I do not hesitate to give them family 

 recognition. The final word as to their exact place must await a 

 much-needed systematic and osteological investigation of all the 

 zeomorph fishes. 



The Grammicolepidae may, then, be defined as Zeomorphi (see 

 Regan, 1910) in which (1) the scales are vertically linear in form, (2) 

 the mouth is small and nearly vertical, (3) the maxillary is ex- 

 tremely short, (4) the anterior trunk muscles just reach the posterior 

 edge of the frontals, (5) the occipital crest is thin, (6) the gills are 

 3I/2? with no slit behind the last, (7) the branchiostegals are 7 in 

 number, (8) the caudal fin is composed of 13 branched rays with one 

 main and several supplementary unbranched rays both above and 

 below, and (9) the pelvic fins are I, 6. 



Gill arches thin, with one thin double row of hemibranchs. The 

 interior, or concave side, of each arch is smooth. Both the anterior 

 and posterior faces of each arch except the last possess a series of low 

 cross ridges, horizontal, or rather perpendicular, to the main line of 

 the arch. These short ridges are studded with spines. On the 

 posterior side of the inner (concave) ramus of the first arch there is 

 a row of small papilliform projections that might be construed as gill 

 rakers. At the upper end of each arch, where it curves around for- 

 ward, the hemibranchs leave the arch proper and run up on the wall 

 of the gill chamber. The gill structure in the two genera is identical, 

 but it cannot be properly seen without excising a complete arch from 

 a specimen. There is no slit behind the last gill. 



Pseudobranchiae of large size are present at the upper end of the 

 outer wall of the gill chamber. In some specimens the filaments are 

 entwined with those of the first gill arch, but they may be separated 

 by a little manipulation. I believe that either this or injury in prob- 

 ing accounts for Barnard's statement that Xenolepidichthys lacks 

 pseudobranchiae. j^ll four examples of this genus before me have 

 them. 



Branchiostegal rays 7 in number, the first three attached to the 

 anterior limb and the last four attached to the posterior limb of 

 the ceratohyal, as in Zeus. Poey, in speaking of Gramrtiicolepis, 

 says, "no he podido descubrir mas que cuatro radios branquiostegos, 

 sin poder asegurar que no haya mayor numero." Evidently he 

 thought there might be more than four; his skeleton of the type 



