148 PROCEEDINGS OF THE NATIONAL MUSEUM vol. 84 



seems to have been in an incomplete condition, from Slmfeldt's re- 

 marks. Shufeldt did not mention the branchiostegals ; they were 

 probably entirely gone when he received the specimen. Barnard 

 gave four branchiostegals for Xenolepidichthys. I myself thought 

 this was correct until I dissected the muscle overlying the first three. 



The frontal and nasal bones are prominent and are covered with 

 rows of fine blunt spines. The preorbital is prominent and its outer 

 face is rough with the spine-studded fluting of what appear to be 

 mucous channels. Cheeks, opercle, subopercle, and, in Gramiru- 

 colepis, the interopercle, scaled. Vertical and lower limbs of pre- 

 opercle rough with fine granules. On the upper corners of the 

 cheeks (in the postorbital region) and opercles, along the predorsal 

 line, at the pectoral base, and on the caudal peduncle, the rough 

 linear scales approach the proportions of normal scales, 



A row of thin, bony bucklers, each bearing a main spine (and, 

 anteriorly at least, one or more smaller, supplementary spines) ex- 

 tends along each side of the entire base of the dorsal and anal fins. 



Eyes large, much greater than interorbital. Body deep and 

 strongly compressed. Caudal peduncle slender. Pectoral fins small. 

 Anal spines 2, separated by an interspace from the first soft ray. 

 Soft dorsal and anal rays unbranched. Greatest body depth at 

 origin of dorsal fin. 



Teeth small, acicular, weak, in a single series on each jaw. 



Besides a 43 mm specimen of Xenolepidichthys (see Smith, 1935), 

 which retains some postlarval characters, no larvae or postlarvae 

 of Grammicolepidae are known. The ^^Acronurus^^ larvae of the 

 Acanthuridae, with their vertically elongate scales (see Liitken, 1880, 

 pi. 5, figs. 4, 5), are likely to be mistaken for young gi-ammicolepids. 

 One young acanthurid, with a most remarkable type of scales, has 

 already been described as a grammicolepid (Smith, 1931, p. 146). 

 These young acanthurids may be distinguished from the Grammi- 

 colepidae both by their different mouth structure and by their 

 metallic "corselet" extending downward and forward from the 

 pectoral base. 



I experienced some difficulty at first in discovering valid char- 

 acters to distinguish the two recognizable forms of Grammicolepidae. 

 The external differences are mostly of a tj'pe unlike those that have 

 been used in related groups, and I present them here in the form 

 of a comparative table. 



