396 REPORT OF COMMISSIONER OF FISH AND FISHERIES. [54] 



difference — the length of the gill-rakers, which, though small, is con- 

 stant, and holds good in all the known species. 



With a view to the discovery of a basis for generic subdivision, we 

 have especially compared the following species: Scicena (Scicenops) ocel- 

 lata, Scicena (Pseudoscicena) aquila, Scicena (Bola) diacantha, and Scicena 

 (Callaus) deliciosa. If these species could be satisfactorily arranged in 

 different genera, it would be comparatively easy to find characters on 

 which to detach the rather more aberrant types of Scicena (umbra), 

 Cheilotrema (saturna nwlfasciata), Ophioscion, and Johnius. 



The four species first mentioned agree in the position of the anal fin. 

 Its second spine is very weak in aquila and adnate to the first ray. It 

 is somewhat so in the others and it is not large in any. In Johnius 

 (dussumieri) it is also small, but in Scicena, Cheilotrema, and Ophioscion 

 it is considerably enlarged. 



The scales are smallest in aquila, largest in ocellata, but the difference 

 is not sharp enough to warrant generic division. In all four of the 

 species first mentioned the preorbital is flat and rather broad, broadest 

 i:i deliciosa (7 in head) and narrowest in aquila — lOi. In the other 

 forms it is generally still broader and more gibbous. 



The slits and pores about the snout are distinct in ocellata and deli- 

 ciosa, little marked in diacantha and nearly or quite obsolete in aquila. 

 In Johnius, Scicena, Cheilotrema, and Ophioscion these are more or less 

 distinct. 



In all the four species the mouth is of moderate size, slightly oblique, 

 with the lower jaw included, the maxillary reaching to opposite the poste- 

 rior border of the eye. The mouth is largest in ocellata, smallest in aquila. 

 In all the others (Ophioscion, &c.) the mouth is still smaller. The upper 

 teeth are nearly alike in all of these ; of the four mentioned they are 

 largest in diacantha, smallest in deliciosa. In some East Indian species 

 (referable to Bola?) these teeth are still larger, some of them almost 

 canine-like. 



The lower teeth are rather large, and chiefly uniserial in diacantha 

 and other species of Bola ; in two or three rows, the inner enlarged in 

 deliciosa and aquila ; in a broad band, some of the inner enlarged in ocel- 

 lata. In Johnius, Cheilotrema, Scicena, and most of the species of Ophio- 

 scion, the lower teeth are in a broad baud and equal. 



The preopercle is sharply serrate in youth, becoming entire with age 

 in ocellata. In aquila it is vaguely crenulatein youth, becoming finally 

 entire. In diacantha it remains more or less crenulate. In deliciosa the 

 preopercle is edged by fine flexible seme. In Ophioscion the preopercle 

 is always sharply serrate. In Scicena, Cheilotrema, and Johnius it is al- 

 ways entire or at least without bony serratures. 



Among the four species first mentioned, the gill-rakers are smallest 

 in diacantha (X + 7), when they are short and thick, the longest not 

 half the pupil. They are longest in deliciosa; when they are slender 

 (X + 12; as long as pupil. In aquila and ocellata they are X + 8 or 9, 



