198 Mr. G. A. Boulenger on the 



lias nothing distinctive ; the second is decidedly misleading, 

 since the heads of Miiripristis and Ilolocentruni, two of the 

 principal genera of Berycidai, show the nuiciferous cavities 

 even less developed than in an ordinary perch, whilst these 

 cavities are excessively larsj^e in the Percid genera Acerina 

 and Percarina. 1'he third character is evidently the leading 

 one ; bnt here again an incon.sist'.'ncy occurs, for, if we turn 

 back a few pages in the book, we Hud in the account of the 

 Perciformes a note to tlie effect that " A North-American 

 freshwater genus, Aphredodirus^ occupies a perfectly isolated 

 position in the system and is evidently the type of a distinct 

 family. It resembles the * Sun-fishes' [Percidai] of the same 

 country w'ith regard to the structure of the vertical fins, but 

 has the vent situated in front of the ventraU, which are com- 

 posed of more than five soft rays." And besides, a new 

 genus, MalacosarcuSj was shortly after, in the " Report on the 

 'Challenger' Deep-sea Fishes," added by the same author 

 to the Berycida;, in spite of its ventrals being described as 

 five-rayed. Giiuther has therefore failed to give us a defini- 

 tion by which his Btryciformes can be distinguished from 

 the Perciformes. Let us see if later authors have been more 

 successful. 



In his ' Memoir on the Families and Subfamilies of Fishes,' 

 which summarizes his views on classification in 1893, Gill 

 adopts a division of Acanthopterygii, named Berycoidea, 

 equivalent to IScombroiden, Percoidea, &c., which contains 

 six families : — Ste})hanoberycidge, Berycida^, Tracliichthyidie, 

 Monocentrida?, Holocentrida?, Anomalopida3. This division 

 was not defined in this paper, but was shortly after by 

 Jordan and Evermann, who have adopted the families of 

 Gill and added to them the Batliyclu|ieidaj, the Polymixiidae 

 (Berycidaj of Giiuther), and the ^lullidaj, the pertinence of 

 the two latter to this group bein<i-, however, regarded as 

 questionable. In their definition the only two distinctive 

 characters, both accompanied by restrictions which impair 

 their value for diagnostic purposes, are the following : — 

 " Ventral fins with 1 spine, usually 7 soft rays, the number 

 of soft rays varying from 5 to 10 ; air-bladder in some species 

 retaining its duct throughout life (a character verified only in 

 JJeryx).'''' The authors add that they regard the group as a 

 valid one, " allied to the Pereoidei and Seombruidei, but 

 characterized as a whole by the retention of the archaic cha- 

 racters of the jjersistent air- duct and the increased number of 

 ventral rays." The character of the persistent air-duct 

 between the swim-bladder and the intestinal canal, first 

 pointed out iiy Kner in Ilolocentrunij by Alcock in Balliy- 



