5 6 DISCOVERY REPORTS 



But such differences, together with those found by Regan, would hardly seem to require separate 

 ordinal status. Regan himself felt that the Xenoberyces were ' probably derived from the same stock 

 as the Berycomorphous fishes '. It would thus be preferable to regard the ' Xenoberyces ' as a sub- 

 order of the Berycomorphi. The other fishes of this order (Regan's Berycomorphi) would form the 

 other suborder. 



The order Berycomorphi (sensu lato) may be defined as physoclistous teleosts with the anterior 

 rays of the vertical fins spinous and with nineteen principal rays (seventeen of which are branched) in 

 the caudal fin (eighteen principal rays with sixteen branched in the Polymixiidae). Pelvic fins sub- 

 abdominal or thoracic, with or without a spine, and with three to thirteen soft rays. Mouth bordered 

 above by protractile premaxillae ; one or two supramaxillae. Orbitosphenoid present or absent. 



Suborder Berycoidea: Families Polymixiidae, Berycidae, Diretmidae, Trachichthyidae, Mono- 



CENTRIDAE, ANOMALOPIDAE, HOLOCENTRIDAE 



Orbitosphenoid present ; a subocular shelf usually developed, palate usually toothed one or two supra- 

 maxillae (Regan, 191 1). Swimbladder with retia mirabilia originating at the anterior end of the sac. 



Suborder Anoplogastroidea : Families Stephanoberycidae, Melamphaidae, Anoplogastridae 1 



No orbitosphenoid; no subocular shelf; palate toothless; a single supramaxilla, triangular in shape 

 (Regan, 191 1). Swimbladder with one or two retia mirabilia, usually posterior in position: an oval 

 at the posterior end of the sac. 



Order Percomorphi, Family Chiasmodontidae 



Like many other percomorph fishes, Chiasmodon niger has a euphysoclistous swimbladder with a 

 diaphragm separating an anterior gas-producing chamber from a posterior resorptive chamber 

 (Text-fig. 30J). Jones and Marshall (1953) and Fange (1953) have drawn attention to percormorphs 

 with this type of swimbladder. 



To conclude, there is the problem of what the structural pattern of the swimbladder may add to 

 present conceptions of the relationships between the major groups of deep-sea teleosts. 



The Isospondyli undoubtedly contain the most generalized forms of teleosts. Although the fishes 

 of each isospondylous suborder are specialized in varying degrees, they still retain certain characters 

 of the early primitive teleosts. While the stomiatoids have special features, such as light organs, that 

 are correlated with a bathypelagic mode of life, Regan (1923) looked on them as being quite similar 

 to the clupeoids. He found the skull of Photichthys to be very like Elops, which is perhaps the most 

 primitive of all living teleost genera. The extreme development of the gas-producing tissues of the 

 swimbladder (or the complete loss of this organ) is also an adaptation to a deep-sea existence, but the 

 structural pattern is quite unlike that in any other group, whether deep-sea or otherwise. 



It is unlikely, as already stated, that the Myctophidae, with a very different form of swimbladder, 

 were derived from a stomiatoid stock. In acquiring an oval, the myctophids have a remarkably 

 advanced type of swimbladder relative to their position in the evolutionary scale of teleosts. Above the 

 Iniomi, an oval is not encountered until the Anacanthini (cod-like fishes) are reached. However, an 

 oval would seem to be no more efficient a resorbent surface than the capillary system in stomiatoids. 

 Astronesthes is no more limited in its vertical migrations than the myctophids on which it feeds (see 

 p. 88). From the functional as well as the structural aspect, the conclusion repeated in the first 

 sentence of this paragraph seems to be justified. 



1 Grey (1955) has confirmed that Anoplogaster cornuta (Val) is the young of Caulolepis longidens Gill and also synonymizes 

 the latter under the former (in accordance with the principle of priority in the International Rules of Zoological Nomen- 

 clature). 





