S4 DISCOVERY REPORTS 



(1953) idea of putting the deep-sea salmonoids into a single family, Argentinidae. (At present they 

 are grouped into the families Argentinidae, Bathylagidae, Microstomidae, Xenophthalmichthyidae, 

 Dolichopterygidae, Winteriidae, Macropinnidae and Opisthoproctidae.) While Hubbs may be want- 

 ing to make too great a leap in the other direction, there would seem to be good reason for regrouping 

 of these fishes to give better expression of their relationships. Perhaps the suborder Salmonoidea 

 might be split into the divisions Salmoniformes and Argentiniformes, the latter containing the deep- 

 sea forms. The last four families, in parentheses above, could be united into one, Opisthoproctidae, 

 following Hubbs's first suggestion in his paper, while Xenophthalmichthys, in view of Bertelsen's 

 (1958) work, could be included in the Microstomidae. 



Order Iniomi 



When defining this order, Berg (1940) wrote that the swimbladder, if present, has a pneumatic duct. 

 The only iniomous fishes with a swimbladder are the Myctophidae and Neoscopelus (see also Marshall, 

 1955) and these have a closed swimbladder. There is nothing resembling a pneumatic duct in the 

 adult, although the ' oval ' in the myctophids may perhaps arise from the posterior part of the larval 

 connection between the gut and the swimbladder. Like the stomiatoids, the iniomous fishes were 

 most probably derived from an early type of teleost fish with an open swimbladder, and as Tracy 

 (191 1) suggested, the oval may have been evolved by a progressive reduction of the pneumatic duct. 

 At all events, it would be interesting to study the development of the 'oval' in a myctophid. 



The swimbladder of iniomous fishes is very unlike that of the stomiatoids and salmonoids. It is 

 physoclistous 1 with three to five retia mirabilia which originate at the anterior end of the bladder. 

 As might be expected, the retia draw their blood from vessels that arise in front of the swimbladder. 

 Each rete usually supplies a corresponding lobe of the gas-gland. 



Suborder Myctophoidea 



The myctophid swimbladder is definitely euphysoclistous, having at the anterior end of the sac 

 an ' oval ' type of resorbent surface. There are always three unipolar retia mirabilia supplying the 

 gas-gland (see Text-fig. 30 c). These individual features provide more evidence for regarding the 

 myctophids as a compact monophyletic group, but the ancestor of the myctophids can hardly have 

 been 'derived from Isospondylous stock very like the Gonostomatidae ' as Fraser-Brunner (1949) 

 supposed. The differences between the stomiatoid and myctophid swimbladders are very consider- 

 able as Table 1 shows. 



Table 1 . Stomiatoid and myctophid swimbladders 



In view of this, it is unlikely that the specialized swimbladder of a stomiatoid could have been 

 transformed into that of a myctophid. As already suggested, the Myctophidae were probably derived 

 from an early generalized form of isospondylous fish. The common origin of the myctophid and 

 stomiatoid lines of evolution would thus be further back in the early history of the teleosts. 



1 I was unable to determine whether Neoscopelus was a euphysoclist or paraphysoclist. 



