52 DISCOVERY REPORTS 



group was later somewhat amplified (Regan and Trewavas, 1929). The stomiatoids were regarded as 

 being most closely related to the clupeoids, but differing from them in possessing photophores. 

 Certain alepocephalids and searsids, which are clupeoids, also have light organs, but the individual 

 feature of the stomiatoid photophore pattern is the presence of one or two series of lights along each 

 side of the mid-ventral line and one or more photophores associated with each eye (Brauer, 1908; 

 Marshall, 1954). 



It is evident that much of Regan's appreciation of the stomiatoids could not be put into words. 1 

 At all events his definition of them has proved to be inadequate. But when the swimbladder characters 

 are considered together with those Regan used, the stomiatoids are revealed as a ' natural ' mono- 

 phyletic group. The implications of this will best be discussed at later stages in this systematic section. 



Turning now to the arrangement of the stomiatoids into families, study of the swimbladder 

 provides no new insight. It will be remembered that the Chauliodontidae, Melanostomiatidae, 

 Malacosteidae and Idiacanthidae have no swimbladder, while such features as can be seen in the 

 rudiments of this organ in certain of the Stomiatidae are merely typical of the suborder. Even in the 

 Gonostomatidae, Sternoptychidae and Astronesthidae, which have swimbladders, the knowledge 

 gained has little bearing on their relationships. This may best be shown by the following illustration : 



Considering the first two families, Hubbs (1953) has cogently urged that they should again be 

 referred to a single family, Sternoptychidae, with subfamilies Gonostomatinae, Maurolicinae and 

 Sternoptychinae. Hubbs sees these subfamilies (in the order given) as forming a series ranging from 

 forms with generalized to those with specialized body-shapes. Comparison of the swimbladder structure 

 in the three groups also shows that there are no clear-cut differences between them, but if swim- 

 bladder characters alone were considered, the Astronesthidae could also be put with them (and into 

 the same family). However, the other characters of the Astronesthidae indicate that they are best 

 regarded as a separate family and that they are closer to the Chauliodontidae and Melanostomiatidae 

 than to the gonostomatids and sternoptychids (Regan and Trewavas, 1929). 



At the generic level, however, swimbladder characters can be of use in definitions, but not always. 

 This may best be shown by short descriptions under the genera that were studied : 



Family Gonostomatidae 



Gonostoma. The differences between the swimbladders of G. denudatum and G. elongatum are so great that no 

 common features can be found to typify the genus. 



Cyclothone. The larval swimbladder (in which the gas-gland is found anteriorly), regresses and becomes 

 invested with fat in the adult. 



Pollichthys. An oval-shaped resorbent area on the floor of the swimbladder is surrounded by a horseshoe-shaped, 

 single-lobed gas-gland. The bladder is ellipsoidal in form. 



Vinciguerria. The resorbent area lies between the rete mirabile and a boomerang-shaped single-lobed gas-gland, 

 which extends over the floor and side walls of the median part of the swimbladder. 



Photichthys. The swimbladder of this gonostomatid is readily distinguished from those of other genera by its 

 elongated tubular form. 



Maurolicus. The gas-gland consists of two pairs of lobes, a smaller pair near the rete mirabile and a larger pair 

 on the floor of the mid-region of the bladder. The resorbent area lies between these lobes. 



Bonapartia. There is a single lobe to the gas-gland, immediately in front of which comes the resorbent area, 

 which invests the floor of the swimbladder. 



Ichthyococcus. The gas-gland has three lobes. 



Family Sternoptychidae 

 Sternoptyx. The gas-gland is three-lobed, while the resorbent area lies between and beyond these lobes. 



1 Pantin (1954) aptly calls such recognition 'aesthetic', as opposed to the analytical deductive methods that the systematist 

 must eventually use. 



