AHLSTROM ET AL.: GONOSTOMATIDAE, STERNOPTYCHIDAE 



185 



Table 47. Summary of Diagnostic Characters for Eggs of Certain Stomiiform Fishes. 



Ahlstrom (1974:672) favored recognition of one family for 

 those stomiiforms with elongate gill rakers in adults. According 

 to the rules of priority this would be the Stemoptychidae. Weitz- 

 man (1974:338) recognized three families, Gonostomalidae, 

 Photichthyidae, and Stemoptychidae, for the same stomiiforms, 

 the last family including the "maurolicin" genera formerly as- 

 signed to the Gonostomalidae and the deep-bodied stemop- 

 tychids traditionally assigned to the family. In a phylogenetic 

 or cladislic analysis this elongate gill raker bearing "group," if 

 recognized as a single family, is paraphyletic if one considers 

 certain of its subgroups as equivalent or higher taxonomic cat- 

 egories. For example, recognition of Ahlstrom's Stemoptychi- 

 dae, which would include the Stomiidae, a monophyletic group 

 with its members having a median barbel attached to the ventral 

 surface of the head in association with the hyoid bone and 

 lacking elongate gill rakers in adults, is incompatible with a 

 phylogenetic classification based on nested monophyletic groups, 

 since the Stomiidae is the sister group of another group within 

 Ahlstrom's Stemoptychidae. Furthermore, the character used 

 here to "define" the paraphyletic Stemoptychidae, the presence 

 of elongate gill rakers in adults, is excellent for use in a key for 

 identification purposes, but cannot be used as a synapomorphy 

 relating these fishes because it is primitive for stomiiforms. 

 Ahlstrom's Stemoptychidae is undefinable in a phylogenetic 

 analysis based on the information at hand. A resolution of the 

 use of familial and subordinal names in stomiiform fishes must 

 await completion of ongoing phylogenetic studies of these fishes. 

 Because these studies are incomplete, it is difficult to make 

 recommendations for names of certain stomiiform subgroups. 

 Among the stomiiforms with elongate gill rakers in adults, the 

 "family" problem is more complex than that recognized by 

 Ahlstrom (1974) or Weitzman (1974). We here recognize two 

 family names but these apply to only some of the 24 genera 

 listed below. We recognize the Stemoptychidae of Weitzman 

 (1974) and the Gonostomatidae in a new and restricted sense. 

 See discussion below. 



The stomiiforms discussed here include the following 24 gen- 

 era, listed alphabetically, which have been variously recognized 

 as belonging to the families Gonostomatidae, Stemoptychidae, 

 Maurolicidae, and Photichthyidae: 



Araiophos Grey (two species), Argyripnus Gilbert and Cramer 

 (four, possibly a few more), Argyropelecus Cocco (about sev- 

 en), Bonapartia Goode and Bean (one). Cyclolhone Goode 

 and Bean (twelve). Danaphos Bruun (one, possibly two), Dip- 

 lophos GUnther (two), Gonostoma Rafinesque (six), Ichthyo- 

 coccus Bonaparte (three), Manducus Goode and Bean (two),' 

 Margrethia Jespersen and Tuning (one, possibly two), Mau- 

 rolicus Cocco (one, possibly two). Photichthys Hutton (one), 

 Pollichthys Grey (one), Polyipnus Giinther (about sixteen), 

 Polymetme McCulloch (one, possibly four), Sonoda Grey (two), 

 Sternoptyx Hermann (two or three), Thorophos Bruun (two, 

 including Neophos Myers), Triplophos Brauer (one), Valen- 

 ciennellus Jordan and Evermann (one, possibly two), Vinci- 

 guerria Jordan and Evermann (five), Woodsia Grey (one), 

 and Yarella Goode and Bean (one). 



' Grey (1964:88) recognized Manducus Goode and Bean, 1896 as a 

 junior synonym of Diptophos GxmXheT, 1873 because, as she stated ". . . 

 the differences appear to be of a specific rather than a generic nature 

 . . ." This was in the context of the kinds of differences Grey noted 

 separating other species of "gonostomatids." She did recognize both as 

 subgenera ot Diphphos. We recognize both as genera. The species were 

 most recently reviewed by Mukhacheva (1978) who recognized four 

 species, D. maderensis (Johnson), D. rebamsi Krefft and Farm, D. greyae 

 R. K. Johnson, and D. taenia Giinther. We have examined all four 

 species and find that D. taenia and D. rebamsi have the cartilages of 

 the two medial proximal pectoral radials, radials III and IV in the 

 terminology of Fink and Weitzman (1982:66), fused while retaining two 

 bony elements separate as reported for D. taenia by Fink and Weitzman 

 (1982:65-67). Furthermore, one of the distal radials is out of line, not 

 in a single series in these two species. These characters are specialized 

 for these species. In Manducus maderensis and A/, greyae there are four 

 completely distinct proximal radials and the distal radials are all in a 

 simple straight series. Because the pectoral radial morphology in Diplo- 

 phos taenia and D. rebamsi may be an intermediate stage of a transition 

 series between radials such as are found in Manducus maderensis and 

 M. greyae and those in the "photichthyid" genera, we recognize Man- 

 ducus &% a genus and apparent sister group of the "photichthyid" genera 

 as well as the Stomiidae, nearly all of which have the radials 111 and IV 

 completely fused to one bone. A few stomiids have an apparent neo- 

 morph condition in which the third proximal radial is divided into two 

 radials, giving a total of four proximal radials. See also text discussion. 



