198 



ONTOGENY AND SYSTEMATICS OF FISHES-AHLSTROM SYMPOSIUM 



condition. The "photichthyids" and stomiids have specialized 

 Type Gamma photophores, although it is not known that all 

 genera in these groups have Type Gamma photophores; this is 

 a problem for further investigation. Manducus and Diplophos 

 retain Type Beta photophores and all of these fishes apparently 

 retain "white" photophore development of one kind or another. 

 These two characters are only useful at the ordinal level as 

 synapomorphies. Again, further research on "white" photo- 

 phore formation is needed since there appears to be more than 

 one form of this development. 



The monotypic Triplophos may or may not be related to 

 Manducus and/or Diplophos. Triplophos has a variety of derived 

 features not shared by Manducus or Diplophos. However, this 

 tells us nothing about its possible relationships with these gen- 

 era. Triplophos has four proximal pectoral-fin radials but with 

 some reduction in radial IV, Type Beta photophores, and prob- 

 ably "white" photophore development, the last two characters 

 synapomorphous only at the ordinal level. Four pectoral-fin 

 radials are not a synapomorphy for stomiiforms at any level 

 since the feature is found in most teleost outgroups. Triplophos 

 appears to be a primitive stomiiform with certain autapomorph- 

 ic features associated with an elongate body. Its relationships 

 are uncertain and there may be indications in the head and 

 pectoral girdle anatomy of a relationship with certain photich- 

 thyid genera. The problem needs much study. 



Cyclothone retains Type Beta photophores and "white" pho- 

 tophore development but has its own specialized features such 

 as only one pectoral-fin radial. It has a modified head and jaws, 

 which resemble and are, in our opinion, synapomorphous with 

 those of Gonostoma. The single pectoral-fin radial might be 

 thought of as a terminal stage in a transition series from Man- 

 ducus (with four pectoral-fin radials) to Diplophos to some mem- 

 bers of the "Photichthyidae" and then to Cyclothone. However, 

 Cyclothone does not have specialized Type Gamma photo- 

 phores of the "photichthyid" genera. The phylogenetic rela- 

 tionships of Cyclothone may not be certain as yet, but in many 

 respects it bears a resemblance to the three gonostomatid genera 

 and we favor its placement with these genera. See also discussion 

 above. 



Although we have perhaps resolved the differences between 

 Ahlstrom (1974) and Weitzman (1974), we have not achieved 

 a useful phylogeny of most stomiiform groups. Rather, we have 

 attempted to outline certain suggested hypotheses of relation- 

 ships to be investigated in the future with additional data. Adult 

 morphological data of the kind used by Weitzman to define and 

 relate the stemoptychid genera are available in abundance and 

 may be useful for other stomiiform groups. A closer look at 

 growth stages with the specific purpose of looking for possible 

 developmental specializations and terminal additions to char- 

 acters found in outgroups should greatly aid in delineating re- 

 lationships among the stomiiform genera. However, problems 

 associated with a high percentage of homoplasy can be expected 

 for some groups. The answers to problems of stomiiform in- 

 terrelationships will not come easily. 



Consideration of certain features is in order. For example, 

 larvae of Diplophos superficially resemble those of Chauliodus 

 with their prolonged development to a large larval size and great 

 elongation with bodies that are circular in cross section. Are 

 these convergent larval specializations or primitive stomiiform 

 features found only in certain stomiiform genera? The ventral 

 pigmentation on the body of developing Diplophos resembles 



that of developing paralepidids and myctophoids. Is this a prim- 

 itive stomiiform feature of Diplophos shared with certain sto- 

 miiform outgroups or a gross convergence of pigment patterns? 



Woodsia and Ichthyococcus share with certain stomiid genera 

 (for example, Eustomias) such developmental features as elon- 

 gate pectoral-fin rays, trailing guts, pigmentation patterns, and 

 bodies with a circular cross section. Some, if not all, of these 

 may be shared larval specializations. But again, independent 

 appearance of these characters indicated by a high degree of 

 homoplasy may be a vexing problem. Larvae of other genera 

 such as Vinciguerria. Pollichthys, and Cyclothone have body 

 shapes and certain other features that closely, but presumably 

 superficially, resemble those of clupeoid larvae. Detailed com- 

 parisons of these similarities may possibly distinguish between 

 homology and convergence among these taxa. 



In summary, a future phylogenetic analysis based on much 

 additional data may clear up many of the problems of stomi- 

 iform generic relationships. However, at present we are left with 

 numerous phylogenetic problems and assignment of certain gen- 

 era to family-level groups at this time would be misleading. The 

 above analysis retains Weitzman's Stemoptychidae. It restricts 

 the Gonostomatidae to the genera Bonapartia. Margrethia, and 

 Gonostoma. and we recommend the inclusion of Cyclothone. 



The other groups of non-stomiid stomiiforms remain unclear 

 as to family relationships. We agree with Fink and Weitzman 

 (1982) that Manducus and Diplophos are primitive stomiiforms, 

 but we cannot provide a stable classification for Manducus. 

 Diplophos. and Triplophos. Manducus and Diplophos might seem 

 to be sister taxa because of their similarity of appearance. How- 

 ever, they share no known specialized character or characters 

 that would unite them as a stomiiform subgroup except the 

 absence of an adipose fin and possibly a short neural spine on 

 the preural centrum. Currently all their other shared characters 

 seem primitive for stomiiforms. Further analysis of this situa- 

 tion is needed. 



Triplophos is again very much like a primitive stomiiform in 

 its head especially, but it has a number of specialized stomiiform 

 features as listed by Grey (1964:106) and may show some re- 

 lationship to some of the "photichthyid" genera. 



That the genera classified in the "Photichthyidae" by Weitz- 

 man (1974) form some kind of related group seems reasonable. 

 However, relationships among these genera are not known. That 

 these "photichthyid" genera are related to Diplophos is possible, 

 and that the stomiids are related to the "photichthyids" is, in 

 our view, very probable. The larval specializations of Woodsia 

 and Ichthyococcus noted above, may be important here because 

 they may be synapomorphies relating these genera to the sto- 

 miids. 



Until the developmental and adult morphological features of 

 many stomiiform genera are analyzed in detail, certain aspects 

 of their developmental stages outlined, and detailed outgroup 

 analysis performed on all putatively useful characters, we can 

 make no certain predictions about relationships and classifi- 

 cation. 



(W.J.R.) National Marine Fisheries Service, Southeast 

 Fisheries Center, 75 Virginia Beach Drive, Miami, 

 Florida 33149; (S.H.W.) Division of Fishes, National 

 Museum of Natural History, Smithsonian Institution, 

 Washington, D.C. 20560. 



