184 



ONTOGENY AND SYSTEMATICS OF FISHES -AHLSTROM SYMPOSIUM 



stomias also belongs to this group, and not with the other "me- 

 lanostomiids." This finding is not particularly radical, since 

 other authors have noted the close morphological resemblance 

 of that genus to the other three and indeed, it has been kept out 

 of the Malacosteidae mostly because the mouth floor is still 

 present, though thin, in members of the genus. The data are 

 insufficient to allow an unambiguous resolution of the interre- 

 lationships of these genera, but numerous characters support 

 the monophyly of the assemblage, including the suborbital pho- 

 tophore being ventral or posteroventral to the eye and the car- 

 tilage of the palatine arch being interrupted between the pos- 

 terior margin of the palatine and the rest of the arch. 



Idiacanthus has usually been placed in a family by itself as 

 was done, for example, by Beebe (1934), primarily on the basis 

 of the specialized stalked-eyed larval stages and the degree of 

 sexual dimorphism. Beebe recognized that the genus was "closely 

 related to the Melanostomiatidae," as did Gibbs (1964b). Nei- 

 ther author suggested more precise relationships, and Beebe and 

 Crane (1939) showed Idiacanthus in a large multichotomy in 

 their figure of "relationships." Regan and Trewavas ( 1 930) con- 

 sidered Idiacanthus to belong with Melanostomias, Echiosto- 

 ma. and Photonectes. but did not say precisely where. My data 

 support placement of the genus as sister group to Tactostoma, 

 a genus described in 1 939. These two are then related to a group 

 of genera as shown in Fig. 97. Note that Melanostomias and 

 Echiostoma are excluded, being the sister group of the entire 

 assemblage. I am confident of the placement of Idiacanthus and 

 Tactostoma together, based on an array of characters, including 

 reduction of the basihyal to a thin, cylindrical element, origin 

 of the dorsal section of the medial division of the adductor 

 mandibulae muscle anterior to the insertion of the levator arcus 

 palatini muscle, and an extremely elongate body. But I am not 

 particularly confident in the placement of these two genera with 

 the others, even though the data appear impressive at first glance. 

 This lack of confidence is attributable to the fact that most of 

 those characters change at least three times in the entire tree, 

 leaving but one, lack of a posttemporal bone, as the only un- 

 reversed character supporting the hypothesis. 



Another possibility is that Idiacanthus and Tactostoma are 

 the sister group of Melanostomias and Echiostoma. as suggested 

 in part by Regan and Trewavas (1930), apparently based on the 

 close morphological resemblance of Idiacanthus with the latter 

 two genera. Such a hypothesis would require some additional 

 reversals or independent losses, but as just noted, most of these 



characters change several times even in the most parsimonious 

 tree. This part of the total phylogeny deserves more critical 

 examination, and it is hoped that larval specializations will be 

 found which will be found which will cause one hypothesis to 

 be clearly preferred over the other. 



Regarding classification of the stomiiform fishes, it appears 

 that most of the traditional groups will cease to be recognized, 

 a move that was initiated by Weitzman (1974). A period of flux 

 should be expected until his curtent work is completed, but such 

 temporary instability is the current state of teleostean classifi- 

 cation at all levels, as phylogenetic methodology is applied with 

 increasing frequency. One might expect, however, that classi- 

 fication within the Stomiiformes will be stable sooner than that 

 in many other groups, because phylogenetic methods already 

 have been applied to it for several years. I will not present a 

 classification here, but I do provide such for the Stomiidae in 

 my revision of the group (Fink, in prep). 



In summary, there is still much to be done in unravelling the 

 phylogenetic history of the main lineages of stomiiform fishes. 

 1 have outlined above areas where our knowledge is either in- 

 complete or poorly developed, and these should be the areas 

 where workers now concentrate their attention— to establish 

 monophyletic groups among the "primitive" stomiiforms and 

 to critically reexamine some of the hypotheses I have produced 

 within the barbelled stomiiforms. Some of this work is under- 

 way, using adult and sub-adult specimens, but the usefulness of 

 larvae is as yet unknown. The data presented in Ahlstrom's 

 (1974) work on patterns of metamorphosis in "gonostomatid" 

 fishes corroborate, when analyzed by phylogenetic methods, the 

 placement by Weitzman (1974) of many of those genera in an 

 expanded Stemoptychidae. An example of this is the presence 

 of photophores in clusters with common bases in those fishes 

 recognized by Weitzman as stemoptychids. Kawaguchi and 

 Moser (this volume) present the most comprehensive infor- 

 mation to date of stomiid larvae. Their data indicate that there 

 should be a plethora of characters for phylogenetic analysis and 

 that study of larvae should indeed prove useful in testing hy- 

 potheses of stomiid relationships. However, even a cursory ex- 

 amination of their data indicates that, as with characters in 

 adults, there appears to be a high degree of homoplasy. This is 

 an interesting phenomenon deserving further study. 



Museum of Zoology, University of Michigan, Ann Arbor, 

 Michigan 48109. 



Families Gonostomatidae, Stemoptychidae, and Associated Stomiiform Groups: 



Development and Relationships 



E. H. Ahlstrom, W. J. Richards and S. H. Weitzman 



A summary of known information about the larvae and re- formation, both published and unpublished, gleaned from early 

 lationships of the stomiiforms with elongate gill rakers in life history stages and from adults. We also append some ten- 

 adults was published by Ahlstrom (1974). The present paper is tative new hypotheses of relationships within this "group" of 

 an addendum to that contribution and includes additional in- stomiiforms. 



