258 



ONTOGENY AND SYSTEMATICS OF FISHES-AHLSTROM SYMPOSIUM 



Table 70. 



Anatomical Differences of Early Stages Betwfen 

 AuLOPvs and Chlorophthalmus. 



Table 71. 



Distribution of Larval Character States among Four 

 Genera of the Ipnopidae. 



Autopus 



Chtorophlhatmits 

 (afler Rosen (1971) 

 and Sulak(1977)] 



D« 



Maxillary teeth 

 Vomerine teeth 



Basihyal 



Fulcral scale 

 Gut morphology 



Peritoneal pig- 

 ment sections 



Present 



Only two widely sepa- 

 rated at opposing 

 anterolateral comer 



Ovoidal with slightly 

 indented tip; teeth 

 absent 



Present 



Moderately elongated, 



straight; intestine 



slightly fat 

 Single; distinct at less 



than 3.5 mm SL 



Absent 



Transverse row of six 

 teeth divided into 

 two rows of three 

 each 



Triangular with similar 

 anterior indentation; 

 a transverse row of 

 six teeth divided into 

 two series 



Present(?) 



Short, compact with 

 slender stomach; in- 

 testine fat 



Single; distinct at more 

 than 5 mm SL 



japonicus is a probable indication of this evolutionary trend 

 (Okiyama, 1974b). Among recent congeners, A. damasi may be 

 the most generalized species in view of its smallest number of 

 vertebrae (20 + 16) similar to the known counts in the fossil 

 aulopids (Goody, 1969; Rosen and Patterson, 1969). Further- 

 more, this species is clearly separable from congeners by the 

 mode of direct association between the first haemal spine and 

 anal pterygiophores (Okiyama, 1979b). A look at the larvae of 

 A. damasi would be enlightening in clarifying the problem in 

 question. 



Myctophoidei: Neoscopelidae, 

 Myctophidae 



The two families of this suborder are readily discriminated 

 from the others by the greatest similarity index value based on 

 a suite of derived characters (1 and 2) not shared by any other 

 families. The smaller sizes at metamorphosis are also peculiar 

 to these families. These larval evidences offer strong support 

 for the views of Moser and Ahlstrom (1974) and Johnson (1982), 

 warranting a distinct subordinal ranking. My observation of the 

 vertebrae of Solivomer (see Table 57 in my Myctophiformes: 

 Development, this volume) also disclosed their closer linkage 

 than assumed by Johnson (1982). 



The similarity matrix in Table 69 would offer little support 

 for Rosen's scheme to transfer these families to a different order. 



Chlorophthalmoidei: Notosudidae, 



Scopelarchidae, Chlorophthalmidae, 



Ipnopidae 



The larval character states indexed in Table 7 1 are less prom- 

 ising in support of this familial assemblage, because only the 

 Notosudidae and Scopelarchidae share a single derived char- 

 acter state (narrow eye). It seems that this ambiguity is also 

 associated with the inadequate numbers of characters in ques- 

 tion. 



Although the admitted cohesiveness of larval characters of 

 Chlorophthalmidae may be altered by the discovery of larval 

 Bathysauropsis or Parasudis. larval characters support the tra- 

 ditional view that it is one of the basal stocks of this order, lying 



Bathytyphlops 

 Ipnops 



Bathymicrops 

 Bathypterois 



10 2 



10 10 4 



10 10 4 



10 15 



' Number of denved character slates. 



at a somewhat advanced place along a line different from the 

 Aulopidae. Trenchant characters in this connection such as the 

 dentition and the mode of anal fin support are shared with the 

 Ipnopidae. 



Members of the Notosudidae, the most cohesive family in 

 this suborder, have the greatest numbers of derived characters 

 of the group. Marshall ( 1 966a) and Bertelsen et al. ( 1976) stated 

 that it seems most closely related to Chlorophthalmidae. The 

 superficial resemblance of larval stages between this and the 

 Paralepididae was also suggested (Ahlstrom, 1972a). On the 

 other hand, the similarity matrix indicates its affinity with An- 

 otopteridae, along with Scopelarchidae. Of these associations, 

 the last grouping based on a single derived state in character 5 

 (narrow eye) appears less arguable. Other features such as the 

 maxillary teeth and the uncommon morphology of the corpus 

 cerebelli suggest the aberrant systematic status of this family. 



Since Table 68 provides few clues to discuss the confused 

 family limits of the Ipnopidae, the same coding of the character 

 states is applied to the four genera of this family (Table 71). 

 Except for the distinct larval status of Bathypterois. derived 

 characters shared among the remaining three genera do not 

 reveal the generic linkages suggested by Sulak (1977). By the 

 same reasoning as discussed before concerning the relationships 

 between Aulopidae and Chlorophthalmidae, the derived state 

 in character 1 (peritoneal pigment sections) shared by Ipnops 

 and Bathymicrops includes the different states of gut morphol- 

 ogy. It seems these genera form a loose but distinct assemblage 

 warranting family rank. Besides the shared dentition mentioned 

 before, the close fit of general larval morphology between Bathy- 

 typhlops and Chlorophthalmus may suggest their relationship. 



The diverse larval characters of Scopelarchidae were elabo- 

 rately enalyzed in the light of adult systematics (Johnson, 1 974b). 

 It is remarkable that this family has no phenetic similarity with 

 Alepisauridae in terms of catalogued characters. On the other 

 hand, two derived states in character 2 (anus location) and 5 

 (eye shape) shared with Evermannellidae give the greatest sim- 

 ilarity index value. Johnson (1982) suggested the independent 

 occurrence of the tubular eyes in adults of both families, but 

 traditional concepts of their close association should be reevalu- 

 ated using larval evidence. 



Synodontoidei: Bathysauridae, 

 Harpadonti[5ae, Synodontidae 



Accepted linkage between Synodontidae and Harpadontidae 

 is clearly substantiated by the larval characters, while familial 

 allocation of Saurida remains to be solved. Synodus lucioceps, 

 having the intermediate state of larval characteristics between 

 these families, may be important here. The relationships among 

 four genera are thus indistinct from the standpoint of the larvae, 

 but Saurida appears to be the most generalized. Possible phy- 



