244 



ONTOGENY AND SYSTEMATICS OF FISHES-AHLSTROM SYMPOSIUM 



Notolychmis valdiviae. here considered a monotypic tribe, 

 could not be placed with certainty in either subfamily. Moser 

 and Ahlslrom ( 1 970: 1 38, 1 974:409) and Paxton ( 1 972:6 1 ) dis- 

 cussed the characters and problems of this enigmatic species. 

 With long jaws and the lack of a cleithral shelf both considered 

 plesiomorphies, the apomorphic number of Pre photophores 

 unknown, and the larval eyes variable and intermediate in shape, 

 future work is required to resolve this trichotomy. 



We recognize three tribes in the subfamily Lampanyctmae 

 (Fig. 125). The tribe Lampanyctini, with nine genera, is defined 

 by the presence of a row of moderately to strongly hooked teeth 

 in the posterior dentary; the only other genus with this feature 

 is the myctophine Diogenichthys. These nine genera are also the 

 only lampanyctines to lack a Dn orbital photophore, but we are 

 unsure if this is a derived state (see discussion of character 6 

 above). Moser and Ahlstrom (1972) and Ahlstrom et al. (1976: 

 148) placed five of these genera (Lampadena. Taaningichthys. 

 Bolinichthys. Lepidophanes, Ceratoscopelus) in the tribe Gym- 

 noscopelini, based primarily on larval photophore pattern. Pho- 

 tophores which appear in larvae of Lampanyctinae are essen- 

 tially the same ones which develop in myctophine larvae (Moser 

 et al., this volume) and, if they are adaptive as Moser (1981) 

 has suggested, it is likely that they have appeared in these typical 

 sites independently in a number of lineages. Moreover, these 

 photophores develop at the end of the larval period, if at all, in 

 Bolinichthys and no photophores develop in Taaningichthys 

 larvae. Likewise, the larval pigment characters do not support 

 the inclusion of these five genera in the Gymnoscopelini. 



In addition to the distribution of hooked dentary teeth and 

 Dn photophores, other features influenced our decision about 

 these five genera. The ischial ligament is medium or long in all 

 Lampanyctini except Taaningichthys (and some species of Dia- 

 phus). while the fifth circumorbital has a ridge or keel in all 

 gymnoscopelines (but is lacking in some species of Diaphus) and 

 no lampanyctines except Bolinichthys (thus the brackets around 

 character 59 in Fig. 125). Finally all of the gymnoscopeline 

 genera except Notoscopelus are restricted to the southern ocean 

 (Moser et al., this volume: Table 59), while the Lampanyctini 

 are found both north and south (except Stenobrachim) of the 

 equator. Placement of the five genera in the Lampanyctini re- 

 quires fewer character reversals and parallelisms. 



Within the Lampanyctini, the development of larval photo- 

 phores in addition to Br, (character 27) unites the five genera 

 discussed above. We recognize Dorsadena as a subgenus of 

 Lampadena until specimens other than the types are available 

 for osteological study and the larvae are discovered. We have 

 not found an apomorphic character that defines the line in- 

 cluding Stenobrachius. Triphoturus. Lampanyctus and Parvilux. 

 We are recognizing Parvilux on the basis of a weakly angled 

 SAO and larval shape and pigmentation. 



We consider the tribe Diaphini to be the sister group of the 

 Gymnoscopelini. The relationships among the three genera of 

 Diaphini are not clear. One of us (HGM) has re-examined the 

 specimens on which the larval features of Idiolychmis urolampus 

 were based (see Moser and Ahlstrom, 1974:405-406; Nafpak- 

 titis and Paxton, 1978), and now thinks they could represent 

 Lobianchia gemellari. with the larvae of Idiolychnus still un- 



known. Two characters shared by Lobianchia and Idiolychnus. 

 the presence of caudal organs and the absence of a luminous 

 patch above the pectoral fin, are considered plesiomorphic, while 

 the absence of a Vn and differences of photophore positions are 

 not clearly apomorphic. The most unequivocal derived state is 

 the presence of a wide pubic plate, indicating Lobianchia and 

 Diaphus are the sister group pair. 



Within the Gymnoscopelini the proposed generic relation- 

 ships are based almost entirely on characters of the photophores 

 and luminous tissue. No consistent osteological or larval fea- 

 tures define generic groupings. Southern ocean larvae require 

 more study. The larvae of Hintonia are unknown and not enough 

 species of Gymnoscopelus have been studied to ascertain if the 

 subgenus Nasolychnus can be defined by any larval characters. 

 The species of Notoscopelus should also be studied to find sup- 

 porting characters of the subgenus Parieophus. 



Within the subfamily Myctophinae (Fig. 126), we also rec- 

 ognize three tribes, the Electronini, Myctophini and Gonichthy- 

 ini. The Gonichthyini is clearly a derived lineage, with a num- 

 ber of osteological, photophore and larval characters 

 distinguishing the four genera from the rest of the subfamily. 

 We think the larval specializations of eyes and pectoral fins arose 

 after the split of the two generic pairs. 



Paxton (1972) was unable to find osteological characters to 

 clearly separate the remaining genera of the Myctophinae into 

 two lineages. We have utilized photophores to distinguish the 

 Myctophini from the Electronini, while recognizing there is a 

 mosaic of osteological and larval characters within these nine 

 genera. We have little question of the sister group relationship 

 of the generic pairs Krefftichthys—Protomyctophum. Mycto- 

 phum — Symbolophorus and Benthosema — Diogenichthys. 

 However two larval features, thin head and body and a bipartite 

 gut, are shared by Metelectrona and some species of Hygophum. 

 Since we think Hygophum is a monophyletic line, we consider 

 these shared larval features parallelisms that do not indicate 

 common ancestry. Paxton (1972) considered Metelectrona a 

 synonym of Electrona. The description of a second species of 

 Metelectrona (Hulley, 1981), coupled with its larval and pho- 

 tophore characters, convinced us to recognize the genus. 



Of the 59 derived characters utilized in our analysis, only 20 

 are restricted to members of the lineage they define, and eight 

 of these are autapomorphic at the generic level. The remaining 

 39 characters are not found in the apomorphic state in any 

 member of the opposite lineage from the defined branching 

 point, but are found in some members of other lineages within 

 the family. This presumed homoplasy of larval, photophore and 

 even osteological characters indicates that the proposed phy- 

 logeny was arrived at with some difliculty. Ten of our proposed 

 lineages are undefined by derived characters. We think that 

 future work will support our proposed phylogeny, although some 

 details may be modified, and that new, less plastic characters 

 and better definitions of polarity will help resolve the problems. 



(J. R. P.) The Al'stralian Museum, 6-8 College Street, 

 Sydney 2000, Australia; (H.G.M.) National Marine 

 Fisheries Service, Southwest Fisheries Center, P.O. Box 

 271, La Jolla, California 92038. 



