FISHERY BULLETIN: VOL. 70, NO. 3 



Figure 7. — Developmental stage of Ceratoscopelus maderensis (Lowe). — A, 13.4-mm larva, RV Meteor Station 



72-39; B, 13.4-mm larva, dorsal view. 



ually during the larval period, which appears to 

 be rather short; most species transform within 

 the size range of 10-15 mm. Also, the charac- 

 teristic shape and basic pigment pattern of lar- 

 val Diaphus and Lobianchia are distinctively dif- 

 ferent from those of Gymnoscopelus. In shape 

 and pigmentation, the larvae of Gymnoscopelus 

 resemble most closely those of group A {Scope- 

 lopsis, Lampichthys, Notoscopelus) . 



Of the remaining six genera of the subfamily 

 Lampanyctinae, Notolychnus develops no photo- 

 phores during the larval period, and Lampanyc- 

 tus, Triphoturus, and Stenobrachkis develop only 

 the Br2 as larvae. Our developmental series of 

 Taaningichthys is incomplete; however, larvae 

 up to 18 mm long have no photophores. In our 

 largest larva (19.3 mm) , the PO5 are just begin- 

 ning to develop. The larvae of this genus are ap- 

 parently neustonic, and this may have an impor- 

 tant influence on photophore development. They 

 are most similar to larvae of Lampadena in pig- 

 ment pattern but are much more slender. The 

 small larvae of Taaningichthys minimus are un- 

 usual in having a conspicuous series of embedded 



melanophores above the vertebral column (Fig- 

 ure 13). The larvae of Hintonia are as yet un- 

 known. 



From our studies of the larvae of the subfam- 

 ily Lampanyctinae, the genera Scopelopsis, 

 Lampichthys, Notoscopelus, Gymnoscopelus, Ce- 

 ratoscopelus, Bolinichthys, Lepidophanes, Lam- 

 panyctodes, Lampadena, and Taaningichthys 

 emerge as a natural assemblage based on se- 

 quence of photophore development, morphology, 

 and pigmentation. The grouping of these genera 

 as a phylogenetic unit conflicts with the phylo- 

 genetic scheme proposed by Paxton (1968, 

 1972). His arrangement of the 17 genera of 

 the subfamily Lampanyctinae into four tribes, 

 based on a combination of adult osteology and 

 photophore pattern, is shown in Figure 14. We 

 agree that Notolychnus valdiviae has an ample 

 array of unique adult and larval characters to 

 warrant its placement in the monotypic tribe 

 Notolychnini. Likewise we concur with the 

 establishment of a separate tribe Diaphini for 

 Diaphus and Lobianchia, two genera with a dis- 



554 



