ATLANT. DEEP-SEA EXPED. 1910. VOL. III]. 



MURAENOID LARVAE. 



31 



There is a row of mucous pores along the upper 

 contour of the head, another row from the anterior nostril 

 backwards along the upper jaw as far as the anterior 

 margin of the eye, and a third row along the lower edge 

 of the lower jaw. 



On all the interspinous elements of the vertical fins, 

 there is a development of true rays. 



The retrogressive metamorphosis may be traced still 

 more distinctly in no. 5, in which the point of the snout 



Fig. 30. Head of L. stylurus, no. 5. 10 /i 



is still more rounded off and protrudes farther forwards 

 beyond the point of the lower jaw than in no. 4; the 

 snout is relatively longer, and the dorsal fin is inserted 

 nearer the head. The most essential difference is the 

 absence in no. 5 of the small embryonic pectoral fin 

 present in nos. 1 — 4. 



The metamorphosis of no. 6 is far advanced, as shown 

 by the shape of the body, which is not so much compressed 

 as in no. 5, but rather oval in cross-section. The height 



w- a a a 



Fig. 31. Head of L. stylurus, no. 6. 10 /i 



is relatively small (see table), the head has become a 

 little longer and the rays more powerful. 



Like no. 5, this larva has no pectoral fin and thus it 

 is highly probable that the parent form is an eel without 

 any pectoral fin. 



The following genera distinguish themselves by the 

 absence of pectoral fins: Heteroconger, Nettastoma, 

 Saurenchelys, Chlopsis, Todarus, and the genera of the 

 family Muraenidae. It is highly improbable that the larvae 

 in question belong to any of the last-mentioned family 

 which have very narrow branchial apertures and other 

 characters differing from those found in our larvae. Neither 

 do they belong to the genus Heteroconger, the species of 

 which are characterized by a very short snout and by 

 the lower jaw projecting beyond the upper. Among the 



remaining four allied genera, the larval forms of the fol- 

 lowing species are known: Nettastoma melanurum (larva 

 named Hyoprorus messinensis or Leptocephalus lon- 

 glrostris), Saurenchelys cancrlvora (larva named L. oxy- 

 rhynchus) and Chlopsis bicolor the larva of which a short 

 time ago is identified by Schmidt (31). 



The larval as well as the adult forms of Nettastoma 

 and Saurenchelys have a much longer snout than those 

 in question, while the larval Chlopsis is of the high, leaf- 

 shaped type, described by Stromman (33) under the name 

 of L. hyoproroides. The adult Chlopsis bicolor has a 

 much shorter snout, according to Supino (34) than our 

 larvae. The remaining genus Todarus I only know from 

 Facciola's and Supino's (36) descriptions; unfortunately 

 they do not give the number of segments, but by com- 

 paring Supino's drawing of the head of Todarus with the 

 head of our most advanced larva I was struck by the 

 likenesses, and it seems to be justified to draw attention 

 to the possibility of the connection between our larvae 

 and this species. 



18. Leptocephalus Saurenchelydis cancrivorae. 



(L. oxyrhynchus Bell.). 



I have referred a larva taken at Stat. 45 in 50 metres 

 to the species L. oxyrhynchus Bellotti because the differ- 

 ences detectable between our larva and Bellotti's may 

 easily be understood as the effects of development, our larva 

 being farther advanced than that of Bellotti. The larva 

 pictured and described by Bellotti (3) is relatively higher 

 on the anterior part of the body, anus is placed further 

 backwards, pectoral-fins are present, the snout is more 

 pointed and the pigment on the brain is lacking. By 

 studying the development of the related species L. stylurus 

 it will be seen, that these differences exist between the 

 least developed and the most developed larvae of this 

 species and consequently also may be regarded as caused 

 by the development at the larva in question. 



On the other hand the number of segments is almost 

 the same in our larva and in Bellotti's (about 249 and 

 240 respectively). 



According to Grassi (17) and Schmidt (31) the species 

 L. oxyrhynchus is the larval form of Saurenchelys can- 

 crivora and not, as supposed by Bellotti (3) of Ophich- 

 thys serpens, which according to his own determination 

 has about 208 vertebrae. 



To the description of Bellotti may be added the 

 following. 



As shown by fig. 4 on pi. V, this larva has a relatively 

 low compressed body, an elongated head and a pointed 

 tail. The anal opening is placed far forward not only in 

 our advanced specimen, but also in the specimen pictured 



