310 



SCANDINAVIAN FISHES. 



iind are more or less helplessly driven ashore ])y wind 

 and tide. We have already (juoted from GCxtiier some 

 (if the most important, general cliai'acteristies -which 

 belong to the deep-sea fishes. One of these character- 

 isties, the loose structure of tlie body, both in the 

 osseous and mnscuilar texture, api^ears in a high degree 

 in the Tracliypteroid family. The larger forms at least 

 — some of tlie Trachypteroids attain a length of at 

 least 6 metres — are so fragile that even with the 

 greatest caution it is scarcely possible to di'aw a spe- 

 cimen out of the water witiiout damaging some part 

 of it. Still less can we expect that a specimen cast 

 ashore b}- the storm and exposed to the action of the 

 waves, may be found in a perfect condition. Further- 

 more, the fins and their appendages undergo changes 

 of growth so great that it is b-\- no means always pos- 

 sible to decide with certaint) whether the loss of a 

 character derived from these parts of the body is the 

 result of some injurv or merelv of the natural develop- 

 ment. The s})ecific distinctum within this family is, 



Fig. 7l>. Young 'rraclii/jitcriif:, 3 inm. in leiigtli. After Emei:Y. 



therefore, far from certain, and the majority of the 

 species estal)lished by formei" Avriters up to the time of 

 Valenciennes" and recognised even in Gunther's Ca- 

 talof/KC, have been rejected or proj^osed for rejection in 

 more recent times. The history of the development has 

 shown a variability scarcely conceivable before, in the 

 supposed specific characters. The pioneer of tliis reform 

 is the Italian Emerv'', who enjoyed the most favourable 

 opportunities of studjing the young of these fishes, as 

 they ajipear in the surface-regions of the Mediterranean. 

 The first stage in the development of a Trachy- 

 ptcnis described to us by Emkky (fig. 7(i), is a larva 3 

 mm. in length. At this period it is perhaps most like 

 the larva of a Coitus or a Lophius. The head is large, 

 and tlie long caudal part sharply contrasted to the much 

 deeper abdominal region. The whole body is trans- 



]jarent. The embryonic vertical tin runs without a 

 break along the margin of the liodv, from the head 

 along the l)ack and round the tail to the end of the 

 abdominal |)art. At the occiput this fin contains the 

 rudiments of three true rays, with this exception it is 

 without rays, and is su])ported onlv bv the primary 

 rods (fibrilhe). The pectoral fins are comparatively 

 large, with a brachiate base, and edged with a rim of 

 the same structure as the vertical fin. The rudiments 

 of the ventral fins appear in the form of a process on 

 each side, projecting back-wards from the lower pos- 

 terior coi'iier of the abdominal part. 



In the next stage described by Emery (fig. 77), 

 the larva has attained a length of (i mm. It is now 

 tadpole-like, and the first three rays of the dorsal fin 

 have grown beyond the fin-membrane, thus reminding 

 us of the Batrachoids. Behind them three new rays 

 have begun to ap|)ear, and the embryonic vertical fin 



Fis. 



Young siiecinieu of Trachi/pterns, 6 niui. in lougtli. 

 After Emery. 



has lirown to sncli a height on the caudal uart that 

 the hind part of the larva is nearly as deep as the 

 forepart. The rudiments of the pectoral and ventral 

 fins have also grown. The ^vhole larva is still trans- 

 parent, but finely pinictated with small, brown collec- 

 tions of pigment, forming two transverse bands on the 

 caudal ]iart. 



In a larva 9 mm. long, which was brought home 

 by Professor Leche from Messina, the caudal part of 

 the liody itself is still terete and of an elongated, con- 

 ical shape. The occiput is raised as in Lophotcs. 

 The dorsal fin has begun to accjuire rudimentary rays 

 throughout the greater part of its length from the 

 occiput. The anterior dorsal rays have attained a length 



Gen. Trachypterus, p. 14, tnv. IX ter), tlic first ventral ray in Trachyptei'Hn rejiandus (iris/), a Mediterranean form, is divided at the tip 

 into four membranous filaments (see fig. 81, </). With these exceptions little is known of the structure of the ventral rays in the older spe- 

 cimens, as they easily and generally break or drop off. During youth, however, and even in specimens 32 cm. long, they arc unbranclied in 

 the s])ecimena we have examined of the Mediterranean sjiecics yrachi/pteriin ir/s. 



" Cuv., Yai.., Hist. Nat. Poiss.. vol. X, pp. 313 etc. 



' Atti Accnd. Line., ser. 3, Mem. Sc. Fis., vol. III. p. .'.'.HI, con tav.: Mittli. Zool. Stat. Neap., Bd. I (187',)), p. 581, tab. XVIII. 



