ROSENBLATT and BUTLER: THE RIBBONFISH GENUS DESMODEMA 



developed, with surface pores present. A juvenile 

 ofl04 mm SV has scales along the dorsal base, and 

 one of 131 mm SV lacks scales and has tubercles 

 and pores over the entire body. 



Desmodema polystictum does not agree with D. 

 lorum in the course of development of the tuber- 

 cles and pore system. None of our specimens has 

 scales. Instead tubercles are developed in a speci- 

 men of 36 mm SV, and tubercles and pores are 

 present in an individual of 42 mm SV. Walters 

 (1963) was unaware of the existence of the two 

 species of Desmodema and his figure 1 was un- 

 doubtedly based on a juvenile of D. lorum. 



In juveniles the first six dorsal rays are elon- 

 gated (broken in all our specimens). These rays, 

 which are borne on the pterygiophores before the 

 first neural spine, are lost, and in adults rep- 

 resented by a stiffening in the skin. The recurved, 

 fanglike lower jaw teeth first appear at a snout- 

 vent length of about 100 mm. 



Life history and behavior. — We lack data from 

 closing nets, and thus have no precise information 

 on depth of capture of our material. Fitch and 

 Lavenberg (1968) inferred that Desmodema 

 "polystictum" lives "500 to 1,000 feet beneath the 

 sea's surface" and Walters (1963) predicated his 

 discussion of energetics on the assumption that 

 Desmodema is mesopelagic. Harrison and Palmer 

 (1968) speculated that Desmodema, which they 

 described as "chocolate brown," might live deeper 

 than its silvery relatives. Actually Desmodema is 

 silvery and turns brown in preservative. 



The number of polka-dotted juveniles of D. 

 polystictum taken at or near the surface indicates 

 that they probably mainly occupy the euphotic 

 zone. The polka-dotted pattern would be maxi- 

 mally useful as protective coloration in the light- 

 dappled environment near the surface. However, 

 records (presumably juveniles) from stomachs of 

 Alepisaurus (Fourmanoir 1969) suggest a consid- 

 erable depth range. A number of juvenile D. lorum 

 have been taken from albacore, Thunnus 

 alalunga, stomachs, and others have been taken 

 by gear fished near the surface. We see no reason 

 to assume that the albacore had been feeding "far 

 beneath the surface" (Fitch 1964); however, Fitch 

 figured a metamorphosing juvenile of D. lorum 

 from an Alepisaurus taken on a longlineand listed 

 four other such specimens, again indicating a wide 

 depth range. Several of the metamorphosed 

 specimens of D. lorum were taken by open nets 

 fished to considerable depths. However, three of 



the largest specimens were taken in purse seines, 

 indicating depths of capture of no more than 100 

 m. We have three adult D. polystictum: two were 

 taken in nets towed in the upper 500 m, and one 

 was taken in a purse seine. 



Fitch's (1964) report on stomach contents pro- 

 vides equivocal evidence; Idiacanthus is a 

 mesopelagic vertical migrator, but Phronima 

 sedentaria occurs in the upper 300 m (Eric 

 Shulenberger, Scripps Institution of Oceanog- 

 raphy, pers. commun.1. There is thus no objective 

 evidence that either species of Desmodema lives 

 below 500 m (although the possibility is not 

 excluded). The species of Desmodema would seem 

 to be members of the deep epipelagic group as 

 defined by Parin (1968). 



Keeping in mind the sketchy nature of the 

 available data on depth distribution, the following 

 hypothetical scheme is suggested for both species. 

 The silvery young have a gas bladder. The large 

 fins and the deep head and rapidly tapering body 

 suggest that they are feeble swimmers. They are 

 probably epipelagic. The polka-dotted pre- 

 juveniles probably occupy the euphotic zone. The 

 tail is short and anguilliform propulsive waves 

 could involve almost the entire body. The very 

 elongate, fanlike pelvic fins and dorsal pennant 

 indicate that swimming is normally slow and 

 probably involves undulations of the dorsal fin, 

 rather than the body. 



With metamorphosis the dorsal pennant and 

 the pelvic fins are lost, the latter dropping off en- 

 tirely. The tail rapidly elongates at this time (see 

 Figure 5). The polka-dotted pattern is also lost, but 

 more gradually. The greatly elongated tail with 

 its associated dorsal rays would produce drag dur- 

 ing active swimming, but probably less so than in 

 Trachipterus, in which the posterior part of the 

 body is deeper. We propose that adult Desmodema 

 normally occupy the twilight zone of a few 

 hundred meters, where they hover, probably in a 

 head-up posture, maintaining position by undula- 

 tions of the dorsal fin. Rapid bursts of anguilliform 

 swimming would accompany prey capture or 

 predator avoidance. The tubercle and pore system 

 might act to maintain laminar flow during burst 

 swimming, as hypothesized by Bone ( 1972) for the 

 oilfish, Ruuettus. 



The elongate tail of Desmodema can be related 

 to the hypothesized mode of life. The lateral line 

 runs the length of the tail, ending at the caudal. 

 The tail then serves the function of greatly extend- 

 ing the lateral line, and in effect provides an an- 



845 



