FISHERY BULLETIN: VOL. 80, NO. 2 



area have been summarized by Maynard et al. 

 (1975). 



Over 2,800 specimens of nine families were 

 examined. Based upon preliminary results, mor- 

 phology, and the literature, the species were 

 separated into two groups, each of which was 

 treated differently. Members of the Photichthyi- 

 dae, Sternoptychidae, and Gonostomatidae were 

 considered planktivores; and those of the Astro- 

 nesthidae, Chauliodontidae, Idiacanthidae, Mel- 

 anostomiatidae, Stomiatidae, and Malacostei- 

 dae as nekton-eating species. 



All specimens of planktivorous species were 

 taken with a 3 m Isaacs-Kidd midwater trawl 

 which terminated in a 1 m diameter cone of ca. 3 

 mm mesh netting with a ca. 2 1 nonfiltering cod 

 end bucket. Towing procedures were the same as 

 described in Clarke (1980). The trawl was low- 

 ered to a given depth as rapidly as possible, 

 towed for 2-3 h at ca. 2 m/s, and retrieved as rap- 

 idly as possible. A time-depth recorder of the 

 appropriate range was attached to the trawls; 

 depth records were accurate to 2-4% of the depth 

 fished. In addition to night tows at 70-170 m 

 described in Clarke (1980), specimens were 

 also taken from day tows at 400-800 m and night 

 tows at 225-250 m made in September 1973 and 

 November 1974 (Table 1). During some of the 

 deeper tows, the trawl changed depth by as much 

 as 50-100 m during the "horizontal" portion of 

 the tow. 



Since the most abundant planktivores were 

 known to feed during the day (Clarke 1978), zoo- 

 plankton were sampled at 400-500 m during the 

 day (Table 1) with opening-closing 70 cm diame- 

 ter bongo nets (505 yum mesh). The nets were low- 

 ered closed, opened for 0.5-1 h atca. 1 m/s ship's 

 speed, then closed, and retrieved. Time-depth re- 

 corders attached to the nets indicated vertical 

 movement of up to 100 m during the open por- 

 tions of the tows. Volume sampled by each net 

 was estimated from the mouth area, duration of 

 the open portion of the tow, and an estimated 

 speed of 1 m/s. 



All material was preserved immediately after 

 capture and held in ca. 4% formaldehyde/sea- 

 water. Except for certain trawl samples where 

 large numbers of Vinciguerria nimbaria were 

 caught and only individuals with obviously full 

 stomachs were selected, all specimens of each 

 species considered were measured (standard 

 length, SL, to the nearest millimeter) and stom- 

 achs examined. Intact prey items were identi- 

 fied, counted, and measured to the nearest 0.1 



mm (prosome length for copepods, total length 

 without telson for malacostracans, and total 

 length or maximum dimension for other prey). 

 Identifiable remains among partially digested 

 material were recorded. Any remains of chae- 

 tognaths (the only gelatinous prey found) were 

 counted as intact since they are probably de- 

 graded much more rapidly than other prey 

 types. 



Items in the mouth or esophagus were not 

 counted; their limbs and bodies were not com- 

 pressed, indicating that they had been taken 

 after capture. Otherwise, there was no evidence 

 of postcapture ingestion by the fishes. Most prey 

 types found intact in the stomach were also re- 

 corded as digested remains that had almost cer- 

 tainly been eaten well before capture, and, 

 conversely, several types of abundant zooplank- 

 ton were rarely or never found in the stomachs, 

 as would be expected if the fish were feeding in- 

 discriminately in the net. There was no evidence 

 that food was regurgitated during or after cap- 

 ture; I found no everted stomachs and no digested 

 remains in the esophagus. 



Zooplankton from the bongo net samples were 

 counted from either the entire sample (euphausi- 

 ids and other relatively large types) or 1/16-1/32 

 aliquots taken with a Folsom plankton splitter. 

 For both plankton and intact prey items, eu- 

 phausiids and most copepods (with the exception 

 of unidentifiable copepodites, which were fairly 

 common in all the plankton samples) were identi- 

 fied to species. Ostracods (mostly Conchoecia 



TABLE 1.— Dates, local (Hawaiian Standard) times, and depths 

 of tows with 3 m Isaacs-Kidd midwater trawl and 70 cm bongo 

 plankton nets off Oahu, Hawaii. Times for trawls are for the 

 period at depth; total times including descent and ascent are in 

 parentheses. Times for bongos are for the open period only. 

 Depth figures are the ranges during "horizontal" portions of 

 tow or modal depth if the range was <20 m. 



288 



