FISHERY BULLETIN: VOL. 76, NO. 3 



minimize bias due to fishes' feeding after capture, 

 the net terminated in a 1-m diameter cone of ca. 

 3-mm (Vs-in) knitted nylon mesh instead of the 

 "normal" plankton netting. 



Specimens were taken in oblique tows which 

 sampled vertically migrating species at nine dif- 

 ferent periods of the 24-h cycle. At night, cable was 

 paid out in increments over a period of 1.5 h such 

 that the trawl fished roughly equal amounts of 

 time at all depths between the surface and ca. 350 

 m. The trawl was retrieved immediately after- 

 wards for a total towing time of about 2 h. Four 

 such tows were made between last light at dusk 

 and first light at dawn. During the day, 1,200 m of 

 wire were paid out initially. This placed the trawl 

 at ca. 350-400 m. Subsequently, cable was paid out 

 in increments such that the trawl fished between 

 this depth and ca. 1,100-1,200 m over a period of 

 1.5 h and then retrieved for a total fishing time of 

 ca. 2.5 h below 350-400 m. Three such tows were 

 made during the day. At dusk, 1,500 m of cable 

 were paid out initially, placing the trawl at ca. 500 

 m. Cable was then retrieved in increments such 

 that the trawl fished between 500 m and the sur- 

 face over 1 .5 h. The trawl reached maximum depth 

 just before sunset and was on deck shortly after 

 last light. At dawn, the process was reversed, and 

 the trawl shot before first light, and fished from 

 the surface to ca. 500 m over 1.5 h such that it 

 reached maximum depth ca. 1 h after sunrise. Ship 

 speed was ca. 1 m/s (2 kn) for cable retrieval and 

 ca. 2 m/s (4 kn) for all other phases. 



In order to collect sufficient numbers of speci- 

 mens for as many species as possible, three 24-h 

 series of nine tows each (dusk, four at night, dawn, 

 and three during the day) were made 27-30 Au- 

 gust 1973. These dates were chosen to bracket new 

 moon (August 28) and minimize avoidance of the 

 trawl at night (Clarke 1973). One day tow of this 

 series was fouled and could not be repeated until 

 13 September 1973. The total range of time fished 

 by equivalent tows of each series (Table 1) 

 overlapped — considerably so for the night series 

 due to one night's fishing proceeding ahead of 

 schedule. The overlap was effectively less than 

 shown in Table 1 since most of the fishes analyzed 

 were probably taken below 50 m (based on previ- 

 ously cited studies of vertical distribution) and not 

 during the first 15 min or the last 5 min of each tow 

 when the trawl was shallower than 50 m. Con- 

 sequently, equivalent tows from each 24-h series 

 were considered replicates and specimens were 

 combined for data from each period. The nine 



496 



sampling periods will subsequently be designated 

 as follows: SS for sunset and SR for sunrise; Nl, 

 N2, N3, N4 for the four night periods in chronolog- 

 ical sequence; and Dl, D2, D3 for the three day- 

 time periods in sequence. 



Danaphos oculatus, a nonmigrating species, 

 was not taken in the shallow night tows described 

 above. Nighttime data for this species were based 

 on specimens from three night series of three tows 

 each taken 30 August-1 September and 13-14 Sep- 

 tember 1973 (Table 1) using the same towing 

 schedule described above for daytime (ca. 400- 

 1,000 m). Thus only eight periods of the diel cycle 

 were considered. The three nighttime periods for 

 D. oculatus were designated dNl, dN2, and dN3. 



In order to obtain more specimens of three 

 species of stomiatoids, I utilized specimens taken 

 24-25 May 1974 in seven tows at the same location 

 with the same net, and with the same procedure 

 and timings as Nl-4 and Dl-3. The numbers of 

 specimens used from this series will be noted in 

 the results. All specimens of the other species 

 came from 1973 collections. 



The catch was immediately preserved in 4-5% 

 formaldehyde in seawater. The specimens re- 

 mained in this solution for up to 2 yr before pro- 

 cessing, but since all specimens of a given species 

 were processed within a period of 2-3 wk, any 

 between-sample differences in weight loss due to 

 leaching can be considered negligible. 



Laboratory Analyses 



The ratio of the dry weight of the stomach con- 

 tents to that of the fish as percent was used as an 

 index of stomach fullness. Where sufficient speci- 

 mens of a given species were available, 20 from 

 each of the nine sampling periods were examined. 

 If possible, the least damaged specimens (or 

 perhaps more appropriately — equivalently dam- 

 aged specimens) were selected from a narrow size 

 range. For many species, however, it was neces- 

 sary to use specimens damaged to various degrees 

 and of all sizes between recently (but fully) 

 metamorphosed juveniles and mature adults. In 

 cases where a specimen was damaged beyond loss 

 of scales or fin rays, i.e., where tissue was missing, 

 I used the median dry weight of other specimens of 

 the same standard length. 



Each fish was briefly rinsed with tapwater and 

 gently blotted; standard length was measured to 

 the nearest millimeter. The stomach (anterior end 

 of the esophagus to the pyloric valve) was removed 



