FISHERY BULLETIN: VOL. 72, NO. 4 



cormorant, red phalarope, glaucous gull, herring 

 gull, ivory gull, black-legged kittiwake, Ross's 

 gull, Sabine's gull, murres, guillemot, Kittlitz's 

 murrelet, parakeet auklet, and horned puffin. 

 Swartz (1966:674) estimated a population of about 

 600,000 piscivorous birds (adults and fledglings) 

 at Cape Thompson, which is south of Cape Lis- 

 burne, in 1960; murres and kittiwake gulls 

 accounted for 90% of the population. He estimated 

 that the birds consumed approximately 13.5 x 10^ 

 metric tons of food during their breeding season. 

 Tuck (1960:166) found that Arctic cod were the 

 most important fish in the diet of murres, as did 

 Swartz (1966:667) in his studies at Cape Thomp- 

 son. Tuck cited depth records for murres which 

 indicate that they are able to feed throughout 

 the entire water column over the shelf portion of 

 the Chukchi Sea. Since the birds at Cape 

 Thompson are undoubtedly only a small fraction 

 of the birds that utilize the southern Chukchi 

 Sea, predation on fishes by birds must be intense 

 over the entire region. However, bird predation 

 could not have been a direct cause of the density 

 structure during WEBSEC-70 because most birds 

 appeared to have left the region prior to the time 

 of our visit. 



The hypothesis of negative phototaxis in ju- 

 venile Arctic cod had further appeal — it seems 

 congruent to presumed needs for lowering vulner- 

 ability to predation by birds during summer when 

 subsurface illumination is high while allowing for 

 vertical foraging by the cod when illumination is 

 low. Because piscivorous birds undoubtedly are 

 visual feeders, directed or passive movement 

 toward the surface by juvenile cod at night should 

 not be counterselective. During the prolonged 

 periods of low illumination of arctic winter, and 

 under ice cover, juveniles presumably would be 

 able to occupy the entire water column over the 

 continental shelf. However, because larger ju- 

 veniles, about 15 cm TL, were commonly seen in 

 daylight near ice that had been broken by the 

 icebreaker, negative phototaxis of the juveniles, if 

 present, must decrease with growth in favor of 

 life near a substrate. Changes from a pelagic 

 to demersal existence during early grow^th stages 

 is a commonplace in fishes. 



Upwelling and downwelling seemed a plausible 

 explanation of the differences in depth noted for 

 the density structure. Salinity of the upper water 

 column and an index of elevation or submergence 

 of the density structure, explained below, showed 



good correlation (Table 6). The index was obtained 

 for each station by algebraically projecting the 

 pooled value for slopes of all regressions of 

 number vs. depth, multidepth and replicate 

 stations, through the station data, to the hypo- 

 thetical apex of the density structure at the 

 station. If the apex lay above the ocean surface 

 (e.g.. Figure 2, stations 41 and 70), the converted 

 amount in meters was negative and was taken as 

 the amount the density structure was elevated. If 

 the apex lay below the ocean surface (e.g., Figure 

 2, stations 88 and 92), the converted amount in 

 meters was positive and was taken as the amount 

 the density structure was submerged. For each 

 replicate station the regression was based on a 

 point determined by depth of trawling and the 

 average number of juvenile cod captured at the 

 station. All indices for salinity were based on data 

 for 10-18 m because this zone was the common 

 depth of the replicate hauls; however, the water 

 column was usually nearly isohaline from the 

 surface to considerably below this depth (Figure 2). 

 Two sources of upwelling seemed possible — an 

 accelerated current around the Cape Lisburne- 

 Point Hope headland and wind over the sampling 



Table 6. — Comparison of hypothetical elevation of the density 

 structure of juvenile Arctic cod at the trawl stations taken at 

 late dusk or after dark, based on density distribution of cod 

 at each station, with surface salinity at a nearby station. 

 Stations in order from greatest submergence (positive values) 

 to greatest elevation (negative values) of the density structure. 

 Projections for the multidepth stations were based on the data 

 of Figure 2 fit to the standard slope of 0.0669; methods for 

 obtaining projections from the replicate data are explained 

 in the text. 



'M = multidepth station; R = replicate station. 



1100 



