WCKSK; KTAl..; KKllAVIOKOF HOWUKAD WllALKS 



nak artificial island (Fig. 1). The clouds represented 

 suspended mud and not plankton because the 

 material was of the same color as mud dredged up by 

 industrial activities. Whales surfaced with large 

 amounts of muddy water streaming from their 

 mouths, indicating they had been feeding from or 

 near the bottom. This behavior occurred in 24-29 m 

 of water and seemed very localized. We saw no in- 

 dication of bottom feeding in the same area on 22 

 August 1980, but we had observed similar mud 

 clouds nearby on 9 August, when prolonged observa- 

 tions were not possible. On 25 August 1981, whales 

 again surfaced with mud streaming from their 

 mouths. The location was 15 km south of the position 

 where we observed such behavior in 1980; water 

 depth was only 10-13 m. 



These are, to our knowledge, the first published 

 behavioral observations of apparent near-bottom 

 feeding by bowhead whales. However, Johnson et al. 

 (1966), Durham (1972), and Lowry and Burns (1980) 

 detected pebbles and bottom-dwelling species in 

 bowhead stomachs. 



Bottom-feeding whales were usually separated 

 from other whales by 150-300 m when at the surface. 

 On 12 August 1980, at least 10 whales were bottom 

 feeding within an area of 3 km radius. Whether they 

 were feeding on inbenthic or epibenthic inverte- 

 brates we do not know. In the eastern Beaufort Sea, 

 the average biomass of inbenthic animals greatly ex- 

 ceeds that of epibenthic animals (Griffiths and 

 Buchanan footnote 7). However, the latter may 

 occur in dense swarms in certain places. For a balae- 

 nid whale, such swarms would seem to be a much 

 more suitable type of food than inbenthic organisms. 

 Mud might be taken inadvertently along with epi- 

 benthic animals. 



Skim Feeding 



The only feeding type that we observed directly 

 was skim feeding. In the third week of August 1980, 

 we observed whales moving slowly and deliberately 

 at the surface with their mouths open wide. Usually 

 the rostrum just broke the surface of the water, and 

 was parallel to it. In these cases, the lower jaw was 

 dropped to varying degrees, as could be seen from 

 the depth of the white chin patch. In 1980, skim 

 feeding was observed along the Tuktoyaktuk Penin- 

 sula in water 12-22 m deep. Whales occasionally 

 skim fed alone, but more often did so in groups of 

 2-10 or more individuals. During any one observation 

 period, they stayed in the same general area by 

 repeatedly turning and did not appear to make any 

 net geographic movement. However, we found 



groups of skim-feeding whales in different locations 

 on different days. 



During 1981 , we witnessed skim feeding on a large 

 scale only on the evening of 18 August, 32 km NNW 

 of Pullen Island in water 25 m deep. About 20-30 

 whales in the 25 km^ area were swimming with 

 mouths open; they travelled slowly, usually just 

 below the surface (~ 2-3 m deep). Copepods were 

 unusually abundant in near-surface waters at this 

 location and time (Griffiths and Buchanan footnote 

 7). On 23 and 24 August 1981, we saw one isolated 

 example on each day of a whale feeding at the sur- 

 face briefly (observed for < 1 min) in approximately 

 the same area as on 18 August. 



Typically, skim-feeding whales were oriented with 

 their backs at the water's surface. However, they 

 occasionally swam on their sides with mouths open at 

 an angle of about 60°, and once we saw two whales 

 separated by three body widths swimming on their 

 sides, belly to back. In one instance, a skim-feeding 

 whale swam inverted for at least 3 min, with the 

 underside of its chin at the surface. 



Frequently, the skim-feeding whales swam in 

 echelon formation, each whale swimming just behind 

 the preceding whale, but offset laterally by one-half 

 to three body widths, reminiscent of geese in V for- 

 mation (Fig. 6). At other times, they swam abreast 

 and parallel to one another. Videotape from 18 

 August 1981 showed that whales within the echelons 

 were a mean of 0.53 whale lengths apart (SD = 

 0.599, n = 66), or about 8 m. These distances were 

 measured from different echelons or from the same 

 echelon at intervals of at least 5 min. We videotaped 

 a recognizable whale for almost 3 h on this day as it 

 skim fed in changing echelon formations, usually 

 taking the lead position. Echelons were clearly 

 dynamic in terms of membership, size, and organiza- 

 tion. In 1981, the mean echelon size was 4.7 animals 

 (SD = 4.05, n = 23). While the largest such forma- 

 tion observed in 1980 contained 5 individuals, the 

 largest in 1981 contained 14 animals. 



We suspect that echelon feeding increases the 

 feeding efficiency of those animals staggered behind 

 and to the side of other individuals, perhaps by help- 

 ing them to catch prey that escape or spill from the 

 mouth of the whale in front, or by reducing the abili- 

 ty of prey to escape to the side. Skim feeding in eche- 

 lon may allow more effective exploitation of concen- 

 trated patches of small prey than would be possible if 

 whales were feeding alone. If so, the change in effi- 

 ciency that accrues when echelons are formed may 

 have an important cost/benefit effect on energy ex- 

 pended per whale. The predominant prey types of 

 bowheads include copepods and euphausiids (Lowry 



365 



