162 



NORTH AMERICAN FAUNA 74 



Time (hours) 



Fig. 101. Percentage frequenc>' of occurrence per hour of \\alruses sighted in the 

 water (versus on ice). Data from aerial and surface sur\eys of the Bering Sea (Feb- 

 ruary-May) and one aerial sur%e>' of the Chukchi Sea (September). Sample sizes: 

 aerial — February-March 14.814. April 23.550. September 11.094; surface — March 

 2.718. April 3.478, May 2.721. (From K. \V. Kenyon. J. A. Estes. and F. H. Fay. un- 

 published data) 



ranged from 4.5 to 5.5 min, and mean surface time was 55 s between dives. 



E. H. Miller, G. C. Ray, and I observed a group of about 50 adult females 

 and young feeding continuously for more than 5 h at a depth of 50 m. The 

 animals surfaced and dove nearly in unison and remained beneath the surface for 

 about 7 to 9 min per dive. One individual's surface interval between dives was 

 1 min 25 s. I watched another adult female feeding at a depth of 55 m, and 

 timed two of her dives. The first was 5 min and the second 6 min 15 s; surface 

 intervals following each dive were 45 and 60 s, respectively. 



J. A. Estes and I observ^ed a single adult male for about 2 h while it fed at a 

 depth of 79 m. During that time, the walrus completed 12 dives, 8 of which 

 ranged in duration from 9 min to 10 min 2 s; the other 4 dives were much 

 shorter (2 min 10 s to 7 min 22 s) . Surface times between the longer dives ranged 

 from 1 min 18 s to 2 min 6 s; those between the shorter dives were from 31 s to 

 1 min 45 s. Loughrey (1959) reported feeding dives of up to 10 min duration, 

 with about 3 min at the surface between dives; Johansen (1912) reported feeding 

 dives of about 10 min duration in water 70 m deep. 



The mean subsurface times indicated b>' these observ ations show a tendency 

 toward positive correlation with depth, but the relation is not close (Table 26). 

 Inexplicably, the surface intervals between dives seem to be negatively correlated 

 with dive duration. The ratio of surface to subsurface time in the samples as a 

 whole ranged from about 1:1 to 1:8. For those in the Bering-Chukchi region, at 

 depths most representative of walrus habitat there, the range was about 1:5 to 

 1:8 (mean, about 1:6.5). This is comparable to the ratio given by Ray and 

 Watkins (1975) for male walruses in mating displays (1:5.3) and suggests that 



