28 



NORTH AMERICAN FAUNA 74 



in and near Anastasia Bay (Pinigin and Prianishnikov 1975; V. N. Gol'tsev, 

 personal communication). 



Nearly all of the migrants that are bound for the Chukchi Sea apparently 

 reach that goal by the end of June. However, some movement of females and 

 young eastward out of the Gulf of Anadyr and northward through the western 

 part of Bering Strait continues for at least another month. This movement was 

 recognized by Belopol'skii (1939) and has been confirmed by some of the St. 

 Lawrence Islanders (personal communication) who formerly resided in those 

 areas. Occasional stragglers and herds of males have appeared irregularly along 

 the Alaskan Bering Sea coast and islands in July. Major concentrations of males 

 have remained at Big Diomede and Arakamchechen islands, on Rudder and 

 Meechken spits, and in Bristol Bay. Repeated sightings of herds along the 

 southern Koryak coast and in northern Karagin Gulf (Fig. 10) suggest that some 

 males regularly summer there, as well. 



The pattern of southward migration in autumn is less clear but seems to be 

 about the reverse of the spring migration. That is, the herds along northeastern 

 Chukotka seem to come down through Bering Strait first, then to eastern St. 

 Lawrence Island and to the western Alaskan coast in the Yukon-Kuskokwim 

 region (Figs. 12 and 13), presumably heading for Bristol Bay. At the same time, 

 those in the northern part of the Gulf of Anadyr leave the hauling grounds and 

 probably head southward. Last to come south are those from the Barrow and 

 Wrangell Island areas, which probably meet in Bering Strait and make their way 

 to the St. Lawrence wintering area. 



This interpretation of the migrations is unavoidably speculative, but the data 

 indicate to me that the three summer concentrations hypothesized by Belopol'skii 

 (1939) do exist. There are others as well along northeastern Chukotka, in Bristol 

 Bay, and in the Koryak region. These concentrations do not appear to be discrete 

 to the degree that he hypothesized, though the data suggest that the amount of 

 interchange between them is not large, even in the Chukchi Sea. The Bristol Bay 

 and Koryak summering groups appear to be entirely isolated from all others. 



In years with average to heavy ice conditions, the wintering population seems 

 to be concentrated mainly in two areas: north-central Bering Sea, to the south 

 and west of St. Lawrence Island, and southeastern Bering Sea, from Nunivak 

 and the Pribilof Islands to Bristol Bay. A third, smaller concentration is found in 

 the western Bering Sea, along the Koryak coast (G. A. Fedoseev, personal com- 

 munication). The amount of interchange between the two main concentrations 

 before their northward migration is unknown but does not appear to be large. 

 Few individuals have been sighted in the intervening area during January to 

 March (Figs. 4, 5, 6). In most winters, the ice in that intervening area is 

 extremely heavy and compact (Brower et al. 1977) and probably inhibits passage 

 of walruses. In years of unusually light ice coverage, when the pack does not 

 extend south of 61 °N, this "barrier" may be absent, and division of the 

 population into two main groups may not occur. Such light icing took place only 

 once in the past 30 years (winter 1966-67), and walrus surveys were not 

 conducted then until mid- April, after the ice had receded northward to 63 °N 

 and the migration was well under way Q. J. Burns, personal communication). 



If my interpretations of the migrations are correct, the spring harvests at 

 Eskimo villages along the Alaskan coast from Bristol Bay to Bering Strait, 



