In the paper the results of the investigations on three mesoscale surveys covering wide area of krill between Antarctic Peninsula and South Georgia Island are considered. Surveys including from 158 to 202 stations were carried out by Soviet research vessels in summer-autumn periods of 1983/84 and 1987/88 and in a spring period 1984/85 by a standard grid of stations according to the standard method by Isaacs Kidd Midwater Trawl in the layer 0-100 m. Krill biomass distribution (g/1000 m3) and its mean values were analysed by Subareas (48.1, 48.2 and 48.3) and for the whole area investigated. The results of surveys confirm modern understanding of general pattern of krill distribution and drift in the western part of Atlantic sector of Antarctic (e.g.: Priddle et al., 1988; Makarov, 1996). The most spatial-temporal stability of the krill distribution field with more higher biomass of aggregations was observed in the western part of the area from Antarctic Peninsula to the eastern part of the South Orkneys area. The most variability of krill distribution in spatial- temporal survey scale, which is characterised by biomass change more than two orders of the magnitude in different seasons, was observed in the region of South Georgia and adjacent waters of the Scotia Sea (Subarea 48.3). Possibility of krill transportation into the region of South Georgia is considered. It is indicated that absence of krill in the Island region at the season of 1983/84 was stipulated exclusively unfavourable oceanological conditions (absence of water flow from south areas). It is also indicated that krill absence in the eastern part of the Scotia Sea may cause its absence in the region of South Georgia independing of favourable oceanological conditions (spring-summer period 1983/84). The possibility of recruitment of the krill stock in the South Georgia region from the central part of the Scotia Sea is considered. The analysis of variability of mean biomass by areas and by seasons indicates that trend of krill abundance and biomass, obtained in the Subarea 48.1 lately (Siegel et al., 1997, 1997a) is not revealed on the whole survey area. The total biomass index in the studied area during all the seasons remained at the same level (76.5-101.7 g/1000m3).