The results of an analysis of haul-by-haul catch statistics from the Soviet krill fishing fleet during the 1989/90 season (sample of 3 614 hauls for the period from 1 November 1989 to 12 June 1990) are presented. During the season, fishing vessels worked one fishing ground off the northwestern tip of Coronation Island. CPUE showed variability with a minimum in November (3.2 tonnes/hour) and maximums in February and March (9.6 tonnes/hour and 11.0 tonnes/hour respectively), followed by a decrease to 6.9 and 7.1 tonnes/hour by May and June. The location of the fishing ground was observed to be stable, and it is considered that the location was determined by certain hydrodynamic characteristics of the area, for example a complex current around the islands together with topographically-induced effects. Except in November 1989, temporally and spatially sustained fishable krill aggregations did not form outside the main fishing area. During the November period the commercial fleet fished over oceanic areas, and targeted krill aggregations found in the major current flowing towards the north-east of the region. The fleet was observed drifting with the current together with krill aggregations. A maximum CPUE (3.8 tonnes/hour) was recorded at the beginning of the period, with a subsequent gradual decrease towards the minimum level acceptable to the fleet (2.5 tonnes/hour). Over the last five days of November, the aggregation disintegrated, having existed for about 25 days, and commercial vessels returned to Coronation Island. Krill drift velocity, estimated on the basis of the fishing ground shift, amounted to 7.4 km/day or 8.7 cm/sec. This is comparable to, though less than, the drift velocity off Elephant Island (11 to 13 cm/sec), evaluated earlier using the same method (Sushin and Myskov, 1992).