10 years worth of recent finescale haul-by haul krill data were used to characterize behaviour of krill fishery. Analysis of travel distance in relation to catch level revealed a pattern that mean travel distances are longer after the least catch levels, and the travel distances decreases as catch level increases to certain catch levels but distance increase again above that catch level. However, this pattern only holds for operations by Japan. Preferred level of catch by Japan derived through the analysis was 15-30 tonnes/haul, which is higher than the 1980s information (5-10 tonne/haul). Other nations do not have a preference for an upper limit of catch. Locations of operations were very close even after considerable number of hauls. The scales of distance after 50 tows from the original tows were 10-15 nm, which is the scale of a single concentration. However, there was considerable year to year variability in the probabilities of operating within a local range after a number of tows. In the 1999/2000 season the probability that hauls would be made within a 30nm range after 300 hauls was only 0.1, but in the 2004/05 season it was 0.7. Fishery behaviour differentiates between market type considerations/strategies which are often the argument for changing fishing patterns and catching efficiency/operational requirements in an area. These kinds of analysis show the importance of good year-round data from observers from all vessels participating krill fishery to assist in interpreting the annual fishing results.