Length, sex and maturity stage of Antarctic krill Euphausia superba Dana occurring in discrete swarms were measured. Over a 14 d period 38 swarms were sampled in a small area southwest of Elephant Island, South Shetland Islands. Analysis indicated that the composition of the swarms in terms of the sex ratio and maturity stage was very variable. Swarms ranged from equal mixtures of male and female krill to nearly all one sex; furthermore krill of similar maturity frequently occurred together. The mean length of each maturity stage also varied between swarms; in some swarms the size of each maturity stage was bigger than average while in others each stage was smaller than average. The range of krill sizes in a swarm was often restricted, indicative of some type of size sorting. However size sorting alone could not account for the observed variation in sex ratio and maturity stage and it is likely that in addition, differential growth and active behavioural responses contribute to the observed variation.

Studies on the characteristics of krill in individual swarms require sampling methods which ensure that the swarms are sampled discretely. A study on the size composition of krill swarms using an enlarged Longhurst-Hardy plankton recorder (LLHPR) and an echosounder reveal that the length frequency distnbutions of krill in adjacent swarnls were often significantly different. In contrast, length frequency distnbutions from within the same swarm were not usually significantly different and any differences were probably caused by several swarms overlapping in the LLHPR samples. The size of krill in the samples showed as much variation over short d~stances (several hundred metres) as over longer distances (several kilometres). Differences in size between adjacent swarms are probably due to a length-dependent sorting mechanism that restricts the size range of krill in the swarm. The presence of a few adults in many juvenile swarms indicate that this mechanism does not always operate.