Samples of Antarctic krill collected from six seabird species and Antarctic fur seal during February 1986 at South Georgia were compared to krill from scientific nets fished in the area at the same time. The length-frequency distribution of krill was broadly similar between predators and nets although the krill taken by diving species formed a homogeneous group which showed significant differences from krill taken by other predators and by nets. There were significant differences in the maturity/sex stage composition between nets and predators; in particular all predator species showed a consistent sex bias towards female krill. Similarities in the krill taken by macaroni (offshore feeding) and gentoo (inshore feeding) penguins and differences between krill taken by penguins and albatrosses suggest that foraging techniques were more important than foraging location in influencing the type of krill in predator diets. Most krill taken by predators were adult; most female krill were sexually active (particularly when allowance is made for misclassification bias arising from predator digestion). Because female krill are larger, and probably less manouverable than males, the biased sex ratio in predator diets at this time of year may reflect some combination of selectivity by predators and superior escape responses of male krill. Overall, adult, sexually active female krill, forming 40 % by number of the local krill population, may comprise 60 - 70 % by number and 75 - 88 % by mass of the krill taken by their main seabird and seal predators at South Georgia at the time of peak local demand in February.