Gut fullness of 3264 Antarctic krill in 38 swarms was estimated using fluorometric determination of chlorophyll a . There was a very wide range of gut fullness (maximum value 3277 ng and median of 190 ng) and 7% of the population was classified as empty. There was a similar variability between swarms. Analysis of this variability attempted to partition it between three sources. First, the swarms differed in their composition and it was thought that intrinsic properties of the individual krill might contribute to interswarm variation. Although certain classes of animal had significantly different gut fullness, this was unimportant for the population as a whole. Animal length only accounted for 0.33% of variation, sex-and-maturity stage for 1.6% and moult stage for 4.2% of the variation. By contrast, differences between swarms accounted for 59% of population variability. Second, examination of characteristics of the swarms themselves, such as size, depth and time of sampling, did not yield any property which explained further variability. The final source of variation was likely to be environmental factors, and it is suggested that variation in gut fullness between swarms reflects patchiness in their phytoplankton food.