Investigations of diurnal variations in the size composition of Euphausia superba were carried out on the commercial trawler Grigory Kovtun near the South Orkney Islands from March to June 1990. Observations at six daily stations were carried out in various locations inside the fishing area. Each station consisted of a series of catches made using a standard commercial trawl (9 to 12 tows per station, over one day). An increase in the total average size of animals caught in periods of light or darkness was noted at several stations. Increases in the proportion of males in the catch, as well as an increased difference between average length of males and females in the layer fished were indicative of these variations. Diurnal variations of females’ size composition were usually less evident. These changes were related to diurnal vertical migrations of krill which were noted on echo sounder recordings. Trawling depths usually corresponded to the depth of the largest concentration of krill. Within swarms males initiate diurnal vertical migrations. In the absence of diurnal migrations (particularly late in the season), diurnal variations in size composition of krill catches were less evident or non-existent. The significance of these observations in relation to krill size composition data obtained from standard surveys (only one sample of krill per station) is discussed. A gradual decrease in the average size of krill from the end of March to June was observed in the fishing area. The following three causes of the observed variations in krill composition are considered: (i) post-spawning mortality of large specimens; (ii) body shrinkage of krill due to a decrease in food availability; and (iii) size selectivity of krill by commercial fisheries.