The various CPUE indices used in Soviet krill fishery statistics and their main sources and primary data for their calculations were considered. The main CPUE indices were catch per fishing day (CFD), catch per extended fishing day (CEFD) and catch per hour (CH). The primary data for all the calculations were catch per haul and haul duration. There is rather high correlation between mean monthly values of CFD and CEFD, and CFD and CH. Instead, correlation between daily values is rather low, due to high fluctuations in hauls duration caused by fishery strategy directed to different final product of fishery: human consumption, krill meal and frozen krill. When considered separately for different strategies and so hauls duration, mean daily CFD and CH are well correlated. Depending on the strategy, number of hauls per day ranged from 1 to 15, mainly 2-4, hauls lasted from 0,1 to 16 hours, mainly 3-8 hours. The higher requirements for the final product quality, the shorter haul duration and so the higher number of hauls per day. The histograms of distribution of various CPUE, hauls duration, number of hauls per day and linear regression equations allowing mutual conversion of the CPUE indices are presented.