Published relationships of various length measurements as predictors of wet or dry weight in Antarctic krill Euphausia superba Dana are standardised with new data on the relationships of three length measurements to wet and dry weight as functions of sex, maturity stage and moult stage. The range of coefficients and exponents for these relationships are examined and an example given to indicate the potential such variation has for introducing error into estimates of biomass based on acoustic data. An alternative approach is examined in which length and additional biological data are assessed in terms of decreases in the residual variance of relationships. We have identified key sources of variability where relatively small increases in the effort of analysis result in large improvements in the precision of prediction. Surprisingly, the stage of the moult cycle of the animal has little effect upon length-weight relationships. The use of categories of sex and maturity stage, however, has marked effect upon the residual variance. Again surprisingly, the simple division of krill into male and female categories is of little practical use in improving the precision of any prediction of mass. However, the separation of gravid females, either from male or non-gravid female krill or, from adult male and other krill does result in a marked improvement in the precision of prediction. Example equations are provided.

(In press: British Antarctic Survey Bulletin No. 79 (May 1988)).