The "best" raising procedure for discards in fishery science is a method everyone talks about but nobody has ever seen it. The specific workshop on discards (Anon. 2003) proposes formulas that are too simplistic as they do not take into account the within trip variance. The literature is full of attempts and results but they are always specific to particular case studies (Stratoudakis et al, 2001, Cotter et al, 2001) or based on inconsistent hypothesis as demonstrated by Rochet and Trenkel (2005). When some generalities are finally found (Borges et al., 2005), the formulas developed do not correspond to multistage sampling because of the unavailability of essential parameters and because the aim of the study is "not to identify optimum raising procedures". The object here is not to discuss the accuracy of such or such document but to build step by step the reasoning for estimating the total volume of discards, in weight or in number, from the hypothesis to the associated variance. With the objective of clarity, the exact reference to the chapters of the two books always cited (Cochran, 1977 and Thomson, 1992) will be provided and when possible the reference to formulas specified in Cochran's sampling theory.