This paper describes in detail a method for obtaining food samples from Pygoscelid penguins and some suggested modifications to CEMP Standard Method A8. It proposes a new method for draining the water from samples, an extension of the sampling period to include the guard phase, and that birds found to be empty of food be included in the data set if they are bona fide breeders. It also suggests that the first regurgitate of each sample be kept separate from the remainder because it often consists of organisms that have been collected adventitiously by the bird on its return journey to the colony. Previous papers which have reported diet composition on the basis of a single regurgitate per bird are likely to be biased toward neritic species (amphipods, Pleuragramma antarcticum and Euphausia crystallorophias) and may underestimate the importance of E. superba in the diet of penguins from some region
