In this investigation we discuss the effects of different procedures during the recovery of samples of waterlogged sediments. Seventy-three samples of a Neolithic cultural layer were recovered by the wet-sieving technique. Different operators processed the samples and although they got the same instructions, they handled the material in different ways. Samples showed quantitative as well as qualitative differences in their macroremain composition. Using statistical methods (analysis of variance

more »

... and correspondence analysis) we could analyse which differences were a consequence of inconsistent wet-sieving. The taxa appearing in the samples were unequally affected: some were almost completely eliminated by intensive wet-sieving, some were only slightly reduced, and some were not affected. It could be demonstrated that the number of fragile remains e.g. cereal chaff remains and Malus sylvestris pericarp fragments were reduced by increasing wet-sieving intensity, but more or less round and robust remains were scarcely affected. To conclude, we suggest a less destructive recovery technique for plant macroremains and give some advice on how to deal with biased data.