Estimating experimenter induced bias in field studies: A cautionary tale.

Biological sampling procedures that remove organisms or litter from the system being studied potentially bias the results. Bias can be estimated by considering the experimenter as a consumer in the system, and comparing experimenter consumption to other organisms feeding at the same trophic level. H. T. Odum's classic work is examined in light of this theory, and it is shown how important it is for workers to consider experimenter impact when evaluating the results of their studies.