How to measure no effect. Part II: Threshold effects in ecotoxicology

The discussion on the use of the no-effect concentration as an appropriate endpoint in ecotoxicological experiments centres around two different questions: 1. The question whether biological systems show a threshold at all in their response to specific toxicants. 2. The question of how to estimate the threshold concentration (assuming there is one) from experimental data on responses of the system at a series of exposure concentrations. The second question often dominates the discussion, and the bad performance of statistical methods for estimating ‘thresholds’ is taken as an argument that no thresholds exist. The first question is, however, more fundamental and deserves due attention before statistical methods are applied. Thresholds may exist at one level of biological organization, but not at another. The blocking of biochemical receptors may be a threshold phenomenon, but thresholds are unlikely to appear in ecosystem studies and in toxicants which act on many different targets. Physiological research has shown that numerous mechanisms exist which help to maintain the integrity of the organism and which often provide strong defence against the toxicant action. There is every reason to assume that all organisms are able to some extent to withstand the exposure of toxicants at the whole-body level. Therefore, the exposure–effect relationship of a toxicant may be interpreted as the response of a system with strong feedback control. In conclusion, there is no a priori biological reason to discard the concept of thresholds in ecotoxicological response analysis. Rather, better models, with appropriate statistical methods, must be developed, based on a biological understanding of the toxicant–receptor interaction. © 1997 John Wiley & Sons, Ltd.