Predictability of the mixture toxicity of 12 similarly acting congeneric inhibitors of photosystem II in marine periphyton and epipsammon communities.

Testing of single chemicals with single species is common ecotoxicological practice in contrast to contaminated environments where highly diverse biological communities are exposed to highly diverse mixtures of chemical compounds. We, therefore, investigated whether mixture toxicity approaches that have been used successfully for single species, might also be applied on a community level of biological complexity. Twelve inhibitors of photosystem II, selected by QSAR and chemometrical approaches as the structurally most similar from a congeneric group of phenylurea herbicides, were tested singly and as mixtures on two types of marine microalgal communities, periphyton and epipsammon. Inhibition of photosynthesis was measured in short-term tests using incorporation of radiolabelled carbon (14C) to estimate photosynthetic rates. Two basic concepts, concentration addition (CA) and independent action (IA), were used to predict the toxicities of the mixtures. Congeneric and similar-acting substances such as the phenylureas are expected to comply with CA rather than IA. The aim of the present study was to evaluate whether these concepts can be used to predict mixture toxicity also to periphyton and epipsammon photosynthesis, i.e. at the level of natural communities. We found that deviations between observed and predicted mixture toxicity were relatively small but that CA predictions were the more accurate ones. The predictions proved to be robust, when based on single substance information even from different seasons, years, and sites. We conclude that the concept of CA for predicting mixture toxicity applies also at the community level of algal testing; at least when a physiological short-term effect indicator is used that matches the mechanism of action of the substances. Copyright 2004 Elsevier B.V.