The toxicities of an oil dispersant (BP 1100X) and a surface active agent (Shell Herder) upon 18 marine species from different taxa, were investigated. The results showed that the toxicity of a product depends very much on the species tested. Some species exhibited high mortality when treated with BP 1100X, and low mortality when treated with Shell Herder, whereas the reverse was true for certain other species. The results therefore indicated a large bias potentially incurred in present procedures, by the use of one or two species in toxicity testing and screening of oil dispersants/surface active agents. It is here suggested that toxicity tests should be carried out on species which are ecologically important (e.g. 'key species' of a community or population with a high energy flow value) in identified receiving environments, rather than on those which are easy to obtain and maintain in the laboratory. Copyright © 1981 Published by Elsevier Ltd.