Predictive models based on quantitative structure activity relationships (QSARs), are used as rapid screening tools to identify potentially hazardous chemicals. Several QSARs are now available that predict the acute toxicity of narcotic-industrial chemicals. Predictions for compounds acting through more specific modes of action, which are generally more toxic, require alternative models. Current efforts are focused on characterizing a toxic mode of action database from which the structural requirements for specific mechanisms can be elucidated. Approaches include application of joint toxicity theory for chemical mixtures and the assessment of fish acute toxicity syndromes, that is, distinct sets of in vivo rainbow trout (Salmo gairdneri) toxic responses that correspond to specific modes of action. By measuring a number of respiratory-cardiovascular variables, response sets associated with nonpolar and polar narcotics, oxidative phosphorylation uncouplers, respiratory membrane irritants, respiratory inhibitors, acetylcholinesterase inhibitors, and central nervous system seizure agents have been described.