Abstract
Acute-chronic ratio (ACR) implicitly needs the assumptions that acute toxicity and chronic toxicity have a log linear relation and that the slope of the log linear regression model is one exactly. The aim of this study was to test whether this assumption is valid or not. Ecotoxicity datasets of algae, daphnia and fish were statistically analyzed by using log linear regression models. The validity of assumption was judged by the slope of the regression model that was selected based on the AIC of the models. The results showed that the assumption was valid for algae, while the assumption was not valid for daphnia and may not be valid for fish. Our results also showed that chemical structures and physicochemical properties can become effective factors which improve predictive ability of regression models.
