ABSTRACT:The potential toxicity of several herbicides-alachlor, diuron and its photo and biotransformation products, glyphosate and its metabolite aminomethyl phosphonic acid (AMPA)-to nontarget cells was assessed using two microorganisms frequently used in ecotoxicology, Vibrio fischeri and Tetrahymena pyriformis. Toxicity assays involved the Microtox 1 test, the T. pyriformis population growth impairment test employing three different processes (flasks, tubes, microplates), and the T. pyriformis nonspecific esterase activities test. Several IC 50 or EC 50 values are reported for each molecule. Alachlor exerted a toxic effect on the two nontarget cells used. The results for diuron and its photo and biotransformation products indicated that most of the metabolites presented nontarget toxicity higher than that of diuron. Glyphosate and AMPA had a less negative effect on T. pyriformis than on V. fischeri. Nevertheless, in all cases, glyphosate was found to be more toxic than AMPA. Comparison analysis of the sensitivity of the different tests showed that, in general, tests using the eukaryotic cell (T. pyriformis) were more sensitive than test using the prokaryotic cell (V. fischeri), and that a population growth criterion is more sensitive than an enzymatic criterion. The three different processes that could be used to evaluate effects on population growth rate were equally sensitive for the herbicides tested. A significant correlation between toxicity data and the hydrophobicity of the chemicals could only be established with the growth population test. This study demonstrates that it is essential to assess the toxicity of the metabolites formed to complete a more comprehensive study of the environmental impact of a polluting agent. # 2007 Wiley Periodicals, Inc. Environ Toxicol 22: 78-91, 2007.