To demonstrate how baseline toxicity can be separated from other more specific modes of toxic action and to address possible pitfals when dealing with hydrophobic substances, the four isothiazol-3-one biocides N-methylisothiazol-3-one (MIT), 5-chloro-N-methylisothiazol-3-one (CIT), N-octylisothiazol-3-one (OIT), and 4,5-dichloro-N-octylisothiazol-3-one (DCOIT) as an example for reactive electrophilic xenobiotics were tested for their cytotoxic effects on the human hepatoblastoma cell line Hep G2, on the marine bacterium Vibrio fischeri, and on the limnic green alga Scenedesmus vacuolatus. In each of the three test systems, toxic effects were observed in a consistent pattern. The two chlorinated compounds and OIT were found to be significantly more toxic than MIT. As compared to baseline toxicants, the small and polar MIT and CIT exhibited pronounced excess toxicity in each of the three test systems that is presumably triggered by their intrinsic reactivity toward cellular thiols. In contrast, OIT and DCOIT showed mainly toxicities that could be explained by their hydrophobicity. Analyzing and comparing these results using the toxic ratio concept and with data that indicate a dramatic depletion of cellular glutathione levels after incubation with DCOIT reveals that for highly hydrophobic substances, baseline level toxicity in an assay for acute toxicity can lead to an oversight of other more specific modes of toxic action that may cause significant effects that might be less reversible than those caused by unreactive baseline toxicants. This possibility should be taken into account in the hazard assessment of chemicals that are both hydrophobic and reactive.