In vitro potency data (e.g. EC(50) values), used to characterise the biological activity of chemicals, are generally based on nominal effective concentrations and thus depend on any factor influencing the availability of a compound. In this study the significance of cell binding for the availability of chemicals in vitro is (i) theoretically investigated by means of a simple equilibrium distribution model and (ii) experimentally examined using a bull sperm assay to measure the cytotoxic potency of selected compounds at different cell concentrations. Compounds were selected either to cover a wide range of hydrophobicity (log K(ow)=2.52-5.69) or to represent modes of cell binding other than partitioning into cellular lipids. With the exception of xylene, the EC(50) values increased with increasing cell concentration. The ratios of EC(50) values determined at about 120 x 10(6) and 15 x 10(6) cells/ml were: pentachlorophenol. 1.2, 1-nitronaphthalene: 1.9, thioridazine: 2.7, dieldrin: 4.1, hexachlorophene: 4.1, digitonin: 5.1, methylmercury chloride: 7.9, antimycin A: 10.1 and p,p'-dichlorodiphenyl dichloroethylene (p,p'-DDE): >19.1. The influence of partitioning into cell lipids was rather well predicted by the equilibrium distribution model, except for p,p'-DDE. The results show that cell binding can significantly affect the availability of compounds in vitro and thus toxic potencies and toxic equivalency factors.