In this study, the effect of hardness on the combined outcome of metal mixtures was investigated using Daphnia magna. The toxic unit (TU) was calculated using modified LC50 values based on the hardness (i.e., LC50-soft and LC50-hard). From a bioassay test, the degree of sensitivity to hardness on the toxicity changes was in the order: Cd < Cu < Zn < Pb, with 25, 66, 77 and 88% decreases in the LC50 values, respectively, when the hard test solution was replaced with a soft test solution. In mixture toxicity tests, the difference in the test solution hardness was found to clearly cause different toxicities, as determined by the TU calculated by the LC50-hard, using the toxicity of a standard culture medium as the reference. That is, approximately four to five times higher toxicity was observed in soft (i.e., 44 +/- 4 mg/L as CaCO3) rather than hard water (i.e., 150 +/- 10 mg/L as CaCO3) test solutions. In the tests where the modified reference toxicity values (i.e., LC50-soft and LC50-hard for soft and hard test solution, respectively) obtained from the individual metal toxicity tests with different hardness were used to calculate the TU, the results showed very similar D. magna toxicities to those of the TU from the mixture of soft and hard test solutions, regardless of the hardness. According to the toxicity results of the mixture, the aquatic toxic effects of the acid mine drainage (AMD) collected from mine areas that contained metal mixtures were investigated using Daphnia magna and the modified LC50 value of the TU hardness function calculated for varying solution hardness. The results of the biological WET test closely matched our overall prediction, with significant correlation, having a p-value of 0.513 in one way ANOVA test (n = 19). Therefore, this study revealed that the predicted toxicity of the metal mixture agreed well with the biological toxicity test when the modified LC50 value was employed as the basis of hardness in the TU calculation. (c) 2006 Elsevier B.V. All rights reserved.