To document the toxicity of copper and nickel in binary mixtures, freshwater amphipods Gammarus pulex were exposed to the metals given independently or as mixtures. Toxicity to Cu alone was relatively high: 96-h LC10 and LC50 were found at 91 and 196 μg L(-1), respectively. Toxicity to Ni alone was very low, with 96-h LC10 and LC50 of 44,900 and 79,200 μg L(-1), respectively. Mixture toxicities were calculated from single toxicity data using conventional models. Modeled toxicity was then compared with the measured toxicity of the binary mixture. Two kinds of mixtures were tested. Type I mixtures were designed as combinations of Cu and Ni given at the same effect concentrations, when taken independently, to identify possible interactions between copper and nickel. In type II mixtures, Cu concentrations varied from 0 to 600 μg L(-1) while the nickel concentration was kept constant at 500 μg L(-1) to mimic conditions of industrial wastewater discharges. Ni and Cu showed synergic effects in type I mixtures while type II mixtures revealed antagonistic effects. Low doses of Ni reduced Cu toxicity towards G. pulex. These results show that even for simple binary mixtures of contaminants with known chemistry and toxicity, unexpected interactions between the contaminants may occur. This reduces the reliability of conventional additivity models.