We examined changes in water chemistry and copper (Cu) toxicity in three paired renewal and flow-through acute bioassays with rainbow trout (Oncorhynchus mykiss). Test exposure methodology influenced both exposure water chemistry and measured Cu toxicity. Ammonia and organic carbon concentrations were higher and the fraction of dissolved Cu lower in renewal tests than in paired flow-through tests. Cu toxicity was also lower in renewal tests; 96 h dissolved Cu LC(50) values were 7-60% higher than LC(50)s from matching flow-through tests. LC(50) values in both types of tests were related to dissolved organic carbon (DOC) concentrations in exposure tanks. Increases in organic carbon concentrations in renewal tests were associated with reduced Cu toxicity, likely as a result of the lower bioavailability of Cu-organic carbon complexes. The biotic ligand model of acute Cu toxicity tended to underpredict toxicity in the presence of DOC. Model fits between predicted and observed toxicity were improved by assuming that only 50% of the measured DOC was reactive, and that this reactive fraction was present as fulvic acid.