In order to evaluate if toxicity of the bile may be used to establish lethal and sublethal exposure of fish to pollutants, rainbow trout (Oncorhynchus mykiss) were exposed to PCP-spiked sediment (4, 8, 16, 32 and 64 mg PCP/kg w.w.) for 1 or 2 days, sacrificed and their bile tested for toxicity to Daphnia magna. Exposure was monitored daily by Daphnia toxicity tests of sediment and water. The 24-h EC50s, expressed as mg PCP/kg sediment w.w., were 20 +/- 6.4 for sediment and 9.6 +/- 5.1 for water phase tests, respectively. The corresponding 24-h LC50 value for sediment to rainbow trout was ca 16 mg/kg w.w. Bile toxicity was tested directly and after hydrolysis by acid and by beta-glucuronidase. Bile and bile extract of surviving trout exposed to the median lethal concentration of sediment (16 mg PCP/kg w.w.) were 10 times more toxic than control fish. Hydrolyzed bile extract of fish exposed to 4 mg PCP/kg w.w. was 10-50 times more toxic than that of control fish. Thus, toxicity tests of extracts of hydrolyzed bile were more sensitive than toxicity tests based on water and sediment. Therefore, at least for chemicals which accumulate in the bile, toxicity tests of the bile may be used to establish sublethal exposure in fish.