In this research a test system was developed that is designed as a tool to evaluate the potential hazard of chemicals to aquatic ecosystems. The system developed is a linear three-trophic level food chain consisting of an alga (Ch/ore/la vulgaris), rotifer (Brachionus calyciflorus), and fish (Pimephales promelas). The chemostat design used for the lower two trophic levels was crucial in being able to supply the top trophic level with sufficient food on a continuous basis. The system was initially evaluated using copper (Cu) and selenium (Se) as toxicants. In the copper experiments, results of a 7 day three-trophic level toxicity test were compared with a series of single species tests. The LOEC was 31.5 µg/L based on a temporary impairment of the algal population growth, with a corresponding NOEC of 16.2 µg/L. The algal population at all initially impaired treatment levels demonstrated recovery to control levels by the end of the test. Single species tests with the same species showed impairment at treatment concentrations lower than the corresponding value from the three-trophic level test. The difference in sensitivity is attributable to the fact that most of the Cu in the single species tests was in the dissolved form (approximately 80 %), whereas in the trophic level test most of the Cu was not ( < 15 % dissolved Cu). The three-trophic level Se experiment lasted for 25 days, with both short-term and long-term impacts evident. At the algal trophic level, growth was not impaired on a daily basis at any of the exposure levels (110.3, 207.7, and 396.1 µg/L Se). However, algal densities were slightly reduced at the 207. 7 and 396.1 µg Sell treatments, although not significantly different when the data was pooled across days. Rotifer populations were impaired at these same levels by day 4, and succumbed to the Se by day 7. Fathead minnow growth was also impaired at these two concentrations by day 7. In addition, sub lethal impairment of rotifer and fish growth was evident at the 110.3 pg/L level after day 20 indicating a more subtle trophic impact. Bioconcentration factors ranged between 100 and 1000 µg/L and were found to be dependent on the species, treatment, and day. / Ph. D.