Generic mixed-flask microcosms were used to evaluate ecosystem responses to aniline and three closely related compounds-2,6-diisopropylaniline, 4-hexyloxyaniline and 2,3,5,6-tetrachloroaniline. Toxicity was determined on both an acute and chronic basis using changes in ecosystem -level variables (pH and dissolved oxygen levels) as indicators of effect. These calculated toxicity values were then compared with reported toxicity data on bacteria, algae, protozoa and Cladocera to evaluate the relative sensitivity of the method. The relative toxicities of the tested compounds were the same in microcosm tests as in the available single-species tests, but the range between the most and least toxic as determined by the microcosm test was smaller by an order of magnitude. Aniline was the least toxic, followed, in order of increasing toxicity, by 2,6-diisopropylaniline, 4-hexyloxyaniline and 2,3,5,6-tetrachloroaniline. The minimum effect concentrations determined for these compounds were generally lower than reported literature values, suggesting that the method does not sacrifice sensitivity in providing an integrated picture of ecosystem-level effects.