Effects of an organic contaminant (a synthetic coal-derived crude oil) were measured in outdoor ponds and indoor pond-derived microcosms and compared with results of laboratory bioassays. Ponds and microcosms were treated with the oil continuously for eight weeks. Concentrations of phenolic compounds (the major water-soluble constituents of the oil) spanned the range of acute and chronic toxicity concentrations determined in single-species bioassays. Effects were similar in microcosms and ponds, implying that microcosms are suitable models for field studies for some purposes. Significant changes in community metabolism and zooplankton populations occurred in microcosms and ponds exposed to less than 0.05 mg/L phenols, near the 28-day lowest observed effect concentration (LOEC) for Daphnia magna. Ponds and microcosms were seriously damaged at concentrations near acute bioassay mean lethal concentration (LC50) values. Indirect effects in the ecosystems occurred at all treatment levels, and included changes in water quality, replacement of sensitive taxa by more tolerant competitors, and changes in abundance of some species because of increases or decreases in their predators or grazers. The safe exposure level determined from the ecosystem experiments was accurately predicted by an application factor of 0.03 in conjunction with the most sensitive acute bioassay result (the D. magna 48-h LC50). Less conservative extrapolation methods overestimated the safe concentration of this material in these ecosystems.