A model ecosystem experiment was conducted to investigate the ability of an in situ Daphnia magna feeding bioassay to assess impairment of ecosystem function. Animals were deployed in model ecosystems dosed with different concentrations of the fungicide carbendazim, and effects on the postexposure feeding rate of D. magna were compared with effects on zooplankton species richness (ecosystem structure) and development of phytoplankton biomass (ecosystem function). In the medium-dosed systems (21 μg/L), a structural change was observed within the zooplankton community, but no indirect effects on phytoplankton development were detected. It appears that at this treatment level, functional redundancy was sufficient to prevent functional impairment despite species loss. The feeding assay did not show any response at this concentration. In the high-dosed systems (221 μg/L), structural changes in the zooplankton community were accompanied by increased development of phytoplankton biomass. The feeding bioassay also showed a significant response at this concentration. At the high treatment level, species loss resulted in functional impairment, indicating that at this level, functional redundancy could not compensate for loss of individuals. The D. magna feeding bioassay matched well with the functional response (i.e., the indirect effects on phytoplankton) in the dosed systems but not with the more subtle effects on zooplankton community structure. These results lend positive support to the use of in situ feeding bioassays in combination with structural indices, such as species richness, to assess the effects of stress on ecosystem functioning in a direct way.