It is often assumed that bioassays are better descriptors of sediment toxicity than toxicant concentrations and that ecological factors are more important than toxicants in structuring macroinvertebrate communities. In the period 1992 to 1995, data were collected in the enclosed Rhine-Meuse delta, The Netherlands, on macroinvertebrates, sediment toxicity, sediment contaminant concentrations, and ecological factors. The effect of various groups of pollutants (polycyclic aromatic hydrocarbons, trace metals, oil, polychlorinated biphenyls) and of ecological variables on the structure of the macroinvertebrate community were quantified. Ecological factors explained 17.3% of the macroinvertebrate variation, while contaminants explained 13.8%. Another 14.7% was explained by the covariation between ecological variables and contaminants. Polycyclic aromatic hydrocarbons explained a larger part of the variation than trace metals. The contributions of oil and polychlorinated biphenyls were small but significant. Elevated contaminant concentrations were significantly associated with differences in the macroinvertebrate food web structure. The response in bioassays (Vibrio fischeri, Daphnia magna, Chironomus riparius) was susceptible to certain contaminants but also to certain ecological factors. There was a weak correlation between in situ species composition and bioassays; 1.9% of in situ macroinvertebrate variation was explained by the bioassay responses. This seems to contradict the validity of using bioassays for a system-oriented risk assessment. Possible reasons for this discrepancy might be the manipulations of the sediment before the test and a higher pollutant tolerance of the in situ macroinvertebrates. Thus, macroinvertebrate field surveys and laboratory bioassays yield different types of information on ecotoxicological effects, and both are recommended in sediment risk assessment procedures.