In the 1990s the Dutch government expressed the need to investigate the impacts of diffuse pollution at (sub)-ecosystem levels. The resulting Netherlands Stimulation Programme on System-oriented Ecotoxicological Research (SSEO programme) ran from 1998 to 2006. Its primary objective was to assess the impacts of low- to medium-level, diffuse, multiple contaminations on ecosystems. The research results were intended as underpinning for policies on environmental, conservation and nature issues. Research was carried out at three sites that were selected because of their importance for nature management and the presence of diffuse contamination. These sites were: a river meadow/floodplain area (Afferdensche en Deestsche Waarden), an estuarine reed-land area (Biesbosch) and an area of lowland peat soils that had been contaminated with urban waste in past centuries (De Ronde Venen). This introductory paper describes the set-up of the programme, the types of diffuse contamination, the interactions between pollutants and other stress factors, the various methodologies used to integrate the effects on (sub)ecosystem level, and the consequences for formulating policies for and the management of these types of locations. The results of the programme are diverse and complicated and show how difficult it is to draw firm, unambiguous, generic conclusions about the effects at the 'total' ecosystem level. It is however, possible to draw conclusions about effects on major components of ecosystems:-The distribution of contaminants, both from a spatial, chemical and ecotoxicological point of view, plays a decisive role in actual effect levels. Even when total contaminant loads are high, such as in estuarine and floodplain areas, bioavailability may be so low that the actual effects are limited. The irregular, heterogeneous, spatial distribution of contaminants in the soil further complicates effect studies, impact assessments and monitoring.-Various stress factors, other than contaminants, both natural and anthropogenic, also play a role. The negative effect of the repeated inundation of floodplain areas, for instance, greatly interferes with the impact of contaminants in the lower soil layers.-A major problem is to find a method to extrapolate the observations from individual and population levels to the ecosystem level. In addition to traditional food-chain models and similar approaches, the potential of other, not yet extensively explored, ecosystem interaction mechanisms is discussed.-Finally, the results have to be interpreted from a policy point of view, both for national soil policies and for implementing the EU Soil Strategy regulations. © 2008 Elsevier B.V. All rights reserved.