摘要：Examining the effects of pesticides is difficult in regions such as Western Europe because of the relatively ubiquitous use of agrochemicals and the lack of unaffected areas. To study a wide gradient of agrochemical stress, we conducted a study in Central Romania, where traditional agriculture, which is assumed to use less agrochemicals, exists adjacent to intensive agriculture. We investigated potential effects of land use related stressors including pesticides on aquatic–terrestrial predator–prey relationships using stable isotope analysis. Therefore, we sampled spiders, as well as their aquatic and terrestrial prey along streams with a pesticide pollution gradient and determined spider and prey stable carbon and nitrogen signals. Aquatic prey contributed 40.8–55.4% to the diet of the orb web weaving spider Tetragnatha sp. and 34.0–53.0% to the diet of the ground-hunting Pardosa sp. The biomass of potential aquatic prey increased along a gradient of increasing riparian habitat conditions and water quality and decreasing agriculture (e.g. arable land) in the catchment. The proportion of aquatic prey in the orb web weavers diet responded positively to the biomass of potential aquatic prey and negatively to this gradient. Increasing potential prey biomass resulted in an increasing proportion of aquatic prey for the orb web weaver. The proportion of aquatic prey in the ground hunters diet increased with in-stream pesticide toxicity and along a gradient of increasing pastoralism in the catchment as well as increasing riparian habitat and water quality. The diet of two riparian predators responded to complex gradients of catchment land use, stream and riparian habitat quality as well as to in-stream pesticide toxicity. The responses to the different environmental variables can be attributed to direct (e.g. change in spider community) and indirect (e.g. changes in aquatic prey and terrestrial prey) effects. Future manipulative field studies or experiments on aquatic–terrestrial food webs are required to examine the causality of our findings and should also consider the quality of prey organisms to foster mechanistic understanding of cross-ecosystem effects..
