Nutrient-insecticide interactions decouple density-dependent predation pressure in aquatic insects

Subsidy-stress responses can decouple density-dependent relationships between taxa within a community by increasing resources for some taxa while removing sensitive species. In this study, we examined the effects of a triad of subsidy-stress responses on predator-prey relationships in a benthic macroinvertebrate community under the action of insecticides, nutrients and predation pressure all at environmentally relevant doses. We stocked 72 outdoor artificial streams with benthic invertebrates to establish a fully factorial experiment to investigate the interactive effects of two anthropogenic factors: (i) three levels of nutrient enrichment (low, moderate and high); and (ii) three doses of imidacloprid, a common agricultural insecticide, at control, low (LOEC) and lethal doses (LC50). These factors were crossed with (iii) another, biotic stressor by stocking half of the streams with a predaceous perlid stonefly at near-natural densities. Fifty percent of the community-level responses were driven by the combined effect of nutrients and the insecticide. An additional 27% of the community variation was best explained by predation pressure, which we estimated as the gape widthxdensity of Agnetina capitata at the end of the experiment. No negative effect of imidacloprid on A.capitata was detected (P=0.62). By the end of the 3-week experiment, the overall gape width and predation pressure of A.capitata had more than doubled (2.1-3.7x) in all moderate nutrient streams (P<0.01). Increased size and predation pressure of A.capitata covaried with the depressed density of all macroinvertebrate taxa as well as EPT (Ephemeroptera, Plecoptera, Trichoptera) taxa and the mayfly Baetis in particular. Structural equationmodels (SEM) highlighted that nutrient enrichment was the main driver of individual and community responses in the artificial streams. Additionally, SEM showed that density but not body size of Baetis was severely hampered by A.capitata predation, suggesting that these stoneflies selectively foraged on other species in the first 10days of the experiment. Our findings highlight the difficulties of assessing the risk of contaminants in aquatic communities where interacting species respond differently to subsidies and stresses.