Landscape context and pollinator composition
The specific landscape context may modulate the observed responses of pollinators. Although we are confident that similar findings (and correlations between landscape predictors) can be expected in other regions, we would like to highlight that our study area belongs to the most intensified landscapes in Europe (for comparison: 39 % of the total land area of the EU is cropland, around our study patches mean of 60%) with a long history of intensive fertilization and pesticides input. Therefore, we assume that the past land-use intensification already had a tremendous effect on the species pool in the area and the species composition in our study is only a small subset of the species pool of some decades ago. The comparison with previous studies in our study area (Hahn 2002, Saure and Berger 2006) indicates that on the one hand oligolectic bee species disappeared (see above), but also common hoverflies. We observed remarkably low ratios between hoverflies and wild bee individuals compared to studies of similar study systems (our study 1419 wild bees vs. 214 hoverflies, ratio = 0.15; Mudri-Stojnić et al. 2012: ratio 0.83, Jauker et al. 2009: ratio: 0.82, Jauker et al. 2019: ratio 0.95). In particular, generalist aphidophagous species (e.g.Eupeodes corollae , Sphaerophoria scripta ) that occur in high densities (e.g. Hahn 2002 for our landscape, Bankowska 1980), are underrepresented in our study, which concurs with the decline of common hoverfly species (Hallmann et al. 2021). These deviations are most likely a consequence of the high land-use intensity in our study region, which is the main factor for the current insect decline (Wagner et al. 2021) and question the hypothesis that hoverflies are less vulnearble than wild bees.