Comments to Dorresteijn et al. (2015)Dorresteijn et al. [1] recently addressed the question what influence humans have on predator-prey interactions in human-modified landscapes. Their study follows increasing evidence from relatively natural landscapes that large carnivores are central drivers of ecosystem structure and functioning through density-mediated and behaviourally mediated effects on their prey [2]. A key question is whether large carnivores will have similar ecological effects in human-dominated landscapes [3]. This is particularly relevant in the light of the carnivore recolonization of European landscapes [4]. We applaud Dorresteijn et al. [1] for highlighting that ecological effects of large carnivores may be quite different in human-modified landscapes. Unfortunately, we feel that there are crucial flaws with their study design, analyses and interpretation. These flaws seriously undermine the strong conclusions that they draw about the role of topdown and bottom-up processes in anthropogenic landscapes. Our main concerns include (i) lack of relevant bottom-up data, (ii) a camera trap design unsuitable for testing their hypotheses regarding community structure, and (iii) top-down and bottom-up processes that are studied at an inappropriate spatial scale.(1) Their main aim was 'to quantify the relative effects of top-down versus bottom-up processes in shaping predator and prey distributions'. It is surprising that the authors articulate this main objective without including actual bottom-up drivers in their analyses. Bottom-up processes were captured by variation in two land cover types, forest versus pasture cover, using the Corine Land Cover data. However, these very coarse categories do not reflect the variation in food resources (i.e. bottom-up processes) that drive the distributions of prey and predator. This variation occurs at finer spatial resolution, nested within the main land cover types that the authors used. For example, inside forest habitat (where all cameras were placed), one important fine-scale factor that drives deer distribution, and hence camera trapping rates, is variation in canopy openness, with deer strongly concentrating inside canopy gaps [5]. In addition, fox abundance (the mesopredator in this study) is driven by variation in prey availability that is linked to microhabitat variation [6]. Therefore, the Corine Land Cover data alone provide no basis for testing the importance of bottom-up effects in driving encounter rates of mesopredators and herbivores. In fact, we would argue that such land cover data may be a better proxy for top-down forces. Ungulate prey species avoid open or closed habitat depending on whether they respond to a large carnivore or human hunter [7]. Although the authors draw conclusions about bottomup effects, we argue that these are unfounded owing to an absence of relevant bottom-up variables in their models. (2) The main results and conclusions of this paper are based on the use of camera traps to estimate the relative occurrence of prey and pred...