Eric Allan, Wolfgang W. Weisser, Markus Fischer, Ernst-Detlef Schulze, Alexandra Weigelt, Christiane Roscher, Jussi Baade, Romain L. Barnard, Holger Bessler, Nina Buchmann, Anne Ebeling, Nico Eisenhauer, Christof Engels, Alexander J. F. Fergus, Gerd Gleixner, Marlen Gubsch, Stefan Halle, Alexandra Maria Klein, Ilona Kertscher, Annely Kuu, Markus Lange, Xavier Le Roux, Sebastian T. Meyer, Varvara D. Migunova, Alexandru Milcu, Pascal A. Niklaus, Yvonne Oelmann, Esther Pasalic, Jana S. Petermann, Franck Poly, Tanja Rottstock, Alexander C. W. Sabais, Christoph Scherber, Michael Scherer-Lorenzen, Stefan Scheu, Sibylle Steinbeiss, Guido Schwichtenberg, Vicky Temperton, Teja Tscharntke, Winfried Voigt, Wolfgang Wilcke, Christian Wirth, Bernhard Schmid

• In order to predict which ecosystem functions are most at risk from biodiversity loss, meta-analyses have generalised results from biodiversity experiments over different sites and ecosystem types. In contrast, comparing the strength of biodiversity effects across a large number of ecosystem processes measured in a single experiment permits more direct comparisons. Here, we present an analysis of 418 separate measures of 38 ecosystem processes. Overall, 45 % of processes were significantly affected by plant species richness, suggesting that, while diversity affects a large number of processes not all respond to biodiversity. We therefore compared the strength of plant diversity effects between different categories of ecosystem processes, grouping processes according to the year of measurement, their biogeochemical cycle, trophic level and compartment (above- or belowground) and according to whether they were measures of biodiversity or other ecosystem processes, biotic or abiotic and static or dynamic. Overall, and for severalIn order to predict which ecosystem functions are most at risk from biodiversity loss, meta-analyses have generalised results from biodiversity experiments over different sites and ecosystem types. In contrast, comparing the strength of biodiversity effects across a large number of ecosystem processes measured in a single experiment permits more direct comparisons. Here, we present an analysis of 418 separate measures of 38 ecosystem processes. Overall, 45 % of processes were significantly affected by plant species richness, suggesting that, while diversity affects a large number of processes not all respond to biodiversity. We therefore compared the strength of plant diversity effects between different categories of ecosystem processes, grouping processes according to the year of measurement, their biogeochemical cycle, trophic level and compartment (above- or belowground) and according to whether they were measures of biodiversity or other ecosystem processes, biotic or abiotic and static or dynamic. Overall, and for several individual processes, we found that biodiversity effects became stronger over time. Measures of the carbon cycle were also affected more strongly by plant species richness than were the measures associated with the nitrogen cycle. Further, we found greater plant species richness effects on measures of biodiversity than on other processes. The differential effects of plant diversity on the various types of ecosystem processes indicate that future research and political effort should shift from a general debate about whether biodiversity loss impairs ecosystem functions to focussing on the specific functions of interest and ways to preserve them individually or in combination.…