Kailen A. Mooney, University of California, Daniel S. Gruner, University of Maryland, Nicholas A. Barber, University of Missouri-St. Louis, Stacy M. Philpott, University of Toledo, Sunshine A. Van Bael, Smithsonian Tropical Research Institute, and Russell Greenberg, Smithsonian Migratory Bird Center.
Background/Question/Methods Insectivorous vertebrates consume both herbivorous and carnivorous arthropods, and in so doing do not occupy a clearly defined trophic position. Because insectivores simultaneously feed upon both herbivores and the enemies of herbivores, it is unclear whether insectivores have net effects on herbivore abundance or an indirect cascading effect on plants. To quantify the multi-trophic effects of insectivores, we conducted a meta-analysis of terrestrial experiments in which birds, bats or lizards were experimentally excluded from plants and the responses of arthropods and/or plants were quantified. In total, we identified 65 appropriate publications. Many of these in turn yielded several multiple independent results (=studies) either by studying a predator type at multiple field sites, or by studying multiple predator types at individual sites. We used natural log response ratios (LLR=Ln[control/exclusion]) to quantify insectivore effects and noted several methodological and biological covariates associated with each response.
Results/Conclusions For the studies that comprehensively sampled arthropods (N=40), the mean effect of insectivores was LLR = -0.69. Breaking this into trophic guilds, the negative effects of insectivores on carnivorous and herbivorous arthropods were nearly identical at -0.49 (N=57) and -0.47 (N=96) reductions respectively. Furthermore, the ratio of predators: herbivores in a community was unchanged by insectivores (N=55). By in large, these effects on arthropods were not influenced by exclusion size, experimental duration, insectivore type, or field site characteristics such as latitude, precipitation or annual primary productivity as indicated by the Normalized Difference Vegetation Index (NDVI). Consequently, it can be inferred that vertebrate insectivores feed indiscriminately upon both herbivorous and carnivorous arthropods, and that these effects are relatively constant across ecological contexts. As for the indirect effects of insectivores on plants, overall there was no detectable effect (N=56). However, the cascading effects of insectivores on plants increased with site NDVI, such that insectivores significantly benefited plants at field sites of higher productivity. In summary, insectivores had strong negative effects on both carnivorous and herbivorous arthropods, and these effects cascaded to impact plants at field sites of high primary productivity.