Global patterns in post-dispersal seed removal by invertebrates and vertebrates

It is commonly accepted that species interactions such as granivory are more intense in the tropics. However, this has rarely been tested. A global dataset of post-dispersal seed removal by invertebrates and vertebrates for 79 native plant species from semi-natural and natural terrestrial habitats ranging from 55° N to 45° S, was compiled from the global literature to test the hypothesis that post-dispersal seed removal by invertebrates and vertebrates is more intense at lower latitudes. We also quantified the relationship between post-dispersal seed removal by vertebrates and by invertebrates to global climatic features including temperature, actual evapotranspiration (AET) and rainfall seasonality. Linear mixed effect models were applied to describe the relationships between seed removal and latitude, hemisphere and climatic variables controlling for the effect of seed mass. Post-dispersal seed removal by invertebrates was negatively related to latitude. In contrast, postdispersal seed removal by vertebrates was positively but weakly related to latitude. Mean annual temperature and actual evapotranspiration were positively related to post-dispersal seed removal by invertebrates, but not to post-dispersal seed removal by vertebrates, which was only marginally negatively related to rainfall seasonality. The inclusion of seed mass improved the fit of all models, but the term for seed mass was not significant in any model. Although a good climatic model for predicting post-dispersal seed predation by vertebrates at the global level was not found, our results suggest different and opposite latitudinal patterns of post-dispersal seed removal by invertebrates vs vertebrates. This is the first time that a negative relationship between post-dispersal seed removal by invertebrates and latitude, and a positive relationship with temperature and AET have been documented at a global-scale. These results have important implications for understanding global patterns in plant-animal interactions, and the factors that shape plant reproductive ecology, and also for predicting how this plant-animal interaction might respond to climate changeThis work was supported by the Spanish Ministry of Economy and Competition (Project CGL2011 24871), the Madrid Government (REMEDINAL2-S2009/AMB-1783), and the Spanish Ministry of Education (Salvador de Madariaga grant PR2011-0491 for the sabbatical of BP in the University of New South Wales). ATM was supported by a QEII fellowship from the Australian Research Counci