Abstract
Aim Biotic interactions have traditionally been predicted to be stronger towards the tropics. However, all previous studies about the latitudinal gradient in seed predation were either based on single species or compiled data from studies that used different methods in different ecosystems. Our goal was to provide the first broad‐scale quantification of the latitudinal gradient in seed predation at both cross‐species and within‐species levels. Location Twenty‐five sites spanning 28° of latitude on the east coast of Australia. Methods Using consistent protocols, we measured pre‐dispersal seed predation for 256 species × site combinations (including a total of 170 species) and post‐dispersal seed removal for 126 species × site combinations (including a total of 91 species), as well as removal of standard rice grains at each site. The relationships between seed predation and latitude were quantified using generalized mixed‐effects models and meta‐analyses. Results Contrary to expectations, the proportion of seeds lost to pre‐dispersal seed predation increased with latitude at the cross‐species level and showed no latitudinal trend at the within‐species level. There was no significant relationship between latitude and the proportion of seeds lost to post‐dispersal seed removal, either within species or across species. We found significantly higher removal of standard rice grains at lower latitudes, but this trend reversed if we excluded the four island sites from analysis. Main conclusions Seed predation on natural seeds was not higher at lower latitudes. Thus, our study joins a growing body of evidence contrary to the idea that tropical species suffer greater losses to herbivores than do temperate species. The latitudinal gradient in removal rates for standard rice grains suggests that natural seeds from low latitudes may have achieved a similar predation rate to high‐latitudinal species through increased investment in defenses. However, additional data will be necessary to test this hypothesis.