Robert M. Pringle1, Corinna Riginos2, Brian F. Allan3, Jacob R. Goheen4, Felicia Keesing5, Douglas J. McCauley1, Wilfred Odadi6, Todd M. Palmer7, Maureen L. Stanton2, and Truman P. Young2. (1) Stanford University, (2) University of California, (3) Washington University, (4) University of British Columbia, (5) Bard College, (6) Egerton University, (7) University of Florida
Complex networks of direct and indirect interactions underpin the structure and function of ecological communities, and the activities of a few strongly interacting species may contribute disproportionately to the maintenance of these networks. Large herbivores have been identified as strong interactors in many ecosystems, and a growing number of exclusion experiments has addressed the direct impacts of large herbivores on plant communities. Many fewer studies, however, have examined indirect effects initiated by large herbivores. The Kenya Long-term Exclosure Experiment has excluded different groups of large herbivores from replicated 4-ha plots since 1995, allowing us to investigate whole-community responses to the selective elimination of different large-herbivore subgroups (e.g., elephants and giraffe vs. cattle vs. native equids and bovids). We provide a synthetic overview of selected results from this experiment, focusing on a suite of indirect interactions initiated by large herbivores. These include: (1) interactions among different large-herbivore species and their consequences for the plant community; (2) interaction cascades linking large herbivores to smaller vertebrates; and (3) relaxation of plant defenses and the breakdown of a multi-species ant-plant mutualism following large-herbivore exclusion. These interactions encompass density- and trait-mediated processes that occasionally work in opposite directions, suggesting that the net effects of large-herbivore exclusion might often comprise disparate components that investigators must disentangle via piecemeal experimentation. Collectively, these results confirm the keystone role of large herbivores in African savannas and allow us to generate several testable hypotheses about the ecological consequences of past and pending megaherbivore extinctions.