Community ecologists are just beginning to study interaction webs that quantify direct and indirect, as well as trophic and non-trophic, interactions.  Until recently, ecological interaction studies have focused on food webs, but we must understand why the dynamics of traditional food web models change when a variety of interaction types are included.   Analyzing specific subsets of the interaction web allows us to identify important species and to quantify the relative strengths of interaction pathways, which are some of the mechanisms that govern community organization.  Mutualistic, non-trophic interactions between ants and honeydew-producing hemipterans have strong community-level effects and have therefore been called “keystone interactions.”  In 2007 and 2008, we experimentally manipulated this interaction, with two objectives:  (1) to determine how an ant/honeydew mutualism affects the abundances of arthropods on rabbitbrush (Chrysothamnus viscidiflorus), and (2) to determine the relative effects of the mutualism on hostplant growth and reproduction.  We manipulated the abundance of ants (Formica obscuripes), honeydew-producing treehoppers (Publilia modesta), and herbivorous beetles (Monoxia schyzonycha) to quantify the interactions among arthropods occurring on rabbitbrush hostplants in meadows near the RMBL, Colorado.  Surveys were conducted throughout the growing season, where abundances for all arthropods and plant herbivory, growth, and reproduction were measured.  

Results/Conclusions

The 2007 results show that ants indirectly benefit plants by deterring beetle herbivory.  This conclusion was not supported in 2008, when beetle densities were very low.  In 2008, ants and treehoppers significantly affected the abundances of phloem feeders, mirid hemipterans, herbivorous lepidopterans and predatory spiders.  Also, ants and treehoppers had negative effects on two plant performance traits, growth and seed production. This result is surprising, given the number of studies that have shown positive or neutral effects of ant/honeydew mutualisms on hostplants.  Multivariate analyses (CCA and SEM) indicate that a significant amount of the variance in the arthropod community can be explained by the presence of the ant/treehopper mutualism.  Though the interaction is important for community organization, the relatively high abundances of treehoppers and ants suggest that the organisms in this study do not constitute a “keystone interaction.”