PS 65-118
Interaction networks linking rare and invasive plant species – a double-edged sword?

Thursday, August 8, 2013
Exhibit Hall B, Minneapolis Convention Center
Diane L. Larson, Northern Prairie Wildlife Research Center, US Geological Survey, St. Paul, MN
Sam Droege, Patuxent Wildlife Research Center, U.S. Geological Survey, Beltsville, MD
Paul A. Rabie, WEST, Inc., Cheyenne, WY
Jelle Devalez, Aegean University, Greece
Jennifer L. Larson, Polistes Foundation
Margaret McDermott-Kubeczko, Department of Plant Pathology, University of Minnesota
Milton Haar, Badlands National Park, National Park Service, Interior, SD
Background/Question/Methods

Eriogonum visheri is a regional endemic found only in sparsely-vegetated badlands formation in western North and South Dakota and southeastern Montana, USA.   It is thought to be at risk from competition with encroaching exotic species such as Salsola iberica and Melilotus officinalis, especially during the seedling stage.  As an annual, E. visheri is dependent on a viable seed bank, so adequate pollination is a key requirement.  The goal of this study was to characterize the interaction network of flowers and flower-visitors surrounding population of E. visheri at Badlands National Park, South Dakota.  Specifically, we ask if the invasive plant species S. iberica and M. officinalis are playing a positive or negative role via their interactions with pollinators shared with E. visheri.  We established four permanent 1-ha plots centered on E. visheri populations in 2010 and sampled insect visitors to flowers on those plots over the course of the flowering period of E. visheri in 2010 and 2011.  In addition, pollen was removed from each captured insect and quantified.  The pollen-based and visitation-based networks were analyzed for basic properties using the Bipartite package in R statistical software, and for modularity using Netcarto (courtesy R. Guimera and colleagues).

Results/Conclusions

As expected, visitation networks in both years were more nested than expected by chance, suggesting that specialized species act preferentially with generalists.  Interaction evenness and diversity were either lower (visitation network in 2011, both pollen networks) or within the expected range.  All four networks were significantly modular.  Melilotus officinalis was among the most highly connected species in each network and shared a variety of mainly hymenopteran species with E. visheri; M. officinalis had high levels of connectivity with species both within and outside its module, a trait that is thought to lend stability to network structure.  Salsola iberica’s roles differed between years: it was well-connected in 2011, but less so in 2010.  Nonetheless, S. iberica shared several hymenopteran species with E. visheri in both years.  The consequence for E. visheri of sharing insect visitors with either invasive species was a greater likelihood of receiving nonconspecific pollen (based on pollen loads of shared pollinators that landed on E. visheri flowers).  It thus appears that the presence of M. officinalis and S. iberica may have positive consequences for network stability, but potentially negative effects on pollination of E. visheri.