COS 35-5
Mongolian steppe pollination: Hymenoptera and Diptera contribute differentially to network specialization and nestedness

Tuesday, August 11, 2015: 9:20 AM
348, Baltimore Convention Center
Daniel S. Song, Department of Biology, PIRE Mongolia Project (, University of Pennsylvania, Philadelphia, PA
Pierre Liancourt, Institute of Botany, PIRE Mongolia Project (, Academy of Sciences of the Czech Republic, Praha, Czech Republic
Bazartseren Boldgiv, Department of Biology, National University of Mongolia, Ulaanbaatar, Mongolia
Laura A. Spence, Faculty in Ecology, PIRE Mongolia Project (, Sterling College, Craftsbury, VT
Peter Petraitis, Biology, University of Pennsylvania, Philadelphia, PA
Brenda B. Casper, Department of Biology, University of Pennsylvania, Philadelphia, PA

Ecological services are influenced by the complexity and stability of interspecific interactions.  This is especially true for insect pollination services.  Understanding the factors that structure networks of plant-pollinator interactions can provide insights to whether and how disturbances will alter them and the services they provide.  We worked with insect pollinator communities at two south-facing locations in the steppe of northern Mongolia.  We constructed plant-pollinator networks from observed visits and quantified two measures contributing to their stability: (1) the degree plants or pollinators interacted with a few or many species, or network specialization and (2) a basic organizational metric known as nestedness.  We constructed multiple networks for the two locations at different times within the summer; we examined numerous variables to determine which best predicted specialization and nestedness: forb species and insect genus richness, network size, flower abundance, total insect visits, and visits made by either Hymenoptera, Diptera, or Lepidoptera.  We used corrected Akaike information criterion to select the best-supported explanatory model: forb species richness and visits made by Hymenoptera and Diptera.  We then used path analysis to test the relationship between these variables and each of the two network indices.


We found Hymenoptera visits were correlated with network specialization (β = 0.70) and not nestedness; Diptera visits were positively related to nestedness (β = 0.69) and not specialization.   Additionally, network specialization varied within the summer and between locations on the south-facing slope.  Nestedness was greater at the lower slope location compared to the higher but did not differ throughout the summer.  Thus, our results show that not all plant-pollinator interactions contribute equally to important aspects of mutualistic network topology.  Generalist pollinators increase network robustness against perturbations while increased network specialization may limit ability to recover from perturbations.  We observed two groups of pollinators contributing differentially to two different measures related to stability.  Furthermore, these network measures varied temporally and spatially within the same system.  Relating complex network topological indices to specific ecological factors allows for better understanding of network structure and its response to potential changes in pollinator or plant species abundances.  Different insect or forb groups may act in concert to either weaken or strengthen network stability.  In the Mongolian steppe network stability may be more sensitive to changes in Hymenoptera abundance at higher elevations and to changes in Diptera abundance at lower elevations.