Linking biodiversity to mutualistic networks – woody species and ectomycorrhizal fungi

Abstract

Mutualistic interactions are currently mapped by bipartite networks with particular architecture and properties. The mycorrhizae connect the trees and permit them to share resources, therefore relaxing the competition. Ectomycorrhizal macrofungi associated with woody species (Quercus robur, Q. cerris, Q. petraea, Tilia tomentosa, Carpinus betulus, Corylus avellana, and Q. pubescens) growing in a temperate, broadleaved mixed forest, from a hilly area near the city of Cluj–Napoca, central Romania were included in a bipartite mutualistic network. Community structure was investigated using several network metrics, modularity and nestedness algorithms in conjunction with C-score index cluster analysis and nonmetric multidimensional scaling (the Kulczynski similarity was index used as most appropriate metric selected by minimal stress criterion). The results indicate that the network presents high asymmetry (hosts are outnumbered by mycobionts at a great extent), high connectance, low modularity, and high nestedness, competition playing a secondary role in community assemblage (non significant difference between simulated and observed C-score). The nestedness pattern is non-random and is comparable to previously published results for other similar interactions containing plants. In the proposed network, woody species function exclusively as generalists. Modularity analysis is a finer tool were identifying species roles than centrality measures, however, the two types of algorithms permit the separation of species according to their roles as for example connectors (generalist species) and ultraperipheral species (specialists). Supergeneralist woody species function as hubs for the diverse ectomycorrhizal community while supergeneralist ectomycorrhizal fungi glue the hubs into a coherent aggregate.

Keywords

• temperate broadleaved mixed forest, ectomycorrhizal fungi, mutualistic bipartite networks, nestedness, modularity, connectance, c-score cluster analysis, non metric multidimensional scaling