Does evolutionary history determine specificity in broad ecological interactions?

Most species engage in broad interactions, in which they interact with multiple partner species. The evolutionary processes that generate such generalized interactions remain unknown, partly due to the difficulty in comparing their breadth. We argue that the interaction specificity of species involved in broad interactions evolves in three ways: (a) assemblage specialization, in which a species specializes on particular host species that contribute unique resources, yielding specialization on the entire host assemblage; (b) apparent generalism, in which a species specializes on one or few host species that contribute unique resources, but also associates with other host species that contribute functionally redundant resources; and (c) true generalism, in which a species associates with multiple hosts that overlap functionally, and that are geographically interchangeable based on opportunity for encounter, leading to frequent host switching. We performed a phylogenetically controlled analysis of data on mycorrhizal fungal associations for approximately 25% of the orchid subfamily Cypripedioideae to determine whether these plants have specialized on their mycorrhizal fungal communities, or whether they are true generalists. We also assessed the impact of environmental factors on these associations. Our results suggested strong support for apparent generalism, suggesting strong specialization on particular, dominant fungi and weak specialization on others. Large orchid clades associated with dominant fungal species, notably Tulasnella cystidiophora for genus Cypripedium, and T. cystidiophora and T. calospora for genus Paphiopedilum. Significant phylogenetic signal in fungal species richness per plant species, but not in the fungal phylogenetic diversity per plant species nor in the composition of fungal assemblages across orchid species suggested that plant phylogeny is an important determinant of fungal association. Mixed linear models showed that environment influenced specificity across plant species, and that observed differences were strongly driven by differences in sampling effort. Synthesis. We found evidence of specialization of plant species on dominant fungal species, and to a lesser extent on their close relatives. The strong dominance of particular fungal species in these associations suggests important ecological roles for them, while environmental gradients in specificity suggest strong environmental filtering of these interactions.