Reinterpreting phylogenetic patterns in communities

Examining the phylogenetic structure of a community in order to understand patterns of community assembly has become an increasingly popular approach. A quick web search of “community”, “phylogenetics”, and “ecology” finds several hundred papers, most written in the last ten years.

Eco-phylogeneticists examine how patterns of evolutionary relatedness within communities may reflect the processes structuring those communities. In particular, a commonly tested hypothesis is the competition-relatedness hypothesis, which suggests that more closely-related species having more similar niches and therefore stronger competitive interactions, making coexistence between them less likely. As a result, if competition is important, communities may exhibit phylogenetic overdispersion, with species being less related on average than if drawn randomly from the regional species pool. The contrasting pattern, phylogenetic clustering, where species tend to be more closely related than expected, is often interpreted as being the result of strong environmental filtering, such that only a closely related group of species, best adapted to that environment, surviving in the community.

Evidence for the competition-relatedness hypothesis has been mixed, and since most tests of this hypothesis focus on patterns in observed data, conclusions about the underlying mechanism driving community phylogenetic patterns are rarely testable, and yet widely made.

In Mayfield and Levine (2010, Ecology Letters), the authors critique the current ecological justification for the competition-relatedness hypothesis, noting that it does not agree with a more current view of the processes driving species coexistence. As established by Chesson (2000, Annual Review of Ecology and Systematics), coexistence can involve both stabilizing forces (niche differences between species), and equalizing forces (fitness differences between species). In a simplistic example, plants using different soil types (niche differences) may coexist, while plants with similar high growth rates may exclude those species with lower growth rates (fitness differences). The final community should reflect the interplay of both these processes.

The implication of this view of species coexistence is that there is no preconceived phylogenetic pattern which should reflect competition: if species with the highest heights are competitively superior and exclude other species (coexistence driven by fitness differences), and height is a phylogenetically conserved trait, the community will appear to be phylogenetically clustered. Traditionally, a clustered pattern would not be considered to indicate the effects of competition. In fact, Mayfield and Levine show that the expected phylogenetic pattern depends entirely on whether niche and/or fitness differences are important and/or related to phylogenetic distance.

This suggest that conclusions in past studies may need to be reinterpreted. It also adds to the list of assumptions about evolutionary relatedness and ecological function which need to be tested: for example, how do niche and fitness differences tend to change through time? Do they tend to be conserved among closely related species? Does one or the other tend to dominate as a driver of coexistence in different systems? If nothing else, we need to be careful about making generalizations which don’t account for the differing evolutionary history, geographical location, and ecological setting that communities experience, when interpreting observed patterns in those communities.