Quantitative evolutionary patterns in bipartite networks: Vicariance, phylogenetic tracking or diffuse co‐evolution?

Abstract

Many pairwise interactions in ecological communities are thought to be structured by co‐evolution, a process difficult to study on a community level. Traditional methods can reveal correlated phylogenies between interacting organisms, suggesting a co‐evolutionary association. However, several processes could lead to cophylogeny, including (1) vicariance, (2) phylogenetic tracking and (3) co‐evolution. We present a framework to investigate diffuse co‐evolution between groups of interacting taxa. Our methodology incorporates two complementary statistical methods to test for evolutionary signals. First, we test for correlated taxonomies between interacting species, and then we test for an association between the pattern of interactions and taxonomy of interacting species. We use these methods to clearly distinguish between three possible explanations of cophylogenies, when present. Our methodology is designed to explore diffuse evolutionary signals among many interacting taxa: it is robust to polytomies and generalization and can use taxonomic trees in place of phylogenies. Thus, we use taxonomic distances to identify groups of associated taxa, as opposed to specific co‐evolutionary relationships. In a case study, we find significant evidence for phylogenetic tracking, vicariance and asymmetry. Generally, our method can be applied whenever relevant taxonomy is well resolved and includes sufficient variation at multiple taxonomic ranks between two groups of interacting taxa. This adaptable method can incorporate many types of data on species traits or behaviours, such as phylogenetic relatedness, pollination efficacy, interaction strengths and visitation rates. The analysis we provide here may be of use to a broad range of ecologists interested in evaluating diffuse evolutionary patterns in groups of interacting species.

Should I stay or should I go? Pollinator shifts rather than cospeciation dominate the evolutionary history of South African Rediviva bees and their Diascia host plants

“…29 Ma, 95% HPD = 19–40 Ma; Kahnt et al, ). Moreover, our findings are in agreement with Russo et al (), who analysed 13,106 unique plant–pollinator interactions and found that vicariance or phylogenetic tracking of one taxon by its interaction partner underlay cospeciation events more often than co‐evolution.…”

“…However, cospeciation might result either (a) from the mutual interaction and reciprocal evolution of interacting populations (strict co‐evolution), (b) from a common biogeographic event that causes the subdivision of the distribution range of both interaction partners which causes them to simultaneously speciate (vicarance), or (c) when one taxon tracks the evolution of its partner and speciates in response to speciation events of its partner (phylogenetic tracking, De Vienne et al, ). We used the approach suggested by Russo, Miller, Tooker, Bjornstad, and Shea () to distinguish between these three alternative hypotheses, (a) co‐evolution, (b) vicariance or (c) phylogenetic tracking, as underlying cause for phylogenetic congruence and cospeciation. Specifically, Russo et al () used a Mantel test to investigate whether phylogenetic relatedness is correlated with ecological similarity, i.e., whether closely related pollinator or plant species are more likely to interact with similar partners than expected by chance.…”

“…We used the approach suggested by Russo, Miller, Tooker, Bjornstad, and Shea () to distinguish between these three alternative hypotheses, (a) co‐evolution, (b) vicariance or (c) phylogenetic tracking, as underlying cause for phylogenetic congruence and cospeciation. Specifically, Russo et al () used a Mantel test to investigate whether phylogenetic relatedness is correlated with ecological similarity, i.e., whether closely related pollinator or plant species are more likely to interact with similar partners than expected by chance. In the case of co‐evolution, one expects a significant cophylogenetic signal (i.e., overall tree congruence) as well as a significant correlation between phylogenetic distance and ecological dissimilarity for both the plant and the pollinator, as the interaction with the partner is important in the evolution of both taxa (Russo et al, ).…”

“…Specifically, Russo et al () used a Mantel test to investigate whether phylogenetic relatedness is correlated with ecological similarity, i.e., whether closely related pollinator or plant species are more likely to interact with similar partners than expected by chance. In the case of co‐evolution, one expects a significant cophylogenetic signal (i.e., overall tree congruence) as well as a significant correlation between phylogenetic distance and ecological dissimilarity for both the plant and the pollinator, as the interaction with the partner is important in the evolution of both taxa (Russo et al, ). If the phylogenies are congruent but the Mantel test is only significant for one taxon, this implies phylogenetic tracking since interactions are only important for one partner (the taxon with a significant Mantel test output), plant or pollinator (Russo et al, ).…”

“…We used the approach suggested by Russo, Miller, Tooker, Bjornstad, and Shea (2018) to distinguish between these three alternative hypotheses, (a) co-evolution, (b) vicariance or (c) phylogenetic tracking, as underlying cause for phylogenetic congruence and cospeciation. Specifically, Russo et al (2018) used a Mantel test to investigate whether phylogenetic relatedness is correlated with ecological similarity, i.e., whether closely related pollinator or plant species are more likely to interact with similar partners than expected by chance.…”