Interactions retain the co‐phylogenetic matching that communities lost

Timothée Poisot,Daniel B Stouffer

doi:10.1111/oik.03788

Abstract

Both species and their interactions are affected by changes that occur at evolutionary time‐scales, and these changes shape both ecological communities and their phylogenetic structure. That said, extant ecological community structure is contingent upon random chance, environmental filters and local effects. It is therefore unclear how much ecological signal local communities should retain. Here we show that, in a host–parasite system where species interactions vary substantially over a continental gradient, the ecological significance of individual interactions is maintained across different scales. Notably, this occurs despite the fact that observed community variation at the local scale frequently tends to weaken or remove community‐wide phylogenetic signal. When considered in terms of the interplay between community ecology and coevolutionary theory, our results demonstrate that individual interactions are capable and indeed likely to show a consistent signature of past evolutionary history even when woven into communities that do not.

Highlights

Ecological interactions often exert important selective pressures on the species involved
Many of the systems that have been described as exhibiting signi cant phylogenetic structure of interactions deviate from this last constraint, and this can occur for a variety of factors that stem from how other species evolved and established, lost, or maintained interactions throughout their joint evolutionary history
Detecting matching phylogenies for interacting clades indicates that their shared evolutionary history is long standing and is suggestive that their extant ecological structure is an outcome of ancestral constraints and/or co-adaptation (Nuismer & Harmon )

Summary

Local network Significant q Non significant qqqqqqqq

Figure Spatial distribution of co-phylogenetic matching across the sites. For each location, we indicate whether or not the structure of regional and local interaction networks is consistent with phylogenetic congruence. One can expect community assembly to promote the cooccurrence of evolutionarily linked species pairs – i.e., a host and a parasite from lineages that interact will tend to co-occur more often because the parasites are ltered to be present in sites where they can nd hosts. ), we see that the relationship between the two is approximately linear ( % con dence interval for the correlation coe cient 0.914–0.971) This ts with the hypothesis of local networks being assembled by a random sampling from regional networks: if interactions between species at matching positions in both trees are maintained by the same set of drivers, this should be re ected in the local networks by a higher degree of co-phylogenetic matching. Relative amount of co−phylogenetic signal 0.00 0.05 0.10 0.15 0.20 0.25 0.30 a q q

Continental Regional Local

None Local only Regional only Both q q