American Society of Naturalists

A membership society whose goal is to advance and to diffuse knowledge of organic evolution and other broad biological principles so as to enhance the conceptual unification of the biological sciences.

“Coevolution creates complex mosaics across large landscapes”

Posted on

Lucas D. Fernandes, Paula Lemos-Costa, Paulo R. Guimarães Jr., John N. Thompson, and Marcus A. M. de Aguiar (Aug 2019)

<i>Greya obscura</i> nectaring in <i>Lithophragma affine</i>. Photo taken at Pinnacles National Park (California, USA) in March 2016.<br />(Credit: Paula Lemos-Costa)
Greya obscura nectaring in Lithophragma affine. Photo taken at Pinnacles National Park (California, USA) in March 2016.
(Credit: Paula Lemos-Costa)

Does our understanding of the interplay between species interactions and patterns of diversity still apply when we go from local to regional or continental scales? It is now widely recognized that species interactions play a major role in influencing evolutionary rates, trajectories, and outcomes, but how does the spatial organization of these interactions affect the way species’ traits are distributed in space? Using mathematical models to describe coevolutionary dynamics in a landscape where the interactions vary from mutualism to antagonism, Fernandes and collaborators investigate how coevolution affects the spatial patterns of phenotypes, when considering large spatial scales and different structures of geographic mosaics of selection. One of the main results of this work is the formation of large clusters of phenotypes, in many cases much larger than the spatial aggregation of sites with the same interaction outcome. This result leads to the implications that (i) phenotype distributions can not be directly obtained from the outcomes of local interactions, and that (ii) local interactions can not be simply inferred from local patterns of trait distributions alone. Focusing on how the magnitude of selection, the spatial distribution of interactions and gene flow influence the spatial patterns of phenotypes, this work calls attention to the importance of comprehending coevolutionary dynamics in this context, and also the many ways in which these dynamics can be affected by human-driven habitat fragmentation in natural landscapes.


The spatial distribution of populations can influence the evolutionary outcome of species interactions. The variation in direction and strength of selection across local communities creates geographic selection mosaics that, when combined with gene flow and genomic processes such as genome duplication or hybridization, can fuel ongoing coevolution. A fundamental problem to solve is how coevolution proceeds when many populations that vary in their ecological outcomes are connected across large landscapes. Here we use a lattice model to explore this problem. Our results show that the complex interrelationships among the elements of the geographic mosaic of coevolution can lead to the formation of clusters of populations with similar phenotypes that are larger than expected by local selection. Our results indicate that neither the spatial distribution of phenotypes nor the spatial differences in magnitude and direction of selection alone dictate coevolutionary dynamics: the geographic mosaic of coevolution affects formation of phenotypic clusters, which in turn affect the spatial and temporal dynamics of coevolution. Because the formation of large phenotypic clusters depends on gene flow, we predict current habitat fragmentation will change the outcomes of geographic mosaics, coupling spatial patterns in selection and phenotypes.