University of California Merced, California, United States
Pollination is a primary process impacting biodiversity and may play an important role in restoration efforts. Though mutualistic interactions are essential for establishing local communities, our understanding of how mutualistic networks assemble is limited, especially from an evolutionary perspective. Coevolution is the reciprocal adaptation between interacting species, and within mutualistic systems can lead to trait matching between species of different guilds, potentially even changing the structure of interactions among species. But how does coevolution shape species traits as mutualistic systems assemble? And how do these traits feed-back to shape the structure of the assembling community? Here we aim to explore the interplay between colonization, extinction, and coevolution, and how these intersecting dynamics shape species’ traits and the structure of mutualistic networks. Integrating fundamental concepts from Island Biogeography Theory with coevolutionary dynamics, we explore this feedback with a stochastic dynamic model. Using empirical pollination networks to capture interaction structure, we explore how colonization and extinction dynamics impact the ability of mutualistic partners to trait-match in a complex evolving community.
As we increase the colonization rate relative to the extinction rate, we observe that community richness stochastically fluctuates around increasingly higher values, where the amplitude of these fluctuation is dependent on the structure of the pollination network. Additionally, our result suggests that the amount of fluctuation a network has is similar for different values of colonization rate, which might indicate that the stability of the networks does not change depending on the proximity of the island. Moreover, we show that when extinctions are random, species’ persistence within assembling communities quickly reaches a threshold as a function of the number of its mutualistic partners. This suggests that the benefits of generalization are achieved with relatively few partners (roughly 5), regardless of which pollination network is evaluated. We suggest that the adaptive benefits of extreme generalization may only be realized if extinctions are non-random. We finish by showing how extinctions dependent on trait-matching impacts the role of generalization within mutualistic systems, and how our model results lend specific insights into the processes governing evolving mutualistic communities.