Understanding the processes that drive the spatial dynamics of range expansion is a fundamental challenge in ecology. Recent theory has demonstrated that evolution can increase the speed of range expansion by selecting for increasingly dispersive individuals at the leading edge of the expansion. This process, termed “spatial selection”, occurs because phenotypes become spatially sorted by dispersal ability as the population spreads outwards, resulting in the aggregation and assortative mating of highly dispersive phenotypes at the leading edge. Additionally, these highly dispersive phenotypes are more likely to experience low conspecific densities, which can result in increased per-capita reproduction, and natural selection for increased reproductive rate. Because range expansion speed is determined by dispersal propensity and population growth rate, these evolutionary mechanisms are thought to accelerate the speed of range expansion over time.
We used laboratory mesocosms of the bean beetle Callosobruchus maculatus to test whether these evolutionary mechanisms generated evolutionarily increased invasion speeds compared to a treatment where they were suppressed. After 10 generations of experimental range expansion, we tested for evidence of evolved differences in dispersal and reproductive rate. Finally, we used simulations to test how genetic correlations between dispersal propensity and reproductive rate can alter these evolutionary outcomes.
Results/Conclusions
In our mesocosms experiments, we found that spatial evolutionary processes increased the mean expansion speed by 6.1%, but made invasion speeds 41-fold more variable overall. This increased variability was large enough to sometimes eliminate any increase in mean expansion speed; in fact, the two slowest experimental range expansions were ones in which spatial evolutionary processes were allowed to operate. Common garden experiments showed that beetles in the spatially evolving populations were more dispersive, on average, than beetles from populations where spatial evolutionary processes were suppressed. We did not find any evidence for the evolution of reproductive rate.
Although we did not measure covariance between dispersal and reproductive rate in C. maculatus, we simulated invasions over a range of potential genetic correlations, to see how these correlations modified rates of range expansion. We find that accounting for evolution while incorporating a range of potential genetic correlations can improve estimates of uncertainty windows for predicting range expansion speeds.