Quantifying non-additive selection on traits caused by indirect ecological effects among species

Background/Question/Methods

In natural biological communities, species interact with many other species. Multiple species interactions lead to indirect ecological effects that have important fitness consequences and can also result in non-additive patterns of natural selection. Given that ecological indirect effects are common in nature, non-additive selection may also be quite common, and thus may be critical for predicting evolution in natural communities composed of many interacting species. Here we develop a clear method for testing for non-additive selection and consider how it might affect adaptation in multispecies communities.

Results/Conclusions

We tested our null hypothesis for non-additive selection using a simulation model. The model validates our null hypothesis, but importantly, indicates that fitness must be standardized across, rather than within, experimental treatments. We used our approach to quantify the strength of non-additive selection relative to pairwise selection in two case studies. In one case, insect herbivores and an invasive plant species impose strong non-additive selection on herbivore resistance in a native plant species. In a second case, three pollinators impose strong non-additive selection on anther exsertion in wild radish. Our results suggest that non-additive selection may be common in nature, and may often be just as important as pairwise selection gradients in determining how traits will evolve in natural communities. We call for further studies using the methodology outlined here to determine how common non-additive selection is and under what conditions it is most likely to occur.