Experimental test of plant defence evolution in four species using long-term rabbit explosures

1. Plant defence traits have evolved over macro- and microevolutionary time-scales in response to herbivores. Although a number of studies have investigated the evolutionary impacts of herbivores over short time-scales, few studies have experimentally examined what defence traits most commonly evolve and whether multiple coexisting species exhibit similar evolutionary responses to herbivores. 2. We addressed these questions using a long-term experiment at Silwood Park, England, United Kingdom, where we excluded rabbits from 38 grassland plots for<1–34 years. To assess the evolutionary impacts of rabbits on plant defence traits, we collected seeds from plots containing the following perennial species: Anthoxanthum odoratum (Poaceae), Festuca rubra ssp. rubra (Poaceae), Holcus lanatus (Poaceae) and Stellaria graminea (Caryophyllaceae). We then grew these plants in a common garden and measured defensive and morphological traits. 3. We found some evidence for evolutionary change of defence traits in three of the four species following the exclusion of rabbits. We observed the clearest changes in F. rubra, which showed a 9% decline in tolerance to herbivory and a 26% decline in leaf number. We also observed weak evidence for a change in all grass species towards a more erect growth form suggesting that grazing selects for plants that grow close to the ground. 4. Although our results are most consistent with evolution due to changes in the frequency of alleles and genotypes, we cannot rule out that epigenetic changes (e.g. methylation) or maternal environmental effects also contributed to or caused the observed long-term phenotypic responses. 5. Synthesis. Our study provides an experimental test of the evolutionary effects of an ecologically important herbivore. We found evidence for plant defence evolution following >20 years of rabbit exclusion. However, the evidence was only strong in one species for multiple traits, weak in all three grass species for avoidance and absent in an herb species. This suggests that the evolutionary effects of an ecologically important herbivore on plants will be variable and difficult to predict in nature.