Predicting invertebrate herbivory from plant traits: evidence from 51 grassland species in experimental monocultures

Invertebrate herbivores can impact plant performance and plant communities. Conversely, plants can affect the ability of herbivores to find, choose, and consume them through their functional traits. While single plant traits have been related to rates of herbivory, most often involving single herbivore-plant pairs, much less is known about which suite of plant traits is important for determining herbivory for a pool of plant species interacting with a natural herbivore community. In this study we measured aboveground herbivore damage on 51 herbaceous species growing in monocultures of a grassland biodiversity experiment and collected 42 different plant traits representing four trait groups: physiological, morphological, phenological, and herbivore related. Using the method of random forests and multiple regression, we identified seven traits that are important predictors of herbivore damage (leaf nitrogen and lignin concentration, number of coleopteran and hemipteran herbivores potentially feeding on the plants, leaf life span, stem growth form, and root architecture); leaf nitrogen and lignin concentration were the two most important predictors. The final model accounted for 63% of the variation in herbivore damage. Traits from all four trait groups were selected, showing that a variety of plant characteristics can be statistically important when assessing folivory, including root traits. Our results emphasize that it is necessary to use a multivariate approach for identifying traits affecting complex ecological processes such as herbivory.