The distribution of herbivory has long been noted for its variability. This heterogeneity has long been a source of speculation and is often used as evidence for a wide array of biological hypotheses, including induced defenses, plant communication, herbivore movement, variable selection regimes, and variable plant qualities. Yet, little is known about how different herbivory patterns are generated. Understanding the distribution of herbivory is important because it has profound impacts on plant population dynamics, stress tolerance, and crop yield.
Inspired by the seminal paper of Edwards & Wratten (1983), I investigate herbivory patterns among species, among individual plants, among leaves, and within leaves. To do so, I rely on methods ranging from small scale experiments to global data synthesis, with an emphasis on using and developing novel quantitative tools for studying herbivory. I focus on two primary questions: (1) How do plant traits and herbivore behavior affect herbivory distribution? (2) How do different patterns of herbivory distribution affect plant fitness and population dynamics?
As a start, I developed a neutral model of herbivory from first principles to separate herbivory patterns caused by hypothesized biological mechanisms from purely neutral stochastic process. Comparison of the neutral model to an extensive global data set of herbivory from the HerbVar Network revealed that neutral stochastic processes dominate natural variation in herbivory, but deviations from neutral patterns are detectable and predictable (Pan et al. 2024). I also developed a high throughput leaf image processing pipeline to extract within leaf herbivory patterns. Using this pipeline, I found that tissue level herbivory tends to be more spatially aggregated when overall food quality is higher, consistent with optimal foraging theory. These tools are available in my R package herbivar with more capabilities to come.