Variability of defensive chemicals in plants and their effects on herbivore behavior

Plants produce a wide variety of chemicals to defend themselves from herbivores and pathogens. While other researchers have noticed that the amount of these chemicals is high variable, few have explored the precise patterns and scales of variability in plant chemical defenses. In my work, I examined patterns of variability in the glucosinolates of wild radish (Raphanus sativus) and explored the implications of this variability on the fitness and behavior of the lepidopteran herbivore, Trichoplusia ni . I examined the spatial variability of these chemicals across multiple scales: (1) among plants with and without prior damage by herbivores, (2) among different plants, (3) among leaves of different ages on a plant, (4) among different leaf parts of a leaf, and (5) within individual parts of a leaf. I also used these data to parameterize mathematical models that explore the effects of this variability on herbivores. This dissertation includes several novel and important findings that provide a new framework in which to consider plant defenses and their interactions with herbivores. First, I found plants increase their mean defenses by increasing the variability, specifically by adding high toxin levels. This same pattern repeats at multiple scales. Second, I found that the highest proportion of variability in plant defenses occurs at the smallest scale (1–2 cm) and the spatial distribution of different concentrations of glucosinolates at this scale is random. Third, I found this small-scale variability is predicted to have the strongest effect of any scale of within-plant variability on small herbivores. These conclusions suggest that plant defenses are like a minefield. Most parts of a plant are only weakly defended but small patches have very high levels of defense, and the location of these highly defended patches is random. However, plants still defend parts with higher risk of herbivory or higher value to the plant more than other parts. I propose that plants use variability of defenses to create uncertainty for herbivores, making it impossible for the herbivores to avoid defenses in ecological time, or to evolve resistance over evolutionary time.