Community-Level Consequences of Plant-Herbivore Interactions

Ecological systems are dynamic, yet many experimental studies examine plant-herbivore interactions as from a simple, static, or single perspective. Reciprocal interactions can have profound effects on communities, and ignoring such feedbacks can result in mismatches between theoretical predictions and experimental results. In this dissertation, I examined reciprocal interactions between two plant species, Solanum carolinense and Solidago altissima and their insect herbivores. In chapter 2, I examined how insect herbivores influenced plant competition and coexistence. Theory suggests that herbivores influence plant communities by altering competitive interactions. Because the outcome of competition is influenced by both the per capita competitive ability of plants and demographic processes including density dependence and intrinsic population growth rates, measuring herbivore effects on all these processes is necessary to understand the mechanisms by which herbivores influence plant communities. I fit alternative competition models to data from a response surface experiment conducted over four years to examine how herbivores affected the outcome of competition between two perennial plants, Solanum carolinense and Solidago altissima . Within a growing season, herbivores reduced Solanum plant size, but did not affect Solidago, which exhibited compensatory growth. Across seasons, herbivores did not affect the density of Solanum but reduced both the density and population growth of Solidago. The best fit models indicated that the effects of herbivores varied with year. In some years, herbivores increased the per capita competitive effects of Solidago on Solanum; in other years herbivores influenced the intrinsic population growth rates of Solidago. I examined herbivore effects on the longer-term outcome of competition (over the time-scale of a typical old-field habitat) using simulations based on the best fit models. In the absence of herbivores, plant coexistence was observed. In the presence of herbivores, Solanum was excluded by Solidago in 60% of the simulations. I demonstrated that herbivores can influence the outcome of competition through both changes in per capita competitive effects and changes in demographic processes. I discuss the implications of these results for ecological succession and biocontrol. In chapter 3, I examined how plant community composition influenced damage patterns on plants. Neighboring plants can increase (associational susceptibility) or decrease (associational resistance) the likelihood of damage to a focal plant but their long-term consequences for plant competition and coexistence are unclear. Neighbor effects on damage can occur through changes in the relative density of the focal plant (i.e., frequency of the focal plant), the absolute density of the focal plant, or through the total density of plants, because the different mechanisms known to influence damage patterns (e.g., pest suppression by predators, herbivore foraging behavior, plant quality) respond to different features of the neighborhood. To examine the long-term consequences of neighbor effects for plant communities, an understanding of how density and frequency of plants influence damage is needed. Using a response surface experimental design, I examined the effects of plant density and frequency on damage to Solanum carolinense. I found non-linear effects of the frequency of heterospecific neighbors (Solidago altissima) on Solanum damage, and a positive effect of Solanum density on damage. The non-linear pattern suggests that multiple mechanisms may be operating to influence damage. Non-linear patterns may be common in other habitats but might be overlooked because traditional neighborhood studies use a very narrow range of densities in their experiments. I encourage future neighborhood studies to use response surface designs to determine the prevalence of non-linear relationships in nature. In chapter 4, I examined how neighborhood composition (i.e. plant density and frequency) influenced four mechanisms known to influence damage to plants (predator suppression, foraging behavior of herbivores, plant quality, and microclimate) using a response surface experimental design. An associational effect was observed between Solanum damage and the frequency of a heterospecific neighbor (Solidago altissima). Predator abundance and richness, soil moisture, and herbivore foraging strategies were all influenced by the frequency of Solidago, suggesting that these mechanisms may contribute to associational susceptibility in this interaction. Other mechanisms (microclimate and plant quality) were influenced by Solidago and total plant densities, respectively. This study showed that different mechanisms can be influenced by different components of the neighborhood and most likely interacts to influence damage to plants. I discuss the implications of these finding for agriculture and for understanding the long-term consequences of damage for plant communities. In chapter 5, I examined how herbivory, herbivore community composition, plant nutrient content, and herbivore performance varied with latitude. A longstanding theory in biogeography is that species interactions, including herbivory, are stronger in southern latitudes compared to those in the north. Because of this, the latitudinal gradients (LG) hypothesis in damage and plant defenses predicts that plants should be better defended in the tropics because selection for plant defenses is greater. Recent empirical studies suggest that the predictions from this hypothesis may be limited to a narrow range of systems (e.g. salt marshes). In efforts to understand why LG in herbivory and plant defenses are not prevalent as once thought, I examined relationships between herbivore abundance and richness, plant nutrient content, and latitude in old-field systems. I also examined latitudinal gradients in herbivore performance using generalist and specialist herbivores. Some relationships with latitude matched predictions from the LG hypothesis (e.g. plant nutrient content, damage to Solidago altissima), while others had opposite relationships (e.g. herbivore abundance and richness, damage to Solanum carolinense), and some relationships varied with leaf longevity. Herbivore responses varied with diet specialization and the exact relationship with latitude (linear, non-linear, positive, negative, or no relationship) varied with herbivore species. These results suggest that the predictions from the LG hypothesis are too simple; a more thorough investigation of relationships between herbivore abundance, damage, and plant resistance in other wide-ranging systems is needed.