| It has been shown in numerous systems that parasitoids are attracted to chemical volatiles produced by herbivore-damaged plants. It has been suggested that by artificially manipulating these volatiles in crop plants, biological control can be enhanced in agricultural systems. Before this technology is implemented, it is important to understand the tritrophic dynamics of the system. I used two different modeling approaches to address this phenomenon.; In the first model, I combined a modified predator-prey functional response equation with an age-structured herbivore population model. I looked at the effects of plant induction delay, plant relaxation delay, herbivore density, and parasitoid host-age preference. Parasitoids following signals had an advantage over randomly foraging parasitoids under the majority of the parameter combinations I examined, with the largest advantage occurring when plants were able to induce within one day of herbivory onset and relax signal production within one day of herbivore removal, when less than 10% of the plants were occupied by herbivores, and when parasitoids were able to attack all feeding instars of their hosts. Under most cases, higher herbivore density had a negative effect, induction delay had no noticeable effect, and shorter relaxation delays had a positive effect on signal relevance to the parasitoid. In cases where parasitoids could only attack first instar hosts, plants with an induction delay longer than two days produced signals that were irrelevant to the parasitoids, and this loss of signal relevance worsened with shorter relaxation delays and smaller herbivore densities.; In the second model, I took a spatially-explicit stochastic simulation approach and examined the Brassica oleraceae, Pieris rapae, and Cotesia rubecula system in more detail. In addition to the variables I considered for the first model, I also looked at a parasitoid distance bias variable. Instead of varying herbivore density over a large range of parameters, I used realistic Pieris rapae dynamics, following three generations of herbivores (a single field season) per simulation. Similar to the previous model, parasitoids gained the most from signals when all herbivore instars were viable hosts, herbivore density was low, and relaxation delays were short. Unlike the general deterministic model, shorter induction delays could lead to considerable gains for the parasitoids in this model.; Together, the models indicate that there are some conditions that favor parasitoids following herbivore-induced plant volatiles, especially when herbivore densities are low, and plant can induce or relax their signal within a day of changes in herbivory. By creating plants that do produce signals in the right time frame, it may be possible to optimize biological control in agriculture. However, it is also apparent from my models, that herbivore-induced volatiles are ineffective during herbivore outbreak conditions, indicating that biological control alone would not be able to contain pest populations because parasitoids are limited by factors other the time it takes to find hosts, which is the primary way herbivore-induced plant volatiles can aid foraging parasitoids.; Improving biological control is one of the practices growers can adopt as part of Integrated Pest Management (IPM), and in the final section of this dissertation I discussed the results of a survey exploring how growers adopt IPM. I found that practices consistent with IPM were adopted in a piecemeal fashion by cotton growers in the eastern part of the state. My analysis indicated that growers did not see all these practices as part of a single management decision, but rather as parts of many independent decisions, dealing with weed management, insect management, crop management, and ecosystem management. |
