Dissertation (Ph.D.)--Auburn University, 2009. / Abstract. Vita. Includes bibliographic references (p.142-151).

Thesis(M.S.)--Auburn University, 2006. / Abstract. Vita. Includes bibliographic references (p.93-105).

The purpose of the present study was to investigate how plant spatial patterns and insect behavior interact to influence the population dynamics of insects using the plants. The study included three phases: l) field experiments using collards (Brassica oleracea) and the crucifer insect fauna; 2) simulation models representing the population dynamics of an insect herbivore as functions of insect dispersal behavior and host plant patch size; and 3) model-field syntheses integrating model predictions and field variability estimates to choose an appropriate spatial scale for future field experiments or applications. In field experiments on surrounding plant type, collards were planted with 1) other collards; 2) collards treated with a systemic insecticide; 3) broccoli, a related host; and 4) tomatoes, an unrelated host whose odor may repel crucifer pests. In three such experiments, eggs and larvae of the imported cabbageworm (Pieris rapae (L.)) were most abundant an collards surrounded by tomatoes. These results were contrary to those from previous experiments with other crucifer pests, but can perhaps be explained by P. rapae's preference for ovipositing on isolated host plants. In three patch size experiments, P. rapae eggs and larvae were more abundant on collards in small patches; in one of these experiments, the diamondback moth (Plutella xylostella (Linn.)) was more abundant in large patches. The results for P. rapae were consistent with previous studies showing more oviposition on plants in smaller patches and on plants at the edge of a patch. The simulation models predicted mean level and variability of an insect herbivore population based on interactions between insect behavior and host plant patch size. Features of insect behavior modeled were: 1) an inverse relationship between distance and dispersal; 2) preferences for host vs. nonhost plants, which differ for generalist and for specialist herbivores; and 3) preference of specialists for larger areas of host plants. Aspects of plant pattern studied were size and number of host plant patches in a background of nonhost vegetation. Constant, exponential and logistic growth of insect populations in the nonhost background and in the host patches were used. The models were designed to provide a theoretical framework for studying interactions between insect behavior and plant pattern, not to duplicate the dynamics of the field system. However, for the model-field syntheses, parameters of one model were fit to field data to establish a correspondence between expected differences in population levels predicted by the model and variability estimates obtained in the field experiments. The fitted model was used to construct a graph of differences in population levels versus patch size. Field variability estimates were used to calculate the least difference in population levels, for a given number of replicates and subsamples, that could be detected in field experiments. The least difference was compared with the differences predicted b y the fitted model to determine what patch sizes will produce significant effects of patch size on insect populations. This method was used to assess feasibility of future field experiments and to select appropriate patch sizes. The method can also be used to evaluate crop field size in agroecosystems as a component of pest management.

No description available.

Diamondback moth, Plutella xylostella (L.) (Lepidoptera: Plutellidae) is a serious pest of crucifer vegetables (Brassica sp.) worldwide because of a lack of effective natural enemies in certain regions and because of insecticide resistance. In 2003, laboratory and field studies were initiated in Virginia to better understand P. xylostella, its primary natural enemies, and their susceptibilities to insecticides in order to develop an economically and environmentally sound integrated pest management program for collards in the state. Ecological life table studies of P. xylostella immature stages on collards located on the Eastern Shore and on Kentland Farm, near Blacksburg at the New River Valley, VA revealed that most (98 to 99%) of P. xylostella died from natural causes. Mortality factors varied between the two regions. Neonates, small larvae, and large larvae disappearing were major mortality factors. Rainfall, predation, and dispersal probably contributed the most to this mortality. Egg mortality played a bigger role at the New River Valley compared with the Eastern Shore. Three parasitoid species were found that contributed to the mortality of P. xylostella: Diadegma insulare (Cresson) (Hymenoptera: Ichneumonidae); Oomyzus sokolowskii (Kurdjumov) (Hymenoptera: Eulophidae); and Microplitis plutellae (Muesebeck) (Hymenoptera: Braconidae). Additional studies conducted in the laboratory using leaf-dip bioassays revealed that P. xylostella collected from the Eastern Shore of Virginia, showed significant tolerance levels to esfenvalerate, acetamiprid, methomyl, methoxyfenozide, indoxacarb, and acephate compared with a susceptible strain of P. xylostella. The highest tolerance ratio (1,876 fold) was to esfenvalerate, a commonly-used pyrethroid. All of the insecticides tested in this study were quite toxic to the adult stage of the parasitoids, D. insulare and O. sokolowskii. The insect growth regulator, methoxyfenozide was considerably less toxic than other insecticides such as esfenvalerate, methomyl, acephate, spinosad, indoxacarb, and emamectin benzoate at field-rate and 1% of field-rate concentrations. The aforementioned insecticides as well as some other insecticides were evaluated several times in the field for efficacy against P. xylostella as well as other pests of collards. The most efficacious insecticides over five field experiments included acephate, emamectin benzoate, esfenvalerate, methomyl, methoxyfenozide, novaluron, indoxacarb, and spinosad. These insecticides were followed in relative efficacy by Bt kurstaki, acetamiprid, and azadirachtin, which provided relatively inconsistent control of lepidopteran larvae over the experiments. Effective insecticide options in collards that are less toxic to natural enemies and that can fit well into integrated pest management programs include indoxacarb, spinosad, novaluron, emamectin benzoate, methoxyfenozide, and Bacillus thuringiensis subsp. kurstaki. / Ph. D.