Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Spatial distribution of fall armyworm, Spodoptera frugiperda (J. E. Smith), (FAW) was studied in cornfields at the St. Gabriel Experiment Station in Louisiana from 1981 to 1983. Corn plants were sampled and FAW larvae counted and divided into categories of small, medium, and large for both single plant and five plant samples. Count data for each category by date were then fitted to Poisson, negative binomial, and Neyman type A distributions. Means and variances for each category by date were also analyzed with the Taylor Power law. Small larvae were more aggregated than the medium larvae which were more aggregated than the large larvae. One plant samples were more efficient than five plant samples. Prevalence of a nuclear polyhedrosis virus (NPV) was studied in 1981 and 1982 at St. Gabriel and in 1982 and 1983 at a private dairy farm 6 km east of Hammond, Louisiana. Information on FAW populations, climate, and virus prevalence were analyzed by multiple regression and regression models made for each location and for combined data. Important factors at St. Gabriel included corn height, number of larvae per plant, and temperature variables. Important factors at Hammond were all environmental--temperature, degree day, rain, and solar radiation. The most important variable for regression models made from the combined data was number of larvae per plant. Parasitism of FAW larvae by three species of parasite--Cotesia (Apanteles) marginiventris, Rogas laphygmae, and Chelonus sp.--was studied concurrently with NPV epizootiology. Regression models were constructed for parasitism by each species using the combined data sets. The R('2) values for the models were low, not exceeding 0.40. The only variable to appear in all models was corn height. The F(,1) offspring of larvae collected in St. Gabriel in 1981 and Hammond in 1982 were tested by bioassay against the locally occurring NPV. Heterogeneity with respect to the NPV increased with time. F(,1) larvae of parents surviving both NPV applications and naturally occurring NPV selection often gave bioassay results too heterogeneous to fit the probit model. Immigration of susceptible individuals was the most likely cause of the heterogeneity.