Cotton varieties genetically engineered to express CrylAc, a delta-endotoxin protein from Bacillus thuringiensis with insecticidal properties to many Lepidopterans, are now commercially available to aid in the control of the tobacco budworm, Heliothis virescens F. (Lepidoptera: Noctuidae). Most of the resistance-management strategies for Bt cotton have focussed on this primary pest, and Helicoverpa zea (Boddie) (Lepidoptera: Noctuidae), to a lesser degree. Because CrylAc is continually expressed in all tissues of the cotton plant, many Lepidopteran insects feeding on Bt cotton may experience selection pressure for improved tolerance of this insecticide. In addition to the major pests of cotton, secondary pests could also develop resistance to CrylAc when feeding on Bt cotton. Currently, there is little available information concerning the best strategy to manage resistance to CrylAc in these secondary pests (Gould and Tabashnik 1998). Soybean loopers, Pseudoplusia includens (Walker) (Lepidoptera: Noctuidae) are occasional foliage feeders of cotton and have the potential to develop resistance to CrylAc (Mascarenhas et al. 1998). We took advantage of a natural infestation of soybean loopers (SBL) in the Mississippi Delta to investigate how well SBL larvae tolerated CrylAc expressed in the leaves of Bt cotton. Comparisons were made of the numbers of larvae found in Bt and non-Bt cotton and also comparisons of the growth of larvae feeding on Bt and non-Bt cotton. We also looked at the variability in tolerances of individuals feeding on Bt cotton. Two varieties of cotton were used in this study. NuCOTN 33B (Bt cotton, treatment = 'BT') and SG125 (conventional cotton, treatment = 'NBT') were planted (9 May 1998) in a randomized design totaling 7 plots of each variety. Plot dimensions were approx. 9 x 9 m and plots were arranged in a 3 x 5 plot grid. Each plot was separated from the others by 2 m of unplanted space. On 6 July 1998, we observed a sizeable infestation of SBL larvae (first and second instars in BT and NBT treatments). Larvae from all plots were manually removed (23 July 1998), placed in plastic bags containing foliage from the plant on which each larva was feeding, and immediately brought back to the lab to be scored and weighed. The number of larvae from each plot was tabulated. Counts of larvae were log-transformed to enhance normality and to homogenize the variances in the BT and NBT treatments. A t-test was used to compare NBT and BT counts. Data are presented as untransformed average numbers per plot. In addition, larvae from 5 and 4 randomly chosen BT and NBT plots, respectively, were weighed to the nearest hundredth of a mg to look for developmental differences in SBL larvae feeding on BT and NBT foliage. Log-transformed weight (mg) data were subjected to ANOVA to determine if there were any differences in the developmental rates of larvae collected from BT and NBT plots. Bt treatment (BT vs. NBT cotton) was considered a fixed effect and plots nested within treatments [plots(treatment)] effects were considered a random source of variation (PROC GLM; SAS 1985). Satterthwaite's approximation was used to estimate the denominator degrees of freedom. Significantly more SBL larvae were collected from NBT plots than BT plots (Fig. 1; t = 3.336, df = 12, P = 0.006). Larvae from NBT plots were also significantly larger than larvae from BT plots (F = 238.96, df = 1, 104.5, P < 0.0001). The weights of larvae
Read full abstract