Jeff Gore, Roger Leonard and Gabie Church

Genetically engineered plants are an important part of integrated pest management (IPM) programs in cotton production. One such plant, Bollgard cotton, includes a gene from a bacterium, Bacillus thuringiensis, that is toxic to caterpillar pests, while being safe for humans, other animals and the environment. Bollgard cotton was introduced into commercial production in 1996, and comprised only 15 percent of the total cotton acreage. Since that time acreage planted to these varieties has increased and in 2001 accounted for more than 80 percent.

Bollgard provides excellent control of the tobacco budworm and usually maintains low to moderate bollworm densities below economic injury levels. Supplemental control with insecticide applications is often needed, however, to prevent economic losses when high bollworm densities persist for several days.

In 1996, bollworm populations were extremely high in most areas of the mid-southern United States, southeastern United States and Texas. Consequently, crop advisers in those regions observed large numbers of bollworm larvae in fields planted to Bollgard cotton. Most of these populations consisted of small larvae feeding within white flowers and on small bolls under dried bloom tags (dried flower corollas). There is little information explaining why bollworms are more commonly found in white flowers of Bollgard cotton than non-Bollgard cotton. Initially, bollworm egg laying was considered to be different on Bollgard plants compared with non-Bollgard plants, but researchers in Mississippi and South Carolina found no differences in tobacco budworm and bollworm egg densities or in the distribution of eggs on Bollgard cottons compared with the non-Bollgard parental cottons.

Alternatively, early larval dispersal may be different on Bollgard cotton plants compared with non-Bollgard cotton plants. Larvae are the developmental stage controlled by Bollgard cotton, and differences in larval behavior could result in feeding preferences on specific plant parts. Therefore, studies were conducted in Louisiana to determine if differences in bollworm larval behavior occur on Bollgard cotton plants compared with non-Bollgard plants. This information is valuable in developing accurate sampling methods in Bollgard cotton. Accurate sampling will improve insecticide application timing against bollworms and reduce unnecessary applications that can be costly and disrupt natural enemy populations.

Bollworm Infestation
Bollgard cotton (NuCOTN 33B) and a non-Bollgard parental cotton (Deltapine 5415) were planted at the LSU AgCenter’s Macon Ridge Station near Winnsboro in 2000 and 2001. Fertilization rates and general agronomic practices for cotton production followed Louisiana Cooperative Extension Service recommendations.

Bollworms were collected from crimson clover in April and sweet corn in June and maintained in the laboratory for at least one generation. Larvae hatching from eggs were provided an artificial laboratory diet. After 48 hours, bollworm larvae were removed from the diet and placed on flowering cotton plants.

With a small paintbrush, first stage bollworm larvae were placed in plant terminals on individual cotton plants (one larva per plant). Individual plants were thinned before infestation so that no interplant movement could occur. Bollworm-infested plants were examined at three, six and 24 hours after infestation. Larvae were located, and the numbers of main stem nodes they moved down from the terminal were recorded.

Bollworm Movement
Typically, bollworm larvae remain near the terminals of non-Bollgard cotton plants. Results from this study indicate that bollworm larvae tended to move a greater vertical distance on Bollgard cotton compared with non-Bollgard cotton. Larvae moved more than twice as far down the plant on Bollgard cotton compared with non-Bollgard cotton at all rating intervals (Figure 1). Cotton plants begin flowering from the bottom and, as they mature, flowers are closer to tops of plants. Therefore, at any given time during the season, the youngest fruiting forms (squares) are generally toward the top of the plant canopy and the older fruit (bolls) are generally closer to the bottom. Consequently, as larvae move down the plant, they are more likely to feed on cotton bolls. Fewer bollworm larvae remained in the terminals and squares on Bollgard cotton than on non-Bollgard cotton at 48 hours after infestation, where multiple larvae were infested in plots (Figure 2). A higher percentage of larvae were lower in the plant canopy within flowers and bolls in Bollgard cotton than in non-Bollgard cotton.

Implications for IPM
Bollworm moths typically use the top one-third of cotton plants for egg laying. Most eggs are usually found in or near plant terminals and small bollworm larvae remain near the terminals of non-Bollgard cotton plants, feeding on small squares. As larvae develop, they typically migrate down the plant’s main stem and injure larger squares and bolls. Most current sampling plans for non-Bollgard cotton are based on numbers of larvae in plant terminals. Cotton pest management consultants have trouble deciding when to apply foliar insecticides to manage bollworms in Bollgard cotton. Action thresholds to initiate bollworm control with foliar sprays are based on numbers of eggs and/or larvae in terminals and numbers of larval infested/damaged squares on non-Bollgard cotton. In Louisiana, insecticide applications are recommended when at least five live larvae per 100 plants plus eggs are present. These thresholds and scouting methods may not be appropriate for Bollgard cotton, because bollworm larvae feeding on white flowers and bolls may be overlooked.

Bollworm larvae disperse more rapidly on Bollgard cotton compared with non-Bollgard cotton. Larvae remained near the top of non-Bollgard cotton plants and damaged terminal foliage and small squares. In contrast, larvae on Bollgard cotton were observed lower in the plant canopy, feeding on white flowers and bolls.

Bollworm larvae began migrating away from plant terminals within three hours after infestation on Bollgard cotton. Therefore, when eggs hatch, there is a narrow period when larvae can still be observed in or near plant terminals. Within six hours after infestation, larvae moved more than four nodes below plant terminals. Field scouts searching for bollworm infestations in plant terminals are likely not to find larvae in the terminals if sampling is begun more than six hours after larval hatching. For the plot infestations, bollworm-infested plant terminals on non-Bollgard cotton exceeded the current action threshold and would require treatment to prevent economic losses. In contrast, bollworm-infested terminals and squares did not exceed the action threshold on Bollgard cotton and would not be treated. However, if the data for infested flowers and bolls are considered, Bollgard cotton may require a foliar insecticide application to prevent economic yield losses.

Current scouting protocols and action levels to initiate insecticide treatments for bollworms on non-Bollgard cotton may not be appropriate for Bollgard cotton. Field scouts should sample white flowers and small bolls in addition to plant terminals and squares when scouting Bollgard cotton. Additional data indicate that, in some instances, bollworms feeding in white flowers can ultimately damage as many as 2.7 fruiting structures in Bollgard cotton. Thus bollworms may cause significant yield losses in Bollgard cotton. Therefore, insecticide applications should be directed at these populations before they become established low in the plant canopy under dried flower corollas.

Acknowledgment
Karla Emfinger, Ralph Sheppard and Rhett Gable, as well as fellow graduate students and numerous student workers, assisted in conducting these studies. Cotton Incorporated and Louisiana’s cotton producers provided funding for this project.

Jeff Gore, Graduate Student, Department of Entomology; B. Rogers Leonard, Professor, Macon Ridge Research Station, Winnsboro, La.; and Gabie Church, Instructor, Department of Experimental Statistics, LSU AgCenter, Baton Rouge, La.

(This article was publishedin the winter 2002 issue of Louisiana Agriculture.)