Department of Entomology, College of Agriculture and Life Sciences, North Carolina State University, Raleigh, NC 27695

Transcription

1 INSECTICIDE RESISTANCE AND RESISTANCE MANAGEMENT Comparative Production of Helicoverpa zea (Lepidoptera: Noctuidae) from Transgenic Cotton Expressing Either One or Two Bacillus thuringiensis Proteins with and without Insecticide Oversprays R. E. JACKSON, J. R. BRADLEY, JR., J. W. VAN DUYN, AND F. GOULD Department of Entomology, College of Agriculture and Life Sciences, North Carolina State University, Raleigh, NC J. Econ. Entomol. 97(5): 1719Ð1725 (2004) ABSTRACT Transgenic cotton, Gossypium hirsutum (L.), expressing either one or two Bacillus thuringiensis ssp. kurstaki Berliner (Bt) proteins was compared with the conventional sister line in Þeld experiments with regard to production of bollworm, Helicoverpa zea (Boddie), and bolls damaged by bollworm. The relative numbers of bollworms that developed on Bollgard (Monsanto Co., St. Louis, MO), Bollgard II (Monsanto Co.), and conventional cotton were estimated under nontreated conditions in 2000 and both insecticide-treated and nontreated conditions in 2001Ð2002 in North Carolina tests. Averaged across seven Þeld studies under nontreated conditions, Bollgard cotton generated statistically similar numbers of large (L4ÐL5) bollworm larvae compared with the conventional variety; however, Bollgard cotton produced signiþcantly fewer damaged bolls and bollworm adults than the conventional variety. Production of large larvae, damaged bolls, and adults was decreased dramatically by Bollgard II cotton as compared with Bollgard and conventional varieties. When comparing insecticide-treated and nontreated cotton genotypes, both Bt cotton sustained less boll damage than the conventional variety averaged across insecticide regimes; furthermore, Bollgard II cotton had fewer damaged bolls than the Bollgard variety. When averaged across cotton genotypes, pyrethroid oversprays reduced the numbers of damaged bolls compared with the nontreated cotton. Insecticide-treated Bollgard cotton, along with insecticide-treated and nontreated Bollgard II cotton reduced production of bollworm larvae, pupae, and adults. However, the addition of pyrethroid oversprays to Bollgard II cotton seemed to be the best resistance management strategy available for bollworm because no bollworms were capable of completing development under these conditions. KEY WORDS Bacillus thuringiensis, bollworm, cotton, Helicoverpa zea, resistance management GENETICALLY ALTERED COTTON, Gossypium hirsutum (L.), expressing the Cry1Ac -endotoxin derived from the soil bacterium Bacillus thuringiensis ssp. kurstaki Berliner (Bt) were planted to 67 and 70% of the total North Carolina cotton acreage in 2001 and 2002, respectively (Bacheler 2002, 2003). The primary targets of these Bt cotton, Bollgard (Monsanto Co., St. Louis, MO), in North Carolina are the bollworm, Helicoverpa zea (Boddie), and to a lesser extent the tobacco budworm, Heliothis virescens (F.). Although Bollgard cotton has provided essentially absolute control of tobacco budworm, supplemental insecticidal oversprays have often been required to provide adequate control of bollworm to prevent yield losses (Mahaffey et al. 1994, 1995; Lambert et al. 1996, 1997). Increased survival of bollworms in Bt cotton compared with tobacco budworms can be attributed to a higher natural tolerance of bollworms to the Cry1Ac protein (Stone and Sims 1993). An Environmental Protection Agency (EPA) ScientiÞc Advisory Panel (1998) (SAP) reported that commercially available Bollgard cotton does not meet the high dose criterion for resistance management of Bt cotton because of the natural tolerance of bollworms to Cry1Ac in Bollgard cotton. The SAP determined that the single toxin-producing varieties expressed only a moderate dose with respect to the level necessary to prevent rapid Bt resistance evolution in bollworms when an effective refuge is used. The current EPA-approved resistance management strategy for Bt cotton is based on a high dose of the toxin and a non-bt cotton refuge planted in proximity to Bt cotton or embedded within Bt cotton Þelds. The nontransgenic cotton refuges serve to produce susceptible moths that reduce the probability of resistant moths mating with one another (Gould 1998). The ScientiÞc Advisory Panel (1998) suggested that 500 susceptible pest adults would be required for each insect carrying at least one Bt resistance allele to successfully delay resistance development to Bollgard cotton. The refuge strategy seems to be inappropriate for bollworms because Bollgard cotton is not a high dose; however, Tabashnik et al. (2003) suggested that refuges may still be useful in delaying Bt resistance /04/1719Ð1725$04.00/ Entomological Society of America

2 1720 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 97, no. 5 evolution even though one or more key assumptions of the strategy are not applicable. Although Bollgard cotton varieties have been commercially available since 1996, quantiþcation of bollworm adults from Bt and non-bt refuge cotton has not been reported. Bollgard II (Monsanto Co.) cotton that produces two Bt endotoxins, Cry1Ac and Cry2Ab, has recently been commercialized. Laboratory studies have con- Þrmed that these dual-gene cotton varieties produce approximately the same level of the Cry1Ac protein as the single-gene Bollgard varieties, but the addition of the Cry2Ab protein provides additional protection against lepidopterous pests (Greenplate et al. 2000, Adamczyk et al. 2001). Stewart et al. (2001) demonstrated in a laboratory bioassay that leaf tissue from Bollgard II plants was more toxic to second instars of bollworms than leaf tissue from Bollgard plants; thus, the dual-gene construct has the potential to increase bollworm efþcacy and may improve Bt resistance management. Caprio (1998) estimated that under certain Þeld conditions pyramiding the two toxin-producing genes into a plant could increase the durability of the toxins 9.3 times in comparison with serial introductions of the same toxins in the presence of a 5% refuge. Therefore, the pyramiding of genes encoding Cry1Ac and Cry2Ab expression in Bollgard II genotypes could increase the effectiveness of the current refuge system for bollworms. Because preliminary reports suggested that Bollgard II cotton exhibited increased efþcacy against bollworms and may provide resistance management beneþts, commercialization of the dual-toxin cotton may delay Bt resistance development in bollworms and sustain this important technology. The pyramiding of two toxins in a variety may not always lead to delay in resistance evolution (Gould 1986). The objectives of this study were to quantify bollworm damage to cotton fruit and to estimate production of large larvae, pupae, and adults in conventional, Bollgard, and Bollgard II cotton under insecticidetreated and nontreated Þeld conditions. Materials and Methods Field studies were conducted at the Tidewater Research Station, Washington County, North Carolina, in 2000 and 2002; the Upper Coastal Plain Research Station, Edgecombe County, North Carolina, in 2000Ð 2002; C. A. Martin Farm, Martin County, North Carolina, in 2002; and Albemarle Beach Farms, Washington County, in The experiment was designed as a randomized complete split-plot with four replicates. Whole plots were cotton genotypes that consisted of Bollgard II (DP50BX), Bollgard (DP50B), and a conventional sister line (DP50), which were 24, 20, and 16 rows (0.91-m centers), respectively, by 15.2 m in length. Unequal plot sizes were established to increase the probability of collecting bollworm larvae from the more toxic Bt genotypes. Larval production would be much less in the highly efþcacious Bollgard II cotton; thus, increased plot sizes were designated to the Bt genotypes in the event that an increase in sample size was necessary to more precisely estimate production. Subplots without insecticides applied for heliothine control (UT) consisted of 20, 16, and 12 nontreated rows for Bollgard II, Bollgard, and conventional genotypes, respectively, and four rows that were treated (T) with an insecticide as required for supplemental bollworm control. Field trials were planted on 15 and 18 May in Edgecombe and Washington counties, respectively, in 2000; on 2 May in Edgecombe County in 2001; and on 14 May in Edgecombe and Washington counties (Tidewater Research Station), and 15 May in Martin and Washington counties. (Albemarle Beach Farm) in Aldicarb (Temik 15G, Bayer CropScience, Kansas City, MO) was applied in-furrow at planting at 0.34 kg (AI)/ha for control of early season insect pests. Acephate (Orthene 97 PE, Valent USA Corp., Walnut Creek, CA) was applied at 0.34 kg (AI)/ha as a midseason overspray for control of tarnished plant bugs and stink bugs, as well as to reduce arthropod natural enemies of bollworm. The purpose of this overspray was to eliminate known sources of variability that may have made unclear the effects of Bt cotton and insecticides on bollworms. Supplemental bollworm control within appropriate subplots was accomplished with applications of cypermethrin (Ammo 2.5 EC, FMC Corp., Philadelphia, PA), lambda cyhalothrin (Karate Z 2.08 CS, Syngenta Crop Protection, Inc., Greensboro, NC), cyßuthrin (Baythroid two 2.0 EC, Bayer Corp.), or spinosad (Tracer 4 SC, Dow AgroSciences, LLC, Indianapolis, IN). Lambda cyhalothrin at kg (AI)/ha was applied to appropriate subplots for supplemental bollworm control at Edgecombe (19 July and 7 August) and Washington counties (27 July and 9 August) in 2000, as well as at Edgecombe County (10 and 16 August) in In 2002, lambda cyhalothrin at kg (AI)/ha plus spinosad at kg (AI)/ha were applied to all test sites on 23 July and 1 August; spinosad was added to control tobacco budworms. Weed control, fertilization, plant growth regulation, and defoliation followed the recommendations of the North Carolina Cooperative Extension Service. The total number of large harvestable bolls was counted in a randomly selected area of 1.5 row meters per treatment per replicate to provide a means of converting numbers of larvae, damaged fruit, pupae, and adults to a per hectare basis. The total numbers of bollworm-damaged bolls and L4ÐL5 larvae were counted on a predetermined number of bolls per plot (100Ð500) Þve to six times from late July to early September. L4ÐL5 bollworm larvae were collected and placed on fresh cotton bolls from the respective genotype in individual 30-ml plastic cups and transported to the laboratory. These larvae were reared on bolls from the respective genotypes until the prepupal stage. Prepupae were then placed singly into 30-ml plastic cups containing non-bt artiþcial diet that served as a pupation medium. Bollworm larvae, pupae, and adults, in addition to boll damage, were estimated in insecticide-treated subplots only in 2001Ð2002. Plants from which L4ÐL5 larvae were collected were

3 October 2004 JACKSON ET AL.: H. zea PRODUCTION IN BT COTTON 1721 Table 1. Estimated mean (SE) numbers of live bollworm larvae (L4 L5), damaged bolls, and bollworm adults produced on a per hectare basis by conventional, Bollgard, and Bollgard II cotton genotypes without insecticides for heliothine control averaged across seven test sites in North Carolina, Genotype No. live larvae/ha a No. damaged bolls/ha a No. adults/ha a Conventional (DP50) 60,283 (9,025)a 367,835 (18,881)a 47,078 (7,030)a Bollgard (DP50B) 42,899 (6,793)a 140,005 (13,353)b 24,589 (4,300)b Bollgard II (DP50BX) 3,910 (1,208)b 16,645 (2,969)c 1,697 (682)c a Means within the same column followed by the same letter are not signiþcantly different, FisherÕs protected least signiþcant difference tested to verify the presence of Bt proteins. Larvae collected from contaminants in Bt plots were excluded from the analyses. Numbers of damaged bolls, large larvae, pupae, and adults were subjected to a log transformation before analysis. All data were then subjected to analysis of variance by using PROC MIXED (Littell et al. 1996). Due to the unbalanced nature of locations within years, all tests (year*location combinations) were analyzed across locations and years. Effects of genotype and insecticide were considered Þxed, whereas effects of test (year*location combinations) and replicates were considered random. Treatments were compared (P 0.05) on the basis of least-squares means (PDIFF option of the LSMEANS statement). Results for data transformed before analysis are reported as untransformed arithmetic means and standard errors. Results Moderate bollworm populations in the non-bt cotton characterized the three experiments in 2000Ð2001; however in 2002, infestations were comparable with the extremely high bollworm numbers encountered by Mahaffey et al. (1995). In comparing UT genotypes averaged across seven Þeld studies, bollworm larval (L4ÐL5) production was statistically similar between the Bollgard and the conventional variety, even though the Bollgard variety had 30% fewer larvae compared with the conventional variety (F 35.79; df 2, 12; P ) (Table 1). However, the Bollgard II genotype reduced larval numbers by 91 and 94% compared with the Bollgard and conventional varieties, respectively. Both Bt genotypes dramatically reduced bollwormdamaged bolls below that of the conventional variety (F 46.11; df 2, 12; P ) (Table 1); in addition, the Bollgard II genotype had signiþcantly fewer damaged bolls than the Bollgard variety. The conventional cotton variety sustained an average of 367,835 damaged bolls per hectare over the 3-yr period of the experiments, whereas Bollgard and Bollgard II cotton exhibited signiþcant reductions of 62 and 95%, respectively. Furthermore, the Bollgard II variety sustained 88% fewer bollworm-damaged bolls compared with the Bollgard variety. Bollworm adult production was signiþcantly reduced by both Bt genotypes versus the conventional variety (F 20.70; df 2, 12; P ) (Table 1). The conventional cotton produced 47,078 bollworm adults per hectare averaged across seven Þeld studies. Bollgard and Bollgard II cotton generated 48 and 96% fewer adults than the conventional variety. Similar to results regarding larval production, Bollgard II reduced bollworm adult production by 93% compared with Bollgard. In 2001Ð2002, estimates of bollworm production were made in both T and UT subplots of each cotton genotype. Bollworm larval production averaged across Þve Þeld studies was characterized by a genotype* insecticide regime interaction (F 31.85; df 2, 6; P ). UT conventional cotton generated 75,510 large bollworm larvae per hectare (Table 2). Neither the UT Bollgard variety nor the application of insecticides to the conventional variety signiþcantly reduced larval production compared with the UT conventional variety. T Bollgard cotton, however, successfully reduced the numbers of bollworm larvae below that of the T and UT conventional variety, as well as the UT Bollgard variety; furthermore, the T Bollgard cotton generated similar numbers of bollworm larvae compared with the UT Bollgard II genotype. The addition of insecticide oversprays to the Bollgard II genotype reduced larval numbers by 99% compared with the T Bollgard and UT Bollgard II genotypes, but it did not statistically differ from the UT Bollgard II cotton with regard to larval production. SigniÞcant main effects of genotype and insecticide regime were observed for the numbers of bollwormdamaged bolls. Averaged across insecticide regimes, damaged boll numbers neared 250,000 per hectare in the conventional variety (Table 3). Bollgard and Bollgard II cotton exhibited a reduction in numbers of damaged bolls of 61 and 95%, respectively, compared with the conventional variety (F 46.14; df 2, 8; P ). The Bollgard II genotype sustained 88% less Table 2. Estimated mean (SE) numbers of bollworm larvae (L4 L5) per hectare produced by conventional, Bollgard, and Bollgard II cotton genotypes with and without insecticides for heliothine control averaged across five study sites in North Carolina, Genotype Insecticide regime No. larvae/ha a Conventional (DP50) Nontreated 75,510 (11,439)a Bollgard (DP50B) Nontreated 55,970 (8,509)a Conventional (DP50) Insecticide-treated 28,973 (4,182)a Bollgard (DP50B) Insecticide-treated 6,005 (1,561)b Bollgard II (DP50BX) Nontreated 5,118 (1,586)bc Bollgard II (DP50BX) Insecticide-treated 32 (32)c a Means within the same column followed by the same letter are not signiþcantly different, FisherÕs protected least signiþcant difference

4 1722 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 97, no. 5 Table 3. Estimated mean (SE) numbers of bollworm-damaged bolls per hectare in conventional, Bollgard, and Bollgard II cotton genotypes with and without insecticides for heliothine control averaged across five study sites in North Carolina, Insecticide regime Genotype Insecticidetreated Mean a Nontreated Conventional 133,669 (10,796) 364,352 (41,997) 249,011 (14,104)a Bollgard 31,327 (5,802) 164,964 (16,643) 98,145 (10,409)b Bollgard II 2,527 (810) 20,118 (3,853) 11,322 (2,095)c Mean 55,113 (5,562)b 181,451 (12,545)a a Means within the same column or row followed by the same letter are not signiþcantly different, FisherÕs protected least signiþcant difference boll damage than the Bollgard variety when averaged across insecticide regimes. When averaged across genotypes, the addition of pyrethroid oversprays significantly reduced the numbers of damaged bolls by 70% compared with nontreated cotton (F 23.15; df 1, 4; P ). Bollworm pupal production was depicted by a genotype*insecticide regime interaction (F 5.08; df 2, 8; P ). As with larval production, UT Bollgard cotton did not statistically differ from UT conventional cotton with respect to pupal production (Table 4). UT Bollgard cotton lowered pupal numbers by only 26% compared with the UT conventional variety. Although the addition of insecticide oversprays to the conventional variety did not statistically lower pupal production compared with the UT Bollgard variety, the insecticide oversprays reduced pupal numbers by 73% below that of the UT conventional variety. Similar to larval production, insecticide oversprays onto Bollgard cotton reduced numbers of pupae below that of T and UT conventional cotton, as well as UT Bollgard cotton. The T Bollgard variety produced statistically comparable numbers of pupae compared with both T and UT Bollgard II cotton, even though no pupae were produced in the T Bollgard II genotype. A genotype*insecticide regime interaction also typiþed bollworm adult production (F 6.91; df 2, 8; P ). An estimated 57,783 bollworm adults per hectare were generated from UT conventional cotton when averaged across Þve Þeld studies (Table 5). UT Table 4. Estimated mean (SE) numbers of bollworm pupae per hectare produced by conventional, Bollgard, and Bollgard II cotton genotypes with and without insecticides for heliothine control averaged across three sample dates in Edgecombe County, North Carolina, 2001 Genotype Insecticide regime No. pupae/ha a Conventional (DP50) Nontreated 65,203 (9,848)a Bollgard (DP50B) Nontreated 48,098 (7,410)ab Conventional (DP50) Insecticide-treated 17,878 (2,719)b Bollgard II (DP50BX) Nontreated 4,525 (1,331)c Bollgard (DP50B) Insecticide-treated 3,826 (1,000)c Bollgard II (DP50BX) Insecticide-treated 0 (0)c a Means within the same column followed by the same letter are not signiþcantly different, FisherÕs protected least signiþcant difference Table 5. Estimated mean (SE) numbers of bollworm adults per hectare produced by conventional, Bollgard, and Bollgard II cotton genotypes with and without insecticides for heliothine control averaged across three sample dates in Edgecombe County, North Carolina, 2001 Genotype Insecticide regime No. adults/ha a Conventional (DP50) Nontreated 57,783 (8,917)a Bollgard (DP50B) Nontreated 32,597 (5,417)ab Conventional (DP50) Insecticide-treated 13,168 (2,235)b Bollgard II (DP50BX) Nontreated 2,260 (899)c Bollgard (DP50B) Insecticide-treated 2,092 (746)c Bollgard II (DP50BX) Insecticide-treated 0 (0)c a Means within the same column followed by the same letter are not signiþcantly different, FisherÕs protected least signiþcant difference Bollgard cotton failed to signiþcantly reduce adult production in comparison with production from the UT conventional variety; however, insecticide oversprays onto the conventional variety lowered bollworm adult production by 77% compared with UT conventional cotton. Although T conventional cotton reduced adult production by 60% compared with the UT Bollgard variety, no statistical differences were evident between these treatment combinations. UT Bollgard II cotton yielded 2,260 bollworm adults per hectare, which was signiþcantly fewer than that generated by the UT Bollgard variety, as well as T and UT conventional cotton. T Bollgard cotton performed comparably to both T and UT Bollgard II genotypes with regard to production of bollworm adults, even though the T Bollgard variety produced no bollworm adults in Þve Þeld studies. Discussion Bollworm production estimates might have been overestimated in the Bt genotypes because of prolonged larval development. However, the authors concluded that the potential overestimates were small because larval counts only included L4ÐL5 stages. Duplicate counts of larvae whose developmental rate had been slowed on Bt cotton would not have been likely because L4ÐL5 larval estimates were made weekly. Overestimates concerning pupal and adult approximations could have been caused by the arti- Þciality of the pupal environments in the study; pupae may have encountered adverse conditions in the Þeld such as predation or parisitization, closing of emergence tunnels, unfavorable soil types, excessive moisture, and others (Williams and Stinner 1987, Kring et al. 1993, Cabanillas and Raulston 1994). However, these known sources of variability were avoided in an effort to determine the effect of the Bt cotton on pupal development and successful eclosion. To the contrary, estimates may have been somewhat conservative in the event that handling effects may have reduced larval survival, pupal development, or eclosion. The reductions in larval production for the Bollgard II genotype were likely due to the increased toxicity of the dual-gene construct compared with the toxicity of the single-gene varieties (Stewart et al. 2001). The

5 October 2004 JACKSON ET AL.: H. zea PRODUCTION IN BT COTTON 1723 higher toxicity of the Bollgard II genotype compared with Bollgard varieties results from the high expression of the Cry2Ab endotoxin in addition to a level of Cry1Ac endotoxin production similar to that produced by the Bollgard variety (Greenplate et al. 2000). Numbers of bolls sustaining bollworm damage within the UT cotton followed a somewhat different trend compared with that of larval production. Although larval production did not differ between UT conventional and Bollgard cotton, damaged boll numbers were signiþcantly reduced by the UT Bollgard cotton because of reduced larval feeding that results from the ingestion of the Cry1Ac endotoxin and increased movement of larvae in the Bollgard variety as noted by Gore et al. (2002). Differences in number of damaged bolls between Bollgard and Bollgard II cotton seemed to result from a difference in the numbers of larvae generated from each genotype; the 91% reduction in larval numbers led to 88% fewer damaged bolls in Bollgard II. Adult production was progressively lower from the conventional to Bollgard to Bollgard II varieties. Because larval numbers were similar between UT conventional and Bollgard cotton, either the Bt toxins negatively affected the pupation/eclosion process or the incidence of disease or parasitization may have been elevated within Bollgard cotton due to prolonged larval development. Storer et al. (2001) reported that pupal mortality for bollworms that originated on Bt corn plants was elevated compared with those from non-bt corn plants. Approximately 78% of large bollworm larvae in the UT conventional variety survived to the adult stage, whereas proportionately fewer (57%) made it to adult emergence in the UT Bollgard variety. Furthermore, only 43% of large bollworm larvae in the Bollgard II variety progressed to the adult stage. The 15-fold reduction of bollworm adults exhibited by Bollgard II compared with Bollgard signiþcantly limits the number of Bt resistance alleles in the general bollworm population, assuming that survivors are resistance allele carriers. Based on reports from Jackson et al. (2002), the portion of the bollworm population that survives on Bollgard cotton may be those that carry minor Bt resistance alleles; therefore, Bollgard II cotton may signiþcantly delay resistance evolution in bollworms by reducing the number of individuals carrying these Bt resistance alleles for the Cry1Ac toxin from the general bollworm population. Bollgard II has the potential to be a high dose for bollworms, which would suggest that the dual-toxin cotton might delay Bt resistance evolution for a longer time than the single-gene varieties. To the contrary, if Bollgard II were not a high dose for bollworms, it could arguably increase the rate of resistance evolution (Gould 1998). However, Gould (1998) also reports that transgenic insecticidal crops with moderate doses would be sustainable with increased refuge size and that other crop hosts for generalists, such as bollworm, could serve as part of a larger refuge. Jackson et al. (2003) demonstrated that Þeld corn, soybean, and peanut generated 92% of bollworm larvae that infested commercial crop Þelds in North Carolina and showed that production from these crops was temporally and spatially synchronous with that of Bt cotton. By supplementing the non-bt cotton refuge with susceptible moths from alternate crop hosts, the dual-toxin cotton would offer a balance between delaying resistance development in bollworms and providing greater caterpillar control for producers. Treatment combinations including Bollgard II cotton and the T Bollgard variety possess an increased potential to delay Bt resistance evolution in bollworms by reducing the number of Bt resistance alleles in the general bollworm population. The increased efþcacy of Bollgard II cotton potentially creates a high dose environment that should effectively maximize the beneþts of current Bt resistance management refuge options. Based on results from Caprio (1998), Bollgard II cotton could effectively delay Bt resistance evolution in bollworms 9.3 times compared with serial introductions of the single-gene varieties in the presence of a 5% unsprayed cotton refuge. Unless producers can relate seed costs and insecticide application expenses to increased proþts, most are unlikely to change their current practices. Producers are likely to embrace Bollgard II technology because it reduced boll damage by 36% compared with pyrethroid-treated Bollgard cotton. Because T Bollgard II reduced boll damage by a negligible amount compared with UT Bollgard II, producers are unlikely to overspray Bollgard II cotton for bollworms. However, insecticide applications for Lygus spp. and stink bugs were made to 56 and 43%, respectively, of the U.S. cotton acreage in 2003 (Williams 2004); these applications also offer varying levels of control of bollworms. In addition, the number of bug sprays in Bollgard II cotton is likely to increase because oversprays for lepidopterous pests will be less frequent. The potential delay in resistance development in bollworms is related to adult production where Bollgard II demonstrates an obvious beneþt. Although adult production in T Bollgard II cotton did not statistically differ from that of UT Bollgard II and T Bollgard cotton, the numerical reduction in adults in T Bollgard II may be biologically signiþcant. Thus, T Bollgard II cotton combined with appropriate refuges seem to be the best Bt resistance strategy available at this time. Bollworm adaptation to the Cry1Ac protein has not been observed over the 7 yr of Bollgard use (Jackson et al. 2002). According to resistance management theory, the production of bollworm moths in conventional cotton as shown in these studies is not optimal for substantially delaying resistance development; however, there are a number if mitigating factors that should be considered as well. Field level resistance has not occurred because of a Þtness cost associated with larval development on Bt cotton and/or substantial temporal and spatial bollworm production from noncotton crop hosts has effectively supplemented the cotton refuge. The latter has likely played the most important role in delaying resistance development in bollworms. Reports from Jackson et al. (2003) dem-

6 1724 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 97, no. 5 onstrated that Þeld corn, soybean, and peanut generated 90% of bollworm larvae that infested commercial crop Þelds in North Carolina. In addition, Gould et al. (2002) suggested that long distance migration from Mexico in the spring and from the north in late summer provide large numbers of susceptible bollworms to mate with counterparts developing on Bt cotton throughout the cotton belt. Thus, the temporal and spatial production of bollworm adults from these crop Þelds more than effectively supplements the 5% unsprayed refuge for Bt cotton. To couple the limited production of bollworm moths from Bollgard II cotton with the enormous numbers from noncotton crop hosts should effectively satisfy Bt resistance management concerns while offering producers another tool for combating lepidopterous pests. Acknowledgments We express appreciation to Cotton, Inc. for providing a graduate research assistantship for R.E.J. and to Monsanto Co. for providing partial project funding. Special thanks are given to Wayne Modlin, Andrew Summerlin, and Phil Threatt for technical assistance. References Cited Adamczyk, J. J., Jr., K. Bew, L. C. Adams, and D. D. Hardee Evaluation of Bollgard (cv. DP50BII) in the Mississippi delta: Þeld efþcacy against various Lepidoptera while proþling season-long expression of cry1ac and cry2ab, pp. 835Ð836. In Proceedings of the Beltwide Cotton Conference, 9Ð13 January 2001, Anaheim, CA. National Cotton Council of America, Memphis, TN. Bacheler, J. S Managing insects on cotton, pp. 126Ð 151. In 2002 Cotton Information. North Carolina Cooperative Extension ServiceÐNorth Carolina State University Publ. AG-417. Bacheler, J. S Managing insects on cotton, pp. 124Ð 150. In 2003 Cotton Information. North Carolina Cooperative Extension ServiceÐNorth Carolina State University Publ. AG-417. Cabanillas, H. E., and J. R. Raulston Evaluation of the spatial pattern of Steinernema riobravis in corn plots. J. Nematol. 26: 25Ð31. Caprio, M Evaluating resistance management strategies for multiple toxins in the presence of external refugia. J. Econ. Entomol. 91: 1021Ð1031. Gore, J., D. R. Cook, M. M. Willrich, and B. R. Leonard Bollworm/Bollgard interactions: implications for management. In Proceedings of the Beltwide Cotton Conference, 8Ð12 January 2002, Atlanta, GA. National Cotton Council of America, Memphis, TN. Gould, F Simulation models for predicting durability of insect-resistant germ plasm: a deterministic diploid, two-locus model. Environ. Entomol. 15: 1Ð10. Gould, F Sustainability of transgenic insecticidal cultivars: integrating pest genetics and ecology. Annu. Rev. Entomol. 43: 701Ð726. Gould, F., N. Blair, M. Reid, T. L. Rennie, J. Lopez, and S Micinski. Bacillus thuringiensis-toxin resistance management: stable isotope assessment of alternate host use by Helicoverpa zea. Proc. Natl. Acad. Sci. U.S.A. 99: 581Ð586. Greenplate, J. T., S. R. Penn, Z. Shappley, M. Oppenhuizen, J. Mann, B. Reich, and J. Osborn Bollgard II efþcacy: quantiþcation of total lepidopteran activity in a 2-gene product, pp. 1041Ð1043. In Proceedings of the Beltwide Cotton Conference, 4Ð8 January 2000, San Antonio, TX. National Cotton Council of America, Memphis, TN. Jackson, R. E., J. R. Bradley, Jr., F. Gould, and J. W. Van Duyn Genetic variation for resistance to cry1ac and cry2ab in bollworm, Helicoverpa zea, in North Carolina. In Proceedings of the Beltwide Cotton Conference, 8Ð12 January Atlanta, GA. National Cotton Council of America, Memphis, TN. Jackson, R. E., J. R. Bradley, Jr., and J. W. Van Duyn QuantiÞcation of Helicoverpa zea populations in eastern North Carolina crop environments: implications for B. t. resistance management. In Proceedings of the Beltwide Cotton Conference, 6Ð10 January 2003, Memphis, TN. National Cotton Council of America, Memphis, TN. Kring., T. J., J. R. Ruberson, D. C. Steinkraus, and D. A. Jacobson Mortality of Helicoverpa zea (Lepidoptera: Noctuidae) pupae in ear-state Þeld corn. Environ. Entomol. 22: 1338Ð1343. Lambert, A. L., J. R. Bradley, Jr., and J. W. Van Duyn Effects of natural enemy conservation and planting date on the susceptibility of Bt cotton to Helicoverpa zea in North Carolina, pp. 931Ð935. In Proceedings of the Beltwide Cotton Conference, 9Ð12 January Nashville, TN. National Cotton Council of America, Memphis, TN. Lambert, A. L., J. R. Bradley, Jr., and J. W. Van Duyn Interactions of Helicoverpa zea and Bt cotton in North Carolina, pp. 870Ð873. In Proceedings of the Beltwide Cotton Conference, 7Ð10 January 1997, New Orleans, LA. National Cotton Council of America, Memphis, TN. Littell, R. C., G. A. Milliken, W. W. Stroup, and R. D. Wolfinger SAS system for mixed models. SAS Institute, Cary, NC. Mahaffey, J. S., J. S. Bacheler, J. R. Bradley, Jr., and J. W. Van Duyn Performance of MonsantoÕs transgenic B.t. cotton against high populations of lepidopterous pests in North Carolina, pp. 1061Ð1063. In Proceedings of the Beltwide Cotton Conference, 5Ð8 January 1994, San Diego, CA. National Cotton Council of America, Memphis, TN. Mahaffey, J. S., J. R. Bradley, Jr., and J. W. Van Duyn B. t. cotton: Þeld performance in North Carolina under conditions of unusually high bollworm populations, pp. 795Ð798. In Proceedings of the Beltwide Cotton Conference, 4Ð7 January 1995, San Antonio, TX. National Cotton Council of America, Memphis, TN. Scientific Advisory Panel Transmittal of the Þnal report of the FIFRA ScientiÞc Advisory Panel, subpanel on Bacillus thruingiensis (Bt) plant-pesticides and resistance management, 9Ð10 February 1998 (docket number OPPTS-00231). Stewart, S. D., J. J. Adamczyk, Jr., K. S. Knighten, and F. M. Davis Impact of Bt cottons expressing one or two insecticidal proteins of Bacillus thuringiensis Berliner on growth and survival of noctuid (Lepidoptera) larvae. J. Econ. Entomol. 94: 752Ð760. Stone, T. B., and S. R. Sims Geographic susceptibility of Heliothis virescens and Helicoverpa zea (Lepidoptera: Noctuidae) to Bacillus thuringiensis. J. Econ. Entomol. 86: 989Ð994. Storer, N. P., J. W. Van Duyn, and G. G. Kennedy Life history traits of Helicoverpa zea (Lepidoptera: Noctuidae) on non-bt and Bt transgenic corn hybrids in eastern North Carolina. J. Econ. Entomol. 94: 1268Ð1279.

7 October 2004 JACKSON ET AL.: H. zea PRODUCTION IN BT COTTON 1725 Tabashnik, B. E., Y. Carriere, T. J. Dennehy, S. Morin, M. S. Sisterson, R. T. Roush, A. M. Shelton, and J. Zhao Insect resistance to transgenic Bt crops: lessons from the laboratory and Þeld. J. Econ. Entomol. 96: 1031Ð1038. Williams, D. W., and R. E. Stinner Respiration and drowning in overwintering pupae of Heliothis zea (Boddie). Can. Entomol. 119: 329Ð338. Williams, M. R Cotton insect loss estimates Ð2003, pp. 1258Ð1312. In Proceedings of the Beltwide Cotton Conference, 6Ð9 January 2004, San Antonio, TX. National Cotton Council of America, Memphis, TN. Received for publication 20 January 2004; accepted 27 June 2004.