FIELD AND FORAGE CROPS
Yield Response of Dual-Toxin Bt Cotton to Helicoverpa zea Infestations J. GORE,1 J. J. ADAMCZYK, JR.,2 A. CATCHOT,3
AND
R. JACKSON4
J. Econ. Entomol. 101(5): 1594Ð1599 (2008)
ABSTRACT Field cage experiments were conducted to determine the impact of bollworms, Helicoverpa zea (Boddie), on yields of Bollgard II and Widestrike cotton, Gossypium hirsutum L. Oneday-old bollworm larvae were infested in white ßowers of Bollgard II and in white ßowers and terminals of Widestrike cotton. The infestation levels included 0, 50, and 100% of white ßowers for each type of cotton. Terminal infestations included one or two larvae per terminal on Widestrike cotton. Larvae were placed in ßowers of Bollgard II cotton each day for 1 to 4 wk during the Þrst 4 wk of ßowering during 2003, 2004, and 2005 seasons and in the ßowers or terminals of Widestrike cotton each day for 1 to 3 wk. Averaged across years and durations of infestation, yields of Bollgard II cotton were signiÞcantly reduced compared with noninfested Bollgard II cotton when 100% of white ßowers were infested. For Widestrike cotton, there was a reduction in yield when 100% of white ßowers were infested in 2005, but not in 2006. There was a signiÞcant relationship for cumulative numbers of white ßowers infested on seedcotton yield of Bollgard II during one of the 3 yr of the experiment. The regression equation during that year had a slope of ⫺0.77. No signiÞcant relationships were observed for cumulative numbers of white ßowers infested on yields of Widestrike cotton. Results of the current experiment suggest bollworms will rarely cause yield losses of Bollgard II and Widestrike cotton. Future research will need to focus on developing speciÞc thresholds for bollworms on Bollgard II and Widestrike cotton. KEY WORDS Bollgard II, Widestrike, Bollworm, integrated pest management, Helicoverpa
The introduction of Bollgard (Monsanto Co., St. Louis, MO) cotton, Gossypium hirsutum L., in 1996 signiÞcantly reduced the importance of the tobacco budworm, Heliothis virescens (F.), in the mid-southern United States (Perlak et al. 2001). In contrast, bollworms, Helicoverpa zea (Boddie), are less susceptible to Bollgard cotton, and one to two insecticide applications are needed annually to prevent economic losses on Bollgard cotton (Gore et al. 2003). Before the introduction and wide-scale adoption of Bollgard cotton, little research was conducted to determine whether thresholds for bollworms were needed on Bollgard cotton or at what level those thresholds should be set. Across the cotton belt, the current thresholds for bollworms on Bollgard are highly variable from state to state, but in general they are based on numbers of live larvae surviving to a given length or numbers of eggs per 100 plants (Gore and AdamMention of a proprietary product is for informational purposes only and does not constitute a recommendation for its use by the United States Department of Agriculture or the Agricultural Research Service. 1 Corresponding author: Mississippi State University, Delta Research and Extension Center, 82 Stoneville Rd., Stoneville, MS 38776 (e-mail:
[email protected]). 2 USDAÐARS, Honey Bee Research Unit, Weslaco, TX 78596. 3 Department of Entomology and Plant Pathology, Mississippi State University, Starkville, MS 39762Ð9775. 4 USDAÐARS, Southern Insect Management Research Unit, 141 Experiment Station and Lee Roads, Stoneville, MS 38776.
czyk 2004). Actual threshold levels were set based on trials that compared yields of insecticide-treated plots versus nontreated plots and not on replicated experiments designed to provide a more precise estimate of yield losses associated with bollworm infestation levels on Bollgard cotton. In response to the lack of scientiÞcally based thresholds, Gore and Adamczyk (2004) designed Þeld cage experiments to investigate the impact of various bollworm levels of infestation for four durations during the Þrst 4 wk of ßowering on yields of Bollgard cotton. Based on results of that research, bollworms are capable of reducing yields when 100% of white ßowers are infested for 1 to 4 wk and when at least 50% of white ßowers are infested for 2 to 4 wk (Gore and Adamczyk 2004). Also, a 1.69-g reduction in yield can be expected for every white ßower infested with a bollworm larva on Bollgard cotton (Gore and Adamczyk 2004). Dual-toxin Bt cotton (Bollgard II, Monsanto Co., and Widestrike, Dow Agrosciences, Indianapolis, IN) was released during the 2003 and 2005 growing seasons, respectively. The Bt insecticidal proteins from Bacillus thuringiensis (Bt) in Bollgard II include the Cry1Ac in the original Bollgard as well as Cry2Ab (Greenplate et al. 2003). In contrast, Widestrike cotton contains a different construct of Cry1Ac and Cry1 F (Adamczyk and Gore 2004). These types of cotton provide better control of bollworms than Bollgard.
0022-0493/08/1594Ð1599$04.00/0 䉷 2008 Entomological Society of America
October 2008
GORE ET AL.: IMPACT OF H. zea ON DUAL TOXIN COTTON
However, some surviving larvae and boll injury may still be observed in Bollgard II cotton (Jackson et al. 2003) and Widestrike cotton (Adamczyk and Gore 2004). The adoption of dual-toxin Bt cotton has been slow, but it is expected to signiÞcantly increase as the trait is infused into varieties with better agronomic characteristics. Thus, the objective of this experiment was to determine the impact of bollworms on yields of Bollgard II and Widestrike cotton. Materials and Methods Field cage experiments were conducted with Bollgard II cotton in 2003, 2004, and 2005, and with Widestrike cotton in 2005 and 2006 at the USDAÐARS, Jamie Whitten Delta States Research Center in Stoneville, MS. Bollgard II cotton (Deltapine 424 BG2/RR) was planted into a large Þeld cage on 17 April 2003, 11 May 2004, and 28 April 2005. Widestrike cotton (2005, Phytogen 470 WR; 2006, Phytogen 475 WRF) was planted into a separate cage on 4 May 2005 and 3 May 2006. Each cage covered 15 rows (101.6-cm centers) that were 35 m in length. All treatments of either Bollgard II or Widestrike were within a single cage. Each plot within a cage included two rows (101.6-cm centers) that were 1 m in length. Separate cages were used for Bollgard II and Widestrike. A 2-m nonplanted border surrounded each plot to minimize interplot movement of larvae. Plots were thinned 2 wk after plant emergence to obtain a Þnal density of 12 plants per plot (six plants per m row). The cages used in these experiments had three separate sections that served as replicates (i.e., three replicates per treatment per year) in a randomized complete block design. The experiments were planted so that all treatments within a replicate were within the same section of a cage. Treatments were arranged as a split-plot with two treatment factors. Duration of infestation (weeks) was the main-plot factor, and level of infestation was the sub-plot factor. Durations of infestation for Bollgard II cotton included 1, 2, 3, and 4 wk during the Þrst 4 wk of ßowering. For Widestrike cotton, the durations of infestation included the Þrst 3 wk of ßowering. The levels of infestation for Bollgard II and Widestrike included 0, 50, and 100% of white ßowers. These levels were based on initial studies that showed no negative impact to Bollgard II cotton at infestation levels ⬍50% (J.G. et al., unpublished data). Additionally, two levels of terminal infestations were included for the Widestrike plots because higher than expected levels of bollworm survival have been reported in Widestrike terminals in the Þeld (Smith et al. 2005). Those levels included one larva per terminal and two larvae per terminal for each of the durations of infestation. Plots were infested each day at the designated levels of infestation for the designated wk of infestation. To determine the proper timing for initiation of infestations, crop development was monitored throughout the season. The entire test areas were treated with insecticides (Orthene 90S, Valent Co. or Bidrin 8 EC, Amvac Co.) weekly up to 2 wk before infestation to control pests and minimize nat-
1595
ural enemy populations. At that time the cages were covered with translucent 32 mesh Lumite screen (Synthetic Industries, Greenville, GA) to prevent natural infestations of bollworms and other insect pests. Colonies of bollworms were established each year from non-Bt Þeld corn, Zea mays L., by collecting 2,000 Ð3,000 large larvae (older than fourth instar) over a period of 1 wk in Washington Co., MS. Larvae were maintained in the laboratory for one generation to obtain sufÞcient numbers of neonates at the proper time for infestations. Larvae were fed a commercially available artiÞcial diet (Heliothis Primix, Stoneßy Industries, Bryan, TX) in individual 29.5-ml plastic cups. Adults were fed a 10% sugar water solution and maintained in 3.9-liter cardboard containers. The tops of the containers were covered with a single layer of cheesecloth to provide a substrate for oviposition. The cheesecloth was removed daily and placed into plastic bags for larval eclosion. Neonates from the Þrst generation were fed artiÞcial diet for 24 ⫾ 4 h before infestation to minimize mortality from handling neonates in the Þeld. All life stages were maintained at 27 ⫾ 2⬚C, 80 ⫾ 5% RH, and a photoperiod of 14:10 (L:D) h. Infestations of larvae were initiated when plots across the test area averaged nine nodes above white ßower. Nodes above white ßower counts were determined by counting the number of main stem nodes above the upper-most Þrst position white ßower as described by Bourland et al. (1992). This corresponded to 7, 20, and 5 July in 2003, 2004, and 2005, respectively, for Bollgard II, and 13 and 20 July in 2005 and 2006, respectively, for Widestrike. The numbers of white ßowers were counted daily within each plot, and these counts were used to calculate the numbers of ßowers to be infested within each plot based on infestation levels. If a plot did not have enough ßowers on a particular day to achieve the desired level of infestation, a running total of ßowers was maintained for each plot to obtain the desired level of infestation on subsequent days. Larvae were placed into white ßowers daily with an artistÕs paintbrush and allowed to feed unhindered. Plots were harvested by hand each year, and seedcotton weights were determined. Data for seedcotton yield were analyzed with analysis of variance (ANOVA) (PROC MIXED version 9.1, SAS Institute 2004). In the model, year, duration of infestation (main-plot), level of infestation (subplot), and the duration by level interaction were designated as the Þxed components of the model. Replication was designated as a random effect. The replication by duration of infestation interaction was designated random as well, and it was the error term for duration of infestation. Residual error (replication by duration by level interaction) was the error term for sub-plots and the duration by level interaction. Means were estimated using the LSMEANS statement and adjusted according to the TukeyÕs studentized range test. In addition to ANOVA, the relationships between the cumulative numbers of white ßowers infested per plot and seedcotton yield in grams per plot of Bollgard
1596 Table 1.
JOURNAL OF ECONOMIC ENTOMOLOGY
Vol. 101, no. 5
Impact of bollworms on seedcotton yield of Bollgard II cotton, 2003–2005 Mean (SEM) seedcotton yield (g/plot) for each duration of infestation
Level of infestationa
1 wk
2 wk
3 wk
4 wk
Mean
0 50 100 Mean
2,071 (137.4) 1,996 (100.6) 1,991 (127.8) 2,019 (69.0)A
2,076 (129.4) 2,128 (136.7) 1,936 (105.9) 2,047 (71.1)A
2,091 (144.4) 1,938 (94.0) 1,989 (100.6) 2,005 (65.4)A
2,016 (148.8) 2,021 (131.1) 1,899 (120.7) 1,978 (75.6)A
2,063 (67.7)a 2,021 (57.4)ab 1,954 (55.4)b
Means within a column followed by the same lowercase letter and within a row followed by the same uppercase letter are not signiÞcantly different (␣ ⫽ 0.05). a Level of infestation represents the percentage of white ßowers that were infested in each plot.
II and Widestrike cotton were analyzed with regression analysis (PROC REG version 8.2, SAS Institute 2004) each year. In the model, grams of seedcotton was the dependent variable, and cumulative number of white ßowers infested was the independent variable. Grams of seedcotton per plot was used for regression analyses so that the slope of the regression equation would reßect the amount of seedcotton lost for each white ßower infested with a bollworm larva. Intercepts and slopes of the regressions for Bollgard II and Widestrike were compared across year by analysis of covariance. Where regressions were signiÞcant, the slope of the equation was calculated to provide a model for predicting yield losses in Bollgard II cotton associated with different levels of white ßower infestation. Results Impact of Bollworms on Yield of Bollgard II Cotton. Bollworm infestations had a signiÞcant impact on seedcotton yields of Bollgard II (Table 1). The year by treatment interactions were not signiÞcant (year ⫻ duration of infestation; F ⫽ 0.76; df ⫽ 6, 92.2; P ⫽ 0.60; year ⫻ level of infestation; F ⫽ 1.39; df ⫽ 4, 87.9; P ⫽ 0.24); therefore, data for yields were pooled across years. There was not a signiÞcant impact for duration of infestation (F ⫽ 0.44; df ⫽ 3, 7.53; P ⫽ 0.73) on seedcotton yields of Bollgard II (Table 1). There was a signiÞcant impact for level of infestation on seedcotton yields (F ⫽ 3.62; df ⫽ 2, 87.9; P ⫽ 0.03). When averaged across durations of infestation, yields were lower when 100% of white ßowers were infested compared with the noninfested plots. Table 2.
Impact of Bollworms on Yield of Widestrike Cotton. For Widestrike cotton, the year by level of infestation interaction was not signiÞcant (F ⫽ 1.34; df ⫽ 4, 90; P ⫽ 0.83). The year by duration of infestation was signiÞcant (F ⫽ 3.11; df ⫽ 2, 90; P ⫽ 0.05), and a separate analysis was conducted for each year. During 2005, there was not a signiÞcant interaction between bollworm duration of infestation and level of infestation on seedcotton yields of Widestrike cotton (F ⫽ 1.02; df ⫽ 8, 36; P ⫽ 0.44). Duration of bollworm infestation did not have a signiÞcant impact on yields of Widestrike cotton (F ⫽ 0.71; df ⫽ 2, 9; P ⫽ 0.52). Level of bollworm infestation did have a signiÞcant impact on seedcotton yields of Widestrike cotton (F ⫽ 2.89; df ⫽ 4, 36; P ⫽ 0.04). Averaged across durations of infestation, bollworms signiÞcantly lowered yield of Widestrike cotton when 100% of white ßowers were infested compared with the noninfested plots and plots where 50% of white ßowers were infested (Table 2). During 2006, there was not a signiÞcant interaction between bollworm duration of infestation and level of infestation on Widestrike yields (F ⫽ 0.83; df ⫽ 8, 36; P ⫽ 0.58). Additionally, the main effects for bollworm duration of infestation (F ⫽ 0.39; df ⫽ 2, 9; P ⫽ 0.69) and level of infestation (F ⫽ 0.34; df ⫽ 4, 36; P ⫽ 0.85) were not signiÞcant (Table 3). Regression Analysis for Bollgard Cotton, 2003–2005. There was a signiÞcant difference in the intercepts of the regression equations relating seedcotton yields to cumulative numbers of white ßowers infested for 2003, 2004, and 2005 (F ⫽ 115.23; df ⫽ 2, 150; P ⬍ 0.01). This difference reßected a difference in yield potential between the year (Fig. 1). The intercepts (SEM)
Impact of bollworms on seedcotton yield of Widestrike cotton, 2005 Mean (SEM) seedcotton yield (g/plot) for each duration of infestation
Level of infestationa
1 wk
2 wk
3 wk
Mean
T1 T2 0 50 100 Mean
1,501 (124.1) 1,582 (51.0) 1,540 (18.3) 1,523 (84.2) 1,522 (66.0) 1,533 (31.3)A
1,511 (71.0) 1,402 (75.4) 1,541 (109.5) 1,611 (126.9) 1,386 (118.5) 1,490 (45.3)A
1,383 (123.2) 1,351 (97.8) 1,492 (94.7) 1,514 (110.8) 1,249 (160.7) 1,398 (52.5)A
1,465 (59.5)ab 1,445 (50.2)ab 1,524 (44.5)a 1,549 (58.3)a 1,386 (71.7)b
Means within a column followed by the same lowercase letter and within a row followed by the same uppercase letter are not signiÞcantly different (␣ ⫽ 0.05). a Level of infestation represents the percentage of white ßowers that were infested in each plot or the numbers of larvae that were infested per plant (T1, one larva per terminal; T2, two larvae per terminal).
October 2008 Table 3.
GORE ET AL.: IMPACT OF H. zea ON DUAL TOXIN COTTON
1597
Impact of bollworms on seedcotton yield of Widestrike cotton, 2006 Mean (SEM) seedcotton yield (g/plot) for each duration of infestation
Level of infestationa
1 wk
2 wk
3 wk
Mean
T1 T2 0 50 100 Mean
985 (85.0) 981 (29.1) 999.9 (57.0) 892 (60.0) 951 (58.6) 962 (25.6)A
924 (76.1) 961 (42.0) 918 (28.6) 1,018 (79.8) 1,034 (11.3) 971 (24.3)A
1,047 (101.4) 1,000 (89.9) 1,047 (60.1) 941 (98.2) 1,026 (53.4) 1,012 (34.2)A
985 (48.5)a 981 (31.5)a 988 (31.0)a 950 (45.0)a 1,004 (26.6)a
Means within a column followed by the same lowercase letter and within a row followed by the same uppercase letter are not signiÞcantly different (␣ ⫽ 0.05). a Level of infestation represents the percentage of white ßowers that were infested in each plot or the numbers of larvae that were infested per plant (T1, one larva per terminal; T2, two larvae per terminal).
of the regressions were 2202 (38.5) g, 2,416 (49.6) g, and 1,676 (47.1) g per plot in 2003, 2004, and 2005, respectively. Additionally, the slopes of the regression equations were signiÞcantly different among year (F ⫽ 3.02; df ⫽ 2, 150; P ⫽ 0.05). Therefore, separate analyses were conducted for each year. The regressions for 2003 and 2005 did not produce signiÞcant relationships between cumulative numbers of white ßowers infested and seedcotton yield (2003: R2 ⫽ 0.02; F ⫽ 1.33; df ⫽ 1, 58; P ⫽ 0.25; and 2005: R2 ⫽ 0.002; F ⫽ 0.11; df ⫽ 1, 58; P ⫽ 0.74). There was a signiÞcant relationship between cumulative numbers of white ßowers infested and seedcotton yield in 2004 (R2 ⫽ 0.16; F ⫽ 6.45; df ⫽ 1, 58; P ⫽ 0.02). However, this was a weak relationship as indicated by the low R2 value. The slope of the regression equation for 2004 was ⫺0.77 g. Regression Analysis for Widestrike Cotton 2005– 2006. There was a signiÞcant difference in the intercepts of the regression equations relating seedcotton yield to the cumulative numbers of white ßowers infested for 2005 and 2006 (F ⫽ 163.12; df ⫽ 1, 68; P ⬍ 0.01), indicating a difference in yield potential between the 2 yr (Fig. 2). The intercepts (SEM) of the regression equations were 1,539 (44.9) g and 953 (29.2) g for 2005 and 2006, respectively. Additionally, the slopes of the regression equations were different between 2005 and 2006 (F ⫽ 5.43; df ⫽ 1, 68; P ⫽ 0.02). Therefore, separate analyses were conducted for each
2004 (P < 0.05)
2003 (NS)
year. For Widestrike cotton, there was not a signiÞcant relationship between cumulative numbers of white ßowers infested and seedcotton yields during 2005 (F ⫽ 3.37; df ⫽ 1, 34; P ⫽ 0.08) or 2006 (F ⫽ 1.59; df ⫽ 1, 34; P ⫽ 0.22). During both year, there was a poor Þt of the data to the model (2005: R2 ⫽ 0.09; and 2006: R2 ⫽ 0.04), indicating little relationship between bollworm infestation and seedcotton yield of Widestrike cotton. Discussion In experiments conducted from 1999 to 2002 in North Carolina, Jackson et al. (2003) demonstrated that mean yields of nontreated Bollgard II cotton were not signiÞcantly lower than yields of Bollgard II cotton treated with pyrethroids for bollworms. In those experiments, ßower bud and boll damage in nontreated conventional cotton averaged 44.3 and 63.0%, respectively, suggesting a high level of infestation. In other experiments, similar results were observed with Widestrike cotton (Jackson et al. 2006, Lorenz et al. 2006). Results from the current experiment suggest that bollworm infestations have little impact on yields of Bollgard II or Widestrike cotton. However, when averaged across year and durations of infestation, signiÞcant yield reductions were observed on Bollgard II cotton when 100% of white ßowers were infested with bollworms for at least 1 wk. SigniÞcant yield reduc-
2005 (NS)
2005 (NS) Seedcotton Yield (lbs per acre)
Seedcotton Yield (lbs per acre)
2006 (NS)
2000
3000 2500 2000 1500 1000
2004: y = -0.77x + 2416,
R2
= 0.16
1800 1600 1400 1200 1000 800 600 400
500 0
100
200
300
400
500
Cumulative Number of White Flowers Infested
Fig. 1. Regression analyses for seedcotton yield of Bollgard II cotton as a function of the cumulative numbers of white ßowers infested per 2-m2 plot.
0
50
100
150
200
250
Cumulative Number of White Flowers Infested
Fig. 2. Regression analyses for seedcotton yield of Widestrike cotton as a function of the cumulative numbers of white ßowers infested per 2-m2 plot.
1598
JOURNAL OF ECONOMIC ENTOMOLOGY
tions also were observed on Widestrike cotton when 100% of white ßowers were infested with bollworms when averaged across durations of infestation in 2005. Because of this, insecticide applications may be needed in some situations to prevent yield losses when extremely high populations of bollworms occur in white ßowers. In contrast, no yield reductions were observed from bollworm infestations in the terminals of Widestrike plants. Although bollworm survival may be higher than expected in the terminals of Widestrike cotton, these data demonstrated that those larvae did not cause signiÞcant damage, and yield losses were not observed. There are numerous factors that may inßuence the ability of bollworms to reduce yields of dual-toxin Bt cotton. Variation in the expression of the Cry proteins among different varieties and environments over time could inßuence the level of control. Parental background has been shown to have a signiÞcant impact on the expression of Cry1Ac in Bollgard cotton varieties (Adamczyk et al. 2001). Also, the level of Cry1Ac in Bollgard cotton declines throughout the season (Greenplate 1999). Although there is no information available about the variability in expression of Cry1Ac and Cry2Ab in Bollgard II cotton or about the variability of Cry1Ac and Cry1 F in Widestrike cotton, similar variation would be expected to that of Cry1Ac in Bollgard. The same variety of Bollgard II was used in all 3 yr of the current experiment; however, differences in expression among different varieties may inßuence bollworm damage in future experiments and in commercial situations. In contrast, different Widestrike varieties were used in 2005 and 2006 because of seed availability, and this may have inßuenced the different responses observed in those years. Results of the current study were likely inßuenced by differences in environmental conditions among yr that may have inßuenced protein expression and subsequent injury. Another important factor that could inßuence the ability of bollworms to damage dual-toxin cotton is variability in the response of bollworms to the Cry proteins. Before the commercial release of Bollgard cotton in 1996, Luttrell et al. (1999) showed that the response of Þeld collected bollworm colonies varied as much as 298.5-fold for puriÞed Cry1Ac and 61.3-fold for puriÞed Cry1Ab. A more recent survey showed a 130.0-fold level of variability to Cry1Ac in lyophilized MVP II (Ali et al. 2006). This documented variability in the response of bollworms to Cry proteins from B. thuringiensis can certainly have a signiÞcant impact on the level of control with both Bollgard, Bollgard II, and Widestrike cotton. Previous research showed that although bollworm survival is lower in white ßowers of Bollgard II cotton than Bollgard cotton, ⬎60% of bollworms survived on Bollgard II ßower anthers (Gore et al. 2001). Some boll damage can be observed from bollworms in Bollgard II cotton (Gore et al. 2003). Boll damage is apparently caused by survival of small larvae in white ßowers; however, relatively few larvae continue to feed and
Vol. 101, no. 5
cause injury after feeding in white ßowers (Gore et al. 2003). Regression equations in the current study suggested that one or more of the previously discussed factors inßuenced our results. However, it was difÞcult to determine which factor had the greatest impact because expression was not measured and different bollworm colonies were used each year. There was a signiÞcant relationship between cumulative numbers of white ßowers infested and seedcotton yield of Bollgard II cotton in one of the 3 yr of the current experiment. Based on the slope of that one regression equation, bollworms are capable of causing a 0.77-g reduction in seedcotton yield for each white ßower infested. The regression coefÞcient was very low (R2 ⫽ 0.16) indicating a relatively weak biological relationship between bollworm infestation and yield. Because of that weak relationship and the lack of a signiÞcant relationship in two of the 3 yr, speciÞc threshold recommendations for bollworms on Bollgard II are not determined from these data. However, these data demonstrated that bollworms may be capable of reducing yields of Bollgard II cotton and provide a starting point for future research to develop accurate thresholds for bollworms on Bollgard II cotton. Similarly, a signiÞcant relationship between bollworm infestation and seedcotton yield was not observed for Widestrike cotton. However, based on the ANOVA, bollworms did signiÞcantly reduce yields of Widestrike cotton when 100% of white ßowers were infested for 1 wk or more. These data provide valuable information about the impact (or lack of impact) of bollworms on yields Bollgard II and Widestrike cotton. Before speciÞc threshold recommendations can be made based on these data, more research to verify these results in large-scale Þeld plots is needed. Based on these data, bollworms can reduce yields of both Bollgard II and Widestrike cotton under extremely high pressure that persists over time. However, it seems unlikely that either Bollgard II or Widestrike cotton will require insecticide applications to prevent economic yield losses from low to high bollworm populations that do not persist for ⬎1 wk. Acknowledgments We thank Donny Adams, Don Hubbard, and John Gordon Campbell for technical assistance with this research and Debbie Boykin for assistance with statistical analyses.
References Cited Adamczyk, J. J., Jr., and J. Gore. 2004. Laboratory and Þeld performance of cotton containing Cry1Ac, Cry1F, and both Cry1Ac and Cry1F (WideStrike䉸) against beet armyworm and fall armyworm larvae (Lepidoptera: Noctuidae) Fla. Entomol. 87: 427Ð 432. Adamczyk, J. J., Jr., D. D. Hardee, L. C. Adams, and D. V. Summerford. 2001. Correlating differences in larval survival and development of bollworm (Lepidoptera: Noctuidae) and fall armyworm (Lepidoptera: Noctuidae) to differential expression of Cry1A(c) ␦-endotoxin in vari-
October 2008
GORE ET AL.: IMPACT OF H. zea ON DUAL TOXIN COTTON
ous plant parts among commercial cultivars of transgenic Bacillus thuringiensis cotton. J. Econ. Entomol. 94: 284 Ð 290. Ali, M. I., R. G. Luttrell, and S. Y. Young, III. 2006. Susceptibilities of Helicoverpa zea and Heliothis virescens (Lepidoptera: Noctuidae) populations to Cry1Ac insecticidal protein. J. Econ. Entomol. 99: 164 Ð175. Bourland, F. M., D. M. Oosterhuis, and N. P. Tugwell. 1992. Concept for monitoring the growth and development of cotton plants using main-stem node counts. J. Prod. Agric. 5: 532Ð538. Gore, J., and J. J. Adamczyk. 2004. Impact of bollworms, Helicoverpa zea (Boddie), on maturity and yields of Bollgard cotton. J. Cotton Sci. 8: 223Ð229. (http://www. cotton.org/journal/2004 Ð 08/4/upload/jcs08 Ð223.pdf). Gore, J., B. R. Leonard, and J. J. Adamczyk. 2001. Bollworm (Lepidoptera: Noctuidae) survival on Bollgard䉸 and Bollgard II䉸 cotton ßower bud and ßower components. J. Econ. Entomol. 94: 1445Ð1451. Gore, J., B. R. Leonard, and R. H. Gable. 2003. Distribution of bollworm, Helicoverpa zea (Boddie), injured reproductive structures on genetically engineered Bacillus thuringiensis var. kurstaki Berliner cotton. J. Econ. Entomol. 96: 699 Ð705. Greenplate, J. T. 1999. QuantiÞcation of Bacillus thuringiensis insect control protein Cry1Ac over time in Bollgard cotton fruit and terminals. J. Econ. Entomol. 92: 1377Ð 1383. Greenplate, J. T., J. W. Mullins, S. R. Penn, A. Dahm, B. J. Reich, J. A. Osborn, P. R. Rahn, L. Ruschke, and Z. W. Shappley. 2003. Partial characterization of cotton plants expressing two toxin proteins from Bacillus thuringiensis: relative toxin contribution, toxin interaction, and resistance management. J. Appl. Entomol. 127: 340 Ð347. Jackson, R. E., J. R. Bradley, Jr., and J. W. Van Duyn. 2003. Field performance of transgenic cottons expressing one or two Bacillus thuringiensis endotoxins against bollworm, Helicov-
1599
erpa zea (Boddie). J. Cotton Sci. 7: 57Ð64. (http://www. cotton.org/journal/2003Ð07/3/upload/jcs07Ð057.pdf). Jackson, R. E., S. Malone, J. R. Bradley, J. Van Duyn, A. Herbert, and R. M. Huckaba. 2006. EfÞcacy of Widestrike and Bollgard II cottons against bollworm under insecticide-sprayed and non-sprayed conditions in North Carolina and Virginia, pp. 1531Ð1535. In Proceedings of the 2006 Beltwide Cotton Conferences, National Cotton Council, Memphis, TN. (www.cotton.org/beltwide). Lorenz, G. M., K. Colwell, R. B. Lassiter, J. Greene, G. Studebaker, J. Hardke, and C. Shelton. 2006. Performance of Widestrike cotton in Arkansas, 2004 Ð2005, pp. 1526 Ð1530. In Proceedings of the 2006 Beltwide Cotton Conferences, National Cotton Council, Memphis, TN. (www.cotton.org/beltwide). Luttrell, R. G., L. Wan, and K. Knighten. 1999. Variation in susceptibility of noctuid (Lepidoptera) larvae attacking cotton and soybean to puriÞed endotoxin proteins and commercial formulations of Bacillus thuringiensis. J. Econ. Entomol. 92: 21Ð32. Perlak, F. J., M. Oppenhuizen, K. Gustafson, R. Voth, S. Sivasupramaniam, D. Heering, B. Carey, R. A. Ihrig, and J. K. Roberts. 2001. Development and commercial use of Bollgard䉸 cotton in the USA: early promises versus todayÕs reality. Plant J. 27: 489 Ð501. SAS Institute. 2004. SAS/STAT 9.1 userÕs guide. SAS Institute, Cary, NC. Smith, R. H., D. P. Moore, R. A. Haygood, L. B. Braxton, and A. R. Parker. 2005. Performance of WidestrikeTM insect protection for control of lepidopteran pests in Alabama from 2001Ð2004, pp. 1330 Ð1335. In Proceedings of the 2007 Beltwide Cotton Conferences, National Cotton Council, Memphis, TN. (www.cotton.org/beltwide). Received 15 February 2008; accepted 11 April 2008.
