Translator Disclaimer
1 February 2009 Three-Year Field Monitoring of Cry1F, Event DAS-Ø15Ø7-1, Maize Hybrids for Nontarget Arthropod Effects
Author Affiliations +
Abstract

Field studies were conducted over a 3-yr period to investigate the potential effects of cultivating transgenic maize hybrids containing a Cry1F insect-resistant protein on nontarget arthropod abundance. The narrow spectrum of activity of Cry1F against a subset of lepidopteran pest species would not suggest broad-spectrum effects on nontarget arthropods. However, because of the insecticidal nature of Bt proteins, an alternate hypothesis is that some nontargets may be affected by exposure to the protein. To examine this hypothesis at the field level, monitoring for nontarget organism abundance was initiated at four locations across the U.S. Corn Belt from 2004 through 2006. At each location, paired fields (≈0.8 ha each) of commercial Cry1F maize hybrids and isogenic nontransgenic control hybrids were planted. Sampling methods used to monitor nontarget organisms included visual surveillance, sticky cards, pitfall traps, and litterbags. Data were analyzed using multivariate analyses to look for a general community level response to the treatments. Analysis of variance was conducted on individual taxa to detect differences distinct from the primary community response. Community level analyses of the nontarget arthropod abundance showed no significant impact on community abundance when comparing Bt with non-Bt maize fields. Analyses of the individual taxa also showed no significant differences in abundance between Bt and non-Bt fields. Results of these studies confirm earlier laboratory testing and support the hypothesis that Cry1F maize does not produce adverse effects on nontarget arthropods occurring in maize fields.

Maize hybrids containing event DAS-Ø15Ø7-1 expressing the cry1F gene for control of lepidopteran maize pests were approved for commercial sale in the United States in 2001 under the trade name Herculex I. Event DAS-Ø15Ø7-1 maize was developed through the insertion of a synthetic truncated cry1F gene from Bacillus thuringiensis (Bt) variety aizawai. Accumulation of Cry1F protein in maize provides effective plant protection against the larval stage of many lepidopteran maize pests including European corn borer, Ostrinia nubilalis (Hübner); southwestern corn borer, Diatraea grandiosella Dyar; fall armyworm, Spodoptera frugiperda (J. E. Smith); western bean cutworm, Striacosta albicosta (Smith); black cutworm, Agrotis ipsilon (Hufnagel); and corn earworm Helicoverpa zea (Boddie). Event DAS-Ø15Ø7-1 maize also contains the selectable marker phosphinothricin acetyltransferase (pat), which confers tolerance to glufosinate-ammonium (Liberty Herbicide; Bayer CropScience, Research Triangle Park, NC). The PAT protein is not a known toxin and/or pathogen of plant or animal species (U. S. EPA 2005). Although some field studies suggest shifts in arthropod abundance based on changes in weed management and tillage (Thorbek and Bilde 2004, Brooks et al. 2005), this study involved standard weed management practices in all treatments and did not involve the application of glufosinateammonium herbicides.

Bt Cry proteins have been shown to have a narrow range of insecticidal activity within a few insect orders (Metz 2003). Since the introduction of Bt crops for commercial use in the United States in the 1990s, a large body of studies designed to evaluate their effect on nontarget organisms has been generated. O'Callaghan et al. (2005) reviewed the effects of Bt plants (primarily maize and cotton) on aerial (pollinators and natural enemies) and soil-borne nontarget biota. They did not detect significant adverse effects in plant feeding and beneficial nontarget insects or in soil-dwelling organisms such as earthworms, collembolans, and other soil microflora. Romeis et al. (2006) summarized >50 field studies conducted to examine the impact of Bt plant incorporated proteins on the natural enemies of crop pests. These studies covered a variety of genetically modified crops including maize, cotton, potato, tobacco, and eggplant. Nearly one half of these studies were conducted on maize expressing the Cry1Ab protein, the first Bt protein to be commercialized in maize and which currently constitutes the majority of Bt maize acreage in the United States. These studies ranged in size, duration, and sampling methodology and mainly focused on comparing the abundance of beneficial arthropods found in Bt and non-Bt plots (Daley and Buntin 2005, Dively 2005, Pilcher et al. 2005, Lopez et al. 2005). Overall, these studies indicated no major effects against natural enemies in Bt fields compared with non-Bt fields, with the occasional exception of taxa that were dependent on Bt-susceptible pests as hosts. These data support the conclusions of lower tier (laboratory) safety tests for Bt maize, which generally have concluded that few negative effects on nontarget organisms are expected because of the narrow spectrum of insecticidal activity of Bt proteins. More recently, Marvier et al. (2007) conducted a meta-analysis of the effects of Bt cotton and maize on nontarget invertebrates using data from 42 field experiments. Their analyses suggest that non-target invertebrates are generally more abundant in Bt cotton and Bt maize fields compared with nontransgenic fields treated with conventional insecticides. Some differences in abundance were detected for certain taxa when Bt fields were compared with non-Bt fields without insecticides, although these differences were event specific and may have been caused (either directly or indirectly) by control of the target pests that serve as prey for the affected beneficial species.

To date, no field scale evaluations of nontarget arthropods in CrylF maize have been published in the literature. As with other lepidopteran-active Bt proteins, the CrylF protein seems to have a narrow range of insecticidal activity (U.S. EPA 2001). However, because of the insecticidal nature of Bt proteins, an alternate hypothesis is that some nontargets may be affected by exposure to the protein. To evaluate this hypothesis, field trials were conducted from 2004 to 2006 to compare nontarget arthropod abundance in fields containing commercial CrylF maize hybrids with abundance in paired non-Bt maize fields.

Materials and Methods

Nontarget arthropod abundance was monitored at four locations across the U.S. Corn Belt with field sites in Nebraska, Iowa, Indiana, and Wisconsin. Table 1 lists field locations, field characteristics, maize hybrids, and planting dates for each study. At each test location, a pair of Cry1F Bt maize and isogenic (maize line of similar genetic background but lacking the DAS-01507-1 event) control maize fields were planted in plots of ≈0.8 ha using cultivation practices typical of each area. Each field was bordered with non-Bt hybrid maize. No conventional insecticides were applied to the fields during the growing season. Test plots in this study were not sprayed with glufosinate-ammonium herbicides, and weed control practices were uniformly applied across test and control plots.

Data collection within the fields was conducted from randomly selected points assigned in the “sampling area” (area excluding the outermost ≈30.5 m) of each field. Arthropod abundance was monitored using multiple sampling techniques: visual observations to monitor taxa occurring on maize plants, sticky cards to monitor aerial taxa, pitfall traps to monitor surface and ground dwelling taxa, and litterbags to measure taxa colonizing ground litter. Table 2 lists the taxa monitored during the study by sampling method, as well as the functional groups they represent. Sampling occurred at three time points during the growing season for visual observations, sticky cards, and pitfall traps: late vegetative stage (V-stage), immediately after anthesis (pollen shed, Rl), and postanthesis (R3) (Ritchie et al. 1993). Litterbags were set in the field during the late vegetative growth stage, and subsets were collected after anthesis and again near crop maturity (R5).

Table 1.

Field and crop characterization summary for 2004–2006 nontarget arthropod field monitoring

t01_381.gif

Table 2.

Monitored taxa summary, 2004–2006

t02_381.gif

Visual Observations. Visual counts of ladybird beetle and lacewing abundance were taken from 10 observation points within each plot at each sampling time. At each sampling point, 10 adjacent plants were selected, and visual counts were made of all life stages of ladybird beetles and lacewings present on the plants.

Sticky Cards. At each sampling time, 10 sticky cards were set in the sampling area of each plot. Cards were 7.6 by 12.7-cm unbaited yellow sticky cards attached to a stake embedded in the soil between rows. Cards were placed at canopy level in vegetative stage maize and at ear height on subsequent sampling dates. Cards were removed 24 h after placement in 2004 and 72 h after placement in 2005 and 2006. Cards were placed in labeled clear plastic bags for storage until identification of trapped arthropods.

Pitfall Traps and Litterbags. Ten pitfall traps per plot were set at each sampling date. Pitfall traps consisted of plastic cups (ranging from 300 to 473 ml) buried upright to the rim containing a small amount of ethylene glycol. Pitfall traps were set for 24 h, and the contents were collected for arthropod identification. Two sets of 10 litterbags were set in each plot at the late vegetative growth stage. Litterbags consisted of plastic 0.9- to 1.4-kg mesh onion bags (General Bag, Cleveland, OH) with a mesh width (unstretched) of ≈1 cm. Litterbags were filled half full with mulched sterile wheat straw. Litterbags were placed between rows on the soil surface and held in place with a stake or wire flag. At each sampling date, 10 litterbags were removed from each plot (one per sample point) and placed in resealable plastic storage bags for transportation to the extraction units. The contents of each litterbag were placed in a Burlese-Tullgren funnel for extraction of arthropods inhabiting the litter.

Climate data (average temperatures and rainfall) were collected for each field location from the nearest available weather station. Thirty-year averages (1971–2000) were taken from National Oceanic and Atmospheric Administration weather stations located closest to the test sites.

Data Analysis. Software programs used for statistical analyses were SAS V. 9.1 (SAS Institute 2002–2003) and CANOCO V. 4.5 (Microcomputer Power, Ithaca, NY).

The PROC MIXED (SAS) procedure was used to test for treatment and interaction effects on individual taxa. Treatment effects were analyzed by averaging data from all years because sample timing was synchronized by crop phenology. Treatments and sampling times were modeled as fixed factors, whereas location was treated as a random effect. A mixed model with repeated measures and the most appropriate covariance structure was used to account for correlations among the observations. Before the analysis, the data were transformed by using the common logarithm of the counts plus 1, and residual plots and Shapiro-Wilks' W test were performed to examine for data normality. Only individual or groups of taxa with sufficient data that satisfied the assumptions of analysis of variance (ANOVA) were analyzed.

Statistical power to detect a 50% impact on population abundance was calculated for the key taxa for each sampling method. The power was computed for detecting a 50% change in the control mean for the difference of means using the estimated variance of a difference of means from the mixed model analysis.

To determine the 50% change, the least squares mean of the control treatment was back transformed to counts. The counts were divided by two (50% change), and this count change was transformed back to log (count + 1) to determine the change in transformed means that correspond to 50% change in the counts. The power was computed by solving for the t-value for a type II error rate from the following sample size equation (Meinert 1986), where the type I error rate or α was set to 0.05:

e01_381.gif
where Δ2 is the change (in this case 50% of control mean), σ2 is the estimate of the variance, α is the type I error rate, and β is the type II error rate. This equation can be solved for tβ,df as
e02_381.gif
where fi01_381.gif represents the SE of the difference of two means. Let s.e.diff = fi02_381.gifdenote the estimate of the standard of the difference of the control mean minus the comparison mean. Then tβ,df can be expressed as
e03_381.gif
where df corresponds to the degrees of freedom used to estimate the standard error of the difference.

Multivariate analysis was used to look for a general community response to the treatments; this level of analysis was implemented to detect trends that occur in the community and to also identify the key taxa influencing those trends. Every taxon for which data were collected for a particular sampling method was included in the community level analysis. The method of principal response curves (Van den Brink and ter Braak 1999) was used to investigate and describe treatment effects at the community level. The advantage of using a multivariate method is that it summarizes all information on the investigated populations simultaneously, and in doing so, it evaluates the effects of a test substance at the community level (Van den Brink and ter Braak 1999). Principal response curves provide an intuitive graphical summarization of the community abundance relative to a specific control. For each year and trap type, the following model was used:

e04_381.gif
for treatment d (d = 1,2), replication (location) j (j = 1, . . . , 4), growth stage t (t = 1, . . . , Ng), and taxa k (k = 1, . . . ,Ns) and for Ng equal to the number of growth stages and Ns equal to the number of taxa present. In the above equation, ydjtk represents the average abundance per trap of the sampling sites in each plot, and it is the log (of these average counts plus 1) that is modeled. In this model example, treatment 2 is taken to be the control treatment and c2t = 0. The terms rjk, gtk, bk, andcdt are estimated model parameters, and εdjtk represents a random error term.

The above equation consists of a set of taxa weights (bk) and a set of canonical coefficients (cdt), which together define the estimated abundance of taxa (k) in treatment (d) at growth stage (t), expressed as a difference with the control abundance in treatment (d) at growth stage (t). The value exp(bk × cdt) gives the estimated relative abundance of the treatment group (d) to the control group for a particular taxa (k) and stage (t). For the analysis of data, the canonical coefficients correspond to the first principal response curve (PRC) for each treatment and are plotted (yl-axis) against time (x-axis), whereas the taxa weights are presented on the y2-axis. When the canonical coefficient (y-axis) is greater than zero and ataxon has a positive weight (or when both the canonical coefficient and the taxon weight are negative), the taxon abundance is expected to be greater in the treatment group than in the untreated control. Van den Brink and ter Braak (1999) reported the taxa weights only when they are above +0.5 or less than -0.5, because taxa having weights with an absolute value <0.5 are unlikely to show a meaningful response that is similar to the overall community response captured in the first principal response curve. All taxa weights are reported in these analyses. Analyses were conducted across all sampling dates in each year using a statistical significance level of P = 0.05.

Estimates of taxon weights and canonical coefficients were obtained using CANOCO V. 4.5. In addition, permutation tests (499 permutations) for significance of the first canonical axis were performed in CANOCO using data from all of the sampling dates.

Results

Climatic Conditions Summary. A summary of average high and low temperatures and rainfall during the 2004–2006 growing seasons for each location is presented in Table 3. The average temperatures for each growing season from April to October were within 2°C of the 30-yr average for all locations. Rainfall averages were within 20% of the 30-yr averages for each location with the following exceptions: York, NE, in 2004, which received 68% of the 30-yr average rainfall during April-October; Scott County, IA, in 2005, which received 41% of the April-October 30-yr average rainfall; and the 2006 Wisconsin location, which received 78% of the April-October 30-yr average rainfall.

Community Level Analyses. The method of PRCs (Van den Brink and ter Braak 1999) was used to study and describe treatment effects at the community level. Taxa counts were summarized and analyzed across visual observations and sticky and pitfall traps; the litterbag analysis was conducted separately because of the asynchrony of sample collection compared with other trap types and the large taxa counts observed by that sampling method. Figure 1 gives the canonical coefficients and taxa weights for the first PRCs for the visual, sticky, and pitfall trap analysis. P values for a test of significance of the first canonical axis are included by year, and taxa weights were calculated based on a meta-analysis of all 3 yr of data. No statistically significant differences were detected between the Cry1F maize hybrids and the non-Bt controls in any year of the study. The first canonical axis accounted for 62.3% of the total variation of the species- environment relationship in 2004, 47.0% in 2005, and 71.7% in 2006. Taxon weights for 16 of the monitored taxa fell between -0.5 and 0.5, indicating no change or an unrelated response pattern than depicted by the PRC. The PRC graph and corresponding taxon weights indicate a general increase in abundance relative to the non-Bt control in collembola, spiders, oribatid mites, and ladybird beetle larvae. A slight decrease in carabid adults and long-legged flies is suggested by the PRC; however, the canonical coefficients and taxa weights are both small, indicating neither were primary drivers of the community response.

Table 3.

Monthly climate summary of field locations monitored from 2004 to 2006

t03_381.gif

Fig. 1.

PRCs and taxon weights of invertebrate communities exposed to CrylF maize compared with isogenic non-Bt control maize as measured by visual observation, sticky cards, and pitfall traps. Responses of taxa with positive weights and positive canonical coefficients (y-axis) showed greater abundance than the control; taxa with negative weights and positive canonical coefficients were less abundant. Organisms with taxon weights between -0.5 and 0.5 generally either show no response or a response that is unrelated to the PRC pattern.

f01_381.eps

Fig. 2.

PRCs and taxon weights of invertebrate communities exposed to CrylF maize compared with isogenic non-Bt control maize as measured by litterbags. Responses of taxa with positive weights and positive canonical coefficients (y-axis) showed greater abundance than the control; taxa with negative weights and positive canonical coefficients were less abundant. Organisms with taxon weights between -0.5 and 0.5 generally either show no response or a response that is unrelated to the PRC pattern.

f02_381.eps

Figure 2 gives the canonical coefficients and taxa weights for the first PRCs for the litterbag analysis. P values for a test of significance of the first canonical axis are included by year, and taxa weights were calculated based on a meta-analysis of all 3 yr of data. No statistically significant differences were detected between the CrylF maize hybrids and the non-Bt controls in any year of the study. The first canonical axis accounted for 61.9% of the total variation of the species-environment relationship in 2004, 88.3% in 2005, and 83.2% in 2006. Species weights for three of the monitored taxa fell between -0.5 and 0.5, indicating no change or an unrelated response pattern than depicted by the PRC. The PRC graph and corresponding table of species weights show that the largest drivers of the PRC were ground beetles in 2005, which generally had higher abundance in 2005 in the CrylF maize treatment than the untreated control.

Table 4.

Statistical summary of individual taxon abundance data 2004 (Nebraska, Iowa, Indiana, and Wisconsin) as monitored by visual surveillance, sticky cards, and pitfall traps

t04_381.gif

Table 5.

Statistical summary of individual taxon abundance data 2005 (Nebraska, Iowa, Indiana, and Wisconsin) as monitored by visual surveillance, sticky cards, and pitfall traps

t05_381.gif

Individual Taxa Analyses: By-year Analyses. Summaries of the statistical analyses for each taxon monitored by visual observations, sticky cards, and pitfall traps (by year) are provided in Tables 46. There were no statistically significant treatment main effects detected for any of the monitored taxa between the CrylF maize hybrids and the isogenic controls in the single year analyses. Investigation of the treatment by stage interaction by individual years showed no statistically significant effects in any taxon with the exception of rove beetle larvae (pitfall) in 2005 (P = 0.04), where abundance was greater in the CrylF maize hybrids. Average counts for rove beetles in 2005 were very low (mean count of less than one organism per plot) in both treatments, and this result does not seem to be biologically significant. Significant sampling stage effects were noted in some of the monitored taxa when studied by year, particularly thrips, which had a significant sampling stage main effect in all 3 yr of study. Examination of the data showed that approximately two to three times more thrips were caught on sticky traps at the vegetative stage sampling period than were caught at the Rl and R3 stage.

Summaries of the statistical analyses for each taxon monitored by litterbags (by year) are provided in Tables 79. There were no statistically significant treatment main effects or treatment by stage interactions detected for any of the monitored taxa between the CrylF maize hybrids and the isogenic controls in the single year analyses. Significant sampling stage effects were noted for several of the monitored taxa, reflecting the increase in colonization of the litterbags that were left in the field until crop maturity.

Meta-analysis. A summary of the meta-analysis of visual observations, sticky card, and pitfall trap abundance data across all 3 yr of study is provided in Table 10. There were no statistically significant treatment main effects detected for any of the monitored taxa between the CrylF maize hybrids and the isogenic controls in the meta-analysis. There were no statistically significant treatment interactions (treatment × stage, treatment × year, or treatment × stage × year) detected between the CrylF maize hybrids and the isogenic controls. Statistically significant sampling stage effects were detected in the herbivorous taxa, the parasitoid taxon, and five of the monitored predators. Investigation of the predator data showed that ladybird beetle larvae, ladybird beetle adults, and Orius counts were very low at the first (v-stage) sampling and increased substantially through the Rl and R3 sampling dates. Conversely, ground beetle adult and spider numbers were higher in the early sampling stage (v-stage) and lower at the last sampling stage (R3). This may relate biologically to the abundance trends of the herbivorous taxa monitored in this study. Ground and plant-dwelling predators like ground beetles and spiders were in greater numbers during the sampling stage where thrips and leafhopper abundance was highest (v-stage, with diminishing abundance through the Rl and R3 sampling stages). Likewise, aphid predators such as ladybird beetles and Orius increased in density over sampling stages as did aphid abundance. Some of these predators are also facultative pollen feeders, which also may explain their increased numbers during the Rl and R3 sampling stages. Ladybird beetle larvae also showed a significant year effect and sampling stage × year interaction, driven by high larval abundance in the Rl and R3 sampling stages in 2004. Ladybird beetle larvae counts were consistently lower in 2005 and 2006.

Table 6.

Statistical summary of individual taxon abundance data 2006 (Nebraska, Iowa, Indiana, and Wisconsin) as monitored by visual surveillance, sticky cards, and pitfall traps

t06_381.gif

A summary of the meta-analysis of litterbag abundance data across all 3 yr of study is provided in Table 11. There were no statistically significant treatment main effects detected for any of the monitored taxa between the Cry1F maize hybrids and the isogenic controls in the meta-analysis. There were no statistically significant treatment interactions (treatment × stage, treatment × year, or treatment × stage × year) detected between the CrylF maize hybrids and the isogenic controls. Significant sampling stage effects were noted for many of the taxa, driven primarily by the increase in litterbag colonization over time.

Table 7.

Statistical summary of individual taxon abundance data 2004 (Nebraska, Iowa, Indiana, and Wisconsin) as monitored by litterbags

t07_381.gif

Table 8.

Statistical summary of individual taxon abundance data 2005 (Nebraska, Iowa, Indiana, and Wisconsin) as monitored by litterbags

t08_381.gif

Discussion

Experimental Design. The experimental design of this study evaluated paired fields at four distinct geographies over 3 yr of study. A perceived weakness is the lack of replication at each location. However, it was considered more important to sample in larger fields to eliminate potential effects of neighboring plots that may occur in smaller research plot settings. This design has been used successfully in other field monitoring studies (Perry et al. 2003, Torres and Ruberson 2005), although the scale of treatments (e.g., independent fields) is debatable and likely influenced by geography and the biology of the taxa of interest (Duffield and Aebischer 1994, Kennedy et al. 2001, Perry et al. 2003). Treating location as a random effect assumes the locations sampled are representative of U.S. Corn Belt maize growing environments.

Table 9.

Statistical summary of individual taxon abundance data 2006 (Nebraska, Iowa, Indiana, and Wisconsin) as monitored by litterbags

t09_381.gif

Table 10.

Statistical summary of individual taxon abundance data 2004–2006 (Nebraska, Iowa, Indiana, and Wisconsin) as monitored by visual surveillance, sticky cards, and pitfall traps

t10_381.gif

Table 11.

Statistical summary of individual taxon abundance data 2004–2006 (Nebraska, Iowa, Indiana, and Wisconsin) as monitored by litterbags

t11_381.gif

The statistical power to detect a 50% change in taxa abundance is included in the “by year” and “across years” summary tables (Tables 411). When considered over sampling stages and all years of the study, analyses of taxa with high abundance such as elongate collembola, mites, leafhoppers, thrips, parasitic hymentoptera, spiders, ground beetles, lacewing eggs, and Orius had very little probability of a type II error. Other taxa such as globular collembola, ladybird beetle adults, and larvae also had high power values, indicating little chance of erroneously failing to reject the null hypothesis. As might be expected, those taxa with the lowest abundance (centipedes/millipedes, ground beetle larvae, lacewing larvae, ladybird beetle eggs, and rove beetles) had the lowest statistical power in the study. However, the 95% confidence intervals reported in the tables are for the differences of the control and Bt means on the transformed scale, all of which contain zero, indicating there are no significant differences between the control and Bt mean counts.

Taxa Grouping and Abundance. Because laboratory safety testing does not suggest CrylF protein has broad spectrum insecticidal activity (U.S. EPA 2001), these field monitoring studies were designed to conduct general surveillance of nontarget arthropod abundance. The selected taxa and subsequent groupings were based on their abundance in corn (Bitzer et al. 2005, Dively 2005, Prasifka et al. 2005, Rose and Dively 2007) and their functional roles to serve as indicators of treatment effects. Although the taxa groupings in this study likely would not detect impacts on individual species, monitoring the abundance of functional groups or family-level monitoring can be an effective means of assessing impacts of insecticidal compounds (Bitzer et al. 2005; Dively 2005, Rose and Dively 2007).

The results of this study are consistent with other field studies evaluating the nontarget effects of Cryl proteins in maize. Field studies such as Dively (2005) and Daley and Buntin (2005) evaluated maize containing CrylAb in field plots using similar nontarget collection techniques (e.g., plant evaluations, sticky traps, pitfall traps) and found few differences between Bt and non-Bt plots. The differences that were noted were attributed to indirect causes such as control of target pests that serve as hosts or prey. This corroborates the conclusions of Pilcher et al. (2005), where no significant differences in generalist predators were noted between Bt and non-Bt plots; however, a reduction in Macrocentrus cingulum, a specialist parasitoid of European corn borer, was noted in Bt plots. Meta analyses conducted on Bt maize containing Cryl proteins such as those reported in Marvier et al. (2007) and Wolfenbarger et al. (2008) also support the conclusion that Cryl activity is limited to selected lepidopteran insects, and impacts to nontarget organisms are largely attributable to reductions in host/prey.

In summary, the multiple test locations used in this study are representative of the major maize cultivation regions of the United States. These studies were conducted in commercial scale agricultural environments over 3 yr using commercially available Cry1F maize hybrids. Nontarget arthropod abundance showed similar dynamics between maize fields containing Cry1F maize and non-Bt isogenic maize hybrids. No differences were detected at either the community level or by individual taxa. The probability of detecting at least a 50% impact in abundance was high for most monitored taxa, especially when considered over all 3 yr of monitoring. Results of these field studies support the hypothesis that maize hybrids containing Cry1F protein are not expected to negatively impact nontarget arthropods. These field studies also confirm the expectation from laboratory experiments that no significant undesirable effects on nontarget arthropods are expected from the cultivation of event DAS-Ø15Ø7-1 Cry1F maize hybrids.

References Cited

1.

R. J. Bitzer , M. E. Rice, C. D. Pilcher , C. Pilcher , and W.-K.F. Lam . 2005. Biodiversity and community structure of soil-dwelling springtails (Entognatha: Collembola) in transgenic rootworm Bt corn. Environ. Entomol. 34: 1346–1376. Google Scholar

2.

D. R. Brooks , S. J. Clark, J. N. Perry, D. A. Bohan, G. T. Champion, L. G. Firbank, A. J. Haughton, C. Hawes , M. S. Heard , and I. P. Woiwod . 2005. Invertebrate biodiversity in maize following withdrawal of triazine herbicides. Proc. R. Soc. B. 272: 1497–1502. Google Scholar

3.

CANOCO V. 4.5. 2002. CANOCO reference manual and CanoDraw for Window's users guide: software for canonical community ordination (version 4.5). Microcomputer Power, Ithaca, NY. Google Scholar

4.

T. Daley , and G. D. Buntin . 2005. Effect of Bacillus thuringiensis transgenic corn for Lepidopteran control on nontarget arthropods. Environ. Entomol. 34: 1292–1301. Google Scholar

5.

G. P. Dively 2005. Impact of transgenic VIP3A x Cry1Ab Lepidopteran-resistant field corn on the nontarget arthropod community. Environ. Entomol. 34: 1267–1291. Google Scholar

6.

S. J. Duffield , and N. J. Aebischer . 1994. The effect of spatial scale of treatment with dimethoate on invertebrate population recovery in winter-wheat. J. Appl. Ecol. 31: 263–281. Google Scholar

7.

P. J. Kennedy , K. F. Conrad, J. N. Perry, D. Powell, J. Aegerter, A. D. Todd , K.F.A. Walters , and W. Powell . 2001. Comparison of two field-scale approaches for the study of effects of insecticides on polyphagous predators in cereals. Appl. Soil Ecol. 17: 253–266. Google Scholar

8.

M. D. Lopez , J. R. Prasifka , D. J. Bruck , and L. C. Lewis . 2005. Utility of ground beetle species in field tests of potential nontarget effects of Bt crops. Environ. Entomol. 34: 1317–1324. Google Scholar

9.

M. Marvier , C. McCreedy , J. Regetz , and P. Kareiva . 2007. A meta-analysis of effects of Bt cotton and maize on nontarget invertebrates. Science 316: 1475–1477. Google Scholar

10.

C. L. Meinert 1986. Clinical trials: design, conduct and analysis. Oxford University Press, New York. Google Scholar

11.

M. Metz 2003. Bacillus thuringiensis: a cornerstone of modern agriculture. Food Products Press. Bacillus thuringiensis: A cornerstone of modern agriculture. 2003. CRC Press LLC, Boca Raton FL, USA. Google Scholar

12.

M. O'Callaghan , T. R. Glare , E.P.J. Burgess , and L. A. Malone . 2005. Effects of plants genetically modified for insect resistance on nontarget organisms. Annu. Rev. Entomol. 50: 271–292. Google Scholar

13.

J. N. Perry , P. Rothery, S. J. Clark , M. S. Heard , and C. Hawes . 2003. Design, analysis and statistical power of the farm-scale evaluations of genetically modified herbicide-tolerant crops. J. Appl. Ecol. 40: 17–31. Google Scholar

14.

C. D. Pilcher , M. E. Rice , and J. J. Obrycki . 2005. Impact of transgenic Bacillus thuringiensis corn and crop phenology on five nontarget arthropods. Environ. Entomol. 34: 1302–1316. Google Scholar

15.

J. R. Prasifka , R. L. Hellmich , G. P. Dively , and L. C. Lewis . 2005. Assessing the effects of pest management on nontarget arthropods: the influence of plot size and isolation. Environ. Entomol. 34: 1181–1192. Google Scholar

16.

S. Ritchie , J. J. Hanway , and G. O. Benson . 1993. How a corn plant develops. Iowa State University of Science and Technology Cooperative Extension Service, Ames, IA. Google Scholar

17.

J. Romeis , M. Meissle , and F. Bigler . 2006. Transgenic crops expressing Bacillus thuringiensis toxins and biological control. Nature (Lond.) Biotechnol. 24: 63–71. Google Scholar

18.

R. Rose , and G. P. Dively . 2007. Effects of insecticide-treated and Lepidopteran-active Bt transgenic sweet corn on the abundance and diversity of arthropods. Environ. Entomol. 36: 1254–1268. Google Scholar

19.

SAS Institute.2002–2003. Version 9.1. SAS Institute, Cary, NC. Google Scholar

20.

P. Thorbek , and T. Bilde . 2004. Reduced numbers of generalist arthropod predators after crop management. J. Appl. Ecol. 41: 526–539. Google Scholar

21.

J. B. Torres , and J. R. Ruberson . 2005. Canopy-and ground-dwelling predatory arthropods in commercial Bt and non-Bt cotton fields: patterns and mechanisms. Environ. Entomol. 34: 1242–1256. Google Scholar

22.

[U. S. EPA] U.S. Environmental Protection Agency. 2001. Biopesticides registration action document—Bacillus thuringiensis plant-incorporated protectants( http://www.epa.gov/opp00001/biopesticides/pips/bt_brad.htm). Google Scholar

23.

[U. S. EPA] U.S. Environmental Protection Agency.2005. Environmental risk assessment of plant incorporated protectant (PIP) inert ingredients ( www.epa.gov/oscpmont/sap/meetings/2005/december/pipinerten-vironmentalriskassessment11-18-05.pdf). Google Scholar

24.

P. Van den Brink , and C.J.F. ter Braak . 1999. Principal response curves: analysis of time-dependent multivariate responses of biological community to stress. Environ. Toxicol. Chem. 18: 138–148. Google Scholar

25.

L. L. Wolfenbarger, S. E. Naranjo, J. G. Lundgren, R. J. Bitzer, and L. S. Watrud. 2008. Bt crop effects on functional guilds of non-target arthropods: A meta-analysis. PLoS ONE 3: e2118. Google Scholar
© 2009 Entomological Society of America
Laura S. Higgins, Jonathan Babcock, Paul Neese, Raymond J. Layton, Daniel J. Moellenbeck, and Nicholas Storer "Three-Year Field Monitoring of Cry1F, Event DAS-Ø15Ø7-1, Maize Hybrids for Nontarget Arthropod Effects," Environmental Entomology 38(1), 281-292, (1 February 2009). https://doi.org/10.1603/022.038.0135
Received: 30 November 2007; Accepted: 1 October 2008; Published: 1 February 2009
JOURNAL ARTICLE
12 PAGES


SHARE
ARTICLE IMPACT
Back to Top