Final Report for GW07-004
Project Information
Limiting unwanted gene flow between transgenic Bacillus thuringiensis (Bt) crop varieties and non-Bt varieties is of paramount importance for sustainable agriculture. Such gene flow creates liabilities for farmers and seed producers, threatens the use of refuges for delaying insect resistance to Bt crops, and complicates the removal of Bt transgenes from the environment if unexpected problems arise. To identify potential sources of Bt contamination, we monitored 15 non-Bt cotton seed production fields throughout the 2007 growing season. Our results demonstrate that honey bee abundance, proximity to Bt cotton fields, and human factors contribute to seed contamination of non-Bt cotton fields.
Introduction
Transgenic cotton that produces insecticidal toxins from Bacillus thuringiensis (Bt) comprised 63% of cotton acreage in the United States in 2008 (Economic Research Service 2008). Bt cotton provides substantial control of several key pests and has reduced broad-spectrum insecticide use (Shelton et al. 2002, Cattaneo et al. 2006). While Bt crops offer economic and environmental benefits, their widespread adoption raises concerns that Bt transgenes will contaminate other cotton varieties (Smyth et al. 2002). Such contamination could have negative economic, political and ecological consequences that are detailed below. Moreover, research conducted on today’s Bt crops could provide valuable information for preventing transgene escape from future pharmaceutical and industrial crops.
In 2006, we discovered low rates of Bt transgenes in the seed supply of non-Bt cotton (Heuberger et al., 2008a). We tested 100 seeds from each of 11 previously unopened non-Bt cotton seed bags purchased by Arizona farmers. Seeds were tested for the Bt toxin Cry1Ac using enzyme-linked immunosorbent analysis (ELISA) (ImmunoStripsTM test system, Agdia, Elkhart, IN). Three of the bags were contaminated, each at a rate of 1%. Although rates were low, this raised concern that widespread use of Bt crops and potential fitness advantages of contaminant plants could lead to accumulation of Bt transgenes in the seed supply through time. Risks associated with such contamination include compromising the refuge strategy for insect resistance management, liabilities for farmers, and complication of transgene removal from the environment, as detailed below.
Growers of Bt cotton commonly plant refuges of non-Bt cotton near their Bt fields to delay insect resistance to Bt crops (U.S. EPA 2006, Matten and Reynolds 2003, Carrière et al. 2005). Refuges produce abundant Bt-susceptible insects that can mate with rarely occurring resistant insects, thus delaying the frequency of resistant insects (Gould and Tabashnik 1998, Tabashnik et al. 2004a, Sisterson et al. 2004). Production of Bt toxins in refuges, resulting from introgression of Bt transgenes, decreases the efficacy of refuges by increasing mortality of Bt-susceptible insects (Chilcutt and Tabashnik 2004, Heuberger et al. 2008b). Insect evolution of resistance to Bt toxins is a threat to sustainable agriculture for two reasons: 1) such resistance could drive conventional growers to return to using conventional, more environmentally harmful insecticides, and 2) organic farmers relying on Bacillus thuringiensis sprays could lose this important tool for organic insect control.
Moreover, the presence of Bt seeds in the non-Bt seed supply creates liabilities for farmers, and could be of particular concern if transgenes entered organic crops (Mellon and Rissler 2004, Andow and Zwahlen 2006). Concern about transgene contamination led to collapsing markets for organic canola production in western Canada and North Dakota (Brummond 2001, Smyth et al. 2002). The USDA National Organic Program prohibits products containing transgenes from bearing the “organic” label in the United States (National Organic Program 2006). Transgene contamination also threatens conventional farmers who want to preserve the integrity of their crop varieties. For example, the recent contamination of rice by an experimental variety known as “Liberty Link” triggered a drop in prices of exported rice from the U.S. and resulted in several lawsuits from rice farmers (Vermij 2006).
Unintended introgression of transgenes into crops could greatly complicate their removal from the environment if unanticipated problems occur, such as detrimental impacts on non-target organisms. Transgenic toxins could persist and proliferate in the environment for many years though cross-pollination between transgenic plants and non-transgenic members of the same species. For example, StarLink corn with Bt toxin Cry9C, was approved in the U.S. for livestock feed but not human consumption. However, after being found in human food, its registration was not renewed in 2000 (Lin et al. 2003). Three years after StarLink corn was withdrawn from the market, the Cry9C transgene was still detectable in commercial corn varieties, illustrating the difficulty of expunging transgenes from the environment (Mellon and Rissler 2004).
Contamination of non-Bt cotton by Bt varieties can result from cross-pollination, emergence of volunteer Bt plants in non-Bt fields, and/or seed mixing by human error. Cross-pollination of non-Bt cotton plants by Bt cotton (i.e., “outcrossing”) results in non-Bt plants with some seeds that produce Bt toxin. While cotton is a predominantly self-pollinating crop, outcrossing can occur when insect pollinators, particularly bees, are present (Free 1970, Umbeck et al. 1991). In fact, cotton plants surrounded by cotton of a different variety can outcross at rates up to 10-48% (Free 1970). Studying cross-pollination between Bt cotton fields and non-Bt cotton seed production fields increases our understanding of gene flow in insect pollinated crops and provides a valuable contrast to the more widely-studied wind-pollinated crops.
In addition to outcrossing, gene flow can enter fields via the accidental cultivation of Bt plants in non-Bt fields. Such plants, which are commonly called “adventitious,” can emerge as volunteer plants from previous years crops, or can result from accidental seed mixing (i.e., human error). Adventitious Bt plants are homozygous for the Bt trait unless previous generations outcrossed with non-Bt varieties.
Seed Contamination in the Literature:
Since the widespread adoption of transgenic crops in the mid 1990’s, several reports have documented transgene contamination in the conventional seed supply. A 2004 Union of Concerned Scientists study found genetically engineered seeds in conventional seed bags of canola, corn, and soybeans (Mellon and Rissler 2004). More than half of all varieties tested contained transgenic seeds. Seeds were pooled in the study, but contamination rates were estimated to range from less than 0.05% to greater than 1% of seeds (Mellon and Rissler 2004). In corn, up to 1% contamination by Bt transgenes was detected, although most transgenes detected in the study conferred herbicide resistance (Mellon and Rissler 2004). Of particular interest, soybean varieties were contaminated at levels similar to corn and canola. This was a surprising result because soybean varieties are almost entirely self-pollinating, in contrast to the highly outcrossed corn and canola crops. The similarity in contamination rates among these crops suggests that sources other than outcrossing were responsible for transgene introductions (Mellon and Rissler 2004).
Another study tested 10 cultivars of conventional canola for contamination by herbicide resistance transgenes and reported that all cultivars were contaminated (Friesen et al. 2003). Nine cultivars had less than 1% of seeds with herbicide resistance transgenes whereas one cultivar contained 3-5% herbicide-resistant seeds (Friesen et al. 2003). Similarly, bags of transgenic glyphosate-resistant canola seed contained up to 0.3% seeds with transgenic resistance to both glyphosate and glufosinate herbicides (Beckie et al. 2003).
Conventional crops have also been contaminated by experimental transgenic events. Between 2001 and 2005, hundreds of tons of seed from Bt10, a transgenic variety of corn not approved for sale, were accidentally sold to farmers (Macilwain 2005). In another event, an experimental transgene from rice was found in the foundation seed stock of a conventional rice variety (Vogel 2006). A common feature of these contamination events is that the source of transgene introduction was not identified. Mechanisms of transgene introgression are extremely difficult to trace in hindsight. Thus, we saw a need for a study that monitored fields throughout the growing season to evaluate multiple potential sources of transgene contamination.
Here, we monitored 15 non-Bt cotton seed production fields throughout the growing season, to determine the pathways by which Bt contamination was entering. We hypothesized that bee abundance, proximity to Bt cotton fields, and presence of adventitious Bt plants in fields would enhance cross-pollination of seed by Bt cotton. We also hypothesized that we would encounter 1) Bt plants emerging as volunteers in fields and, 2) pre-existing contamination in the seed bags used to plant fields; and that these factors would result in adventitious plants in seed lots. Our goal was to develop a predictive model to identify fields with particularly high risk of Bt contamination.
Knowledge gained in this project will be useful for seed companies and policy makers in designing methods that limit Bt contamination of non-Bt cotton seed. Insights from this study could also be broadly applicable to understanding crop contamination by other transgenic plants, including those producing pharmaceutical or industrial compounds.
- 1) Determine the importance of seed bags as a source of adventitious Bt plants.
2) Identify factors that influence the level of Bt outcrossing in non-Bt cotton seed production fields. Specifically, examine adventitious Bt plants, landscape attributes, and bee activity as potential explanatory variables. Also examine the potential for human error to contribute to the presence of adventitious Bt plants.
3) Identify changes needed in seed production guidelines to reasonably limit contamination.
Cooperators
Research
Objective 1:
Two-hundred cotton seeds were tested from each of the six seed lots that were used in planting the non-Bt cotton seed production fields surveyed. Seed bags originated from Arizona, California, and Texas. We used ELISA (enzyme-linked immunosorbent assay) to test seeds for presence/absence of the Bt toxin Cry1Ac. Commercially available lateral flow test strips (Immunostrips®, Agdia Inc., Elkhart, Indiana) were used to perform ELISA tests according to the manufacturer’s guidelines. Pools of 25 cotton seeds were tested together, with pools of 24 non-Bt seeds plus one Bt seed serving as positive controls, and pools of 25 non-Bt seeds serving as negative controls. All controls produced expected results, supporting the accuracy of these tests. For pools testing positive, archived portions of the individual seeds were tested, to quantify Bt seeds in the seed pool. Because a low rate of contamination was expected, this pooling technique greatly decreased the cost of testing compared to testing individual seeds. Results from these assays were used in calculating the overall percent contamination of seed bags.
Objective 2:
With permission from growers, we conducted surveys of contamination in 15 non-Bt cotton fields in Arizona (i.e., private farms) that were being grown under seed contract. We focused on Bt contamination at the edge of fields, to increase our likelihood of detecting cross-pollination and, therefore, our ability to understand factors that augment cross-pollination. Multiple studies have detected maximum outcrossing at the edge of fields (Umbeck et al. 1991, Llewellyn and Fitt 1996, Zhang et al. 2005).
As outcrossing in cotton is predominantly mediated by bees, we monitored bee activity throughout the flowering season with trapping and visual monitoring. Propylene glycol traps were used to catch bees throughout the flowering period in fields. Traps, which consisted of quart-sized paint cans with a yellow funnel on top to attract bees and propylene glycol (pet safe antifreeze) to preserve them, were placed at the four corners and four edges of each field. Traps were emptied every two weeks. Visual monitoring was also conducted in fields every two weeks throughout flowering. I walked along three rows per field, including the two edge rows, counting the number of open flowers, and the number of flowers that contained a foraging bee. The overall density of bees per flower was calculated for each field at the end of the season by summing the total number of bees sighted foraging in flowers, and dividing that by the total number of observed flowers. This density was also calculated individually for honey bees (Apis mellifera), and then for native bees (any bees that were not A. mellifera).
The vast majority of bees identified in visual monitoring (88.3%) were honey bees (A. mellifera). As it turned out, our propylene glycol traps were not efficient at catching honey bees. Moreover, we did not end up placing traps in fields where beekeeping was present, due to concerns about negatively impacting the beekeeping operation. Therefore results from visual monitoring, and not from trapping, were used in all data analyses.
We examined the possibility of volunteer Bt plants using field observations. Soon after plant emergence and before rows were cultivated, we walked along several rows of each field, quantifying residual lint in the soil and cotton plants that emerged outside of the rows.
GIS maps of all Bt and non-Bt cotton grown in Arizona for the 2007 season were obtained from the Arizona Cotton Research and Protection Council (ACRPC). These maps were used to identify the Bt cotton field nearest to each non-Bt cotton seed production field. The shortest distance between each field and the nearest Bt cotton field was measured using GIS (MapInfo ProfessionalTM, version 6.5).
For each field, immediately prior to harvest, we sampled 25 cotton bolls from the four outer edges (25 bolls x 4 edges = 100 bolls per field). Bolls were selected from a 25m section at the center of each edge, as defined using a GPS unit (eTrex Legend, Garmin®). No two bolls were ever collected from the same plant. Although we only collected bolls that appeared to be full-sized and mature, we found that some bolls did not contain mature, viable seeds of a size that could be tested. Such bolls were particularly common from fields that were harvested early in the season by growers or that had severe infestations by the pink bollworm. Thus, the number of bolls tested from fields ranged from 67 to 96.
Seeds from bolls were tested for the Bt toxin Cry1Ac to estimate the percentage of Bt-outcrossed cotton seeds and adventitious Bt plants in fields. ELISA analyses were conducted with ImmunostripsTM. We followed the suggested protocol included with the Immunostrips, but increased extraction time of seed pools to 2 h, because the increased extraction yielded clearer test lines on the strips. For a boll, we began by testing subsampled portions of 10 impartially selected seeds as a pool. For bolls with fewer than 10 mature seeds, the reduced number of tested seeds was recorded. The total number of seeds tested from individual fields ranged from 670 to 960. When bolls tested positive for the toxin, we tested archived portions of the 10 seeds individually to quantify Bt seeds. We also tested maternal tissue (i.e., the fruit wall) from bolls with Bt positive seeds to differentiate between outcrossing and adventitious Bt plants. Bolls with Bt positive seeds were defined as “outcrossed” if the maternal tissue was negative for Bt, and as “adventitious” if maternal tissue was positive. Positive and negative control seed pools or fruit walls were run alongside all sample replicates.
Statistical Analysis:
Using individual fields as the experimental unit, logistic regression was used to explore outcrossing as a function of adventitious plants, distance from the nearest Bt field, and bee density. Logistic regression, which predicts the odds of an event (e.g., outcrossing), is the preferred method for analyzing data with a binomial distribution (i.e., proportions composed of yes/no data) (Ramsey and Schaffer 2002). Logistic regression was also used to determine whether contamination of the planted seed and volunteer plants likely resulted in adventitious plants.
Objective 3:
Using data from Objectives 1 and 2, we outlined conditions that led to increased contamination of seed fields. Also, to draw more specific conclusions regarding the effects of various factors on gene flow, we are currently working on a computer simulation model that predicts accumulation of transgenes in fields. The model, designed in Visual Basic (Microsoft Excel), calculates ratios of Bt versus non-Bt pollen in fields based on their levels of adventitious Bt plants. Resulting “pollen clouds” outcross existing plants in the field, and set proportions of pollen clouds can be exchanged between fields to simulate bee movement between fields. The model assigns proportions of outcrossed bolls and adventitious Bt plants to fields based on cross-pollination, and predicts changes over multiple field seasons (i.e., outcrossed seeds become adventitious plants the following year). This model is not yet complete. Before publishing the model, we plan to add a component where the fitness of Bt versus non-Bt seed in fields is adjusted for insect damage. Insects are expected to damage non-Bt seeds more than Bt seeds, which could favor accumulation of transgenes. We are currently working on an experiment with insects on Bt and non-Bt seeds that will help us to gain the needed estimates.
Outreach:
An extension publication with our results and suggestions will soon be distributed to extension professionals, growers, and members of the seed production industry.
RESULTS AND DISCUSSION
Of the 15 monitored non-Bt cotton fields, 13 were planted with varieties that are transgenic for resistance to glyphosate herbicides. Cotton varieties that are totally free of transgenes are rare in Arizona. The 15 monitored fields were planted with six seed lots, of which two were contaminated with detectable levels of Bt transgenes. Sampled bags from these seed lots contained 0.5% and 20% Bt seeds, respectively. After encountering the seed bag with 20% contamination, we sampled seed from a second seed bag from the same seed lot. We tested 25 seeds from the second bag, and detected a contamination rate of 28%, indicating that contamination is relatively consistent throughout a seed lot. The presence of contamination in seeds that are used to plant seed production fields indicates that Bt contamination persists over multiple generations.
Field boll samples collected at the end of the season revealed both adventitious Bt plants and Bt-outcrossed cotton seeds. However, the extent of outcrossing was low. Although all samples were collected from the edge of fields, the average percentage of Bt outcrossed seeds produced in non-Bt cotton fields was only 0.23% of tested seeds (95% confidence interval: 0.092% to 0.37%). With logistic regression, we modeled outcrossing as a function of honey bee density and proximity to Bt cotton fields (χ2 = 6.63, p = 0.036). We excluded the two fields that were planted with an extremely contaminated seed lot (20% contamination) in this model, as contaminant plants may enhance cross-pollination and diminished the influence of other factors (see below).
In an analysis assessing the effect of honey bees and distance of Bt fields, density of honey bees per flower was positively associated with the odds of outcrossing in fields (honey bee density, arcsine square root transformed: χ2 = 4.75, p = 0.029), while the odds of outcrossing decreased as the distance to the closest Bt cotton fields increased (χ2 = 6.60, p = 0.010). When the effect of native bees (bees other than honey bees) was added to the analysis, there appeared to be no effect of native bees on outcrossing, after accounting for honey bees and distance to Bt fields (native bee density, arcsine square root transformed: χ2 = 0.32, p = 0.57). Native bee densities were always low (< 5 bees per 1000 flowers), which may have prevented us from detecting an effect. Honey bees comprised 88.3% of bees sighted in flowers in this study. The vast majority of insects sighted moving from flower to flower were bees, with moths and wasps spotted only on rare occasion.
As expected, fields with adventitious Bt plants had increased numbers of Bt-outcrossed non-Bt seeds. In a logistic regression analysis that included all 15 fields, adventitious plants (arsine square root transformed) significantly increased the odds of non-Bt seed becoming outcrossed (χ2 = 4.5, p = 0.027), after accounting for proximity to Bt cotton and honey bee density. However, in this model, proximity to Bt fields and honey bee density were not statistically significant (p = 0.27 and 0.25, respectively), indicating that adventitious plants, when abundant, overwhelm the usual association between bee abundance and proximity to Bt cotton fields and outcrossing. In the two fields planted with contaminated seed, 0.40% (95% confidence interval = -2.5% to 3.3%) of seeds from non-Bt plants were outcrossed by Bt. This outcrossing rate was within the range discovered in other fields, but was higher (particularly for one of the fields) than would be expected based on distance from Bt fields and bee densities alone. Thus, when a non-Bt field contains a high proportion of Bt plants, it appears that significant outcrossing between Bt and non-Bt plants in that field will take place.
Adventitious Bt plants were encountered in eight of the 15 fields, but the vast majority of adventitious bolls sampled (88.7%) were from three of the fields. Thus, adventitious plants were rare in most fields, but highly abundant in some fields. Two of the three fields with high instances of adventitious plants were planted from the highly contaminated (20% contaminated) seed lot. These fields were contaminated at rates of 16.9 and 22.8% of tested plants (average = 19.9%); corresponding almost perfectly to the percent contamination in the planted seed. Adventitious plants in the third highly contaminated field had a different origin. In our sample from that field, only one field edge contained adventitious Bt plants, and all plants sampled from that edge were Bt. Since bolls were only tested from field edges, we do not know how many rows at that end of the field were contaminated. This observation indicates accidental planting of the wrong variety on one edge of the field, and underlines the importance of limiting human error in seed production. The remaining five fields with adventitious plants averaged only 1.6% of sampled plants in the fields. The source of the low contamination levels is unknown.
We note that the misplanted field was not excluded from logistic regression analyses for outcrossing, because the contaminated edge of the field was all Bt, and thus there was no potential for outcrossing of non-Bt plants at that edge. Moreover, the contaminated field edge bordered a Bt field. Thus, the misplanting probably did not significantly change the distance of other plants in the field from Bt cotton.
Bolls from adventitious Bt plants belonged to two categories: those containing only Bt seed in the subsample of seeds tested, and those with both Bt and non-Bt seed in the subsample of seeds tested. Bolls with only Bt seed likely resulted from homozygous Bt plants. Bt cotton varieties marketed in the U.S. are homozygous for Bt genes (Jayaraman 2005, Adamczyk and Meredith 2006). Production of mature non-Bt seeds in Bt plants demonstrates hemizygosity, because Bt expression is dominantly inherited (Zhang et al. 2000, Heuberger et al. 2008b). Hemizygous plants must have resulted from unwanted cross-pollination between Bt and non-Bt cotton in previous generations. Seven hemizygous bolls, originating from four fields, were encountered, and contained an average of 77% Bt cotton seeds. This fits the expected 3:1 ratio of Bt expression in seeds from self-pollinating hemizygous Bt plants (Zhang et al. 2000).
Our discovery of hemizygous Bt plants in fields lends further support to our hypothesis that cross-pollination is an important factor in Bt introgression into the non-Bt cotton seed supply. However, of adventitious plants identified in the non-Bt cotton seed production fields, only 11.3% were hemizygous, indicating that other mechanisms of transgene entry in previous generations, such as human error, contributed as well.
We did not find compelling evidence that volunteer plants are an important contamination source. While eight of the sampled fields contained residual cotton lint in the soil at the time of planting, we only observed plants occurring outside of rows in four of the fields during our observations soon after planting, and such plants were very rare (> 2 plants per monitored row). Moreover, these plants could have resulted from imperfections in the planting machinery. When all fields were included in a logistic regression analysis, it appeared that fields with more residual lint at planting had higher odds of adventitious plants (χ2 = 46.3, p < 0.0001). However, the three fields with markedly high rates of adventitious plants also had residual lint at planting and, as noted above, adventitious plants in those fields probably resulted from contamination in the planted seed and from planting errors. When the three highly contaminated fields are omitted from analysis, there is no longer evidence of a trend between residual lint and adventitious plants (χ2 = 0.530, p = 0.467), suggesting that volunteer plants did not play an important role in contamination.
To conclude, we have found that non-Bt cotton fields that are near Bt cotton fields or have high abundances of bees have the highest probability of becoming contaminated by outcrossing. Thus, we recommend that growers space their non-Bt cotton seed production fields as far from their Bt cotton fields as is feasible, particularly if fields occur in an area where bees are abundant. Honey bees may be more important cross-pollinators of cotton than native bees, but more data are needed to support that hypothesis. We have also found that contamination occurs through human error, as was demonstrated in the sampled field where an entire edge was planted to the wrong cotton variety. Once contamination has entered a seed field via cross pollination or other mechanisms, it may result in contamination of subsequent years’ seed crops if it goes unnoticed. Moreover, adventitious Bt plants can cross-pollinate surrounding non-Bt plants. If resulting seeds have an advantage against insect pests, this may lead to an accumulation of transgenes over time, even if new transgenes are not introduced to the field after the initial contamination event. We are working on a model to predict contamination over the years in seed lots, based on these mechanisms.
While this study reports the presence of Bt plants in non-Bt cotton fields that are grown under seed contract, it is important to note that this cotton may not have been sold for seed. Cotton fields that are grown under contract for seed production are commonly rejected by the seed companies due to product volume or quality concerns. At least one of our cooperating growers has had seed rejected in the past due to transgene contamination. We do not have information on whether or not seeds from this study were accepted. Nevertheless, our data are useful for understanding the mechanisms by which Bt transgenes enter the seed supply.
LITERATURE CITED
Adamczyk J. J. Jr and W. R. Meredith (2006) Selecting for efficacy of Bollgard cotton cultivars against various Lepidoptera using forward breeding techniques. J. Econ. Entomol. 99: 1835-1841.
Andow, D. A. and C. Zwahlen (2006). Assessing the environmental risks of transgenic plants. Ecol. Lett. 9: 196-214.
Beckie, H. J., S. I. Warwick, H. Nair and G. Séguin-Swartz (2003). Gene flow in commercial fields of herbicide-resistant canola (Brassica napus). Ecol. Appl. 13: 1276-1294.
Brummond, B. (2001). Genetically modified organisms: How they affect organic agriculture in North Dakota. Situation Statement. Walsh County Extension, North Dakota State University. http://www.ag.ndsu.nodak.edu/biotech/papers/brummond.pdf. Accessed 12/21/06.
Carrière, Y., C. Ellers-Kirk, K. Kumar, S. Heuberger, M. Whitlow, L. Antilla, T. J. Dennehy and B. E. Tabashnik (2005). Long-term evaluation of compliance with refuge requirements for Bt cotton. Pest Manag. Sci. 61: 327-330.
Cattaneo M. G., C. Yafuso, C. Schmidt, C. Huang, M. Rahman, C. Olson, C. Ellers-Kirk, B. J. Orr, S. E. Marsh, L. Antilla, P. Dutilleul and Y.
Carrière (2006). Farm-scale evaluation of transgenic cotton impacts on biodiversity, pesticide use, and yield. Proc. Natl. Acad. Sci. USA 103: 7571-7576.
Chilcutt, C. F. and B. E. Tabashnik (2004). Contamination of refuges by Bacillus thuringiensis toxin genes from transgenic maize. Proc. Natl. Acad. Sci. USA 101: 7526-7529.
Economic Research Service, U.S. Dept. of Agriculture (2008). Adoption of genetically engineered crops in the U.S.: Extent of adoption. http://ww.ers.usda.gov/data/BiotechCrops/adoption.htm. Accessed 8/25/2008.
Free, J. B. (1970). Insect pollination of crops. Academic Press Inc., New York, N.Y.
Friesen, L. F., A. G. Nelson and R. C. Van Acker (2003). Evidence of contamination of pedigreed canola (Brassica napus) seedlots in western Canada with genetically engineered herbicide resistance traits. Agron. J. 95: 1342-1347.
Gould, F. and B. E. Tabashnik (1998). Bt-cotton resistance management, pp. 67-97. In M. Mellon and J. Rissler (eds.), Now or never: Serious new plans to save a natural pest control. Union of Concerned Scientists, Cambridge, MA.
Heuberger, S., C. Yafuso, B.E. Tabashnik Y. Carrière, and T.J. Dennehy (2008a) Does outcrossing between Bt and non-Bt cotton refuges decrease with distance? Environ. Biosafety Res. 7:87-96.
Heuberger, S., C. Kirk, C. Yafuso, A. Gassman, B.E. Tabashnik, T.J. Dennehy, and Y. Carrière (2008b). Impacts of Refuge Contamination by Bt Transgenes on Pectinophora gossypiella (Lepidoptera: Gelechiidae) Resistance. J. Econ. Entomol. 101:504-514.
Jayaraman K. S. (2005) Monsanto’s Bollgard potentially compromised in India. Nat. Biotechnol. 23: 1326.
Lin, W., G. K. Price and E. W. Allen (2003). Starlink: Impacts on the U.S. corn market and world trade. Agribusiness 19: 473-488.
Llewellyn, D. and G. Fitt (1996). Pollen dispersal from two field trials of transgenic cotton in the Namoi Valley, Australia. Mol. Breed. 2: 157-166.
Macilwain, C. (2005). News: Stray seeds had antibiotic-resistance genes. Nature 434: 548.
Matten, S. R. and A. H. Reynolds (2003). Current resistance management requirements for Bt cotton in the United States. J. New Seeds 5: 137-178.
Mellon, M. and J. Rissler (2004). Gone to seed: Transgenic contaminants in the traditional food supply. Union of Concerned Scientists, Cambridge.
National Organic Program (NOP) (2006). NOP regulations (standards) and guidelines Q&A document. U.S. Department of Agriculture, Agricultural Marketing Service. http://www.ams.usda.gov/NOP/Q&A.html. Accessed 1/5/07.
Shelton, A. M., J. Z. Zhao and R. T. Roush (2002). Economic, ecological, food safety, and social consequences of the deployment of Bt transgenic plants. Annu. Rev. of Entomol. 47: 845-881.
Ramsey, F. L., and D. W. Schafer (2002). The statistical sleuth: a course in methods of data
analysis, Second edition. Duxburry Press, New York.
Sisterson, D. M., L. Antilla, Y. Carrière, C. Ellers-Kirk and B. E. Tabashnik (2004). Effect of insect population size on evolution of resistance to transgenic crops. J. Econ. Entomol. 97: 1413-1424.
Smyth, S., G. G. Khachatourians and P. W. B. Phillips (2002). Liabilities and economics of transgenic crops. Nat. Biotechnol. 20: 537-541.
Umbeck, P. F., K. A. Barton, E. V. Nordheim, J. C. McCarty, W. L. Parrott and J. N. Jenkins (1991). Degree of pollen dispersal by insects from a field test of genetically engineered cotton. J. Econ. Entomol. 84(6): 1943-1950.
U.S. Environmental Protection Agency (2006). Introduction to biotechnology regulation for pesticides. http://www.epa.gov/pesticides/biopesticides/regtools/biotech-reg-prod.htm. Accessed 1/7/07.
Vermij, P. (2006). Liberty Link rice raises specter of tightened regulations. Nat. Biotechnol. 24: 1301-1302.
Vogel, G. (2006). Tracing the transatlantic spread of GM rice. Science 313: 1714.
Zhang, B., T. Guo and Q. Wang (2000). Inheritance and segregation of exogenous genes in transgenic cotton. J. Genet. 79: 71-75.
Zhang, B., X. Pan, T. Guo, Q. Wang and T. A. Anderson (2005). Measuring gene flow in the cultivation of transgenic cotton (Gossypium hirsutum L.). Molec. Biotechnol. 31: 11-19.
Research Outcomes
Education and Outreach
Participation Summary:
We have prepared a draft of a newsletter that will be distributed within the next few months. We are waiting for it to be reviewed by a few more colleagues for their feedback before distributing it to a broader list of growers, extension agents, and seed production professionals. We are also currently preparing a manuscript that we plan to submit to a peer reviewed journal. Before we submit the manuscript, we plan to perform spatially explicit analyses using our GIS maps of the fields. Please refer to our previous publications (Heuberger et al. 2008 a,b) for a description of our previous research that led to the development of this project.