Refuges of non-Bacillus thuringiensis (Bt) cotton are used to delay Bt resistance in several key insect pests. In 2004, I discovered that Bt cotton plants sometimes enter refuges via the seed bag, and hypothesized that this type of gene flow could have important effects on resistance evolution in insect pests. In the research described herein, I investigated the sources of Bt plants in the non-Bt cotton seed supply and assessed the potential implications of this gene flow on pest resistance evolution. I report results from an empirical study of gene flow in 15 non-Bt cotton seed production fields, as well as results from simulation modeling studies of gene flow from one-toxin and two-toxin Bt cotton. The current policy on gene flow from genetically engineered crops in the United States is also reviewed, including the implications of my research findings for policymakers. Key findings of this study included the prominent role of seed-mediated gene flow in the seed-production setting, and the utility of a geographic information system (GIS) ring analysis approach for describing pollen-mediated gene flow in cotton fields. Modeling results indicated that high rates of gene flow of Bt cotton into refuges could have large effects on pest resistance evolution under certain sets of assumptions, particularly in parts of the world where farm-saved seed is planted year after year in cotton fields. It appears that some of these effects could be mitigated by using non-cotton refuges or by using plants that contain linked transgenes that confer multiple toxins. There are no clear regulations in the United States regarding gene flow of Bt cotton into refuge seed or into seed production fields of non-Bt cotton, as Bt cotton has been deregulated following extensive safety testing. Nevertheless, results from this research suggest that limiting gene flow into refuge seed could be important for sustaining the efficacy of Bt cotton against targeted insect pests in regions where refuges are used.

This study aimed to examine farmers’ descriptions of their experiences of cultivating Bt cotton and to see whether or not the farmers’ descriptions of their experiences are in coherence with the purposes with Bt cotton. This was carried out through qualitative semi-structured interviews with Bt cotton farmers in Vidarbha, Maharashtra, and a qualitative content analysis of the farmers’ descriptions. Our conclusions are among other, that the experiences of the farmers are various. Some farmers have experienced what Bt cotton was aimed to lead to; higher yields and reduced use of pesticides while other farmers have experienced the opposite.

Trichogramma limit pest damage to Ord River Irrigation Area (ORIA) cotton crops by killing the developing embryo of their insect host at the egg stage, effectively reducing the number of emergent pests ingesting transgenic tissue. Their impact on the potentially resistant species, Helicoverpa armigera (Hübner), is considered integral to the Insect Resistance Management (IRM) strategy for transgenic cotton production in the ORIA. This thesis examines aspects of Trichogramma ecology pertinent to this strategy. The dominant species of egg parasitoid in ORIA cotton crops is the introduced Trichogramma pretiosum Riley. Other species make up less than one percent of collected specimens. Surveys revealed T. pretiosum has been introduced or adventitiously dispersed to all developed agricultural regions of northern Australia. Several new species were discovered during surveys in more pristine habitats. Based on field collected eggs, T. pretiosum appears to prefer H. armigera over H. punctigera Wallengren as a host in ORIA cotton. Measured as percent parasitism, Trichogramma activity appears highly variable and does not necessarily coincide with periods of peak insect pest density. Host abundance alone does not define conditions suitable for Trichogramma activity. Environmental constraints on wasp survival, such as the impact of temperature, humidity and insecticide applications, limit their effectiveness in biological control. Despite consistently high rates of percentage egg parasitism (60-99%), acceptable pest control is not readily achieved in ORIA cotton without the aid of insecticides as Helicoverpa numbers exceed damage thresholds. Insecticides inhibit Trichogramma considerably, hence the conundrum regarding initiation of insecticidal control. Trichogramma activity is relatively high early season (May to July), significantly limiting the buildup of pests. Trichogramma effectively stifle Helicoverpa population increase following initial pest egg lay at least during high density years. The impact of farming practices, especially insecticide applications, should be avoided early season to ensure pest mortality attributed to Trichogramma egg parasitisation is maximised. Pre-season habitat manipulation to establish large populations of Trichogramma in alternative hosts is advised. Despite being rare in surrounding habitats and suffering near 50% immature mortality in the field, Trichogramma effectively disperse into young crops attractive to ovipositing hosts and display a high intrinsic rate of increase. Spatial patterns of parasitism tend toward heterogeneity and do not necessarily coincide with host spatio-temporal dynamics. Both host abundance patterns and mean rates of parasitism are not good indicators of parasitoid patchiness. Parasitism rates are highest within the middle strata of the plant canopy prior to complete canopy closure despite a similar number of host eggs being available elsewhere in the plant. Pest density declines as the season progresses. However, insecticide applications become necessary if Bt expression in cotton plants wanes and larval damage increases. Measuring parasitism during periods of declining and low host density is prone to inaccuracy due to small sample size, but can be overcome with the use of egg cards. A better indication of parasitoid activity is achieved using egg cards during periods when insecticide applications are possibly required. As Trichogramma are most active in ORIA cotton from morning to early afternoon, insecticide applications if needed should occur outside of these periods. Trichogramma survival is constrained by environmental influences. Adult female T. pretiosum were exposed to ambient conditions in dialysis tubing sleeve cages to test survival and fecundity in cotton fields. Peak survival and fecundity occurs mid season with both life history variants displaying an inverse relationship to temperature. Adult female T. pretiosum survive longest in the field when provided with sustenance however mortality attributed to handling was considerable. Sleeve cages are not effective for survival measurement of small parasitoids if the cages are frequently moved. Trichogramma effectively reduce pest abundance but are clearly hindered by insecticides and hot dry conditions in ORIA cotton crops. The decision to initiate insecticide applications is best delayed unless absolutely necessary to avoid disruption of Trichogramma impact on pests. Parasitoid activity must be carefully monitored if chemical control becomes imminent. The impact of Trichogramma on pest species can then be optimally exploited.

Genetically modified (GM) crops were introduced in South Africa in 1989 and commercially available by 1998. Legislation to control the use of GM crops was only implemented in 1999, with the genetically modified organisms (GMO) act (15 of 1999). In 2012 2.9 million ha of GM crops were planted in South Africa alone. GM Crops, such as Bt maize, are promoted as safer for the environment since no chemical pesticides are needed. However, recently GM crops have been making headlines as more and more studies find adverse effects of these crops on non-target organisms. The effects on aquatic environments have not yet been fully determined, even though traces of Bt residue have been found in water systems surrounding agricultural lands. The aim of this study was to establish the effects of the Bt toxin on fecundity, development and growth of Bulinus tropicus, a freshwater snail. The experiment made use of a static renewal tests to expose B. tropicus to 50 cm2 Bt maize and cotton leaves in 900 ml of synthetic freshwater. The snails were exposed for the duration of one full life cycle (embryo to adult). Endpoints measured included the development, growth, fecundity, and deformities of the reproductive organs. The results obtained showed retarded development and low embryo survival when the snails were exposed to cotton leaves, irrespective of the presence or absence of Bt, indicating to the possibility of trace residues of chemical pesticides may have been present on the leaves. Initial stimulated growth of hatchlings was observed for both Bt cotton and maize exposures, but after sexual maturity has been reached, ‘surplus’ energy was probably shared between growth and fecundity, resulting in a reduction of growth rate. Energy is gained from their diet, thus a sub-optimal diet would result in less energy available to functions such as growth and fecundity. Signs of developmental instability were found in the formation of the shell opening of the snails exposed to Bt. Fecundity decreased significantly after snails had been exposed to Bt maize / cotton leaves. No differences were found in the penis sheath-preputium length ratio, indicating that Bt had no deleterious effects on the reproductive organs. / MSc (Environmental Sciences), North-West University, Potchefstroom Campus, 2014

Genetically modified (GM) crops were introduced in South Africa in 1989 and commercially available by 1998. Legislation to control the use of GM crops was only implemented in 1999, with the genetically modified organisms (GMO) act (15 of 1999). In 2012 2.9 million ha of GM crops were planted in South Africa alone. GM Crops, such as Bt maize, are promoted as safer for the environment since no chemical pesticides are needed. However, recently GM crops have been making headlines as more and more studies find adverse effects of these crops on non-target organisms. The effects on aquatic environments have not yet been fully determined, even though traces of Bt residue have been found in water systems surrounding agricultural lands. The aim of this study was to establish the effects of the Bt toxin on fecundity, development and growth of Bulinus tropicus, a freshwater snail. The experiment made use of a static renewal tests to expose B. tropicus to 50 cm2 Bt maize and cotton leaves in 900 ml of synthetic freshwater. The snails were exposed for the duration of one full life cycle (embryo to adult). Endpoints measured included the development, growth, fecundity, and deformities of the reproductive organs. The results obtained showed retarded development and low embryo survival when the snails were exposed to cotton leaves, irrespective of the presence or absence of Bt, indicating to the possibility of trace residues of chemical pesticides may have been present on the leaves. Initial stimulated growth of hatchlings was observed for both Bt cotton and maize exposures, but after sexual maturity has been reached, ‘surplus’ energy was probably shared between growth and fecundity, resulting in a reduction of growth rate. Energy is gained from their diet, thus a sub-optimal diet would result in less energy available to functions such as growth and fecundity. Signs of developmental instability were found in the formation of the shell opening of the snails exposed to Bt. Fecundity decreased significantly after snails had been exposed to Bt maize / cotton leaves. No differences were found in the penis sheath-preputium length ratio, indicating that Bt had no deleterious effects on the reproductive organs. / MSc (Environmental Sciences), North-West University, Potchefstroom Campus, 2014

Evolution of Helicoverpa armigera resistance to Bacillus thuringiensis (Bt) cotton producing Cry1Ac is progressing in northern China, and replacement of Cry1Ac cotton by pyramided Bt cotton has been considered to counter such resistance. Here, we investigated four of the eight conditions underlying success of the refuge strategy for delaying resistance to Cry1Ac+Cry2Ab cotton, a pyramid that has been used extensively against H.armigera outside China. Laboratory bioassays of a Cry2Ab-selected strain (An2Ab) and a related unselected strain (An) reveal that resistance to Cry2Ab (130-fold) was nearly dominant, autosomally inherited, and controlled by more than one locus. Strong cross-resistance occurred between Cry2Ab and Cry2Aa (81-fold). Weaker cross-resistance (18- to 22-fold) between Cry2Ab and Cry1A toxins was also present and significantly increased survival of An2Ab relative to An on cotton cultivars producing the fusion protein Cry1Ac/Cry1Ab or Cry1Ac. Survival on Cry1Ac+Cry2Ab cotton was also significantly higher in An2Ab than in An, showing that redundant killing on this pyramid was incomplete. Survival on non-Bt cotton did not differ significantly between An2Ab and An, indicating an absence of fitness costs affecting this trait. These results indicate that a switch to three-toxin pyramided cotton could be valuable for increasing durability of Bt cotton in China.

Genetically modified (GM) cotton expressing Cry1Ac proteins was released in South Africa in 1997 for control of the bollworm complex on this crop. No reports of the failure of Bollgard® cotton to control these pests have yet been made. Throughout the world there are concerns about the development of resistance of target pests to Bt cotton due to the use of only one Bt gene. The aim of this study was to determine if Helicoverpa armigera (Lepidoptera: Noctuidae) and Diparopsis castanea (Lepidoptera: Noctuidae) developed resistance to Bt cotton in South Africa. To determine if H. armigera developed resistance, laboratory experiments were conducted to determine the levels of larval survival and development time when feeding on Bt and non–Bt cotton. Bollworm populations were collected on maize and cotton at different sites in South Africa and reared on Bt and non–Bt cotton under laboratory conditions. Results showed that some populations survived on Bt cotton and that a significant proportion of the individuals successfully completed their life cycles on Bt cotton. Surveys were also conducted amongst cotton farmers to determine the levels of compliance to the refuge strategy that has to be implemented by farmers as an insect resistance management (IRM) strategy to delay resistance development. The levels of compliance to refugia requirements were low and farmers generally only started planting refugia several years after they planted Bt cotton for the first time. The development of resistance of H. armigera to Bt cotton in South Africa can possibly be ascribed to non–compliance to the prescribed refuge requirements. No conclusions can be made on resistance of D. castanea to Bt cotton but the relatively long time to mortality of larvae could indicate development of tolerance to Cry1Ac proteins. The new generation Bollgard II® cotton, expressing both Cry1Ac and Cry2Ab2 proteins, has been released in South Africa during the 2010/11 growing season and field observations showed effective control of the bollworm complex at several sites in the country. Monitoring of refuge compliance levels as well as resistance development in the bollworm complex to Bollgard II® cotton is necessary to ensure the future success of GM cotton. / Thesis (M. Environmental Science)--North-West University, Potchefstroom Campus, 2012.

Genetically modified (GM) cotton expressing Cry1Ac proteins was released in South Africa in 1997 for control of the bollworm complex on this crop. No reports of the failure of Bollgard® cotton to control these pests have yet been made. Throughout the world there are concerns about the development of resistance of target pests to Bt cotton due to the use of only one Bt gene. The aim of this study was to determine if Helicoverpa armigera (Lepidoptera: Noctuidae) and Diparopsis castanea (Lepidoptera: Noctuidae) developed resistance to Bt cotton in South Africa. To determine if H. armigera developed resistance, laboratory experiments were conducted to determine the levels of larval survival and development time when feeding on Bt and non–Bt cotton. Bollworm populations were collected on maize and cotton at different sites in South Africa and reared on Bt and non–Bt cotton under laboratory conditions. Results showed that some populations survived on Bt cotton and that a significant proportion of the individuals successfully completed their life cycles on Bt cotton. Surveys were also conducted amongst cotton farmers to determine the levels of compliance to the refuge strategy that has to be implemented by farmers as an insect resistance management (IRM) strategy to delay resistance development. The levels of compliance to refugia requirements were low and farmers generally only started planting refugia several years after they planted Bt cotton for the first time. The development of resistance of H. armigera to Bt cotton in South Africa can possibly be ascribed to non–compliance to the prescribed refuge requirements. No conclusions can be made on resistance of D. castanea to Bt cotton but the relatively long time to mortality of larvae could indicate development of tolerance to Cry1Ac proteins. The new generation Bollgard II® cotton, expressing both Cry1Ac and Cry2Ab2 proteins, has been released in South Africa during the 2010/11 growing season and field observations showed effective control of the bollworm complex at several sites in the country. Monitoring of refuge compliance levels as well as resistance development in the bollworm complex to Bollgard II® cotton is necessary to ensure the future success of GM cotton. / Thesis (M. Environmental Science)--North-West University, Potchefstroom Campus, 2012.

This thesis looks at the divergent policy decisions on the commercial release of two genetically modified (GM) crops in India. Bt cotton was introduced in India in 2002, and has spread widely across the country, though not without controversy. In 2010, the first GM food crop – Bt Brinjal (eggplant) – was put forward for approval. In contrast to the Bt cotton decision, and following heated debate and a series of public consultations across the country, an indefinite moratorium was placed on the crop. In this thesis, I unpack the various factors that shaped both decisions and the politics that accompanied them. To facilitate this analysis, I use a conceptual framework that combines four key forces that are often the focus of food governance analysis, but which are not always considered together. I argue that the divergence in outcomes can be explained by a confluence of shifts in key elements of the policy process. These dimensions of influence can be understood in four related categories: corporate actors, institutional mechanisms, science and science networks, and discursive elements. Changes that took place in each of these dimensions in the period between the two decisions align to explain divergent outcomes that no individual influence could. The conceptual framework I develop in this thesis presents a useful structure to analyse the often-complex and multi-causal processes and outcomes related to food and environmental issues. The results of this research have implications for the future direction of agricultural GMO policy in India, as well as in other countries in the global South.

This thesis looks at the divergent policy decisions on the commercial release of two genetically modified (GM) crops in India. Bt cotton was introduced in India in 2002, and has spread widely across the country, though not without controversy. In 2010, the first GM food crop – Bt Brinjal (eggplant) – was put forward for approval. In contrast to the Bt cotton decision, and following heated debate and a series of public consultations across the country, an indefinite moratorium was placed on the crop. In this thesis, I unpack the various factors that shaped both decisions and the politics that accompanied them. To facilitate this analysis, I use a conceptual framework that combines four key forces that are often the focus of food governance analysis, but which are not always considered together. I argue that the divergence in outcomes can be explained by a confluence of shifts in key elements of the policy process. These dimensions of influence can be understood in four related categories: corporate actors, institutional mechanisms, science and science networks, and discursive elements. Changes that took place in each of these dimensions in the period between the two decisions align to explain divergent outcomes that no individual influence could. The conceptual framework I develop in this thesis presents a useful structure to analyse the often-complex and multi-causal processes and outcomes related to food and environmental issues. The results of this research have implications for the future direction of agricultural GMO policy in India, as well as in other countries in the global South.