Modeling evolution of resistance to pyriproxyfen by the sweetpotato whitefly (Homoptera: Aleyrodidae)

We used computer simulations to examine evolution of resistance to the insect growth regulator (IGR) pyriproxyfen by the sweetpotato whitefly, Bemisia tabaci (Gennadius), biotype B [=Bemisia argentifolii (Bellows & Perring)]. Consistent with trends seen in cotton (Gossipyium spp.) fields in Arizona and Israel, results suggest that evolution of resistance to pyriproxyfen may occur rapidly in this haplodiploid insect. Similar to results from models of diploid insects, resistance evolved faster with increases in toxin concentration, dominance of resistance in females, the initial frequency of the resistance allele, and the proportion of the region treated with pyriproxyfen. Resistance was delayed by fitness costs associated with resistance. Movement between treated and untreated cotton fields had little effect, probably because untreated cotton leaves provided internal refuges in treated fields and whiteflies were controlled with other insecticides in external refuges. Resistance evolved faster when susceptibility to pyriproxyfen was greater in susceptible males than susceptible females. In contrast, resistance evolved slower when susceptibility to pyriproxyfen was greater in resistant males than resistant females. Results suggest that growers may be able to prolong the usefulness of pyriproxyfen by applying lower toxin concentrations and promoting susceptible populations in refuges.     

Haplodiploidy, sex, and the evolution of pesticide resistance

The assumption that males and females are equally tolerant to pesticides in haplodiploid arthropods led to the prediction that the evolution of resistance is faster in haplodiploid than in diploid arthropods. However, in this review, it was found that the ratio of male to female tolerance is substantially smaller in haplodiploid than in diploid arthropods, indicating that resistance alleles are not strongly up-regulated in haploid males. In addition, males were generally less tolerant than females in both haplodiploid and diploid arthropods. Factors such as sexual size dimorphism and sex-dependent selection may account for the lower tolerance in males than in females. Little among-population variation in the ratio of male to female tolerance was found in three species. Moreover, the tolerance ratio generally remained unchanged by selection for resistance to pesticides, although significant among-species variation was present within arthropod orders. This indicates that sexual dimorphism in pesticide tolerance evolves at a slower rate than resistance to pesticides. Simulations considering between-sex differences in pesticide tolerance showed that resistance evolution can be slower in haplodiploids than in diploids. Recessive resistance, low male tolerance to pesticides, fitness costs expressed in males, and the use of refuges contributed in substantially delaying the evolution of resistance in haplodiploid arthropods. These findings cast a new perspective on the evolution of pesticide resistance in haplodiploid herbivores and natural enemies.     

Effects of entomopathogenic nematodes on evolution of pink bollworm resistance to Bacillus thuringiensis toxin Cry1Ac

The evolution of resistance by pests can reduce the efficacy of transgenic crops that produce insecticidal toxins from Bacillus thuringiensis (Bt). However, fitness costs may act to delay pest resistance to Bt toxins. Meta-analysis of results from four previous studies revealed that the entomopathogenic nematode Steinernema riobrave (Rhabditida: Steinernematidae) imposed a 20% fitness cost for larvae of pink bollworm, Pectinophora gossypiella (Saunders) (Lepidoptera: Gelechiidae), that were homozygous for resistance to Bt toxin Cry1Ac, but no significant fitness cost was detected for heterozygotes. We conducted greenhouse and laboratory selection experiments to determine whether S. riobrave would delay the evolution of pink bollworm resistance to Cry1Ac. We mimicked the high dose/refuge scenario in the greenhouse with Bt cotton (Gossypium hirsutum L.) plants and refuges of non-Bt cotton plants, and in the laboratory with diet containing Cry1Ac and refuges of untreated diet. In both experiments, half of the replicates were exposed to S. riobrave and half were not. In the greenhouse, S. riobrave did not delay resistance. In the laboratory, S. riobrave delayed resistance after two generations but not after four generations. Simulation modeling showed that an initial resistance allele frequency > 0.015 and population bottlenecks can diminish or eliminate the resistance-delaying effects of fitness costs. We hypothesize that these factors may have reduced the resistance-delaying effects of S. riobrave in the selection experiments. The experimental and modeling results suggest that entomopathogenic nematodes could slow the evolution of pest resistance to Bt crops, but only under some conditions.     

Insect resistance to Bt crops: evidence versus theory

Evolution of insect resistance threatens the continued success of transgenic crops producing Bacillus thuringiensis (Bt) toxins that kill pests. The approach used most widely to delay insect resistance to Bt crops is the refuge strategy, which requires refuges of host plants without Bt toxins near Bt crops to promote survival of susceptible pests. However, large-scale tests of the refuge strategy have been problematic. Analysis of more than a decade of global monitoring data reveals that the frequency of resistance alleles has increased substantially in some field populations of Helicoverpa zea, but not in five other major pests in Australia, China, Spain and the United States. The resistance of H. zea to Bt toxin Cry1Ac in transgenic cotton has not caused widespread crop failures, in part because other tactics augment control of this pest. The field outcomes documented with monitoring data are consistent with the theory underlying the refuge strategy, suggesting that refuges have helped to delay resistance.     

Effects of four nematode species on fitness costs of pink bollworm resistance to Bacillus thuringiensis toxin Cry1Ac

Evolution of resistance by pests can reduce the efficacy oftransgenic crops that produce insecticidal toxins from the bacterium Bacillus thuringiensis Berliner (Bt). In conjunction with refuges of non-Bt host plants, fitness costs can delay the evolution of resistance. Furthermore, fitness costs often vary with ecological conditions, suggesting that agricultural landscapes can be manipulated to magnify fitness costs and thereby prolong the efficacy of Bt crops. In the current study, we tested the effects of four species of entomopathogenic nematodes (Steinernematidae and Heterorhabditidae) on the magnitude and dominance of fitness costs of resistance to Bt toxin CrylAc in pink bollworm, Pectinophora gossypiella (Saunders) (Lepidoptera: Gelechiidae). For more than a decade, field populations of pink bollworm in the United States have remained susceptible to Bt cotton Gossypium hirsutum L. producing CrylAc; however, we used laboratory strains that had a mixture of susceptible and resistant individuals. In laboratory experiments, dominant fitness costs were imposed by the nematode Steinernema riobrave Cabanillas, Poinar, and Raulston but no fitness costs were imposed by Steinernema carpocapsae Weiser, Steinernema sp. (ML18 strain), or Heterorhabditis sonorensis Stock, Rivera-Ordu?o, and Flores-Lara. In computer simulations, evolution of resistance to Cry1Ac by pink bollworm was substantially delayed by treating some non-Bt cotton refuge fields with nematodes that imposed a dominant fitness cost, similar to the cost observed in laboratory experiments with S. riobrave. Based on the results here and in related studies, we conclude that entomopathogenic nematodes could bolster insect resistance management, but the success of this approach will depend on selecting the appropriate species of nematode and environment, as fitness costs were magnified by only two of five species evaluated and also depended on environmental factors.     

Fees or refuges: which is better for the sustainable management of insect resistance to transgenic Bt corn?

The evolution of resistance in insect pests will imperil the efficiency of transgenic insect-resistant crops. The currently advised strategy to delay resistance evolution is to plant non-toxic crops (refuges) in close proximity to plants engineered to express the toxic protein of the bacterium Bacillus thuringiensis (Bt). We seek answers to the question of how to induce growers to plant non-toxic crops. A first strategy, applied in the United States, is to require Bt growers to plant non-Bt refuges and control their compliance with requirements. We suggest that an alternative strategy is to make Bt seed more expensive by instituting a user fee, and we compare both strategies by integrating economic processes into a spatially explicit, population genetics model. Our results indicate that although both strategies may allow the sustainable management of the common pool of Bt-susceptibility alleles in pest populations, for the European corn borer (Ostrinia nubilalis) one of the most serious pests in the US corn belt, the fee strategy is less efficient than refuge requirements.     

Effects of gossypol on fitness costs associated with resistance to Bt cotton in pink bollworm

Fitness costs associated with insect resistance to Bacillus thuringiensis (Bt) crops may help to delay or prevent the spread of resistance alleles, especially when refuges of non-Bt host plants are present. The potential for such delays increases as the magnitude and dominance of fitness costs increase. Here, we examined the idea that plant secondary chemicals affect expression of fitness costs associated with resistance to Bt cotton in Pectinophora gossypiella (Saunders). Specifically, we tested the hypotheses that gossypol affects the magnitude or dominance of fitness costs, by measuring performance of three independent sets of pink bollworm populations fed artificial diet with and without gossypol. Each set had an unselected susceptible population, a resistant population derived by selection from the susceptible population, and the F1 progeny of the susceptible and resistant populations. No individuals completed development on diets with gossypol in one set, suggesting that these individuals partially lost the ability to detoxify this chemical. In the other two sets, costs affecting survival did not support the hypotheses, but costs affecting pupal weight did. Adding gossypol to diet increased the magnitude and dominance of costs affecting pupal weight. In one of the two sets with survivors on diet with gossypol, costs affecting development time were less recessive when gossypol was present in diet. These results indicate that gossypol increased the magnitude and dominance of some fitness costs. Better understanding of the effects of natural plant defenses on fitness costs could improve our ability to design refuges for managing insect resistance to Bt crops.     

Fitness costs associated with Cry1F resistance in the European corn borer

Crops producing insecticidal toxins derived from the bacterium Bacillus thuringiensis (Bt) are widely planted to manage insect pests. Bt crops can provide an effective tool for pest management; however, the evolution of Bt resistance can diminish this benefit. The European corn borer, Ostrinia nubilalis Hübner, is a significant pest of maize and is widely managed with Bt maize in the Midwest of the United States. When Bt crops are grown in conjunction with non‐Bt refuges, fitness costs of Bt resistance can delay the evolution of resistance. Importantly, fitness costs often vary with ecological factors, including host‐plant genotype and diapause. In this study, we examined fitness costs associated with Cry1F resistance in O. nubilalis when insects were reared on three maize lines. Fitness costs were tested in two experiments. One experiment assessed the fitness costs when Cry1F‐resistant and Cry1F‐susceptible insects were reared on plants as larvae and experienced diapause. The second experiment tested resistant, susceptible and F1 heterozygotes that were reared on plants but did not experience diapause. Despite some evidence of greater adult longevity for Cry1F‐resistant insects, these insects produced fewer fertile eggs than Cry1F‐susceptible insects, and this occurred independent of diapause. Reduced fecundity was not detected among heterozygous individuals, which indicated that this fitness cost was recessive. Additionally, maize lines did not affect the magnitude of this fitness cost. The lower fitness of Cry1F‐resistant O. nubilalis may contribute to the maintenance of Cry1F susceptibility in field populations more than a decade after Cry1F maize was commercialized.     

Early detection of field-evolved resistance to Bt cotton in China: cotton bollworm and pink bollworm

Transgenic crops producing Bacillus thuringiensis (Bt) toxins kill some major insect pests, but pests can evolve resistance and thereby reduce the effectiveness of such Bt crops. The main approach for slowing pest adaptation to Bt crops uses non-Bt host plants as "refuges" to increase survival of susceptible pests. To delay evolution of pest resistance to cotton producing Bt toxin Cry1Ac, several countries have required refuges of non-Bt cotton, while farmers in China have relied on "natural" refuges of non-Bt host plants other than cotton. This strategy is designed for cotton bollworm (Helicoverpa armigera), which attacks many crops and is the primary target of Bt cotton in China, but it does not apply to pink bollworm (Pectinophora gossypiella), which feeds almost entirely on cotton in China. Here we review evidence of field-evolved resistance to Cry1Ac by cotton bollworm in northern China and by pink bollworm in the Yangtze River Valley of China. For both pests, results of laboratory diet bioassays reveal significantly decreased susceptibility of field populations to Cry1Ac, yet field control failures of Bt cotton have not been reported. The early detection of resistance summarized here may spur countermeasures such as planting Bt cotton that produces two or more distinct toxins, increased planting of non-Bt cotton, and integration of other management tactics together with Bt cotton.     

Sustained susceptibility of pink bollworm to Bt cotton in the United States

Evolution of resistance by pests can reduce the benefits of transgenic crops that produce toxins from Bacillus thuringiensis (Bt) for insect control. One of the world's most important cotton pests, pink bollworm (Pectinophora gossypiella), has been targeted for control by transgenic cotton producing Bt toxin Cry1Ac in several countries for more than a decade. In China, the frequency of resistance to Cry1Ac has increased, but control failures have not been reported. In western India, pink bollworm resistance to Cry1Ac has caused widespread control failures of Bt cotton. By contrast, in the state of Arizona in the southwestern United States, monitoring data from bioassays and DNA screening demonstrate sustained susceptibility to Cry1Ac for 16 y. From 1996-2005, the main factors that delayed resistance in Arizona appear to be abundant refuges of non-Bt cotton, recessive inheritance of resistance, fitness costs associated with resistance and incomplete resistance. From 2006-2011, refuge abundance was greatly reduced in Arizona, while mass releases of sterile pink bollworm moths were made to delay resistance as part of a multi-tactic eradication program. Sustained susceptibility of pink bollworm to Bt cotton in Arizona has provided a cornerstone for the pink bollworm eradication program and for integrated pest management in cotton. Reduced insecticide use against pink bollworm and other cotton pests has yielded economic benefits for growers, as well as broad environmental and health benefits. We encourage increased efforts to combine Bt crops with other tactics in integrated pest management programs.     

Managing insecticide resistance by mass release of engineered insects

Transgenic crops producing insecticidal toxins are now widely used to control insect pests. The benefits of this method would be lost if resistance to the toxins spread to a significant proportion of the pest population. The primary resistance management method, mandatory in the United States, is the high-dose/ refuge strategy, requiring toxin-free crops as refuges near the insecticidal crops, and the use of toxin doses sufficiently high to kill insects heterozygous for a resistance allele, thereby rendering resistance functionally recessive. We propose that mass-release of harmless susceptible (toxin-sensitive) insects could substantially delay or even reverse the spread of resistance. Mass-release of such insects is an integral part of release of insects carrying a dominant lethal (RIDL), a method of pest control related to the sterile insect technique. We show by mathematical modeling that specific RIDL strategies could form an effective component of a resistance management strategy for plant-incorporated protectants and other toxins.     

Delaying evolution of insect resistance to transgenic crops by decreasing dominance and heritability

The refuge strategy is used widely for delaying evolution of insect resistance to transgenic crops that produce Bacillus thuringiensis (Bt) toxins. Farmers grow refuges of host plants that do not produce Bt toxins to promote survival of susceptible pests. Many modelling studies predict that refuges will delay resistance longest if alleles conferring resistance are rare, most resistant adults mate with susceptible adults, and Bt plants have sufficiently high toxin concentration to kill heterozygous progeny from such matings. In contrast, based on their model of the cotton pest Heliothis virescens, Vacher et al. (Journal of Evolutionary Biology, 16, 2003, 378) concluded that low rather than high toxin doses would delay resistance most effectively. We demonstrate here that their conclusion arises from invalid assumptions about larval concentration-mortality responses and dominance of resistance. Incorporation of bioassay data from H. virescens and another key cotton pest (Pectinophora gossypiella) into a population genetic model shows that toxin concentrations high enough to kill all or nearly all heterozygotes should delay resistance longer than lower concentrations.     

Fitness cost of resistance to Bt cotton linked with increased gossypol content in pink bollworm larvae

Fitness costs of resistance to Bacillus thuringiensis (Bt) crops occur in the absence of Bt toxins, when individuals with resistance alleles are less fit than individuals without resistance alleles. As costs of Bt resistance are common, refuges of non-Bt host plants can delay resistance not only by providing susceptible individuals to mate with resistant individuals, but also by selecting against resistance. Because costs typically vary across host plants, refuges with host plants that magnify costs or make them less recessive could enhance resistance management. Limited understanding of the physiological mechanisms causing fitness costs, however, hampers attempts to increase costs. In several major cotton pests including pink bollworm (Pectinophora gossypiella), resistance to Cry1Ac cotton is associated with mutations altering cadherin proteins that bind this toxin in susceptible larvae. Here we report that the concentration of gossypol, a cotton defensive chemical, was higher in pink bollworm larvae with cadherin resistance alleles than in larvae lacking such alleles. Adding gossypol to the larval diet decreased larval weight and survival, and increased the fitness cost affecting larval growth, but not survival. Across cadherin genotypes, the cost affecting larval growth increased as the gossypol concentration of larvae increased. These results suggest that increased accumulation of plant defensive chemicals may contribute to fitness costs associated with resistance to Bt toxins.     

Insect resistance to transgenic Bt crops: lessons from the laboratory and field

Transgenic crops that produce insecticidal toxins from the bacterium Bacillus thuringiensis (Bt) grew on >62 million ha worldwide from 1996 to 2002. Despite expectations that pests would rapidly evolve resistance to such Bt crops, increases in the frequency of resistance caused by exposure to Bt crops in the field have not yet been documented. In laboratory and greenhouse tests, however, at least seven resistant laboratory strains of three pests (Plutella xylostella [L.], Pectinophora gossypiella [Saunders], and Helicoverpa armigera [Hübner]) have completed development on Bt crops. In contrast, several other laboratory strains with 70- to 10,100-fold resistance to Bt toxins in diet did not survive on Bt crops. Monitoring of field populations in regions with high adoption of Bt crops has not yet detected increases in resistance frequency. Resistance monitoring examples include Ostrinia nubilalis (Hübner) in the United States (6 yr), P. gossypiella in Arizona (5 yr), H. armigera in northern China (3 yr), and Helicoverpa zea (Boddie) in North Carolina (2 yr). Key factors delaying resistance to Bt crops are probably refuges of non-Bt host plants that enable survival of susceptible pests, low initial resistance allele frequencies, recessive inheritance of resistance to Bt crops, costs associated with resistance that reduce fitness of resistant individuals relative to susceptible individuals on non-Bt hosts ("fitness costs"), and disadvantages suffered by resistant strains on Bt hosts relative to their performance on non-Bt hosts ("incomplete resistance"). The relative importance of these factors varies among pest-Bt crop systems, and violations of key assumptions of the refuge strategy (low resistance allele frequency and recessive inheritance) may occur in some cases. The success of Bt crops exceeds expectations of many, but does not preclude resistance problems in the future.     

Early warning of cotton bollworm resistance associated with intensive planting of Bt cotton in China

Transgenic crops producing Bacillus thuringiensis (Bt) toxins kill some key insect pests, but evolution of resistance by pests can reduce their efficacy. The predominant strategy for delaying pest resistance to Bt crops requires refuges of non-Bt host plants to promote survival of susceptible pests. To delay pest resistance to transgenic cotton producing Bt toxin Cry1Ac, farmers in the United States and Australia planted refuges of non-Bt cotton, while farmers in China have relied on "natural" refuges of non-Bt host plants other than cotton. Here we report data from a 2010 survey showing field-evolved resistance to Cry1Ac of the major target pest, cotton bollworm (Helicoverpa armigera), in northern China. Laboratory bioassay results show that susceptibility to Cry1Ac was significantly lower in 13 field populations from northern China, where Bt cotton has been planted intensively, than in two populations from sites in northwestern China where exposure to Bt cotton has been limited. Susceptibility to Bt toxin Cry2Ab did not differ between northern and northwestern China, demonstrating that resistance to Cry1Ac did not cause cross-resistance to Cry2Ab, and implying that resistance to Cry1Ac in northern China is a specific adaptation caused by exposure to this toxin in Bt cotton. Despite the resistance detected in laboratory bioassays, control failures of Bt cotton have not been reported in China. This early warning may spur proactive countermeasures, including a switch to transgenic cotton producing two or more toxins distinct from Cry1A toxins.     

Modelling the spatial configuration of refuges for a sustainable control of pests: a case study of Bt cotton

The 'high-dose-refuge' (HDR) strategy is widely recommended by the biotechnology industry and regulatory authorities to delay pest adaptation to transgenic crops that produce Bacillus thuringiensis (Bt) toxins. This involves cultivating nontoxic plants (refuges) in close proximity to crops producing a high dose of Bt toxin. The principal cost associated with this strategy is due to yield losses suffered by farmers growing unprotected, refuge plants. Using a population genetic model of selection in a spatially heterogeneous environment, we show the existence of an optimal spatial configuration of refuges that could prevent the evolution of resistance whilst reducing the use of costly refuges. In particular, the sustainable control of pests is achievable with the use of more aggregated distributions of nontransgenic plants and transgenic plants producing lower doses of toxin. The HDR strategy is thus suboptimal within the context of sustainable agricultural development.     

Increased frequency of pink bollworm resistance to Bt toxin Cry1Ac in China

Transgenic crops producing insecticidal proteins from Bacillus thuringiensis (Bt) kill some key insect pests, but evolution of resistance by pests can reduce their efficacy. The main approach for delaying pest adaptation to Bt crops uses non-Bt host plants as "refuges" to increase survival of susceptible pests. To delay evolution of pest resistance to transgenic cotton producing Bt toxin Cry1Ac, the United States and some other countries have required refuges of non-Bt cotton, while farmers in China have relied on "natural" refuges of non-Bt host plants other than cotton. The "natural" refuge strategy focuses on cotton bollworm (Helicoverpa armigera), the primary target of Bt cotton in China that attacks many crops, but it does not apply to another major pest, pink bollworm (Pectinophora gossypiella), which feeds almost entirely on cotton in China. Here we report data showing field-evolved resistance to Cry1Ac by pink bollworm in the Yangtze River Valley of China. Laboratory bioassay data from 51 field-derived strains show that the susceptibility to Cry1Ac was significantly lower during 2008 to 2010 than 2005 to 2007. The percentage of field populations yielding one or more survivors at a diagnostic concentration of Cry1Ac increased from 0% in 2005-2007 to 56% in 2008-2010. However, the median survival at the diagnostic concentration was only 1.6% from 2008 to 2010 and failure of Bt cotton to control pink bollworm has not been reported in China. The early detection of resistance reported here may promote proactive countermeasures, such as a switch to transgenic cotton producing toxins distinct from Cry1A toxins, increased planting of non-Bt cotton, and integration of other management tactics together with Bt cotton.     

Application of pyramided traits against Lepidoptera in insect resistance management for Bt crops

Since initial launch of insect protected transgenic crops, the most effective strategy to manage the potential for target pests to evolve resistance has been the use of a single mode of action with "high dose" and structured refuge. However, the effectiveness of this strategy is limited if mortality of certain pests does not reach "high dose" criteria, inconsistent implementation of refuges and non-rare resistance alleles. More recently, several pyramided trait products, which include multiple modes of action against key target pests, have been developed. These products offer the potential for dramatically improved resistance management with smaller refuges and less dependence on high mortality of susceptible and heterozygous insects and rare resistance alleles. We show that products such as SmartStax and PowerCore offer compelling resistance management benefits compared with single mode of action products and allow for the option of products containing refuge seed mixtures rather than structured refuges to effectively delay resistance. We conclude that all stakeholders, including technology developers, growers, crop advisors, extensions services and regulatory authorities should continue to encourage the development, deployment and adoption of pyramided trait products for improved pest management and improved resistance management.     

Evolution of resistance to transgenic crops: interactions between insect movement and field distribution

The refuge strategy is designed to delay evolution of pest resistance to transgenic crops producing Bacillus thuringiensis Berliner (Bt) toxins. Movement of insects between Bt crops and refuges of non-Bt crops is essential for the refuge strategy because it increases chances that resistant adults mate with susceptible adults from refuges. Conclusions about optimal levels of movement for delaying resistance are not consistent among previous modeling studies. To clarify the effects of movement on resistance evolution, we analyzed simulations of a spatially explicit model based partly on the interaction of pink bollworm, Pectinophora gossypiella (Saunders), with Bt cotton. We examined resistance evolution as a function of insect movement under 12 sets of assumptions about the relative abundance of Bt cotton (50 and 75%), temporal distribution of Bt cotton and refuge fields (fixed, partial rotation, and full rotation), and spatial distribution of fields (random and uniform). The results show that interactions among the relative abundance and distribution of refuges and Bt cotton fields can alter the effects of movement on resistance evolution. The results also suggest that differences in conclusions among previous studies can be explained by differences in assumptions about the relative abundance and distribution of refuges and Bt crop fields. With fixed field locations and all Bt cotton fields adjacent to at least one refuge, resistance evolved slowest with low movement. However, low movement and fixed field locations favored rapid resistance evolution when some Bt crop fields were isolated from refuges. When refuges and Bt cotton fields were rotated to the opposite crop type each year, resistance evolved fastest with low movement. Nonrecessive inheritance of resistance caused rapid resistanceevolution regardless of movement rate. Confirming previous reports, results described here show that resistance can be delayed effectively by fixing field locations and distributing refuges uniformly to ensure that Bt crop fields are not isolated from refuges. However, rotating fields provided better insect control and reduced the need for insecticide sprays.     

Evolution of resistance to Bt crops: directional selection in structured environments

Abstract Widespread use of transgenic crops that express an insecticidal endotoxin from Bacillus thuringiensis increases the risk of evolution of resistance by the European corn borer and other insect pests. To delay resistance evolution, the high‐dose/refuge strategy is being implemented for Bt maize and Bt cotton. We develop a general modelling framework to understand the invasion and spread of alleles conferring resistance. We show that at least three processes are involved in explaining the effectiveness of the high‐dose/refuge strategy: the intensity of selection, assortative (non‐random) mating due to spatial subdivision, and variation in male mating success also due to spatial subdivision. Understanding these processes leads to a greater range of possible resistance management tactics. For example, efforts to encourage adults to leave their natal fields may have the unwanted effect of speeding rather than slowing resistance evolution. Furthermore, when Bt maize causes high mortality to susceptible target pests, spraying insecticides in refuges to reduce pest populations may not greatly disrupt resistance management.