Main

The refuge strategy has two critical assumptions: that inheritance of resistance is recessive, and that random mating occurs between susceptible and resistant insects. If resistance is recessive, hybrid firstgeneration (F1) offspring produced by matings between susceptible and resistant adults are killed by eating Bt plants. If mating is random, initially rare homozygous resistant adults emerging from Bt plants are likely to mate with the more abundant homozygous susceptible adults emerging from non-Bt plants, producing hybrid F1progeny that cannot survive on Bt plants. Mathematical modelsand limited data from laboratory and greenhouse studies indicate that resistance can be delayed substantially when these assumptions are valid4,7.

Previous work on the feasibility of the refuge strategy has focused on inheritance of resistance, spatial proximity of refuges relative to transgenic crops, and refuge size4,7. To achieve random mating, however, resistant adults from Bt plants and susceptible adults from non-Bt plants must emerge synchronously4,7. We tested the inheritance of resistance and synchrony for the pink bollworm by measuring survival and developmental rates of a laboratory-selected resistant strain, a susceptible strain, and their hybrid F1progeny on Bt cotton and non-Bt cotton.

Consistent with one of the assumptions of the refuge strategy, we find that pink bollworm resistance to Bt cotton is recessive. Survival of the hybrid F1progeny (2%) was not higher than survival of the susceptible strain (6%), and both were markedly lower than survival of the resistant strain (37%) (G-test, Gadj=24.8, d.f.=1, P<0.001). in the only other case in which inheritance of resistance was studied using Bt plants, resistance was also recessive8. These results differ from the non-recessive resistance to Bt toxins in artificial diet seen in a laboratory-selected strain of European corn borer9.

Resistant larvae on Btcotton required an average of 5.7 days longer to develop than susceptible larvae on non-Bt cotton (Fig. 1). Field data suggest that the median longevity of male pink bollworms is less than a week10, and laboratory results show that 80% of moths mate within three days of emergence2. This developmental asynchrony therefore favours assortative mating among resistant moths from Bt plants. In the field, the extent of developmental asynchrony and assortative mating would be affected by variation in toxin expression, weather and overlap between generations.

Figure 1: Development of pink bollworm larvae on Btand non-Bt cotton plants.
figure 1

Pieces of paper, each with 20-40 eggs, were placed under the bracts of 74 bolls of 16 Bt cotton plants (Delta Pine 50B) and 10 non-Bt cotton plants (Delta Pine 50) in a greenhouse. After a week, we counted a total of 1,411 entrance holes made by neonates. After two weeks, bolls were caged, and twice a week we counted mature larvae that had exited from bolls. For each strain, survival on Bt cotton was estimated by adjusting for mortality on non-Bt cotton using Abbott's correction. The resistant strain (APHIS-98R) developed significantly more slowly in bolls of Bt cotton (29.8±1.1 days) than the susceptible strain (APHIS-S) did in bolls of non-Bt cotton (24.1 ± 0.9) (t-test, t=3.89, d.f.=47, P<0.001). developmental time on non-Bt cotton did not vary significantly among the resistant strain, the susceptible strain and the F1progeny.

Full size image

Assortative mating would generate a disproportionately high number of homozygous resistant insects, accelerating the evolution of resistance. This effect would be diminished if the slower development of resistant larvae increased mortality associated with overwintering or other factors. Computer simulations show that interactions between developmental asynchrony and season length increase uncertainty because they either hasten or slow the evolution of resistance11. There are no reports of resistance to Bt crops in the field, but our results indicate that developmental asynchrony must be considered in efforts to sustain this technology.