Abstract
Widely grown transgenic crops producing insecticidal proteins from Bacillus thuringiensis (Bt) can benefit agriculture, but adaptation by pests threatens their continued success. Refuges of host plants that do not make Bt toxins can promote survival of susceptible insects and delay evolution of resistance, particularly if resistance is inherited as a recessive trait. However, data have been lacking to compare the dominance of resistance when Bt and non-Bt seeds are planted in random mixtures versus separate blocks. Here we report results from greenhouse experiments with transgenic cotton producing Bt toxin Cry1Ac and the bollworm, Helicoverpa zea, showing that the dominance of resistance was significantly higher in a seed mixture relative to a block of Bt cotton. The proportion of larvae on non-Bt cotton plants in the seed mixture was also significantly higher than expected under the null hypothesis of random distribution. In simulations based on observed survival, resistance evolved 2- to 4.5-fold faster in the seed mixture relative to separate blocks of Bt and non-Bt cotton. These findings support previous modelling results indicating that block refuges may be more effective than seed mixtures for delaying resistance in pests with mobile larvae and inherently low susceptibility to the toxins in Bt crops.