Abstract
Over the past decade, the high-dose refuge (HDR) strategy, aimed at delaying the evolution of pest resistance to Bacillus thuringiensis (Bt) toxins produced by transgenic crops, became mandatory in the United States and is being discussed for Europe. However, precopulatory dispersal and the mating rate between resident and immigrant individuals, two features influencing the efficiency of this strategy, have seldom been quantified in pests targeted by these toxins. We combined mark-recapture and biogeochemical marking over three breeding seasons to quantify these features directly in natural populations of Ostrinia nubilalis, a major lepidopteran corn pest. At the local scale, resident females mated regardless of males having dispersed beforehand or not, as assumed in the HDR strategy. Accordingly, 0–67% of resident females mating before dispersal did so with resident males, this percentage depending on the local proportion of resident males (0% to 67.2%). However, resident males rarely mated with immigrant females (which mostly arrived mated), the fraction of females mating before dispersal was variable and sometimes substantial (4.8% to 56.8%), and there was no evidence for male premating dispersal being higher. Hence, O. nubilalis probably mates at a more restricted spatial scale than previously assumed, a feature that may decrease the efficiency of the HDR strategy under certain circumstances, depending for example on crop rotation practices.