Insecticidal control and genetic structure were studied in a strain of the saw-toothed grain beetle, Oryzaephilus surinamensis (L.), initially comprising 69.5% susceptible phenotypes. Populations were exposed for 20 wk to fabric treatments of pirimiphos-methyl in model grain bins under simulated field conditions, which included the presence of 16 kibbled wheat-filled peripheral refuges mounted in 2 rows (4 to a wall) and a central bulk of untreated wheat. Substantial control, related to the extent of treatment (number of walls sprayed), was achieved with all treatments. Between 43.4 and 80.0% of individuals died within 3 d of being introduced into treated bins. By 20 wk, all treated populations were reduced by ≥85.7%, whereas no live beetles were recovered from the most extensively treated bin. More beetles were found in wheat bulks than in all refuges from the same bin. The final distribution of live and dead beetles was predominantly the result of the swift action of the insecticide, especially at the boundaries of treated and untreated areas. Microhabitat preferences also significantly influenced population distribution. Treatments selected for an increased proportion of resistant phenotypes among surviving beetles and even higher proportions in their offspring. This response suggests that the exposure to insecticide in the grain bins was qualitatively different from that experienced in the standard laboratory bioassay. The beetle populations within treated bins were divided into subpopulations with heritable differences in their knockdown proportions. This observation suggests that the insecticide treatment reduced dispersal between subpopulations.