For delaying evolution of pest resistance to transgenic corn producing Bacillus thuringiensis (Bt) toxins, limited data are available to compare the effectiveness of refuges of non-Bt corn planted in seed mixtures versus blocks. Here we addressed this issue in the ear-feeding pest Helicoverpa zea Boddie by measuring its survival and development in the laboratory on ears from field plots with 90% Cry1A.105 + Cry2Ab corn and 10% non-Bt corn planted in a seed mixture or blocks. We compared a strain of H. zea selected for resistance to Cry1Ac in the laboratory, its parent strain not selected in the laboratory, and their F1 progeny. The relative survival of the F1 progeny and dominance of resistance were higher on ears from Bt plants in the seed mixture than the block. Half of the kernels in ears from non-Bt plants in the seed mixture produced both Cry1A.105 and Cry2Ab. However, survival on ears from non-Bt plants did not differ between the block and seed mixture. In simulations based on the observed survival, resistance to Cry1A.105 + Cry2Ab corn evolved faster with the seed mixture than the blocks, because of the higher dominance of resistance in the seed mixture. Increasing the refuge percentage improved durability of Cry1A.105 + Cry2Ab corn more for the blocks than the seed mixture. These findings imply that, for a given percentage of non-Bt corn, resistance of H. zea and other ear-feeding pests to multi-toxin Bt corn is likely to evolve faster for seed mixtures than blocks.