The wheat stem sawfly, Cephus cinctus Norton, causes severe losses in wheat grown in the northern Great Plains. Much of the affected area is planted in monoculture with wheat, Triticum aestivum L., grown in large fields alternating yearly between crop and no-till fallow. The crop and fallow fields are adjacent. This cropping landscape creates pronounced edge effects of sawfly infestations and may be amenable to trap cropping using existing agricultural practices. The behavioral preference for two wheat varieties was assessed in the context of developing trap crops for this insect. In field nurseries, stem lodging assessments indicated that the cultivar ‘Conan’ was infrequently damaged, whereas ‘Reeder’ was often heavily damaged. In laboratory choice and no-choice tests, ‘Reeder’ was significantly preferred by ovipositing wheat stem sawfly females. These two cultivars did not differ significantly in height or developmental stage, factors known to impact sawfly preference. Although Conan received fewer eggs than Reeder in no-choice tests, oviposition was further reduced in choice tests, indicating that females clearly preferred Reeder. In field trials where the overall dimensions of the spatial structure in choice tests was varied, females always selected Reeder over Conan in alternating block, row, and interseeded planting scenarios. Reeder releases greater amounts of the attractive compound, (Z)-3-hexenyl acetate than Conan but is similar to Conan for three other known, behaviorally active volatile compounds. The results are discussed in terms of cultivar selection for large scale trap crop experiments for the wheat stem sawfly.