Corn (Zea mays L.) and soybean (Glycine max Merrill) were planted in monoculture and strip-intercropping plots (18.3 times 15.5 m each) under conventional tillage and no-tillage to create differences in habitat structure. The effects of these agronomic practices on carabid beetle species composition were measured by pitfall-trap collections during three years. Although neither the number of individuals or number of species were significantly affected by treatments, analyses of variance indicated that four of the six dominant species (75.7% of the total number of carabids captured in three years) responded to tillage treatments. Harpalus pensylvanicus DeGeer and Cyclotrachelus sodalis (LeConte) were more abundant in no-tillage plots, whereas Pterostichus chalcites Say and Scarites substriatus Haldeman were more common in conventional tillage plots. Pterostichus chalcites was also significantly more abundant in soybean plots than in intercropping or corn plots. H. pensylvanicus, a primarily herbivorous carabid, was the most common species collected. However, P. chalcites, a predaceous carabid, was the dominant species in soybean/conventional tillage plots. The results suggest that characteristics of individual species should be carefully considered for studies of communities such as carabid beetles, which have few strongly dominant species and a wide range of resource utilization.