Few plant breeding programs involve attempts to develop varieties specifically adapted to low-input cropping systems. Plant breeders generally select varieties in highly fertile, weed-free, densely seeded environments. However, alleles needed for achieving maximum yield in low-input environments often differ from those required in highinput conditions. Thus, when effective selection can be undertaken under low-input conditions, breeding programs specifically targeted at low-input environments should produce the best varieties for those environments. Experimental protocols for identifying such situations are described in this review. These involve (a) testing breeding lines under low- and high-input conditions, (b) estimating both the degree to which yields from the two types of environment are controlled by the same alleles and the relative accuracy with which superior genotypes can be identified in low- and high-input environments, and (c) predicting the direct and indirect responses to selection at each input level. On the basis of published data, it seems feasible to develop corn hybrids specifically adapted to production with lower rates of N application than are now commonly used in the United States.