Identification of sources of favorable alleles to improve existing hybrids is one of the most important problems facing a maize (Zea mays L.) breeder. Previous work has demonstrated the effectiveness of a procedure developed by Dudley for identifying populations containing favorable alleles not present in an elite hybrid. However, previously reported work involved at most two elite hybrids. The objective of this study was to evaluate the potential of 20 improved populations to improve the three hybrids made from three inbreds in commercial use. Each of the populations was crossed to LH195, LH212, and LH216. The population × inbred crosses, the three hybrids among the inbreds, and the three inbreds were evaluated in seven U.S. midwestern environments in 1993 and four in 1994. Traits measured were grain yield, grain moisture, plant height, ear height, and concentration of protein, oil, and starch in the grain. For grain yield, 15 of the 20 populations had significant estimates of dominant favorable alleles not present in the highest yielding target hybrid (LH195 × LH212). None of the populations showed potential for reducing ear height. However, seven populations had more favorable recessive alleles than unfavorable dominants for plant height when LH195 × LH212 was the target hybrid. None of the populations tested appeared to have potential for increasing starch concentration in any of the target hybrids. Eight populations showed potential for increasing protein concentration in all three target hybrids. Assumptions required to identify parents were not met for grain moisture, oil concentration, and stalk and root lodging.