Numerous open-pollinated maize (Zea mays L.) accessions are maintained in the world's gene banks. These accessions need to be evaluated, particularly for quantitative traits, but they tend to be adapted only to specific environments. Temperate maize accessions may contain new alleles that can be used for maize improvement in the U.S. Corn Belt. Testcrossing unadapted populations to a series of testers has been proposed as an evaluation method. This study compared the ranking of temperate unadapted populations via testcrosses with eight testers consisting of broad-based populations, single crosses, and inbred lines. Also, the accessions were evaluated per se. Agronomic performance of the testers is important in their use. Testers B14xB37, B14, Oh43, and Mo17 had non-significant environmental interactions for grain yield. Inbred line tester B37 was a poor pollen producer. Broad-based testers BS13 and BS26 had the highest root lodging while testers B14xB37 and B14 had the lowest stalk lodging. Similarities were observed in ranking of some accessions by the testers for grain yield but extreme differences occurred in the ranks of other accessions. Inbred lines Oh43 and Mo17 and single crosses B14xB37 and Oh43xMo17 ranked several accessions in the opposite order for grain yield. Accession per se evaluation also showed extreme differences in ranking when compared to the testcrosses. Less extreme differences in ranking the accessions were evident for root and stalk lodging. Factor analysis was used on the rank correlation matrices in an attempt to further explain the differences in ranking among the testers and accession per se evaluation. One approach was to use factor analysis to compare testers and accessions per se for a single trait. Another approach analyzed each set of testcrosses or accessions per se for all traits and then compared factor loadings across all sets. Neither of these methods were effective in determining which tester or group of testers ranked the accessions differentially. Differences observed in ranking the populations were not extreme enough for factor analysis to detect.