Comparative testcross and per se performance of heterogeneous maize lines with their homozygous derivatives

The objectives of this study were to compare the testcrosses of S[subscript]2 lines from BS13(S2)C1 with two S[subscript]2 lines from Lancaster Composite with their S[subscript]8 x S[subscript]8 counterparts, to compare S[subscript]2 per se performance with their S[subscript]8 derivatives, to compare S[subscript]2 per se performance with their S[subscript]8 testcross performance, and to evaluate the performance of S[subscript]2 x S[subscript]2 hybrids. Twenty-seven S[subscript]2 lines and their S[subscript]8 derivatives from BS13(S2)C1 were crossed to two S[subscript]2 lines and their S[subscript]8 derivatives from Lancaster Composite. The testcrosses and inbreds, along with checks, were evaluated. The traits measured were grain yield, grain moisture, stand, root and stalk lodging, dropped ears, and ear height. Only ear height differences were significant when comparing the S[subscript]2 testcrosses with that of the S[subscript]8 testcrosses. Trait-for-trait correlations between the S[subscript]2 and S[subscript]8 testcrosses for each tester were all positive, but generally nonsignificant. Correlations for yield between the S[subscript]2 and S[subscript]8 testcrosses for each tester were 0.13 and 0.15. The correlations of traits between the two generations of BS13(S2)C1 lines per se were all positive and were significant for grain yield, root lodging, and ear height. The trait correlations between the S[subscript]2 lines and their S[subscript]8 testcrosses were nonsignificant and not large enough to be of predictive value. The mean yields for the S[subscript]2 and S[subscript]8 testcrosses were very similar. The mean grain yield for the BS13(S2)C1 S[subscript]2 and S[subscript]8 lines was significantly different, as were moisture and ear height differences. Although the trait correlations were largely positive, the S[subscript]2 x S[subscript]2 testcrosses were not satisfactory predictors of their S[subscript]8 x S[subscript]8 counterparts. The S[subscript]2 line per se performance was a good indicator of the S[subscript]8 derivative\u27s per se performance. While the trait correlations between the S[subscript]2 lines and their S[subscript]8 testcross performance were generally larger than between the two testcross generations, they were not of a predictive level. The performance of the S[subscript]2 x S[subscript]2 hybrids was very comparable to the S[subscript]8 x S[subscript]8 hybrids. Several significant rank changes occurred between the two generations of testcrosses, indicating that one or both lines underwent significant genetic changes during inbreeding