Evaluation of Recurrent Selection for Specific Combining Ability in Two Open‐pollinated Maize Cultivars1

Abstract

We evaluated recurrent selection for specific combining ability in two open‐pollinated maize (Zea mays L.) cultivars, ‘Kolkmeier’ and ‘Lancaster’. A common, singlecross tester, WF9 ✕ Hy, was used in the first cycle, but inbred Hy was used in the four subsequent cycles. The principal basis of selection was grain yield of half‐sib progenies. Progress for population improvement was evaluated in five environments by measuring yields in the Kolkmeier and Lancaster populations per se, the Hy ✕ Kolkmeier Cn (n = 0, 1,2 …, 5) and Hy ✕ Lancaster Cn test‐crosses, the Kolkmeier Cn ✕ Lancaster C. population crosses, and the Cn and C5 populations of both cuitivars in crosses with Alph C0, Alph C5, BI4A, B73, and Mo17. Kolkmeier per se did not show a significant change in yield; Lancaster per se, however, had a significant negative linear trend, −0.95 ± 0.45 q/ha/cycle. The rates of yield gain (q/ha) for Hy ✕ Kolkmeier Cn and Hy ✕ Lancaster Cn were 2.75 ± 0.45 and 2.23 ± 0.45, respectively. The rate of yield gain (q/ha) for Kolkmeier Cn ✕ Lancaster Cn was linear and significant, 2.45 ± 0.45. Similar yield gains were expressed with the C0 and C5 populations in testcrosses with other unrelated testers, except B73 for which there were no differences between testcrosses of the C0 and C5 populations. The recurrent selection program has been successful in increasing frequencies of genes with favorable yield effects in crosses, and overdominance was relatively unimportant in the change in yield potential of the two cultivars. Recurrent selection by using an inbred line tester seems to be an efficient method to improve breeding populations. The obvious tester would be a line that is already widely used in the seed industry. As the program progresses, this line can be replaced by a better line with no deleterious results relative to the population improvement achieved by the earlier tester.