Culling soybean crosses in early generations based on yield tests of bulks in microplots

Because of genetic heterogeneity and a limited seed supply, selection for yield in soybeans is not initiated until the F₅-F₇ generations. This study was conducted to (i) compare effectiveness of standard 6 m plot size with 3 and 1.5 m plots for yield testing of early generation bulks and (ii) determine if the superior crosses could be identified on the basis of tests of this type in early generations. This study was conducted at three locations in Tennessee. An incomplete block design was used with plot size as the main- and cross or line as the sub-plot effect. Single row plots 6, 3, and 1.5m with a .91 m row spacing and seeding rate of 32 seed m^-1 were used. The F₂ and F₄ cross bulks were evaluated and advanced without selection in 1987, the F₃ and F₅ bulks in 1988. Parentage of these crosses was all group V maturity with the exception of two which were of group IV. Sixteen lines were also utilized as benchmarks at each location. There were two, four, and eight replications of the 6 m, 3m and 1.5m plots, respectively. By increasing the number of replications as plot size decreased, the harvested linear footage of row was kept constant. The 6 m plots were harvested in four 1 m subplots, the 3 m plots in two 1 m subplots, and the '1.5 m plots in one 1 m subplot. The patterns of the ranks of the breeding lines were similar in all plot sizes, suggesting no difference in relative yield performance when tested in different size micro-plots. This implies that all three plot sizes provided equal opportunity to discriminate the highest yielding from lowest yielding lines. The CV values generally decreased as plot size in creased. Based on the ranks of the five crosses from all data combined over generations, plot size, and locations, the top 3 crosses could have been identified in the F₂ generation and two of the top 3 crosses would have been identified from the F₃ and F₅ tests of bulks. It appears that early generation yield testing of individual cross bulks is a feasible way to eliminate crosses with the least potential.