Plot Configuration in Corn Yield Trials

It is generally assumed that bordered plots are efficient for reducing error variance and the associated bias from border effects. Conflicting results in the literature regarding use of borders in corn ( Zea mays L.) grain trials and limited information on use of borders in corn silage (herbage) trials prompted studies on the most efficient plot size and the need for bordered plots in corn grain and silage yield trials. The studies encompassed five and six environments over a span of 4 yr for silage and grain, respectively. Silage trials included from 24 to 28 hybrids each year. Grain trials included three different maturity groups with 18 to 64 entries. Four‐row plots were used with the center two rows representing bordered two‐row plots and the outside two rows considered as unbordered two‐row plots. Bordered two‐row plots were 1.06 and 1.02 as efficient in reducing error variance as unbordered two‐row plots for silage and grain, respectively. The hybrid by border interactions generally were not significant, and highly significant yield and rank correlations suggest that very little border bias was occurring. Four‐row unbordered plots were only 1.30 and 1.48 as efficient as the two‐row bordered plots for silage and grain trials, respectively. The greatest gain in efficiency, while not compromising estimation of relative hybrid yields (1.89 and 2.26) came from using two‐row unbordered plots and doubling the number of replicates for silage and grain yield trials, respectively, vs. the two‐row bordered plots.