Resource Allocation in Experiments Measuring Soybean Tolerance to Soybean Cyst Nematode

Abstract

Genotypic variability in soybean, Glycine max (L.) Merr., for tolerance to soybean cyst nematode (SCN), Heterodera glycines Ichinohe, has recently been reported. The measurement of soybean tolerance to SCN requires twice the land area and is more expensive compared with yield testing. The objectives of this research were to estimate the coefficients of variation (CV) and error variances (σ2) for different plot designs from two experiments measuring tolerance to SCN and to determine efficient resource allocation for future experiments. The CV for tolerance index [(yield in untreated plot ÷ yield in nematicide treated plot) ✕ 100] ranged from 0.15 to 0.48 for various hill‐plot designs (one to eight hills/plot) and ranged from 0.18 to 0.33 for various row‐plot designs (1.78 to 7.10 2 harvested). The error variances ranged from 108 to 1131 for the hill‐plot designs and ranged from 106 to 579 for the row‐plot designs. Unbordered hill and row plots had similar CV compared with bordered plots. Assuming no genotype ✕ environment interaction (σ2GE), six and two replications would be required to detect a 50% difference in the experimental mean between two geneotypes using one and eight hills/ plot without borders, respectively. Unbordered row plots of 1.8 m and 7.32 m length (row width = 0.97 m) would require four and two replications, respectively, to detect the same level of difference. The σ2GE expressed as a fraction of σ2 was below 0.10 for all plot designs except one. Assuming σ2GE = 0.1 σ2, a 30% difference in the experimental mean between two genotypes can be detected with three replications in each of two environments with the 7.32‐m row plot design. These results are discussed in regard to efficient allocation of resources for measuring tolerance to SCN.