小麥品種育成 方法에 關한 硏究 第1報 實驗誤差, 品種, 環境 및 品種×環境의 分散 成分 比較와 地域 및 反復數 推定

Effect of selection on plant improvement Ⅰ. A Comparison of Variance Components of Error, Genotype × Environment, Genotype and Environments, Estimation of Optimal Number of Location and Replication

Eleven wheat lines were grown at 6 locations for 2 years to analyze the variance components of error, genotype, genotype-location, genotype-year and genotype-location-year, and to estimate their implication for the optimum number of replications and locations, explaining the interpretation of appropriate parameters related to biometrics. In analyses of variance for grain yield and its components, there were non-significant in location and year, but highly singificant in genotypes. As far as G-E interactions are concerned, there showed significant variations, especially indicating the higher significance for 2nd order interaction rather than 1st order. Mean square expectations such as σ²gsy, σ²gy, σ²gs and σ² were computed by subtraction of pooled mean squares, and variance of mean for a genotype(σ²/p) and environmental variance(σ²/g) were obtained from these mean square expectations. Among the variance components for G E interaction, there appeared highest variation in σ²gsy, which would be desirable for plant improvement. Thus, σ²e and σ²gsy, were more important than that of σ²gs and σ²gy because of their implication for estimation of optimum number of replications and locations. It was more suitable to be grown at 6 locations with 2 replications for this experiment, thus indicating that increase of locations would be much more important than that of replications. However, estimate of number of years should be conducted upon basis of gain from selection(Gs) relating to selection intensty and heritability as well as long and short term selections.