论文部分内容阅读
用方差分析法研究了1981～1985年吉林省大豆联合区域试验中95个点次的产量试验资料,计算了各点次的机误方差和机误变异系数及基因型×环境互作方差组分等参数。计算了检验不同水平的品种间差异需要的总重复数,研究了各种互作条件下同一环境内的区组数和环境数对试验准确性的影响及两者之间的相互关系,讨论了区域试验的合理设计问题。各熟期组平均机误均方为1.325,平均变异范围为0.606～1.957。各熟期组的平均机误变异系数为11.93％。所研究的试点中约有57.7％的点次机误变异系数小于10％。平均基因型×环境互作方差组分约是机误方差的0.112倍,其大小变化于0.000～1.553之间。假定无基因型×环境互作,当α=0.05,β=0.50时,如试验的机误变异系数为10％,要检验出两个品种间10,20,30％的差异分别需要10,3和2欢重复。当有基因型×环境互作时,增加试验地点内的重复数或试验环境数均可有效地降低试误差。在试验点内的重复数一定时,试验需要的环境数随互作的增大而增多。
The variance test was used to study the 95-point yield test data in the joint area experiment of soybean in Jilin Province from 1981 to 1985. The machine error variance, error variance coefficient and genotype × environment interaction variances Other parameters. The total number of repetitions needed to test the differences among varieties at different levels was calculated. The effects of number of zones and environments on the test accuracy and the relationship between the two groups under different interaction conditions were studied. Rational design of regional trials. The mean machine-error-averaged square of each maturity group was 1.325, and the average range of variation was 0.606 ~ 1.957. The average error coefficient of each machine group was 11.93%. Approximately 57.7% of the pilots studied were less than 10% chance of error. The average genotype × environment interaction variance component is about 0.112 times the machine error variance, and its size changes between 0.000 ~ 1.553. Assuming that there is no genotype × environment interaction, when α = 0.05 and β = 0.50, if the machine error coefficient of variation is 10%, it is necessary to check that 10,20,30% difference between the two varieties requires 10,3 And 2 Huan complex. When there is genotype × environment interaction, increasing the number of replicates in the test site or the number of test environments can effectively reduce the test error. When the number of replicates in the test point is constant, the number of environments required for the test increases with the increase of interaction.