AMMI模型在小麦区域试验产量组成性状分析中的应用

为了解AMM I模型在小麦区域试验产量组成性状分析中的应用效果,利用AMM I模型对2002年四川省小麦区域试验资料进行了分析。结果表明,在分析基因型与环境互作时,AMM I模型优于多元回归和相关分析。AMM I模型中的极显著和显著项能分别解释产量、生育期、有效穗、株高、穗粒数和千粒重等性状92.6%、76.6%、80.1%、76.4%、68.4%和72.9%的交互作用平方和。AMM I模型双标图和稳定性参数Di可以分析产量性状在不同地点下的稳定性。稳定性Di值分析表明,品系和环境的变化均能对穗粒数的稳定性产生较大影响,而品系的变化对生育期影响较小,地点对株高的影响也较小;千粒重不稳定的品系更易检测出产量的不稳定性;有效穗差异大的地点,其对产量的鉴别力相对较弱;而穗粒数变化大的地点,则对产量的鉴别力较强。

Application of AMMI Model on the Yield Trait Analysis of Wheat Regional Trials

To detect the performance of yield traits in wheat regional trials,the data collected from Sichuan wheat regional trials in 2002 were analyzed by AMMI model.The results indicated that the AMMI model was superior in explaining genotype by environment interactions to joint linear regression and correlation analysis for yield traits.AMMI model explained most of the genotype by environment interactions,such as 92.6% for yield,76.6% for days to maturity,80.1% for spikes per hectare,68.4% for kernels per spike and 72.9% for 1000 grain weight.Based on the biplots and the D_i value,which showed the scores of genotype and environment,it was found that genotype and environment had the strong effects on kernels per spike,while genotype and environment had relatively lower effects on days to maturity and plant height,respectively.The genotypes with larger variances for 1000 grain weight had relatively lower stability.The environments with higher variance for spikes per hectare had the low ability to discriminate the yield performance,whereas the environments with the higher variance for kernels per spike had the higher ability to discriminate the yield performance.