Abstract: The AMMI model was used to select locations for rice cultivar trials with parameter D j which was got from the environmental value of significant IPCA. The results indicated that the location with low D j value has little genotype×environment interactions. When four locations with the lowest D j value was deleted from 15 locations of Zhejiang Province early rice regional trial, the square sum of genotype×environment interactions just lost 10.87%.

Key words: regional trial, site selection, selection methods, statistical analysis, rice

摘要： 提出利用AMMI 模型中达显著水平的主成分轴上环境分量作为水稻区试点对品种判别能力的估算依据, 并设计了估算参数D j 。试验表明, D j 值小的试点, 其与基因型互作效应弱, 包含较少的互作信息。以浙江省早籼稻区域试验15 个区试点为例, 剔除D j 值最小的4 个试点后, 交互作用值仅损失10. 12%。

关键词: 区域试验, 试点选择, 选择方法, 统计分析, 水稻