用AMMI模型分析作物区域试验中的地点鉴别力  

Analyzing Site Discrimination in Crop Regional Yield Trials by AMMI Model

作　　者：李本贵[1] 阎俊[2] 何中虎[2] 李仲来[1] 

机构地区：[1]北京师范大学数学系,北京100875 [2]中国农业科学院作物育种栽培研究所/国家小麦改良中心

出　　处：《作物学报》2004年第6期593-596,共4页Acta Agronomica Sinica

基　　金：国家自然科学基金项目 ( 3 993 0 110和 3 0 0 60 0 43 )资助

摘　　要：为了比较不同试验点在区域试验中鉴别力的大小 ,提出了AMMI模型中地点得分向量长度的衡量方法 ,并与传统联合线性回归模型中的斜率进行了比较。以 1995— 1996和 1997— 2 0 0 0年度黄淮南片春水组小麦区域试验 4年的产量数据为例进行了地点鉴别力分析。结果表明 ,AMMI模型比联合线性回归模型能更好地解释基因型与环境互作效应 ;多年间地点鉴别力参数的可重复性分析表明 ,AMMI模型中地点的得分向量长度比联合线性回归模型中斜率的可重复性程度更高 ,因此AMMI模型中地点得分向量长度可作为地点鉴别力的一个指标 ,也可以将其作为取舍试验地点的标准之一。To compare the discrimination of different sites in regional trials, the length of score vector in AMMI model was used as indicator for site discrimination and comparison was made between AMMI model and joint linear regression model. The yield data from South Yellow and Huai River Valley's wheat yield trials from 1995—1996 and 1997—2000 were used to illustrate the site discrimination. The results showed that the length of score vector in AMMI model was better to explain genotype by environment interaction than by joint linear regression model. The site discrimination analysis between years showed the length of score vector in AMMI model had higher repeatability than the slope of joint linear regression model. The length of score vector in AMMI model can be used as an index to measure the site discrimination and a criterion for judging testing location.

关 键 词：AMMI模型 作物区域试验 地点鉴别力 地点得分向量长度 基因型环境互作效应 联合线性回归模型 

分 类 号：O29[理学—应用数学]