作物品种区域试验试点区辨力评价方法的比较(英文)

试点区辨力是作物品种区域试验试点选择的重要依据,可以用不同的方法进行评价。本研究分别对西南玉米区域试验2003和2004两年的产量数据进行联合方差分析,然后用变异系数(CV)、遗传变异系数(GCV)、环境区分指数(DI)、B型生态价(WE)、回归系数(RC)、基于AM MI分析结果的试点互作主成分点到原点的距离(DA)、AMMI稳定性值(ASV)和基于GGE双标图分析的试点到原点的距离(DG)对西南玉米区域试验试点的区辨力进行评价,进而用秩次相关分析和主成分分析揭示试点区辨力评价方法间的内在关系。结果表明,两年试验品种间、试点间均存在极显著差异,且品种和试点间互作也极显著;8个统计数也许可以用来度量试点的区辨力,但是不同统计数对同一年度内试点的区辨力评价结果不完全一致;统计数CV、GCV、WE、DA和DG间都存在极显著的秩次相关,但是统计数DG也许更有利于评价试点的区辨力。

Comparison of Methods to Assess the Discriminating Ability of Testing Locations in Crop Cultivar Regional Trials

The discriminating ability of testing locations is an important issue for the selection of testing locations in crop cultivar regional trials.Several procedures are available to assess the discriminating ability of testing locations,each of them allowing for different interpretations.The objectives of the present study were to assess the discriminating ability of testing locations in National Maize Cultivar Regional Trials in Southwestern China using the eight statistics including coefficient of variability（CV）, coefficient of genetic variability（GCV）,discriminating index（DI）,type-B ecovalence （WE）,regression coefficient（RC）,distance（DA） of interaction principal component point of a location with the origin in space and stability value（ASV） based on the results of AMMI analysis,and distance（DG） between a location marker and the biplot origin based on the results of GGE biplot analysis,and to investigate their interrelationships. Two yield datasets were obtained from the 2003 and 2004 National Maize Cultivar Regional Trials in Southwestern China.A combined analysis of variance,assessment of the discriminating ability of testing locations and rank correlations among the discriminating ability statistics were determined.Effects of location,cultivar and cultivar by location interaction were highly significant（P0.01）.Different statistics could be used to determine the discriminating ability of the studied locations,but the discriminating ability of testing locations evaluated by different methods was dissimilar.The statistics CV,GCV,WE,DA and DG were positively correlated with each other.The statistic DG would be useful to assess the discriminating ability of test locations.