玉米品种多环境测试数据的最优相对转化方法

相对评价是农作物品种多环境测试的主要评价方式。该文设计了5种测试数据的相对转化方法，同时，提出标准差比值这一指标，用于判别相对转化方法的优劣，并根据8 a中国玉米品种区试数据进行了实证分析。结果表明：该文提出的5种试验数据相对转化方法，均可适用于不同观测性状，相对值信息量丰富、易于理解，并有助于进行多指标加权与作图分析。标准差比值可直接反映经不同方法转化后，相对值与品种遗传特征的接近程度，可用于多环境测试数据相对转化方法优劣的判别。对于大部分性状，最优的数据相对转化方法依次为：以组均值为参照系的3S和2S转化法>标准位次法>以组均值为参照系的类NDVI法>以对照值为参照系的类NDVI法>以对照值为参照系的3S和2S转化法>原始观测值。其中，以组均值为参照系的3S和2S转化值与遗传特征最为接近，可作为品种多环境测试数据的主要相对转化方法。

Optimal method of transforming observables into relative values for multi-environment trials in maize

Relative evaluation is commonly used in multi-environment trials of crop varieties. Five methods of transforming observables into relative values were proposed, and the corresponding relative values were measured by the new index standard deviation ratio to find the optimal one. Then, 5 relative values and standard deviation ratios were calculated using 8-year maize variety regional trial data of China. The results show that: the 5 relative values transformation methods can be applied to various traits, and relative values were informative and understandable, and moreover making weighted multi-traits and graphic analysis much easier. The new index standard deviation ratio can directly response the closeness of each relative value to the genetic characteristic, and can be used to measure these relative values transformation methods. For most traits, the optimal methods of transforming observables into relative values were: 3S and 2S value compared to the sample mean value (short for A) > standard order value (C) > the imitated NDVI value compared to the sample mean value (D) > the imitated NDVI value compared to the check variety (E) > 3S and 2S value compared to the check variety (B) > observables (F). 3S and 2S value compared to the sample mean value, which is closest to the genetic characteristic, can be used as the main method of transforming observables into relative values for multi-environment trials in maize.