Abstract:

The existence of extensive crossover genotype by environment interactions (GEI) clearly suggests that the efforts are necessary to identify the discriminating and representative test locations over multilocation data for specific trait selection and new cultivar evaluation.Genotype+Genotype×Environment interaction (GGE) biplot analysis is an efficient technique in graphically displaying and visualizing the interrelationship between the actual locations and the virtual ideal location.In the “discrimination and representativeness view” of GGE biplot,the vector length of a test location stands for its discriminating ability in differentiating the genetic differences among candidate genotypes.The angle between a test location vector and the average environment coordinate (AEC) is defined as its representativeness to the average performance in the target planting region.Further,in the latest version of heritabilityadjusted GGE biplot,the test location vector length and the cosine value of the angle to the AEC are approximated by the genetic correlation(r〖KF（〗H〖KF）〗) and the square root heritability (〖KF（〗H〖KF）〗) respectively.Meanwhile,the product of the genetic correlation(r〖KF（〗H〖KF）〗) and the square root heritability(〖KF（〗H〖KF）〗) is defined as the ideal index of test location.However,the Euclidean distance to the “ideal” location is more precise in evaluating the ideal degree of test locations.In the present study,the heritabilityadjusted GGE biplot analysis was employed to analyze twentyseven independent oneyear datasets of national cotton variety trials from 2000 to 2010 in the Yangtze River Valley to evaluate trial locations based on cotton fiber length selection.The trial locations were comprised of Janyang and Shehong in Sichuan province; Changde and Yueyang in Hunan province; Huanggang,Jingzhou,Wuhan and Xiangfan in Hubei province; Nanyang in Henan province; Jiujiang in Jiangxi province; Anqin in Anhui province; Nanjing,Nantong and Yancheng in Jiangsu province and Cixi in Zhejiang province.Judging by the desirability indices and superiority parameter,we could conclude that Jingzhou,Jiujiang and Changde test locations were most efficient in broad adaptive breeding for cotton fiber length selection and also ideal to act as regional trial sites for the enhancement of cultivar selection efficacy.However,the coastal region test location in Jiangsu and Zhejiang province (Nantong,Yancheng and Cixi) and the test locations located in Sichuan basin(Jianyang and Shehong) were not suitable in cotton fiber length selection targeting at the whole cotton planting region in the Yangtze River Valley.The other trial locations including Yueyang,Huanggang,Wuhan,Xiangfan,Nanyang,Anqin and Nanjing were of moderate performance in the regional trial scheme for cotton fiber length selection.The results above were positive as a case study in displaying the GGE biplot application efficiency in cotton regional trial environment evaluation and also in providing the theory background for the decisionmaking of national cotton scheme configuration in the Yangtze River Valley.〖