Stablity Of Wheat Varieties For Dryland And Its Assessment

Dry-land wheat is an important grain crop in arid and semiarid areas in China. Drought and water-shortage is the main limiting factor in the production of dry-land wheat. As a result of the complex environment conditions related to water use of dry-land wheat, the great restriction of geography, landform, precipitation and its space-time distribution, and the remarkable variation of genotype×environment interaction effects, the yield of wheat in dry-land areas is low and unstable, and the water resource on dry-lands is not utilized efficiently. With the purpose to solve this problem, many researchers consider the stability of varieties as an important subject in the field of drought-resistant and water-saving breeding, and try to discover the productivity and stability as well as the environmental adaptability of different varieties. In order to evaluate the stability of varieties to environment, the researchers in the world have conducted related investigations since 1930's, and have brought forward more than 20 evaluating methods. This paper is aimed to probe into the ideal analysis methods of stability of dry-land wheat by adopting the experimental results in Shaanxi Province in 2003~2004, and by conducting difference analysis, parameter estimation, quantitative evaluation and practical application to the models set up by four common methods, and to provide theoretical references to the selection, examination and extension of improved wheat varieties on dry-lands. The results show that: 1. Eberhart-Russell Method can predigest G×E interaction into linear reaction, and the analysis precision can be improved by applying the two statistical parameters (bi and sdi2) for stability estimation. However, the linear relation is usually difficult to be satisfied since the environment effects and genotype effects of dry-land wheat are not independent each other. Therefore, there are some obstacles for using regression coefficients and regression errors in the stability evaluation of wheat varieties on dry-lands. 2. Shukla Method is satisfied for its ideal statistical characters, concise procedures and definitude meanings; there is a set of practical method for G×E difference test and intercomparison; the stability is evaluated by directly computing the interaction difference between varieties and environment; the suitability of the models is not considered. However, it is difficult to conduct homogeneity test and difference comparison, since the sum of interaction value is 0 and the Shukla differences are not independent each other. 3. AMMI Method owns the advantages of both variance analysis and principal components analysis; the analysis of G×E interaction effects and the evaluation of stability can be achieved by using chart and mathematic formula; the estimation precision of parameters can be improved by regarding the unremarkable IPCA differences as residues. However, some real interaction ingredients may be also cancelled when deleting the error ingredients. 4. HSC Method shows the ranking orders of both stability and yield of varieties, which are approximately the same with each other; its computing formula is simple and it's easy to compare the stability of varieties by using the parameters of X, S and Xc k. However, this method only considers the environment factors and genetic factors respectively, and ignores their interactions. The HSC value can be used as a primary index to evaluate stability of varieties, but it will be better to use it along with other related parameters. 5. All the above methods can be used to evaluate the stability of varieties but none of them is perfect. In order to get a precise evaluation result, it is suggested to use multiple methods according to the breeding goals, the evaluating objects and the environment conditions. Generally, Eberhart-Russell regression analysis is the method to be used first. If the model is unavailable, or cannot make a precise judgment on the stability, it is necessary to try Shukla Method. If a high estimation precision for the varieties in different places is needed, it will be effective to use AMMI model that is relatively complex. HSC is a simple and feasible method for ranking the stability but it can not make a judgment on the difference significance among varieties and the G×E interaction effects. All in all, it is essential to comprehend the practical meanings of different methods and their differences and relations, so as to make a comprehensive evaluation on the productivity, stability and adaptability of varieties under various environments. 6. The G×E interaction is remarkable for the 12 tested varieties in the field experiment, in which the yield of T105, Xinong 797, Qinfeng 208 and Q2911 is higher than the contrast (Jinmai 47). According to the comprehensive evaluation with the four methods, Xinong 797 is highest in productivity, stability and adaptability; T105 takes second place in stability and adaptability and is prominent in productivity.