Using AMMI Model To Analyzing Yield Characters Stability Of Some Super Rice Varieties

The growth of super rice was effected by lots of environment factors such as themoisture, illumination, temperature, topography, and Cultivation factors such as fertilizer,planting density,seedling-age and so on. The variation of rice genotype-environmentinteraction（G×E） was significant and phenotype was great influenced by environmentfactors.Based on the data of five super rice varieties collected from six differentecological conditions regions in2006^-2011,using variation analysis, correlationanalysis, path analysis and AMMI model to analyze the relationship ofgenotype-environment interaction（G×E）. The variation and stability of yield charactersand yield of different rice varieties and the environment discrimination power in differentyears and regions were discussed and the main results were as follows:1. Variation of yield characters and yield:The CV and variation range of yield characters were different in different varieties.The CV of ganxin203was largest for yield and the lowest for the1000-grain weight. TheCV of total grains per panicle and filled grains per panicle of ganxin688was larger,effective panicles was the lowest.The CV of luliangyou996was the largest for effectivepanicles and the lowest for the1000-grain weight. The CV of total grains per panicle ofY-liangyou1and liangyou293were the largest, the1000-grain weight were the lowest.The largest sum of CV of yield characters in six years and regions was Y-liangyou1（62.7%,76.66%） variety and the lowest was ganxin203（33.56%,31.03%）. The largestsum of CV was total grains per panicle and yield and the lowest was1000-grain weightin the five varieties. The result of Duncan’s multiple range test showed that the variationof years and regions have reached significant difference or extremely significantlevel.The variation range of yields in years was8698.56（2010）⁹527.12（2008）kg·hm^-2and in regions was8613.71（guilin D6）⁹324.76（liuyangD5）kg·hm^-2. That indicatedthe fluctuation levels of yield was larger.2. Correlation analysis, path analysis of yield characters and yield:The correlation of super-rice between yield characters and yield have reachedextremely significant level,but correlation coefficient were different in varieties.Pathcoefficient analysis also showed that the yield characters promotion which could directlyenhance yield or indirectly promoting were also different.3.Stability of yield characters and yield, discrimination of environment analysis:Genes, environment, and genotype-environment interaction（G×E） effects on yieldcharacters and yield were significant or extremely significant level.Complex variance analysis indicated that genes are the main reason for the variation of effectivepanicle,total grains per panicle,filled grains per panicle and the1000-grain weight;genotype-environment interaction（G×E） was the main reason for the yield,seed settingrate.Using AMMI model to analyze the stability of yield characters indicated that thestability of Y-liangyou1and liangyou293were poor; ganxin688was middle; ganxin203and luliangyou996were the best varieties. It also showed that,with the variation rangesmall,the yield stabilities of varieties which changed with the stabilities of yieldcomponents were good,but with the variation range large, the yield stabilities werepoor.The stability of the super rice between yield and yield components has largerrelevance and the stability of yield components was the premise and basis of the yieldstability.The environment discrimination power of liuyang region was the highest, lowestin guilin region.