Study On Source-sink Characteristics And Evaluation System Of High-yielding Rice Varieties In Jilin Province

Thirteen physiological indices, such as leaf area index (LAI) and number of grain per panicle (NPG) were measured for 76 high-yielding rice varieties in Jilin Province. Subsequently, principal component analysis, cluster analysis, discriminant analysis and other statistical methods were conducted. Source-sink germplasm resource database of high-yielding rice varieties in Jilin Province was established. And also, a scientific and standardized evaluation system of the source-sink was established. The results were as follows:1. Using dynamic clustering method, we classified source-sink traits and source-sink types into 5 main components and 4 types (source-sink interaction type, increase production by increase sink type, improving production by increase source type, and saturated source-sink type) for 76 high-yielding rice varieties in Jilin Province. The characteristic differences were also observed among the 4 source-sink types. The results indicated that 76 rice varieties were mainly distributed in 3 types including source-sink interaction type, improving production by increase sink type, and increase production by increase source type, and the number of varieties were distributed evenly in each type. The results of multiple comparisons of key indicators showed that significant differences existed between different source-sink types. Source-sink saturated type only includes 3 varieties, Jijing 88, Yutong 221, and Jijing 501, however, it has the highest yield potential because its LAI, storage capacity, NGP, and harvest index were significantly higher than those of other types.2. Each variable of rice source-sink characteristics was selected by using principal component analysis, and the variables closed related with source-sink characters were selected. The evaluation system for the source-sink germplasm resources of high-yielding rice varieties in Jilin Province was firstly established by discriminant analysis. The discriminant analysis results showed that the posterior probability range of source-sink interaction type was 0.4619～1.0000; the range of increase production by increase sink type was 0.4685～1.0000; the range of increase production by increase source type was 0.7544～1.0000; the range of saturated source-sink type was 0.9999～1.0000. The simulation results of discriminant function analysis on the 4 source-sink type showed that only Tongyuan 6 had the differences between the simulation result and clustering analysis among 76 varieties. Simulation results showed that Tongyuan 6 belonged to source-sink interaction type, however, the clustering analysis showed that it belonged to yield increase library type. Therefore, the discriminant accuracy of this discriminant function was up to 98.9%.In addition, the results of discriminant analysis on 17 varieties in 2010 were completely consistent with the results of in 2009, so the discriminant accuracy was 100%. It means this discriminant can effectively divide the source-sink types of high-yielding rice varieties in Jilin Province.