Correlation Analysis on Grain Characters of Japonica Hybrid Rice
Yang Liansong et al (Rice Research Institute,AAAS,Hefei 230031)
The experiment in rice grain shape traits of eight Japonica rice varieties from different countries and their 28 combinations were studied with the method of correlation and regression analysis. The results showed that there were greatly significant different between rice grain weight and height(0.840 9 ** ), and thickness(0.670 7), between rice grain length and thickness(-0.668 2), which indicated that rice grain weight weight was mainly affected by rice grain length and thickness, and environmental factors. Regression analysis of rice shape traits showed that line regression equation between rice grain weight ( y ) and length( x 1) , wideness( x 2) ,thickness( x 3 ) was y=4.175x 1+0.838x 2+19.630x 3-52.038, the multiple coefficient of correlation was R 2 =0.730. It showed that rice grain thickness was most important in affecting rice grain weight,while the variation of rice grain length,wideness, thickness had a same amount.In the principal analysis of grain characters showed that first principal component was weight factor which eigenvalue was 1.804,cumulative percent was 45.092%;second was thickness factor which eigenvalue was 1.369, cumulative percent was 34.225%; and third was wideness factor which eigenvalue was 0.700, cumulative percent was 17.50%. All of those showed that in order to get superior quality and yield, rice grain traits must be selected.
【CateGory Index】： S511.22