Evaluation of Agronomic Traits Associated with Grain Yield in Rice (Oryza sativa) Using Regression and Path Analysis

Facilitate the selection procedure to improve performance of new genotypes of rice is essential. In present study in order to evaluate relationship between grain yields with other agronomic traits, a global rice collection including 100 accessions was seeded in an incomplete block design (triple lattice 10×10). Afterward, grain yield and 14 morphological traits were evaluated. Correlation coefficient displayed a positive and significant correlation between grain yield and number of filled grain per panicle, 1000-grains weight and grain width. To select affective traits in grain yield step wise regression was done and 1000-grains weight, number of filled grain per panicle, number of fertile tillers, grain length and flag leaf length were entered to regression model. Path analysis was used to understand the direct and indirect effect of these traits with grain yield. Results showed 1000-grains weight and number of filled grain per panicle had the largest positive direct effect and lowest negative indirect effect on grain yield. Therefore, the results can be pointed out the importance of these traits in improvement grain yield. However, 1000-grains weight and number of filled grain per panicle can be appropriate criteria for indirect selection for grain yield in rice breeding projects.