Correlation and Path Analysis of Some Yield Determine Traits in Rice Genotypes

Rice is one of the most important food crops, thus identification of definitive traits on the yield can be used in breeding programs. In order to determine the best suitable morphological markers for selection of superior genotypes, seeds of 15 rice genotypes including 5 varieties (namely: Binam, Dasht, Neda, Mashhad Domsiah and IR62871-175-1-10) and their 10 F1 hybrids of corresponding variates from a 5 × 5 one-way diallel cross were evaluated using a randomized complete block design with 3 replicates. This research was performed at Gonbad Kavous University Research Farm in 2016. Twelve traits including plant height, length and width of flag leaf, panicle length, number of unfilled and filled grain, days to 50% flowering, 1000 grain weight, number of fertile tillers, panicle weight, number of spikelets per spike and yield were evaluated. Results of correlation analysis showed that the grain yield had significant and positive correlation with flag leaf width, number of filled grain and panicle weight, while showed a negative significant correlation with flag leaf length. Stepwise regression analysis (correlation coefficient 69%) indicated that panicle weight and flag leaf width are the most important components of grain yield. Also, path analysis revealed that the number of filled grains per panicle (I=1.27) showed the highest positive direct effect on yield. Despite this, the panicle weight (I= -0.81) showed the highest negative direct effect, however, this trait had the highest indirect effect through number of filled grain. Genotypes were clustered in 3 groups based on UPGMA method. Also, cophenetic correlation coefficient (0.86) indicated a high correlation between similarity matrix and the matrix resulted from the cluster analysis. Overall, the characteristics of the studied traits such as flag leaf width, number of filled grain and panicle weight can be important criteria for selection of high yield genotypes.