Study of agronomic traits in a number of promising rice lines by multivariate statistical methods

Abstract

The objectives of the experiment were to study the relationships between the agronomic traits and yield components and identify effective traits for the improvement of grain yield in seven promising lines of rice plus two native cultivars (Shiroodi and 843) as control. The experiment was conducted as a randomized complete block design (RCBD) with four replications at Rice Research Station of Tonekabon, Iran in 2011. In this study, 17 traits including plant height, total tiller number, fertile and sterile tiller number, length, width and area of flag leaf, panicle length, total grain, number filled and empty grain, 100-grain weight, length, width and the ratio of length to width of grain and grain yield. The result of correlation showed that the grain yield has a positive and significant correlation with plant height and negative correlation with empty grain. Stepwise regression showed that the effective characters related to grain yield were total number of grain that about 49.48 percent of the regression model variation was allocated to itself. The results of path analysis indicated that number of empty grain and plant height were the most important components of grain yield and they had the most directed effects on grain yield. For synchronizing of these effects on rice lines, cluster analysis was done and results of this method showed that there was 3 classes between these genotypes.