Studying Effect of Phenological and Morphological Traits on Seed Yield of Bitter Vetch by Path and Biplot Analyses

Path and biplot analysis based on linear and bilinear models, respectively, are of the effective statistical methods for expressing relationships between traits in plant. In order to study of effective traits on seed yield of bitter vetch, 20 ecotypes were evaluated in a randomized block design with three replications during 2017-2018 growing season in the Research Field of Faculty of Agriculture at Vali-e-Asr University of Rafsanjan. The analysis of variance revealed significant differences among the ecotypes for 16 studied traits except plant and root height in maturity stage and one seed pod length suggesting high variation among the ecotypes for the majority of the studied traits. Correlation analysis showed that there were a positive and significant correlation between seed yield and the most of traits except phenological traits including days to 50% flowering and days to maturing start and days to 50% maturing and two morphological traits including root height in maturity stage and two seed pod length. Stepwise regression analysis indicated that thousand kernel weight, number of seed per plant and number of lateral branch entered into model and explained about 65.7 of the total variation. The results obtained from path analysis demonstrated that thousand kernel weight had the most direct effects on seed yield (an increase of 0.657 g in seed yield per 1 g increase in thousand kernel weight). Biplot of first two components divided the successful ecotypes into two groups and identified the most effective traits association with the success of each group.