Evaluation of the Effect of Morphological Traits on Grain Yield in Introduced Faba Bean (Vicia faba L.) by Multivariate Analyses

Abstract

Determining the relationships between morphological traits is very important in plant breeding. For this purpose, path analysis and biplot analysis are among the most effective statistical methods. To study the morphological traits affecting faba bean seed yield, 12 introduced faba bean genotypes were studied under the influence of 3 concentrations of gibberellic acid based on split plot in a randomized complete block design with three replications in the research greenhouse of Shahrekord University, Shahrekord, Iran. The analysis of variance showed that there was a significant difference between genotypes in terms of all studied traits except for number of internodes, internode length, stem diameter, number of seeds per pod and root dry weight, indicating high genetic variation among the studied traits. The study of correlation coefficients showed that there was significant correlation between seed yield and all studied traits except for plant height, number of internodes, internode length, stem diameter, number of seeds per pod, and shoot dry weight. The results of multivariate regression based on stepwise method showed that among the studied traits, 3 traits including number of fertile pods, pod length, and number of seeds per plant were entered into the model, and altogether justified 86% of the variations of faba bean seed yield. The results of path analysis based on just identified model and over identified model showed that the most direct effect on faba bean seed yield was through the number of fertile pods, so that one unit change in this trait causes 0.784 units increase in faba bean seed yield. Thus, in selection for yield the number of fertile pods in the studied genotypes is of special importance. Biplot analysis as a bilinear model, identified the C2, C7 and C10 as suitable genotypes in terms of yield and yield related traits.

Keywords

• direct effect
• harvest index
• linear relationship
• principle component
• seed yield