selection for yield improvement using of multivariate statistical methods

in order to providing selection indices using heritability and correlation effective traits on yield and multiple regression an experiment was conducted by 265 f3 families as well as parents and f1 related to gharib × khazar population in 2009 at gonbad high education center fields. days to repining (0.97) and panicle number and flag leaf length (0.66) had maximum and minimum heritability, respectively. positive and significant correlations were detected between plant yield and flag leaf width (0.265**), plant height (0.193**), panicle number (0.734**) and biomass (0.828**). biomass, days to heading and plant height were explained about 98% of total variation of yield and inserted to model respectively. phenotypic and genotypic correlations, genetic and phenotypic direct effect in path analysis, heritability were used for construct selection vectors. according to this study, increasing of traits is not result of relative efficiency and compares parameter. selection indices were showed that yield, significant genetic correlation with yield and high heritability are three important part of selection index. fifth, sixth and fourteenth are the most important between discussed indices.