To study variations and correlations between different quantitative traits in 72 double haploid (DH) barley lines, seeds of parents and lines were increased in greenhouse in fall 2000 to obtain enough seeds for field studies. Parents were Steptoe cv. (high yielding and drought sensitive) and Morex cv. (drought tolerant). The seeds obtained from green house were sown in field conditions in fall 2001 and 2002. The experiment, designed in RCBD with three replications was carried on research farm of the University of Tehran at Karaj, Iran. Each doubled haploid line was sown in a 2 m. long row (40 seeds per row) and scored for the following characteristics: seed yield (gr), date to flowering, spike initiation and maturity (days), plant height (cm), seeds per spike, tillers per plant, 1000 seed weight (gr), spike length (cm) and protein content (%). Covariance analysis was performed for yield, seeds per spike, tiller number per plant and spike length. Significant differences were found for all this traits. The adjusted data were analysed by F-test. The results showed that the above-mentioned traits were statistically significant. There was a positive correlation between seed yield and plant height (r=0.325). The coefficient of determination (R2) for these four traits of plant height (cm), seeds per spike, tillers per plant, 1000 seed weight was 20, 35.2, 44.7 and 44.3%, respectively. The highest variations were observed for seed yield and tillers per plant, whereas the lowest ones recorded for date to maturity and flowering. Multiple regression analysis showed that 71% of seed yield variation (Y) was expressed through tillers per plant (X7), plant height (X4), number of seeds per spike (X6) and 1000 seed weight (X8). The best fitted equation for seed yield estimation, based on stepwise regression analysis, was found as:
Y=-428.8 + 17.22X7+5.82X8+3.52X6+1.7X4
Path coefficient analysis showed that the indirect effects of number of tillers per plant, 1000 seed weight and number of seeds per spike on seed yield were both positive and high. Parental lines were shown to be located a part by Euclodean distances. Using Hotelling’s T test for comparing trait means, T2 and F values were found to be 178.5 and 16.8 respectively, which revealed, the two were highly significant for recorded traits. The results have also revealed that, traits like number of seeds per spike and 1000 seed weight could be used as selection indexes for improving grain and forage yield in the studied barley cultivars.
)