Evaluation of safflower germplasm of National Plant Gene Bank of Iran for tolerance to safflower fly ( Acanthiophilus helianthi)

Tolerance of 22 selected safflower genotypes of National Plant Gene Bank of Iran was studied against safflower fly (Acanthiophilus helianthi) in a randomized complete block design with four replications. The relationship between three fly tolerance characters (percentage of infected heads, number of damaged seeds in infected heads and total number of damaged seeds in plot) and number of agromorfological characteristics were analyzed. The results of combined analysis of variance showed highly significant differences among genotypes and revealed a considerable genetic variation for tolerance to fly among them. The lowest coefficient of variation was observed in damaged seeds in plot and the highest in percentage of damaged seeds in infected heads. The genotypes TN679 and TN796 showed the highest and the lowest number of damaged seeds in plot respectively. Genotypes Esfehan and TN690 also showed some levels of tolerance to the fly. Minimum seed damage was observed in genotypes with late flowering, particularly with a longer period for buding to flowering resulting in asynchronization between the host and the pest. Traits such as seed number in head, head size and bract width showed a high negative correlation with percentage of total damaged seeds in plot and percentage of damaged seeds in infected heads. Based on genotypic correlations between characters related to fly damage and other charecteristics, it seems selection of genotypes with early emergence and shooting stage, late flowering, high plant height, big heads and wide bracts may result in pest damage reduction and yield increase. Results of stepwise regression analysis on percentage of damaged seeds in plot against other traits showed that number of seeds in head, head number in plant, days to shooting stage and bract width constitiutes 72% of the variance. Also seed number in head, days to shooting stage, bract width with two traits such as seed hull thickness and period of flowering to seed maturity determined 90% of the variance. Path analyses showed that seed number in head had the highest direct effect on percentage of damaged seeds in infected heads while bract width had the highest indirect effect.The highest direct effect on percentage of damaged seeds in plot was related to bract width. According to path analyses, selection for increased seed number in head and bract width were most effective to reduce pest damage and increased seed yield.