correlation between traits and path analysis of safflower grain yield under water stress conditions

introduction safflower is native to iran, and is cultivated for high adaptability, low water requirement, resistance to drought and useful unsaturated fatty acid (ashri et al., 1977). identification of effective agronomic traits and relationship between them are important aims under water stress conditions. among these characters, grain yield is the most important trait that is genetically polygenic and its direct selection has not been effective to improve this character, while this selection method can be useful for yield components. ashri et al., (1977) by using path analysis found that seed yield variation among safflower genotypes was mostly related to diversity in 1000-seed weight and number of heads per plant under limited moisture conditions. therefore, this study aimed to identify hidden factors that form traits, their impact on seed yield and also recognition of yield components, which playthe greatest role in determining yield under different water stress conditions. materials and methods present research was arranged at bu-ali sina university research farm during 2012. the experimental design was split plot based on randomized complete block design with three replications. the main plot was allocated to three levels of water stress including no irrigation (control), cutting irrigation from %50 flowering to ripening (flowering stress) and no irrigation from onset of seeding stage to ripening (seeding stress). six safflower cultivars such as pi, local ajabshir, mec11, faraman, local zarghan 6 and sina as the second factor were placed in subplots. each plot contained of 5 rows with 5 meters length, 50 cm row spacing, 5 cm spacing on each row. at the end of the growing season, after removing marginal rows, plants were harvested at the extent of two square meters from three central rows of each plot. then seed yield was determined in the scale of gram per square meter, and statistical analysis was performed using sas and minitab softwares, path analysis was also carried out using path2 software, graphs were drawn with excel software, and means compared by lsd test at the 5% level of probability. results and discussion the results showed that correlation coefficient of seed yield with harvest index at three moisture regimes including non-stress, flowering stress and seeding stress were respectively, 0.42*,0.76**and 0.46**, and seed yield with 1000 seed weight were also 0.53**, 0.88** and 0.43** respectively. in general it can be concluded that at flowering and seeding water stress conditions, the correlation between seed yield and harvest index was greater than non-stress condition. under flowering stress, 1000 seed weight was more affected by stress impacts and the consequence of this reduction decreased seed yield. stepwise regression method was used to evaluate the effect of each trait on seed yield as dependent variable, and also to reduce the number of independent variables. regression analysis of effective traits on grain yield showed that harvest index and biological yield under three water stress conditions entered into the regression model. under non-stress condition number of capitule per plant, seed weight before harvest index and biological yield also entered into the regression model, while at flowering stress 1000 seed weight joined the model prior to harvest index and biological yield. stepwise regression method, path analysis and correlation coefficient estimations were considered for different traits in spring safflower. results showed that 1000-seed weight, number of seeds per plant, number of seeds per head and biological yield were suitable indicators that could genetically improve seed yield under water stress conditions (golparvar and ghasemi pirbaluti, 2010). results also indicated that biological yield (0.99) and harvest index (0.90) under non-stress condition, while 1000 seed weight (1.82) at flowering stress, and harvest index (1.60) at seeding stress condition had the highest direct effects on grain yield. conclusions improvement of traits that have the most direct effects on grain yield and the highest coefficient correlations with grain yield can be considered to increase safflower grain yield at each level of water stress treatments.