Abstract: Two field experiments were conducted in 2010/2011 and 2011/2012 growing seasons at the experimental farm of the Faculty of Agriculture, Cairo University, Giza, Egypt. Twenty bread wheat Egyptian cultivars were evaluated in alpha lattice design with three replications for nine traits. The aims were to determine relationships between yield and its components and examine the efficiency of such components in building yield capacity by using four different statistical methods. Highly significant differences were detected among cultivars for all studied traits. Highly significant and positive correlation estimates were detected between grain yield plant-1 and each of number of number of tillers plant-1, number of spikelet’s spike-1, number of grains spike-1, 1000-grain weight and harvest index. On the other hand, days to 50% heading and plant height showed negative association with grain yield plant-1. Based on simple regression analysis, linear regression of number of tillers plant-1, spike length, number of spikelet’s spike-1, number of grains per spike, thousand grain yield and harvest index it leads to increase the grain yield plant-1 by 0.67, 0.52, 0.32, 0.30, 0.64 and 0.63 units, respectively. Path analysis showed that maximum positive direct effect on grain yield plant-1 was contributed mostly by number of tillers plant-1, followed by number of grains spike-1, harvest index and 1000-grain weight were the major contributors towards grain yield. Also, stepwise multiple linear regression analysis revealed that four traits included number of tillers plant-1, number of grains spike-1, harvest index and 1000-grain weight with R2 = 97.29%, had justified the best prediction model. Results of stepwise regression and path analysis revealed that the two methods are equivalent in determine the dependence relationship between grain yield and yield component characters. Also, results in the study with respect to four statistical methods which have been used in this study showed that that the number of tillers plant-1, harvest index, number of grains spike-1 and 1000-grain weight were the most important characteristics and they were highly effective on grain yield. These characters have to be ranked the first in any breeding program to improve wheat grain yield.
Keywords: Wheat, Grain yield, Statistical procedures, Simple correlation, Path analysis, Stepwise multiple linear regression analysis.