DIFFERENT METHODS OF ESTIMATING THE PERFORMANCE OF DOUBLE CROSSES IN SUNFLOWER (Helianthus annuus L)

Abstract

Fifteen single crosses, 60 three-way crosses and 45 double crosses were produced from six inbred lines of sunflower (EMB, EUR, L3, L6, L10 and PERE12). All these genotypes planted through 10th of July 2013 at a field in Al-Hawija, west of Kirkuk Governorate using randomized complete block design with three replications. The single and three-way crosses values were used to test five methods of double cross hybrids performance prediction for characters head area, number of seeds per head, 100 seeds weight, seeds yield per plant and oil percent, as compared with actual observed values of these double crosses characters. The results showed highly significant differences among double crosses means (expressed as mean of observed and predicted values from the five methods of prediction) and among the different predicting methods for all studied characters. In general, it was shown through comparison between the means of observed values of double crosses and the predicted one from the five prediction methods used in this study, that the first mehods of prediction which depend on single crosses was found to be the most consistent method for for predicting all characters, followed in importance by the methods that depends on three-way crosses, the second for number of seeds per head and oil percent, the third for 100 seeds weight and both secnd and third for head area. These results were supported through simple correlation coefficients among observed and all predicted values, as well as through stepwise regression test.