A half diallel cross among 9 inbred lines of maize (Zea mays L.) was evaluated under two locations (Fac. Agric. Moshtohor and Quesna menofiya Governorate) for some quantitative characters. General and specific combining ability were estimated according to Griffings (1956) diallel cross analysis designated as method 2 and 4 model 1 for each location as well as the combined over them. Location mean squares were significant for all traits under study except for no. of rows/ear, in both methods, with mean values in L1 being higher than those in L2 for all traits in both methods of analysis (M4 and M2) except no. of days to 50% pollen sheding. In both methods, genotypes mean squares were significant for all traits in both locations as well as the combined analysis. The mean squares due to general and specific combing ability were highly significant for all traits in both methods as well as the combined analysis. In the method 4, high ratios which largely exceeded the unity were obtained for all traits in both locations as well as the combined analysis, except days to 50% pollen shedding at L2 only. The parental inbred lines no. 9, 1 and 8 seemed to be good cominers for yield and some of its components in both methods of analysis. The crosses P1xP4, P1xP8, P1xP9, P4xP7, P6xP9 and P8xP9 showed the highest values for SCA effects for grain yield and its components in both methods of analysis. In method 2, no. of rows/ ear at L2 had GCA/SCA ratio equal unity. High ratios which largely exceeded the unity were obtained for days to 50% tasseling, 100-kernel weight and grain yield/ plant in both locations and the combined analysis. This finding coincided with that already reached from the combining ability analysis of Griffing method-4. For the other remain cases, GCA/SCA ratios, were less than unity. The mean squares of interaction between locations and both types of combining ability were significant for all traits in both methods of analysis. The correlation coefficient between ( ) and ( ) effects in two methods (Griffing’s method-2 and 4) was significant for all traits. For additive gene effects, insignificant F-test was detected between the two methods used in this concern for all traits. The correlation coefficient of additive effects in the two methods for all traits used in this study was highly significant between the two methods i.e. Griffing’s method-2 and 4 indicating strength relation between the two diallel methods in the estimation of additive components. The efficiency of the two diallel methods of analysis for additive x location interaction based on F. test. Insignificant F. test was obtained between the two methods (Griffing’s method-2 and4) used in this study for all traits. The correlation coefficient of additive x location interaction for all traits, significant correlation values between the two methods were detected. The efficiency of the both diallel methods of analysis for non additive gene effects based on F-test. The results indicated that significant F. test was detected for all traits in both locations as well as the combined analysis. The correlation coefficient of non-addiitive effects derived from two methods of diallel analysis over all traits in both locations as well as the combined analysis were highly significant indicating the strong relation between the Griffing,s methods in estimating non-additive genetic variance. The efficiency of the two diallel methods of analysis for non-additive x location interaction based on F. test, insignificant F. test was obtained between the two methods (Griffing m-2 and m-4) used in this study for all traits. For error mean squares, insignificant F. test was detected between Griffing’s method-2 and 4 for all traits in both locations and the combined analysis. |