In order to compare the efficiency of different half-diallel methods in the analysis of bread wheat(Triticum aestivum L.), 15 genotypes from a six-parental diallel cross, excluding reciprocals, were grown in the field using a randomized complete block design (RCBD) with three replications under two different water regimes (irrigated and rainfed). Comparison of Griffing’s model I, method II; Morley- Jones method; Gardner and Eberhar’s method and Hayman’s method indicated regardless of the restrictions of the Griffing model, it has the advantages of the definition of formulas for estimating the effects, their variances and the variances of contrasts of effects, as well as for the calculation of orthogonal sums of squares. Therefore, it is generally quite safe to use the Griffing model. Gardner and Eberhart appears to have the following advantages over the others: (1) it provides general combining ability (GCA) and specific combining ability (SCA), (2) information on the additive effects of varieties and their average and individual contribution to heterosis in crosses, (3) a clear-cut idea about the genetic aspect of heterosis by partitioning the total sum of squares of heterosis (hij) into different components, (4) since in this analysis parents are also included, and there is a simple relationship between heterosis (hij) and specific combining ability (sij) as hij = (2 sij - sii - sjj) / 2, heterosis can be easily calculated, (5) the results further indicated the possibility of dominance variance being confounded with the additive variance of general combining ability.
