Half diallel cross among six parental bread wheat divergent genotypes were evaluated under rainfed and saline stresses. Type and relative amounts of genetic variance components that interacted with environments were detected. Highly significant differences between environments were recorded for all traits under consideration i.e. mean squares for genotypes, parental lines or varieties, F1 hybrids and parents vs. hybrids were significant for all traits in both locations as well as the combined analysis except parents vs. hybrids at Maryout and the combined analysis for number of grains / spike. The collected data revealed that predominance of additive gene effects in the genetic control of all traits except plant height under Siwa conditions and combined analysis. Combining ability x environment interaction revealed that the GCA (additive and additive x additive genetic effects) was more distorted by environmental fluctuations than specific combining ability (SCA) effects for all traits except number of spikes / plant and 1000-grain weight.
Mean performances of the parental genotypes and its derived F1 crosses at both locations showed that P1 x P5 and P2 x P3 under rainfed andP1 x P4 and P3 x P4 under Siwa saline conditions were the superior combination for grain yield and most of its components. For grain yield / plant; eleven, fifteen and thirteen crosses expressed significant positive heterotic effects relative to mid parent in Maryout, Siwa and the combined analysis, respectively. While, six, five and six crosses from the previous hybrids exhibited significant positive heterotic effects relative to better parent in the same order. The best crosses were P1 x P5, P2 x P3, P2 x P5, P3 x P5 and P3 x P6 over both locations.
With regard to yielding capacity, P4 was the best combiner under saline environment as well as the combined data. Also, P4 had a significant gi effects for two or more of the yield attributes under each of the two stress types tested. Such parental genotype (P4) was developed under similar aimed conditions through the previous segregating generations among Desert Research Center breeding program.
For grain yield / plant, six, three, and five crosses gave significant positive Sij effects in Maryout, Siwa and the combined analysis, respectively. The best crosses were P1 x P5 and P4 x P5 in both locations and combined analysis and P3 x P5 in Maryout and the combined analysis.
Insignificant associations between parental means and their gi effects where detected in all cases revealing that it is not necessarily that good combiner must have a good index of intrinsic performance.