COMBINING ABILITY ANALYSIS OF EARLINESS, SEED YIELD AND RELATED TRAITS IN SUNFLOWER

Twenty-six genotypes of sunflower (16 F₁-hybrids, four female lines, four restorer lines and two check varieties; Sakha 53 and Giza 102) were evaluated under two contrasting environments, i.e., loamy sand and clay soils in season 2016. Genotypes mean squares of the studied traits was significant (P<0.01) either in the separate or in the combined analysis. The differences between the two environments were significant for all traits except head diameter (HD).  The genotype x environment interaction was significant for all traits. Furthermore, most of the variability was for environment, except for husk %. Mean squares of the combined analysis of female and testers lines was significant (P<0.01) for all traits. These results indicate the presence of additive variance. Mean squares of parents vs. crosses and lines x testers were significant for all traits, indicating the presence of non-additive in the inheritance of these traits. The combined analysis indicated that mean squares of lines x environment was significant for all traits, except for HD. Mean squares of testers x environment was significant except for days to 50% flowering and HD.  The interaction mean squares of LxTxE were significant for all traits, indicating the interaction of non-additive gene effects with environment. The results of the combined analysis indicated that the ratio s²A/s²D was less than unity for all traits, and the role of dominance was more important than that of additive effects. The results of GCA indicated that none of the female or male lines was the best combiner for all traits. Thirteen out of the 16 hybrids were significantly (P<0.01to P<0.05) earlier than the earliest check cultivar Giza 102.  The performance of the F₁-hybrids in days to 50% flowering were mostly related to the GCA of the parents rather than the SCA of the hybrids. The combined analysis of plant height showed that eight hybrids gave negative SCA effects. All the F₁-hybrids were significantly (P<0.01) shorter than the two check cultivars. Based on the combined analysis; eight hybrids had positive SCA for head diameter; but none exceeded the check variety in head diameter. Based on the combined analysis 8 hybrids showed negative SCA for husk%, the performance of all hybrids was significantly (P<0.01) lower in husk % than the better check Sakha 53. The combined analysis of oil % indicated that five hybrids showed significant positive SCA, four of them exceeded significantly (P<0.01 to P<0.01) the better check cultivar Giza 102. The combined SCA effects of seed yield/head (SY/P) were positive and significant for three hybrids (A7 x RF1, A15 x RF3 and A21 x RF5). The performance of the first hybrid (46.45 g/head) was significantly (P<0.01) better than the better check Giza 102 (41.21 g/head). The hybrids performance was not in accordance with sign and significance of SCA of SY/P. Furthermore, the GCA of the parents was far from yielding ability. The combined SCA of five hybrids for oil yield/head were positive and significant (P<0.01). The performance of the first hybrid (A7 x Rf1) (18.18 g) exceeded significantly (P<0.01) the better check Giza 102 in oil yield/head (15.43 g). It could be concluded that the performance of the hybrids was not in accordance with the sign and significance of the SCA effects. This could be due to that the ratio of s²A/s²D was less than unity and the dominance effects were more important than additive in the inheritance of all traits, and evaluation of hybrids should be at a variety of environments