ABSTRACT
Three cycles of pedigree selection from the F2 for grain yield/plant (GY/P) and one cycle from the F4 were achieved in four seasons under normal irrigation and drought stress environments. The objectives were to study the relative merits of pedigree selection for (GY/P) under drought stress and normal irrigation environments, to compare the efficacy of three cycles of selection started in the F2 with one cycle started in the F4 generation, and to study the sensitivity of the selected families. The genetic materials were 440 F2 individual plants stemmed from the cross Sids4 x Giza164. The reductions in GY/P caused by drought stress were 42.88, 13.64 and 25.81% for the F2 generation, Sids4 and Giza164; respectively. The reduction% in the other traits ranged from 0.58% for harvest index to 24.97% for number of spikes/plant (NS/P). The phenotypic correlations among traits in the F2 were larger under drought stress than under normal irrigation and in the same direction. Grain yield/plant showed positive correlation with NS/P and harvest index. Days to heading showed significant correlations with plant height, NS/P, GY/P and harvest index under both environments. The PCV and GCV of GY/P decreased from 76.01 and 45.81% in cycle0 (C0) to 20.15 and 19.16% after C3 under drought stress, and from 48.31 and 28.59% in C0 to 18.67 and 18.38% after C3 under normal irrigation; respectively. The realized heritability of GY/P increased from 23.02 and 41.06% in C1 to 67.24 and 42.92% in C3, under drought stress and normal irrigation; respectively. The observed genetic gain (OG) in GY/P of drought selected families in C3 was 20.69 and 17.98% under drought stress, and 31.18 and 30.13% under normal irrigation evaluation of the bulk sample and the better parent; respectively. While, the OG in GY/P of the normal irrigation selections was 17.12 and 14.49% under drought stress, and 20.43 and 19.47% of the bulk sample and better parent; respectively under normal irrigation conditions. The results of OG of the three cycles indicated that antagonistic selection was better than synergistic selection in changing GY/P mean. The OG of the one cycle of selection from the F4 generation under both environments was mostly better than the three cycles started from F2 generation. Furthermore, the 10 selected families from the F4 traced back to different F2 plants from the 10 selected families after three cycles under both environments, indicating no correlation between GY/P of the individual plants in the F2 and in the F4 families. The results suggest delaying selection for GY/P to F4 and F5 generations until the recombinants reached acceptable level of homozygosity, to save efforts and expenses. Antagonistic selection for GY/P (under drought stress) was better than synergistic selection (under normal irrigation) in decreasing sensitivity to environments of the selected families.