Independent culling is useful when aimed to selecting for multiple traits. Thus, the present work aims to study the efficiency of selection by independent culling levels (ICL) for the three traits, No. of spikes/plant, 100-kernel weight and spike length under normal and late planting dates and to estimate observed and expected response to selection and genetic parameters after two cycles of phenotypic selection in the F2 population. The F5 selected families under normal and late planting were evaluated under both conditions. The genotypic variance was slightly less than the phenotypic one under both planting dates and generally decreased from the base population (F3) to the F5-generation. After the second cycle, heritability estimates were high, and reached 90.14, 81.22 and 85.71% under normal planting date, and 75.00, 78.97 and 85.71% under heat stress for No. of spikes/plant, 100-kernel weight and spike length, respectively. After the second cycles in the F5-generation, evaluation under normal planting date showed observed gain of 0.57, 3.97 and 2.19% was achieved for the normal planting group, and 0.90, 22.21 and 4.85% for the late planting group, from the bulk sample, for number of spikes/plant, 100-kernel weight and spike length respectively, however, the observed gain was significant (P<0.01) increase for number of spikes/plant, 100-kernel weight and spike length under normal and late planting conditions from the better parent. Evaluation under late planting date, showed that the observed gain from the bulk sample was 1.69, 6.18 and 2.38% for normal group, and 7.55, 11.12 and 4.04% for the late planting conditions, for the same respective traits, respectively. The relative merits of selection under late planting to selection under normal planting conditions in changing the mean number of spikes/plant indicated that antagonistic selection was better than synergetic selection.