Six rice genotypes with different drought tolerance were crossed. Six populations (P1, P2, F1, F2, BC1 and BC2) of three rice crosses namely Giza 177 x Giza 178 (cross I); Sakha 103 x WAB 880 SG 33 (cross II) and Sakha 104 x IET 1444 (cross III) were raised in a randomized complete block design with three replications during the three successive summer seasons from 2010 to 2012 at the Farm of Rice Research and Training Center Sakha, Kafer El-Sheikh, Egypt. This study aimed to determine the heterosis, gene action, heritability, genetic advance and phenotypic correlation response to selection and prediction by the new lines for the traits of days to 50 % heading, plant height, panicle length, number of panicles/ plant, number of filled grains/ panicle, 100-grain weight, sterility % and grain yield/ plant. Flashing water irrigation was added every 10 days intervals.
The results obtained from the mean of parents, F1 and F2 generations showed wide range in mean values between the parents and the presence of partial and over-dominance were found for all studied characters. Significant and highly significant positive estimates as a deviation from mid and better parents in all crosses for panicle length, number of panicles/ plant and grain yield/ plant, also highly significant positive estimates as a deviation from mid and better parents in the first two crosses for plant height. Scaling test provide evidence of non-allelic interaction in controlling all the characters studied in most crosses, the additive gene effect (d) was more important in the genetic system for all studied characters, dominance gene effects (h) and additive x additive gene effects (i) were played an important role in the inheritance of plant height, panicle length, number of filled grains/ panicle, 100-grain weight, sterility % and grain yield/ plant. The additive x dominance (j) were significant and involved in the inheritance of days to 50 % heading, plant height, number of panicles/ plant, number of filled grains/ panicle, 100-grain weight, sterility % and grain yield/ plant. However, the dominance x dominance (l) were involved in the genetic control of all characters. Heritability estimates in broad sense were high. Meanwhile, heritability estimates in narrow sense were mostly low. The maximum genetic advance of the mean values was found to be high for number of panicles/ plant, plant height and days to 50 % heading characters. Highly significant and positive correlation was found for grain yield/ plant with panicle length, number of panicles/ plant, number of filled grains/ panicle and 100-grain weight in all studied crosses.
From the foregoing results, cross II (Sakha 103 x WAB 880 SG 33) and cross III (Sakha 104 x IET1444) could be recommended for growing under water deficiency to obtain the highest rice grain yield and the highest values of saving water at the same time.