Identification of selection traits for yield improvement of bread wheat using path analysis

desconocido. The present study was carried out at Assiut Univ. Exp. Farm during the three successive seasons of 1997/98 to 99/2000, to identify traits useful as selection criteria in improving yield using path analysis and to compute observed response to pedigree selection. The breeding materials were the F3, F4 and F5-generations of the wheat multiple cross (Giza 160 x Pitic 62) x Sakha 69 x Sonora 64) x (Yecora Rojo x Sonora 64). Highly significant differences among F3 families (base population), sufficient genotypic coefficients of variability and high broad sense heritability were obtained for all the studied traits. In the base population, grain yield/plant showed positive correlations with all studied traits except number of grains/spike. However, spikes/plant, grain weight and harvest index were negatively correlated. Path analysis revealed that grain weight had the highest direct effect on yield followed by spikes/plant and harvest index at both phenotypic and genotypic levels. indirect effect of grain weight on grain yield via number of spikes was negative. The indirect effect of days to heading via no. of spikes on grain yield was negative reflecting negative correlations between days to heading and no. of spikes. Selection for earliness resulted in earlier (-6.29) families from the bulk sample. However, earliness reduced no. of spikes, grain yield, grain weight and harvest index by -25.39, -23.17, -7.43 and -4.88, respectively, after two cycles of pedigree selection. Pedigree selection for each of 1000- grain weight, no. of spikes/plant, biological yield and harvest index increased the selection criterion by 14.49 , 16.26, 18.45 and 20.21 and also increased grain yield by 9.11, 7.38, 9.56 and 6.71, respectively, from the best parent after the two cycles. No. of grains/spike was negatively affected by increased grain weight and high number of spikes/plant and were reduced by -13.24 and -24.04, respectively from the best parent. The observed response to selection for grain yield was 21.42 from the best parent. The results indicated the relative importance of grain weight and harvest index in selection for high yielding wheat families. Results also indicated that path analysis could be successfully used to identify appropriate selection criteria for improving grain yield of wheat.