الفهرس 
INTRODUCTIONOnly 14 pages are availabe for public view
 |  | from | 139 |  |  |
| | | from | 139 | | |
Abstract
The present study was carried out at El-Mattana Agric. Res Station, Agric. Res. Center, Egypt, during the three successive seasons of 2005/ 2006, to 2007/2008.
The objectives of the present investigation were to study the response to selection in two durum wheat populations using the pedigree selection method, also to study the interrelationships among the studied traits for improving grain yield and its components and some quality traits.
The materials used in these studies were two wheat populations, as follows:
1- Population I : Triticum turgidum var. durum (kuck x Rascon/kitti).
2- Population II: (Sohag 2 x Sohag 3).
The F3 plants were used as a base population for selection to drive F4 and F5 generations. Selection was based on one of the following criterion:
1) Grain yield/plant.
2) Number of spikes/plant.
3) 1000-kernel weight (g) and
4) Number of kernels/spike.
In addition to protein content in the F5 generation. In addition, the parents of each cross, and the bulk population, were used for comparison of the selected families in each generation.
The studied traits were:
1- Days to heading. 2- Plant height, cm.
3- Number of spikes/plant. 4- 1000-kernel weight (g).
5- Number of kernels/spike. 6- Grain yield/plant (g).
7- Biological yield/plant (g). 8- Protein content (%).
9- Semolina (%). 10- Flour (%).
11- Bran (%).
The obtained results could be summarized as follows:
1. Selection for grain yield/plant:
A. Variances and means:
The results indicate significant or highly significant differences among selected families in the population I and II. The average grain yield/plant of the selected families of populations I and II increased from 18.4 and 17.7 in base population to 24.3 and 26.6 g/plant in F5 generation (cycle 2), respectively. The average grain yield/plant of the bulk populations was 16.2 and 20.7 g/plant for population I, and 20.9 and 19.5 g/plant for population II, in F4 and F5 generations, respectively.
The results indicated that pedigree selection decreased the genetic variability measured as a genotypic coefficient of variability of the selection criterion and other studied traits after two cycles of selection in both populations. For example, the GCV of grain yield/plant decreased from 28.60% in the base of population I to 3.8% after the two cycles of selection, number of spikes/plant from 26.2 to 8.5% and biological yield/ plant from 31 to 7.8%.
Also in population II, the GCV of grain yield/plant, number of spikes/ plant and biological yield/plant decreased from 31.3, 27.3 and 32.3% to 3.8, 7.3 and 5.9%, respectively.
B- Heritability estimates:
Estimates of broad sense heritability after two cycles of selection in population I and II, were moderate or high for grain yield/plant (47 and 54.4%), heading date (54.11 and 71.2%) and number of spikes/plant 73.19 and 78.4%), respectively.
The realized heritability estimated from the realized gain in both cycles of single trait selection in the two studied populations also decreased from C1 to C2 for most studied traits.
Narrow sense heritability values computed by regression in population I for the first cycle were low for most of the selection criterion. The same conclusion was obtained in population II except number of spikes/plant which was moderate (0.40) while grain yield/plant was high (0.55).
C- Realized and correlated response to selection:
The results showed that direct selection for grain yield/plant increased it by 17.39% from the bulk sample after two cycles of selection. Such increase accompanied with increase in biological yield/plant (26.36%), 1000-kernel weight (6.1%), number of kernels/spike (6.65%) and number of spikes/plant (8.89%) and decrease in heading date (-1.78%) and plant height (-3.68%), in population I. The results in population II behaved the same as in population I.
D- Phenotypic and genotypic correlations:
The positive phenotypic and genotypic correlations were obtained between grain yield/plant and all studied traits except genotypic correlation for heading date (-0.05) in population I, and it was positive between grain yield/plant and all studied traits in population II.
E- Means of the selected families:
After two cycles of the direct selection for grain yield/plant, two high yielding families (No. 7 and 4) were obtained. Family No. 7 yielded 36.71, 44.38 and 63.58% more than the bulk sample, the best parent and the check cultivar, respectively, and family No. 4 yielded 35.26, 42.85 and 61.85% more than the bulk sample, the best parent and the check cultivar in population I, respectively.
In population II, the direct selection for grain yield/plant gave family (No. 2) which out yielded the bulk sample, the best parent and the check cultivar by 49.74, 68.78 and 68.78, respectively. In addition, all selected families in the two populations significantly out yielded the check cultivar. These results proved that the pedigree method of selection was more effective in improving grain yield/plant in these materials.
2- Selection for number of spikes/plant:
A- Variances and means:
The over all mean of the selected families for number of spikes/plant after two cycles of selection in population I increased from 9.7 in cycle 1 to 10.2 spikes in cycle 2.
The genotypic (GCV%) coefficient of variability of number of spikes/plant in the base population decreased from 26.2 to 4.7% after the two cycles of selection. Also, grain yield/plant decreased from 28.6 to 12.6% and No. of kernels/spikes from 13.66 to 9 for population I, also, in population II, The GCV of number of spikes/plant, grain yield/plant and number of kernels/spikes decreased from 27.3, 31.3 and 17.39 to 7.7, 12.1 and 5.1%, respectively.
B- Heritability estimates:
The results indicated that heritability estimates decreased from (88.9 and 88.4% in the F3 (Base population) to 52.6 and 63.7% in the F5 generation (C2) for number of spikes/plant in populations I and II, respectively. The realized heritability estimates and the narrow sense heritability computed by regression showed the same trend observed for the selection criterion in other studied traits.
C- Realized and correlated response to selection:
Selection for number of spikes/plant improved it by 13.33 and 12.09% from the bulk sample and the check cultivar after two cycles of selection in population I, respectively. Such increase accompanied with increasing number of kernels/spike (5.29%), 1000-kernel weight (4.88%), grain yield/plant (12.08%) and biological yield/plant (19.45%), and decreasing heading date (-1.59%) and plant height (-2.13%) from the bulk sample.
D- Phenotypic and genotypic correlations:
The results showed that positive phenotypic and genotypic correlations between number of spikes/plant and all studied traits in the two populations except heading date (-0.26 and -0.84) in population I.
E- Means of the selected families:
Two cycles of pedigree selection for number of spikes/plant as a selection criterion gave family No. 5 which was higher in grain yield/plant by (35.26 and 61.84%), number of spikes/plant (27.77 and 26.37%), number of kernels/spikes (18.7 and 25.51%), and biological yield/plant (41.1 and 77.16%) than the bulk sample and the check cultivar, respectively, in population I.
For population II, family No. 4 was higher in grain yield/plant by (42.56 and 60.69%), number of spikes/plant (35.1 and 39.56%), 1000-kernel weight (16.81 and 19.53%) and biological yield/plant 39.65 and 86.52%) than the bulk sample and check cultivar, respectively.
3- Selection for 1000-kernel weight:
A- Variances and means:
The analysis of variance revealed that highly significant differences among genotypes for all studied traits in the two cycles of selection of the two population. The overall means of the selected families after two cycles of pedigree selection in the two populations for 1000-kernel weight ranged from 43.9 and 50.1 in cycle 1 to 45.9 and 50.1 in cycle 2. Also, the bulk sample ranged from 40.4 and 43.4 in cycle 1 to 41 and 44 in cycle 2 in population I and II, respectively. Pedigree selection for 1000-kernel weight in population I reduced the genotypic coefficient of variability from 10.9 to 7.2% after the two cycles of selection and from 13.6 in base population to 5.8% after two cycles in population II.
B- Heritability estimates:
Estimates of broad sense heritability in the two cycles were for 1000 kernel weight (64.8 and 60.3%), plant height (75.4 and 72.8%), number of spikes/plant (75.4 and 82.8), number of kernels/spike (76.6 and 78.7%), grain yield/plant (62.4 and 72%) and biological yield/plant (69.6 and 79.4%) for populations I and II, respectively.
Realized heritability values for the 1000-kernel weight were 28 and 37% in C1 and C2 of population I, while they were 50 and 0.0% of population II, respectively. In the first cycle, narrow sense heritability estimates computed by regression for 1000-kernel weight were 47 and 91% in population I and II, respectively.
C- Realized and correlated response to selection:
1000-kernel weight increased this trait by 11.95 and 6.74% from the bulk sample and the check cultivar, respectively, after two cycles of selection in population I. Such increase accompanied with increasing in grain yield/ plant (1.45 and 21.39%), number of kernels/spike (3.32 and 9.9%) and biological yield/plant (8.0 and 35.62) and decrease in heading date (-1.56 and -3.71%). The results in population II behaved the same trend as in population I.
D- Phenotypic and genotypic correlations:
The results of the phenotypic and genotypic correlations indicate positive correlation between 1000-kernel weight and all studies traits except No. of kernels/spike which was negative for genotypic correlation in population I. In population II, the results of the phenotypic and genotypic correlations indicate positive correlations between 1000-kernel weight and each of number of kernels/spike and plant height but it was negative with heading date, number of spikes/plant and grain yield/plant.
E- Mean of selected families:
Two cycles of pedigree selection for 1000-kernel weight as a selection criterion gave some superior families. For example, in population I family No. 10 was higher in 1000-kernel weight by (35.36 and 29.07%), grain yield/plant by (24.63 and 49.13%), biological yield/plant by (24.72 and 56.62%) and number of spikes/plant by (13.33 and 12.08%) than bulk sample and check cultivar, but it was less in heading date (-1.56 and -3.7%) and plant height (-3.97 and 0.0%) than the sample and check cultivar, respectively.
For population II, family No. 6 was higher in 1000-kernel weight by (27.05 and 30%), number of kernels/spike (5.9 and 15.94%), number of spikes/plant (12.77 and 16.48%) and grain yield/plant (0.0 and 12.7%) while it was loss in plant height (-5.33 and -2.28%) and heading date (-2.6 and -5.99%) than the bulk sample and the check cultivar, respectively.
4- Selection for number of kernels/spike:
A- Variances and means:
The analysis of variance revealed significant or highly significant differences among genotypes for all studied traits in the two cycles of selection in the two populations except heading date which was non significant in F5 generation for population I.
The average of the selected families in the two populations were 70.6, 73.9, 67.4 and 71.6 for number of kernels/spike in the F4 and F5 generations, respectively.
On the other hand, the average of bulk population was 68, 66.2, 68.7 and 68.6 for number of kernels/spike in the F4 and F5 generation in population I and II, respectively. Pedigree selection for number of kernels/ spike in population I reduced the genotypic coefficient of variability from 13.66% in the base population to 8.1% after the two cycles of selection for number of kernels/spike. On the other hand, the magnitude of the phenotypic variability ranged from 8.4% of the F5 to 14.26% in the base population. The same trend was observed in population II.
B- Heritability estimates:
Estimates of broad sense heritability in the two cycles for number of kernels/spike were 94.1 and 49.5%, in population I and II, respectively. Realized heritability values varied according to the generation and they were 35 and 25% in the F4 families and 76. and 66% in the F5 families in population I and II, respectively.
Moreover, heritability estimates by regression method were 72 and 23% for (F4/F3) and 67 and 67% for (F5/F4) generations in the two population, respectively.
C- Realized and correlated response to selection:
Number of kernels/spike increased by 15.47, 11.63 and 17.86% from the mid parent, the bulk sample and the check cultivar, respectively, after two cycles of selection in population I.
The results in population II were significantly higher in number of kernels/ spike by 14.19% than the check cultivar but it was insignificant from the mid parent and the bulk sample (10.24 and 4.37), respectively.
In population I, the correlated response computed for number of kernels/spike revealed that number of kernels/spike varied in correlated response with the other traits in the range of negative significant of -11.45 to positive of 0.83% with plant height, -6.44 - -0.78% with heading date, -3.23-5.81% with number of spikes/plant, -7.21-6.61% with 1000-kernel weight, -4.35-26.52% with grain yield/plant and it was positively significant correlated with biological yield/plant.
These results indicated that direct selection caused an improvement of kernels/spike and indirectly improved grain yield/plant. The results in population II behaved the same as in population I.
D- Phenotypic and genotypic correlations:
The results of the phenotypic and genotypic correlations indicate positive correlation between number of kernels/spike and all studies traits in population I and II.
E- Mean of selected families:
Two cycles of pedigree selection for number of kernels/ spikes gave 15 F5 families which were significantly higher than bulk population, the best parent and the check cultivar in population I.
In population II, 5 F5 families exceeded significantly the check cultivar.
5- Protein content:
A- Variances and means:
The analysis of variance revealed significant or highly significant differences among families for protein content, semolina and other studied traits in the F5 families in the two populations. Except heading date it was insignificant in population I, also, 1000-kernel weight was insignificant in population II.
The overall mean of the selected families in F5 for protein content and semolina in the two populations were 16.6, 17.6, 68.8 and 68.2%. Also, the bulk sample were 14.6, 14.9, 67.3 and 67.7% in population I and II, respectively.
B- Heritability estimates:
Estimates of broad sense heritability in F5 generation were high for protein content (94.97 and 95.39), semolina (82.02 and 80.31), flour (85.21 and 83.59) and bran (71.99 and 76.20 for population I and II, respectively.
C- Realized and correlated response to selection:
Protein content increased by (16.9, 13.7 and 20.29%) from the mid parent, the bulk sample and the check cultivar, respectively, in population I. Such increase accompanied by increase in semolina (3.85, 2.23 and 0.88%) and decrease in flour (-17.28, -19.39 and -11.24%).
The results in population II behaved the same as in population I.