الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 184 |  |  |
| | | from | 184 | | |
Abstract
Wheat is the world’s most important cereal crop both in area and production. Wheat is the strategic cereal crop not only in Egypt, but also all over the world. Egypt produces about 8.2 M tons and consumes 17.9 M tons. The gap between total production and consumption is met by imports. Also, demand of wheat is increasing with increasing population. Consequently, the maximum crop yield via selected new genotypes is an important objective in most breeding program in wheat.
The objectives of the current study were:
1- Estimate the efficiency of single trait selection as grain yield/plant in increasing yield for two segregating populations of bread wheat under late planting date.
2- Study the relationship between yield and some yield components through the correlation coefficients across base and two successive generations in the two segregating populations.
3- Assess the contributions of highest correlated traits with yield via path coefficient analysis in base population as well as the next two successive generations across pedigree selection for the two studied population.
4- Test and determine the better models had the significance of the independent variable affecting the grain yield via the stepwise regression analysis in the two studied populations.
The basic genetic materials in the current investigation were a bulk sample in F3-generation of two segregating bread wheat populations. The first population were produced from the cross Depeira x Sahel and the other one from cross Giza 165 x Sakha 93. The current study was carried out during the three successive seasons i.e., 2012/2013, 2013/2014 and 2014/2015 at Fac. Agric. Farm, Assiut University, Egypt.
The obtained results could be summarized in five main topics as following:
I- Description of the base population
The two base populations were presented by 497 and 210 F4 replicated families for population 1 and 2, respectively along with their parents and the unselected bulk sample.
1- The analyses of variance showed that the F4-families possessed high significant differences for all the studied traits in the two base populations. These results reflect the genetic differences among the F4-families for the studied traits in the two base populations, consequently, the presence of sufficient genetic variation that could be used for further practice of pedigree line selection especially for grain yield/plant.
2- High values of genotypic coefficient of variation (gcv) were recorded for plant height (14.89 and 12.15%), spike length (11.97 and 13.20%), number of spikes/plant (12.05 and 18.39%), number of spikelets/spike (7.01 and 12.92%), biological yield/plant (14.51 and 20.47%), weight of spikes/plant (14.84 and 23.13%), grain yield/plant (14.90 and 21.93%), harvest index (11.86 and 8.64% and threshing index (10.43 and 11.57%) in population I and II, respectively. These results were coupled with high estimates of heritability in broad sense which were more than 70.0 and 84.0% for all the studied traits in population I and II, respectively. The values of genotypic coefficients of variation coupled with the estimates of heritability only would seem to give the best picture of the amount of genetic advance expected from selection.
3- The values of phenotypic correlations between grain yield/plant and each of number of spikes/plant, biological yield/plant and weight of spikes/plant were high and accounted 0.652, 0.891 and 0.960 in population I and 0.786, 0.959 and 0.982 in population II, respectively. Also, weight of spikes/ plant possessed high phenotypic correlations with each of number of spikes/plant (0.751) and biological yield/plant (0.968) in population I, as well as (0.819) and (0.975) in population II, respectively.
4- Weight of spikes/plant was superior trait for its contribution (0.920 and 0.965) in grain yield/plant, followed by (in rank) biological yield/plant (0.794 and 0.920) and numbers of spikes/plant (0.425 and 0.617) in populations I and II, respectively. Moreover, the R2 values for all traits together on grain yield/plant were 0.987 and 0.995 in populations I and II, respectively. The path coefficient analyses revealed that about 95 and 97% of variance in grain yield/plant could be explained by these three traits in populations I and II, respectively. Also, weight of spikes/plant out of them had the major direct and indirect effects on grain yield/plant.
II- Pedigree line selection in the F4 and F5-generations for grain yield/plant, g
1- All entries as well as selected families after one cycle (F5) and two cycles (F6) of pedigree line selection showed significant or highly significant differences for grain yield/plant (selection criterion) and its correlated traits in both segregating populations.
2- The genotypic (gcv) and phenotypic (pcv) coefficients of variation for grain yield/plant were decreased from 14.90 and 17.12 in base population (F4) to 10.73 and 12.23 after one cycle (F5) and to 3.98 and 6.16% after two cycles (F6) of selection in population I . Also, the viewed values were decreased from 21.93 and 22.65 (F4) to 17.33 and 19.60 (F5) and to 2.68 and 8.85% (F6) in population II. The same trend could be found for correlated traits in both populations.
3- The estimates of broad sense heritability calculated from the expected mean squares were high (more than 70.0%) for most the studied traits after two cycles of selection in both segregating populations, except for grain yield/plant as selection criterion and its correlated traits of number of spikes/plant and harvest index were low and accounted 41.67, 41.55 and 48.67% in population I, respectively.
4- After two cycles of pedigree line selection for grain yield/plant, means over the F6 selection of population I ranged from 16.39 to 19.92 with an average of 17.82 g compared to the less values of their parents Depeira (15.83) and Sahel (11.94) as well as bulk sample (13.38 g). Also, the selections of population II varied from 16.66 to 22.03 with an average of 18.78 g compared to the less values of 12.41, 15.09 and 14.19 g for both parents Giza 165, Sakha 93 and bulk sample, respectively. It is clear that the selections of both populations surpassed their respective parents and unselected bulk sample in cycle 2 (F6) of pedigree line selection for grain yield/plant.
5- The selected families in F6 surpassed the better parent and unselected bulk sample for correlated traits i.e. number of spikes/plant, biological yield/plant, weight of spikes/plant, plant height, spike length and number of spikelets/spike in both populations.
6- The observed direct responses of pedigree line selection for grain yield/plant (selection criterion) in population I were 15.98 and 33.18, 13.68 and 12.57, and 26.94 and 28.39% over one and two cycles of selection as accounted from unselected bulk sample, better parent and mid parents. These direct responses were 8.89 and 32.35, 8.33 and 24.45, and 17.21 and 36.58% for same respective items in population II. It is clear that the direct response for grain yield/plant in cycle two (F6) was larger than cycle one (F5) in both populations indicating to the effectiveness of direct pedigree selection for grain yield in those studied populations of bread wheat.
7- It is clear that the three traits i.e. number of spikes/plant, biological yield/plant and weight of spikes/plant had the highest positive values of correlated response in cycle one (F5) and cycle two (F6) in both populations. The positive estimates for number of spikes/plant were 11.18 and 22.83, 10.70 and 21.01, and 17.57 and 23.62% in population I, and 9.07 and 37.47, 14.38 and 28.91, and 17.76 and 30.37% in population II, for biological yield/plant were 13.62 and 22.61, 9.51 and 15.20, and 20.62 and 24.77% in population I, and 17.86 and 36.51, 15.40 and 28.98, and 22.84 and 39.44% in population II and for weight of spikes/plant were 15.09 and 27.04, 11.11 and 13.45, and 22.09 and 29.75% in population I, and 13.03 and 34.44, 14.47 and 31.70, and 22.25 and 42.12% in population II in relative to the bulk, better parent and mid parents, respectively.
8- Direct response to pedigree line selection for grain yield/plant in population I revealed that all selections (20-F6 families) in cycle two of selection exceeded significantly the mid parents and unselected bulk sample. The direct response for grain yield/plant ranged from 18.08 to 43.52 with an average of 28.39% over the mid parents, and varied from 3.54 to 25.84 with an average 12.57% over the better parent. Also, it ranged from 22.50 to 48.88 with an average of 33.18% over the bulk sample. All the largest values of direct response were correlated to the Fam. no. 536. Moreover, seven families among all F6 families surpassed significantly the mean of better parent. These superior families were 28, 41, 108, 296, 460, 518 and 536 and accounted direct response of 15.35, 19.27, 14.97, 21.98, 22.05, 24.13 and 25.84% over the better parent (Depeira), respectively.
9- The direct response for grain yield/plant in population II could be the same trend as in population I, and exhibited that all selected F6 families surpassed significantly the mid parents and bulk sample. Moreover, eight selected families out of the ten families were exceeded significantly the better parent. The direct response for grain yield/plant ranged from 21.16 to 60.22 with an average of 36.58% over the mid parent and from 10.40 to 45.99 with an average of 24.45 over the better parent. Also, it varied from 17.41 to 55.25 with an average of 32.35% over the bulk sample. The highest values of direct response could be correlated with Fam. no. 249. The superior families surpassed significantly in rank the better parent (Sakha 93) in values of 45.99 (no. 249), 32.94 (no. 1), 32.60 (no. 70), 29.89 (no. 289), 23.06 (no. 192), 21.47 (no. 266), 20.34 (no. 236) and 16.04% (no. 154).
10- In population I, the averages of correlated responses for number of spikes/plant were 23.62, 21.01 and 22.83% over the mid parent, better parent and bulk sample, respectively. The highest correlated response matched the family no. 41. There were five families (nos. 41, 291, 474, 518 and 536) surpassed in highly significant means the mid parents, better parent and bulk sample. These five families surpassed the better parent and bulk sample by 59.13 and 61.52, 42.87 and 45.02, 35.16 and 37.19, 28.62 and 30.55, and 28.83 and 30.76%, respectively. The average of correlated responses for number of spikes/ plant in population II accounted 30.37, 28.91 and 37.47% over the mid parents, better parent and bulk sample, respectively. There were six out of the ten selected families i.e. 59, 154, 192, 236, 249 and 289 highly significant exceeded the bulk sample, better parent and mid parents. Their values of correlated response over the bulk sample and better parent were 35.08 and 26.66, 37.25 and 28.70, 67.76 and 57.30, 42.37 and 33.50, 58.28 and 48.42, and 44.55 and 35.55%, respectively.
11- The correlated response for biological yield/plant in population I revealed that the means of all the twenty selected F6 families were significantly surpassed the bulk sample and mid parent. The mean values of correlated response were 24.77, 15.20 and 22.61% over the mid parents, better parent and bulk sample, respectively. The highest correlated response was correlated with Fam. no. 536 over the three scales. The highest ten superior F6 families in respective rank surpassed highly significant the better parent by 36.10 (no. 536), 34.40 (no. 41), 25.75 (no. 401), 24.66 (no. 28), 22.03 (no. 460), 20.86 (no. 261), 18.84 (no. 518), 18.41 (no. 291), 18.03 (no. 296), and 16.37% (no. 313) in biological yield/plant. In population II, the means of all the ten F6 families were exceeded the bulk sample, mid parents and better parent in highly significant differences. The averages of correlated response were 39.44, 28.98 and 36.51% over the mid parents, better parent and bulk sample, respectively. The family no. 249 had the highest correlated response over the three scales. The ten selected F6 families were exceeded in highly significant differences the better parent and recorded values in respective rank of 52.35 (no. 249), 48.90 (no. 1), 40.07 (no. 192), 31.39 (no. 289), 26.64 (no. 226), 26.61 (no. 70), 22.48 (no. 236), 20.01 (no. 59), 11.96 (no. 154) and 9.43 (no. 279).
12- Means of selected F6 families in population I for weight of spikes/ plant exhibited that all families exceeded significantly the bulk sample and mid parents. Moreover, two and seven families surpassed in high significant and significant differences the better parent. The averages of correlated response were 27.04, 29.75 and 13.45% over bulk sample, mid parents and better parent, respectively. The nine families surpassed significantly the better parent accounted correlated response over it in respective rank by 31.96 (no. 536), 26.59 (no. 460), 26.28 (no. 41), 23.70 (no. 28), 20.73 (no. 518), 19.62 (no. 401), 17.49 (no. 296), 13.14 (no. 108) and 12.21% (no. 261). For population II, the ten selected F6 families were exceeded significantly the bulk sample, mid parents and better parent for weight of spikes/plant. The means of correlated response accounted 34.44, 42.12 and 31.70 % over the bulk sample, mid parents and better parent, respectively. The ten selected F6 families surpassed the better parent in correlated response for weight of spikes/plant with respective rank of 57.82 (no. 249), 49.90 (1), 34.85 (no. 289), 34.08 (no. 192), 32.82 (no. 70), 27.52 (no. 226), 24.61 (no. 59), 22.38 (no. 236), 17.72 (no. 154) and 15.19% (no. 279).
13- The results revealed that the mean of family score for selection response (FSSR) over all selected F6 family of Model 1 recorded the highest value (34.57 and 36.19%), followed by Model 2 (31.50 and 33.12%) and Model 7 (29.52 and 31.12%) in population I and II, respectively. These results may be due to that those models were including traits possessed high selection response such as grain yield/plant, biological yield/plant and weight of spikes/plant. Moreover, means of FSSR over all selected F6 families in relative to bulk sample were high in rank for model 4 (30.11) and model 6 (35.19%); followed by model 3 (28.01) and model 3 (34.91%); and model 2 (27.89) and model 8 (34.57%) for population I and II, respectively. These models had traits accounted high selection response such as grain yield/plant, biological yield, number and weight of spikes/plant.
14- The respective rank for FSSR in comparing to the better parent exhibited that the model 3 (16.79) and model 6 (28.51%) were the first order, followed by model 8 (15.68) and model 7 (28.38%) in the second rank and then model 6 (15.56) and model 8 (28.35%) in population I and II, respectively. Same previous traits in different combinations were belonged to these models. These results revealed that the genes controlled these traits expressed high genetic direct (GY/P) and indirect (BY/P, NS/P and WS/P) responses to pedigree selection in wheat.
15- The FSSR for individual selected F6 families scaled that the family no. 536 in population I ranked to be the highest order in four models (nos. 2, 4, 5 and 7) and second order in four models (nos. 1, 3, 6 and 8) as a mean and also in relative to bulk sample, better parent and mid parents. Its FSSR ranged from 21.65 (mod. 1) to 48.32% (mod. 4), from 15.31 (mod. 1) to 31.30% (mod. 7) and from 22.79 (mod. 1) to 47.28% (mod. 7) in relative to bulk sample, better parent and mid parent, respectively. Also, the family no. 41 in population I arranged to be the first order in four models (nos. 1, 3, 6 and 8), second order in two models (nos. 2 and 7) and the third one in two models (nos. 4 and 5). Its FSSR varied from 23.79 (mod. 1) to 51.31% (mod. 3), from 16.00 (mod. 5) to 39.20% (mod. 3) and from 25.01 (mod. 1) to 49.30% (mod. 3) comparing to the bulk sample, better parent and mid parents, respectively. It is remark result that both families nos. 41 and 536 were in successes order through the first and second order. Also, the first rank for family no. 536 was correlated with the models having GW/P, BY/P and WS/P and the second order connected with models which had NS/P and vice versa for family no. 41.
16- Also, in population II, the FSSR for individual selected F6 families (cycle two) graded that the family no. 249 ranked to be in the first order in all models for mean of FSSR and relative to the bulk, better parent and mid parents. Its FSSR ranged from 30.43 (mod. 1) to 59.20% (mod. 7), from 27.03 (mod. 1) to 52.05% (mod. 7) and from 32.56 (mod. 1) to 65.26% (mod. 4) comparing to unselected bulk, better parent and mid parents, respectively. Moreover, the family no. 1 exerted to be in the second order for FSSR in five models (nos. 1, 2, 4, 5 and 7) and third order in two models (nos. 6 and 8). Its FSSR varied from 20.68 (mod. 1) to 50.66% (mod. 7), from 17.66 (mod. 1) to 43.91% (mod. 7) and from 22.82 (mod. 1) to 56.21% (mod. 7) in relative to bulk sample, better parent and mid parents, respectively. Also, family no. 192 ranked in second order in two models (nos. 3 and 8) and third order in two models (nos. 2 and 7). In general, it is clear that the highest response was exerted from model 7 which includes grain yield/plant, biological yield/plant and weight of spikes/plant.
17- The superior families of 41 and 536 in population I and 1 and 249 in population II and others in both populations appear to be in grate order evaluation as a new genotypes exerted from this study.
III. Phenotypic correlation in one (F5) and two (F6) cycles of pedigree selection for grain yield/plant.
1- Grain yield/plant had positive and high phenotypic correlation with each of number of spikes/plant (0.582 and 0.305), biological yield/ plant (0.807 and 0.708) and weight of spikes/ plant (0.918 and 0.884) in one and two cycles of selection for grain yield/plant in population I, respectively. It is clear that the greatest correlation occurred between grain yield/plant and weight of spikes/plant in one and two cycles of selection.
2- Weight of spikes/plant was correlated in high and positive values with each of spike length (0.539 and 0.650), number of spikes/plant (0.752 and 0.345) and biological yield/plant (0.953 and 0.866) in one and two cycles of selection for grain yield/plant in population I, respectively. Moreover, high and positive phenotypic correlation coefficients were recorded between biological yield/plant and each of spike length (0.636 and 0.640), number of spikes/plant (0.832 and 0.527), and number of spikelets/spike (0.390 and 0.541) in one and two cycles of selection for grain yield/plant. Also, the phenotypic correlation between harvest index and threshing index (0.881 and 0.680) was in same picture. Also, the values between number of spikelets and spike length were 0.651 and 0.545in respective cycles. It is clear that the obtained values of correlation in one and two cycles of selection had the same trend as in base population with a little decrease in values with the cycles of selection. The same view of correlation results could be found in population II. These results could be expressed as change in genetic makeup from generation to other.
3- Consequently, the results of correlation indicate that the most effective components in grain yield of wheat would be number of spikes/plant, biological yield and weight of spikes/plant in major issue and also spike length and number of spikelets/spike in minor role. It is concluded that these traits could be selected for wheat improvement and would be beneficial for the grain yield.
IV. Path analysis in base, one (F5) and two (F6) cycles of pedigree selection for grain yield/plant
1- The partitioning of phenotypic correlation into direct and indirect effects by path analysis revealed that the highest direct effect on grain yield/plant was exerted by weight of spikes/plant in base population, cycle one and cycle two of pedigree selection for grain yield in both studied populations. The values of these direct effects accounted 1.5185, 1.5870 and 1.1181 in population I and 0.9460, 0.7508 and 1.4333 in population II relative to base population (F4), cycle one (F5) and cycle two (F6) of pedigree selection for grain yield/plant, respectively.
2- Moreover, the highest indirect effects were correlated also with the weight of spikes/plant across the base population, cycle one and cycle two of selection in both studied populations. The estimates of these indirect effects were larger via biological yield/plant than number of spikes/plant across all generations in both populations. Their values were 1.4700, 1.5124 and 0.9683 via biological yield/plant followed by 1.404, 1.1934 and 0.3857 via number of spikes/plant in population I and 0.9223, 0.7088 and 1.3831 via biological yield/plant followed by 0.7748, 0.6254 and 0.5547 via number of spikes/plant in population II across their base population, cycle one and cycle two of pedigree selection for grain yield/ plant.
3- These results provided that the weight of spikes/plant has exhibited to be powerful trait as a yield component and must be given preference in selection to superior genotypes of wheat.
4- It is remark and clear conclusion that the path coefficient analysis revealed that about 95, 89 and 80% of phenotypic variance (1-R2) in grain yield/plant could be explained by the selected traits of path analysis in base population, cycle one and cycle two of selection for grain yield/plant in population I, respectively. Also, about 97, 93 and 87% of phenotypic variance in grain yield/plant could be exerted by these selected traits for path analysis in base, cycle one and two of selection for grain yield/plant in population II, respectively. Meanwhile, most of these phenotypic direct and indirect effects were explained by weight of spikes/plant as revealed by path analysis.
5- Also, there are two remarks showed be taken an interest issue i.e. a) the direct and indirect effects of weight of spikes/plant were decreased from base population to cycle two of selection in both populations. This result was coupled with b) decreasing of variances (1-R2) exhibited from the path analysis. Otherwise, the residual factors were increased from base population to cycle one and cycle two of selection. The values were 0.2327, 0.3245 and 0.4485 in population I and 0.1851, 0.2707 and 0.3792 in population II across the base, cycle one and cycle two, respectively.
6- These results are indicating that the genes controlling the weight of spikes/plant and other traits in the current path analysis tend to the maximum genetic expression. Consequently, the selection should be directed to other traits in next generations.
V. Simple, multiple and stepwise multiple regression analyses
* Simple regression
1- Simple regression analysis revealed that the weight of spikes/plant (WS/P) (model no. 6) was superior to other traits and its relative contributions in grain yield were 0.921, 0.843 and 0.782 in population I and 0.956, 0.922 and 0.840 in population II for base (F4), cycle one (F5) and cycle two (F6) of pedigree selection for grain yield/plant, respectively.
2- Biological yield/plant (BY/P) (model no. 5) and number of spikes/plant (NS/P) (model no. 3) were in the second and third order, respectively, after WS/P for their contributions in grain yield of both populations.
3- Also, it is clear results that the value of each contribution of the three traits i.e. WS/P, BY/P and NS/P was decreased from the base population to cycle one and cycle two of selection in remark trend for both populations. These results are in line with correlation values, direct and indirect effects as revealed by path analysis of those traits on grain yield/plant, also, in both populations.
4- The obtained results indicated that the genetic expression of these traits closed to the maximum values and the next cycles of selection should be directed to different traits as a tandem selection.
5- Spike length in both populations and plant height and number of spikelets/spike in population II exhibited to may be having an importance in the next cycles of selection due to their contributions in grain yield increased from base population to cycle one and cycle two of selection. For example, the spike length which contributed by 0.036 and 0.067 in base population, increased to 0.123 and 0.108 in cycle two in population I and II, respectively. Also, number of spikelets/spike recorded value of 0.100 in base increased to 0.299 in cycle two of selection in population II.
* Multiple regressions
6- Multiple regression analysis revealed that, in population I, the highest contribution in grain yield as more than 98% in base, 97% in cycle one and 99% in cycle two of selection were obtained via seven models nos. 9, 10, 13, 14, 17, 19 and 20. The model no. 9 included only two traits i.e. WS/P and TI. Meanwhile, those two traits were the main predominant elements in previous seven models.
7- Also, in population II, ten models nos. 9, 10, 11, 13, 14, 16, 17, 18, 19 and 20 had contributed in grain yield by 99% or more in base, cycle one and cycles two of selection. All these models included mainly WS/P and TI as revealed in seven models as in pop. I and the other three models namely 11, 16 and 18, the HI was found instead of TI.
8- Consequently, the WS/P and TI had justified being more harmonized with grain yield variations. It is remark results that the model no. 9 which included only the WS/P and TI in both populations is superior to other combinations and preferred by wheat breeders.
9- It could be concluded that the coefficient of regression for weight of spikes/plant (WS/P) showed decreased linear relationship from base population to cycle two of selection in both of simple and multiple regression analyses, except models no. 19 and 20 in both populations and models nos. 15, 16, and 18 in population II, revealing that the other traits in multiple regression analysis could be change the trend of the coefficient of regression for WS/P according the relationship with these traits across the different generations via cycles of selection.
* Stepwise regression
10- The stepwise regression analysis revealed to three fitted models for each of base and cycle one as well as two superior models for cycle two of selection in population I. Meanwhile, two fitted models were exhibited for each of base and cycle one and four efficient models released from cycle two of selection in population II. The model no. 1 in all cases included only the WS/P in both populations. It is remark results that the relative contribution of WS/P in grain yield/plant was decreased from 0.921, to 0.843 and to 0.782 and from 0.965, to 0.922 and to 0.840 for base, cycle one and cycle two of selection in population I and II, respectively.
11- Moreover, the model no. 2 in all cases, also, included two traits of WS/P and TI in both populations. It is remark results that the insertion of TI with WS/P in this model (no. 2) were increased the