Genetic variability of South African spring wheats for milling and breadmaking quality

English: The main aim of this study was to use a data with regards to milling and baking quality characteristics generated under similar environmental locations and seasons, in order to determine the source, size and magnitude of genotype x environment interaction; the optimal locality, year, and replication combinations; to compare the various available statistical procedures in assessing performance stability and to evaluate the different existing methods available regarding their suitability for grouping different genotypes into homogeneous groups. In order to achieve this, five spring wheat cultivars (Triticum aestivum L.) were grown in five environments under irrigation from 1997 through 1999. The cultivars used in this study displayed a broad range in milling and baking quality. 1. The source, size and magnitude of genotype x environment interaction for spring wheat under irrigation were determined. From the results, the best and poorest cultivar in overall milling performance was Kariega and SST 822 respectively, while the best and poorest overall baking performance were displayed by Kariega and SST 876 respectively. All the characteristics related to gluten strength and the elastic properties namely SDSsedimentation test, wet gluten content standardized at 12% protein, mixograph mixing time, alveograph strength, alveograph distensibility and alveograph PIL ratio displayed genotypic variance contributions above 30% and even as high as 57.65% for mixograph mixing time. Genotype x location and genotype x year interaction components for most of the milling and baking characteristics was insignificant. However, genotype x location x year interaction was significant for most milling and baking characteristics. 2. The optimal locality, year, and replication combinations regarding the screening of spring wheat genotypes for milling and baking quality under irrigation were also determined. From the data it was clear that there is not a large improvement in relative efficiency when a second and third replicate is used in the determination of the optimal amount of localities in order to distinguish between genotypes. A similar pattern is also observed regarding the number of years needed and almost no differences were observed if either three, four or even five year's quality data are taken into account. It has been proven that for both the milling and baking characteristics five trials are needed over a two-year test period. 3. Various available statistical procedures were used in assessing performance stability regarding milling and baking characteristics in order to identify the most suitable method. The following statistical analysis were conducted and procedures followed in order to determine stability for the milling and baking characteristics: (i) Wricke's - Wj (1962) ecovalence, (ii) Eberhart and Russell's - S2dj (1966) deviation parameter, (iii) Shukla's - 0-2j (1972) stability variance, (iv) Linn and Binn's - Pj (1988a) cultivar superiority performance and (v) AMMI stability value - ASV (Purchase, 1997). Lower significant rank correlation coefficients were found in the pairwise comparisons of both Eberhart and Russell's as well as that of Lin and Binn, but in the latter to a lesser extent. From this study, it would appear that all three the following appropriate methods namely Wricke's ecovalence, Shukla's stability variance as well as the AMMI Stability Value can be used to describe the stability of a genotype. Looking at the stability of the cultivars per se it can be concluded that Marico is very unstable for its hectolitre mass (HLM) values. Marico is however very stable for the expression of its inherent kernel hardness values (VK, SKCS-HI and BFLY). Kariega is however very unstable regarding the expression of its inherent kernel hardness value. SST 825 displays unstable hectolitre mass and hardness values. SST 825 has very stable characteristics that influence kernel size. SST 822 expressed very stable flour yield as well as hardness values. SST 876 produced very stable values for hectolitre mass as well as the traits that influence kernel size but it produced an unstable flour yield. Marico is very unstable regarding its rheological as well as baking properties. Kariega on the other hand displays remarkable stable rheological as well as baking performance properties. SST 825 is unstable in the expression of its grain (GPC-Leco) and flour protein (FPC-Leco) levels. SST 825 produced stable alveograph strength (AS) and alveograph distensibility (AD) values. SST 822 however has very stable GPC-Leco as well as FPC-Leco values. Although SST 876 showed very stable GPC-Leco as well as FPC-Leco values, its rheological properties is weak. 4. Different selection methods were used to group different genotypes into homogeneous groups according to their milling and baking characteristics respectively. How effective these methods identify superior genotypes according to their milling and baking characteristics was also determined. The suitability of these methods were evaluated, in order to determine which is the appropriate method for evaluating the milling and baking quality characteristics from multilocation field trials for new testing lines to be considered for release. Therefore, two test lines were also used in order to verify the accuracy of the evaluation methods, which were the combination of the independent culling with absolute limits of acceptability relative to a check method from the SAGL, Lukow's (1991) multiple linear regression analysis method, Howard's (2003) milling and baking worth procedure, canonical variate analysis (CVA) and cluster analysis. It can be concluded that Lukow's method has the tendency to identify lines that has stronger gluten strength characteristics, which makes sense because is was developed for the Canadian wheat market. Howard's milling and baking worth procedure has relative value for ranking cultivars, but the weighted contributions should be changed according to environmental influence on milling and baking characteristics. It was found that the CVA analysis is an objective and efficient procedure to evaluate new breeding lines regarding their milling and baking potential in comparison to a biological quality standard. The cluster analysis gave similar clusters when compared with the groupings of the CVA for baking results but not for the milling characteristics.