Marie F. Smith

Multivariate assessment of environmental effects on hard red winter wheat. I. Principal-components analysis of yield and bread-making characteristics

The attributes of yield and bread-making quality of winter wheats produced in the Free State are subject to environmental effects. The result is a supply to the baking industry of flour of inconsistent quality, which in turn is detrimental to the market value of wheat. Various genetic and environmental factors may affect bread-making quality. In the first part of this study we demonstrated the magnitude of environmental and genetic factors, and used principal-components analysis to examine the underlying source of variability in the bread-making quality of winter wheat cultivars. Yield and protein concentration responded to environmental influences, whereas the dough-mixing requirement was primarily inherently determined. Variation in protein concentration accounted for variation in the mixing requirement of one cultivar and the SDS sedimentation volume of most cultivars.

A single point sorption test for the routine determination of the phosphorus requirement of low to moderate P-fixing soils

The phosphorus requirement (P0.11) of low to moderate P-fixing soils from the South African tobacco industry and KwaZulu-Natal was interpolated from multiple point P sorption isotherms. The relationship between P0.11 and a single point sorption measurement, consisting of the P concentration left in solution after equilibrating for 24 h additional rates of either 70 or 100 mg P kg−1 soil, over the soils of the tobacco industry was investigated using non-linear regression analysis. The results showed that a high degree of correlation existed between both variables, suggesting that P0.11 could simply and accurately be predicted from a single P sorption measurement instead of a multiple point sorption isotherm. Confirmation of the predictive value of the functions best fitting the data from the soils of the tobacco industry was validated using data of the soils from KwaZulu-Natal. The results showed that values of P0.11 predicted for the soils of KwaZulu-Natal, using the single point P sorption test in conjunction with the regression functions developed for the soils of the tobacco industry, were highly significantly correlated to the P0.11 values interpolated from multiple point P sorption isotherms. The implication of these findings for the routine determination of the phosphorus requirement of soils by soil testing laboratories and P fertiliser recommendations is discussed.

A quality assurance strategy for wheat (Triticum aestivum L.) where growth environment predominates

The variation in the processing quality of South African winter wheat cultivars over growth environments is detrimental to its market value. Variation in protein content has previously been identified as a primary contributor to this variation and should therefore be considered in the interpretation of quality data presented during the cultivar release procedure. The aim of the study was to analyse such effects on the interpretation of quality data of advanced wheat lines. Our results indicated a distinctive segregation between entries having higher (≥10%) and lower (<10%) flour protein contents for the three critical phases of commercial processing. The incorporation of a range of localities where both protein levels are sufficiently represented, appeared to be of more importance than the number of harvest seasons over which the entry is evaluated. Finally, multivariate statistical techniques were used to compute user-friendly models for the objective characterisation of variation in milling and rheologi...

Phosphorous sorption study of selected South African soils

Phosphorous isotherms for 21 selected soils from the Republic of South Africa were constructed and the Langmuir, Freundlich and Temkin sorption models were compared using regression analysis. Over a range of solution concentrations of 0.025 - 2.00 mg P I−1, all three sorption equations were found to fit the data well (R2 = 0.934 - 0.996). The values of the coefficients of the different equations were more accurately estimated by the Freundlich and Temkin models, whilst the Langmuir model was a slightly better indicator of adsorbed P. A multiple regression analysis between sorption coefficients and selected soil properties showed citrate bicarbonate dithionite-AI to be an important factor in P sorption. Other soil constituents important in P sorption included clay percentage, organic matter, citrate bicarbonate dithionite-Fe and Bray 2 P content. Although useful in identifying soil properties responsible for P sorption, the regression equations were too cumbersome and imprecise to be of any practical use in characterising P isotherms routinely. It was also apparent from the regression analysis that, except for the Langmuir maximum and Temkin P buffer power index, the other sorption coefficients require the inclusion of a measure of native P content to be true soil characteristics, when constructing isotherms. Limitations, which this constraint imposes on the routine use of isotherms for the purpose of determining sorption coefficients, are discussed.

Two-step approach to determining some useful phosphorus characteristics of South African soils: a review of work done at the ARC-Institute for Soil, Climate and Water

Work carried out at the ARC-Institute for Soil, Climate and Water on soil phosphorus characteristics was reviewed. Regression functions were established in order to test the correlations between P parameters, single point indices and soil properties. The parameters investigated were selected indices of P sorption, including the Langmuir adsorption maximum (Lm), as well as indices of the P buffer power of the soil, including the slope of each of the Langmuir (Lbp), Freundlich (Fbp) and Temkin (Bt) P sorption isotherms. Other parameters included an index of the P requirement of the soil (P0.11), the phosphorus requirement factor in terms of the Ambic 2 extraction procedure (PRFAmbic2) and an index of the slow reaction of P sorption with soils (dx/dt)24. The results showed that to obtain a measurement of these parameters, cumbersome multiple point procedures could mostly be substituted by single point indices more suited for routine determination purposes. Lm, Lbp, Fbp and Bt could readily be estimated from measuring the sorption index (SI) as determined by the National Soil Survey in modal soil profiles. P0.11 could be estimated by combining, in multiple regression equations, SI with a measure of the native P content in terms of either the Bray 2 or Ambic 1 extraction procedures. Alternatively P0.11 could be estimated from measures of either S1 or S2, the concentration of P in soil solution remaining after incubating the soil for 24 h with additional rates of 70 or 100 mg P kg−1 soil respectively. PRFAmbic2 was found to relate to the Ambic 2 phosphorus desorption index. The rate of the slow reaction, (dx/dt)24, was found to relate to both the clay and the citrate bicarbonate dithionite-Fe contents of the soil. The implication of these findings for the routine determination of soil phosphorus characteristics by soil testing laboratories involved in crop nutrition or environmental advisory work is discussed.

Determination of the preharvest sprouting resistance of South African winter wheat (Triticum aestivum L.) cultivars

Preharvest sprouting significantly reduces the quality of the South African wheat crop. This study classified the preharvest sprouting resistance of 17 South African winter wheat cultivars sampled over a range of typical winter wheat growth environments. Sampling was performed under field conditions and sprouting responses were evaluated under controlled conditions. Variation in sprouting response between cultivars was predominantly genetically determined and varied from 1.7 to 7.4 on a scale from 1 (no visual sprouting) to 8 (fully sprouted). Canonical variate analysis and AMMI analysis, as well as the use of hierarchical clustering of cultivars over environments using the AMMI estimates, identified four distinct groups ranging from resistant to susceptible. The AMMI model was used as it combines the additive main effects of the analysis of variance with the interaction effects of principal components analysis. In general, Betta, and certain cultivars derived from Betta, showed good to moderate resistanc...

AMMI analysis of potato cultivar yield trials

Potatoes, Solanum tuberosum (L), are cultivated in South Africa under a variety of different climates and seasons. The wide environmental variation encountered and the confirmed occurrence of genotype-environmental interaction necessitates the use of criteria, other than averages, in the evaluation and selection of new potato cultivars. Yield data from cultivar trials countrywide, were used to evaluate the use of the AMMI model against the existing Finlay-Wilkinson regression technique. The AMMI model was found to be suitable for determining the reaction of cultivars/lines in an environment. The usefulness of the technique improves with a graphical presentation of the results to display the stability of a genotype in different environments.

The success of the AMMI model in predicting lucerne yields for cultivars with differing dormancy characteristics

The AMMI model, which incorporates both additive and multiplicative components into an integrated, powerful, least squares analysis, is used for the analysis of cultivar trials at different localities. These trials are a major experimental component in crop production and plant breeding research. Multilocation trials have as objectives the assessment of the success of yield predictions, the grouping of sites for evaluation and the interpretation of genotype-environment interactions. The data used originated from the National Lucerne Evaluation Program and the lucerne cultivars were representative of dormancy groups 4–9, i.e. semi-, intermediate-, non- and highly non-dormant. Expected trends in results from the yield trials were not obvious using the additive main effects model (ANOVA), while the application of the AMMI model resulted in rankings of cultivars in different environments which could readily be explained by the breeding history and dormancy of the cultivars.

An improved phosphorus desorption index for the routine determination of the Ambic 2 phosphorus requirement factor of South African soils

A short (24 h) laboratory incubation study was conducted and plots of P desorbed in Ambic 2 extraction solutions versus added P were constructed for a wide range of soils of varying P fixing capacity representative of the province of KwaZulu-Natal, South Africa. The plots were found to yield straight lines and, using linear regression analysis, equations of best fit describing these desorption functions were established for each soil. A two point test, the Ambic 2 Phosphorus Desorption Index (PDIAmbic2), was found to correlate closely with the slopes of the desorption functions (R2 = 0.943). This result showed the proposed quick two point PDIAmbic2 test to be a reliable index of the slope of the multiple point desorption functions. It established that, although simpler, it could be used advantageously as a time saver to estimate the slopes of the desorption functions routinely. PDIAmbic2 was further shown to correlate significantly to the Ambic 2 Phosphorus Requirement Factor (PRFAmbic2) obtained from a previous long term (six weeks) glasshouse incubation study and designed to simulate the fate of additional P fertiliser under field conditions. PRFAmbic2 is a key factor used to convert soil-P deficits in quantities of additional fertiliser required by crops, and is a useful characteristic for making P fertiliser recommendations, albeit one that is difficult to measure routinely. Non-linear regression analysis was used to establish the function best describing the relationship between PRFAmbic2 and the more user-friendly PDIAmbic2 test. Confirmation of the predictive value of the equation of best fit was validated using data obtained from soils not originally included in the regression. Implications of this test for routine laboratories involved in P fertiliser advisory work are discussed.

The relationship between the phosphorus requirement and some phosphorus characteristics of selected soils of the South African tobacco industry

Estimates of the phosphorus requirement (P0.11) of 20 selected soils of the South African tobacco industry were interpolated from phosphorus sorption isotherms. The relationships between P0.11 and various soil phosphorus characteristics, including Bray 2 P content and various indices of P-fixation, were investigated using regression analysis. The results showed that P0.11 varied widely and is influenced by both the level of Bray 2 P content and the P-fixation capacity of the soil. Measurements of the P-fixation capacity of the soil included the more formal buffer power indices derived from multiple point P sorption isotherms and both the phosphorus desorption index (PDI) and the sorption point index (SI), each a single point test easily determined routinely. SI was shown to relate closely to the buffer power indices and to be a quick and useful index of the soil P-fixation capacity. PDI, on the other hand, was found to be of limited value to predict the phosphorus fixation of the soils. An equation, allowing for P0.11 to be routinely predicted from Bray 2 P content and SI is proposed. Its predictive value was validated from data obtained from five additional soils that originated from the tobacco industry. The implications of this equation for P-fertiliser recommendations and future field calibration programmes are discussed.