Thesis (Msc Food Sc (Food Science))--University of Stellenbosch, 2009. / Triticale (X Triticosecale Whittmack), a cross between durum wheat (Triticum sp.) and rye, is a crop with an increasing agronomic and economic potential Though studies on the functional and compositional quality of triticale have been conducted in other parts of the world, little is known regarding cultivars developed in South Africa in terms of these aspects. South African triticale cultivars from various localities in the Western Cape, obtained for two subsequent harvest seasons, were analysed for moisture, protein and ash contents, as well as falling number (an indication of α-amylase activity), hardness (particle size index), 1000-kernel mass and baking potential (SDS sedimentation). These triticale samples were derived from a breeding program that was not focused on baking quality. The results obtained were found to compare well with those reported on in previous studies. Significant differences were observed between both cultivars and localities within years, illustrating the effect of genetic as well as environmental factors. Significant differences were also observed between localities when comparing the two harvest seasons, whereas differences between the cultivars for the two seasons were in most cases not significant; illustrating the effect of environment. Interactions between cultivars and localities were found to be significant for all parameters, and trends were observed between protein content and both particle size index (PSI) (negative) as well as SDS sedimentation (positive) results for both years. Near infrared (NIR) spectroscopy is a rapid method of analysis and is widely used for the quality evaluation of wheat. Limited research has been reported on calibration models for the quality evaluation of triticale, and thus NIR spectroscopy was applied to develop models for the prediction of moisture, protein and ash contents, as well as hardness and baking potential for South African cultivars. Spectra were collected in diffuse reflectance mode and partial least squares (PLS) models developed for both triticale flour and wholegrain using two different instruments (Büchi NIRFlex N-500 and Bruker MPA Fourier transform NIR spectrophometers) and software packages (The Unscrambler and OPUS). Full cross-validations were performed, after which the best prediction models obtained (R2 > 0.66) were validated using an independent test set (n = 50). The best prediction results were obtained with flour for moisture (Bruker: SEP = 0.08%; R2 = 0.95; RPD = 4.65) and protein (Büchi: SEP = 0.44%; R2 = 0.96; RPD = 5.23 and Bruker: SEP = 0.32%; R2 = 0.96; RPD = 4.88). For whole grain, acceptable results were obtained for protein (Büchi: SEP = 0.55%; R2 = 0.94; RPD = 4.18 and Bruker: SEP = 0.70%; R2 = 0.90; RPD = 3.23). Though results for ash content, PSI and SDS sedimentation prediction did not yield models that can be applied as yet, these models form a good basis for further calibration model development and possibly use in early generation screening. The current limited ranges could be expanded by adding samples from subsequent harvest seasons. By adding more data, a better quality profile for South African triticale can be obtained, which will facilitate better interpretation in terms of the effect of genetic and environmental factors. It would also enable the development of improved NIR prediction models.