The particle size distribution produced from first break roller milling of wheat determines the flows through the rest of the mill and hence the quality of the final flour, and is affected by debranning and by the operation of the roller mill. The Double Normalised Kumaraswamy Breakage function (DNKBF) gives a quantitative basis to describe breakage during first break milling of wheat and to interpret effects. Previous work developed and extended the breakage equation in order to understand and predict wheat breakage based on distributions of the grain characteristics and the operating parameters of the mill. However, broken particles vary in composition as well as size; therefore the primary objective of the current work was to extend the DNKBF during first break milling to include particle composition, using fingerprints of pericarp, aleurone, endosperm and germ. Meanwhile, debranning is a technology that has enhanced flour milling in recent years, leading to improvements in quality that are not well understood but that start with the effect on milling. A second objective of the current work was therefore to apply the DNKBF to describe and interpret the effects of debranning on wheat breakage and, in so doing, to clarify the physical significance of the DNKBF parameters.Samples of Mallacca (hard wheat) and Consort (soft wheat) were debranned for nine different times, at three roll gaps and under S-S and D-D dispositions. The DNKBF successfully described the normalised particle size distribution at different debranning times. The DNKBF describes wheat breakage in terms of Type 1 and Type 2 breakage, where Type 1 describes a relatively narrow distribution of mid-sized particles, whilst Type 2 describes a wide size range of predominantly small particles extending to very large particles. The proportion of Type 1 breakage increased at longer debranning times, while Type 2 breakage decreased, for both wheats under both dispositions. S-S milling tended to produce more Type 1 breakage than D-D. A mechanism of wheat breakage is proposed to explain the co-production of very large and small particles via Type 2 breakage, and hence the effect of debranning. The proposed mechanism is that small particles of endosperm arise from scraping of large flat particles of wheat bran under the differential action of the rolls; removal of the bran reduces the production of the large bran particles and thus reduces the opportunity for the scraping mechanism that produces the very small particles.The composition of broken particles can be characterised considering the four major wheat components, pericarp, aleurone, endosperm and germ. Kernels of Mallacca and Consort wheats were manually dissected to isolate these components. FTIR spectroscopy was able to distinguish the different components in milled fractions. However, attempts to quantify the relative contribution of each wheat component in milled fractions (by measuring specific peak heights and by Partial Least Squares, PLS) were compromised by technical limitations. An alternative approach aimed to fingerprint the components using sugar analysis by HPLC, with some success; however the technique was too complex and limited by the detection limit of HPLC, in particular for arabinose and xylose.Instead, the botanical distributions within eight milled fractions of Mallacca and Consort wheats milled under S-S and D-D dispositions were analyzed by PLS models developed by Barron (2011). The concentration functions were then found by applying the DNKBF to the particle size distributions and to the compositional distributions, the ratio of the DNKBFs giving the concentration function. The DNKBF was able to describe the data well for the four botanical components studied in both wheats: pericarp, aleurone, intermediate layer and starchy endosperm. The analysis clarified the nature of the particles produced on breakage, showing that for Mallacca wheat, the pericarp and aleurone layer compositions mostly varied with particle size in similar ways. Intermediate layer showed broadly similar results to those for pericarp and aleurone in the Mallacca wheat despite being the least accurate component predicted. However, for Consort wheat, the intermediate layer behaved differently from pericarp and aleurone, suggesting a different breakage mechanism, perhaps associated with how the wheat hardness affects breakage of the bran and the production of large flat bran particles. Creation of pericarp/intermediate layer/aleurone dust during milling was notable, in particular for Mallacca wheat. The relative uniformity of the Mallacca compositions in relation to pericarp, intermediate layer and aleurone, which varied in consistent ways with particle size, was also notable. By contrast, for Consort wheat, the relative proportions of these three components appear to vary substantially in particles of different size, pointing to very different breakage origins. It seems that in the hard wheat, the breakage patterns are dominated by the endosperm physical properties, while for the soft wheat, the behaviour of the large bran particles produced is dictated much more by the properties and structure of the bran layers than by the hardness of the endosperm.The approach presented is practical to describe, quantify and interpret the effects of breakage on component distributions, in order to understand the fate of kernel components during milling and hence the origins of flour quality.