Acacia mearnsii (black wattle) is an important South African commercial forestry species, providing a source of high quality raw material (fibre) for both the domestic and international pulp and paper industries. Compared with many Pinus and Eucalyptus species, there has been very little research into the wood and pulping properties of black wattle. The ability to assess pulp yield in a non-destructive manner, using near infrared (NIR) spectroscopy, is vital from a tree improvement perspective. Destructive sampling and analysis, results in the loss of the genotype, while also being very expensive and time consuming. In order to assess some of the important characteristics that make the species desirable from a fibre perspective, this study investigates growth characteristics, wood density and pulp yield of ten trees grown on each of three different sites namely, Bloemendal, Glen Echo and Phoenix, and from each of three different age classes being 7- , 9- and 11-years-old. In total, 90 trees were sampled for this stage of the study. In general, physical characteristics such as utilisable height and diameter at breast height of the trees differed between sites and increased with age, this age effect trend was not reflected in the pulp yield or wood density results. Pulp yield measurements ranged from 52.61 to 59.91% across all sites and age classes, which, when compared to the pulp yield from many other forestry species, is relatively high. Laboratory pulp yield data was used in conjunction with NIR spectra obtained from the same wood samples to calibrate a NIR spectrophotometer to predict pulp yield. Thirty 11- year-old trees were then chosen from the Bloemendal site and sampled extensively to investigate the within-tree variation in pulp yield. The NIR model developed was used to measure the pulp yield from the numerous samples taken from within the trees. In general, pulp yield decreased from pith to bark and from the base of the tree to about 20% of the tree height and then decreased towards the top of the tree. The within-tree variation data for pulp yield was analysed to identify the best position for non-destructive sampling, and a model was then developed to predict whole tree pulp yield based on this sample, which was taken at 1.4 m up the tree. The analyses of small samples of wood meal, using near infrared spectroscopy, enabled the prediction of whole tree pulp yield. / Thesis (M.Sc.) - University of KwaZulu-Natal, Pietermaritzburg, 2009.