The aim of this study was to examine the wood-acetic anhydride reaction by studying the reaction with wood substrates such as wood blocks, ground wood and fibre. Isolated wood polymers such as cellulose, holocellulose, hemicelluloses and lignin, were also examined. One lignin model compound was investigated in order to examine the relative rates of reaction of the different hydroxyl groups. The principal focus of the work was, however, on the acetylation reaction and kinetics of wood blocks. The reaction between acetic anhydride and solid wood blocks under uncatalysed conditions was found to be diffusion controlled, rather than activation controlled. The rate-determining diffusion was found to be that of the reagent moving through the wood cell wall. Diffusion was also significant in influencing the rate of reaction in the ground wood sample (0.250 -0.425 mm). It was not possible to form a model for the reaction of solid wood based on its chemical composition. However, it was possible to mathematically model the wood block (and ground wood) reaction based on reaction time and temperature. A relatively low activation energy (Ea) was obtained for radiata pine wood blocks (34 kJ/mol). The equivalent value for ground wood was higher at around 53-62 kJ/mol. The reaction level was lower for the ground wood compared to the wood blocks. It was not possible to obtain reliable kinetic data for the cellulose or holocellulose reactions. Three ground wood samples were partially delignified to examine the effect of a decreasing lignin content (in situ). However, this approach was not very successful. This may be due to the increasing acid soluble lignin content with decreasing lignin content, due to the chlorite delignification process used, or perhaps the removal of lignin created more micropores which aided diffusion. The reaction of a commercial alkali lignin (AL) from mixed softwoods gave surprisingly similar results to the milled wood lignin (MWL) isolated from radiata pine. Although the lignin reaction (29% of the MWL reaction) was similar in shape to that of the wood block reaction profile, the isolated lignin reaction was not enough to explain or model the wood reaction. This may have been because of the diffusion control of the latter and the reaction over longer reaction times of the hemicelluloses. The ~-0-4 lignin model compound studied showed that the relative rates of acetylation of the hydroxyl groups were: the phenolic hydroxyl reacted more rapidly than the primary hydroxyl, and both reacted a great deal faster than the secondary hydroxyl. The results were similar to those found for four milled wood lignin (MWL) acetylation products : the phenolic and the primary hydroxyl reacted at a similar rate, and that both reacted more rapidly than the secondary hydroxyl. Overall, this study demonstrated the importance of the wood ultra-structure in determining the rate of reaction of acetylation in wood and wood-based substrates. The isolated wood components were therefore found to be of less importance due to the diffusion control of the reaction.