The primary objective of this study was to find out if changes in chemistry could be used to quantify Pinus radiata compression wood severity or degree of compression wood development. Basic chemical composition and the lignin structure was assessed for a range of different compression wood samples sourced from juvenile wood, mature wood, earlywood, latewood, branches, knots, 2-year and 1-year old Pinus radiata. Fluorescence microscopy was used as the reference method to assess the degree of compression wood development. Lignin structure of compression wood was studied by thioacidolysis, size exclusion chromatography, and thioacidolysis/31P NMR spectroscopy. Variation in the basic chemical composition and lignin structure with compression wood severity was ascertained. Results showed that, as the severity of compression wood changed, progressively from normal through mild to severe, all chemical parameters commonly associated with compression wood changed concurrently. With increasing severity lignin and galactose levels increased while glucose and mannose levels decreased. Lignin structural changes were also associated with changing severity of compression wood. Levels of p-hydroxyphenyl (H) releasable β-ethers increased and guaiacyl (G) releasable β-ethers decreased. Similarly, levels of uncondensed p-hydroxyphenyl units increased, while uncondensed guaiacyl units decreased. Similar proportions of condensed guaiacyl units were present in compression wood and normal wood. Similar trends in chemical composition were observed between the compression wood and related opposite wood in branches, knots and young wood of Pinus radiata. A number of chemical parameters changed linearly with compression wood severity. They were: the amount of lignin and galactose, the galactose/glucose ratio and p-hydroxyphenyl content in lignin. Parameters based on the p-hydroxyphenyl unit content in lignin, the H/G releasable β-ether ratio, releasable p-hydroxyphenyl β-ether units and uncondensed p-hydroxyphenyl C9 units are most suitable indicators of compression wood severity as they spanned a larger range relative to the normal wood levels and were not influenced by the morphological origin of wood samples. Chemical methods for quantifying compression wood severity should focus on the detection and measurement of these parameters. Galactan present in Pinus radiata compression wood was isolated and characterised. Structural investigation by methylation analysis and NMR spectroscopy revealed that this galactan was largely composed of (1→4)-linked β-D-galactopyranose residues. No evidence was found to indicate the presence of any branches. Characterisation of lignin in cell wall fractions of Pinus radiata normal wood revealed that middle lamella lignin has a higher lignin content, a lower amount of releasable β-ethers and a more condensed lignin than the secondary wall lignin. Levels of releasable p-hydroxyphenyl units were not higher in middle lamella lignin. A new method based on thioacidolysis and 31P quantitative NMR spectroscopy for estimation of the degree of lignin condensation of the phenolic and etherified C9 units in in situ wood lignin is described. Using this method it was found that phenolic C9 units in in situ lignin were considerably less condensed than etherified C9 units in both compression wood and normal wood.