This thesis describes a new analytical method to determine the levels of uncondensed (protonated C5) and condensed (bonded through C5 to other phenyl propane units) guaiacyl (G) units in lignin. This method is important because the proportion of condensed and uncondensed units in lignin determines, in part, how the lignin will behave during mechanical and chemical processing. Currently there are few analytical techniques that can simultaneously determine the proportions of condensed and uncondensed units in isolated lignins and even fewer that can elucidate this information for lignin in situ in wood. The method combines two well-accepted techniques, thioacidolysis and quantitative ³¹P NMR spectroscopy. Thioacidolysis involves depolymerisation of lignin with ethanethiol and boron trifluoride at elevated temperatures. This leads to an organic-solvent soluble product in which almost all the phenylpropane units are phenolic. Subsequent quantitative ³¹P NMR spectroscopic analysis, of the thioacidolysis product derivatised with 2-chloro-4,4,5,5-teramethyl-1,3,2-dioxaphospholane gave the total amount of condensed and uncondensed units in the lignin. The method proved highly successful at determining the amounts of uncondensed G, condensed G, syringyl (S) and p-hydroxyphenyl (H) units in milled wood lignin (MWL) and in situ lignins from softwoods, hardwoods, softwood compression woods and softwood pulp fibres. Also, for MWL samples the method proved useful for determining the amounts of uncondensed G, condensed G, S and H units present as free phenolic and etherified moieties. A further benefit of performing the method was that it allowed the levels of individual condensed units, such as β-5, 5-5 and 4-O-5, in MWL and in situ lignin to be determined. Initial investigations with model compounds found that quantitative ³¹P NMR spectra were collected even when thioacidolysis was performed prior to ³¹P NMR spectroscopy. This was significant, as prior to this work we had been concerned that thioethyl incorporation into sterically crowded 5-5 and β-5 structures may have hindered quantitative analysis. Dibenzodioxocin structures have recently been reported as important units in lignin polymer crosslinking, therefore the thioacidolysis of a dibenzodioxocin model compound was studied. The dibenzodioxocin reacted primarily via an acid catalysed ring rearrangement to yield a dibenzooxepine. The incomplete cleavage of dibenzodioxocin ring during thioacidolysis suggested that the method might be under reporting the levels 5-5 units. Thioacidolysis/³¹P NMR spectroscopy was initially applied to MWL and in situ lignin from radiata pine wood. In both samples around 35% of the phenylpropane units were condensed. As MWL originates primarily from the secondary wall (SW), this finding suggests that the SW and middle lamella (ML) lignin in radiata pine are more similar than previously believed. Thioacidolysis/³¹P NMR spectroscopy was also applied to lignin in situ in wood from slabwood, corewood, earlywood and latewood regions of a mature radiata pine. In all samples around 35% of the phenylpropane units in the total lignin were condensed. While the concentration of lignin in these different regions of the tree has been shown to vary markedly, these results suggest that the proportion of condensed units remains the same. Thioacidolysis/³¹P NMR spectroscopy was next applied to MWL and in situ lignin samples from a hardwood and a softwood compression wood. For eucalypt MWL, 37% of the free phenolic groups were S units but 73% of the phenolic groups released by thioacidolysis were S units. On the other hand for compression wood around 90% of the H units were found to be present as free phenolic groups, with few H units released during thioacidolysis. These results can be seen as further confirmation of the preference of S units to be present as etherified moieties and H units as free phenolic moieties in lignin. The method was next applied to lignins in in situ thermomechanical and chemithermomechanical pulp fibres. The lignin in both thermomechanical pulping (TMP) and medium density fibreboard (MDF) fibres contained around 35% condensed units. This indicated that mechanical refining at elevated temperatures did not significantly change the proportions of condensed units in the lignin. However, for thermomechanical refining in the presence of added sulphite, results showed that around 38% of the C9 units in lignin were condensed moieties. This increased proportion of condensed units relative to radiata slabwood, was due to a decrease in the amount of uncondensed units, rather than an increase in the amount of condensed units. This lower yield of uncondensed units was attributed to α-sulphonation of the lignin during the pulping process, which may block the degradation of lignin by thioacidolysis. This effect was highlighted in the heavily sulphonated Ontario Paper Co. (OPCO) pulp, where a very low overall yield was observed. Finally thioacidolysis/³¹P NMR spectroscopy was applied to a number of different kraft spent liquor lignins (KSLLѕ). KSLLѕ were found to contain up to 45% condensed G units. This increased proportion of condensed structures, relative to in situ pine lignin was consistent with numerous literature reports. However, thioacidolysis/³¹P NMR spectroscopy was found to be less applicable to these heavily modified lignins, as they were much less comprehensively degraded on thioacidolysis than wood lignins.