Lignocellulosic biomass is a potential source of fermentable sugars such as glucose. Enzymatic hydrolysis of cellulose is a viable method of solubilizing the glucose from biomass, but the cellulose fraction of native lignocellulosic material is shielded from enzymatic attack by the lignin-hemicellulose matrix surrounding it. Pretreating lignocellulosic biomass with dilute sulfuric acid at high temperatures solubilizes hemicellulose, rendering the cellulose fraction more susceptible to enzymatic hydrolysis. Evaluation of dilute-acid, high-temperature pretreatments depends on polysaccharide analysis of the two fractions resulting from a pretreatment, prehydrolyzed solids(PHS) and prehydrolyzate liquid(PH). The polysaccharide analysis is based on a method described by the National Renewable Energy Laboratory and involves a two-stage sulfuric acid hydrolysis followed by HPLC quantification using ion-moderated partition chromatography and refractive index detection. The subject of this thesis is identifying and quantifying the sources of error associated with the polysaccharide analysis and the error associated with the evaluation of the effects of pretreatment on the polysaccharide fractions of switchgrass and poplar. This was addressed by conducting replicate polysaccharide analyses on single samples of native biomass, PHS, and PH. The variability associated with these measurements was compared to the variability associated with replicate analyses of identically pretreated biomass. It was found that the use of sugar standards to correct for sugar destroyed during the analysis adds error and most likely overestimates the amount of sugar from biomass actually destroyed. It is evident that assuming a volume after neutralization of the hydrolyzed biomass sample is more reproducible than measuring the volume. When using a batch-type reactor and the temperature and acid parameters used in this study,140°C-180°C/ 0.6-1.2 % sulfuric acid (w/w), it is evident that the major source of error in evaluating pretreatment conditions is the pretreatment itself, not the analysis. / Graduation date: 1995