all rights reserved the properties and chemistry of wood are increasingly being measured by coupling near infrared (nIr) spectroscopy with multivariate analysis to build predictive models based on analytical data for specifically targeted properties and/or chemical compositions. advantages of the technique are that it is non-destructive, relatively inexpensive and can be rapidly applied after models have been developed. common physical (for example, density, microfibril angle) and mechanical (for example, stiffness) properties have been predicted for both hardwoods and softwoods. chemical composition of wood cell wall polymers (cellulose, hemicellulose, lignin) and extractives have also been predicted by using models with high correlations. 1 Spectroscopic analyses have been carried out by diffuse reflectance 2-5 and transmittance, 6,7 the latter technique requiring much less material for analysis. physical properties and chemical composition play an integral role in the performance of wood in service, so models have been developed to predict mass losses and/or changes in composition as a result of thermal treatment 8 or exposure to wood decay fungi. 9-11 recently, increasing interest in biofuels raises questions about the utility of nIr spectroscopy for rapidly assessing feedstock qualities, in particular, fuel value. the higher heating value (HHV) or gross calorific value (GcV) of a fuel is defined as the amount of heat released by a specified quantity (initially at 25°c) once it is combusted and the products have returned to a temperature of 25°c, which takes a School of renewable natural resources, lSu agricultural center, Baton rouge, louisiana, uSa. e-mail: cso@fs.fed.us b uSda-forest Service, Southern research Station, pineville, louisiana, uSa Gross calorific value (GCV) has been predicted by building models based on near infrared (NIR) spectroscopy and multivariate analysis; however, to date, the impact of feedstock chemical composition on the models has not been directly assessed. In the present study, 20 longleaf pine trees were sampled at two positions (breast and mid-height) for calorimetric and spectroscopic analyses. The GCVs, which ranged from 20 MJ kg −1 to 24 MJ kg −1 , showed a strong correlation with the wide-ranging values of acetone-soluble extractives content.After extraction of the samples with acetone, the range for the GCV was both lower and slightly narrower (19-21 MJ kg −1 ) and was poorly correlated to lignin content with its narrow range of values. Near infrared (NIR) spectroscopy coupled with multivariate analysis was applied to the samples and provided a strong coefficient of determination (R 2 ) between the values predicted by NIR and those determined by calorimetry for the unextracted, but not the extracted, samples. Plotting the regression coefficients validated the results by showing very similar plots for GCV and extractives content, thereby indicating that the same molecular features are driving the models.NIR spectroscopy coupled with multivariate analysis can predict GCV fo...