For analysis of carbon isotope discrimination in wood, cellulose or holocellulose is often preferred to whole tissue because of the variability in isotopic composition of different wood components and the relative immobility of cellulose. Most currently used methods for the preparation of wood components for stable isotope analysis (e.g., the Jayme-Wise method) produce a residue of holocellulose. The Jayme-Wise method was initially developed to extract holocellulose from small (~1 g) samples of wood, and, despite subsequent modifications, the method requires specialized glassware, considerable time and entails the risk of sample loss. For carbon isotope analysis, we adapted an acid-catalyzed solvolytic method for preparing crude cellulose by treating wood meal with acidified di-glycol methyl ether (diglyme). The one-step process requires no special glassware, is complete within 24 hours and enables over 100 samples to be processed in a day. This method gives similar delta(13)C values to the Jayme-Wise method for wood of Eucalyptus globulus Labill., Pinus radiata D. Don and Pinus pinaster Ait. The relationship between delta(13)C of wood and crude cellulose is as strong as that observed between wood and alpha-cellulose and stronger than that observed between wood and holocellulose in other species. These relationships suggest that variation in delta(13)C of wood may result from hemicellulose and that analysis of stable carbon isotopes in crude cellulose is preferable. If the consistent -0.3 bias in the value of delta(13)C of cellulose resulting from residual lignin is corrected for, then the relationship between delta(13)C of wood and crude cellulose may be used to predict delta(13)C of cellulose from a small sub-sample. The method is well suited to species with low concentrations of extractives, but further testing is needed to assess its general applicability.