Crop residue decomposition is controlled by several factors, of which the biochemical composition of the plant material is particularly important. Different approaches have been proposed to characterize the quality of crop residues and predict their kinetics of C and N mineralization. Amongst these, the chemical extraction proposed by Van Soest (1963) is one of the most commonly used. This extraction separates the plant components into four fractions (soluble, cellulose, hemicellulose and lignin) differing by their recalcitrance to a chemical attack. These fractions are then assumed to decomposed at various rates in soils, the lignin pool being the most recalcitrant to decomposition. Several C and N biotransformation models utilized this concept and represent C and N fractions in crop residues by several compartments decomposing at differing rates (Garnier et al., 2003). However, the prediction of C and N mineralization kinetics remained difficult, and this was particularly true for the roots (Abiven et al., 2005; Rasse et al., 2005). Roots constitute one of the main source of C to soil organic matter and the understanding of their decomposition in medium and long term studies should be improved. Our hypothesis is that a better knowledge of biochemical characteristics of root residues would further our understanding of the decomposition process in soil. The main aim of this work is thus to verify that specific biochemical properties of root cell walls could contribute to explain and predict the mineralization of C in soils. We used maize roots to test the importance of cell wall composition on the kinetics of mineralization of C in soil. To do so, 16 different maize genotypes were sampled in the field and roots were separated, washed and dried pending analyses. Biochemical characterization consisted on classical determination such as C-to-N ratios and Van Soest extractions. In addition to these, cell wall composition was determined by measuring the amount and nature of neutral sugars by acid hydrolysis (Lequart et al., 1999), of lignin by thioacidolysis (Lapierre et al., 1986), of phenolic acids following the method described by Beaugrand et al., 2004. Roots were then cut into small pieces (5-8 mm) and incubated in a soil in controlled temperature and moisture conditions. Potassium nitrate was added to the initial N concentration of the soil to ensure that N would not limit decomposition (Recous et al., 1995). Carbon mineralization was measured in soil samples incubated in the presence of a CO2 trap. The different genotypes used presented variations in some specific cell wall components such as lignin and phenolic acid contents. Phenolic acids which are the precursor of lignin, play a key role in cross-linking xylans to each other and to lignin resulting in less degradable walls (Bertrand et al., 2005). The cross link between polysaccharide and lignin could also play a key role in understanding soil residue decomposition because the lignin-polysaccharide complexes could reduced polysaccharides accessibility to soil micro-organisms. Phenolic acids could also induced a toxicity to soil micro-organisms (De Ascensao and Dubery; 2003). This range of plant material is thus interesting to test the relative importance of specific cell wall components on C mineralization process. The biochemical changes induced in the cell wall network of maize roots were related to their kinetics of C mineralization in soil and the results will be discussed in detail. References: (1) Abiven S et al., 2005 Soil Biol. Fert. 42, 119-128. Beaugrand J et al., 2004 J. Agric. Food Chem. 52, 7108-7117. (2) Bertrand I et al., 2005 Plant Soil, In Press. De Ascensao ARFDC and Dubery IA, 2003 Phytochemistry 63, 679-686. (3) Garnier P et al., 2003 EJSS 54, 555-568. Lapierre C et al., 1986 Holzforschung 40, 113-119. (4) Lequart C et al., 1999 Carbohydrate Research 319, 102-111. (5) Rasse DP et al., 2005 Plant Soil 269, 341-356. (6) Recous S et al., 1995 Soil Biol. Bioch. 27, 1529-1538. (7) Van Soest PJ, 1963 Assoc. Off. Agr. Chem. Jour. 46, 825-829.