A simple model of the decomposition and nitrogen mineralization of plant material from two unfertilized grassland ecosystems has been developed, with only the proportion of leaves and stems in the original material, the initial nitrogen contents of these plant parts and temperature as input data. The model simulates carbon losses from stems and leaves, using a double exponential decay function, with the temperature sum as independent variable. Mineralization of nitrogen is not calculated via microbial growth rates, but simulated on the basis of the carbon utilization efficiency of the microorganisms and the critical C/N ratio, i.e. the C/N ratio of the litter at which the microbial demand for nitrogen is met exactly. The parameter values for leaching fractions of carbon and nitrogen, relative decay rates, microbial carbon utilization efficiencies and critical C/N ratios were derived from a litter bag experiment with 12 litter types (species) including both green and dead materials, carried out in two unfertilized grassland ecosystems differing in production level. The model was evaluated using a cross-validation method, in which one species was omitted from the parametrization procedure, and its decomposition and mineralization were predicted by the resulting model. In general there was good agreement between the observed and predicted amounts of carbon and nitrogen remaining for all green litter types/species, but carbon and nitrogen dynamics in the dead material of Festuca rubra were poorly predicted. This disparity has been attributed to the proportion of leaves in the material of Festuca rubra (95 being far beyond the range of leaf proportions in the three litter types the calibration set consisted of (When the data of all litter types were used to determine the model parameters, good agreement was obtained between measured and simulated values for the changes in nitrogen and carbon in all litter types of both the green and dead material series. Optimization yielded parameter values for microbial carbon utilization efficiencies of 0.30 for microorganisms associated with green litter and 0.35 for those associated with dead litter. The critical C/N ratios for green and dead material were found to be 29 and 36, respectively.