The objective of this work was to determine if the impact of nitrogen (N) on the release of organic carbon (C) into the soil by roots (rhizodeposition) correlated with the effect of this nutrient on some variables of plant growth. Lolium multiflorum Lam. was grown at two levels of N supply, either in sterile sand percolated with nutrient solution or in non-sterile soil. The axenic sand systems allowed continuous quantification of rhizodeposition and accurate analysis of root morphology whilst the soil microcosms allowed the study of C-14 labelled C flows in physico-chemical and biological conditions relevant to natural soils. In the axenic sand cultures, enhanced N supply strongly increased the plant biomass, the plant N content and the shoot to root ratio. N supply altered the root morphology by increasing the root surface area and the density of apices, both being significantly positively correlated with the rate of organic C release by plant roots before sampling. This observation is consistent with the production of mucilage by root tips and with mechanisms of root exudation reported previously in the literature, i.e. the passive diffusion of roots solutes along the root with increased rate behind the root apex. We proposed a model of root net exudation, based on the number of root apices and on root soluble C that explained 60% of the variability in the rate of C release from roots at harvest. The effects of N on plant growth were less marked in soil, probably related to the relatively high supply of N from non-fertiliser soil-sources. N fertilization increased the shoot N concentration of the plants and the shoot to root ratio. Increased N supply decreased the partitioning of C-14 to roots. In parallel, N fertilisation increased the root soluble C-14 and the C-14 recovered in the soil per unit of root biomass, suggesting a stimulation of root exudation by N supply. However, due to the high concentration of N in our unfertilised plants, this stimulation was assumed to be very weak because no significant effect of N was observed on the microbial C and on the bacterial abundance in the rhizosphere. Considering the difficulties in evaluating rhizodeposition in non sterile soil, it is suggested that the root soluble C, the root surface area and the root apex density are additional relevant variables that should be useful to measure along with the variables that are commonly determined when investigating how plant functioning impacts on the release of C by roots (i.e soil C, C of the microbial biomass, rhizosphere respiration).