Backgrounds and Aims Nitrogen nutrition of legumes, which relies both on atmospheric N(2) and soil mineral N, remains a major limiting factor of growth. A decade ago, breeders tried to increase N uptake through hypernodulation. Despite their high nodule biomass, hypernodulating mutants were never shown to accumulate more nitrogen than wild types; they even generally displayed depressed shoot growth. The aim of this study was to dissect genetic variability associated with N nutrition in relation to C nutrition, using an ecophysiological framework and to propose an ideotype for N nutrition in pea. Methods Five pea genotypes (Pisum sativum) characterized by contrasting root and nodule biomasses were grown in the field. Variability among genotypes in dry matter and N accumulation was analysed, considering both the structures involved in N acquisition in terms of root and nodule biomass and their efficiency, in terms of N accumulated through mineral N absorption or symbiotic N(2) fixation per amount of root or nodule biomass, respectively. Key Results Nodule efficiency of hypernodulating mutants was negatively correlated to nodule biomass, presumably due to the high carbon costs induced by their excessive nodule formation. Root efficiency was only negatively correlated to root biomass before the beginning of the seed-filling stage, suggesting competition for carbon between root formation and functioning during the early stages of growth. This was no longer the case after the beginning of the seed-filling stage and nitrate absorption was then positively correlated to root biomass. Conclusions Due to the high C costs induced by nodule formation and its detrimental effect on shoot and root growth, selecting traits for the improvement of N acquisition by legumes must be engineered (a) considering inter-relationships between C and N metabolisms and (b) in terms of temporal complementarities between N(2) fixation and nitrate absorption rather than through direct increase of nodule and/or root biomass.