Résumé : Modifying root architecture is a strategy that aims to develop plants, which capture nutrients more efficiently and thus suitable for sustainable agriculture with fewer fertilizer inputs. The focus is on nitrate, since it is a major nutritional determinant of root morphology and because of its importance in determining yield. Low concentrations of nitrate in soil stimulate the development of lateral roots, thereby increasing the root surface available for the uptake. Conversely, uniformly concentrations of nitrate inhibit the elongation of lateral roots. The aim of the thesis was to study the natural variation of the root morphological responses to nitrate in Arabidopsis thaliana. First, the manuscript starts with the screening of a core collection of twenty-four accessions, which maximize the genetic diversity within the species, on agar medium at low and moderate nitrate levels. Our results showed that the variability for production and allocation of biomass and root architecture traits exists within the species. Second a detailed characterization is done with eleven accessions showing contrasting root morphological responses to nitrate supply. We demonstrated that at an early development stage, the nitrate uptake efficiency is not implicitly correlated with root system architecture. Third, the genetic determinism of the natural variation of the root system architecture is studied. A combination of genome-wide association mapping on a larger number of accessions (> 300), a linkage mapping with existing recombinant inbred lines and a bulk segregant analysis was carried in order to identify candidate genes involved in root morphological responses to nitrate. The perspectives of this work would be, through a model species to crop pipeline, to translate knowledge from Arabidopsis to Brassica crops that would have a root architecture redesigned to increase the acquisition of nutrients.