Résumé
Root System Architecture (RSA) plays a central role in plant performance by regulating water and nutrient uptake. As agriculture faces increasing challenges from environmental variability, nutrient limitation and water scarcity, identifying adaptive root traits in wild relatives is critical for developing resilient crop varieties. We screened a diverse panel of cultivated and wild lentil (Lens spp.) accessions using the Rhizoscope, a high-throughput root phenotyping system developed by CIRAD. In total, 42 wild accessions and eight advanced breeding lines were evaluated for RSA traits and quantified at 30 days after sowing using a rhizobox-based phenotyping platform. Our objectives were to assess RSA variation within wild species and compare RSA traits between cultivated and wild genotypes. Cultivated lentil showed higher values for traits such as root mass, diameter, root volume, root angle (RA) and maximum root depth (MRD), suggesting greater resource acquisition efficiency. In contrast, wild accessions exhibited higher root:shoot ratios and Collar-First Ramification length (CRL), consistent with adaptation to resource-limited environments. To understand the drivers of RSA variation, we incorporated environmental variables from the center of origin of each accession, including Aridity Index, soil type and bedrock depth, into multivariate analyses using Linear Discriminant Analysis and Classification and Regression Trees. Results showed that variation in traits such as MRD, RA and CRL was more strongly linked to environmental conditions than species classification. Deeper roots were associated with arid regions and deep bedrock, while wider RAs and shorter CRL lengths were typical of genotypes from compacted or shallow soils. These findings suggest that RSA traits in wild lentil species are shaped primarily by local environmental selection rather than taxonomic identity. This highlights the importance of integrating ecological provenance with phenotypic assessments when evaluating wild germplasm. Relying solely on species classification may overlook key adaptive traits. Incorporating environmental data can improve the identification of genotypes with root traits conferring tolerance to drought and edaphic stress, thereby supporting the development of more resilient lentil cultivars.