Plants have evolved arrays of responses to abiotic constraints, many of which involve the root phenome. Presumably, these responses were shaped by natural selection to maximise adaptation and reproductive success. However, it is generally unclear whether this plasticity also contributes to improved agronomic performance of cultivated species under abiotic stress. In the context of the SolACE EU project, we monitored root growth and architecture of 250 bread wheat and 250 durum wheat genotypes in the 4PMI platform (rhizotubes) under combined low N and water and in the RootPhAir platform (aeroponics) under unconstrained conditions. Phenotypic plasticity in response to low N and water was estimated by comparing data from the two platforms. In parallel, conventional phenological and yield component variables were recorded for the same genotypes in replicated field trials involving control conditions vs combined low N and water. AMF colonisation data were also recorded on roots sampled in control conditions from one of the durum wheat trials. The genotypic effect (within each species) was significant for many variables and experiments. Root plasticity was predominant in the architectural data, as indicated by extremely low correlations between analogous variables in the two platforms. The root:shoot ratio was higher in 4PMI than in RootPhAir, which is consistent with the abiotic constraint present in the former. However, the genotypic variance for the root:shoot ratio was larger in RootPhAir than in 4PMI, as if the constraint in the latter was reducing the genotypic variability in root:shoot allocation. Field data were summarized by computing the rank of genotype responses to combined water and N deficits across the different trials, low ranks being assigned to generally sensitive genotypes. After assembling field and platform data, the sensitive genotypes in field trials tended to be those which have a low root allocation in RootPhAir (compared to 4PMI), while the tolerant genotypes in the field were those for which the root allocation in RootPhAir was close to that of 4PMI. Our interpretation is that the sensitive genotypes are able to reduce their root allocation when the root constraints are small, which makes them sensitive to soil-related constraints that would develop later during the crop cycle. Interestingly, the AMF data also points that the same genotypes are less keen to support the symbiotic association. Our results suggest that the ability to achieve a large root:shoot ratio is an important component of field crop tolerance to abiotic stress and call for reconsidering the value of architectural variables estimated in phenotyping platforms as direct predictors of crop tolerance.