Excess water can induce flooding stress resulting in yield loss of crops, even in wetland plants such as rice. However, traits from species of wild Oryza have already been used to improve tolerance to abiotic stress in cultivated rice. This study aimed to establish root responses to sudden soil flooding among 8 wild relatives of rice with different habitat preferences benchmarked against 3 genotypes of O. sativa. Plants were raised in hydroponics, mimicking drained or flooded soils, to assess the plasticity of adventitious roots. Traits included were apparent permeance (PA) to O2 of the outer part of the roots, radial water loss (RWL), tissue porosity, apoplastic barriers in the exodermis and root anatomical traits. These were analysed using a plasticity index and hierarchical clustering based on principal components analysis. For example, O. brachyantha, a wetland species, possessed very low tissue porosity compared to other wetland species, whereas dryland species O. latifolia and O. granulata exhibited significantly lower plasticity compared to wetland species and clustered in their own group. Most species clustered according to growing conditions based on PA, RWL, root porosity and key anatomical traits, indicating strong anatomical and physiological responses to sudden soil flooding.

Aims Root tissue water can be lost to the dry topsoil via radial water loss (RWL) resulting in root shrinking and loss of contact with the rhizosphere. The root barrier to radial oxygen loss (ROL) has been shown to restrict RWL, therefore we hypothesized that the inducible barrier can be formed as a response to low soil water potential and play a role, together with other root traits, in restricting RWL. Methods Rice and wheat were grown in hydroponics with contrasting water potential to diagnose ROL barrier formation and to explore how key root traits (ROL barrier, root diameter, root porosity) affect RWL. Moreover, we developed a numerical model predicting RWL as a function of root diameter, root porosity and presence of a barrier to ROL. Results Methylene blue staining showed that low water potential induced a ROL barrier formation in roots of rice, and also resulted in an apoplastic barrier, as identified by the apoplastic tracer periodic acid. The barrier significantly restricted RWL, but root diameter and tissue porosity also influenced RWL. Our numerical model was able to reflect the empirical data and clearly demonstrated that thick roots and a barrier to ROL restricts RWL while cortical porosity accelerates RWL. Conclusions Our modelling approach highlighted that increase in root tissue porosity, a common response to drought, conserves water when new roots are formed, but the higher desiccation risk related to high-porosity roots can be effectively counteracted by forming thick roots or even better, by a barrier to ROL.