Two grass species, Holcus lanatus and Deschampsiajexuosa, were grown in soil containing available nitrogen at 4 pg g-' ('Low N') or 58 pg g-' ('High N'), in a controkdenvironment. Water was supplied via an auto-irrigation system, which also minimized the vertical redistribution of soil nutrients, particularly NO, -. Growth, root morphology and distribution, and N uptake were measured during a 70-day period.H. lanatus produced more dry matter than D.jexuosa in the high N treatment. Yields were similar in both species in the low N treatment. The growth of H. lanatus in the low N treatment was not limited by a complete exhaustion of soil N. Root density of H. lanatus was likely to have been inadequate to allow this species to attain its maximum potential N-inflow rate. Thus, N uptake per plant and dry matter production were restricted. The growth of D.flexuosa was not limited by the low N treatment, compared with that in the high N treatment. This was attributed to its low maximum potential relative growth rate and, hence, its low demand for N compared with H . lanatus, rather than to any major 'compensatory' adjustments in root morphology or N absorption efficiency.