The growth of the grass Brachypodium pinnatum (L.) Beauv. in Dutch nutrient-poor chalk grasslands increases with enhanced nitrogen supply, whereas other grass species also require an enhanced phosphorus supply for a similar response (e.g. Dactylis glomerata L.), or are competitively suppressed at any increase in nutrient supply (e.g. Briza media L.). We investigated whether this interspecific variation in response to N and P supply is caused by differences in P productivity (PP), i.e. the instantaneous rate of biomass production per unit of P present in the plant. We hypothesized that PP is highest in Brachypodium pinnatum, in contrast to N productivity which is known to be the highest in Dactylis glomerata. Phosphorus productivity and its components were studied using a growth analysis with four exponential P addition rates of 0.03, 0.06, 0.09 and 0.11/0.15 mg P mg-1 P d-1. Although Brachypodium pinnatum allocated more P to its leaf blades, it had a lower P productivity at high N and low P supply than did Dactylis glomerata. This was associated with a higher productivity per unit leaf P in Dactylis glomerata. Across all species and treatments, leaf PP showed a distinct negative correlation with P concentration per leaf area, regardless whether the variation in area-based leaf P concentration was caused by variation in leaf thickness, leaf tissue mass density or mass-based P concentration. A possible explanation for this would be a positive correlation between leaf chlorophyll concentration and P concentration, leading at high concentrations to shading within the leaf and to a low photosynthetic rate per unit leaf P. We conclude that a high PP is determined by the ability of a plant to distribute its P over a large leaf area, rather than by greater allocation of P to the leaves. Interspecific relationships for P productivity are similar to those known for N productivity.