The phosphorus nutrition index (PNI) method proved to be suitable for diagnosing P nutrition level over a wide range of grass swards. It is based on relative nutrient concentrations and relies on a critical curve for optimum P concentration which is a function of sward N concentration. The objectives of this paper were to verify to what extent this method was valid for P nutrition diagnosis when white clover was present in the sward in view of the fact that white clover N concentration is frequently reported to be much higher than that of grass and to remain constant whatever the amount of biomass produced. For this purpose we analysed the effect of N addition on N and P concentrations in white clover and rye grass grown in pure stands with unlimited P supply in a garden collection and compared it with clover behaviour in a field experiment on a natural grassland with a similar P status. Clover displayed a different affinity for P according to whether it grew in a pure or mixed sward. In a pure stand, we observed a very significant effect of N on N and P concentrations measured in individual organs and whole plants; the highest N concentrations were measured in clover. This species behaved differently from ryegrass: for a given level of N supply, plant N concentration decreased when petioles were added to the laminae samples, while P concentrations remained constant; for rye grass both N and P concentrations decreased when sheaths were added to laminae. Clover in mixed swards had a lower P nutritional status than when grown in a pure stand, due to its lower capacity to compete with grasses for light capture and P acquisition. Analysis of data from the literature demonstrated that differences in response between clover and grass were more pronounced under limited P supply, when clover displayed a lower ability than grasses to adjust P concentration to that of N due to its lower capacity to adapt to low P availability. As a consequence it became even less competitive. We concluded that there are limitations to the use of nutrient diagnosis systems based on relative concentrations when the herbage contains a large proportion of white clover and proposed that only the non-clover fraction should be used for diagnosis to avoid excessive P application when PNIs are used for fertiliser recommendations.