The 2 previous papers in this series have dealt with the response to phosphorus (P) of populations of white clover collected from soils of different P status, in experimental conditions ranging from grazed mixed swards to solution culture. This paper compares the results obtained under the various conditions. Differences between populations for some plant attributes, such as leaf size, root/shoot ratio, and proportion of stolon, were consistent across experimental conditions. However, differences in plant yield were consistent only under similar experimental conditions (e.g., grazed sward/cut sward or sand culture/solution culture). Differences between populations in response to P were even less consistent between experiments, indicating an overriding effect of experimental technique on P response. The problems of screening for low P tolerance are discussed. It is concluded that field tolerance of soils with a low P status can be accurately identified only by screening in realistic field conditions, i.e., in grazed, grassdominant swards on low P soils.