A model for the impact of ozone on grain yield in spring wheat (Triticum aestivum L.) was constructed. The mode was based on the observations in the literature as well as in the present experiment, that the productivity of the above-ground biomass and the carbon partitioning of the above-ground biomass do not respond to ozone above the same threshold concentrations. The model was applied to a data set in which spring wheat was exposed to different ozone concentrations in open-top chambers. It was concluded that the nonlinearity of a simple dose-response curve for grain yield can be explained in terms of the product of two linear responses: (1) the response of total above-ground biomass to an ozone dose above 10 nl 1−1, and (2) the response of the harvest index to an ozone dose above 40 nl 1−1 The nitrogen content of the grain increased linearly with the dose above 40 nl 1−1.It is hypothesized that the effect of comparatively small ozone concentrations on total above-ground biomass is caused by a reduction of leaf duration by ozone, and that the effect on carbon partitioning is caused by ozone affecting the sink capacity or activity of the growing grain. The maximum coefficients of determination (r2) for the linear regression between grain yield and accumulated ozone dose above thresholds in the range of 0–70 nl 1−1, were never as great as those for total above-ground biomass and harvest index. Nevertheless, statistically significant linear regressions were obtained for the thresholds 20, 30 and 40 nl 1−1, the highest correlation being with 30 nl 1−1 Thus, it seems acceptable to use linear regressions between grain yield and the accumulated dose, although these do not take into direct account the complexity of the effects of ozone on wheat.