A three-year integration of a global three-dimensional model including dynamics and simple photo- chemistry is used to predict ozone. Distributions of NO3 and odd hydrogen deduced by McConnell and McElroy are used to incorporate in a simple way the chemical effect of these species. Good agreement with observation is obtained for stratospheric motion patterns, meridional circulations, ozone density as a function of height and latitude, eddy transports of ozone, surface destruction of ozone, and correlations of ozone with other variables. The annual cycle of columnar ozone in high latitudes is present, but at a smaller amplitude than observed. Vertical transport of ozone downward from the main generation level at 30 km is accomplished primarily by small-scale eddy diffusion between 20 and 30 km and again near the ground; large-scale vertical transport dominates inbetween. The model predicts a secondary maximum in ozone mixingg ratio at 45 km somewhat equatorward of the winter-polar-night zone. This feature, recently observed from satellite measurements, is thought to he caused by the temperature-dependence of reaction rates in the Chapman scheme. The principal deficiency of the model is an underprediction of the spring ozone concentration in high latitudes in the lower stratosphere.