Observations show that the monthly zonal mean total ozone between 1978 and 1982 varies by as much as 5% with respect to the 4-year average. These variations have been analyzed with the aid of a two-dimensional chemistry transport model. Observed temperatures and meridional stream functions, calculated from data, were used as inputs for the model. Deviations of the model-calculated total ozone from the 4-year mean were compared with the observed variations. The data sets used consisted of 4 years (November 1978 to October 1982) of solar backscattered ultraviolet (SBUV) and total ozone measurements and National Meteorological Center (NMC) temperature data. The diabatic circulation was obtained from radiative heating rates, which were calculated for each of the 48 months of the observations. It is shown that a substantial part of the observed variability of total ozone is reproduced by the model, despite uncertainties in the heating rates and without any special attempts to obtain consistent eddy diffusion or otherwise to tune the model. Our results indicate that the temperature dependence of the ozone gas phase chemistry plays a negligible role in explaining the observed interannual variations in total ozone. A large part of the variability can be explained by the year to year differences in the transport circulation. The variability of the diabatic circulation is shown to be largely independent of the ozone fluctuations and to be a consequence of the observed temperature variations. ¿1991 American Geophysical Union