Low-frequency stationary wave variability, having periods of greater than 8 months, has been studied in the total ozone column measured by the Nimbus 7 total ozone mapping spectrometer (TOMS) over the period 1979--1988. This has been compared with the total ozone column variability obtained from a general circulation model (GCM) having prescribed observed sea surface temperatures (SSTs) for the same period. The GCM results demonstrate that the total ozone stationary waves are strongly anticorrelated with the 200-hPa eddy geopotential height, suggesting that total ozone can be used to study upper level geopotential height stationary waves and their low-frequency variability. The response to El Ni¿o/Southern Oscillation (ENSO) events is clearly seen in both the TOMS and the GCM total ozone stationary waves. In the tropics, there is a tropical dipole pattern response to ENSO similar to the negative of the dipole response seen in the upper level geopotential height, attributable to the east-west ENSO see-saw in tropical convection. The model total ozone response is about a half of that of the TOMS response consistent with the model geopotential height response being too weak in comparison with analyses from the European Centre for Medium-Range Weather Forecasts (ECMWF). South of 25¿S, the TOMS data show a zonal wave 1 ENSO response in antiphase with the response in the tropics and the response is such that during El Ni¿o (La Ni¿a) events the southern hemisphere stationary waves are weaker (stronger) than normal. No such response is evident in the GCM results or ECMWF analyses, which instead have a zonal wave 3 ENSO response in the southern hemisphere. A zonal wave 1 analysis reveals that the model and analyses have quite different time-mean zonal wave 1 structures in the topics and in the southern hemisphere compared with those from the TOMS data reflecting possible model errors. ¿ American Geophysical Union 1995