In the present paper we examine the effect of long-term solar variability on total column ozone using the longest available ground-based (1964--1994) and satellite (1979--1992) zonally averaged total column ozone records. Numerical simulations with a two-dimensional (2-D) model, which incorporates heterogeneous chemistry, transport, and the representation of the long-term solar variability, are compared to the observations. Our analysis of the total column ozone records shows that the solar activity signal in total column ozone is better seen in the tropics and during periods with no volcanoes and no synergistic effects from El Nino-Southern Oscillation (ENSO) and quasibiennial oscillation (QBO). On decadal timescales the solar activity component in total column ozone is confirmed to have a relatively small amplitude (1--2% of the 1964--1994 mean). The slowly varying decadal solar activity component in total column ozone has an amplitude 3 to 5 times larger than that of the 27-day solar rotation oscillation as seen in total ozone. The observations confirm model calculations of a larger amplitude in total column ozone at solar maximum as compared to solar activity minimum conditions for the 27-day period. The heterogeneous 2-D chemical transport model results are in good agreement with the long-term ground-based observations. It is noted that in the tropics both the interannual variability of total column ozone and the long-term impact from chlorofluorocarbons (CFCs) through the action of polar stratospheric clouds (PSCs) are small and therefore the 1--2% solar activity signal in the data can be more clearly seen above the noise level at these lower latitudes.¿ 1997 American Geophysical Union