The first 18 months of altimeter data from the TOPEX/POSEIDON (T/P) mission were analyzed to study the large-scale mean surface circulation and its variability in the south Indian Ocean. A T/P sea level time series corrected for ocean tides using the Texas model (Eanes, 1994) was first validated against in situ sea level measurements from a bottom pressure recorder near Amsterdam Island. T/P data are a little contaminated by a 60-day tidal aliasing effect, very near the alias periods of M2 and S2 constituents. When this 60-day effect is removed, the data agree well with the bottom pressure-derived sea level within 2.0 cm rms difference. This indicates that only low-frequency ocean signals of periods greater than 60 days can be safely studied from T/P data. A T/P-derived mean dynamic topography relative to the Ohio State University 91A geoid (Rapp et al., 1991) was compared to the output of the fine resolution Antarctic model (Webb et al., 1991) and to historical hydrographic data. The altimetric solution shows excellent agreement with the numerical model solution and indicates an anticyclonic subtropical gyre north of the Antarctic Circumpolar Current and two cyclonic subpolar gyres south of this current, respectively, west and east of the Kerguelen Plateau. The western subpolar gyre corresponding to the eastern boundary of the Weddell Gyre appears to extend as far east as 60 ¿E in the Weddell-Enderby Basin. However, the hydrography-derived dynamic topography does not reveal any clear evidence of these gyre features because of its inability to detect the barotropic component of currents. Finally, the T/P-derived sea level variability reveals clearly the well-known, strong current-topography interaction in the Southern Ocean. There appear no significant interannual variations of the mean position and magnitude of the major current systems in the south Indian Ocean. ¿ American Geophysical Union 1995