Atmospheric density measurements near 200 km from the Satellite Electrostatic Triaxial Accelerometer (SETA) experiment are analyzed for geomagnetic and solar ux variability e ects. Data from the SETA experiment, onboard two satellites, are available for the periods of May to November 1982, and July 1983 to March 1984. The data utilized the span ยฑ79:5 โข latitude, and are available for both day (1030 LT) and night (2230 LT). Annual and semiannual density variations are removed and regression analyses are performed on the residuals using a series of lagged 3 h Kp indices to determine and remove geomagnetic uctuations. Densities are found to increase by as much as 134% in response to an increase in the Kp index from 1 to 6. Monthly curves are generated for the Kp regression coe cients to delineate seasonal-latitudinal and day=night dependences, which re ect the e ects of mean meridional advection of disturbances from high to low latitudes. Further analyses are performed comparing measured densities with MSISE-90 predictions. Results show that the model is able to capture many of the prominent features, but does not fully predict the level of variability for the individual disturbance periods analyzed. After the geomagnetic e ects are removed, the residual densities are interpreted in terms of solar ux variability. The daily-averaged SETA density residuals are strongly correlated with long-term solar ux variability, and exhibit a much greater dependence on the 27-day solar rotation period than MSISE-90 predictions. Variations in residual density of the order of 10 -20% occur in association with day-to-day and 27-day solar ux variations. The MSIS model does not accurately predict the magnitude of these short-term density variations in response to solar activity.