ABSTRACT:We present results of multiple regressions of the leading mode of atmospheric variability at southern high latitudes: the Southern Annular Mode (SAM). It is regressed against indices with large interannual variability, and one of several trend indices in order to determine which trend term gives the optimum fit. We use SAM in sea-level pressure from station data in order to provide a long time series, from 1957 to 2005. The regression indices are stratospheric volcanic aerosol, solar activity, the quasi-biennial oscillation (QBO), the El Niño-Southern Oscillation, together with either a linear trend, or the effective equivalent stratospheric chlorine (EESC) that depletes polar ozone, or ozone mass deficit (OMD) in the Antarctic vortex. We find a statistically significant signal for volcanic aerosol, and, when the solar and QBO indices are represented by their product, a highly significant response to the product. We find a significant linear trend in SAM, but there is a major increase in significance using EESC 2 and a further increase using OMD. We make no direct attempt to identify cause and effect, but if the trend is due to human influence then ozone loss is at least 9 times more likely the principle cause of the trend in SAM than greenhouse gases, although we do not exclude greenhouse gases making a smaller contribution. Monthly and seasonal regressions show a maximum correlation with OMD between December and May (summer and autumn), consistent with previous work on stratospheric change as a cause of change in the troposphere.