The temperature trend in the lower stratosphere was negative between 1964 and 1993. There were, however, large fluctuations on a decadal scale, and the slope of the trend can be appreciably different depending on whether the time series begins or ends in an extreme of the decadal variations. These fluctuations affect the trends in quantities which are temperature dependent, such as ozone. For instance, the steep drop in ozone measured by TOMS between 1979 and 1990 north of about 30°N occurred in a period with a similarly steep drop in temperature, but over a longer period the temperature trend differed considerably from that of the TOMS observations. The change in the ozone during the latter period is therefore not likely to be representative of a long term ozone trend. The negative temperature trend in the stratosphere below 30 km peaked at 50 hPa (18-20 km near the level of highest concentration of ozone) at all latitudes in summer, and in the polar regions also in early winter. This is in contrast to model predictions of where the level of largest negative trend in the stratosphere associated with increasing levels of CO₂ should be. In January a positive trend began in the arctic above 30 hPa which expanded downward to the tropopause during the following months. The thermal west wind in the area of the cyclonic vortex thus weakened from above during the period. This trend pattern resembles the development of a warming in the stratosphere in individual years. Marked variations of stratospheric geopotential heights on a 10-12 year scale were well correlated with the sunspot cycle, and the overall trend in the heights of the lower stratosphere was similar to the rise in the surface air/tropospheric temperature averaged over the hemisphere.