Although only 15 years of continuous global satellite data are available, existing measurements are consistent with a significant, in-phase solar cycle variation of upper stratospheric ozone and temperature. Here we investigate the latitude and seasonal dependences of this variation using 15 years (1979--1993) of combined solar backscattered ultraviolet (SBUV) and SBUV/2 ozone profile measurements and 14 years (1980--1995) of National Meteorological Center (NMC) temperature analyses. These dependences are estimated by applying a multiple regression statistical model to monthly zonal mean time series extending from 60 ¿S to 60 ¿N latitude and approximately 25- to 50-km altitude. Solar variability is represented in the statistical model by the Mg II index, a close proxy for solar UV variations at wavelengths that affect the photochemical production of ozone. In agreement with earlier studies, although an apparent solar cycle variation of both ozone and temperature is present in the upper stratosphere, no detectable solar cycle variation is present in the middle stratosphere (30- to 35-km altitude). Ozone increases of 4--6% from solar minimum to solar maximum are found near 2 mbar at middle latitudes, with comparable amplitudes in both summer and winter hemispheres. While the presence of a large positive ozone response in the northern hemisphere during winter appears to be qualitatively consistent with theoretical predictions, the amplitude of the observed ozone variation is nearly twice as large as estimates based on current two-dimensional models of the middle atmosphere. Temperature increases of 2.5 K are found near 1 mbar at low latitudes throughout the year, in addition to a seasonally varying temperature response of 2.5--3 K near 5 mbar at middle and high latitudes in the summer hemisphere. A one-dimensional radiative model is used to calculate the expected change in equilibrium temperature associated with the observed solar cycle variability in the ozone profile under the assumption of fixed dynamical heating (FDH). Near the equatorial stratopause, the temperature response calculated with the FDH model is within 0.5 K of the observed response. At higher latitudes near 4 mbar, the FDH model cannot account for the large amplitudes and strong latitude dependences of the NMC-derived temperature variation. This would suggest that changes in stratospheric dynamics over the 11-year solar cycle may also be important for understanding the observed temperature and ozone variations. ¿ American Geophysical Union 1996