A one-dimensional photochemical model of the Martian atmosphere has been developed to study variations in the atmospheric composition with solar activity. The low sensitivity of (\mathrm{O}) and (\mathrm{CO}) number densities near the ionospheric peak to solar activity suggests that eddy mixing on Mars is proportional to the solar activity index (\left(K(z)=\left(F_{10.7 \mathrm{~cm}} / 30\right) \times 10^{13} n^{-1 / 2}\right)). Due to the higher escape rate and the stronger eddy diffusion at solar maximum the (\mathrm{H}{2}) densities above the ionospheric peak are smaller by a factor of 6 than those at the solar minimum. The solar cycle variation of (\mathrm{CO}{2}^{+})by a factor of 2.2 cannot compensate for the variation of (\mathrm{H}{2}), and the atomic hydrogen escape flux at solar minimum is larger by a factor of 1.7 than that at solar maximum (the variation of the (\mathrm{H}) number density is a factor of 30()). The total hydrogen ((\mathrm{H}+) (2 \mathrm{H}{2}) ) escape flux is stable (though not diffusion limited) and equal to (2.5 \times 10^{8} \mathrm{~cm}^{-2} \mathrm{sec}^{-1}). If eddy diffusion does not depend on solar activity, then the (\mathrm{H}) escape is stable while the total escape flux varies by a factor of 2 . The main reason for the hydrogen escape rate being close to the diffusion limit in the early models is that the rate coefficient of the (\mathbf{H}_{\mathbf{2}}) dissociation reaction was larger in the early models by a factor of 3 than the current value. Solar cycle variation of the lower and middle atmosphere is also studied. The CO mixing ratio in the lower and middle atmosphere would vary by a factor of (\mathbf{4}) if the species lifetime were a few months. A mechanism of this variation is analyzed. However, since the real lifetime is 6 years the variation is only a factor of 1.35 , with a time lag of 2 years. 1993 Academic Press, Inc.