Several middle atmospheric simulations have been carried out from January 1979 to December 1997 including most effects important to stratospheric O3. Results of these simulations for several species and species ratios have been compared in detail with observations made by the Halogen Occultation Experiment, the Polar Ozone and Aerosol Measurement II, the Total Ozone Mapping Spectrometer, the Jet Propulsion Laboratory Mark IV Interferometer, and during the Photochemistry of Ozone Loss in the Arctic Summer mission. For the simulation including all effects, comparisons of all species and ratios show excellent agreement. Comparisons of simulated sunset NO2 with and without the effects of energetic electron precipitation show excellent agreement with observations with the effects of the electrons included but poor agreement when they are excluded. The validated simulations indicate that the effects of a polar source of NOy must be included for an adequate simulation of stratospheric O3 and NOy. A comparison of simulations, during the 11-year solar cycle, of the relative effects on O3 of solar UV flux variations and the energetic electron precipitation has been made. For global total O3 the effects are comparable. For the global column above 25 km, the effects of energetic electron precipitation are significantly larger. The implications of, and some issues raised by, these findings are briefly discussed. ¿ 2001 American Geophysical Union