A one-dimensional aerosol model is used to investigate the sensitivity of the stratospheric distributions of gaseous sulfur compounds and sulfate aerosol particles to changes in OH and CS2 concentrations, in eddy diffusion coefficients, and in key chemical rate constants. By comparing model predictions with recent observational data for So2, OCS, and particulates, the following conclusions are reached. With regard to atmospheric sulfur, it appears that CS2 is only a secondary source of sulfur for the stratosphere relative to OCS and that background tropospheric CS2 concentrations are likely to be <70 pptv. It is also found that under stratospheric conditions, the rate coefficients for the reactions of OH with OCS and CS2 may be substantially smaller than the room temperature laboratory values of Kurylo (1978). The most important conclusion of the present study, however, is that OH concentrations below 30 km may be over-estimated by a factor of 3 or more in current photochemical models. A variety of observational evidence is discussed which supports this conclusion, including measurements of HNO3/NO2 concentration ratios and CIO mixing fractions. Plausibility arguments for lower atmospheric OH concentrations are supported by recent laboratory studies of the pressure, the temperature, and the humidity dependences of OH an HO2 chemical kinetic systems. New reaction rate data bearing on this problem are reviewed. A reduced level of stratospheric OH has strong implications for ozone perturbations attributable to aircraft NOx emissions, to nitrogen fertilizer usage, and to halocarbon releases. Such ozone perturbations are simulated using a one-dimensional atmospheric photochemistry model in which the OH concentrations are reduced in accordance with the sulfur sensitivity tests highlighted above. Owing to the reduction in OH, SST NOx injections at 20 km act to decrease ozone, not increase it as had been thought previously. Nitrous oxide produced by fertilizer decomposition is also found to be more destructive of ozone. On the other hand, fluorocarbon-induced ozone depletions are somewhat smaller.