During operation of smog chambers, the gas mixture is normally well-stirred although the extent of stirring varies from chamber to chamber. Experiments conducted at Battelle’s Columbus Laboratories in a 200-liter and a 610-cu ft smog chamber have shown that stirring can decrease aerosol formation. The faster the chamber contents are stirred the greater the reduction in aerosol formation observed. Sufficiently rapid stirring can completely prevent observation of aerosol formation by light scattering. However, thermal gradients in these chambers have been adequate to maintain homogeneity after initial mixing and satisfactory chamber experiments are being performed by turning the stirring off prior to turning on the lights. The differences in aerosol formation between stirred and nonstirred chamber operation is dependent on the type of system being studied and the stirring rate. In the 200-liter chamber during dynamic operation (4-hour average residence time) aerosol formation from 1-heptene/NOx and l,3,5-trimethylbenzene/NOx was completely suppressed by stirring at 1 600 rpm, while aerosol formation from 1 -butene/NOx/SO2 was decreased by a factor of 25. In the 610-cu ft chamber the following decreases were observed by stirring at 218 rpm: toluene/NOx, 73%; 1 -heptene/NOx, 37%; SO2/clean air, 20%; 1-heptene/NOx/SO2, no change. The differences in sensitivity to stirring between olefins and aromatics, and between systems which do or do not contain SO2 account for many of the discrepancies in the literature with respect to aerosol formation. Many of the early studies of aerosol formation need to be re-interpreted because the role of stirring was not recognized. In future studies in which aerosol formation is measured, stirring effects should be well defined.