Meteorological conditions in the Atlantic Stratocumulus Transition Experiment (ASTEX) region favored advection of clean air from the central Atlantic during the early part of the experiment that was replaced by polluted air of European origin during the latter part of the experiment. Marked differences in the aerosol size distribution, composition, and state of mixing existed in these air masses. Pronounced differences in their vertical structure also demonstrated that surface measurements often do not represent average boundary layer or column concentrations. Clean subsiding air from the free troposphere had concentrations of condensation nuclei that significantly exceeded concentrations in the boundary layer and had very low mass concentrations and volatility consistent with homogeneous nucleation aloft, supporting the hypothesis that these nuclei can provide a source for new nuclei into the marine surface layer. This finding was in contrast to polluted air characterized by high concentrations of aged aerosol but having no evidence for significant recent nuclei formation. Particles in polluted air consisted of more than 90% volatile mass (mostly sulfate) and a refractory residual remaining at 300 ¿C. The refractory mass varied with the concurrently measured light absorption coefficient, associated with combustion-derived soot. In spite of 2 orders of magnitude more volatile sulfate in polluted air, most of the particle number in polluted air remained after volatilization at 300 ¿C, in contrast to less than 10% in clean air. This finding suggests that the number of particles in polluted air may reflect the combustion process producing the primary soot aerosol more than the SO2 emissions responsible for much of the sulfate aerosol mass. The accumulation mode mass mean diameter was found to increase with total accumulation mode mass, suggesting that a size parameterization based upon the more commonly measured sulfate mass could be included in current aerosol models. ¿ 1997 American Geophysical Union