Abstract: Proton Induced X-ray Emission, (PIXE), is capable of high precision analysis for trace element components of aerosol particle size fractions sampled by cascade impactor. A statistical evaluation of data quality has been carried out in order to distinguish between analytical uncertainties in the PIXE procedure, errors caused by cascade impactor performance and by other factors in the sampling procedure, and geophysical causes of differences in composition and particle size distributions of the elements in aerosols. Replicate analyses and simultaneous samplings taken in north Florida and St. Louis have been used for the data evaluation. In addition to the analytical error the sampling procedure contributes an error of ~ 10% to be added quadratically. The resulting precision is sufficient to evaluate the data in geophysical terms. This is illustrated by means of sample sets taken simultaneously in an urban, forest and coastal environment of the same region.