Stratospheric aerosol concentrations obtained on seven balloon soundings by in situ photoelectric counting of individual particles and jet impaction with later electron microscope analysis have been compared. The potential error sources in the two sizing techniques appear sufficient to explain the observed differences in concentration. This implies an underestimate in the impactor-derived precollection radius (for sulfuric acid aerosol) of about 25% relative to the photoelectrically determined radius in the two ranges r>0.15 μm and r>0.25 μm. The variance in impactor-derived column loadings is determined by comparison of the impactor and photoelectric detector loadings: over the 10- to 28-km altitude range, for example, this gives an error estimate of ?¿18% for a 95% confidence level. An estimate of the flight-to-flight accuracy of the impactor sizing models is also determined from the loading comparison.