Lyophilised birch pollen allergen extracts, reconstituted with different diluents (H2O, saline, Albumin diluent (AD] were investigated to determine whether the allergen activity and quality of the extracts deteriorated by nebulization with different nebulizers (Pari, Wright, and Sandoz). Allergen activity was measured by IgG4 RAST inhibition technique and allergen quality was analysed by crossed immunoelectrophoresis (CIE). The distribution of particle sizes of aerosols of different allergen solutions was determined by a TSI Aerodynamic Particle Sizer. A significant difference (P less than 0.05) in allergen activity was found between the AD and H2O diluents before and after using a Sandoz nebulizer and a Wright nebulizer equipped with a small chamber. This suggested greater allergen activity in AD-diluted solutions, and the pattern was repeated with the other two nebulizers, but was not statistically significant. The samples diluted with saline showed no significant differences in quality after nebulization except for the impacted aerosol in which one of the precipitates was slightly diminished. In the AD-diluted sample one of the precipitates disappeared from the impacted aerosol and from the nebulization chamber after 2 min nebulization. To elucidate whether the extract proteins had been bound by albumin, samples were submitted to CIE with rabbit anti-human albumin in an intermediate gel. By this procedure one precipitate was transformed to a precipitate in the intermediate gel, indicating that one or more proteins in the extracts may associate with albumin. No significant difference in output was observed between the nebulizers. The particle size distribution curves for each diluent (saline, AD) were identical for the Wright nebulizer with 99% of the dry particles distributed within 0.5-2.0 microns.(ABSTRACT TRUNCATED AT 250 WORDS)