In an experimental field study, airborne particulate birch and grass pollen allergens were sampled with a cascade impactor on the roof of an office building in the centre of a middle-sized town in Sweden. The impactor separated particles into eight size fractions. Simultaneously, inhalable pollen allergens in the indoor air were sampled in offices of the building. Significant amounts of the outdoor pollen allergens were found on particles much smaller than the pollen grains. These small particles could penetrate the fine filters (quality F6) of the building’s ventilation system. By taking into account the effectiveness of the installed ventilation filters, the pollen allergen concentration indoors could be well predicted, thus pointing out the supply air as the main source of indoor pollen allergens. Comparison of data from workdays with those from weekends indicated that secondary emissions of pollen allergens brought indoors by people (on clothes, hair, shoes etc.) were insignificant.