Pollen and spore counts from Burkard traps for routine pollen and spore sampling placed at 15 m above ground and at ground level were compared. Daily counts of most pollen types were higher on the ground than at roof level, but the counts were significantly correlated. The ratios of pollen frequencies at high and low levels varied between 1.0 and 11.5. The most prominent differences were recorded for herbaceous pollen (e.g. Artemisia counts 11.5 and Poaccae counts 4.4 times higher at ground level) and in Botrytis and Ustilaginales spores. Tree pollen grains and basidiomycetous spores were more equally distributed. Wind speed did not affect the variation of pollen frequencies at either height. Large spores are not so unevenly distributed as previously supposed. Artemisia and grass pollen was detected 1 to 2 weeks earlier at ground level than on the roof. It is therefore concluded that especially the beginning of flowering should be monitored at a low level.