The objective of this study was to determine the frequency of reactivity to a series of commonly used fragrances in dermatological patients. A total of 48 fragrances (FF) were chosen, based on the publication of Fenn in 1989 in which the top 25 constituents of 3 types (1. perfumes, 2. household products, 3. soaps) of 400 commercial products on the US market had been determined. In a pilot study on a total of 1069 patients in 11 centres, the appropriate test concentration and vehicle were examined. For most fragrances, 1% and 5% were chosen, and petrolatum proved to be the best vehicle in comparison to isopropyl myristate and diethyl phthalate. In the main study, a set of 5 to 10 fragrances at 2 concentrations was patch tested in each centre on a minimum of 100 consecutive patients seen in the patch test clinic. These patients were also patch tested to a standard series with the 8% fragrance mix (FM) and its 8 constituents. In patients with a positive reaction to any of the 48 FF, a careful history with regard to past or present reactions to perfumed products was taken. A total of 1323 patients were tested in 11 centres. The 8% FM was positive in 89 patients (8.3% of 1072 patients). Allergic reactions to the constituents were most frequent to oak moss (24), isoeugenol (20), eugenol (13), cinnamic aldehyde (10) and geraniol (8). Reactions read as allergic on day 3/4 were observed only 10X to 7 materials of the new series (Iso E Super (2), Lyral (3), Cyclacet (1), DMBCA (1), Vertofix (1), citronellol (1) and amyl salicylate (1)). The remaining 41 fragrances were negative. 28 irritant or doubtful reactions on day 3/4 were observed to a total of 19 FF materials (more than 1 reaction: 5% citronellol (2), 1% amyl salicylate (2), 1% isononyl acetate (3), 0.1% musk xylol (2), 1% citral (2), and 1% ionone beta (2)). Clinical relevance of positive reactions to any of the FF series was not proved in a single case. This included the 4 reactions in patients who were negative to the 8% FM. In conclusion, the top 25 fragrances commonly found in various products caused few reactions in dermatological patients and these few appeared to be clinically irrelevant, with the possible exception of Lyral. However, this data should be interpreted in the light of the relatively small number of patients tested (only 100 in most centres).