The constancy, both temporal and spatial, of the profile of polycyclic aromatic hydrocarbons (PAHs) relative to benzo[a]pyrene (BaP) is a prerequisite to using the BaP-indicator approach in the carcinogenic risk assessment for PAHs. The principal aim of this study was to provide a contribution to validate this approach, by studying the variability of the profile at a typical urban site through a multi-year data set and by comparing the profiles available for different cities. Seven carcinogenic PAHs (benz[a]anthracene, benzo[b + j + k]fluoranthenes, BaP, indeno [1,2,3-cd]pyrene, dibenz[a,h]anthracene) were determined in PM10 24-h samples collected every third day at a road site; moreover, benzo[e]pyrene was determined as a reference PAH due to its stability. The profile was found stable from year to year. Besides, it was similar to those recently found in other European cities (observed differences within a factor of four) and to those elaborated from earlier (1970s-1980s) investigations. The substantial similarity of profiles, both temporal (on an annual basis) and spatial, supports the validity of the BaP-indicator approach. Large PAH-to-PAH differences were, however, found in the seasonal pattern of profile: they were explained by the different atmospheric degradability of PAHs, whose effect is enhanced under the meteoclimatic conditions typical of the European Mediterranean countries. PAH annual means showed a slight declining trend since 1994. In the last sampling year, mean concentration of BaP was 1.2 ng m(-3). Within-year differences among monthly averaged PAH concentrations were observed, as large as up to 44-fold for BaP, underlining the need for whole-year monitoring. (C) 1999 Elsevier Science Ltd. All rights reserved.