The comparative examination of experimental and epidemiological data, together with risk evaluations concerning coke-oven, diesel, and gasoline motor emissions, as well as specific PAH fractions and single PAHs, indicates that, at least under risk additivity hypothesis, benzo(a)pyrene (B(a) may account for only a relatively small amount of the whole carcinogenic potential of PAH mixtures. This finding emerges from the comparison of risk evaluations of coke-oven and diesel emissions (based on epidemiological and/or experimental data) with risk evaluations of B(a)P as a single substance (based on carcinogenic bioassay data), as well as from the results of experimental studies evaluating the carcinogenic potency of motor exhaust condensates, of specific condensate fractions (particularly, the 4-7 ring PAH fraction), and of B(a)P. These data indicate that B(a)P might account for a percentage of the whole mixture carcinogenic response, at least in the hypothesis of an additive effect of PAHs. However, the overall carcinogenic response and risk still appear at least approximately related to the B(a)P concentration in the examined PAH mixtures. Finally, recent findings on PAH relative potencies (compared with B(a)P potency) indicate that several PAH present in mixtures commonly detected in occupational and in the general environment are characterized by a carcinogenic potency comparable with or even higher than that of B(a)P. This confirms the above hypothesis with respect to the significant, but minor role, of B(a)P. This relatively coherent picture appears to confirm that the reference to B(a)P as a complex mixture index may still represent a reasonable solution. However, the selection of a few other indicators from among the PAH whose relative concentrations and relative carcinogenic potencies are higher might improve the analysis.