Mutagenic activity on the Ames test was evaluated in 15 samples of naphthenic high-viscosity minerals oils and 12 samples of used lubricants (recovered and pooled) and their recycled products. Bacterial mutagenesis was assayed using both the standard technique and Blackburn's modification. The contents of polycyclic aromatic hydrocarbons (PAH) was also evaluated, as polynuclear aromatic fraction (PAF) and total PAH, determined respectively with the semi-quantitative dimethylsulphoxide-refractive index method and the Grimmer method. Only four samples (three acid-treated naphthenic oils and one recycled fraction of a used oil) showed mutagenic activity higher than 6 revertants/mg of oil, considered by Blackburn and coworkers as indicating a potential carcinogenic risk for these compounds. Limited mutagenecity was found in all used and recycled oils, but also in samples of acid- or solvent-treated oils. No hydrogen-treated naphthenic oils turned out to have any mutagenic activity. PAF contents of oils were closely correlated with those of total PAH ( n = 15, r = 0.83; n = 12, r = 0.91; p<0.01 for both naphthenic and used/recycled oils respectively). No recycled oil had high PAF contents. Eleven samples had PAF contents higher than 3%, the arbitrary danger threshold suggested by the CONCAWE (1988). Of these 11 samples, the majority were acid-treated products, although there was one hydrogen-treated oil and one used and recycled oil. No mutagenic activity could be demonstrated in almost half the oils with PAF > 3%. In this study, the presence of mutagens was not correlated either with PAF or with total or mutagenic PAH. The difficulty of predicting the mutagenecity of mineral oil is stressed. Most naphthenic and some recycled oils clearly have components which inhibit the metabolizing system in the bacterial mutagenesis test, with consequent possible false negative results.