The cytotoxicity and mutagenicity of several polycyclic hydrocarbons and their K-region derivatives were tested in a clone of Chinese hamster cells; the production of clones resistant to 8-azaguanine was used as the marker for mutagenesis. In the series related to benz(a)anthracene, the K-region epoxide was highly mutagenic, and the phenol was less mutagenic; the hydrocarbon and cis and trans-dihydrodiols were not mutagenic. Seven resistant clones were isolated and retained their drug-resistance; three of these could be reverted to the wild type. There was no difference in the chromosome numbers among the parent and mutant clones. The results in the methylcholanthrene series were similar to those for benz(a)anthracene. However, in the dibenz(a,h)-anthracene series, the phenol was more mutagenic than the epoxide. 7-Methylbenz(a)anthracene epoxide and 7-bromomethylbenz(a)anthracene were highly mutagenic, 7-bromomethyl-12-methylbenz(a)anthracene was less mutagenic, and the parent hydrocarbons were inactive. These results demonstrate that metabolic activation of polycyclic hydrocarbon is required for mutagenic activity in mammalian cells.In 1950 it was postulated by Boyland (1) that epoxides are intermediates in the metabolism of polycyclic hydrocarbons. Since then, epoxides have been isolated as intermediates in the microsomal hydroxylation of naphthalene (2) and dibenz(a,h)anthracene (3). Moreover, we have shown that K-region epoxides of polycyclic hydrocarbons react with DNA and proteins (4), and are much more active than the corresponding hydrocarbons, dihydrodiols and phenols at producing malignant transformation in cultures of hamster embryo and mouse prostate cells (5). Therefore, it appears highly probable that epoxides are the metabolically activated and ultimate carcinogenic form of polycyclic hydrocarbons, and in common with other ultimate carcinogens (6), are chemically reactive as electrophiles. An alternative mechanism for the metabolic activation of polycyclic hydrocarbons that contain methyl groups has been proposed, which involves carbonium ion formation on the methyl groups (7,8 Because of this chemical reactivity with DNA, it appeared that epoxides of polycyclic hydrocarbons might possess mutagenic activity. Although in the past there has been a very poor correlation between the mutagenic and carcinogenic activities of a number of potent carcinogens, it was recently demonstrated in this laboratory that the ultimate metabolically active forms of the hepatocarcinogenic aminoazo dyes and acetylaminofluorene are mutagenic to bacteriophage T4, whereas the parent compounds and proximal metabolites are not (9). Thus, it appears that one of the major limitations to the development of correlations between the processes of carcinogenesis and mutagenesis is the fact that the metabolic activation that now appears to be a prerequisite for both processes is not performed by many of the bacterial and phage systems used for the assay of mutagenic activity.Recently, considerable progress has been made in d...