Abstract Antibodies specific for polycyclic aromatic hydrocarbon (PAH)-DNA adducts have previously been reported in human sera. In this study, we examined the association between mixed PAH exposure and PAH-DNA adduct specific antibodies in BALB/c mice. Mice were treated either by i.p. injection or by intragastric (i.g.) intubation with a mixture of seven different PAHs [benzo(a)pyrene (BP), benz(a)anthracene (BA), fluoranthene (FA), dibenz(a,h)anthracene (DBA), 3-methylcholanthrene (MC), chrysene (Ch), benzo(b)fluoranthene (BF)] at three doses (0, 15, 150 μg of each PAH) twice a week for 8 weeks. Sera were screened by direct ELISA for antibodies recognizing DNA modified by diolepoxides or epoxides of each PAH injected. In i.p. treated mice, the sera were slightly more reactive to DNAs modified with diolepoxides of BP, BA, or Ch or an epoxide of DBA than to unmodified DNA. In i.g. treated mice, the sera were more reactive to DNAs modified with diolepoxides of BA or BF than to unmodified DNA. For some PAHs, a dose-response effect was observed between sera reactivity to PAH metabolites and the dose of PAH administered. However, there was considerable variation in the immune responses among individual mice within each treatment group. When tested by competitive ELISA, none of the sera could discriminate between modified and unmodified DNA. This animal study suggests that an assessment of previous carcinogen exposure by measuring DNA adduct-specific antibodies requires further validation prior to its application to the human monitoring of carcinogen exposure.