Development of oral cavity cancer in man has been linked to alcohol consumption and use of tobacco products. In order to understand the underlying carcinogenic mechanisms in the oral cavity a method is needed to monitor exposure of this site to various environmental insults. In this pilot study we evaluate the use of the 32P-postlabeling assay to detect adducts in DNA from exfoliated oral mucosa cells. Exfoliated cells were collected from the cheek and tongue of 27 men aged 35-69 years. DNA was extracted from the cells and analyzed by the enhanced 32P-postlabeling technique using butanol extraction. A variety of adduct spots were detected but none was consistently associated with exposure to alcohol or tobacco products. Some of the adducts detected had migration patterns in TLC very similar to the major deoxyguanosine adducts formed by the diol epoxides of benzo[α]pyrene and 5-methylchrysene, suggesting that they may have been formed from polynuclear aromatic hydrocarbons. Adduct spots with migration patterns similar to polynuclear hydrocarbon adducts accounted for only about one third of the total adduct spots observed. Relative adduct labeling (RAL) values were determined for samples from 12 of the 27 individuals. RAL values ranged from 1.6 × 10-6 to 7.7 × 10-11 adducts per nucleotide. The RAL values for adducts from the cheek or tongue were not significantly different. Adduct levels in smokers (median RAL of 4.8 × 10-8) were significantly higher (P<0.001) than adduct levels in non-smokers (median RAL of 2.9 × 10-9). Adduct levels in drinkers (median RAL of 9.1 × 10-10) were significantly lower (P<0.001) than adduct levels in non-drinkers (median RAL of 3.7 × 10-8). Four of the subjects in this study have subsequently developed squamous cell carcinoma of the oral cavity. 32P-Postlabeling analysis of DNA from the oral cavity of these subjects did not demonstrate unique patterns or RAL values. Lack of information on the structure of the majority of adducts observed in this study was a serious limitation. Further improvements in adduct identification will be needed before 32P-postlabeling can be a useful tool for monitoring exposure of the oral cavity to carcinogens.