The study of gamma-glutamyl transpeptidase (GGT) activity in transformed tissues has conventionally involved the killing of an animal or excision of the lesion, preventing longitudinal study. Oral exfoliative cytology allows longitudinal examination; however, it has been shown to be an unreliable criterion of malignancy. GGT activity has been demonstrated histochemically to involve the full thickness of the epithelium, including the keratin layer during carcinogenesis of hamster buccal pouch epithelium. This study correlates the GGT-stained foci in tissue sections and the proportion of GGT-positive cells in a superficial smear during a 13-week regime of tri-weekly topical application of 7,12-dimethylbenz[a]-anthracene (DMBA) in mineral oil. GGT-positive cells were detected in smears 3 weeks after application of the carcinogen, coincident with GGT-positive foci in tissue sections involving the keratin layer. The proportion of GGT-positive cells in each experimental period increased during the first 7 weeks of the experiment and plateaued thereafter. The number of GGT-positive foci in tissue sections in each experimental period also increased during the experiment. GGT activity was not detected in either smears or tissue sections of untreated or mineral oil treated mucosa. The correlation between the proportion of GGT-positive cells in smears and GGT-positive foci in tissue sections suggest the possibility of studying the GGT activity in an experimentally induced lesion without its elimination. Furthermore, oral exfoliative cytology using GGT staining may be useful in detecting precancerous lesions clinically.