A canal in the lower lip to function as a reservoir for snuff was surgically created in 150 male Sprague-Dawley rats. The animals were randomized into five groups of 30 each: Group I received snuff twice a day, 5 days a wk; Group II was painted with propylene glycol (solvent control) on the hard palate 3 times a wk during 4 wk; Group III underwent painting on the hard palate with 4-nitroquinoline N-oxide (4-NQO) dissolved in propylene glycol, 3 times a wk for 4 wk; Group IV received 4-NQO as in Group III followed by snuff application as in Group I; and Group V received a cotton pellet dipped in saline twice a day, 5 days a wk. Treatment continued for up to 108 wk. There was no significant difference in mean survival time between the groups. Squamous cell tumors of the lip, oral and nasal cavities, esophagus, and forestomach were seen only in Groups I, III, and IV. Nine tumors of these organs were found in Group I (six carcinomas and three papillomas), nine in Group III (seven carcinomas and two papillomas), and ten in Group IV (eight carcinomas and two papillomas). The difference between each of these groups and the control groups (II and V) with regard to tumor incidence is statistically significant (P less than 0.05). In Group I, four oral cavity or lip carcinomas were found in 29 rats, a significant difference in relation to control rats (P less than 0.05). In addition, hyperplastic lesions of the lip, palate, and forestomach were significantly more common in Groups I and IV compared with Groups II, III, and V. The study has shown that snuff and 4-NQO by themselves have the potential to induce malignant tumors. Initiation with 4-NQO followed by snuff did not significantly enhance tumor formation.