Fish or vegetable oils were fed during the promotion stage of a mouse skin carcinogenesis model in order to investigate the effects of dietary fat on tumor development. Two weeks after initiation with 10 nmol dimethylbenz[a]anthracene, SENCAR mice were divided into five groups and maintained on one of the following semipurified diets containing 10% total fat and varying the type of fat: 8.5% coconut oil (CT)/1.5% corn oil (CO); 1% menhaden oil (MO)/7.5% CT/1.5% CO; 4% MO/4.5% CT/1.5% CO; 8.5% MO/1.5% CO; or 10% CO. Promotion with twice-weekly applications of 40 mg benzoyl peroxide was begun 2 weeks later and continued for 52 weeks. No statistically significant differences in kcal food consumed or body weights were observed between diet groups. Papilloma latency, incidence and yield differed among the diet groups with the group fed the 8.5% CT/1.5% CO diet having the shortest latency and highest papilloma incidence and number. In addition, carcinoma latency and incidence was assessed and the first carcinoma appeared in the group fed 8.5% CT/1.5% CO after 20 weeks of benzoyl peroxide treatment; this group yielded the highest carcinoma incidence throughout the study. In comparison, the group fed the 10% CO diet had the longest latency period, and among the lowest papilloma and carcinoma incidence and fewest tumors. In parallel studies, ornithine decarboxylase activity, vascular permeability and hyperplasia were elevated in the epidermis of benzoyl peroxide-treated mice but the extent of the response did not correlate with the different rates of tumor formation observed among the diet groups. These data indicate that dietary fat modulates tumor promotion by benzoyl peroxide in this skin carcinogenesis model with the predominantly saturated fat diet producing the highest rates of papilloma and carcinogen formation and the polyunsaturated fat diet the lowest.