The administration of caffeine to the drinking water (250 mg/l and 500 mg/l) of female Sprague-Dawley rats after treatment with 7,12-dimethylbenzanthracene (DMBA) (5 mg i.g.) resulted in an increase in mammary carcinoma incidence. The confidence levels of statistical significance for the groups of rats receiving the 250-mg and 500-mg doses of caffeine were greater than 0.70 and greater than 0.99, respectively. This increase in mammary carcinoma incidence was observed when caffeine treatment was initiated commencing 3 days after DMBA treatment and continued for 21 weeks. This effect was also observed when caffeine treatment was initiated commencing 20 weeks after DMBA treatment and continued for 6 weeks in rats relatively refractory to carcinogen treatment (mammary tumor-free at onset of treatment) and in rats relatively sensitive to the carcinogen (mammary tumor bearing at onset of caffeine treatment). Caffeine treatment of rats prior to and during carcinogen treatment did not significantly affect mammary carcinoma incidence. Thus caffeine consumption has been shown to significantly enhance the promoting phase but is without effect on the initiating phase of this carcinogenic process.