The inhibitory activity of short-term feeding of one of four concentrations of dietary selenium against the induction of mammary gland carcinomas by 7,12-dimethylbenz(a)anthracene (DMBA) was studied in female Sprague-Dawley rats. When 28 days old, the animals were placed on a Torula yeast diet formulation which contained, by analysis, either 0.05, 0.15, 1.05, or 2.06 microgram of selenium, as sodium selenite, per g of diet. Mammary cancer was induced by a single p.o. administration of either 7.5 or 15.0 mg DMBA at 50 days of age. The animals were maintained on the above diets until 14 days after carcinogen treatment at which time all animals were transferred to a chow diet containing 0.21 microgram of selenium per g of diet. The study was terminated 120 days after DMBA administration. The concentrations of selenium in the liver and mammary tissue measured at the time of DMBA treatment increased with increasing levels of dietary selenium (p less than 0.05). At the low dose of DMBA, there was a trend towards reduction in the number of cancers with increased amounts of selenium, but the only significant difference occurred between groups fed the next to lowest and the highest level of selenium. At the high dose of DMBA, the number of observed cancers showed a strong dose effect (p less than 0.05). In addition, tumor load was significantly reduced in selenium-supplemented rats (p less than 0.05), and there was a significant delay (p less than 0.05) in the time to appearance of the cancers of animals receiving the highest level of selenium when compared with those receiving the lowest level. The dietary concentrations of selenium shown to inhibit the early stage(s) of cancer induction in this system were both significantly lower and fed for a shorter time interval than that which was previously reported.