Chemoprotective effect of dietary selenium (as sodium selenite or as Se-rich egg) on mouse skin tumor induced by topical application of 2'-(4-nitrophenoxy)oxirane (NPO) as tumor initiator and 12-O-tetradecanoylphorbol-13-acetate (TPA) as tumor promoter was evaluated in relation to the dietary source and levels of selenium. Selenium supplementation (0.3 p.p.m.) to the basal diet (0.07 p.p.m. Se) as sodium selenite or as Se-rich egg brought about a 40 or 37% reduction respectively, in the incidence of papilloma formation at 12 weeks after NPO treatment. Tumor yield (number of papillomas per mouse) at 14 weeks after NPO treatment in the basal diet group, basal diet supplemented with 0.3 p.p.m. Se as sodium selenite group and basal diet supplemented with 0.3 p.p.m. Se as Se-rich egg group were 7.5 +/- 2.1, 2.7 +/- 2.3 and 4.1 +/- 3.5 respectively. Dietary supplementation of 1.0 p.p.m. See as Se-rich egg to the basal diet reduced the incidence and the multiplicity of papillomas during the early phase of promotion (11 weeks) but its antitumor activity decreased thereafter, indicating that the accumulation of tissue selenium above the saturated level may not be beneficial. Selenium concentrations in blood, liver and skin tissue of mice in basal diet group (0.33 +/- 0.02, 0.54 +/- 0.10 and 0.21 +/- 0.03 p.p.m. respectively) increased significantly ( P < 0.05) by the supplementation of 0.3 p.p.m. Se as selenite (0.58 +/- 0.02, 1.17 +/- 0.10 and 0.31 +/- 0.05 p.p.m. respectively), 0.3 p.p.m. Se as Se-rich egg (0.59 +/- 0.02, 1.25 +/- 0.11 and 0.33 +/- 0.06 p.p.m. respectively) and 1.0 p.p.m. Se as Se-rich egg (1.20 +/- 0.05, 2.32 +/- 0.28 and 0.51 +/- 0.01 p.p.m. respectively). Glutathione peroxidase activity in blood of mice of the basal diet group (2.4 +/- 0.4 EU/mg) increased significantly (P < 0.05) by dietary selenium supplementation (3.8 +/- 0.6 EU/mg in 0.3 p.p.m. Selenite; 3.7 +/- 0.8 EU/mg in 0.3 p.p.m. Se as Se-rich egg; 4.9 +/- 0.9 EU/mg in 1/0 p.p.m. Se as Se-rich egg) and the enzyme activities in liver and skin tissue also increased by 0.3 p.p.m. Se (as selenite or Se-rich egg) supplementation, but no further increases in their activities were obtained by 1.0 p.p.m. Se (as Se-rich egg). It is, therefore, concluded that a moderate level of dietary selenium (0.3 p.p.m.) has an efficient chemopreventive activity at the promotional stage of carcinogenesis and that dietary selenium rich egg as well as dietary selenite exerted antitumor activity.