Selenium toxicity was first confirmed in 1933 to occur in livestock that consumed plants of the genus Astragalus, Xylorrhiza, Oonopsis, and Stanleya in the western regions of the United States. In 1957 selenium was identified as an essential nutrient for laboratory rats and soon thereafter for chickens and sheep. Essentiality for mammalian species was established in 1973 with the discovery that the enzyme glutathione peroxidase contained selenium. During this same period of time, human epidemiological evidence suggested that selenium possessed anticarcinogenic effects. Since the 1970s, many animal studies have confirmed the human epidemiologic evidence that selenium compounds possess carcinostatic activity. Less progress has been made in explaining why many of these compounds of selenium are toxic and why these same compounds are carcinostatic. In 1988 the observation was made that oxidation of glutathione by selenite produced superoxide, opening a new area for selenium research. This present paper, drawing information from the literature on selenium metabolism in plants and animals, selenium toxicology, selenium cytotoxicity, and selenium carcinostatic activity in animals over the last sixty years, sets forth a probable biochemical catalytic mechanism that encompasses both selenium toxicity and selenium carcinostatic activity. The thesis presented here for scrutiny is that compounds of selenium are toxic owing to their prooxidant catalytic activity to produce superoxide (O2.-), hydrogen peroxide, and very likely other cascading oxyradicals. The toxicity of selenium compounds is countered by plant and animal methylation reactions and antioxidant defenses. As carcinostasis is mostly known to occur at supranutritional levels of selenium in animals, carcinostasis appears to be directly correlated to selenium toxicity. The catalytic toxic selenium specie appears to be the metabolic selenide (RSe-) anion.