Selenium has been recognized as an essential nutrient in animals since the 1950s. Demonstration of the role of dietary selenium in protection from oxidative stress followed in the early 1970s, and was largely attributed to its presence as an integral part of cellular glutathione peroxidase. However, the functions of this enzyme did not explain many of the other effects of selenium deficiency. The identification of other mammalian selenoproteins during the last few years has provided new insights into the functions of this trace nutrient. The discovery that type 1 deiodinase (D1) is a selenoenzyme, in addition to unveiling an essential role for selenium in thyroid hormone action, has had more far-reaching implications. Studies of this protein opened the door for investigation of the requirements for eukaryotic selenoprotein synthesis, and the features that distinguish this pathway from the corresponding prokaryotic pathway. Selenium is present in a number of prokaryotic and eukaryotic proteins in the form of the unusual amino acid, selenocysteine. Incorporation of selenocysteine into these proteins requires a novel translation step in which UGA specifies selenocysteine insertion. Since UGA codons are typically recognized as translation stop signals, an intriguing question is raised: How does a cell recognize and distinguish a UGA selenocysteine codon from a UGA stop codon? In this review, we will focus on what is known about selenocysteine incorporation in eukaryotes, briefly summarizing initial studies and discussing a few recent advances in our understanding of this unique "recoding" process.