Teachers of language have turned to authentic texts to improve the communic
ative competence of their students. Traditional sources, including curricul
um materials, categorical book lists, teachers' prior knowledge of texts, a
nd lists of graded texts using readability formulas, are incomplete in toda
y's technological world. Computer corpora could provide additional sources
of authentic texts. This article examines a variety of teaching implication
s in using existing corpora and explores the need to develop additional cor
pora and databases.