This thesis examines physical science textbooks and evaluates the language of science textbooks as it relates to the history and purpose of science education. It is based on the premise that such language can be characterised not only by structural factors (such as sentence length and vocabulary) but also by what the language attempts to achieve. The thesis explores the history of physical science texts since 1800, and uses that history to develop both a classification system for textbooks, and a framework against which to view the changing purposes of science textbook authors. It is based on an analysis that provides information about the development of the modern science text, uses past and present texts as mirrors of the long standing debate about the place of the textbook in science education, and provides the classification and characterisation of textbooks needed for any further rhetorical consideration of the textbook as prose.