The point of view which I wish to develop' was suggested by a passage in Michel Foucault's L'arch6ologie du savoir. Foucault has proposed to replace arbitrary historical unities that are often imposed on the past with genuine unities he calls "fields of discourse." One way to delineate a field of discourse, he asserts, is to find "the play of rules" which for a given period renders possible the appearance of whatever objects the field deals with. These rules, rather than the shifting objects themselves, provide the unity of discourse.2 It occurred to me that such an approach might illumninate the formation of a field of scientific investigation that was a central concern to early nineteenth century biologists: that is, the subject of animal nutrition. After rereading part of Foucault's discussion I suspect I may have misunderstood him, and that the use I planned to make of his statements may not adequately reflect his own meaning. Nevertheless I will appropriate his words and, I hope, something partially resembling one of his ideas. I had already for some time felt that it was more fruitful to describe the emergence of this science in terms of definitions of problems than in terms of demonstrable knowledge of specific nutritional processes and of particular experimental discoveries. My earlier view, however, stressed conceptual factors more than investigative modes, reflecting perhaps the unrecognized influence on me of what was a prevalent notion a decade ago, that the history of science is the history of ideas. Foucault's statement may help to focus the problem on the sites of interplay between ideas and actions, on the rules by which operations cause "objects," or "phenomena," to show up, to be distinguishable and delimitable. My interest in applying Foucault's statement to the field of nutrition