D ISCUSSION of the problems inherent in classifying archival material has traditionally centered around an assumption that all documents must be arranged in groups corresponding to the administrative units of the government which created or last made use of them. This theory requires that, before any documents be examined, minute research establish the history of all the departments, bureaus, divisions, sections, and other units which have at one time or another, for longer or shorter periods, constituted the government. Only thereafter, as a second step, may documentary material be examined and each file assigned to its proper organizational unit. If this method is adopted, say its advocates, the final classification of any collection of documents will faithfully reflect the administrative history of the governmental unit from which it stemmed. The organizational history thus reflected may conceivably be very interesting, but this theory overlooks the fundamental purpose of classification, whether of books or of documents, i.e., to make the material readily available to searchers. Necessarily it depends on an unwarranted assumption that the searcher knows as much about the administrative history as did the classifier on the day that he created the scheme. Of course, the very fact that this history is usually so obscure that its discovery takes months of detailed research indicates the fallacy of this hypothesis. Only rarely will the searcher have a knowledge of administrative history-or an interest in it-comparable to that of the classifier triumphantly emerging from months spent in studying it for the specific purpose in hand. Even the classifier soon forgets details and must constantly refer to his copious