Now the mechanism consists of certain apparatus necessarily independent with each library--administrative records, charging systems, etc.; but also of classification, catalog and bibliography. All of these may be standardized,--but the opportunity for a co-operation which may save expense occurs chiefly in the three last named. The extravagance, the needless extravagance, of an absence of it represented by the old conditions was little apparent to the general public or to boards of control. It becomes obvious when one considers that thousands of libraries receiving hundreds of identical books,--and hundreds of libraries receiving thousands of identical books--were each undertaking independently the expense of cataloging and classifying these: thus multiplying by exactly their number the total cost of the community. As against this, the economy of a system under which a particular book shall be cataloged--and perhaps classified--at some central point once for all, and the result made available in multiple form to all libraries receiving copies of it--needs only to be stated to be convincing. A condition of it is, in the case of classification, identity in the basic scheme and notation, in the case of catalog identity in the form, and uniformity in the practice. The general availability of bibliographic lists does not depend upon either, though convenienced by both. Identity in classification seems still remote, nor does the undoubted vogue of the Decimal scheme assure it: for this is chiefly among the smaller libraries. In the larger, the Decimal scheme, where adopted, is apt to be accompanied by variations of detail, which mean a variation in the place and symbol assigned to a particular book, and thus bar the general adoption of a decision in the classification of it made at any central bureau. So far as this variance affects the direct administration of a particular library it may be unimportant: for the arrangement of its own books upon its own shelves--provided this is based on a subject scheme, consistently carried out--may be sufficiently effective for its own purposes, even though purely individual with itself. What it implies, however, in multiplication of an expense that might be avoided by the adoption of an identical scheme, is of an import very serious. The construction of a scheme which should suit equally all libraries and all librarians is not to be expected. The best that can be hoped for is a scheme sound in its fundaments and upon which the concessions of individual preference necessary will be only as to detail. The reluctance--of librarians--to make such concessions is due, I think, to an exaggerated estimate of the importance of classification as such--that is to say, of the precise location of a particular book in a given collection; a failure to realize--what experience should have taught--that in many groups no location can be absolutely permanent, owing to changes in