Like most libraries, that of Princeton University has expanded at definite periods, usually in connection with new building facilities, and a reorganization of administrative methods. The first reorganization epoch was connected with the provision of the ample room in Nassau Hall in 1755.

The second impulse was on the accession of President Witherspoon and involved a large increase and reorganization, but was not associated with a new building.

The third cataclysm, which came with the restoration of the Library after burning in 1802, was associated with a new building, a great increase of books and reorganization in cataloguing.

 
On the rebuilding of Nassau Hall in 1855-6, the old chapel in Nassau Hall was made into a greatly improved library room, and the return of the books was followed a few years later by a reorganization effected by Mr. Harris and by Professor Cameron, who introduced the classification of books on shelves. Professor Cameron was especially thanked for this work of reorganization by the Trustees in 1866.

About the time of the accession of President McCosh in 1868, a series of generous gifts from the Green family began. At this time the Library numbered about 14,000 volumes, but by these gifts it grew rapidly until in 1873 it numbered more than 20,000.

At this time, also by the generosity of the Green family, the Chancellor Green Library building was erected, shelved, nominally for 100,000 volumes, and actually accommodating about 80,000. At this time the first permanent exclusive Librarian was appointed in Mr. Frederick A. Vinton, who had had large experience at the Boston Public Library and at the Library of Congress, and who proceeded at once to the removal of the books to the new building, and to reorganization.

The Library was reclassified and a shelf list and subject catalogue made, the latter being completed and ready for printing in 1884. At that time the Library was said to number 60,000 volumes. Mr. Vinton introduced the card catalogue system, having both an author and an alphabetical subject list. By the time the subject catalogue was printed in 1884, the Library was already overcrowded, and the crowding went on until in 1888. Mr. Vinton was obliged to report the Library to the Trustees as overcrowded, have shelving erected in the cellar, and a considerable number of the less used books moved thither. Moreover, the books had been placed in fixed location, and the attempt to utilize the shelves under these circumstances resulted in a good deal of disturbance of the classified order and a good deal of alteration in shelf arrangement, so that the shelf list had been much disturbed.

On Mr. Vinton's death in the winter of 1889-90, still farther adjustments of books were made, and on the accession of the present Librarian in 1890, the problem of reorganization had become pressing, but extremely difficult in a crowded building with growing collection. The first step in this reorganization, undertaken at once, but only brought to a climax this year, was, naturally, to bring the collection strictly into accord with the shelf list. This shelf list was then made the basis for an accessions catalogue, a consecutive number being put on the books included in the list, while those not shelf-listed, to the number of about 20,000, were written up in regular accessions catalogues. The various departments of periodicals, etc., having been meanwhile adjusted to the enlarging problem of card catalogue was next attacked. The introduction of the standard card in place of a longer printed portion of the subject catalogue made it necessary to consult four alphabets in order to exhaust a subject reference and two to exhaust an author reference. By cutting and pasting and some copying, the four subject alphabets were first reduced to one, and later by the rude process of cutting the ends off the long cards and interlining the cut-off words, the author catalogue was brought into the same state.

In the meantime, the Library was growing rapidly. The accession had increased from 1,200 to 2,500 annually, and the Library, already crowded for ten years, was becoming more and more crowded. The cellar was nearly full of books, the reading-room accommodations had been reduced to nothing, and administration was located wherever it could find a clearing, when a generous friend, through the agency of M. Taylor Pyne, Esq., provided the sum of $600,000 for a new Library building.

This new building is a hollow quadrangle, 160x155 feet square, connected with the Chancellor Green Library by the delivery room, 20x50 feet. It is in English collegiate style of the fifteenth century, and allows of indefinite light and indefinite extension. It is equipped with the latest systems of heat, light and ventilation, with telephone, electric elevators, etc., and contains delivery room, stack room for one million volumes (of which one-half is now shelved), and about forty smaller rooms, ten for administration, sixteen for seminar work, etc. Moreover, the new building released the Chancellor Green building for reference work. The whole building, including the Chancellor Green Library, fully shelved has a capacity, i.e., shelf space for 1,250,000 volumes, which means, of course, to the Librarian, that it will probably be comfortably workable up to seven or eight thousand volumes. In brief, the new building affords every facility for proper storage, administration and use. It was begun in 1896 and finished in 1897.

Under the crowded condition of the old building, reclassification had been impossible. At the same time, it had become annually more and more necessary. The books having been arranged in a fixed location, classification had, through overcrowding, become pretty well annihilated. But while direct work had been impossible, lines had been laid for such work by the preparation meantime of a new card shelf list of the whole Library, with the exception of the (say) 30,000 unclassified books. With the abundant room in the new Library building, reclassification became at once possible, but was being made annually more difficult by reason of the avalanche of books which the new building brought with it. In the year of completion and transfer (1897-8) the accession rose to 5,000 books and 1,000 substantial pamphlets. The next year it numbered 12,000 volumes and 6,000 pamphlets. This year there have been added in eight months 23,000 numbers (including dissertations but not pamphlets) requiring regular cataloguing and classification.

In the spring of 1899, the removal having been fully accomplished and ordinary adjustments made, tentative experiments were made in the matter of reclassification, beginning with the Classical Seminary, and a method was developed. The cataloguing since 1890 had been substantially according to the A. L. A. rules in the Library School interpretation and according to the ordinary methods of the modern cataloguing, but of the books added before that time, though many were catalogued in an excellent, if not uniform, method, there were 20,000 or 30,000 volumes, old and new, which were not catalogued at all. The question was first considered whether it might not be best at once to recatalogue as well as reclassify the entire collection, but apart form the fact that there was no money in sight for either, it was decided that the work could be done more quickly and on the whole more thoroughly, with less disturbance to ordinary use, by making the problems of reclassification and recataloguing or catalogue revision entirely distinct. It was estimated that the interest on the $50,000 which would be required for complete recataloguing would itself nearly pay for the work of reclassification, and far more than pay for all the title cataloguing--more than 100,000 cards at the least estimate--which would have to be done in the process of reclassification and verification.

The method adopted contemplated, (1) a complete reclassification and card shelf listing of every accessioned book in the Library; (2) the providing of a new author card and at least one subject card for every book (about 30,000 in all) not hitherto so treated; (3) the changing of numbers on all author and subject catalogue cards; (4) the comparisons of these cards with one another and the book to ensure, first, that all entries are correct, chiefly that the main entry shall be identical in shelf, author and subject catalogue; second, that each card shall bear the accessions number, and third, that the main author card shall have on the back a list of all subjects under which subject cards are provided.

The method having been worked out in the spring, and the ground well cleared, tentative work was begun, nominally the first of August, on the basis of $1,000 given by a friend of the Library. At the meeting of the Trustees in October, it was reported by the Librarians as practicable, "with $5,000, (1) to entirely reclassify all the present working portions of the Library, (2) to make as much progress as possible with those portions for which no cards had ever been written." If this amount of money were obtained they proposed to push the work "at the rate of 20,000 volumes per month until the essential parts are done, when the rate may be slackened and the rest of the year given to completing and perfecting the work done." This estimate, it will be noted, did not make provision for the uncarded portions of the Library, nor for the large amount of cataloguing required for the greatly increased additions to the Library. Through the active exertions of Messrs. Pyne, Green, Morgan, and other, this amount of $5,000 was soon placed at the disposal of the Librarians, and on the 15^th of November the work which had been going on at the rate of about 8,000 volumes per month was increased to the 20,000 a month, a rate which was maintained for three months, when the rate was reduced to 10,000 per month. The 100,000 mark was passed on March 15^th, and at the time of writing (April 20^th) 115,000 volumes had been done. Although all the cards for these have not had their numbers changed, yet, on the other hand, the original promise to the Trustees involved only about 100,000 volumes classified, no books catalogued that there were then uncarded, and exclusively paper labels; whereas, up to the present time, there have already been done 15,000 volumes more than promised, most of which required carding as well as classification, and about 3,000 have had numbers gilded instead of tagged on. If the problem had stood still it would be now within ten thousand volumes of complete solution. Meantime, however, the number of articles received and calling for complete cataloguing and classification has been, including a collection of 17,000 dissertations, nearly 25,000. While these fall under a different head and are not involved in the original undertaking, a strong effort will be made to bring these into complete organization by August 1. As a matter of fact, nearly all except the dissertations have been already carded and classified and cards have been prepared for the dissertations. There is every indication, therefore, that before the expiration of the time set and the exhaustion of the money, there will have been completely classified and provided with shelf, author and at least one subject card, not less than 130,000 and perhaps 150,000 volumes, of which 30,000 (or 50,000) have been treated completely, except for accessioning.

In brief, the work has already gone so far beyond the minimum promised as to give hope of the maximum hoped for, but any eagerness to make a record with the new material will not be allowed to interfere with the completion of the cards, the elimination of conflicts and the general rounding out of the work on the (say) 127,000 volumes of the original problem.

The particular technical features of the operations have been (1) careful preparation beforehand; (2) strict routine with division labor, rigid subordination of function, and careful assignment of task; (3) the discardal of minute and scrupulous carding for the rough title card.

The routine is as follows: (1) Classification. In each group the work is done by the chief Librarian himself until the special assistant assigned tot he class is familiar with the interpretation of its divisions. It is found that one classifier thoroughly familiar with the idea of the system can number five to six and even ten times as many books as the average well-trained assistant. In the work of classification, the classifier reads off to an assistant the number, and this is written in in lead pencil in the inside front cover of the book. A rapid classifier, in work without snags, can keep two assistants writing in steadily, and can classify two to three hundred volumes per hour. (2) Second numbering of volumes. This is done by a corps of skilled workers, largely library school graduates, and consists in assigning the author number, date number, locality number, individual work number or biography number, as the case may be. It is, in brief, the complete number following the class number, and is written in in lead pencil like the first. (3) Preparation of shelf list card either by withdrawing from the cards already done or writing a new one--in the latter case the work being done by a tolerably well trained cataloguer. (4) Blank-labeling--blank tag on outside and on upper left corner inside cover. (5) The crossing off of old and writing on of new numbers in ink, (a) on the card, (b) on inside tag, (c) on outside tag; also writing on of accessions number on back of card. (6) Verification. This is done by a corps which includes only those who are expert in rules of entry. The card is compared with the title, with inside and outside numbering and with accession number, and is withdrawn from book, which is then (7) turned over to the boys who return to the shelves or take for gilding number on, as the case may be. (8) Cards are now alphabetized and the corresponding card withdrawn from the alphabetical author catalogue. Cards are then compared, all the entries are made identical, the author cards returned to their places, while the shelf card are arranged as shelf list. (9) Changing numbers on subject cards. A considerable amount of this work was done by combining with author cards and doing all together, but the method was found uneconomical and unnecessarily disturbing to use, so that the remaining subject cards will be re-arranged alphabetically by authors and the changes made by comparison with the author cards. In all cases where there are not author cards or subject cards, careful copies in disjointed library hand are made of the shelf card, and suitable subjects assigned.

This completes the process, and result being the books classified, numbered inside and out, arranged on shelves and provided with (a) shelf list card, (b) author card, (c) subject card;--these cards having passed one or more times under the careful inspection of skilled entry cataloguers and containing the essentials of description. Although very little analyzing or cross-referring is done, except where the cards were previously provided, or where the entries are essential, the cataloguing is complete in a sense and self-consistent. The verification feature of the routine is one on which great stress is laid, and however rigidly the work may have been done, no book goes to the shelves without having been inspected by a supposedly high class assistant. This by no means prevents mistakes, and there will be no doubt many errors to be eliminated, but the various cross-checking elements of the method are such that there have at least been eliminated a great many previous as well as current errors, and the net result has actually many of the characteristics of a catalogue revision. There are at least gained besides reclassification and besides the actual provision of say 122,000 new cards, (1) Unity of entry, so that the same card begins with the same words in every place, (2) The subjects are now written on the backs of each author card. It is, therefore, possible at any time to take a class and assemble all author and subject cards of each book as it is revised and so carry on the complete work of revision in the most approved style without disturbing at all the routine of use. The chief Librarian is personally of the opinion, although not committing his colleagues or anyone else to the doctrine, that this careful revision should not be done at present, as this simple cataloguing answers most purposes perfectly well, and that we should wait for the perfecting of some scheme of cooperation cataloguing such as is now being considered by the Cooperation Committee of the A. L. A. and such as must come sooner or later and gradually replace our rough cards by these cards.

The chief technical lesson of the work is that any library may be put into complete orderly business form, provided with a triple catalogue at a cost not exceeding 7 or 8 cents per volumes. This result is not the same as that of careful cataloguing costing six or seven times as much, but for the net advantage of average use will yield at least 95 percent, in our own case probably 98 percent of actual efficiency over the other. In this case it is probable that by this simple method costing $6,000, we shall get 98 percent of the value of what would have cost $50,000, and for a longer time. The question raised is this: Is it worthwhile to let a library wait for five, ten, or twenty years, in an unfinished condition, waiting for money enough to do the thing on a complete scale, with all the incidental disturbance of use for so long a period, when it can be put into shape in a short time and with small means by simply regarding the work as invoice work and not as perfected cataloguing?