more, and be worse done than the preservation
and cataloging of the whole."

These remarks referring to local English
publications, apply with even greater force to
similar publications, past and present, in both
the United States and Canada.

Many state bibliographies have been pub-
lished. For the purpose of demonstrating the
importance of coordinated activity in collect-
ing and cataloging local printing, I have made
certain inquiries using as a basis three pub-
lications of this character, namely Norman
Fee's very important Catalogue of pamphlets,
journals and reports in the Public Archives of
Canada, Earl G. Swem's Bibliography of Vir-
ginia, one of the best available works of its
kind, and the catalog of the Calvin Morgan
McClung historical collection of books, pam-
phlets, manuscripts, pictures and maps relating
to early western travel and the history and
genealogy of Tennessee and other southern
states, published by the Lawson McGhee Li-
brary, Knoxville, Tennessee.

Fee's catalog contains entries for 2,931
pamphlets relating to Canadian history and
published prior to Confederation, Swem's bib-
liography, 6,981 entries relating to Virginia,
and the McClung catalog, 297 Tennessee im-
prints.

A careful checking of these 10,209 titles
against a depository catalog of Library of
Congress cards reveals that printed cards are
available for about 35 per cent.

Of the 2,931 entries in Fee's catalog about
50 per cent are to be found entered in the cata-
logs of three other important Canadian libra-
ries, although these three libraries have over
twice as many separate titles for other Cana-
dian pamphlets not found in the Archives
collection.

This reveals the difficulties in the path of
the research worker in the field of American
and Canadian history and how impossible it is
to reach an assurance that all the available
source material has been consulted.

The remedy undoubtedly lies within the
scope of well-organized and adequately sup-
ported cooperative cataloging.

When the immense field to be covered is
considered it seems quite evident that there
must be a distribution of labor. It is beyond
the capacity of any single agency, under exist-
ing conditions, to do the necessary checking

involved in the creation of a union catalog,
approximately complete for source material
in the field of American history, contained in
American libraries.

Spasmodic efforts will accomplish little
towards reaching such an objective.

Let us then consider possible means by
which a union catalog, to cover all printed
matter relating to the history of the United
States and of Canada, may be created.

I. Relation to the general subject of cooper-
ative cataloging.

It is obviously desirable that any activity in
the field of cooperative cataloging be carried
on as a unit in a general scheme. This postu-
lates uniform cataloging practice and the
mechanical production of cards uniform in
size and style. Since by far the greater
number of printed cards already available for
use in a union catalog are those issued by the
Library of Congress, few will dissent if these
cards are adopted as standard and the cata-
loging practice of this library followed with
such modification as is made necessary by an
enlargement of the field covered by the printed
card.

II. Delimitation of the field.
Although it may be worth while consider-
ing delimitation on the basis of period, subject,
or type of material, in each case there exists
the objection of overlapping. This may be
avoided by restricting the activity of each co-
operating agency within a limit determined by
the place of publication. Assuming that the
active cooperating agency in any state or
province was to be the state or provincial
library working in cooperation with the state
or provincial historical society, such an agency
would confine its efforts to books, pamphlets,
and journals published within its own state
or province. The most complete collections
relating to the history of any particular state
would naturally be expected to exist within
the borders of that state. The state libraries
are already, in a majority of the states, in
close touch with the other libraries of the
state, and with local publishing activities.
Many state historical societies possess very
fine libraries and are in a position to render
invaluable advice and encouragement. Where
it is impracticable to utilize the state library,

the central agency for a particular state may
be, instead, the library of the state university,
or possibly the principal public library.

III. Financial consideration.

Printed cards produced in the cooperative
cataloging of the local literature of history
would be utilized principally (1) in union cat-
alogs, (2) in cataloging in other libraries, and
(3) as bibliographical tools by research work-
ers. The revenue to be derived from the sale
of cards would be limited. The average edi-
tion would probably not exceed fifty copies.
It is unlikely that the sale of cards would
bring in enough even to meet the cost of cata-
loging and printing, aside from the expense
involved in their distribution. Inadequate
financial support would be fatal and this
should be assured before work is undertaken
in any state or province. Possible sources of
support include (1) government grants, (2)
subscriptions from libraries and others, (3)
subsidies from historical societies, (4) endow-
ment by educational corporations or interested
individuals, (5) contributions by publishers to
cover the expense of cataloging their own
publications, and (6) revenue derived from
the sale of cards.

IV. Procedure.

1. Material published prior to the begin-
ning of the plan in any particular state.
Preparation of copy covering items printed
in the state and not already in the Library of
Congress Union Catalog. This would include
first material in the state library and finally
material in other libraries both within and
without the state.

2. Current material. Preparation of cards
for uncopyrighted material published in the
state (assuming that copyrighted material
would be largely covered by Library of Con-
gress cards). The extensive checking neces-
sary would probably require the service of
library assistants definitely assigned to such
work. It might be necessary to send such
assistants into the field to visit the various
libraries of the state, as well as local printers,
for the purpose of gathering data concerning
material not in the state library. Some con-
tribution towards the final completion of an
accurate record of all printed matter originat-
ing in any particular state might be made by

local historians, book-collectors, and particu-
larly by graduate students at various universi-
ties, working in the field of American history
as a part of the requirements for advanced
degrees.

V. Utilization of results.

The results of cooperative enterprise, along
these lines, would be available to scholars in
the form of printed cards, filed in various
union catalogs at centers of research and in
partial sets of cards devoted to specific phases
of American history where required, also in
the form of check-lists and bibliographies,
based upon the printed cards, and to libraries
for use in cataloging their own holdings and
acquisitions and as a guide to desiderata in the
development of their collections.

VI. Unlocated Material.

Many books and pamphlets known to have
existed from references occurring in contem-
porary works, particularly in early newspapers,
are today unknown even in single copies. In
the case of newspapers it is much easier to
compile a check-list of newspapers published
in the smaller cities and villages of the
United States or Canada, than it is to locate
complete or even partial files of such pub-
lications. When such titles are discovered it
would be a valuable contribution to historical
research, if tentative entries on cards could be
prepared, thus putting such items upon per-
manent record. This would increase the pos-
sibility that copies of such missing publica-
tions might ultimately be located and acquired
by the libraries most interested.

In this paper, I have purposely refrained
from any discussion of such aspects of the
problem of cooperative cataloging as relate
to the agency or agencies that would actually
handle the printing, storage, and sale of printed
cards. Other contributors to this symposium
will no doubt consider this matter in detail.

Personal experience has shown me that any
competent printer, with a little effort and
working under the direction of competent
catalogers, can produce creditable printed
cards comparing favorably with those of the
Library of Congress. An edition of fifty
copies costs between sixty cents and eighty-
five cents, depending upon local conditions.
Should more libraries produce printed cards

within limited fields and not duplicating those of the Library of Congress, such cards might be readily added to a central stock either at the Library of Congress or elsewhere and sold upon relatively the same basis. The revenue accruing from such sales less an overhead charge made by the central agency could be placed to the credit of the producing library and periodical settlements made in cash less debits for cards purchased by the producing library for its own use. When the economy, effected by the use of printed cards in routine cataloging, is considered in comparison with the cost of independent cataloging and the production of typewritten cards, the present charges for printed cards are extremely low. A higher charge for cards produced in connection with a well-developed scheme of cooperative cataloging would be justified and might operate to make the extension of such a scheme more possible.

References.

Esdaile, Arundell. The preservation of a national literature. Library Association Record, n.s.,4:214-225:1926.

.

Fee, Norman. Catalogue of pamphlets, journals and reports in the Public Archives of Canada. 1611-1867. Ottawa, 1916. Lawson McGhee Library. Calvin Morgan McClung historical collection of books, pamphlets, manuscripts, pictures and maps relating to early western travel and the history and genealogy of Tennessee and other southern states. . Knoxville, 1921. Swem, Earl G. A bibliography of Virginia, part 1, containing the titles of books in the Virginia State Library, which relate to Virginia and Virginians, the titles of those books written by Virginians, and of those printed in Virginia. Richmond, 1916. van Patten, Nathan. Local printing. Library, Journal. January 15, 1927. Webb, W. P. (Texas local history) Quoted in "Building a rare book department," by Morris H. Briggs. Publishers' Weekly, CXI: 1634-1635:1927.

THE EXTENSION OF COOPERATIVE CATALOGING AND GENERAL LIBRARY COOPERATION

By ERNEST CUSHING RICHARDSON, Library of Congress

There has been such a surprising increase of cooperative cataloging in 30 years, such a wealth of experience gained, and such a

variety of aspects of the matter developed, that it is hard to choose what to say at this time.

The main thing to say is perhaps that it has been and is already extending at breakneck speed in several directions, that it is a question whether to put on brakes or put on more current, and in either case at what points to apply.

Probably the best approach is to review the actual demands for extension now being made.

The present active demand takes two main lines: (1) the extension of the published card system, (2) the extension of the Union Catalog. To these lines may be added, (3) cooperative subject cataloging and the closely related topic, (4) cooperative analytical cataloging. These 4 problems in cooperative cataloging are so intertwined with one another and with other lines of interlibrary cooperation; selection, purchase, storage, classification, lending, reference, information, that they can only be practically discussed as aspects of a general problem, of which published cards and union finding lists are simply the most urgent aspects.

1. The extension of the published card system. This paper has its origin in a general demand by catalogers associations for more published cards. The matter has been publicly discussed at many meetings. Various publishing expedients have been suggested and experimented with e. g., Mr. Hall's plan of lending cards for copying. The situation reminds one of the similar situation in the years just preceding the undertaking by the Library of Congress to publish its printed cards. Now, as then, the commonest idea is that the A. L. A. should do it if the L. C. cannot. The fact is that the L. C. system, although it now serves 3,900 libraries, includes 1,000,000 titles and has 90,000,000 cards in stock, has long been unable to meet the legitimate demand.

Many local libraries have therefore been forced to supplement by printing their own cards. Some have done this on a rather large scale, e. g., John Crerar, New York Public, Boston, Harvard, University of Chicago, etc. A few of these publish their cards in the sense that they sell full sets for bibliographical purposes. A very few (John Crerar) furnish individual titles for cataloging, but most decline to be bothered with the detail of marketing individual cards. The net result is that only

those large libraries which use enough copies of the card to justify printing have the benefit of printed cards and these libraries duplicate printing to a considerable extent as well as duplicate to a very large extent the most expensive types of cataloging.

In the light of the experience of nearly 30 years of card publishing what seemed an irrational situation to begin with, seems doubly

so now.

Though the point has been reached after nearly 30 years experience where there are 1,000,000 titles in stock and 33,000 or more titles are issued in a year, a vast number of titles are still cataloged independently in manuscript, in from 2 or 3 to 2 or 3 score duplications, at an average cost of from 50 cents to $2 or even $3 per title, plus an average of 3 cents per card for copying from 3 to 30 cards per library, while if introduced into the published system, even 2 or 3 libraries cooperating would produce the more legible printed card at a considerable saving. This is precisely the waste which the printed card system started to do away with nearly 30 years ago.

Incomplete as the system still is, nearly 4,000 libraries use it and consumed some 10,000,000 cards last year.

It is hard for various reasons to judge of the savings effected, but it has been roughly estimated that with something allowed for quality it now means an annual saving to American catalog budgets of $1,000,000.

The fact that the system is practically complete for current American copyright books gives a concrete demonstration of what it would mean to cataloging to have it entirely complete so that it might be used automatically for all titles.

Three ways have been suggested for organizing this, (1) simple extension of present system by the Library of Congress, (2) a supplementary organization under the direction of the A. L. A., (3) the whole matter turned over to some commercial concern like the R. R. Bowker Company or the H. W. Wilson Com

pany.

Before discussing the matter further, however, let us consider briefly the three other lines of cooperative cataloging extension, Union, Subjects, and Analytical because of the intimate relationship of the four lines to

one another and to plans for cooperation in the selection, purchase, storage, classification, lending, and information services.

This brings us to:

2. The extension of union finding lists. The gist of this problem is the need of directing library clients to the location in other libraries of books which cannot be found at home. In the older library practice it was the custom to let the reader hunt such books for himself. Modern practice is deeply concerned with helping readers to find elsewhere, if not at home. Pending better aids, this is now carried on laboriously and imperfectly by means of correspondence and the printed catalogs of other libraries, with such fragmentary union lists as now exist. The matter is simplified by lists of special collections in various libraries, but its right solution is to be found only in the extension of union cataloging of specific individual titles.

In union cataloging it is convenient to make distinction between union finding lists and union catalogs, between cumulative card lists and printed lists, between lists of special collections and lists of individual titles. The union finding list or union catalog is the indicated method for locating books for entire library visiting or borrowing. These catalogs may be distinguished into: (1) guides to groups of books, such as the Johnston Mudge, and (2) guides to individual titles. The latter may be again distinguished into union finding lists which aim to locate one or a few copies and union catalogs which include all books in a group of libraries or all copies of a given title in all libraries.

The active demand for the extension of union finding lists is shown by the actual extension of union cumulative card finding lists both in America and Europe, the increase both in the number and importance of printed union lists, the increasing demand by university libraries through correspondence on behalf of their research students, and by the formal official demand of such bodies as the Woods Hole biological conference, the American Medical librarians, and the League Committee on Intellectual Cooperation, for some means of locating copies.

Perhaps the most significant and certainly the most extensive development of union finding lists is the 56 full depository sets and 44

partial sets of cards published by the Library of Congress.

These are all contributions to the machinery of the local library in finding for its clients books in other libraries. They are strictly union lists in that they not only include (in separate alphabet) titles in the local library and titles in the Library of Congress, but even the L. C. cards include perhaps 10 per cent of material not in the Library of Congress but in other government libraries and other libraries which furnish copy for L. C. cards.

Further than this many of these depository collections now introduce large numbers of printed cards from other libraries which print these cards, John Crerar, University of Chicago, Harvard, etc. Several of these union catalogs must number towards 1,500,000 titles, while the Library of Congress Union Catalog which adds by exchange the printed cards of all printing libraries, now contains in its 2 alphabet lists twice as many. As most of these are in full bibliographical form and printed, this forms the largest bibliography in existence except the Brussels Repertory and by far the largest which is edited to strict bibliographical standards. It forms an invaluable basis for the extension of American union cataloging.

The number and size of large union catalogs abroad has also increased. The Brussels list contains in author and subject form something like 13,000,000 cards, representing perhaps 5,000,000 titles. The Frankfurt and Berlin union lists contain perhaps 3,000,000 titles each and the various local catalogs such as those of Zurich, Basle, Berne, etc., contain smaller but considerable numbers.

The printed union lists have also rapidly increased. Most of these are of rather limited subjects or forms of books. The favorite field for this is periodicals and the greatest example to date, the A. L. A. Union list of serials which covers perhaps 75,000 titles. Berlin now proposes to print its union catalog of Prussian libraries with its 3,000,000 titles.

The vigorous demand for the extension of union finding lists is not confined to simple author finding lists, but is carried to classified lists such as that of Brussels, or the more popular modern form of alphabetical subject lists. This has recently been urged by serious investigations abroad and at the Library of Congress.

The matters of cooperative subject cataloging and cooperative analytical cataloging may be dismissed very briefly in order to save time for our main point of discussion and because they are bound to be discussed more fully at some other place and time. As to:

3. The extension of cooperative subject cataloging. This is of course involved in all public card work, or should be. The subjects are printed on the cards. The subject catalog is formed automatically by writing on and alphabetizing the subjects. If therefore the supply of public cards should be carried and near completion, it would by the same token provide material for the universal subject catalog asked for by Mlle. Bonnvie in the League Committee and by many other European and American research students. A few depository sets of these would make by far the most useful and most time-saving as well as the most extensive bibliography ever gathered. Meantime, there are two shortcomings to which practical attention might be given. In the first place, the subject headings are not yet sufficiently standardized to allow of either automatic application of printed cards or the freest cooperation by other libraries in the cataloging of new titles. The greatest contribution to extension at the present moment would be therefore the carrying through of the standardizing of these subject headings as thoroughly as the cataloging rules have been standardized.

In the second place, many published cards are not provided with subject headings at all. The first extension would seem to be the systematic provision of such cards with their subject headings. It is a matter of constant embarrassment in the ordering of cards to find that they are incomplete, and it is perhaps not too much to say that it would be better not to publish than to publish before cataloging is completed by subject headings, and class numbers.

A simple first extension of union subject cataloging would be to have somewhere one subject catalog with all L. C. cards and information. There is, I believe, no subject catalog which contains even all L. C. published card titles. The Library of Congress catalog contains of course cards of all its own printed titles, but one tenth of the cards published by