LC's memorandum is being posted in two parts, due to its length. --ed.
______________

Date: Mon, 07 Nov 1994 17:30:55 -0400 (EDT)
From: Sarah Thomas <stho@LOC.GOV>
Subject: LC Decision on Series Authorities

UNITED STATES GOVERNMENT
Memorandum
                                              LIBRARY OF CONGRESS

                                         DATE:   November 4, 1994

TO      : Everyone with a Stake in the Series Authorities Decision

FROM    : Sarah E. Thomas
          Director for Cataloging

SUBJECT : Series Authorities

          The Library of Congress Cataloging Directorate proposed in
1993 to eliminate series added entries and series authorities.  As a
result of over a year's analysis of the issue, including the broad
solicitation of comments from libraries and librarians and review of
recommendations of the Cooperative Cataloging Council's Series Task
Group, the Library of Congress will continue to provide added entries
and series authorities.  The attached paper formulated by Barbara
Tillett, Chief, Cataloging Policy and Support Office, provides details
on the decision and recommendations that the Cataloging Directorate
Management Team has endorsed.

Series Description and Access Decision
October 18, 1994, revised Oct. 28, 1994
Prepared by Barbara B. Tillett, Chief, Cataloging Policy and
Support Office, Library of Congress

      The much awaited report of the Cooperative Cataloging
Council's Task Group on Series Authority Records is in.  The clear
message from this Task Group and from the majority of respondents
to LC's proposal for eliminating some categories of series added
entries and series authority work is that
      * numbered and unnumbered series are important to many types
of users and we must not diminish controlled access and that
      * we must continue to expand the cooperative creation and
maintenance of series authority records to reduce the cumulative
workload for catalogers everywhere.

There also is a clear directive to continue the economic savings of
having the Library of Congress (and cooperating libraries) create
bibliographic and authority records once for all other libraries to
use.

      At the Library of Congress, we have an imperative to increase
the efficiency of cataloging operations and a mandate both to
eliminate arrearages and to stay current with new receipts.  We
seek to improve ways of doing cataloging to increase productivity
while maintaining accuracy and improving timeliness of record
creation and maintenance, and also to reduce complexity wherever
possible.  For the Library of Congress to simplify, it must
vastly improve its automated support of cataloging operations for
internal productivity increases, continue to pursue the Program for
Cooperative Cataloging to increase the availability of authority
records for series for catalogers everywhere, and continue the
current process of reviewing workflow and rewriting policies to aid
catalogers with series decisions.

Background

      At least as far back as the mid-1800's debates between
Panizzi and the Trustees of the British Museum, library managers
have called for cataloging simplification to speed up the
cataloging process and reduce costs.  In August 1993, the Library
of Congress embarked on an investigation of the impact of
eliminating certain categories of series authorities work and
series added entries.  The premise behind the analysis was that
catalogers would realize considerable savings in the creation of
bibliographic and authority work without substantial loss of
access.  LC's Cataloging Management Team directed a Series Group to
Propose implementation plans.  By December 1, 1993 the Series Group
submitted its recommendations, summarized as follows:

      1. Series will not be traced on analytic records of
unnumbered series, unnumbered multipart items, and numbered series
classified separately; the series information, transcribed
as found on the publication, will be in a 490 field with a first
indicator of "0."  Series will be traced on analytic records of
numbered series classified as collections; technical reports; and
numbered, analyzable multipart items, since in these instances
series authority records will already exist -- see #3 below.

      2. No series authority records will be made for unnumbered
series, unnumbered multipart items, numbered series classified
separately, and series-like phrases.  Existing series
authority records (SARs) for these categories will not be regarded
or updated.

      3. Series authority records will be made for numbered series
classified as collections, technical reports, and numbered,
analyzable multipart items.   Existing SARs for these
categories will be updated.

      4. Two additions could be made to the analytic bibliographic
records to increase access to the series information and to
compensate for the loss of information given now in SARs: statement
of responsibility associated with the series; other title
information especially when the other title information could be
mistaken for the title of the series or when the
statement of responsibility is included in the other title
information.

      5. LC catalogers/technicians and LC serial catalogers would
accept series added entries in other libraries' records (copy
cataloging records, NCCP records, CONSER records).  NCCP and CONSER
participants would have to make their own policy decisions
concerning series added entries and series authority records.
NACO/NCCP participants may continue to contribute SARs to the NAMES
file.

      6. LC should continue to provide documentation for the full
range of series activities to both cooperative partners and other
libraries.

      Following LC's distribution of the Series Group report LC
received responses from a wide variety of academic, public, and
special libraries, vendors, bibliographic utilities, and
individuals in the United States, Canada, and England.  The vast
majority (72%) of responses were against LC's proposal.  It was
felt that LC's proposed change was directly in conflict
with the cooperative philosophy being fostered for the Program for
Cooperative Cataloging, and would shift the cost burden to the
local level for providing controlled series access.
Respondents made it clear that acquisitions and collection
development activities would be adversely impacted, that keyword
searching of uncontrolled series statements would not be a
substitute for controlled access, and that some areas would be
particularly hard hit by the loss of controlled series access for
unnumbered series.  There are still thousands of libraries
throughout the world that rely on our cataloging and do not have
automated catalogs (OPACS) for bibliographic retrieval, so we must
not abandon controlled access for series.

      In considering compromises, it was pointed out that we, LC
and the nation's libraries, could save time and effort if we
eliminated control over unnumbered series.  This is very
appealing to those focusing only on reducing the cataloging
workload.  Approximately 40% of LC's series authority records are
for unnumbered series, and about 13% of bib records
created each year have unnumbered series (Fig. 1), so there would
definitely be some savings of LC's cataloger's time were we to
discontinue controlled access and stop making series
authority records for these.

      From the cataloger's view, it would be clear that when there
was no numbering, there would be no controlled access (no tracing)
and no series authority record would be made. For series with a mix
of numbered and unnumbered items, we would provide controlled
access only to the numbered items.  Series-like phrases that were
unnumbered would be treated exactly the same as other series, so
when numbering was present, there would be a series authority
record to indicate treatment, but otherwise not.

      However, this would mean that the nation's catalogers would
need to rely even more on cooperating libraries other than LC for
shared resource authority records for unnumbered series.  Many
libraries already have their own series authority files to indicate
local decisions that vary with LC's decisions, and by relying on
the cooperating libraries, they would not necessarily have to
individually establish the authority record for themselves.  This
past year the NACO libraries contributed 23% of all series
authority records added to the national series authority file.
This reflects a 144% increase in contributions from NACO
participants, from 1,281 series authority records in FY93 to 3,216
in FY94, and there is every indication they will continue to
increase their percentage of shared resource authority records for
series.  Why should LC bow out of this activity when other
libraries continue to feel it is important?

      It was also pointed out that in the area of access, LC's
online systems (and some online catalogs) can search on the
uncontrolled descriptive field (490 0) either as a text string
or as keywords.  However, such retrieval may result in confusing
displays, because unnumbered series titles often consist of
nondistinctive words or are forms that would be in conflict with
the controlled series yet be displayed together with them.
Searching for the list of items sharing a given unnumbered series,
would require the searcher to know the variant titles used over the
history of that series, which would involve multiple searches.
This puts a greater burden on many searchers rather than taking
care of the control once by a cataloger. It was also pointed out in
responses to LC's proposal on series, that not all libraries have
indexed the 490 fields in their online catalogs, and many more
libraries do not have online catalogs at all.

      Another consideration is that there are important unnumbered
series.  Series are sometimes a shorthand for publishers to use for
their inventory, and as such are a useful mechanism for ordering
ongoing materials on topics of interest.  Providing controlled
series access in library catalogs enables bibliographers,
collection development officers, and acquisitions staff to quickly
check for duplication and to verify the holdings of titles already
selected for the library.

      The policy could be to make series authority records for only
important unnumbered series, but exceptions lead to more decision
making time, which slows down cataloging operations, which is
something we are trying to avoid in a time of emphasis on
production.  As the CCC Task Group on Series Authority Records
noted in their September 1994 report (p. [19-20]), "It seems
simpler to be consistent in requiring authority records for all
series than to introduce exceptions which will provide new decision
points for catalogers."

      The CCC Task Group on Series Authority Records also
recognized the tremendous benefits to be gained in productivity
with automation of portions of series authority work,
particularly the machine-assisted generation of a base authority
record.  This is the area that holds the most hope for
significantly reducing cataloging time and increasing
productivity.  At the Library of Congress, the current level of
automation is a serious problem that hinders operational
efficiencies.  However, work is underway, but still several
years before being realized, for record creation
      - duplicate detection for headings and records
      - machine assisted generation of base authority records
      - full call number browse capability for completing call
           numbers (one part of shelflisting)
      - item level holdings records for inventory control (the
           other part of shelflisting:
           location, volume, copy, barcode, and associated
           notation)
      - online access to essential cataloging tools for all
           catalogers

      record maintenance
      - automated serials check in
      - integrated automated system to eliminate current redundant
           work, building instead on
           shared data (bibliographic, authority, holdings
           information)
      - global update

all of which would greatly assist catalogers and improve
productivity and cataloging operations.

      At the same time as we seek computer assistance, we must
guarantee that our catalogers and the catalogers in cooperating
libraries are given the necessary foundation and continuing
training needed to sustain the high quality of the records we
create and share with the world, and that our workflows are as
efficient as we can make them.  We are concerned about the accuracy
of the information we provide and the timely availability of the
records to library patrons and other libraries (acquisitions,
catalogers, circulation, ILL, preservation, binding, remote
storage, etc.).