This message was posted on Autocat some time ago.  Reading Cathy Wolfson's
recent message made me think it advisable to post it here as well.  This is
UCSD's official reply to LC about the series proposal, signed by Bruce Miller
since it is official correspondence, but actually drafted by me.

Crystal Graham
cgraham@ucsd.edu
-----------------------

February 14, 1994

Sarah Thomas
Director for Cataloging
Library of Congress
Washington, DC   20540

Dear Sarah:

The University of California, San Diego Libraries welcome the
opportunity to respond to your call for comments on the Library of
Congress' proposal to cease tracing series on separately classified
monographs and to cease making series authority records for those
untraced series.

I.  The need for series access.

While we acknowledge that some series have little value, there are
others which are essential for user access.  In the following
example from the Medline indexing and abstracting service, the
article is cited only by series title:

17. Sung RJ; Fan W; Huycke EC.
      Ventricular arrhythmias in the absence of organic heart
disease.
    Cardiovascular Clinics, 1992, 22(1):149-63.
      (UI:  92097160)
      Pub type:  Journal Article; Review; Review, Tutorial.

This article appears in the library catalog as a monograph in the
series Cardiovascular clinics:

TITLE        Contemporary management of ventricular arrhythmias /
             (editors) Arnold J. Greenspon, editor, Harvey L.
             Waxman, editor.
PUBLISHED    Philadelphia : F.A. Davis, c1992.
DESCRIPTION  xxiii, 368 p. : ill. ; 27 cm.
SERIES       Cardiovascular clinics ; 22/1.
NOTE         Includes bibliographical references and index.
ISBN         0803643438.
SUBJECT      Arrhythmia --Treatment.
BIOMED SUBJ  Arrhythmia.
             Cardiovascular Diseases --collected works.
             Heart Ventricle.

The series tracing is the access point which enables the user to
locate the publication containing the article sought.

II.  The need for authority control in series access.

Some members of the library community have suggested that one could
"index" series without "tracing" series, in other words, provide
access to series without bibliographic control over them.  In a
simple case like the Cardiovascular clinics cited above, such
indexing of the series statement would suffice.

In other cases, indexing uncontrolled access points gives ambiguous
results.   In UCSD's online catalog we indexed our untraced series
(primarily pre-AACR2 records) and have encountered a number of
difficulties, such as lack of collocation of variant forms, filing
on initial articles, and undifferentiated search results for series
of the same name.

A search on the title Documentos de trabajo without any
bibliographic control retrieves a hodgepodge of entries from many
series of that name:

 1   Documentos De Trabajo  ...............................   30
 2   Documentos De Trabajo 152  ...........................    1
 3   Documentos De Trabajo 167  ...........................    1
 4   Documentos De Trabajo 182  ............................   1
 5   Documentos De Trabajo 186  .............................  1
 6   Documentos De Trabajo 191  .............................  1
 7   Documentos De Trabajo 206  .............................  1

as contrasted with the orderly file retrieved when qualifiers are
used:

17   Documentos De Trabajo (Centro De Economia Internaciona     5
18   Documentos De Trabajo (Centro De Estudios De La Mujer      1
19   Documentos De Trabajo (Centro De Estudios De La Reali      1
20   Documentos De Trabajo (Centro De Estudios Del Desarrol     1
21   Documentos De Trabajo (Centro De Estudios Macroeconom      1
22   Documentos De Trabajo (Centro De Estudios Publicos) 182    1
23   Documentos De Trabajo (Centro De Estudios Publicos) 191    1
24   Documentos De Trabajo (Centro De Informaciones Y Estud     1

The absence of a filing indicator in the 490 series statement field
also hinders retrieval of series.  In our online catalog, we have
indexed series which were not traced.  A search of Libros de la
frontera in our catalog retrieves no records, because the series
was not traced (according to pre-AACR2 practice).  If searched with
the initial article, however, one finds 5 items.

1   Los Libros De La Frontera  4  ........................    1
2   Los Libros De La Frontera  9  ........................    1
3   Los Libros De La Frontera 26  ........................    1
4   Los Libros De La Frontera 27  .........................   1
5   Los Libros De La Frontera 28  .........................   1

Keyword searching has been proposed as a way to ameliorate the
inconsistencies of series transcription and the obstacle of initial
articles.  We have this feature in our online catalog and find it
helpful as an additional form of access, but all too often we find
that it yields so many results that it is difficult to ferret out
the desired entry.

For example, the heavily-used series Advances in chemistry can be
easily retrieved in our catalog through a phrase search, yielding
a result of 203 entries neatly sorted in numerical order.  A
keyword search on the same title yields 405 entries in random
order, intermixing Advances in carbocation chemistry, Advances in
carbohydrate chemistry, Advances in free-radical chemistry, etc.

III.  The need for series authority records.

The series authority record (SAR) performs a number of functions.
First of all, it gives the authoritative form of the series and
numbering caption, so that all items in a series can be collocated
in an orderly fashion.  It also provides a reference structure so
that catalog users will be directed to the established form of the
series.  For example, a citation in the INSPEC indexing and
abstracting service to the Proceedings of SPIE uses a variant form
from that on the piece:

31. CONFERENCE PAPER
    Zhang, X.X.; Bass, M.; Villaverde, A.B.; Lefaucheur, J.; and
others.
      Spectroscopy and lasing performance of a new solid state
laser crystal-Nd:GdLiF/sub 4/. (Growth, Characterization, and
Applications of Laser Host and Nonlinear Crystals II, Los Angeles,
CA, USA, 17-18 Jan. 1993).
    Proceedings of the SPIE - The International Society for Optical
    Engineering, 1993, vol.1863:2-8.
      Abstract available.

A search of Proceedings of the SPIE will not retrieve any of the
thousands of entries in the catalog, but the series authority
reference directs the user to the correct form:

     You searched for the TITLE: proceedings of the SPIE

     Proceedings of the SPIE is not used in this library's catalog;

     Proceedings of SPIE--the International Society for Optical
     Engineering is used instead

     Do you wish to search for
     Proceedings of SPIE--the International Society for Optical
     Engineering?  (y/n)

Another function of the series authority record is to record
classification decisions.  The Library of Congress proposal states
that series authority records would continue to be provided for
series and multivolume sets which are classified together,
recognizing the need to record the classification number to be used
on items in such series.  An important feature of the SAR is that
it allows libraries to record classification decisions which vary
from institution to institution and even among branches within
a single library system.   One of the most effective cost-reduction
strategies we have adopted at UCSD is to classify together
items in series, particularly working papers and other small
publications, saving the labor of assigning separate call numbers
to each one.  This also enables us to analyze selective volumes
rather than taking an all-or-nothing approach.

IV.  The need to analyze cost savings

Since LC will continue to utilize SAR's for numbered series
classified as collections, technical reports, and numbered,
analyzable multipart items, catalogers must continue to search the
SAR to determine whether a publication falls into one of those
categories.  The proposal states that the cataloger should
interpret failure to find an SAR as a directive to treat the series
as "classified separately, analyzed in full, not traced."

The failure to find an SAR may indicate that the search was not
done correctly and additional searches for variant forms will be
performed.  The absence of a series authority record could also
indicate that no decision has yet been made.  In order to determine
whether the series is untraced or whether it is new to the library,
an additional search of the bibliographic file may be required.
Careful analysis is needed to determine whether this redundant
searching will offset any savings realized by not tracing series.

V.  The need for cooperative cataloging

The Library of Congress has recently made a strong commitment to
cooperative cataloging, endorsing a Program for Cooperative
Cataloging and beginning to accept cataloging records from other
institutions for many types of materials at LC.  In contrast to
this exciting progress, the series proposal seems to disregard "the
ripple effect" of LC's policies on other libraries.

The fundamental premise of cooperative cataloging is that we will
input records according to a common standard and therefore accept
each other's cataloging as found in the shared database.  Not all
libraries will be able to follow LC practice as outlined in the
series proposal.  Libraries that still maintain card or fiche
catalogs cannot rely on keyword searching or indexing in lieu of
tracing series.  Libraries that identify specific series as crucial
to their patrons or that need to record classification decisions
must adopt variant practices.

Most of the nation's libraries use OCLC, which historically has
mandated adherence to LC's rule interpretations.  If OCLC requires
us to follow LC's policy not to trace most series, we will need to
enter original records according to one standard and then edit them
locally to meet our needs.  If, as seems more likely, OCLC allows
libraries to make individual treatment decisions, the series on a
single record will be edited locally time and again in libraries
throughout the country.  The redundant searching described above to
determine whether the absence of an SAR in the national authority
file indicates faulty searching, a new series, or a decision not to
trace will be repeated in scores of institutions.  Additional
confusion will be caused by the fact that NACO/NCCP participants
will continue to contribute SAR's to the national authority file,
yet the concomitant LC bibliographic records will show a different
treatment from the SAR.

A more constructive approach to cost-reduction in series processing
lies in the expansion of the series portion of the NACO program.
This should be a fundamental part of the expanded Program on
Cooperative Cataloging.  Contribution of series to NACO could be
simplified by eliminating the 670 field, which would realize
savings at LC and contributing libraries.  The bibliographic
database shared by the Program on Cooperative Cataloging will
ameliorate the need for a bibliographic citation in the record.  In
addition, dropping the requirement to include a 670 field would
enable libraries like UCSD to contribute the series already
established in our local system since they lack only that field to
meet the national standard.

We thank you for soliciting opinions from the library community
before implementing your series proposal.  We hope that these
comments will assist you in your deliberations.

Sincerely,

R. Bruce Miller
Assistant University Librarian
University of California, San Diego Libraries