Harvard Wiki has been integrated with Group Services.
Wiki administrators: visit IT Help for an overview of the changes to managing groups in your wikis.
Skip to end of metadata
Go to start of metadata

Originals vs. Reproductions

It often happens that Houghton will hold an original copy of an item for which another Harvard library owns a microform. Per Harvard policy (see Harvard University Library Bibliographic Standards), all formats generally appear on a single bibliographic record (which describes the original print version), with information about the physical characteristics of the reproduction recorded in the 3xx fields in the appropriate holdings record. [Though see Section XI.6 (Paperbacks and Book Club Editions) for treatment of paperbacks and book club editions.] However, when the card catalog was converted, many of Houghton's original printed editions were added to records for microforms, and many duplicate records for microforms were entered in HOLLIS. In addition, HOLLIS now contains an ever-growing number of records for networked resources, such as EEBO. When upgrading these bibliographic records, delete duplicates where possible and edit the bibliographic record to reflect the print original. If the maintenance is onerous or impossible (in the case of Holdings records of many other libraries) a Trigger can be sent to LTS. Some Houghton holdings records will have "papercopy" in an 866 field---this field should be deleted.

Duplicate Records

Many duplicate records appear in HOLLIS as a result of the batch-loading of series, microfilms, and electronic records, as well as many other causes. 

Existing holdings records may be relinked to different bibliographic records only if there is no item or order records attached (see Moving Records).

Bibliographic records with item records attached (with or without a holdings record) should not be altered, but, instead, they should be referred to the Head of Technical Services for resolution.

NOTE: To see whether a bibliographic record has an item or order record attached to it, go into the search module, highlight an individual record and go to the left-most pane and change from Functional to Overview. The entire set of links attached to the bibliographic record (holdings, items, and orders) will then appear below.

If more sorting of multiple records or if the issue involves a library for which we do not have editing privileges, the problem and suggested resolution can either be submitted using the Trigger feature in Aleph or forwarded to OIS' Database Management Team by using the Aleph Support Center. 

Duplicate bibliographic records may be created in HOLLIS if useful for specific purposes. An example: the stub records created for Hyde Collection books prior to their arrival. Normally such records will be given an STA field with value SUPPRESSED so that they will not display in the OPAC.

For more on the cleanup of duplicate records, see Resolving Duplicate Records in HOLLIS.

Duplicate Copies

Houghton's collections are full of "duplicate" copies which vary in some copy-specific or bibliographically significant way. In the past, Houghton actively deaccessioned genuine duplicates, either on its own or in cooperation with the former Gifts & Exchange unit of HCL. Duplicate copies are not being deaccessioned at present. Catalogers have multiple options when it comes to classifying duplicate items:

  • Add as a copy B, C, etc. (and designating the first as copy A)*
  • Add to a different classification within a given scheme (e.g. a work of Tasso translated by Thomas Kyd and owned by Wordsworth might be classed variously for Tasso, Kyd, or Wordsworth – though keep in mind that an item accessioned for P&GA or Theatre should not be classed in the general Houghton scheme)
  • Leave under old accession number or Widener number

It is acceptable to not reclassify duplicate copies, though accuracy in all holdings needs verifying. (This is in direct contrast with earlier practice, which emphasized getting every item into a Houghton class.) The bottom line is to try to avoid getting bogged down in systematizing classification if it does not improve accessibility.

Given Houghton's serious space constraints, duplicates should be sent to HD when appropriate. In all cases, potential "duplicates" should be compared directly with other copies and appropriate notes added to the record as to issue/state, completeness, provenance, and condition.

*A time-saving option is to not mark the first copy as A, but just mark B, C, etc. copies. In this case (A) would not be in the holdings record, but (B), (C), would be. This option should be used with discretion, since for many books, it is essential to retrieve copies that appear to be duplicates.

Multiple Printings and Impressions

When cataloging modern books, the general advice is not to make multiple records for each printing and/or impression of a book. If the copy in hand belongs to a later printing/impression (as per notice on the t.p. verso), include this information in a 562 note in the holdings record, with a note in the bibliographic record if warranted. All printings/impressions should be added to the same bibliographic record unless another aspect makes it desirable to create separate record(s).

Both ALA's Difference between, Changes within and OCLC's policy on when to make a new record  are important guides when deciding on whether or not to make a new record. The latter says that "variation in printing, manufacture or distribution date alone" does not justify a new record. However, Harvard's response to the first says: "A different date of publication, distribution, etc., including a copyright or printing date used as a substitute for the publication dates, and including an inferred date, is MAJOR: CREATE NEW RECORD.” It is simplest to follow this dictum. This is especially the case when the gap between the date on the title-page and the date of printing are more than a few years apart. When there are no changes to the text itself, the decision to create a new record may involve cataloger’s or curatorial judgment. DCRM:B's Appendix E. Variations requiring a new record provides an even more nuanced approach.

Undifferentiated names                           

In general, catalogers will follow RDA rules and do their best not to use undifferentiated names, with the exception of any local headings (i.e. those that have a subfield 5 at the end), which can be left undifferentiated. This occurs most frequently with former owners. Questions about individual problematic headings can be addressed to the RBT for problem-solving.