The Collation

Research and Exploration at the Folger

24,000 “preliminary” catalog records are better than nothing!

At least, we hope the approximately 24,000 “preliminary records” added to the Folger’s online catalog yesterday are better than nothing, which is what Hamnet had for most of these books since going live in 1997. Today’s Collation post explains where this big batch of records came from, and how to navigate their perils and pitfalls if you come across them in your research. They are easily recognizable: most have an advisory statement saying “This is a PRELIMINARY RECORD copied from an old card (which is better than nothing!). Please email for assistance.” The rest have a nearly identical advisory statement explaining that the information was “copied from an old purchase order.”

Preliminary Hamnet record with the advisory statement “This is a PRELIMINARY RECORD copied from an old card (which is better than nothing!)”

Catalog cards

The vast majority of these new better-than-nothing records contain (extremely) brief information copied (with varying degrees of accuracy) from the existing card catalog by a data-entry company. Why the low bar? Because we needed to get it all done. For over twenty years, we’ve been chipping away at getting it done properly, finding grants to hire experts to update card catalog records to current international standards one-by-one, while enhancing them with additional information online. We still do want to get it all done properly, of course, we’ve just let go of the mantra of “Do it once. Do it right.” That’s appropriate for brain surgery, but people’s lives aren’t at stake here. If all you want to know is whether or not the Folger has the 1589 edition of John Sanderson’s Institutionum dialecticarum libri quatuor, the record at the start of this blog post will do fine. If you want to browse the subject heading “Dialectic” to see where it fits in with other books the Folger has on the topic, though, you’re out of luck (for now) but you’re tough, you can handle it.

Here are some of the more important characteristics of the “copied from an old card” records:

  1. Most are Continental publications dated between 1501 and 1799; the rest are 18th-century British and American publications.
  2. They don’t have “Subjects” for what the book is about. For example, they won’t come up in a subject search for “Medicine” or “Horology” even though some cover those topics.
  3. They don’t have “Genre/Form” for what the book is a representative example of.1 For example, they won’t come up in a genre/form search for “Psalters” or “Armorial bindings” even though there’s at least one of each among them.
  4. They don’t have “Associated names” (e.g. former owner, printer), only a “Main name” (usually the name of the author, with surname first; occasionally some other piece of information at the top of the card that happens to have been mistaken for a name.)
  5. Search limits available through the “Set Filters” function will mostly work:
    • Language: usually correct, but a few bad guesses (e.g. Dutch mistaken for German, English works with Latin in the title mistaken for Latin).
    • Location: usually correct, because they should all be for “Vault” items (but a few open-stacks items will have slipped through the cracks, and say “Vault” even though they’re modern reprints).
    • Date: usually correct, but a few later reprints might have the original date in the machine-readable “date” field, and anything without a publication date on the title page will sort as “unknown” (and therefore first in lists) rather than current Folger practice of including an estimated date or date range.2
    • Item type: usually correct because “Books” was used as the default, and except for a few “Serials” and a few books consisting entirely of engravings that the Folger would now catalog as “Prints”, that’s what they are.3
    • Place of Publication: hopeless. The machine-readable information wasn’t entered in the first place.4
  6. Initial articles are not ignored in searching and sorting. On the plus side, this is less jarring than it used to be: when everything was filed by hand, it was unthinkable to have The Boston Gazette, The Nightingale, and The Poetical Works of Joseph Hall all listed under “The” but now, it’s merely disappointing.5
  7. Symbols of contraction carried over from the manuscript tradition will not be expanded in square brackets (for example, “Almanach perpetuuʒ… nuper emẽdatũ omniũ celi motuum cum additionib’ in eo factis tenens complementum” instead of “Almanach perpetuu[m] … nuper eme[n]datu[m] omniu[m] celi motuum cum additionib[us] in eo factis tenens complementum”).6
  8. Superscripts in signature statements will usually be regular size and on-the-line (for example, “A-F6” instead of “A-F⁶”).
  9. Copy-specific information, such as binding material or missing leaves, will usually be in the main “Note” field, not in the “Folger copy” or “Item note”.
  10. Many are duplicates: in some cases, the book had already been cataloged online, in other cases multiple copies on the same card, or lengthy descriptions that continued across several cards, ended up getting separate records. If you find an extremely brief record with the parenthetical phrase “(card 3)” you can be pretty sure the matching “(continued on next card)” and “(card 2)” records are also in Hamnet somewhere.

Purchase orders

The rest of the new preliminary records come from smudgy carbon copies of acquisition records, and have the warning “copied from an old purchase order.”

Preliminary Hamnet record with the advisory statement “This is a PRELIMINARY RECORD copied from an old purchase order (which is better than nothing!)”

Technically, the information comes from old “accession slips” rather than from the purchase orders themselves, but conflating the two gives a better idea of the kind of information they contain. These records represent books acquired by the library, but not yet fully cataloged.7 In the Folger’s card catalog, the pink copies of these five-part accession slips are filed under author (or title for anonymous work), and the green copies are filed under date.8

Source of the “preliminary record” shown in this blog post (an accession slip from the “Continental Chronological File” of the Folger card catalog).

These slips were transcribed as found, including cryptic-looking information such as the dealer’s catalog number and item reference, and the initials of the staff member who assigned an accession number to the book.

Almost all of the caveats given above under “Catalog cards” apply to the “Purchase order” records, too. The exceptions are:

  1. They are all Continental publications from 1501 to 1799.
  2. There will only be one “preliminary record” per call number (because the slips never continue to a second page), but again, some will represent books that have since been fully cataloged online, so there definitely will be duplicates.
  3. Only author, title, and imprint are given (except very rarely, when the general note contains brief copy-specific information such as “Bound with 2 other titles” or “Copy of Thomas Wentworth, Earl of Strafford, with his signature.”)

Future plans

Digital images of the cards and accession slips that were scanned for this project will be made available to researchers, but we’re still working on the logistics. Meanwhile, staff can access specific images, one-by-one. If there’s a preliminary record from an old card or an old purchase order where seeing the scanned image would be helpful, just send the “URL for this record” from Hamnet to and we’ll send you the image(s) that the person doing the re-keying worked from.

Folger staff will continue to fix up these records, including removing duplicates. If you’ve saved a “URL for this record” for a record that no long exists, don’t worry. We can find out what happened to it. Just send the URL to, and we can send you a link to the new record.

Questions? Comments?

We’ve never tried something like this before, and we know there will be problems. If you have any questions or comments, please feel free to post them here, since other people are probably wondering the same thing. You can also, as always, email Meanwhile, we thank you for your patience, and we beg your indulgence if our response to an error report ends up being a sad “Yeah… we know…” followed by a chipper “But it’s better than nothing!”


  1. The prohibition on ending sentences with a preposition is a Victorian conceit over which I wish pedants would get.
  2. We considered replacing all the “unknowns” with a “questionable date” range of 1501 to 1799, but decided it was better to completely exclude them from search results that use date filters. It’s technically true that they were all published some time between 1501 and 1799, but if you’re trying to find things published between 1643 and 1645, it’s just cruel to throw in hundreds of records that are “maybe, but almost certainly not” from those years.
  3. Folger cataloging policy used to consider any publication with a title page as a “Book”, even if it’s a set of engravings with an engraved title page. That decision is now made on a case-by-case basis. Keep in mind, too, that “Book” includes single-sheet printed texts such as broadsides, and excludes all manuscripts (for example, a bound volume that consists entirely of text hand-copied to mimic a publication is a Manuscript, not a Book, for cataloging purposes).
  4. Changing them all to the machine-readable code “unknown or undetermined” is on my to-do list, but so is writing this blog post, and the blog post has a fixed deadline. The coding might or might not have been changed by the time you read this.
  5. We will be fixing these up by machine as much as possible, based on the most common initial articles in different languages. For example, you do not want to ignore the initial “De” in the Latin title De casibus virorum illustrium, but you do want to ignore it in the Dutch title De geestelyke stryd, and the fate of a “Der” in German depends on whether it’s being used as a nominative singular or a genitive plural.
  6. This particular one has been updated in Hamnet, but there are many more still needing work.
  7. Like most libraries, the Folger has a cataloging backlog. Rare books, manuscripts, and prints can be acquired much more quickly than they can be formally described. Librarians often call these materials “uncataloged” but in general terms, they are just extremely “undercataloged.” Items that are fully cataloged on cards or online can be searched by author, title, date, subject, genre, former owner, printer, etc., and contain accurate expert descriptions. Items in the cataloging backlog can only be searched by the bare-bones information created in the Acquisitions Department when the item was ordered, so if the dealer misdescribed something, it won’t have been corrected.
  8. The other three parts were buff, white, and yellow. Buff and white slips were filed in drawers in the Acquisitions department: buff alphabetically by “main entry” (usually author, but sometimes title) and white numerically by six-digit accession number. The yellow copy used to stay with the item until it had been cataloged, but for conservation reasons, many yellow slips have since been removed and kept separate.

Leave a Reply

  • (will not be published)