Find an Archival Film Item and Request the Copy You Actually Need

A catalog description is a lead, not footage. It tells you that something exists under a title, in a repository, with an identifier attached to it. It does not tell you that the passage you want is in the film, that anyone can show it to you, or that you may use it in a pitch, a trailer, or a finished cut.

Two things turn that lead into research you can honestly build on: an identifiable object — repository, local identifier, title, and the reel or part if there is more than one — and a route to inspect it. Everything below is the work of getting those two things, and of keeping your notes about what you've requested separate from what you've actually seen.

Capture the object, not just the search result

When a record looks promising, copy it before you write anything down about what it shows. Take the repository and the collection or series it sits in. Take the local identifier — the item number, accession, call number, or whatever the catalog uses — because that, not the title, is what a reference archivist can act on. Take the title exactly as it is written, including the parts that look like filing noise: dates in parentheses, "miscellaneous," "unidentified," a reel count. Take the extent and format, the production date as given, and the description text itself.

Then note the two things that are easy to skip. One: the URL, and the date you looked at it. Catalog records get revised and occasionally disappear into a migration; a note that says "found in [ARCHIVE] catalog, seen [DATE]" survives that. Two: mark which fields came from the description and which are still unclear. A three-column habit is enough — in the record, my inference, unknown. If the record says "4 reels of 4," that's in the record. If your note says "reel 3 has the parade," that's your inference, and at this stage you have no grounds for it.

Titles deserve particular suspicion. A familiar title can name a whole collection, a newsreel issue containing several unrelated stories, a series rather than an item, or a single reel whose description was written decades ago from a donor's memory. A phrase match in a search result is not an order form. You are not looking for the title you recognize; you are looking for the object the repository can retrieve and put in front of you. If the identifier is missing or ambiguous, that's the first thing you solve, not the last.

Work out what can actually be inspected

Most catalog leads sit in one of three states, and they call for different next moves.

A digital object you can play. Something is linked, and it streams or downloads. Before you treat this as the item, check what the object represents. It may be reel 1 of 4. It may be a low-resolution access copy made years ago. It may be an outtake, a trailer, or a compilation assembled for a different purpose. The object's own description usually says — and when it doesn't, that is a question for the repository, not a detail to assume.

A description with no object. The record exists; nothing is linked. This is a lead, not a dead end. Many holdings are described but undigitized, and the route to seeing them runs through a person.

A record that needs the repository to clarify it. "Unprocessed," "reel count to be determined," a title that matches two collections, an identifier that points to a series rather than an item. Here the useful move is a short, specific question to the reference desk, not a guess that hardens into a presumption.

An unresolved lead is not unavailable footage. Those are different claims, and only the repository can move a lead from the first category to the second. Reel identification is the same kind of thing. If the description doesn't say which reel holds what you need, ask. Inventing "Reel 2" because you need there to be a Reel 2 is how a pitch ends up with a citation that its own source contradicts.

Write the viewing or research-copy inquiry

A good inquiry gives the reference staff everything they need to identify the object and a short set of specific questions. It usually has six parts.

1. Who you are and what the research is, in one line. A documentary researcher working on a project about a specific subject is enough context.
2. The item, as fully identified as you can manage: repository, collection or series, title as it appears, local identifier, date, format, extent, and a reel number if one is known.
3. What you're looking for, in words. The place, the period, the event, the person, and what you're trying to establish or show. Descriptive terms, not timecodes.
4. What you have already viewed. Say it plainly. "I have viewed the catalog record only; no digital object is linked." Or "I have viewed reels 1 and 2 at the reference desk." This is the sentence that keeps you honest later.
5. The specific questions. Does a viewing copy exist? Which reel covers the material I described? Is reproduction done at reel or item level, or can a shorter section be copied? Is there an approval step before duplication? Anything the repository can tell you about reel contents, leader notes, or an older shot list?
6. An explicit separation of inspection from reuse. "I'm asking about viewing or a research copy at this stage. If the material proves useful, I'll write separately about permissions and reuse."

Here is a constructed example. It is not an actual holding; the placeholders are deliberately conspicuous so that nobody pastes them into a real request or mistakes them for a real accession number.

To: Reference desk, [ARCHIVE NAME] Subject: Research copy inquiry — [TITLE AS IT APPEARS], [LOCAL IDENTIFIER]

Hello,

I'm a documentary researcher working on [PROJECT IN ONE LINE]. I'd like to ask about one item and ask a few bounded questions about it.

The item. [TITLE AS IT APPEARS], [DATE AS GIVEN], in [COLLECTION OR SERIES], local identifier [LOCAL IDENTIFIER], extent listed as 4 reels, format [FORMAT]. I found it through [CATALOG NAME] on [DATE]; the record has no digital object linked.

What I'm looking for. The description mentions [PLACE], [EVENT OR SUBJECT], around [PERIOD]. I'm trying to establish whether the item shows [PASSAGE, DESCRIBED IN WORDS] and roughly where in the footage it falls.

What I've viewed. Nothing yet beyond the catalog record. I have not seen any part of the item, and I don't know which reel the description refers to.

What I'd like to know.

1. Does a viewing copy exist, and can I view it? If so, what is the usual route — on site, by appointment, or something else?
2. If no viewing copy exists, is a research copy available, and at what level? I would rather request one reel than all four if a reel can be identified first.
3. Is there an approval step before duplication, and is there anything in the reel contents, leader notes, or an older inventory that would point to which reel to check first?
4. What should I expect in terms of process from here?

I'm asking about inspection only at this stage. If the material turns out to be useful, I'll come back separately about permissions and reuse.

Thank you, [NAME], [ROLE OR AFFILIATION], [CONTACT]

Read the letter as an archivist would. They can act on the identifier and the title. They can answer the reel question or tell you it isn't answerable. They know you haven't viewed anything, so they won't assume you've already confirmed the passage exists. And they can see where the boundary sits between inspecting and licensing, which is not always the same office and never the same approval.

What the letter cannot do is answer the filmmaker's own remaining unknowns: whether the passage is in the film at all, which reel holds it, whether a copy exists on any terms, and whether a clip-level transfer is even offered. Notice the fourth point carefully. A repository that duplicates at whole-reel or whole-item level will not necessarily make you a transfer of the one moment you described, no matter how precisely you describe it. For the US National Archives, the order-copies page — checked on 18 September 2026, and previously on 9 September 2026 — asks for item-level citations including reel information where relevant, separates approval for reproduction from the duplication arrangements themselves, and describes transfer at whole-reel or item level rather than for individual scenes. That is a dated observation of that page, not a rule for archives in general, and not a current price or turnaround quote. The page carried an older administrative notice and a February 2024 review date when it was checked, so anything operational on it should be confirmed with the repository before you build a schedule around it.

One more distinction, because it is the one that gets blurred in a hurry: a research copy is evidence. It is something you can watch, log, and cite. It is not permission to use the material, and receiving it does not settle any rights question.

Keep the request state attached to the pitch claim

Requests move through a short ladder, and each rung is a separate fact. Advance only when the event actually happens.

A reply is not a copy. A copy is not inspection. Inspection is not permission. Each of those gaps has produced a documentary pitch that promised something the film couldn't deliver.

When a copy does arrive, four checks before you write a word of narration. Identity: is this the identifier you asked for, or the similarly titled item from the neighboring collection? Coverage: how many reels arrived, and which ones? Completeness: heads and tails, missing reels, a silent reel where the record promised sound, a duplicate of a reel you already had. Contents: watch for the passage you described, and note the reel and approximate position so that your own citation means something later. Sometimes the answer to the fourth check is that the material is real, is held, is viewable, and does not contain what the description implied. That is a finding, and it belongs in your notes in exactly those words.

Now the branch that catches everyone. You send the letter, and the reply says the records don't allow reel-level identification: the finding aid is collection-level, there's no shot list, and staff can't confirm which of the four reels holds the material you described. The next move is more clarification, not a stronger claim in the deck.

A few things are actually available here. Ask whether viewing can cover all four reels, or whether the reels can be checked at their heads for titles or slates that distinguish them. Ask whether an older inventory or donor file exists outside the online catalog. Look for published sources that describe the footage — a history of the event, a memoir, a previous film that used the same collection — and see whether any of them place the material in a particular reel. Then decide, in hours and money, whether watching four reels to find one passage is worth it. If it isn't, the pitch keeps the uncertainty: a catalog record identified, contents not yet confirmed. That sentence is weaker than "we have the parade footage," and it is the one that will still be true when the film goes into production.

What to do next

Take the record you actually found. Fill in the placeholders from it — repository, title as written, identifier, extent, date — and send the inquiry to the reference staff named on the repository's own moving image or research page. If the repository publishes a separate route for copies and a separate one for rights, use the copies route now and leave the other for later.

Until the copy is in hand and watched, the honest pitch line is short: a record located in [REPOSITORY] under [LOCAL IDENTIFIER], contents not yet viewed, inquiry under way. Success at this stage is a traceable inquiry that a responsible person can act on — the identifier they need, the reel question asked out loud, and a clear statement of what you have and haven't seen. That is a smaller achievement than securing a shot. It is also the one you can build a documentary on, because the next step after it is real.

Frequently asked questions

What details should I capture from a catalog record before making a request?

Capture the repository and collection or series, the local identifier such as item number, accession, or call number, the title exactly as written including filing noise, extent and format, production date as given, description text, URL, and the date you looked. Keep fields separated as in the record, my inference, and unknown. A title alone can name a collection, a newsreel issue, a series, or a single reel described from memory.

What if the catalog record has no digital object linked?

That is a lead, not a dead end. Many holdings are described but undigitized, so the route runs through a person. An unresolved lead is not the same as unavailable footage; only the repository can move a lead to unavailable. If the description does not say which reel holds what you need, ask rather than inventing a reel number.

What should a viewing or research-copy inquiry contain?

The body gives six parts: who you are and what the research is, the item as fully identified as possible, what you are looking for in words, what you have already viewed, specific questions, and an explicit separation of inspection from reuse. Specific questions can cover whether a viewing copy exists, which reel covers the material, whether reproduction is at reel or item level, whether there is an approval step, and whether reel contents, leader notes, or an older shot list can help. The body also cites a dated observation of the US National Archives order-copies page: item-level citations including reel information where relevant, approval for reproduction separated from duplication arrangements, and transfer at whole-reel or item level rather than individual scenes. That is not a rule for archives in general or a current price or turnaround quote.

Does receiving a research copy give me permission to use the material?

No. A research copy is evidence you can watch, log, and cite. It is not permission to use the material, and receiving it does not settle any rights question. The body puts it as a ladder: a reply is not a copy, a copy is not inspection, and inspection is not permission.

What can I honestly say in a pitch before the copy is inspected?

Match the claim to the state that is actually true. The ladder runs from described, inquiry prepared, inquiry sent, reply received, copy available, to contents inspected. Until the copy is in hand and watched, the honest line is short: a record located in a repository under a local identifier, contents not yet viewed, inquiry under way. If the repository cannot identify which reel holds the material, ask about viewing all reels, checking heads for titles or slates, older inventories or donor files, and published sources; then decide whether the hours and money are worth it, or keep the uncertainty.

Further reading

• Pitch a Documentary Before You Know How It Ends — Pitching before the material's state is settled needs the same honesty; this labels what is captured, reported, established or only hoped for, in its own words.
• Does Your Film Need an Archivist, a Historical Adviser, or More Reporting? — If the inquiry needs more than a reference desk can answer, this separates locating material from provenance, interpretation and rights, and how to brief whoever takes it on.