Archive Visit Checklist: A Four-Stage Guide

This is a working archive visit checklist for researchers travelling to consult manuscript material — the four stages that run from confirming access rules to the last look at your images before you stand up. It matters because almost nothing on this list can be fixed once you are home: a folio you did not photograph the back of stays unphotographed, and a legible image with no reference attached cannot be cited.

A good archive visit checklist has four parts, and they run in order. Confirm the institution's access rules before you travel — reader ticket, appointment, advance ordering deadline, daily production limits, and whether photography is permitted at all, at what cost, and with what equipment. Pack to those rules, not to a generic list. Photograph in a fixed sequence that keeps the archival reference attached to every image. Check before you leave — reference codes recorded, foliation noted, images reviewed at reading size, gaps identified while the box is still on your desk. The costly mistakes are almost never dramatic. They are a legible image nobody can cite, and a folio you did not photograph the back of.

What follows is that sequence in detail, aimed at a researcher travelling to consult manuscript material — English wills, French notarial acts, German parish registers, Dutch account books, or whatever else your project turns on.

Part one: confirm before you travel

Reading-room policy is local. There is no cross-archive norm on fees, limits, or permissions, and the published examples below vary widely enough that you should treat all of them as illustrations of what to ask, not as answers.

Access and identity

Most repositories require some form of registration before you touch an original. The National Archives (UK) issues a free reader's ticket but requires two forms of identification and a photograph, plus a document-handling video you can watch online in advance. NARA requires researcher registration and, under 36 CFR Part 1254, re-registration in each research room, each day. Registration is not bureaucracy for its own sake; it is how the institution tracks who handled what.

Ask, specifically: is a ticket required before arrival or issued on the day? Do I need a letter of introduction from my institution? Is there an age or affiliation restriction on the material I want?

Appointments, orders, and production limits

An appointment reserves a seat. An advance order tells closed-stack staff what to retrieve. A production limit governs how much can be pulled and returned safely in a day. These are three separate constraints, and each one can silently halve your reading time.

The numbers vary, and the variation is instructive. TNA asks that original documents be ordered at least four working days before a visit for guaranteed access, allows three documents at a time, notes that on-site orders take up to an hour to arrive, and offers a bulk-order visit for 20–40 documents from a single series. NARA recommends contacting the facility at least four weeks ahead, and for research rooms outside Washington, DC, advance consultation and an advance pull are required — the consultation is also how you confirm the records are actually at that facility. The German Federal Archives requires an appointment and a reserved workstation, and will not accommodate unregistered walk-ins. The Nationaal Archief allows a maximum of three boxes on your desk at once. Meanwhile Warwickshire County Record Office requires no appointment at all and lets you pre-order three items free.

The practical upshot: send exact references early. Not "the borough court rolls" but the repository's own reference code, in the repository's own format. The Archives nationales puts it bluntly — the cote is the unique identifier, indispensable for ordering, and its exact form must be respected. Also ask whether anything you want is stored off-site; remote holdings routinely need extra notice.

Photography: permitted, prohibited, or paid

Self-service photography is common now, but it is conditional and it is not universal. Assume nothing.

TNA permits handheld photography without flash and will lend a camera stand on request. The Library of Congress permits phone and camera photography with flash off. Essex Record Office charges £14.00 per day (£7.30 on Tuesday evenings after 5pm), requires a valid reader's ticket before you may even apply for permission, prohibits your own tripod, may confine you to a fixed photography station, and requires you to log the reference and image count on a form before each exposure. Warwickshire sells a daily permit at £8.50 or a weekly at £15.50. The German Federal Archives prohibits photography outright in at least its Bayreuth and Berlin-Reinickendorf reading rooms. The National Gallery of Art Archives encourages photography but requires a paper identification target in every frame.

Confirm the room, the material, the fee, the equipment, and the payment method. And confirm one more thing that catches people out: permission to photograph is not permission to publish. The NGA states plainly that obtaining a reproduction is not authorization to publish, and that written permission must be sought separately in advance. Access, copyright, donor restrictions, privacy, and publication rights are five distinct questions with five distinct answers.

Part two: what to pack

Pack against the confirmed policy, then add the friction items. The guidance on batteries and pencils below is practical researcher advice rather than controlled study, but the failure modes are real and cheap to prevent.

Pencils only. Pen and ink are near-universally prohibited. Some rooms are specific: Beinecke specifies yellow pencils.
Camera or phone, plus a spare battery, a formatted spare card, and a power bank if the room allows one. Sustained page-turning capture is heavy use; phone cameras throttle and batteries die mid-series.
A printed reference list. Every item you intend to consult, with its full reference code, in ordering sequence. This becomes your capture log.
A written metadata template. Repository, collection or fonds, series, box, folder, item title and date, reference code, folio or page locator. One row per item, filled in as you go.
Laptop or tablet, if permitted, for reviewing images at full size before you shoot a run.
Nothing else on the desk. Coats, bags, and cases usually go in a locker. Food and drink stay out.

Leave the cotton gloves at home unless you are told otherwise. TNA's conservators are explicit: they do not require gloves for handling documents, because gloves reduce finger sensitivity, catch on torn edges, transfer dirt, and make heavy items harder to control safely. Clean, dry hands and good handling practice protect paper better. Gloves genuinely belong with photographs, negatives, metal objects, lead seals, mould-affected material, and anything with hazardous residue — and staff will provide them.

Supports come from the archive, not your bag. Stanford's special collections require all books to sit in a foam cradle with staff-supplied weights; loose manuscript material lies flat on the table, never in your lap. Weights and snakes hold safe margins or protective sheets — never text, never fragile edges. NEDCC's handling guidance is the reference if you want the reasoning behind the enclosures and supports you'll be handed.

Part three: the capture sequence

This is the part that decides whether the trip was worth making. A disciplined shot order costs nothing at the desk and saves weeks afterwards. Our field method for photographing archival documents goes deeper on lighting, geometry, and handling; the sequence below is the checklist version.

Identifier shot first. Before the document, photograph the reference: the box label, the folder, the archive's target sheet, or a card you have written yourself with repository, collection, series, box, folder, and reference code. This single frame is what stops a perfectly legible image from becoming an orphan you cannot cite. Tropy's own metadata guidance makes the same point from the other end: archive, collection, box, folder, and identifier must be recorded, or the citation cannot be built later. This is workflow advice rather than a published archival standard, but the cost of skipping it is asymmetric.
The whole item, then the parts. Shoot the complete document or the full opening before any detail. Digital Scriptorium's imaging standards put the principle well: represent the manuscript as a three-dimensional object as a reader would see it, showing the edges of the book rather than cropping to a flat rectangle — and always photograph the full page alongside any detail, so the detail can be situated.
Recto and verso, always. A leaf has two sides. Endorsements, dorse notes, later docketing, and continuation text live on the back, and the back is the thing people forget. For a folded English will or deed, the endorsement is often where the probate act or the filing note sits.
Everything attached and everything loose. Seals, tags, wrappers, inserts, slips, marginalia, and continuation pages. A French notarial act needs its complete margins — the marginal notes recording later acts are frequently the evidentiary point.
Closing identifier or note when the item changes. A trailing frame, or a written line in your log, marking where one item ends and the next begins.
Check one image at full reading size before you shoot a run. Zoom to the level at which you will actually read the text. Focus, glare, and motion blur are invisible on a phone preview and fatal at the desk three months later.

Bound registers, loose papers, oversized sheets

Physical form changes the shot list more than language does. German parish books and Dutch account books are bound-register cases: cradle angle, safe opening, recto/verso sequence, and visible folio evidence in the frame. Digital Scriptorium recommends photographing bound leaves with the gutter running horizontally so light runs parallel to it, which helps with gutter shadow if the cradle and the room allow it. Loose papers lie flat. Maps and plans need staff-approved oversized support and an explicit scale or recorded dimensions.

Control glare by changing your angle or your position, never by pressing the object flatter. And do not use flash on the theory that one exposure is harmless. The Library of Congress's guidance on light damage is unambiguous: light damage is permanent, irreversible, and cumulative, and dim light over a long period does as much harm as bright light briefly. That reciprocity principle is why so many rooms ban flash even where ambient photography is welcome.

Targets, RAW, and the scanner-app trap

A greyscale, colour patches, and a linear scale in frame are genuinely valuable when colour, material, dimensions, or later calibration matter — Digital Scriptorium requires them for contributed manuscript images. For a text-legibility capture of a parish register, they may be unnecessary, and some rooms will not permit them anyway. Decide by use.

On file format, there is no cross-institution rule for handheld manuscript photography, and it would be a mistake to invent one. RAW keeps more adjustment latitude; a good full-frame JPEG is often serviceable for text-only work when focus and exposure are sound. The defensible rule is not about format at all: retain the least-processed original and create derivatives separately. Never overwrite the camera file with an enhanced version.

Which is the argument against running everything through a scanner app at the desk. Adobe Scan advertises exactly what it does — perspective correction, sharpening, glare and shadow removal, OCR. Those are useful derivatives for a modern business document. On a manuscript, auto-cropping removes marginalia and seals; contrast and binarisation obscure faint iron-gall ink; sharpening invents edge artefacts that look like strokes. If you want the app output, make it in addition to the camera original, and compare the two before trusting either. The Digital Library Federation's imaging guidelines are the standard statement of the master-file principle behind this.

Part four: what to record — foliation and reference codes

The image is half the record. The other half is where it came from.

Build each reference from the archive's own hierarchy, not your paraphrase of it: repository and location, collection or fonds or record group, series, box, folder, item title and date, the institution's exact reference code, and the page or folio locator. The Society of American Archivists' annotated finding aid shows how those hierarchical units map onto one another. Fonds, collection, record group, and series are not interchangeable synonyms — keep the repository's terminology.

Then get the locator right. Foliation numbers leaves; pagination numbers sides. One leaf has a recto and a verso; a page is one side, and conflating the two produces citations nobody can follow. If material is unfoliated or unpaginated, do not invent a number. Record "unfoliated," then add recto/verso, position in the sequence, opening words or heading, date, folder, and your image filename. Never infer folio numbers from the order of your image files. When you get home and start building citations you can defend, these are the fields you will wish you had.

Before you stand up

Run this while the material is still in front of you, because none of it can be fixed remotely:

Every item on your order list either photographed or explicitly logged as not consulted.
Identifier frame present at the start of every item.
Verso captured for every leaf that carries anything.
One image per item inspected at full reading size.
Reference code, folio range, and date written in the log for every item.
Anything you could not photograph — restricted, too fragile, oversized — noted with the staff member's name and the reason, so you can order a reproduction later.
Material returned in original order, nothing rearranged.

What happens to the images afterwards

Capture is one stage of a longer sequence, and each stage constrains the next: capture, organise, transcribe, analyse, cite. An image with no reference attached cannot be organised. A badly focused image cannot be transcribed, by you or by a machine. A transcription that has drifted from the page cannot be cited.

Organisation comes first. Tropy is free, open-source, and built for precisely this: turning a card full of IMG_5261.JPG into described items with archive, box, and folder fields. What it does not do is read them — Cornell's guide notes plainly that Tropy has no built-in text recognition and does not generate citations from its metadata.

Reading is the next stage, and it is where the bottleneck actually sits. Handwritten text recognition can now take a competent first pass at early-modern hands: Transkribus advertises handwritten and printed recognition across 100+ languages, which is a capability claim rather than an accuracy guarantee for your particular Jacobean will or Dutch ledger, and most specialist platforms expect you to select or train a model for your hand before you get useful output. Leo takes the other route — a specialist transcription model, ATR-1, that runs on Latin-script material out of the box, with no per-collection training step, in any language written in that alphabet: English, French, German, Dutch, Spanish, Italian and the rest (Greek, Cyrillic, Hebrew, Arabic, and East Asian scripts are outside its range). It reads printed matter as well as manuscript hands, which matters for the mixed folders most visits produce. The reason it belongs in a capture checklist is source integrity: it transcribes what is on the page — strikethroughs, insertions, marginalia, archaic spelling — rather than smoothing it into modern prose, and the reference metadata you recorded at the desk travels with the document as structured fields you can search across later. It also means the identifier frame you shot first is not decoration; it is what lets a whole trip's worth of images stay attached to their provenance while the text becomes searchable. You can upload straight from the phone in your pocket, which some researchers do before they have left the building. Transcription is not translation, incidentally — that is a separate step, run afterwards, that leaves the base text untouched.

Whatever reads the pages afterwards, it can only work with what you brought back. Machine transcription rewards a sharp, complete, full-frame image and punishes a cropped, contrast-crushed one, exactly as your own eyes do. Which is the closing argument for the discipline of the reading room: the checklist is not administrative overhead, it is the moment at which the evidence for everything you will later argue is either secured or quietly lost. Shoot the verso. Write down the cote. Look at one image properly before you shoot two hundred.

Frequently Asked Questions

What should be on an archive visit checklist?

An archive visit checklist has four stages, in order. First, confirm the institution's access rules before you travel: reader ticket or registration, appointment, advance ordering deadline, daily production limits, and whether photography is permitted, at what cost, and with what equipment. Second, pack to those confirmed rules — pencils, camera with spare battery and card, a printed reference list, a written metadata template. Third, photograph in a fixed sequence that keeps the archival reference attached to every image. Fourth, check before you stand up: reference codes recorded, foliation noted, images reviewed at reading size, gaps identified while the box is still on your desk.

Can you take photographs in an archive reading room?

Sometimes, but never assume it. Self-service photography is common and conditional. The National Archives (UK) permits handheld photography without flash and lends a camera stand; the Library of Congress permits phone and camera photography with flash off. Essex Record Office charges a daily fee, requires a valid reader's ticket before you may even apply, prohibits your own tripod, and may confine you to a fixed station. The German Federal Archives prohibits photography outright in at least two of its reading rooms. Confirm the room, the material, the fee, the equipment, and the payment method — and remember that permission to photograph is not permission to publish.

Do you need gloves to handle archival documents?

Usually not for paper and parchment. The National Archives (UK) conservators are explicit that gloves are not required for handling documents, because they reduce finger sensitivity, catch on torn edges, transfer dirt, and make heavy items harder to control safely. Clean, dry hands and careful handling protect paper better. Gloves do belong with photographs, negatives, metal objects, lead seals, mould-affected material, and anything carrying hazardous residue — and in those cases staff will provide them. Supports come from the archive too: cradles, weights, and snakes are supplied, and weights hold safe margins or protective sheets, never text or fragile edges.

What is the correct order for photographing a manuscript?

Shoot the identifier first: the box label, folder, target sheet, or a card recording repository, collection, series, box, folder, and reference code. That frame is what stops a legible image becoming an orphan you cannot cite. Then the whole item or full opening before any detail, so details can be situated. Then recto and verso for every leaf — endorsements, dorse notes, and later docketing live on the back. Then everything attached and loose: seals, tags, wrappers, inserts, marginalia, continuation pages. Mark where one item ends with a closing frame or log line. Finally, inspect one image at full reading size before shooting a run.

What is the difference between foliation and pagination in archival citations?

Foliation numbers leaves; pagination numbers sides. One leaf has a recto and a verso, so a folio covers two sides, whereas a page is a single side. Conflating the two produces citations nobody can follow, so record which system the material actually uses and keep the repository's own terminology throughout — fonds, collection, record group, and series are not interchangeable. If material is unfoliated or unpaginated, do not invent a number. Write "unfoliated," then add recto or verso, position in the sequence, opening words or heading, date, folder, and your image filename. Never infer folio numbers from the order of your image files.