Working from archives without damaging the record
How to photograph an archive document safely, check an old photograph's provenance, cite the source in a caption, and let a research notebook sort your
Photograph an archive document with the page flat, the camera parallel to it, and no pressure on the paper. Light it from two sides at a low angle, never with a flash aimed straight down. Then write down what you shot, where it came from, and who owns it, because the photograph is only half the record.
How do you photograph an archive document without damaging it?
Start with permission and a clean surface. Ask the archivist what the item can take. Some bound volumes will not open flat, and forcing them cracks the spine and breaks the sewing. If the book resists, shoot the page at the angle it allows and correct the perspective later. A slightly keystoned image of an intact volume beats a flat image of a broken one.
Use a copy stand or a tripod with the camera directly above the page. Handheld shots at an angle introduce distortion you cannot fully undo, and they tempt you to lean on the document. Set the camera parallel to the paper, fill the frame with the page plus a small margin, and keep the same distance for every frame in a folder. Consistency matters more than resolution when you are shooting two hundred pages.
Light from both sides at roughly forty five degrees. Two small LED panels on stands, or the room's own indirect light, will do. A single hard source from one side throws the texture of the paper into shadow and buries faint pencil. Never use the camera's built in flash. The burst of heat and light is small, but repeated flashes on the same page are exactly the kind of exposure an archive is trying to avoid.
Weights go outside the text block. Use snake weights or beanbags on the blank margins, never on the writing. Cotton gloves for photographic prints and negatives, clean dry hands for most paper. Do not lick a finger to turn a page. Do not use tape, clips, or adhesive notes on anything.
Shoot a target frame first. A sheet with the collection name, box number, folder number, and date goes in front of every set. When the images come off the card six months later, that frame is the only thing that tells you what you have. This habit is standard in naval history work, where a single folder can hold forty documents from four different commands, and it is the reason a project like this mine warfare research journal keeps its deck logs and muster rolls indexed by box before anything else.
Check focus at full size on the back of the camera, not on the small screen preview. Shoot raw. Name files the same day.
How is the provenance of an old photograph checked?
Provenance is a chain, and you check it link by link. Start with the object itself. A photographic print carries physical evidence: the paper base, the surface finish, the border width, the stamp on the back, the handwriting, the crop marks. A scan carries almost none of that, which is why you ask for the back of the print before you ask for anything else.
Ask who held the picture and when. A family photograph that stayed in one household for eighty years has a short chain and a strong story. A picture bought at a flea market has no chain at all, and the caption someone wrote on the envelope is a claim, not a fact.
Match the image against the record. If the photograph shows a ship, compare hull numbers, radar fit, and paint scheme against the dates those features were current. If it shows a place, compare buildings and shoreline against charts and other photographs from the same year. A single detail that cannot exist on the date claimed is enough to move the picture.
Look for the same image elsewhere. Prints were duplicated and mailed, so a copy may sit in a museum collection, a local archive, or a published history with a caption that names the photographer. Finding a second copy with a documented source is the fastest way to settle a date.
Write down what you cannot settle. A caption that says "probably 1967, photographer unknown" is honest. A caption that says "1967" is a guess wearing a suit. Keep the doubt in the file, because the next researcher may be the one who resolves it.
How is a source cited in a caption?
A caption carries the credit, the collection, and the identifier, in that order, in as few words as the page allows. Credit names the photographer or the institution that made the image. Collection names the archive and the file. Identifier is the box, folder, negative number, or accession number that lets someone else order the same item.
A working caption looks like this: "USS Phoebe, Sasebo, 1966. U.S. Navy photograph, Naval History and Heritage Command, NH series, box 14, folder 3." If the photographer is unknown, say unknown. If the date is uncertain, say circa and give the range you can defend.
For a document rather than a photograph, cite the record group, the series, the box, and the folder, plus the repository. For an oral history, cite the narrator, the interviewer, the date of the interview, and the collection. For a published book, cite author, title, publisher, year, and page.
Keep the same citation format across the whole project. A reader who learns your format on page one should never have to relearn it on page ninety. Put the full citation in the caption when the image stands alone, and a short form in the caption with the full entry in a source list when the images run in sequence.
What a research notebook changes in the way pictures are sorted
A notebook turns a pile of images into a set of answers. Without one, you sort by what a picture looks like: ships, docks, crew, sunsets. With one, you sort by what a picture proves: a date, a hull number, a port, a refit, a deployment.
The change is practical. Every image gets a page in the notebook with the file name, the source, the date range, the people and places identified, and the open questions. Images that share a question get filed together even when they look nothing alike. A photograph of a pier and a page of a deck log end up in the same folder because both answer where the ship was in March.
This is how a chronology gets built. You stop asking whether a picture is good and start asking what it settles. The notebook also records the dead ends, which saves the next person from walking them again.
The sorting rule that survives every project: file by question, not by subject. Subjects multiply. Questions end.
What to do before you leave the reading room
Check every frame for focus, glare, and cut off edges while you still have the document in front of you. Re shoot the bad ones now. You will not get another appointment tomorrow.
Confirm the citation details with the archivist before you pack. Ask how the repository wants to be credited, and write it down exactly. Repositories differ, and a wrong credit line is a small error that costs a real relationship.
Back up the card before you do anything else. Two copies, two drives, one of them off site. Then rename the files to match the notebook, and only then start editing.
Return the materials the way you found them, in the order you found them. The next researcher gets the same access you did because someone before you left the box as it was.