How to Date an Old Photograph From Clues

A photograph of a street corner will usually tell you where it was taken. Look a little longer and it will often tell you when. The two questions feel like one question, and they are answered by completely different evidence. Place lives in what varies across space: the script on a shopfront, the pitch of a roof, the plants at the edge of the frame. Time lives in what changed on a schedule — a car that went on sale in a particular year, a registration format replaced overnight by law, a phone kiosk pulled out of every high street inside a decade.

Dating an undated picture is a human skill, and a quietly satisfying one. It is also the half of the puzzle Raven cannot do for you. The model is built to guess place, not year, and it will happily name a country while saying nothing at all about the decade. So the method below is yours to run by eye. The useful part is that the two halves feed each other: nearly every clue that brackets a date also narrows a location, and the reverse holds too.

The distinction matters because it changes what you look at. Someone placing a photo scans for things that differ between countries and stay stable through time: a mountain profile, a writing system, the colour of a road sign. Someone dating a photo wants the opposite — objects that are much the same everywhere but were only made for a while. A chunky mobile handset in somebody's grip is useless for place and excellent for period. A limestone cliff is the reverse. Most of this blog is about the first kind of reading, and the long story of working out where photos were taken is where that thread begins. Dating runs on a parallel track, borrowed mostly from archivists and museum cataloguers.

One more difference is worth stating early. Placing a photo can succeed on a single decisive clue: one legible sign in the right script is sometimes enough. Dating almost never works that way. Every date clue is a boundary rather than a point, and the answer only appears once several boundaries are stacked against each other.

Cars are the workhorse of photo dating, because they are dense with dated information and people photograph streets full of them by accident. A model has a launch year, a facelift year and an end of production. Even a partial view often carries enough: a grille shape, a tail light cluster, the proportions of the glasshouse. Once you can name the newest vehicle in the frame, you have an earliest possible date for the whole picture, since nothing in a photograph can predate the newest thing in it.

Number plates sharpen that further, and in Britain they are unusually generous. Registration plates in the United Kingdom carried a year suffix letter from 1963, switched to a prefix letter in 1983, and moved to the current two-letter, two-digit format in September 2001. That modern age identifier changes twice a year, in March and September, so one legible plate can bracket a photograph to six months with no other evidence at all. Other countries offer their own versions: a colour change, an EU band added or removed, city codes redrawn after a border moved.

This is the clearest case of dating and locating reinforcing each other. A car model tells you which market you are probably standing in and which years are plausible, in a single glance. The geographic half of that reading is covered in what a car and licence plate say about location; the temporal half uses exactly the same observations, held against a calendar instead of a map.

Fashion is the clue everyone reaches for first and the one that misleads most often. A style has a peak, a long tail and, eventually, a revival, and the person in your photograph may be a decade behind all three. Trouser cut, collar width, spectacle frames, trainers and hair volume do carry period, but they carry it loosely, and a single subject is weak evidence on its own.

Crowds are far better than individuals. In a street scene, a dozen half-visible people average out personal taste and leave you with the shape of a decade rather than the taste of one wardrobe. Children and teenagers are the sharpest group, because their clothes are replaced constantly and follow trends quickly. Uniforms, sports kit and workwear help too, since institutions change their designs on documented dates and keep records of it.

If you are holding a physical photograph rather than a scan, half the evidence is in the object itself. Format changed steadily through the twentieth century, and each shift is a boundary. A carte de visite mounted on stiff card belongs to the 1860s and 1870s; a cabinet card is later. A square print about 3.5 inches across usually came from an Instamatic cartridge, so 1963 or after. Smaller square prints, closer to two inches, came off 120 roll film and can be decades older, so measure before you conclude. The move from black and white to colour in ordinary family albums happened gradually across the 1960s and 1970s, so a holiday album entirely in colour is rarely earlier than that.

Paper stock and finish carry their own dates. Deckled white borders were common on domestic prints through the 1950s and 1960s and fell away afterwards. Many processors stamped a month and year in the margin or on the reverse, which is as close to a free answer as this hobby offers, though it dates the printing rather than the exposure and reprints do exist. Colour dyes also fade unevenly: prints made in the 1970s often drift towards magenta after 40 years or so, while a Kodachrome slide from the same era can still look startlingly correct, because its dyes were formed during processing rather than stored in the film.

• —The back. Processor stamps, date codes, handwritten captions and album glue all carry period information.
• —The border. Deckled edges, a white margin, a printed frame number or a studio name each narrow the years.
• —The size and shape. Square, 110 format, standard enprint or a modern 6x4 — every format had its stretch of dominance.
• —The surface. Glossy, matte, silk and textured finishes moved in and out of fashion at the lab.
• —The damage. Silvering, foxing and album-corner marks suggest a genuinely old print rather than a recent reprint.

Street furniture and consumer technology are the most reliable clues after vehicles, precisely because nobody composes a photograph around them. Rooftop television aerials appear in British streets from the 1950s. Satellite dishes on ordinary houses arrived far later, barely anything before 1989 in the UK and then suddenly everywhere. Street lighting is datable by colour: the flat amber of low-pressure sodium, then the warmer orange of high-pressure sodium, held British streets for decades before white LED replaced both through the 2010s. The colour of a night photograph is a rough decade marker in itself.

Then there is the vanishing furniture. The red telephone box is the obvious British example, but every country has equivalents: postboxes with old markings, cigarette machines bolted to pub walls, bus shelters in a livery that lasted exactly one contract. Shopfronts are excellent, because chains rebrand on documented dates and a single logo in the background can cut a decade in half. Indoors, look for the screen. A deep cathode ray tube television, a beige computer monitor and a flat panel each belong to a different stretch of years.

A skyline caught mid-construction is one of the most precise clues available, and it is also the clue that needs the location first. Once you know the city, a tower topped out at twenty floors instead of forty gives you a specific season rather than a decade. The same applies at street level: resurfaced roads, tram tracks being relaid, scaffolding on a church, a terrace that has since been demolished but is still standing in your frame. Cross-checking that kind of detail against public imagery is a discipline of its own, and the approach closely follows the one set out in how to verify where a photo was taken.

The method archivists use is simple and worth copying. Work through the frame listing every dateable object with the window it implies, phrased as 'not before' or 'not after'. The newest item sets the floor: a photograph containing a car launched in 1998 cannot be from 1996, however old everything else looks. The ceiling comes from whatever later disappeared — a building since demolished, a shop that closed, a plate format replaced.

Stack those boundaries and a range appears. Suppose a British street scene shows a prefix-letter plate, an orange sodium lamp, a stub-aerial mobile phone in somebody's hand, and a chain shopfront in a livery adopted in 1996. The plate puts the photograph before September 2001 and the livery puts it after 1996. The lamp and the handset both sit comfortably inside that window rather than arguing with it, which is what agreement looks like. Four independent clues, one five-year range. That is a good result. A single clue producing a single year is not a better result; it is usually an error waiting to be found.

Two habits keep the process honest. Scan the back of a print as well as the front before you file it away, and write the reasoning down beside the range instead of only the conclusion. If you are working through a box of family prints, doing both while you scan saves repeating the work later, and the practical side of that job is covered in digitising and exploring old travel albums.

It is worth being blunt about the limit, because the question comes up constantly. Raven looks at a photograph, weighs visible clues and names a place. It has no notion of the year, it does not read the file's metadata, and it would gain little from that metadata anyway, since most scans of old prints carry the date of the scan rather than the date of the exposure. Nothing in the pipeline is trying to answer 'when'.

Where the tool earns its place is upstream. Knowing the country turns a vague date hunt into a specific one, because almost every clue in this article is national: plate formats, lighting rollouts, telephone kiosks, satellite television launches, school uniforms. Get the place first, then read the period against that country's own timeline. The wider story of how photographs themselves changed, from film to sensor and print to camera roll, sits in how technology changed the travel photo.

Dating a photograph rewards patience more than expertise. The evidence is nearly always there, sitting in a bumper, a lamp, a border or a badly chosen jumper, and it only asks that you notice the things the photographer never meant you to look at. Start with the newest object in the frame, work outward, and be content with a range. A picture honestly dated to within four years is worth far more than a confident wrong year written on the back in biro.