How to date a photograph from the object, not the face.

Some evidence cannot be argued with. Start there.

Taken together, a family photograph offers four kinds of evidence. They are not equally strong, and people habitually use them in exactly the wrong order.

1. The process. Nearly decisive. A tintype cannot have been made in 1935; a colour print on resin-coated paper cannot have been made in 1955. If you can identify the process, you have bracketed the date before you look at a single person. If you cannot name it, the formats reference identifies nine processes from one glance at the object.
2. The mount and the back. Strong, and usually ignored. Card sizes, studio imprints and postcard back formats changed on documented dates, because they were commercial products with suppliers and standards.
3. The surface, edges and corners. Moderate. Deckled edges, scalloped borders, glossy resin-coated paper and square corners all cluster in particular decades.
4. The subject. Weakest. Clothing, cars and hairstyles tell you when a style existed, not when this photograph was taken — and photographs were routinely printed years or decades after the negative was exposed.

That last point is the one that undoes most amateur dating. The object in your hand is a print, and a print has its own date. A 1950s print from a 1920s negative is a 1950s object. If you are dating it for a caption, say "photographed circa 1925, printed later" — that is usually the honest answer, and it is far more useful than a confident wrong year.

The formats, and the dates they cannot escape.

If you can place the process, you have already beaten most of the guesses on the internet. The tool below does this systematically, and the rest of this page explains the evidence it uses.

Turn it over. The back is usually the most informative side.

Card-mounted photographs were a product, and products carry manufacturer marks.

• Studio imprints. From the 1860s, studios printed their name and often their address on the back. A street number or a photographer's name is a lookup, not a guess — town directories record exactly when a studio operated at a given address, and that can pin a cabinet card to a five-year window. This is the single most productive thing you can do with a card-mounted photograph.
• Undivided versus divided back. Early postcards reserved the whole back for the address. Divided backs, with a separate message space, were permitted later — the change dates around 1901–1907 depending on country. An undivided back puts a postcard print before roughly 1907 in Britain.
• Stamp boxes on postcard-format photographic paper. Papers sold for making your own postcards carried a printed stamp box naming the manufacturer, and the corners of those boxes carry coded designs. The well-known four-triangle AZO mark sits at roughly 1925–1940. Other marks run back to the turn of the century and forward into the 1950s.
• Backprint on the paper itself. From the 1940s, manufacturers printed the paper type across the reverse in repeating text. "A Kodak Paper" style backprints give a manufacturer and a period, and resin-coated paper entering that text is a strong post-1968 signal.

A back with pencil handwriting and nothing else is not dateless. The hand, the ink and the names are genealogical evidence — that is the labelling problem, not the dating problem, and it is worth solving while whoever can read the handwriting is still with you.

Borders, corners and finishes cluster by decade.

None of these is decisive alone. Two of them agreeing is a real signal.

• Deckled or scalloped edges, 1940s–1960s. The wavy, cut-edge look on small snapshot prints is a mid-century fashion, and it is one of the more dependable surface clues.
• White border with square corners, roughly 1920s–1950s. The border came from the printing process; the square corners are what change.
• Rounded corners, mid-1950s onward. The commonest snapshot print of the 1960s and 1970s.
• Resin-coated (RC) paper, 1968 onward. RC paper replaced fibre paper for amateur printing. It looks slightly plastic, is thinner, does not cockle the same way and feels slick. A colour print on RC paper is late 1960s at the earliest.
• Fibre paper, glossy or matte. The textured, faintly fibrous surface of earlier prints. Matte fibre prints from the 1930s–1950s are often beautifully made and often mistaken for older than they are.
• Borderless full-bleed prints. Mostly 1990s and after, when minilabs and then home printers stopped needing a border.

The surface also tells you how the print has aged, which matters for what you do next. Glossy RC colour prints of the 1970s and 1980s are the classic dye-fade cases — the ones that have gone pink, cyan or yellow-green — and that fading is not dirt. It is chemistry, and it is why colour cast correction is cheap while reconstruction is not.

Colour is a very good clock, if you read it correctly.

Colour processes arrived in a specific order, and amateur colour arrived later than people expect.

• Hand-tinting, 1850s–1930s. Painted over a black-and-white print. Unmistakable up close: the colour sits on the surface, does not align perfectly with the edges beneath, and is usually restricted to faces, flowers and dresses.
• Sepia and warm-toned monochrome, before c.1930. Sepia is a toning process, not age. A warm brown print with even tone is usually deliberate; a yellowed print with a purple cast is usually deterioration.
• Kodachrome, 1935, and Kodacolor, 1942. Slide film made colour ordinary for enthusiasts first; colour negative film for amateur prints followed. A genuinely vivid colour print almost certainly post-dates the 1950s.
• The Instamatic era, 1963 onward. Cartridge-loading cameras put cheap colour square prints into ordinary households. Millions of 1960s and 1970s family photographs are 126-format squares.
• Instant colour, 1963 onward, and the SX-70 format from 1972. Square, thick, often with a white frame and a slightly soft, warm rendering.
• Faded colour, roughly 1965–1985. Paradoxically useful: the dye instability of that period means a colour print that has shifted dramatically is very likely from it. The faded prints are the ones that date themselves.

One warning, and it is the same warning as everywhere else on this site: do not date a photograph by its quality. A soft, poorly exposed photograph is not necessarily old, and a sharp, well-made one is not necessarily recent. Quality tracks the photographer, not the decade.

Narrow your photograph to a decade range.

Four questions about the object. The tool intersects the date ranges of everything you can answer, and tells you which single clue did most of the work — so you know what to check next if the answer feels wrong.

A date you do not record is a date you will lose again.

The most common failure in family photograph collections is not damage. It is that someone worked out what a photograph was, told two people at a kitchen table, and died.

1. Write on the back, in pencil, on the border. Not on the image, not in ink, and never with a marker. A soft pencil on the border of a print, or on the mount beneath it, is reversible and has been the archive convention for a century. Include the words "circa" when you are estimating — "c.1925" is more honest than a false precision.
2. Record the reasoning, not just the answer. "Cabinet card, studio imprint of H. Ellis, Falmouth; directory puts him there 1888–1896" is a record. "1890s" is a guess that will be re-guessed by the next person.
3. Put the date in the filename of the scan. A file called 1925-c-trebarwith-harbour.jpg survives; a file called IMG_4471.jpg does not.
4. Keep one copy of the record outside the collection. A single text file, a note in your phone, an email to yourself. The photographs can be lost in one event; the index should not be lost in the same event.

Then store it properly. A photograph you have just dated is a photograph you now know is worth protecting, and the mid-century colour prints in most collections are both the most recently dated and the most urgently in need of better storage.