Anna Atkins had a problem, and it was an art problem before it was a science one.
She was a botanist in Kent, in England, and she studied algae — seaweed. To publish what you knew about a plant in 1843, you had to have it drawn, then engraved, then printed. That worked well for an oak leaf. It did not work at all for seaweed.
Look closely at a piece of seaweed sometime. Some of it is a tangle of threads finer than hair. An engraver copying that by hand has to decide which threads to keep, and the answer is always most of them, but not all. Every copy loses a little truth.
So Atkins did something else. She stopped trying to draw the seaweed and made the seaweed draw itself.
The paper does it
The year before, a family friend named John Herschel had invented a new way of making pictures. Herschel was an astronomer, and a famous one — he is the person who gave photography the words negative and positive.
His process is called the cyanotype, and there is no camera in it at all.
You brush two chemicals onto a sheet of paper and let it dry in the dark. Both contain iron. Then you lay an object right on top of the paper and put the whole thing in the sun. Wherever the light lands, the iron changes. Wherever the object sits, it does not.
Then you rinse the paper in plain water. And this is the surprising part — the blue arrives in the water, not in the sun. In the sun the paper only goes a dull bronze. The rinse washes away everything the light never touched, leaves behind everything it did, and what is left behind is a deep, permanent blue.
What the paper is really measuring
Here is the idea worth carrying away, and it is a science idea rather than an art one.
The paper cannot see color. It only counts how much light got through.
Lay something solid on it — a thick stem, a shell, a coin — and no light reaches the paper underneath. That patch rinses away to bare white. Lay something you can half see through, like a flat blade of kelp, and some light gets through. That patch comes out pale blue. Lay a few hair-thin threads down and almost all the light gets past them, so they print as the faintest lines.
Which means a cyanotype is not a picture of what seaweed looks like. It is a measurement of how much light each part of it blocks, printed at life size. A red seaweed and a green one of the same thickness come out exactly the same.
It has to be the sun
The paper is not fussy about beauty, but it is very fussy about which light you give it.
The iron chemistry only responds to the short, high-energy end of the spectrum — ultraviolet and the deepest blues. It is completely blind to yellow, orange and red. That single fact explains three things at once.
You can coat the paper under a dim indoor lamp without ruining it, because a warm bulb gives out almost nothing the paper can feel. A window on a cloudy day still prints, just slowly, because ultraviolet comes through clouds. And a sheet left under a bright lamp all evening may do nothing at all, while ten minutes of real sunshine does the whole job.
One blue, keeping busy
That deep color is not new in 1842. It is Prussian blue, and it has a life story.
Around 1706 a color-maker in Berlin was trying to make a red. Something in his batch was contaminated, and the mixture turned blue instead. The accident turned out to be the first modern synthetic pigment — a strong, cheap blue at a time when the best blue on the shelf was ground from a stone mined in Afghanistan and priced like gold.
It traveled. By the 1820s it reached Japan through the only trade Japan then allowed, where printmakers called it bero-ai — “Berlin blue.” Japanese prints had used indigo before, which fades in sunlight and sits pale on the page. Prussian blue did not fade and did not sit pale. Hokusai used it to build the most reproduced wave in the world.
Twenty years after that, Herschel worked out that the same blue could be grown on a sheet of paper by sunlight. Thirty years after that, engineers noticed that this was the cheapest way to copy a drawing, and copied their plans this way for the better part of a century. That is why we still say blueprint.
How much work was that, exactly?
Atkins wrote out her reason in the introduction, by hand, and it is worth reading slowly:
The difficulty of making accurate drawings of objects as minute as many of the Algae and Confervae, has induced me to avail myself of Sir John Herschel’s beautiful process of Cyanotype, to obtain impressions of the plants themselves.
Impressions of the plants themselves. Not a picture of the plant. The plant, pressing its own shadow onto the page.
She began publishing in October 1843 and finished in 1853. Ten years. Three volumes. Somewhere around 400 plates in a complete set, and every single plate is an individual print — laid out, exposed, rinsed and dried one at a time. She wrote the text by hand too, because there was no way to combine printed type with these pages.
Now do a little arithmetic, the rough kind scientists do on the back of an envelope.
Seventeen copies are known to survive. Nobody knows how many she actually made. So take what we can count: 17 copies × about 400 plates ≈ 6,800 prints. Spread across ten years, that is roughly 680 a year, or about 13 every week, for a decade.
Then say out loud what that number is worth, because that is the real lesson. It leans too high in one direction — some surviving copies were never complete. It leans too low in the other, and by more, because copies certainly got lost, damaged or thrown away in 180 years. Two errors, pushing opposite ways, neither one measured.
So the number is not 6,800. But the size of it holds: thousands of prints, by one pair of hands. An estimate you can defend is worth much more than a precise number you cannot.
Who was first
You will see Atkins called the first woman photographer. That one is genuinely uncertain — Constance Fox Talbot has a claim too, and no photograph survives from either to settle it. Honest history has to leave that open.
The other claim is much firmer. Henry Fox Talbot published The Pencil of Nature in 1844 and called it the first attempt at a book of photographic plates. Atkins had put out her first installment in October 1843, several months ahead of him. Hers is the first book illustrated with photographs, and it is a book about seaweed.
Try it, with no chemicals at all
You can buy sun-print paper that does exactly what Atkins did. But there is a free version that teaches the same idea, and it works on any sunny windowsill.
- Take a sheet of colored construction paper. Dark blue, purple or red fade fastest.
- Lay flat things on it — leaves, a fern, keys, scissors, a paper snowflake you cut. Flat is what matters, so the shapes sit tight against the paper.
- Tape the sheet in a window that gets direct sun. Leave it. A few days is usually enough; a week is better.
- Lift everything off in one go.
You will get pale ghosts on a faded sheet. That is the same physics: light is doing chemistry to the paper, and whatever blocked the light is now a record of its own shape.
Two things are worth trying on purpose. Put one object down on day one and an identical one on day four, and lift both together — the difference between the two ghosts is a picture of time. And prop a piece of clear plastic over one corner. Say out loud what you think will happen before you lift it — will that corner fade like the rest, or less? Either answer tells you something real about what the plastic is letting through, and writing your guess down first is what makes it an experiment instead of a look.
Where this comes from
- Anna Atkins’s cyanotypes: the first book of photographs — Natural History Museum, London
- Rhapsodies in Blue: Anna Atkins’ Cyanotypes — The Public Domain Review, on her father, her method, and the Talbot comparison
- Photographs of British algae: cyanotype impressions — New York Public Library, the plates themselves, scanned
- From blue skies to blue print — Harry Ransom Center, on Herschel’s invention and the words he coined
- The Great Wave: spot the difference — British Museum, on Hokusai and imported Prussian blue
The diagrams above are original drawings made for this post, not reproductions of any plate.