Skip to content
clickidy
🎯 ALL GAMES

READING · 26 MIN

WRITE SOMETHING
NO ONE CAN SEE.

A blank page that isn't blank at all - invisible ink is one of the oldest tricks in espionage, and it's simpler than you'd think. Here's how it actually works, and how it got real spies caught.

The oldest trick in the letter ◢

The first record of it comes from Pliny the Elder in the first century AD, who wrote that the milky sap of the tithymalus plant could be used to write invisibly: it dried clear, then darkened when gently heated. Everything since - lemon juice, milk, onion juice, diluted honey - works on the same idea. These are organic inks, and they all rely on one property: heat.

Why lemon juice actually works ◢

Write with lemon juice and the water evaporates, leaving a thin film of citric acid soaked into the paper fibres - invisible to the eye. Hold the page near heat and the acid weakens those fibres, so the written areas scorch and oxidise faster than the paper around them, turning brown well before the rest of the sheet catches. You're not revealing ink; you're revealing damage. It's real chemistry, it's safe to try with things already in your kitchen, and it's exactly why organic inks were only ever a parlour trick to a professional spy - any censor with a hot iron could find them in seconds.

The spies who got caught ◢

By World War I, both sides had moved past kitchen chemistry. German intelligence recruited the American journalist George Vaux Bacon and equipped him with ink impregnated into black wool socks - soak one end in water and it produced a pale liquid he wrote with on rough paper. The active ingredient was Argyrol, a commercial silver compound, diluted as far as 1 part in 500,000 to defeat testing. It didn't work. British physicist Thomas Ralph Merton used spectroscopy to detect trace silver in the treated cloth, while French chemist Gaston Edmond Bayle's team developed a way to deposit silver out of the paper using electrical current. Bacon was arrested in 1915. In 2011 the CIA declassified its six oldest classified documents, all secret-writing files from 1917 and 1918. One of them lists five inks - a solution of nitrate of soda and starch among them - each paired with the chemical wash that brings it out, which is what a genuine reagent ink is: chemically inert until a specific developer makes the writing appear.

A war fought by chemists, not writers ◢

That split - heat-reveal inks anyone could make, versus reagent inks only a lab could develop - defined the whole discipline. Britain's Special Operations Executive treated invisible ink as a backup channel at best, training agents to avoid relying on it when a slip meant execution. It still turned up in the field: WWII entertainer and French Resistance agent Josephine Baker reportedly carried notes on German troop movements written in invisible ink on her sheet music, hidden behind attention nobody thought to search. A message like that hides from anyone but the person meant to read it - so does a good cipher; try one yourself →

Why it stopped working ◢

Invisible ink lost the arms race it started. Ultraviolet light made some chemical inks fluoresce under a lamp, letting censors screen huge volumes of mail in seconds. Iodine vapour reacted with residue an eye alone would miss. Heat, the original giveaway, never stopped working on the organic inks. By the mid-20th century, cryptography and radio had made physically hiding a message far less reliable than scrambling one - which is really the same idea invisible ink started with, just moved from chemistry into math.

Play the ciphers in this story

CODESCaesar Cipher Decoder

Shift the alphabet to encode - or brute-force all 25 to crack it.

SHIFT 1-25 · ROT13CODESCodes & Ciphers

Morse, binary, Caesar, more - encode anything.

10 CIPHERS

More Codes & Ciphers history →

WHAT THE RECORDS SHOW ◢

Professional invisible ink was never one bottle. It was two, sent separately, and the second one was the difficult half. George Washington called them the stain and the counterpart, and when he shipped them to Long Island in July 1779 he numbered the phials so nobody would confuse them.

Everything that made the technique worth using, and everything that eventually killed it, follows from that split.

Key Takeaways

  • Washington sent Benjamin Tallmadge two phials in 1779: the ink, and "the Counterpart which renders the other visable by wetting the paper with a fine brush after the first has been used & is dry".
  • The supply of the counterpart was a standing crisis. Robert Townsend, Washington's confidential correspondent in New York, reported in December 1779 that he was "now entirely out of the Counter-part of the stain".
  • The German formula the Americans were sent in June 1918 was a pharmacy tablet dissolved in water. What the file guards is not the ink; it is the two-bath developer that reads it.
  • A wartime examiner's file released in 2011 sets out twelve tests to run on a suspect sheet, escalating from non-destructive to destructive.
  • The same file's method for opening sealed letters carries a two-line note in the margin: "Tried", then "not successful".

What did Washington's invisible ink actually consist of?

Two liquids, from a chemist who never wrote the recipe down.

The introduction came from John Jay, who told Washington on 19 November 1778 that his brother would explain "a mode of Correspondence, which may be of use, provided proper agents can be obtained. I have experienced its Efficacy by a three Years Trial." He added a condition: "We shall remain absolutely silent on the Subject."

The brother was James Jay, a physician working in chemistry. The editors of the Washington Papers record flatly that he never disclosed the recipe (Founders Online), and no recipe survives. The most consequential American secret ink of the Revolutionary War is, chemically, a blank.

What does survive is the shipping paperwork, and it is more useful than a recipe. Writing to Major Benjamin Tallmadge on 25 July 1779, Washington set out the system in one sentence:

"All the white Ink I now have (indeed all that there is any prospect of getting soon) is sent in Phial No. I. by Colo. Webb. the liquid in No. 2 is the Counterpart which renders the other visable by wetting the paper with a fine brush after the first has been used & is dry"

Phial one writes. Phial two reads. The reader wets the sheet with a fine brush, and only after the writing has dried. The naming is the editors' too: a note to Washington's letter of 5 February 1780 records that on the draft he wrote "Stain and counter-part of the stain", and glosses the pair as invisible ink and reagent.

The 1779 letter carries a warning that has nothing to do with chemistry and everything to do with sourcing. Washington had heard that Governor Tryon "has a preparation of the same kind, or something similar to it which may lead to a detection if it is ever known that a matter of this sort has passed from me." A rare reagent identifies its owner. Possessing it is itself evidence.

Why did a two-part ink beat one you could read with a candle?

Because of who else can read it.

A heat-developed ink has no gatekeeper, and the consequence is a distribution problem rather than a chemical one. Everyone who handles the letter is equally equipped to read it: the courier carrying it, the searcher at the ferry, the clerk who opens it. A reagent ink moves the secret off the paper and into a bottle held by one person at the far end of the chain.

Washington spelled out the operational payoff on 5 February 1780, correcting his own agent's habit. Writing on a bare sheet was the mistake: "he should avoid making use of the stain upon a Blank sheet of paper (which is the usual way of its coming to me)". His reasoning:

"Such a letter would pass thro' the hands of the enemy unsuspected, and even if the Agents should be unfaithful, or negligent, no discovery would be made to his prejudice, as these people are not to know that there is concealed writing in the letter, and the intelligent part of it would be an evidence in his favor."

That is the whole case for a reagent, in eighteenth-century terms. The couriers do not know they are carrying anything. If one of them defects, he has nothing to sell. And the visible letter is not merely cover: it is exculpatory, because a chatty note about family matters is what an innocent man writes.

The cost of that security is logistics. Two chemicals must reach two different people and stay in stock at both ends. They did not. On 9 April 1780 Washington told James Jay that "The liquid with which you were so obliging as to furnish me for the purpose of private corrispondence is exhausted", adding that he still had "a sufficiency of the materials for the counterpart on hand". Half a system is no system.

Four months earlier the shortage had run the other way, at the far end of the chain. Robert Townsend, whom Washington described to Congress as "my confidential Correspondent in New York", closed a dispatch of 27 December 1779 with one flat sentence: "I am, now entirely out of the Counter-part of the stain."

A fortnight before that, Abraham Woodhull had already reported the same failure and its consequence, in the ring's own vocabulary: "Culper Junr is intirely out of the Counterpart. and Unable to read any future Blank that may come." Woodhull added the part that closed the channel completely, since the obvious workaround was not available either: "we cannot get any Person to run the resque writen with common Ink." Both letters are preserved among Washington's papers, and both reach a modern reader through the notes to his report to Congress of 18 January 1780.

A "blank", in the ring's usage, was a sheet already written on in stain and still looking empty. Woodhull uses the word that way in the same letter, forwarding "a blank Just come to hand from C. Junr Which I conclude contains Some important Intelligence". Without the counterpart, a blank stayed blank, and the intelligence in it was as good as destroyed.

Jay's reply of 13 April 1780 explains why resupply was so slow. "The composition requires some assistance from Chemistry", he wrote, and he had nowhere to put a small brick furnace. Washington's answer in May was to build him one. He told Jay on 12 May 1780 that he had directed help "in erecting a small Elaboratory", and wrote the next day to Lieutenant Colonel Udny Hay to furnish "the necessary hands" and, if needed, boards "to compleat the building which is to be of Logs", explaining that Jay "has already furnished me with some Chymical preparations from which I have derived considerable advantages". A commander in the field stopped to have a log hut built for a chemist, because the alternative was a dead channel.

Where do you put writing that must not look like writing?

Not on a blank page, which is the one place everyone looks.

Washington's instructions of 24 September 1779 are the clearest surviving statement of secret-writing tradecraft in his correspondence. His agent should "occasionally write his information on the blank leaves of a pamphlet", or on the first pages of a pocket book, or on the blank leaves at either end of a yearly register, an almanac, any new publication, or a book of small value.

Then comes the detail that shows chemistry driving tradecraft. The choice of book, Washington wrote, should be governed "principally by the goodness of the blank paper as the ink is not easily legible unless it is on paper of a good quality." A cheap absorbent sheet spreads the ink and ruins the message. The carrier had to be worthless enough to discard and well made enough to write on.

His justification is a piece of intelligence about the opposition's search regime: books can travel "without risque of search, or the scrutiny of the enemy", because inspection "is chiefly directed against paper made up in the form of letters". Where the search is not, the message goes. That is the same instinct behind every physical concealment from a hollow tree to a dead drop.

Where a letter had to be used, he prescribed interlining: write an ordinary note "a little in the tory stile, with some mixture of family matters", and put the real report between the lines and on the unused part of the sheet.

The idea kept working for a very long time, and the files show it running at both ends of the skill range. A briefing compiled for American postal inspectors during the First World War notes that "A solution of common table salt or urine is often used by convicts in prison, to write between the lines of a letter, with a match. The confederate heats the paper to make it legible." Same trick, no chemist, no supply chain, and no protection at all against a warm iron.

What was in the German invisible ink of 1918?

The file says a headache tablet, dissolved in water. The interesting part was never the ink.

Six declassified files, which the CIA announced on release as "the United States Government's six oldest classified documents", are all about secret writing and all from 1917 and 1918. Each carries a stamp reading "APPROVED FOR RELEASE DATE: 31-Mar-2011". One of them is a report on a form pre-printed "Replying to O. N. I. No.", numbered 433 and dated 14 June 1918, headed "German Secret Ink. Formulas used in preparation of." It is written in French. A companion file reproduces it and adds a page headed "Translation from the French", and the English quoted below is that official translation rather than anything rendered here.

The ink is one compressed tablet, and the file offers a choice of two: "Compressed Pyramiden" or "Compressed or Powdered asperine", the file's own spelling. Whichever the writer used, "All of this must be mixed with 400 cms of pure water", the unit written that way in the document.

The developer takes up more of the page than the ink does. It is two separate washes. The first is a mixture based on camphorated alcohol, distilled water, nitrate of potash and acetic acid, and the file says it "Must be well mixed and applied with wet cotton". The second, built on acetic acid and alcohol with tincture of capsicum and quinine, "Must be applied after the first bath, in the same manner but one has to wait until the first solution is completely dry."

Burden The writer's side The reader's side
What it needs One compressed tablet, "Pyramiden" or aspirin Two prepared baths, five listed ingredients each
Water 400 cms of pure water Distilled water in both baths
Skill Dissolve and write Mix, apply with wet cotton, dry fully, apply again

Every line of that table comes from the file. The conclusion drawn from it does not: nothing in the document discusses portability, but a tablet travels in a pocket and two prepared baths do not.

Read the ratio rather than the chemistry. An agent needs one tablet and clean water. The person receiving the letter needs two baths of five listed ingredients each, mixed correctly, applied in the right order, with a full drying stage in between. Everything hard has been pushed to the receiving end, which is exactly where you want it: the end that is not being searched.

That inversion is what the phrase "invisible ink" hides. Across these files the ink is rarely the secret. The developer is.

Why the developer, not the ink, was the secret

Because the files are organised around the developer, and treat the ink as the interchangeable part.

Another of the six documents, headed simply "SECRET WRITING", makes the point in its layout. It gives five inks, numbered first to fifth, and then a separate numbered list of four developers under the heading "For developing secret writing is used". The third ink is named, but its developer is given only by cross-reference, as "2, or 4", which tells you what kind of paper this is. Developers are the stable, numbered, shared items. Inks are what you pick to suit them.

The fifth entry is the one worth stopping on. Lemon juice appears in this file too, but it is not filed under heat. It is listed as "Lemon juice, developed with Ferro cyanite of potas", a chemical reagent. A professional service treating the most domestic ink on the list as a reagent problem is a small thing that says a lot: by 1918 the development step was where the craft lived.

The same ordering shows up at the head of the briefing compiled for postal inspectors by Theodore Kytka, a San Francisco handwriting expert, and printed as "INVISIBLE PHOTOGRAPHY AND WRITING, SYMPATHETIC INK, ETC." with "CONFIDENTIAL" and "NOT FOR PUBLICATION" side by side at the head. Of its fifty numbered entries, the first ten are inks brought out with heat or sunlight. Entries eleven to seventeen each name a specific chemical developer instead. That sort holds only for the head of the list, which then turns to practical field notes and returns more than once to kitchen tricks and a warm iron, but for those seventeen entries the line between what anyone can read and what only a bench can read falls squarely between ten and eleven.

Meanwhile the United States was formulating its own. On 30 October 1917 an assistant chemist at the Bureau of Standards, part of the Department of Commerce, sent a memorandum transmitting "seven samples which seem to be the most suitable in the list which you submitted". Then this, which is the most human line in the whole set:

"Samples 1, 4, and 5 will doubtless exert a very corrosive action on steel pens, and therefore if suitable in other respects would have to be used with a quill pen."

A national standards laboratory, working out that three of its candidate spy inks would eat a nib. The chemist closes by offering to keep going: "if you will tell me just what characteristics you desire such a product to have, I will try to reproduce it." Secret writing had become a procurement specification.

How did examiners find invisible ink at scale?

By no longer looking for a particular ink, and testing every letter for all of them in a fixed order.

The most striking of the six declassified files is not a formula at all. It is a numbered examination sequence, twelve steps long, that reads like laboratory triage. Compressed, and keeping the file's own wording where it matters:

  1. Examine through a powerful beam of light at different angles, then photograph excluding certain rays.
  2. Expose to daylight for one or two hours. Discloses salts of gold and silver.
  3. Dust powdered charcoal over the surface and brush off well. Discloses paraffin.
  4. "Run a warm iron over the surface." Discloses a run of metal salts, sugar with sulphuric acid, and "the juice of Lemons, Onions, Leek, Cabbage, Artichoke".
  5. "Run a hot iron over the surface being careful not to scorch the paper." Discloses potassium hydroxide, sulphuric acid, potassium nitrate, copper nitrate.
  6. Wet with water. Discloses camphor.
  7. Expose to hydrogen sulphide gas. Discloses lead acetate and compounds of antimony, arsenic and tin.
  8. Dry in the air and wet with ammonia water. Discloses mercury and copper salts.
  9. Add hydrogen sulphide water to the part wet with ammonia. Discloses iron, antimony, tin, copper.
  10. Rinse with water and dry in the air.
  11. Wet with a solution of iron sulphate. Discloses gallic acid and potassium ferrocyanide.
  12. "To another part of the paper" add potassium ferrocyanide or tannin. Discloses iron salts.

Two features of that sequence are worth reading closely, and the reading is ours rather than the file's, since the document supplies the order without ever explaining it.

The first is the escalation. The opening steps do nothing to the sheet at all: light, daylight, a dusting of charcoal brushed off again. Heat arrives in two graded stages, warm before hot, with the caution "being careful not to scorch the paper" attached to the hotter one. Wet chemistry comes last. Whatever the compiler's reasoning, the effect of that order is that an early step never destroys what a later step would have found, which is exactly what you need if you do not yet know which ink you are chasing.

The second is the ending. Step twelve completes a division the sequence had already begun at step eight, when the ammonia treatment was confined to one part of the sheet: it sends the examiner to "another part of the paper" for the last reagent. By then the sheet is being spent rather than examined, and the procedure divides it instead of ruining all of it at once.

No individual test here is clever. The sequence is what does the work, because it turns a chemist's judgment into a numbered list that a clerk can follow. That is the shift that matters. Skill has moved out of the examiner and into the procedure, and a procedure scales to a mail sack in a way that a chemist never could.

The infrastructure followed. Kytka's briefing was, its imprint says, "Printed by the San Francisco Division for the information of Post Office Inspectors", and its entry forty-nine tells them to scrutinise "Type-written sheets, printed music, printed newspapers and pamphlets" and to examine suspect paper in sunlight before resorting to the heat test. Washington's best hiding place, the blank leaf of a printed book of small value, is now item forty-nine on a postal inspector's checklist.

Britain built the same machinery. The National Archives at Kew holds the Security Service file KV 3/2, catalogued as "Invisible ink and secret writing: first volume on World War 1 secret writing" and covering 1917 to 1919. The India Office records at the British Library include a file covering 1915 to 1919 on "Censorship arrangements for mails from prisoner of war camps in India; testing of mails for writing in invisible ink". Note the noun in the catalogue title: arrangements. Testing of mails had stopped being an investigation and become a standing procedure with staff, premises and a budget.

Once that exists, the arithmetic turns against the writer, and it is the arithmetic rather than any single discovery that settles it. A spy has to get one letter through, and has to hold two chemicals, a suitable carrier and good paper to do it. An examining service has to test every letter, which sounds like the far harder job until the test is a numbered list, a warm iron and a few seconds of a clerk's time. The cost of writing stayed high and the cost of finding collapsed. Every technique that survives that comparison works by making a message unreadable in plain sight rather than invisible, which is the game a sculpture like Kryptos is still playing in a courtyard where anyone can photograph it.

What the files do not settle

Three honest gaps, because the documents are working papers and working papers hedge.

The German formula reached the Americans at second hand. The file says only that it is "a copy of the formulas used preparation of German secret ink", attributed to a source given as initials. Nobody in the file claims to have tested it, and an intelligence report of a recipe is not proof the recipe worked.

The method in the same set for opening sealed letters without detection, a mixture of copper acetol arsenate, acetone and amyl alcohol heated in a water bath so the rising steam would dissolve the sealing material, carries a two-line note in the margin of the handwritten page: "Tried", and beneath it, "not successful". Somebody ran it and it failed. Who, and when, the file does not say.

And James Jay's stain remains unrecovered. Washington used it for years, had a laboratory built for its inventor, and never recorded what was in it. The letters name no ingredient, so the honest answer to "what was in Washington's invisible ink" is that we do not know, and a reconstruction offered anywhere else is not coming from this correspondence.

Frequently asked questions

What is the difference between the stain and the counterpart?

They are the two halves of one system. The stain is the ink the agent writes with, and the counterpart is the chemical developer that makes it visible. Washington described the counterpart in July 1779 as the liquid "which renders the other visable by wetting the paper with a fine brush after the first has been used & is dry", and the editors of the Washington Papers gloss the pair as invisible ink and reagent. Sending them separately is the point: a courier carrying only the written letter has nothing that can be developed, and nothing to betray.

Do we know what was in George Washington's invisible ink?

No. It was supplied by James Jay, a physician working in chemistry, who according to the Washington Papers editors never disclosed the recipe. Jay's own letter of 13 April 1780 says only that "The composition requires some assistance from Chemistry" and that he needed a small furnace to make more of it. Washington had a small laboratory built for him the following month. Any specific formula you see attributed to the Culper ring is a later reconstruction rather than something the correspondence records.

What was in the German secret ink formula the CIA declassified?

The file reports the ink as a single compressed pharmaceutical tablet dissolved in 400 cms of pure water, offering Pyramiden or aspirin interchangeably, and gives far more space to the developer: two separate washes applied with wet cotton, the second only after the first had dried completely, built on camphorated alcohol, acetic acid, nitrate of potash, tincture of capsicum and quinine among others. Treat it as a report rather than a proven formula, because it reached the Americans at second hand and nobody in the file claims to have tested it. It is dated 14 June 1918 and carries a release stamp dated 31 March 2011, alongside five companion files.

How did wartime examiners test letters for invisible ink?

With a fixed sequence rather than a hunch. One of the declassified files sets out twelve numbered steps that begin with raking light and photography, move through daylight exposure and powdered charcoal, then apply a warm iron and a hot iron, then water, gases and specific reagents, and finally turn to "another part of the paper" for a destructive test. The file gives the order and never explains it, but the effect of running gentle tests before harsh ones is that an early step does not destroy what a later step would have found, and the whole thing can be worked through by a clerk rather than a chemist.

Why did secret ink stop being a serious method?

Because the examiner's job got cheap and the writer's did not. Writing needed a scarce reagent kept in stock at the receiving end, a carrier that survived inspection, and paper of good enough quality to hold the ink. Examining needed a numbered checklist and a warm iron, applied to everything. Kytka's briefing for postal inspectors ends with the line that sums up the losing side of the trade: "The rule is to suspect and examine every possible thing. The war between the spy or forger and the expert is continually bringing out new methods."

If the two-bottle problem is the interesting part, the next thing to read is what happens when the reading key stops being a liquid and becomes an object you can hold, in the history of the decoder ring. Working from the other end, from a message rather than a method, the cipher identifier is the place to start.

Sources