The Enigma machine was a German electromechanical cipher machine: press a key, and current ran through a plugboard, three turning rotors and a reflector to light a different letter. Polish mathematicians broke it first, in 1932; Bletchley Park built on their work and read Enigma traffic for the rest of the war.
1918
filing date of the first Enigma patent
1932
the year Polish codebreakers first read military Enigma
103 sextillion
possible daily settings, by Bletchley Park’s count
211
British bombes built to find those settings
Key takeaways
- Enigma grew out of a German patent filed on 23 February 1918 and credited to Arthur Scherbius; commercial machines were on the market from 1923.
- Each key press sent current through the plugboard, the rotors, a reflector and back, so the machine could never encipher a letter as itself.
- The daily setting worked like a combination: rotor order, ring settings, start positions and plugboard cables. Its key space dwarfs a three-wheel safe dial’s 1,000,000.
- Marian Rejewski, Jerzy Różycki and Henryk Zygalski broke it in Poland and handed their methods to Britain and France at Pyry on 25 July 1939.
- Turing and Welchman’s bombe, captured key material and German operators’ habits did the rest. German operating habits gave away as much as the machine’s design.
In This Article

The Enigma machine at a glance
| What it was | An electromechanical rotor cipher machine with a keyboard, lampboard, rotors, reflector and (military models) a plugboard |
| Inventor | Arthur Scherbius; patent DE 416 219 filed 23 February 1918 |
| Reflector idea | Willy Korn, Scherbius’s colleague |
| Commercial sale | From 1923; glow-lamp Enigma A from 1924 |
| Main military model | Enigma I, final version in service from 1 June 1930; more than 20,000 made |
| Power | 4.5 V battery |
| Rotors | Three in the machine, chosen from five (army and air force) or eight (navy) |
| Naval four-rotor model | M4, introduced on 1 February 1942 |
| First broken | By the Polish Cipher Bureau, December 1932 |
| Bletchley Park’s first wartime breaks | January 1940 |
Sources for the table: Crypto Museum’s histories of the Enigma, the Enigma I and the M3 and M4,1234 the patent itself,5 the Polish Foreign Intelligence Agency6 and Bletchley Park Trust.7
How the Enigma machine worked
An Enigma looks like a typewriter that has lost its paper. There’s a keyboard at the front, a panel of 26 lettered lamps above it, three small windows showing a letter on each rotor, and on military machines a row of sockets along the front for the plugboard. Most Enigma I machines travelled in an oak case with the machine bolted to the bottom.2 Nothing is printed. The operator reads each lit letter and someone writes it down.
Here’s what happened when an operator pressed a key on an Enigma I, following Crypto Museum’s walk-through.8
- The rotors move first. The right-hand rotor steps on before the contact closes, so it turns on every key press. A notch carries the next rotor on, much like an odometer.
- Plugboard. Current from the 4.5 V battery passes through the plugboard, where each cable swaps a pair of letters.
- Entry disc and rotors. It enters the right-hand rotor, then the middle and left ones. Each rotor is a wired scramble of the alphabet.
- Reflector. The reflector (Umkehrwalze) joins the 26 letters in pairs with 13 wires and sends the current back through the rotors by a different path.
- Back out. The current goes through the plugboard a second time and lights a lamp.

Two consequences fall straight out of that reflector. The machine was reciprocal: with the same settings, typing the ciphertext gave back the plain text, so one machine did both jobs. And because the return path always differs from the way in, a letter could never be enciphered as itself.8 The Germans got convenience. The codebreakers got a rule they could test every guess against, as we’ll see.
The rotors also explain why the same letter typed twice came out differently. The right-hand rotor had already moved before the second key press closed its contact, so the current took a new path each time. With three rotors, the positions repeat only after 16,926 key presses, slightly fewer than 26 × 26 × 26 because of a quirk called double stepping.8
The setting was a combination: Enigma vs a safe dial
This is where Enigma meets the rest of this site. A three-wheel safe lock opens when you dial the right three numbers in order. An Enigma operator couldn’t read anything until the machine was set to the day’s key, and that key was a combination with four parts. Each day’s sheet listed the rotor order (Walzenlage), the ring settings (Ringstellung) and the plugboard pairs (Steckerverbindungen).9 The starting position of the rotors completed it. Even the words overlap: the British called the rotors wheels, and a safe lock’s wheel pack does a similar job of storing a secret position.10
Crypto Museum works out the size of that combination for the army’s Enigma I, using three rotors chosen from five and ten plugboard cables, and assuming the enemy already knows how the rotors are wired.8
| Part of the setting | Enigma I | Three-wheel safe lock |
|---|---|---|
| Order of the wheels | 60 (5 × 4 × 3) | Fixed by the maker |
| Ring settings | 676 (26 × 26; the left ring makes no difference) | None |
| Starting positions or numbers | 17,576 (26 × 26 × 26) | 1,000,000 (three wheels, dial numbered 0 to 99) |
| Plugboard, 10 cables | 150,738,274,937,250 | None |
| Total | 107,458,687,327,250,619,360,000 (about 1.07 × 1023) | 1,000,000 |
Sargent & Greenleaf sells its three-wheel Model 6730 with “1,000,000 possible combinations”, which is 100 dial numbers on each of three wheels.11 Even that figure flatters the dial a little, because S&G’s own instructions tell you to avoid certain numbers when you set a new combination.12 Put the two side by side and a single day’s Enigma key had roughly a hundred million billion times as many possibilities as the safe. Bletchley Park puts it at 103 sextillion settings, usually changed every 24 hours.713 If you’ve only ever met a dial on a padlock, our guide to opening a combination lock shows the right, left, right routine that stores those three numbers.
The two machines kept their secret in the same way, though. A safe dial and an Enigma are only as good as the people holding the numbers. Our page on how old safe combination locks work covers the dial side; on the Enigma side, the paperwork is where the story gets human.
The key sheet and the soluble ink
Army and air force key sheets held a month of settings, one line per day. An officer tore the day’s strip off the bottom of the sheet for the radio operator, and after use the strip was destroyed. The navy split its keys in two: the rotors, ring settings and reflector changed only on odd-numbered days, the plugs and basic setting every day.9 On most military machines an officer, not the operator, set the inside of the machine and could lock the front panel.9
The U-boat arm went further. Its codebooks were printed in red water-soluble ink on pale red paper, with a printed warning, in German, that the print was water-soluble. Thrown into the sea, they wiped themselves.14 We set out to confirm that the daily Enigma key sheets were printed this way too, but the sources we read confirm the soluble ink only for the naval codebooks and bigram tables, so that’s all we’ll claim. It’s a lovely piece of design either way: a document built to fail safe. The modern equivalent is the SF-700 form for a US government safe, where the written combination is itself classified, as our explainer on SCIFs and the safes inside them describes.
Who invented the Enigma machine?
Arthur Scherbius, a German engineer, is credited with the Enigma. His first patent application, for a “Chiffrierapparat”, was filed on 23 February 1918, and a prototype had been built in 1917, according to Crypto Museum.1 The patent, DE 416 219, was published only in 1925, in the name of the company Gewerkschaft Securitas. It describes wired “intermediate line carriers” that can be drums or cylinders, stepping on one place each time the one in front completes a turn, with lamps at the receiving end. Its own example uses three rotors of 25 positions for 15,625 key settings.5 Small numbers by later standards, but the idea is all there.
Scherbius wasn’t alone. Crypto Museum lists several rotor-machine inventors working between 1917 and 1919, among them the Dutchman Hugo Koch, whose patent Scherbius later bought. The reflector, which gave Enigma its reciprocal action and its fatal flaw, was the idea of his colleague Willy Korn, who would later run the company.1 The firm’s names changed several times: Scherbius & Ritter, Gewerkschaft Securitas, then Chiffriermaschinen Aktiengesellschaft from 1923.1
Enigma began as a business machine. A large printing model, the Handelsmaschine, was marketed from 1923, followed by the smaller glow-lamp Enigma A in 1924 and a run of commercial models through the 1920s. The German navy bought a version of the commercial Enigma C in 1926, and development of a military machine began the same year. The final version of the Enigma I, with its double-ended plugboard, went into service on 1 June 1930. Scherbius didn’t see any of the war: he died in an accident in 1929.1
The Polish break: Rejewski, Różycki and Zygalski
The story of breaking Enigma starts in Poznań, not Bletchley. Early in 1929 Polish military intelligence asked Professor Zdzisław Krygowski of the University of Poznań to pick students for a secret cryptology course. The choice of Poznań was deliberate: the city had been in the Prussian partition, so German was widely spoken. About 20 mathematicians took the course, and the Cipher Bureau kept three: Marian Rejewski, Henryk Zygalski and Jerzy Różycki.6 Breaking ciphers had long been seen as work for linguists. The Poles bet on mathematicians who also knew German.
Rejewski started on Enigma late in 1932 and soon worked out the wiring of the machine’s rotors.1 He had help from an unexpected quarter. A German in the cipher office, Hans Thilo Schmidt, was selling documents to French intelligence, and the French passed copies on to Warsaw. With material brought by Captain Gustave Bertrand on 8 December 1932, the Polish team decrypted Enigma messages for the first time in the last days of that December, in the Cipher Bureau’s offices in the Saxon Palace.6 Bletchley Park’s own summary is blunt about the order of events.7
“The Poles had broken Enigma in as early as 1932.”
Bletchley Park Trust
The Germans kept changing the procedure, and the Poles kept building answers. A Warsaw radio firm, AVA, made a dozen or so replica Enigmas, in use by 1934. In 1936 came the cyclometer, which could recover the current key within minutes. In 1938 Rejewski’s team built the bomba, a machine that searched for settings automatically, and Zygalski designed perforated paper sheets to find the rotor order.6 The Germans dropped their common basic setting on 15 September 1938, and around the same time added two more rotors, IV and V, to the three in use.1 Choosing three rotors from five meant 60 possible rotor orders instead of six.
By the summer of 1939 war was weeks away, and Poland chose to share. At a first meeting in Paris in January 1939 the Poles had said nothing of their success. On 25 July 1939 they took the French and British delegations, including Alastair Denniston, head of the Government Code and Cypher School, and Dilly Knox, to their decryption centre at Pyry, outside Warsaw. The guests were shown how the Poles recovered keys, and each delegation was given a Polish-built Enigma replica with the materials to work it.6 After the German invasion the Poles destroyed their bomby.15
Timeline
- 1918 Scherbius’s cipher machine patent is filed (23 February).
- 1923 The printing Enigma goes on the commercial market.
- 1930 The final Enigma I, with plugboard, enters German service (1 June).
- 1932 Rejewski reconstructs the wiring; first Polish decrypts in December.
- 1938 The Polish bomba is built; Germany adds rotors IV and V.
- 1939 The Pyry meeting hands Polish methods to Britain and France (25 July).
- 1940 Bletchley Park’s first wartime breaks (January); the first British bombe arrives (March).
- 1941 U-110 is captured with its Enigma material (9 May).
- 1942 The four-rotor M4 goes into U-boat service (1 February); U-559’s codebooks are seized (30 October).
Bletchley Park, Turing and the bombe
At Bletchley Park, Dilly Knox, a codebreaker since the First World War, set up the Enigma Research Section, starting with Tony Kendrick and later with Peter Twinn, Alan Turing and Gordon Welchman, working in the stable yard. The first wartime Enigma messages were broken in January 1940, and from then on Enigma was broken routinely for the rest of the war.7 Turing was lodging near the park from 4 September 1939, the day after Britain declared war, and he worked on naval Enigma in Hut 8.16
In late 1939 Turing and Welchman designed a new machine, the British bombe.16 It took its name from the Polish bomba and was partly based on it, but it worked differently. The Polish machine exploited the Germans’ habit of enciphering each message key twice. Turing’s bombe instead started from a crib: a guess at a piece of plain text, matched against the ciphertext.15 Here the reflector’s flaw paid off. Since no letter could stand for itself, a crib could only sit where none of its letters matched the ciphertext underneath.17 The bombe then ran through rotor positions looking for settings consistent with the guess, and the codebreakers tested what it found.
The first bombe, named Victory, was working at Bletchley Park in March 1940: The National Museum of Computing gives 14 March, Crypto Museum 18 March.1715 It was slow. Welchman’s contribution, the diagonal board, used the plugboard’s pairing of letters to cut down false stops dramatically, and the second machine, Agnus Dei (soon Agnes), had it from August 1940.1715 The British Tabulating Machine Company built the bombes at Letchworth under Harold “Doc” Keen, and about 211 were built and deployed during the war, many at outstations such as Eastcote and Stanmore.1517
The people who ran them are easy to forget. Hundreds of bombes were operated by Wrens, women of the Women’s Royal Naval Service, in long shifts in dark, stuffy rooms; Bletchley Park’s own description of the work is “boring and oppressive”.13 A rebuilt bombe today weighs a ton and holds 12 miles of wire.17 We can only imagine the clatter of a hut full of them at three in the morning.
Captures and the four-rotor problem
Naval Enigma was the hardest. The navy’s machines drew three rotors from a box of eight, not five.3 Tony Sale, the first curator of the Bletchley Park Museum, wrote that it had “just too many unknowns” to be solved by cryptography alone, so captured material, known as pinches, was essential.18 Keys taken from a disguised German trawler off Norway in April 1940 and from the Krebs in March 1941 helped Hut 8 fill in the gaps.18
The best-known pinch came on 9 May 1941. In the mid-Atlantic a boarding party from HMS Bulldog, under Commander Joe Baker-Cresswell, got aboard the disabled U-110. Led by 20-year-old Sub-Lieutenant David Balme, they brought out an Enigma machine, its rotors and months of cipher material.19 The haul included key tables, officer-only procedures, a bigram book and instructions for emergency keys.18 The capture was kept so secret that few people on the British side knew of it.19 The safes and strongrooms that kept Britain’s own secrets in the same years are the subject of our page on safes and vaults of the Second World War.
Then the door shut again. On 1 February 1942 the U-boats switched to the four-rotor Enigma M4. Its extra thin rotor, Beta or Gamma, sat beside the other three and didn’t step. Set one particular way, the M4 could still work with three-rotor machines. The U-boat procedure was called Triton; Bletchley Park called the traffic Shark, and it stayed unreadable for about nine months.4
The way back in cost two lives. On 30 October 1942 HMS Petard forced U-559 to the surface in the eastern Mediterranean. First Lieutenant Anthony Fasson and Able Seaman Colin Grazier boarded the abandoned boat with Tommy Brown, a canteen assistant. Fasson found the documents and passed up the short weather codebook and the short signal book. Fasson and Grazier went back below and drowned when the U-boat went down; Brown got clear. Fasson and Grazier received posthumous George Crosses and Brown the George Medal. The books reached Bletchley Park on 24 November 1942 and were vital in getting back into Shark.10
Four rotors needed faster bombes. In the United States, National Cash Register in Dayton, Ohio, built US Navy bombes to a design by Joe Desch; prototypes ran by mid-1943, and 120 machines were installed by December 1943.15
Why Enigma failed: procedure as much as maths
No codebreaker ever tried 103 sextillion settings. They didn’t need to, because the way Enigma was used leaked information. The pre-war procedure had operators encipher their chosen message key twice, so the first and fourth letters of every message were linked, and many messages on the same day started from the same position. That was the opening the Poles used.9 The Germans changed it in 1940, but by then other habits were doing the damage.
Operators were supposed to choose random message keys. Many didn’t. Bletchley veteran Susan Wenham remembered operators using their girlfriends’ or their own initials, and Mavis Lever said the hunt for these “cillies” became something of an art. John Herivel guessed that tired operators would leave the rotors near the day’s ring settings and send what was almost the ring setting in clear; the “Herivel tip” paid off. One operator sent a dummy message by pressing nothing but the L key. Because Enigma couldn’t turn L into L, the ciphertext had no L in it at all, and that gave away a new rotor’s wiring.20 Add the reflector’s rule that no letter could be itself, which made cribs testable, and the safe-lock comparison holds. A big number doesn’t help if people write the combination on the door.
Alan Turing broke Enigma.
Polish mathematicians broke it first, in 1932. Turing’s bombe, Welchman’s diagonal board, Bletchley’s huts and captured material kept it broken through the war.
The British got Enigma by capturing a machine.
The Poles gave Britain a working replica at Pyry in July 1939. Later captures mattered for keys and naval procedures.
Enigma was beaten by trying every setting.
The bombe tested settings against a guessed piece of plain text. Operator habits and captured keys cut the search to something possible.
The Imitation Game vs history
The 2014 film is why many people search for Enigma at all, and it’s a drama, not a documentary. Its biggest compressions are the ones this page has already covered. The machine Turing builds is called Christopher; the real ones were bombes, named after the Polish bomba, and Victory was the first.2115 Gordon Welchman doesn’t appear, and the diagonal board is credited to Hugh Alexander.2221 The bombes weren’t built by Turing in a hut but by the British Tabulating Machine Company at Letchworth.15 The film’s subplot with the Soviet spy John Cairncross is, in the words of Turing’s biographer Andrew Hodges, invented: Cairncross worked at Bletchley but not in Turing’s section.21
The Polish contribution gets a passing line. That’s the one we’d most like viewers to look up, because it changes the shape of the story: Bletchley didn’t start from nothing in 1939, it started from Warsaw.226
Where to see an Enigma machine
- Bletchley Park, Buckinghamshire. The wartime codebreaking site, run as a museum by Bletchley Park Trust.7
- The National Museum of Computing, also at Bletchley Park. Home of the rebuilt Turing-Welchman bombe, finished in 2007, in Block H; a working Enigma is regularly demonstrated in the bombe gallery.1715
- National Cryptologic Museum, Fort Meade, Maryland. Holds the one US Navy bombe that survived the war.15
- Imperial War Museum, London. The machine in our labelled photo was photographed there in 2004; check the museum’s current displays before you go.
When you stand in front of one, look at the rotors. Each has an alphabet ring round its edge, read through a little window like the numbers on a dial, and a notch that carries the next wheel round. You’ll find the same principle in the Yale and Wilson combination lock patent of 1843, which held its bolt until notches in its disks lined up.23 The wider story of how wheels, gates and levers came to guard safes is in our history of lock mechanisms in safes.
Frequently asked questions
Who invented the Enigma machine?
Arthur Scherbius, a German engineer. His first patent application was filed on 23 February 1918, and commercial machines were sold from 1923. The reflector was his colleague Willy Korn’s idea.
Who broke the Enigma code first?
The Polish Cipher Bureau. Marian Rejewski worked out the rotor wiring late in 1932, and with Jerzy Różycki and Henryk Zygalski the bureau read Enigma for years before handing its methods to Britain and France in July 1939.
How many settings did the Enigma machine have?
For the army’s Enigma I with three of five rotors and ten plugboard cables, about 1.07 × 1023 before allowing for double stepping, or the 103 sextillion Bletchley Park quotes after it.
Why couldn’t Enigma encipher a letter as itself?
The reflector sent the current back through the rotors by a different path from the one it came in on, so the lit lamp was always a different letter from the key pressed. Codebreakers used that rule to place cribs.
Did Alan Turing break Enigma on his own?
No. He designed the British bombe with Gordon Welchman, building on the Polish bomba, and worked on naval Enigma in Hut 8. Many other people, captured material and the Poles’ head start all played a part.
Is an Enigma setting like a safe combination?
In principle, yes: both are secret wheel positions that have to be set exactly. A three-wheel safe lock has 1,000,000 combinations; one day’s Enigma key had about 1023.
Sources:
- Crypto Museum, “History of the Enigma” (accessed 12 November 2026). ↩︎
- Crypto Museum, “Enigma I” (accessed 12 November 2026). ↩︎
- Crypto Museum, “Enigma M3” (accessed 12 November 2026). ↩︎
- Crypto Museum, “Enigma M4” (accessed 12 November 2026). ↩︎
- German patent DE 416 219, “Chiffrierapparat”, applicant Gewerkschaft Securitas, filed 23 February 1918, published 8 July 1925, via Google Patents. Crypto Museum identifies it as Scherbius’s first Enigma patent. ↩︎
- Agencja Wywiadu (Polish Foreign Intelligence Agency), “Enigma decryption” (accessed 12 November 2026), based on the agency’s archives and the Polish literature it lists, including Kozaczuk and Rejewski. ↩︎
- Bletchley Park Trust, “Enigma” (accessed 12 November 2026). ↩︎
- Crypto Museum, “How Enigma works”, including the section on the number of possible settings (accessed 12 November 2026). ↩︎
- Dirk Rijmenants, Cipher Machines and Cryptology, “Enigma message procedures” (accessed 12 November 2026). ↩︎
- Tony Sale, Virtual Wartime Bletchley Park, “The Breaking of German Naval Enigma: U559”. ↩︎
- Sargent & Greenleaf, Model 6730 mechanical combination lock, product page (accessed 12 November 2026). ↩︎
- Sargent & Greenleaf, Mechanical Safe Lock Guide (© 2006), “Model 6730, 6735, 6738, 6741 Mechanical Combination Lock: Operating/Changing Instructions”, PDF, archived copy of January 2026. ↩︎
- Bletchley Park Trust, “6 facts about the Bombe” (accessed 12 November 2026). ↩︎
- Crypto Museum, “German naval bigram tables” (accessed 12 November 2026). ↩︎
- Crypto Museum, “Turing Bombe”, with the sections on the US Navy Bombe and the rebuild (accessed 12 November 2026). ↩︎
- Andrew Hodges, The Alan Turing Internet Scrapbook, “Alan Turing and the Second World War” (accessed 12 November 2026). ↩︎
- The National Museum of Computing, “The Turing-Welchman Bombe” (accessed 12 November 2026). ↩︎
- Tony Sale (first curator of the Bletchley Park Museum), Virtual Wartime Bletchley Park, “The Breaking of German Naval Enigma: early pinches”. ↩︎
- Royal Museums Greenwich, Caird Library catalogue record for David Balme, Enigma: The Untold Story of the Secret Capture (Whittles Publishing, 2016), with its summary of the U-110 capture. ↩︎
- NOVA (PBS), “Decoding Nazi Secrets: the mind of the codebreakers”, interviews with Bletchley Park veterans including Mavis Lever, Susan Wenham and John Herivel (updated November 2000). ↩︎
- History vs Hollywood, “The Imitation Game (2014)”, quoting Turing’s biographer Andrew Hodges (accessed 12 November 2026). ↩︎
- Christos Triantafyllopoulos, Christos military and intelligence corner, “Fake history: The Imitation Game movie review” (9 January 2015). ↩︎
- US Patent 3,312, Linus Yale and Charles Wilson, “Combination safety-lock for doors”, 20 October 1843, via Google Patents. ↩︎
The History of Safes letter
One old safe a month, in your inbox
Once a month I send one safe worth a closer look, the new maker histories and a find from the patent and advert collection. Nothing else, and you can leave with one click.
You'll get a confirmation email first. One email a month, unsubscribe any time. Privacy policy.



