The successes of French and British linguistic cryptanalysts, working on methods perfected during the First World War, persuaded their masters to ignore the problems of mathematical cryptanalysis. This was why the Enigma machine defeated them. The Poles had superior mathematicians, more men familiar with the German language, and the will to succeed at a task no one else believed possible.
By 1933 the Poles had rigged up a reproduction of the Enigma. They kept the French informed about their progress and the French faithfully passed to Warsaw the new codes and whatever mechanical changes to the machine Hans-Thilo could discover. However over a period of five years few of these messages got to the codebreakers. The Polish high command wanted its men to crack the German codes without outside help, and it doled out the material from Hans-Thilo only in small amounts when the codecrackers were stuck.
The Germans improved their coding machines while the Poles improved their codecracking ones, constructing what they called a bombe, a computer consisting of six linked Enigma machines. In September 1938 the fears of the Polish intelligence chiefs came true. Hans-Thilo was transferred to Göring’s Forschungsamt and the supply of codes ended. But by now the codecrackers had learned to manage without his help.14
On 24 July 1939, with war not far away, the Poles invited the French and the British to Warsaw to show them in great detail the work they were doing breaking the German codes. They showed them the bombe, a method of using overlaid perforated sheets, and calculations about wiring. Mathematical talent was at the heart of the Polish work. Some of the most notable breakthroughs had been made by Marian Rejewski, a young mathematician who has been described as one of the greatest cryptanalysts of all time.15
As a going-away present both French and British representatives were given a ‘replica’ Enigma machine. Although as the Germans got nearer to war they changed the codes, and added an extra rotor to their Enigmas to make the machine codes more complex, this gift was beyond price. Gustave Bertrand, a senior French intelligence officer, described how he brought the machine to the chief of Britain’s Secret Intelligence Service.
On 16 August 1939 I was on my way to London accompanied by Uncle Tom – the diplomatic courier of the British Embassy in Paris – who was carrying a diplomatic bag with the Enigma machine. At Victoria Station Colonel Menzies, head of the [S]IS wearing the rosette of the Legion of Honour in the buttonhole of his dinner-jacket (he was going to a soirée) was waiting for us: triumphal welcome! Which occasioned him to say one day [that the French intelligence service] had done him a ‘considerable service on the eve of the war’.16
Considering how little accurate information his SIS was able to supply about even peacetime Nazi Germany, the ill-judged rejection of the French offer of the Enigma secrets and a total lack of any preparations for war, Colonel Menzies did not exaggerate.
Poland was invaded, but keeping ahead of the Germans the Polish codecrackers moved to France. Then France fell too. The French team escaped the German invaders, set up shop near Uzès in the unoccupied sector of France, and continued their work, sending their solutions to London (enciphered in Enigma!). Many Polish and French cryptologists ended up as captives of the Germans, but all managed to convince their interrogators that Enigma was beyond their abilities. The Germans believed them; such is the power of self-deception.
In 1943 some Polish members of the original team reached England after harrowing experiences. According to one notable historian of the Enigma story: ‘The Poles reaped the customary reward of the innovator whose efforts have benefited others: exclusion. The British kept Rejewski and the others from any work on Enigma, assigning them instead to a signals company of the Polish forces in exile, where they solved low-level ciphers. It was not one of Britain’s finest hours.’17
Bletchley Park was sited halfway between Oxford and Cambridge universities, and with more and more big wooden huts built in the grounds, GC&CS were able to recruit and accommodate academics including the junior dean and mathematical tutor of Sidney Sussex College who arrived on the first day of war. Hurrying back from an International Chess Olympiad in Buenos Aires came Britain’s chess team. More such men and women followed.
No one in the world had ever attempted to break machine-enciphered messages on a regular basis with all the urgency that war brings to such a task. The resulting intelligence – eventually to be called Ultra – came solely from radio transmissions in Morse code. No teleprinter message or telephone conversation was included, and since most of the vital messages were sent by those means, the greater proportion of secret enemy communications were never intercepted. Sometimes a radio message in Enigma was answered by telephone or teleprinter, or vice versa, so that only one side of a dialogue was available. Radio reception was often subject to interference, and errors were commonly included, making the job even more perplexing.
The British took up where the Poles left off. The first break-throughs came from analysis of the uncoded message prefixes that told the recipient the key settings for the Enigma machine. Analysing the electrical wiring in rotors and plug boards, mathematicians and ‘probability specialists’ soon ‘reduced the odds against us by a factor of 200 trillion’.18 However there was still about a million to one against the men trying to conquer the Enigma messages. Wheel order and ring settings, the two most vital secrets, were sometimes guessed at by ‘sheet stacking’, a technique the Poles had pioneered. Holes – one per letter – punched in large sheets of paper allowed light through to reveal the pattern of the day’s key.
The repetitive phrases used in much of the traffic also helped: especially the formal way in which people and organizations were addressed in full. Often the codebreakers were waiting for the Germans to describe something that had already happened, such as a bombing attack or a weather report. Sometimes the same message was sent in a low-grade code – already cracked – and in Enigma too. In that case the two messages could be compared.
From time to time a rotor or two was retrieved from the pocket of a rescued U-boat crewman. By summer 1941 much-improved bombes came into use at Bletchley Park. These were 10-feet-high electro-mechanical machines like calculators. Using rotors like those inside the Enigma machines, the bombe searched rapidly through, not every possibility but the limited number chosen by the operator. The clickety-clickety sound came to a sudden stop when the bombe found a letter substitute for which the operator was looking. This would provide a setting that could be tried on a British encoding machine that had been adapted to perform like an Enigma. The printer started and produced a long strip of text. If it was ‘a good stop’ it might mean a batch of messages could be broken.19
If it was a navy message it would then go to the Royal Navy section in one of the huts, where a dozen or more men juggled with the German military jargon to make an intelligible message in English. This done, English and German versions might be sent by teleprinter to the Submarine Tracking Room in the OIC (Operational Intelligence Centre) in the Admiralty’s Citadel near Trafalgar Square in London.
The ‘key’, the settings of the rotors and the plugs, changed daily, or sometimes every two days. It was the most difficult problem, but even after the settings were discovered, the deciphered messages still had to be translated and made intelligible. Many messages were long, and included codenames and complicated service references that would be baffling to any civilian. There were technical aspects of stores, requisitions, meteorology, aviation or maritime matters. There were newly minted technical words and acronyms. A great deal of the intercepted material was banal and of little or no possible use as intelligence. And of course everything was in German. For some messages a result in two or three days was too late, and this was usually the case with the constantly moving war at sea.
The military