Gordon Welchman Page 4
Arthur Scherbius, was an electrical engineer born in Frankfurt-am-Main, Germany, on 20 October 1878. He had invented a new system of machine cryptography at the same time as several others were making similar inventions. Their purpose was to mechanize the process of encryption so that when any letter of the alphabet was entered into a machine, it produced as randomly as possible another letter of the same alphabet. At the heart of the machine was one or more rotors or encryption wheels. Each wheel had a set of electrical contacts on each side, one for each letter of the alphabet. Inside the body of the wheel, each contact on one side was connected to a contact on the other side by a strand of wire. The other significant advance on previous machines of this type was that the wheels could rotate inside the machine. After having an earlier version of his machine rejected by the German Navy in 1918, Scherbius, with a patent secured,3 began to offer it for sale commercially. His hope was that organizations wanting to encrypt information, such as banks, might be interested in it. The machine was not a huge success until the German Navy renewed its interest in it and decided to begin using it around 1926. It is believed by some authors that Dilly Knox bought a commercial version of the machine in Vienna around 1925.4 In 1928, the Enigma machine went into German Army service and eventually it would also be put into service by the German Air Force. In 1926 a single Enigma cost 6,000 Reichsmarks or $144. Once the German military decided to adopt the Enigma machine, it gradually disappeared from public sale as production of the commercial version was discontinued.
The German military were interested in the Enigma machine because of its portability. They expected that their forces would be in constant motion in wartime which meant that good communication between German commanders and their troops could only be maintained by wireless communications. The machine came in a wooden box, weighed around twenty-five pounds, and could run off an internal battery which held its charge for some time. The machine was an ‘offline’ device in that it was only used to encrypt and decrypt messages; their transmission was a separate process. Encrypted messages would be sent by radio in Morse code.
Before an operator encrypted or decrypted a message he would have to set up a number of components on the machine. The security of the system was based on the instructions provided to the operators to do this. Each month, operators were issued with explicit instructions for setting up each component of the system for each day of the month. The daily setting was known as the key. The number of possible keys was very large and the number usually quoted is 158.9 million, million, million. Compared to the odds of winning the jackpot on the National Lottery in the UK of 14 million to 1, this is a number far too large even to contemplate. The problem facing the cryptanalysts at BP was not only the huge number of ways the machine could be set up by the operator, but also that the German Air Force and Army had a number of different communication networks, each with its own daily setting of the Enigma machine.
Some information (known as the preamble) would be put at the beginning of each message in plain German. This included call signs which identified the sending and receiving German radio station, a discriminant identifying which networks would be able to read the message, the time of origin, the number of letters in the message and whether it was a single or multi-part message. Finally it would include letters, referred to as the indicator, which defined the indicator setting and message setting (see Appendix 1). These settings added a final layer of security before encrypting a message.
The Germans’ operating procedures exploited the reciprocal nature of the machine. When machines on a communication network were set up in the same way using the daily key and the same message setting, then the sending operator would type the message in so-called plain-text and his machine would encrypt each letter. Other operators on the same network would receive the message, then type the encrypted letters on the keyboard, and the machine would output the original plain text.
After a very short and fairly unmemorable stay, in Welchman’s words, ‘I was turned out of the Cottage and sent to Elmers School, where I was to study call signs and discriminants.’ Due to a shortage of office space at BP, Elmers School, a private grammar school for boys located nearby, had been acquired for the Commercial and Diplomatic Sections. When Welchman arrived there, the place was empty but Knox sent Tony Kendrick5 along to work with him. Kendrick normally gave fundamental explanations to newcomers about Enigma, something that the very busy and secretive Knox would no doubt have been loath to do.
As Welchman said in The Hut Six Story:
Alex [Welchman got his name wrong] Kendrick, a civilian member of Dilly’s prewar staff, was sent along to get me started. He was fairhaired, walked with a stick as a result of a paralyzed leg, and was noted for the holes burnt in his trousers by a cigarette, or possibly by ash from a pipe. We occupied a fairly large room with bare walls and no view from the windows; its only furniture was a long table and a few wooden chairs. Nobody else was working in the School, so Kendrick and I felt a bit lonely.
When he arrived at BP, Welchman had known nothing about the operation of the radio networks of that time, never mind call signs, indicators and discriminants. He and Kendrick were given several collections of Enigma traffic to analyse and, ironically, his separation from the main message-breaking activity going on in The Cottage, may have been the start of the Hut 6 organization that was to follow. Colonel John Tiltman, who was in charge of Army operations at BP, had instructed his sergeant to assist the work going on in Elmers School and all bundles of messages intercepted at the Chatham intercept station, along with a report on the day’s traffic, were made available to them. Again, in Welchman’s words:
Kendrick started to work on the large collection of material from Chatham, and set me a good example by beginning to analyse its characteristics in a methodical manner. His approach was reminiscent of the period some five to ten years earlier when I had been doing research in algebraic geometry and had often been faced with the problem of thinking of something to think about. In those earlier days I had found the best approach to this problem was to force myself to start writing, and here was Kendrick dealing with the same problem in the same way.
As no one else at BP had studied any of this material thoroughly, it was inevitable that they would soon start to break new ground. Unfortunately, Kendrick was moved on to other work and Welchman was left on his own. GC&CS’s inactivity in this area was, to a large extent, a result of the ongoing controversy about the function of signals intercept (‘Y’ material) and the directly related task of cryptanalysis. The Air Ministry stated at the time that:
Cryptographers are not Intelligence Officers, but only exist to provide the material from which Intelligence is produced; and it is as well to keep the intelligence side as far divorced from the cryptographic side as possible.6
Edward Travis, GC&CS’s Deputy Director, had disagreed and expressed the view that: ‘It is quite obvious that cryptographers will always know more of interception than the interceptors can possibly know of cryptography.’7
John Tiltman was not the only senior BP person interested in Welchman’s work. The head of BP’s Air Section, Josh Cooper, also understood its potential value to BP. He had suggested that the Air Ministry should itself be involved in the interception of high-grade traffic. In the early stages of the war the Army ‘Y’ Service was intercepting a considerable amount of German Air Force Enigma traffic. He was promptly rebuked by Group Captain Blandy, then Deputy Director of Signals (Y): ‘My Y-services exist to produce Intelligence, not to provide stuff for people at Bletchley to fool around with.’8
Cooper gave Welchman BP’s sole collection of decrypted German Enigma messages, which helped him understand the task at hand. Cooper, according to former colleagues, was even more eccentric than Turing and Knox. Like them he was a King’s College man and had joined GC&CS as a Junior Assistant in October 1925 to specialize in Russian codes and ciphers. Stories abounded about his eccentric behaviour (many untrue) which apparently included ta
king his evening cup of tea by the lake at BP and then throwing his cup and saucer in.9
Three key figures in the history of BP, John Tiltman, Josh Cooper and Oliver Strachey (in his initial briefing in 1938), thus helped inform Welchman’s first breakthrough of major significance for the future of the organization he had only joined several months earlier. What he was starting to understand was that call signs, indicators and discriminants (part of the message preamble) could, along with the message itself, help to build up a picture of the German order of battle. What he didn’t know at this stage was where the intercepted messages were coming from.10
Through Tiltman’s good auspices, Welchman was able to make his first visit to the Army’s radio intercept station at Chatham in Kent at the end of September 1939. He quite quickly befriended Commander Ellingworth who was to prove to be ‘a tower of strength throughout the war’. It was Ellingworth who taught him how German radio nets operated and they were able to establish a system of co-operation between BP and the intercept stations. Ellingworth ensured that the preambles of messages were sent to BP by teleprinter to give codebreakers advance material to work on.
Commander M. J. N. Ellingworth was an ex-naval man and First World War veteran, known behind his back as ‘Truncy’, and apparently well-liked by all his staff. Heavy bombing raids on Chatham during 1940 and the fact that German forces were just across the Channel, prompted a move to Beaumanor Hall in Leceistershire. It had been requisitioned from a William Curzon-Herrick and was about sixty-seven miles from BP. The move was interrupted by a short stay at the RAF signals station at Chicksands Priory in Bedfordshire, just twenty miles from BP. The Chatham interceptor operators worked alongside their RAF and WAAF counterparts until Beaumanor was ready. Beaumanor then became the main Army listening station and supplier of raw intelligence to BP.
Chatham, Beaumanor and Chicksands were part of Britain’s ‘Y’ Service. In 1924 GC&CS had established an interdepartmental ‘Y’ committee dealing with wireless interception, though each service retained control of the personnel and installations for their own interception stations. In the years leading up to the Munich agreement, British intelligence organizations were keen to gain information about the German government’s intentions. This was achieved both by conventional intelligence-gathering methods and by monitoring German diplomatic and military wireless traffic. At the start of the Second World War there were three ‘Y’ stations in the UK, at Scarborough for the Admiralty, at Chatham for the Army and at Cheadle for the RAF. As the war progressed, the main UK intercept stations were based at Scarborough and Flowerdown (Winchester) for the Navy; Chatham and then at Beaumanor Priory and Forest Hill for the Army; and Cheadle and Chicksands for the RAF. There were also ‘Y’ stations in the Middle East, the Far East and behind the battle lines in North Africa, Sicily, Italy and mainland Europe as Allied forces advanced after D-Day. BP trained most of the decryption staff for the overseas centres and the intercepted encrypted signals from the ‘Y’ stations were all sent to BP. The allocation of frequencies to be monitored for German Army and Air Force signals was determined at BP by an intercept control team in Hut 6 and for German Naval signals in Hut 4.
Operators at intercept stations would tune in on a specific communications network (‘net’) and write the signals down character by character, at the rate of at least twenty words per minute, as they were being transmitted on short-wave frequencies in Morse code by German operators. They would then be sent by pneumatic tube to a control hut for logging and collating and then on to BP by despatch rider and later in the war by teleprinter, on a daily basis. The teleprinters operated at only 10 characters per second and were prone to errors so despatch riders augmented the teleprinter links, hurtling down dark roads to BP on their BSA M20 side-valve machines. No attempt was made to decrypt the signals at the intercept stations but considerable information could be gleaned from the frequencies, call signs and times of transmission. This work was handled by log-readers who were part of MI 8 (Military Intelligence, Section 8), the cover designation for the Radio Security Service (RSS), a department of the Army’s Directorate of Military Intelligence and part of the War Office.11
Beaumanor was the one of the largest intercept stations with 196 radio sets in use at the end of the war and some 1,300 operators. Beaumanor was known as War Office Y Group or WOYG – the staff referred to it as ‘Woygland’ and to themselves as ‘Woygites’. Other terms used were ‘Beaumaniacs’, with the more dashing males as ‘Manor Beaus’. By all accounts, the tone of Beaumanor was set by its likeable commander whose personality is best exemplified by his foreword to the first Beaumanor staff magazine in October 1941 and the cartoon of him which accompanied it.12
Without the Y Service and the selfless dedication of around 10,000 people who staffed it, the work of BP would not have been possible.
For the first six months of the war, the German Army and Air Force continued to use the Enigma indicator system that had been introduced towards the end of 1938. Before the war the Polish Cipher Bureau had developed an elegant method for recovering part of each Enigma key by exploiting a weakness in that indicator system. Although the Poles had disclosed the details of their methods to the British late in July 1939, no information about the Polish achievements was passed on to Welchman when he joined Knox’s team in September 1939. However, he was told about the operation of the Enigma machine and of the indicator system then still being used by the Germans. After a few weeks, Welchman undertook a private investigation on this indicator system. Remarkably, he re-invented one of the key Polish ideas for reducing the possibilities for part of the daily key.13
Welchman described in The Hut Six Story, the less than favourable reaction from Knox when he told him of his ‘discovery’.14 While Knox did not give him any details about the state of that development, he did say that it was being led by John Jeffreys. As they had known each other before the war, Welchman was confident that the work was in good hands. More importantly, it left him free to start thinking about the wider implications of decrypting Enigma traffic on a large scale. Knox’s approach was typical of the First World War cryptanalysts, that is that each decryption of a message was in itself a cryptological success. Welchman recognized that real success would lie in its exploitation leading to military gains.
The following description of Knox that Welchman gave in his book has often been quoted by other authors:
Dilly was neither an organization man nor a technical man. He was, essentially, an idea-struck man. He was not interested, as I was, in the administration and automatic routine needed to handle the enormous volume of Enigma traffic generated by the German army and air force.
Interestingly, Welchman saw very little of Knox at BP, even when he was part of his team. He certainly knew nothing of his pre-BP life. A book by Knox’s niece, Penelope Fitzgerald, The Knox Brothers, published in 1977, helped to fill in the gaps.
The story of the early battle against the Enigma machine and the Polish contribution is a fascinating one and Welchman would spend much of the last few years of his life trying to unravel it and put the record straight.15 At the end of the First World War, the newly reborn Polish state had taken control of former German territory in parts of Silesia, Pomerania and around Poznań (Posen in German). This had been part of Poland before the eighteenth-century partitions of the country by its powerful neighbours. The return of this territory to Polish control was a cause of considerable anger in Germany and an atmosphere of enmity and continual tension persisted, fuelled by the long history of numerous and bloody wars between Poland and Germany. The reborn Polish state felt seriously threatened by its neighbour to the west.
Some Polish historians claim that in December 1927 Polish military intelligence dismantled and photographed a commercial version of the Enigma machine which was discovered accidentally by customs officers at Warsaw’s Okecie Airport in a misaddressed German package.16 On 15 July 1928 Polish radio-monitoring stations at Starogard in Pomerania, Gdańsk, Poznań an
d Kraków-Krzeslawice intercepted the first German messages which were in a machine-generated cipher. While Polish intelligence in Germany was trying to find out about German cipher changes, a course in cryptology was started at the University of Poznań for twenty of its most advanced mathematics students who could also speak German. The course was set up at the initiative of the radio intelligence department and some of its specialist officers would lecture to the students, the most gifted of whom would be asked to volunteer to continue their studies within military intelligence. During 1931 a cipher bureau was created which was an amalgamation of the radio intelligence and cryptography sections. The new bureau was headed by Major (later Lieutenant Colonel) Guido Langer. As the new department was being organized, the course at Poznań was coming to an end and included three students who had frequently managed to solve the German ciphers that had been set for them. It was decided to set up a small section of the department for them in Poznań. Their names were Marian Rejewski, Henryk Zygalski and Jerzy Różycki.
The German military had introduced a new version of an Enigma machine with a plugboard at the beginning of June 1930 and the following year the Poles purchased a commercial model on the open market in Germany. At the same time, Captain Gustave Bertrand, who worked for French intelligence, initiated Polish-French co-operation in cryptographic matters. At the beginning of September 1931 Rejewski, Zygalski and Różycki were assigned to Section BS 4 (German intelligence) of the Cipher Office of Department II of the General Staff in Warsaw. This section was under the command of Captain (later Major) Maksymilian Ciçżki.
In June 1931 a 43-year-old employee of the German Defence Ministry Cipher Office, walked into the French Embassy in Berlin and into history. Hans Thilo Schmidt had access to, and was prepared to sell, manuals and other information relating to the Enigma machine being used by the German Army. Bertrand and a colleague, who went by the fictitious name of Rodolphe Lemoine, met with Schmidt on 1 November 1931 at the Grand Hotel in Verviers, a small Belgian town on the German border, Schmidt produced some of the material that he had promised. Bertrand was delighted, and while Lemoine arranged for Schmidt’s payment, he photographed the material. He then rushed back to Paris and offered the material to French cryptanalysts who promptly said that it was of no use to them. An approach to the British experts for a second opinion yielded the same answer. The material only dealt with the encryption of a message, not how to decrypt one. However, when Bertrand got permission to approach his Polish counterpart, he boarded a train to Warsaw. The Polish Cipher Bureau, perhaps not surprisingly, received Bertrand and the material from Schmidt, with open arms.17