Gordon Welchman Read online

  By the end of 1940, Welchman’s plan for the subdivision into huts was gathering pace and a staff list produced on 2 December 1940 shows staff numbers as: Hut 6 – 93, Hut 3 – 60, Hut 8 – 37, Hut 4 – 40. The number of staff listed in Knox’s Research Section in The Cottage was 8!10 By July 1941 Welchman needed more staff and he wrote to Travis on the 4th, stating his requirements:

  As you know, the work of Hut 6 is getting more and more difficult. We need more staff and more space. The male staff will probably not need many additions, but I should like permission to engage up to 6 seniors and 6 juniors in addition to our present staff. The girls are divided into 4 sections which are likely to require an average staff of 25 each. In fact we shall need at least 4 Temporary Junior Assistant Principals and 100 Temporary Assistants or Grade II clerks. At present we have 3 girls capable of being Temporary Junior Assistant Principals, and 64 others of whom 12 are probably not up to the work. I should like permission to engage up to 48 more girls either as Temporary Assistants or as grade II clerks and up to three more Temporary Junior Assistant Principals.

  Of course more space will be needed. We are very overcrowded already. I should think that an addition equivalent to about one quarter, or possibly one third of our present space would be sufficient.11

  Failure by senior BP management to deal quickly with urgent requests such as this led Welchman to take more dramatic action. In late October 1941, Welchman, Turing, Alexander and Milner-Barry all signed a letter to the Prime Minister, Winston Churchill. Milner-Barry then proceeded to London where he took a taxi to No. 10 Downing Street and handed in the letter to one of Churchill’s staff. The consequence was a marked improvement in the situation because on receipt of the letter Churchill had put an ‘Action This Day’ stamp on it with a handwritten note to his chief military assistant, General Ismay, saying: ‘Make sure they have all they want on extreme priority and report to me that this has been done.’ The letter may well have led indirectly to Denniston’s replacement as Operational Director of GC&CS by his deputy, Edward Travis in February 1942. This followed a review of the administration of signals intelligence by Sir Stewart Menzies, Chief of SIS.12

  Turing’s work at the start of the war on a machine solution had quickly reached the build stage and the British Tabulating Machine Company at Letchworth had been contracted to carry out the work. Their Research Director, Harold ‘Doc’ Keen, needed to be briefed on BP’s requirements and Peter Twinn was given the task, as he explained in a letter to Welchman in 1984:

  Keen’s first introduction to the problem was from me and took place in the White Hart at Buckingham. Travis asked me to put the matter to Keen first because he thought, erroneously in my opinion, that Alan Turing’s manner and occasional incoherence might put Keen off and create the wrong impression.

  By early 1940, the manufacture of the prototype bombe, a complex electro-mechanical machine, was well under way. The first prototype, named Victory, was installed in Hut 1 on 18 March 1940 but did not prove particularly effective. It was only when a second and improved machine was installed in Hut 11 in August that real results were achieved. This machine, named Agnes, incorporated a brilliant design modification invented by Welchman and called the diagonal board (see Chapter 4). Victory was then moved to Wavendon as a training machine as it was unreliable and six more machines were on order by November. The first of an even more sophisticated model, the Jumbo bombe, arrived in Hut 11 in March 1941.

  With the arrival of the machines, the whole Hut 6 operation changed and fundamental to its success was the concept of the crib. The idea was to guess a plain text phrase that was embedded somewhere in an intercepted encrypted message. If the plain text phrase could be matched to a run of encrypted characters within the message, it provided a crib which could be used with a bombe to help deduce some of the key settings such as the wheel configuration. A crib was not a new concept and Knox and his team had used it quite effectively several years earlier. However, they were using hand methods and it was now proposed to use a machine to speed up the process. A new structure to do so was gradually put into place during 1941.

  Up to the end of 1940, Welchman had personally directed all of the Hut 6 activity. But a more complex infrastructure was now required to deal with the growing Enigma traffic. In early 1941 the Governing Body was formed and included the heads of Control (Colman) the Crib Room (Milner-Barry) and Machine Room (Babbage) with Welchman in the chair. Fletcher, as Head of the Registration Room, Decoding Room and Netz Room was added to the group in early 1942. He handled most of the administration of Hut 6 throughout the war and had the crucial responsibility of liaising with BTM at Letchworth, which had been contracted to manufacture the bombes.

  The group had weekly meetings and considered weekly reports from section heads (compiled into a Hut 6 weekly report), but had little operational decision-making authority. They concentrated on laying down policy and had grown to nine members by 1944. The Governing Body continued to meet until March 1945, when it was dissolved.

  Hut 6 Administration was created to enable individual members to carry out their jobs with the minimum amount of distraction from the administration of GC&CS. The office had five staff, including three secretaries, and their duties consisted of the supply of furniture and office equipment, accommodation, statistics and staffing.

  The position of duty officer was introduced in 1941 and in theory he was Hut 6’s chief executive officer. It was filled by a senior member of the hut who carried out other duties until August 1942 when a permanent assistant duty officer was appointed on the day shift. The roles of the day duty officer and the head of the Decoding Room were combined in April 1944 and eventually it was almost solely responsible for liaison with Hut 3 on matters of decoding priorities.

  Welchman had recognized by early 1940 that the problem of interception required a special section and the Control Room had been created under John Colman. The section had a wide-ranging role which included recognizing radio set requirements; pressing for suitable interception and communications facilities; organizing and maintaining relations between Hut 6 and intercept stations; maintaining a continuous service to direct cover and assist interception. It consisted of a body of watch officers, each with an assistant, who maintained a round-the-clock rota keeping control by direct line of the main home intercept stations. Policy on cover was set by the Governing Body of Hut 6. Priorities between different requests were determined by the control officer (order of priority usually cryptography, intelligence, traffic analysis), with Colman, as head of Control, the final arbitrator.

  The scope of minute-to-minute management of interception from the Hut 6 Control Room developed considerably and staff numbers grew to around twenty-five. By late 1941, Hut 6 maintained liaison officers at two main intercept sites. In the summer of 1943, a small special section, the Overseas Party, was created to tackle problems associated with overseas interception. Control had a special telephone exchange set up with direct lines to the main intercept station. In late 1943, a larger switchboard was installed and was in use twenty-four hours a day, seven days a week. A priority teleprinter system was put into place with a performance target of twenty minutes from time of intercept to time of teleprinting. While despatch riders were rarely necessary to send important traffic, Chicksands, only thirty minutes away, sent riders to BP hourly. Hut 6 Control communicated with intercept stations at Beaumanor, Bishop’s Waltham, Chicksands, Forest Moor, Harpenden, Denmark Hill, Whitchurch (Shropshire), ‘Santa Fe’ (Bexley, Kent, US Army), Wick and Montrose.

  In his memoir, John Herivel observed that:

  Nobody was recruited to the Machine Room in Hut 6 who was not a mathematician; mathematicians, moreover, who were of two sorts only: those who had been students of Welchman’s at Sidney Sussex College, and those who had been friends of his at Cambridge; in other words, he had packed the house!

  With the arrival of effective bombes, the Machine Room was set up to produce instructions for the women (usually members o
f the Women’s Royal Naval Service or WRNS) who set up the bombes, operated them and then tested the results that were produced. The instructions were in a diagrammatic format and called menus, apparently because they were the diet that was fed to the bombes! A separate Testing Room was set up and as a direct descendant of the old Netz Room, carried on with the name until 1943.

  The Crib Room was formed on 1 October 1940 under Milner-Barry. The art of cribbery, that is finding usable cribs, was steadily developed and the members of the section, the cribsters, became gradually surer in their touch. They gained experience which would prove to be the foundation stone of the later success of Hut 6. Sometimes the cribsters would be helped by German Air Force and Army Enigma operators’ use of stereotyped language. An Army unit stationed in North Africa but seeing little action used the same phrase in its encrypted daily report to headquarters every day for a month ‘KEINE BESONDEREN EREIGNISSE’ (this loosely translates as ‘No special occurrences’). In 1940, when the German Air Force was operating in France, each day its airfields would receive instructions for their targets. Each message would begin with the phrase ‘BESONDERE ANORDNUNGEN FUER DIE …’ (‘Special Orders for the …’) followed by each airfield’s designation.

  The Research Machine Room began informally in autumn 1940 as a separate body of people entrusted with the specific task of trying to break ‘odd keys’ while Red and another Luftwaffe key known as Brown, used to send information about navigational beams, were broken daily by routine shifts of the Machine Room and Crib Room. The Research Machine Room had variable membership of at most two or three people at a time who were seconded from the routine shifts for a week or fortnight. They worked permanent days rather than shifts. Once progress was made with a key, and if its contents were sufficiently important, it would be quickly transferred to the routine shifts. An example was Light Blue, used by the Luftwaffe in Africa. It was broken using cillies in March 1941 and at once became a current key for the Machine Room. Another example was Orange, which was used by the SS and broken initially by the Research Machine Room on 10 December 1940.

  A Research Crib Room was created in April 1942 and became permanent in September 1942 when Derek Taunt was assigned to it.

  The Decoding Room decrypted all of the traffic using British Typex encryption machines, which had been converted to emulate Enigma machines. Once a key was broken, Typex machines could be set up using that key and all traffic that had been intercepted on it could then be decrypted. The section operated much like a typing pool with women typing in thousands of encrypted 200–250 character messages. The machine would produce the German text on thin strips of paper tape which were then stuck on to sheets of paper and passed to Hut 3 for translation and analysis. They also decrypted broken naval traffic for Hut 8 until it established its own Decoding Room.

  The Typex machine was the British equivalent of Enigma.13 In 1926, the British Government had set up an Inter-Departmental Cipher Committee to investigate the possibility of replacing the book systems then used by the armed forces, the Foreign Office, the Colonial Office and the India Office by cipher machines. O. G. W. Lywood, a signals officer with the RAF, believed that it would be possible to develop an improved version of the commercial Enigma machine. He proposed to incorporate parts from Creed teleprinters, in order to produce printed text. The committee refused to proceed so the RAF went it alone with Lywood’s ideas in 1934. Unlike the military Enigma, which was a standard machine with virtually no variations between machines apart from the additional three wheels used by the German Navy and two settable reflectors, Typex had a minimum of 120 different wheels or inserts in service. As five wheel inserts could in effect be selected from a set of 28, the Mk. VI version operators could arrange their wheels inside the machine 7,687,680 different ways, compared to 60 or 336 with Enigma. Typex had 20 or more completely different sets with no common wiring between them. The Germans would have had to find the wirings for between 120 and 250 wheels – a huge task. Ironically, the Typex infringed several patents on Enigma held by the German company Chiffriermaschinen Aktiengesellschaft. Of course no royalties were ever paid!

  The old Registration section became the Registration Room, recording all Air and Army traffic in a form suitable for the Crib and Decoding Rooms. The registration of traffic remained its main function until September 1942. From summer 1941 until autumn 1943, it also documented operational keys. A separate group of staff within the section formed the Research Registration Room and dealt with research keys, that is those that were problematic. All messages were registered on blist forms and numbered. Details of incoming discriminants were recorded on a chart called the ‘Hanky-Panky’ or just ‘Hankey’, which also directed messages bearing identified discriminants to the correct blist. The ‘Hankey’ was named after its designer John ‘Hank’ Hancock.

  Before autumn 1942, the training of staff in Hut 6 came from doing the work itself. The Decoding Room was the first to find a more specific requirement and a school was set up on 14 September. A Registration Room school followed in late September, which laid the foundations for systematic training for all new staff. The need for training was driven by the large influx of staff and the scope of the work becoming more difficult. Therefore, a formal training syllabus for Hut 6 staff was put in place.14

  About 200 distinct Enigma keys were identified, named and broken in the history of Hut 6 and details can be found in Appendix 4. With a very large number of live keys at any one time a system was needed to manage it effectively. A parentage system was created to make best use of the available cryptographic resources and the special talents of each person. It also ensured that each key got its fair share of attention, and incidentally proved useful in countering boredom and staleness. One or sometimes two people (the parent) took responsibility for a key or set of keys and had specialized knowledge of related cribs and cillies on a key. Changes in allocations of keys to a parent were made at intervals.

  Space was an ongoing problem until Hut 6 moved to the spacious Block D in February 1943. The research branches of the Machine, Crib and Registration Rooms had to be housed in the Mansion until Block D was completed. All organizations, no matter how harmonious, have their problems and Hut 6 was no exception. According to the authors of the ‘Official History of Hut 6’,15 there were artificial and arbitrary divisions of responsibility. Every day questions arose which neither party could settle on its own. As cribs became more important, the balance of power shifted from the Machine Room to the Crib Room and members of the former ‘felt degraded from their former proud position to little more than menumakers and testers of stories’. Meetings were held in the autumn of 1941 but the situation persisted into 1942. A solution was deferred until the move to Block D but a fusion process ensued in which Machine Room members spent time in the Crib Room and vice-versa. In February 1943, the Machine and Crib Rooms merged as the Watch and the Netz Room became the Machine Room. A similar merger took place between their Research Section counterparts with a similar information exchange. Milner-Barry took over the Watch, and Babbage the Research activities. The ‘Official History of Hut 6’ summarized the merger as follows:

  The final moral of the Machine Room/Crib Room story may be stated thus: While in a complex cryptographic organization like Hut 6, a considerable degree of specialization is unavoidable as between interception, traffic analysis and cryptography, it is undesirable that there should be any watertight divisions in the initial processes of breaking. Any specialization that is necessary here should arise from divisions of the material to be broken – e.g. Watch/Research and later Air/Army – not from different lines of approach to the same material.

  Over a few years, Welchman had managed to put in place a remarkably efficient organization. It would provide a model for its sister organization, Hut 8. Hut 6 was run in a very loose and informal style with only an indispensable minimum of formal routine meetings. Heads of departments had a free hand with their advice and it was almost always accepted without question. B
y the end of the war, Hut 6 had around 550 staff.

  As the architect of the hut-based process at BP, Welchman knew that some problems would eventually emerge. While he had established an excellent working environment in Hut 6, the same could not be said for Hut 3, led by Malcolm Saunders. This was of concern to him as he had envisaged Hut 6 and Hut 3 working as an inter-service organization under Foreign Office administration.

  R. V. Jones was a frequent visitor to BP in his capacity as Assistant Director of Intelligence (Science) in the Air Ministry. Travis had given him the impression that he regarded Hut 3 as ‘a lot of exploiters of Hut 6’s efforts’ and asked that he come to BP in November 1940 to give Hut 6 a talk on his work. This was to be exclusively for Hut 6 because on a previous visit during the summer he had given a talk to Hut 3 and they had not invited Hut 6. After his morning talk to Hut 6, Jones encountered Saunders and others from Hut 3 at lunch. They were furious that they hadn’t been invited. As Jones recalled in a letter to Welchman:

  The only way I could placate them was to stay and repeat the talk to Hut 3 during the afternoon, which I knew would involve driving back into London after the nightly attack had started. As I described in my book, I crashed into an unlit lorry abandoned outside St. Albans Hospital, and Charles Frank and I went through the windscreen. I feel that the rivalry between the huts still owes me for the car, for this was the first year in which I had not insured it comprehensively and it was a write-off. I am sure that you with your point about reading German, did not feel that Hut 6 were unnecessary middlemen, but my impression may have been unduly biased by Travis, and relations between him and Saunders may not have been too good.16