Seventy Years Ago This Month at Bletchley Park

August 1939

The Germans Prepare to Invade Poland.

In March 1939 Hitler told his Army staff that Poland was to be the next target, and set 1st September as the date for the attack. Hitler resented the loss of direct access to East Prussia because of the creation, under the Versailles treaty, of the Polish Corridor to the Baltic. The integrity of Poland was guaranteed by France by a treaty dating back to 1921, and was reinforced by Britain in March 1939. But the abrogation in April 1939 of the German-Polish non-aggression treaty, and the German discussions with the Russians for a pact made it inevitable that Poland would soon be at war. On 23rd August an Anglo-Polish Mutual Assistance Pact was signed and given publicity. Poland had mobilised 1 million men by the time the Germans attacked with five armies totalling 1½ million men on 1st September 1939. With the help from their code breaking over the preceding years, it is said that the Poles were able to identify 80-90 % of all the German units assembled on their frontiers.

The Polish Code-breaking Achievement.

The Poles had been reading some of the German Enigma keys since 1933. But the Germans increased the rotor wheel combinations by a factor of ten in December 1938, providing a choice of five, rather than the previous three, wheels from which to fill the three slots in the Enigma machine. This was adopted by most of the German services and thereafter the flow of Polish decrypts reduced to some 10% of the 1938 rate, and then almost dried up after July 1939 when the German Security service finally adopted the new procedures. On 24/25 July 1939 at the invitation of the Poles, the British and French cryptographers went to meet them at their site near Pyry in the Kabackie Woods, a few miles south of Warsaw. The Poles gave them all details of the wiring of the machine and the five code-wheels. They explained their methods, including the use of the Zygalski sheets, and showed them their machine, the Polish Bomba, that they employed for finding the key settings using the encoded, but twice repeated, wheel start positions sent by the Germans at the beginning of their messages. The team from GC&CS included the ageing but still brilliant code-breaker Dillwyn Knox, and Alastair Denniston, the first Director of Bletchley Park. The third Briton present, “Professor Sandwich”, was the head of the Admiralty’s Interception Service, DNI 9, though there are some improbable suggestions that it was Col. Stewart Menzies, deputy head, soon to become head of MI6, “C” himself. This meeting gave the British the confidence of knowing that breaking Enigma could be achieved, and shortened the time before Bletchley Park could first read Enigma messages by giving them the internal wiring of the machine and its rotor-wheels.

Poles hand over their copy of the Enigma machine.

At the meeting on 24th July the Poles had offered France and the UK a Polish copy of the German Enigma machine. The Allies knew the wiring and other details of the Enigma commercial machine, but, unlike the Poles, did not know the wiring of the current German military version of the Enigma machine. After the meeting the Poles sent two of their Enigma machines to their primary French contact, Gustav Bertrand now promoted to Major. They had been co-operating with Bertrand for some eight years, since he was the controller of the German spy, Hans-Thilo Schmidt, who had provided the French with the material that they had offered to the Poles, so helping them to work out the wiring of the Enigma wheels in 1932/3. (Neither the French nor the British Intelligence had taken advantage of the material). It seems it was Bertrand himself who brought one of these machines to London on the 16th August 1939. According to Bertrand he was met on Victoria station by Col. Stewart Menzies, “C” being “en smoking, avec rosette de la Légion d’Honneur à la boutonnière. Accueil triumphal!” He may also have brought with him some of the important details of the wiring of rotor wheels, etc, received from the Poles, though it seems more likely that Dilly Knox would have brought this vital information back with him. It is said that Dilly phoned the crucial information on the wiring, that from the keyboard to the first contacts of the code-wheels, through to his assistant, the young Peter Twinn; Peter had worked out the wheel wiring from this information by the time Dilly reached home.
Occupation of Bletchley Park by GC&CS.

In May 1938 Bletchley Park had been acquired to provide a wartime home for the Government Code & Cypher School. This body nominally reported to the Foreign Office, but was a tri-service operation. It acted under instructions from the Joint Intelligence Committee of the Chiefs of Staff but with direct relations with the three Service Ministries as well as the Foreign Office. By 1939 it was clear that all the German armed forces, as well as other bodies such as the railways, were using closely related cyphers based on the Enigma machine, so the case for the UK retaining a single body to handle the attack on the enemy cyphers was strong. It was decided in the spring of 1938 that it would be wise to move GC&CS from London to Bletchley Park on the outbreak of war. GC&CS had temporarily occupied the Park in the autumn of 1938 at the time of the Munich crisis, and preparation for the wartime move had been going on for some time. On 1st August 1939 some of the Service teams moved down from London, followed by the Commercial & Diplomatic teams on 15th August, just two weeks before the out-break of war. Some 140 GC&CS staff started to work in BP (Bletchley Park), of whom perhaps up to 30 were code-breakers, the rest being Intelligence and support staff. The numbers grew rapidly as staff engaged for service in the event of a war joined up. A further 40 of the GC&CS staff moved to the requisitioned Mansfield College, Oxford, for work preparing the British codes and cyphers.

The Initial Staff.

They were led, as BP Director, by Alastair Denniston who had been in Room 40 in the Admiralty during the First World War and who had stayed on in GC&CS between the wars. Amongst his long serving colleagues who came to BP in August were Dilly Knox, Oliver Strachey, Josh Cooper, Nigel de Grey, and Col. John Tiltman. Frank Birch was another of the First World War team who now rejoined from Cambridge. The Deputy Director and head of the Service cryptographers, Commander Edward Travis had been in Naval Intelligence for some years in World War I before joining the newly created GC&CS in 1919. Tony Kendrick, one-time head boy at Eton, had joined a few years before the Second World War. Peter Twinn, from Oxford University was another of the younger generation who had joined GC&CS in February 1939, as what is said to have been their first mathematician. And then there were the new recruits; Alastair Denniston had been recruiting in Oxford and Cambridge since the Munich crisis in the autumn of 1938, for suitable volunteers who would join GC&CS on the outbreak of war. Amongst these were two further mathematicians, Alan Turing, from Kings College, Cambridge, and Gordon Welchman from Sidney Sussex College, Cambridge who came together to BP by car on 4th September; both were to play vital parts in the great story that was about to unfold.

The Building of the first Huts

probably dates from the time of the Munich crisis. Hut 4, alongside the Mansion, immediately housed part of Naval Section, and at least one of the two huts on what is now the Mansion car-park was soon occupied. Probably the first six huts were up, if without internal partitions, by September 1939, certainly by the end of the year.

The Enigma Machine.

The basic Enigma machine was conceived and patented by a Dutchman, Hugh Koch, in 1919, though the concept of using a cypher machine employing code wheels had been invented 100 years and more before. But it was a German engineer, Arthur Scherbius, who developed and exploited the Koch patent, putting an ‘Enigma machine’ on the market in 1923. It was bought by some banks and other large companies to secure their communications, and so the design became widely known. The British Government bought a machine in about 1926 after a demonstration at the Foreign Office, though the view in the War Office was that the ‘large’ machine was too bulky and heavy for use in the field. An improved ‘small index’ version of it was adopted by the German Navy in 1926, by their Army in 1928, and then by the Luftwaffe in 1935. At this point it was withdrawn from the open market, and the Germans instituted a series of modifications with a view to increasing its security.

GC&CS progress on Enigma.

It was in 1927 that GC&CS first studied the Enigma machine, primarily to see if it was suitable for British use. Hugh Foss, who had been recruited in 1924, no doubt primarily because he was that rarity, a Briton who spoke excellent Japanese as he had grown up in Tokyo where his father was a missionary. He worked on the ‘small’ Commercial machine, which lacked a stecker, or plug-board, but was similar in other respects to the German military model, though its internal wiring differed from the wiring of the military model in one elementary yet very important respect. He wrote a paper showing how, if the internal wiring was known, it might be broken, using cribs. As the threat of war mounted in 1936, some of the best BP codebreakers turned their attention to breaking the steckered Enigma, failing to do so for lack of knowledge of the internal wiring between the keyboard and the input to the code-wheels. Then on 24th April 1937 Dilly Knox broke an unsteckered Enigma machine, essentially the Commercial machine used by the Italian (and Spanish) navy during the Spanish civil war. He used what was very like the method proposed by Foss, but employing a set of what he called ‘rods’, which essentially provided a slide rule type representation of the wiring between the various Enigma wheels. After that break, William Bodsworth took over handling some of the, now readable, current traffic from the Mediterranean, while Dilly concentrated on breaking the German military machine.

Dilly knew that the Germans had changed this wiring from the Commercial model where the wiring followed the keyboard order, what Dilly always called the ‘QWERTZU…’. (Note that the German keyboard, to this day, has Z in the place occupied by Y on the British typewriter keyboard). He naturally assumed that the Germans had used, at random, one of the vast number of different wirings that were possible. When the Poles told him at the Pyry Forest meeting that all the Germans had done was to substitute ‘QWERTZU…’ with ‘ABCDEFG…’. Dilly was furious and lost his temper to the embarrassment of his two British colleagues, as Denniston described years later. Luckily, by the next morning Dilly had recovered his normal charming self (that he could lay on if he wanted to do so, which was certainly not always!) and the Poles were intrigued by his use of the ‘rods’. However, he seems to have remained convinced that the Poles were not telling the full story about their first break, and indeed it soon emerged that they had received some help from a German spy, Hans-Thilo Schmidt. However, this does not destroy the remarkable achievement of the Poles in making that break, way back in 1932, and in defeating the various complications subsequently introduced by the Germans, until they were over-whelmed by the increase in the number of possible code-wheels introduced in December 1938. There seems little doubt that the very few British codebreakers who were in the know about the Polish achievement, gained considerable confidence from the knowledge that it had been done, though now that war was about to break out they assumed the Germans would change the wiring, so in a sense making the Polish work on recovering this wiring even more important. It is said that, ignoring the security risk, Dilly phoned the news about the Enigma wiring to his young assistant, Peter Twinn back home, and that Peter was now able to work out the wiring of each code-wheel, which he had achieved by the time Dilly reached London.

Naval Enigma.

The top priority for GC&CS as war broke out, was to break the German naval codes. The importance of the defence of the sea needed no reminders for an island race like the UK, where fighting on land would be postponed, if at all possible, to allow the Armed Forces to build up to their war-time strength. Moreover, it was the Admiralty who had had the greatest successes due to their code-breaking in the First World War, still very fresh in memories. But unfortunately, perhaps for the same reason, the Navy was the German Service with most security consciousness. Unlike the land forces, which may be able to employ cables, naval ships at sea have no choice but to send their communications by radio. They had been the first Service to adopt Enigma, [and virtually throughout the war they proved to be most adept at maintaining good security]. Partly for this reason, and partly because naval matters were not so important for the Poles as land operations, the Poles had made little progress in breaking German Naval Enigma, though they did pass over to the British a few Naval decrypts dating from May 1937. According to Denniston in a post war memo, it was in May 1939 that the German Navy had adopted an elaborate way of encrypting their Enigma ‘indicator’ settings using tables, unlike the other Services who continued to send them encrypted on Enigma but otherwise without further disguise. And when war broke out the German navy, who, until then, used just the same, now five, code-wheels as the other Services, introduced three extra code-wheels, making eight in total for use in naval communications. So Bletchley Park in the summer of 1939 is faced with a far more difficult problem to break the German Naval Enigma than they do to break the Enigma communications of the other German Services.

The Bletchley Park Trust welcomes the preparation of these notes, but the authors are responsible for
the statements and the views expressed.