The Bletchley Park Codebreakers Read online

  Hut 6’s successes ultimately depended upon the comprehensive and accurate coverage of the Heer and Luftwaffe radio nets by the Y service, but regrettably the story of interception in 1939 and for much of 1940 ‘is not a pretty one’. In September 1939, there were twentyfive sets, manned by civilians, at the War Office intercept station at Chatham, which initially was the main intercept station for Enigma. In November 1939 and February 1940 the War Office warned Group Captain L. F. Blandy, the head of AI1e (which was responsible for interception in the Royal Air Force (RAF)), that if a major offensive began on the Western Front, Chatham would drop Luftwaffe Enigma traffic, which the RAF would then have to intercept. In April 1940, Denniston informed Blandy and Colonel D. A. Butler, of MI5, that ‘it is now of the highest importance that Y stations should concentrate as far as possible on German Enigma traffic (Air and Army) … I should be very grateful if you would issue orders to that effect’.

  The Air Ministry agreed to intercept the Enigma traffic on 2 March 1940, but Hut 6 received nothing from it for many months. The root of the problem was that AI1e belonged neither to the RAF’s Director of Signals, who had some receivers and men but needed them for communications purposes, nor to the Director of Intelligence, who had no responsibility for intercept facilities and seems not to have realized that there was an intercept problem, since he was already receiving a great deal of intelligence – intercepted by the army. In April 1940, Denniston urged Menzies ‘most urgently to call a meeting of your Main Committee [on intercept co-ordination]’, adding that ‘no action has been taken to improve the position (in regard to interception of Enigma traffic)’. But despite Denniston’s plea, little or nothing seems to have happened: the Y Committee did not consider the subject until July. In July, 85 per cent of the Chatham intercept facilities were allocated to Enigma, mostly Luftwaffe, yet no Luftwaffe Enigma was yet being intercepted by the RAF, which was just about to send forty operators to Chicksands (a naval Y station) for training.

  In 1940, the services, mainly the War Office, and not the user (Hut 6), controlled the tasks undertaken by the stations. Around August 1940 Chatham even removed six sets from Enigma cover without notifying GC&CS. Hut 6 protested, but received scant sympathy from the military and Air Force authorities, who considered it ‘an act of grace on their part to allow GC and CS any voice in the allocation’ of sets. Hut 6 even had to give battle in order to prevent Enigma coverage being transferred from highly skilled Army operators to unskilled RAF personnel, with potentially disastrous results for breaking Enigma. Although RAF Y had not yet started to intercept any Luftwaffe ground to ground traffic, Blandy wrote patronizingly about Hut 6’s protest: ‘My only comment is that the authors of this document have just begun to understand the niceties of wireless interception. If they had all done a course at Chatham or Cheadle they would not have wasted so much paper.’

  As more ciphers were broken and as the cover on them improved, more cribs and re-encipherments were discovered, which in turn required more intercept sets to exploit them. Hut 6 ideally needed coverage on every Enigma cipher, however unimportant it appeared to be, but in 1941 it had to complain about ‘the [lamentable and inexcusable] failure of the Army authorities to provide an adequate number of trained operators’. It is therefore not surprising that two postwar histories described the expansion of intercept facilities during 1940 as ‘astonishingly and lamentably slow’ and ‘the machinery of the Y services [as] not then functioning well’. By March 1941, Chatham’s complement of sets had only increased by two, to twenty-seven, although there was also a second Army station at Harpenden. A newly formed ‘E’ Sub-Committee of the Y Committee decided in March 1941 that 190 sets were needed for Enigma. One hundred and eighty Army, RAF and Foreign Office sets in Britain and abroad were taking Enigma traffic by November 1941. However, there was a shortage of skilled operators for many months, partly because ex-Post Office operators proved unsuitable. Militarizing personnel at the stations, and forming an ATS intercept unit, was largely to prove the answer for the Army’s purposes. It turned out to be a slow process, and it was only in November 1942 that a set room, with thirty-six sets, was completely staffed by ATS at a new Army station, Beaumanor.

  In January 1942, the Chiefs of Staff authorized a major expansion of all the Y services, including the recruitment of 7,150 additional personnel, together with the establishment of a much-needed radio network to handle traffic from overseas Y stations and outposts of GC&CS. The sets on Enigma at home and abroad increased from 210 in April 1942 to 322 in January 1943. The Army and Foreign Office provided 64 per cent of the intercept sets, but 60 per cent of the daily keys solved were Luftwaffe, and only 26 per cent Heer; 10 per cent were Railway daily keys, and 5 per cent were SS. On 3 March 1943, the Chiefs of Staff authorized a second Y expansion programme. By May 1943, the WOY G (War Office Y Group) at Beaumanor had 105 sets on Heer and Luftwaffe Enigma traffic, and the main RAF station at Chicksands, ninety-nine receivers. There were also sixty sets on Enigma at Foreign Office stations at Whitchurch, Denmark Hill, London (formerly a Metropolitan Police station) and Cupar, an Army station at Harpenden and RAF stations at Shaftesbury and Wick, plus seventy-five sets in the Middle East and the Mediterranean. By June 1944, 515 sets were tasked on Luftwaffe and Heer Enigma, but even as late as January 1945, Hut 6 was ‘as short of sets as ever’: minor Heer and Luftwaffe ciphers were not being intercepted, and no attempt could therefore be made to break them.

  The service authorities came to accept Hut 6’s view that Enigma interception was an indivisible problem, and allocated Army sets to Luftwaffe traffic, and RAF sets to Heer signals. Later, they also relinquished control of tasking the Army and RAF sets. In May 1943 Hut 6’s intercept control section under John Coleman specified the Enigma tasks to be taken by the intercept stations, and the number of sets to be allocated, but the stations remained under the administrative control of their parent services. Coleman’s section also co-ordinated interception in the Middle East with interception in Britain.

  Accurate interception was essential when attacking Enigma, since hours could be lost because of a single wrong letter or call sign, or even a mistake in measuring the frequency. Hut 6 was unable to break Yellow for 5 May 1940, because of a single incorrect letter in the intercept, although the Poles solved it on 7 May. When resources allowed, radio nets were therefore often covered by two (or sometimes even six) sets to ensure that signals were precisely recorded. Only first-rate operators could deal with very faint signals coupled with interference, and drifting or split frequencies, but they were still in short supply even in December 1942. Hut 6 estimated that only about one third of the operators then at Beaumanor and Chicksands were first-rate, and another third, second-rate. Taking a burst of Morse from ‘a distant signal underneath the cacophony of different Morse transmissions, a diva singing grand opera in German, [and] a high-speed Morse transmission’ required a very high degree of skill.

  Breaking Red was Hut 6’s most important task throughout the war, as can be seen from the number of radio receivers allocated to intercepting it: in July 1941, sixty-eight sets were taking Red – over half the 119 sets in Britain tasked on Enigma, although the proportion had declined slightly, to about 35 per cent (fifty sets) in October. The average daily traffic on Red of 380 Teile (message parts) from June to November 1942 was 65 per cent of the total combined traffic (590 Teile) intercepted on all the Army and SS cipher nets, while the average daily total of the Luftwaffe traffic (1,400 Teile) was over double that of the Heer and SS. Red was easy to break once continuity had been established: the net was so widespread that if one crib went down there was a good chance of finding another to replace it. Red’s links to many other Luftwaffe keys made it possible to penetrate them by cribs from re-encipherments, which were known as ‘kisses’ in Bletchley Park parlance, because they marked the relevant signals with ‘xx’. Red was also an invaluable source of intelligence on Heer topics in North Africa and elsewhere.

  Hut 6 solved the Luftwaffe
Light Blue cipher, which provided intelligence about the Heer and Luftwaffe in Libya, within about eight weeks of its introduction in January 1941, and read it daily until it went out of service at the end of 1941. The only other Luftwaffe cipher intercepted in 1941, Mustard, the field cipher on the eastern front of the Luftwaffe Sigint service, the Funkaufklärungsdienst (Radio Reconnaissance Service), was solved for twelve days in the late summer, and from April 1942. It proved useful in giving the order of battle of the Soviet Army and Air Force, and by revealing the very considerable insecurity of Soviet ciphers.

  In 1942 all Luftwaffe Enigma key-lists except Brown were apparently prepared by one man, who often merged different components of previous keys (for example, Stecker) when preparing new key-lists. The April Foxglove keys combined parts of the January Red and March Mustard keys, with the remainders being used to make up the April Red key-list. In June, Locust, Mosquito, Snowdrop, Hornet and Garlic also consisted of partial key repeats. Key repeats died out at the end of 1942, but were briefly revived between March and June 1943, when ciphers used by certain Luftgau (Luftwaffe administrative units) were allotted different discriminants, but some key components were identical, for example, Daffodil, Gorse, Lily and Speedwell formed one group.

  Primrose, the cipher for Luftgau XXVIII, the Luftwaffe supply unit in Africa, yielded between 140 and 290 decrypts daily, mainly on logistics, on most days between June and November 1942 – almost as many decrypts as the prolific Red, whose total varied from 350 to 445 daily. Primrose was doubly useful since it also provided complete key repeats, although in random order, for Scorpion (employed by Luftwaffe close support units and Luftwaffe liaison officers with the Afrika Korps) for each month from June to November. Scorpion was very difficult to intercept in Britain, since it was transmitted with low power on medium frequencies by units near the fighting. Special arrangements were made to decipher it at Heliopolis, near Cairo, as well as at GC&CS, since it produced a great deal of intelligence about Luftwaffe plans and the ground fighting. The daily keys were available in advance so that messages in it could therefore be deciphered immediately they reached the Heliopolis centre.

  In complete contrast to the Luftwaffe, Heer traffic in 1941 was divided into small, unrelated groups, with few re-encipherments, and light traffic. Hut 6 broke very little Heer Enigma before September 1941. Thus, although Vulture (used on the Russian front) was first read on 27 June 1941, shortly after the invasion of Russia, Hut 6 broke no Vulture keys for August, and only one for September. It solved only twelve Heer daily keys in August, all in Kestrel. By November Hut 6 was breaking both Vulture, which was the first Heer cipher to be broken regularly, and Chaffinch, but then lost them for four months after the Vulture traffic was reduced when the Heer began to rely more on land lines.

  Heer ciphers in North Africa were initially very difficult to solve. Hut 6 broke Chaffinch II (used between Panzerarmee Afrika and Rome and Berlin) for nine days in September 1941, and for part of October and from 2 November to 6 December, although often a week or more later. Traffic on Phoenix, which rose from twenty-five to over 200 messages daily during fighting in Libya, could only be taken by forward intercept sets with the 8th Army, since it was also transmitted at low power on medium frequencies. When it was broken, delays in sending the traffic back from 8th Army to Heliopolis for transmission to GC&CS led to considerable hold-ups, until arrangements were made for Hut 6 to send key-lists to Heliopolis. Hut 6’s solution of the Heer Mediterranean ciphers, such as Chaffinch, Phoenix and Thrush (which carried information about the air transport of supplies and reinforcements) increased considerably after April 1942. Even so, only 2,800 Heer signals were decrypted in May 1942 out of 10,300 Heer intercepts, compared with 19,400 Luftwaffe decrypts from 31,000 intercepts, illustrating just how difficult it was to solve Heer Enigma.

  Hut 6 found that many of the twenty-two Luftwaffe ciphers it was regularly breaking by the summer of 1942 were interrelated: the fifty Heer and Luftwaffe Enigma ciphers then being used presented it with a single, indivisible problem. It therefore had to try to break every single cipher, however unimportant for intelligence purposes, since there was no way of telling in advance which would prove the entry point to a different cipher. Even Heer Mediterranean ciphers, such as Mallard (a Rome administrative cipher), were sometimes broken by re-encipherments from Red and Scorpion. All the Mediterranean Heer keys solved in 1942 were linked by re-encipherments, but the only Heer cipher to produce cillies consistently in 1942 was Osprey which, although classified as a Heer cipher by Hut 6, was really the cipher of the Todt construction organization, which may explain its poor operating practices.

  No repeats of key-lists were discovered in Heer traffic in 1941 or 1942. The success rate (decrypts as a percentage of the signals intercepted) against Luftwaffe traffic fell from 92 per cent in November 1941 to 50 per cent in November 1942, largely because Light Blue went out of service in January 1942, and the main Luftwaffe traffic started to use eight or more ciphers. During 1942, the success rate against Heer traffic was only 0.6 per cent (a mere thirty decrypts) in January but rose to 27 per cent in May. Intercepts of all Hut 6 traffic increased from 32,000 per month in September 1941 to 82,000 in November 1942. The overall percentage of the traffic (Luftwaffe, Heer and Railway) broken by Hut 6 remained around 50 per cent of the total intercepts in 1942. Unidentified traffic, where Hut 6 did not know which cipher was being used, was reduced from 17 per cent of the total traffic in September 1941 to only 2.4 per cent in November 1942, illustrating Hut 6’s and the Y service’s increasing mastery of the Luftwaffe and Heer radio nets, and their skill in reconstructing the extensive lists of cipher discriminants used by them.

  SS decrypts increased considerably in 1942 following the discovery in April that daily reports (codenamed HOR-HUG by Bletchley Park) of concentration camp numbers provided half the Orange Stecker. The HOR-HUG reports gave the numbers of people in certain concentration camps, which were encoded by a letter for digit substitution (so providing ten letters – the first five Stecker pairs), but ceased in early 1943, after which the reports were sent by land line. Orange keys could often be broken with an unused part of a bombe, while the other parts were being used for a different cipher.

  Hut 6 also solved the Railway traffic in eastern Europe (later codenamed Rocket) in 1941 and 1942. The traffic used a rewired version of commercial Enigma, without Stecker, and was therefore relatively easy to solve using hand methods. Lt. Col. John Tiltman first broke the traffic around the end of July 1940, after which the wiring of the machine was solved by Hut 8, which became responsible for reading the traffic until the work was later transferred to Hut 6. All 2,300 messages received in July and August 1940 were solved. The traffic ceased at the end of August, but began to be intercepted again in February 1941. The 90 per cent success rate against it in 1942 was the highest for all Hut 6 Enigma.

  Hut 6 largely consolidated its position during 1943. However, it faced a significant challenge when the Heer stopped using discriminants in Enigma traffic on 1 September, and the Luftwaffe followed suit on 1 November. The small change represented a major improvement in security, since Hut 6 was now faced with 3,000 or so signals daily, all of which apparently used the same cipher but which in fact employed up to ninety different ciphers. The new procedure slowed Hut 6 quite considerably, and required it to increase its registration and decoding room staff substantially, from about 220 to 300. Hut 6 and the Y service were able to meet the challenge only because they had become highly experienced and very flexible organizations, capable of responding to any change quickly and efficiently.

  In addition, Hut 6 had to prepare for the second front in late 1943 and early 1944. It identified many of the Heer’s Enigma ciphers and their related frequencies and call signs with the help of the Y service, although it had little success in solving them before the Normandy landings in June, and thought that most of its intelligence on the Heer might have to be derived from Red and Flivo (Heer-Luftwaffe liaison) ciphers. Fortunate
ly, some of the Heer ciphers being used in France were solved after the landings, although initially there were not many decrypts. However, following the Allied break-out in Normandy at the start of August, there was a massive increase in Heer traffic, which led to about five Heer ciphers being read daily and a huge rise in Heer decrypts, although the decrypts declined in September 1944 when the land war became static. So did Luftwaffe decrypts, when Luftwaffe operations were restricted, largely due to a shortage of fuel and trained pilots.

  Throughout 1944 Hut 6 was very apprehensive that a new rewirable reflector, Umkehrwalze D (Dora – UKD), which had first appeared in some Luftwaffe ciphers in January, would be brought into general use, and blind it almost completely. Providentially, despite warnings that UKD would be widely used on 1 August, that did not happen. UKD was employed with about twenty-five Luftwaffe ciphers by March 1945, but seldom to the total exclusion of the standard reflector in the same cipher, which was a classic Luftwaffe blunder of the first order. But for the Luftwaffe bungles UKD would have had a devastating effect on Hut 6 and the production of Ultra.

  The Heer brought an unbreakable system of enciphered call signs into service on 1 November 1944. This was a much-feared step, but it led to an only temporary drop in solutions of Heer Enigma, such were the combined capabilities of the Y service and Hut 6, who were greatly helped by the fact that major nets retained their fixed frequencies, confining the problem to about half of the Heer traffic. Almost four times as much bombe time was required to break Heer Enigma in November 1944, compared with November 1942, and more than twice as much for Luftwaffe Enigma (see Figure 4.3). Hut 6’s production of decrypts in 1944 would almost certainly have declined significantly if US Navy bombes had not been available, since they more than doubled the three-rotor bombe capacity available to Hut 6.