The Secret War: Spies, Codes and Guerrillas 1939–1945

These men, and a few hundred others throughout the armed forces, spent much of the war exploiting and assessing information derived overwhelmingly from interception and decryption of the enemy’s wireless traffic. Bill Williams, who served in the Mediterranean until 1943 and in Europe thereafter, stated in an important 1945 report: ‘It must be made quite clear that Ultra and Ultra only put intelligence on the map.’ Until decrypts began to become available in bulk in 1942, ‘Intelligence was the Cinderella of the staff … Information about the enemy was frequently treated as interesting rather than valuable [though] of course this attitude varied according to the commander.’

Scepticism was often merited, because much material was downright specious. The 1940 war diary of the army’s Middle East intelligence section in Cairo included comically frivolous snippets: ‘All Hungarian cabaret artistes have been ordered to leave the country by the end of May.’ Data about the Italian army was scanty, so that on 9 August the section recorded: ‘The present location and organisation of Libyan troops in Eastern Cyrenaica is obscure.’ A despondent staff officer added a week later: ‘There has been no further reliable information of fresh [Italian] ground units or formations arriving in Libya from overseas.’ On 27 September, the British high command’s weekly intelligence summary included a paragraph on domestic conditions in Germany: ‘A neutral traveller to the Leipsic fair, whose personal observations are believed reliable, reports that relations between the [Nazi] Party and the Army are not good.’ Three months later, the head of MI6’s Political Section wrung his hands: ‘It is piteous to find ourselves in this state of ignorance’ about both Germany’s internal condition and economy.

Only when Allied warlords were empowered to read the messages being exchanged between enemy generals in the field and their higher headquarters was scepticism about the value of ‘intelligence’ replaced by increasingly fervent belief. Ultra forced commanders-in-chief, not to mention the prime minister, to treat senior intelligence officers with a respect they had seldom received in the pre-Bletchley universe. Brigadier Ian Jacob of the war cabinet secretariat said: ‘My impression is that once the Ultra business got well-established, Churchill didn’t look at anything else.’ Eisenhower’s intelligence chief Kenneth Strong wrote in 1943, in a memorandum on training staff officers: ‘We no longer depend on agents and cloak-and-dagger sources for our information. Modern methods have completely transformed intelligence.’

He meant codebreaking, of course, and in Britain the fountainhead of such activity was the Government Code & Cypher School at Bletchley. In the months following the outbreak of war, GC&CS expanded dramatically with the arrival of a stream of academics, many of them earmarked by its recruiters before the war. Though some were seconded from the armed forces, it was understood that there was no need to train the universities’ contingent to march, blanco webbing, and name the parts of a rifle. They remained their sallow, tweedy, pipe-smoking young selves when housed in lodgings around the dreary suburban town, and enlisted on the government payroll without uniform or ceremony. Twenty-year-old mathematician Keith Batey found his landlady demanding an assurance from his employer that he was not a despised ‘conchy’ – conscientious objector – before he joined the growing body of academics working on a task of supreme importance to their country, fulfilment of which might do something to assuage its shocking vulnerability. What was the task? Bletchley’s little band, 169 strong in 1939 including support staff, understood only that the nation’s enemies communicated in a multitude of codes and ciphers, vulnerable to interception. If even a portion of these combinations of numbers and letters could be rendered intelligible, information might be gained of priceless value to the war effort.

Nobody knew, in the beginning, whether a given message hijacked from the airwaves might be an order from Hitler for his armies to march on Warsaw, or a request from a Luftwaffe airfield in eastern Germany for a delivery of filing cabinets. Ahead of the codebreakers lay a mammoth menu of requirements which could only be addressed as mobilisation sluggishly made available ears, brains and hands to monitor the enemy’s frequencies around the clock, log some of his vast output of messages, fix the locations and possible identities of the senders – diplomatic, police, military, naval or air force. Then came the much greater challenge, of discovering what the messages meant.

All radio communications involved a trade-off between speed and security. At the simplest level, battlefield direction by land, sea and air required some voice linkage. This enabled the instantaneous passage of orders and information, at the cost of being overheard by anybody else who cared to tune to a given frequency. Crude security could be introduced by using coded callsigns in place of names and suchlike – during the Battle of Britain fighter controllers added 5,000 feet to indicated altitudes, to confuse eavesdroppers. But voice messaging was inherently insecure: sensitive information should never be passed verbally, though it often was.

Most military messages were instead wirelessed by Morse key. Low-level material could be rapidly encrypted under battlefield conditions by relatively unsophisticated personnel using so-called hand- or field-ciphers, usually involving groups of two or three letters or numbers – the Kriegsmarine employed twenty-seven variants. More sensitive traffic, issuing from higher echelons, was translated by machine-generated or manual ciphers, usually involving combinations of four or five letters or numbers. The British thought justifiably highly of the security of their Type-X machines, though they never had enough of them.* (#ulink_69655881-16d6-5703-885c-8e3fca7501d7) The Americans rightly trusted their Sigaba, a fifteen-rotor system.

For substantial periods between 1939 and 1943 the Germans broke some Allied codes, including those of the US State Department and military attachés, along with the traffic of several exile governments, notably the Poles and Free French. They sometimes also accessed messages of all three British services, including the RAF’s four-character cipher, and later had successes in attacking products of the US Army’s M-209 field-ciphering machine. It deserves emphasis that Allied code-security weaknesses, and enemy achievements in exploiting them, gave the Germans much more operational assistance than some Western historians acknowledge, especially in the Battle of the Atlantic. However, higher British, American and Russian communications defied enemy scrutiny: Nazi eavesdropping on transatlantic telephone conversations between Churchill and Roosevelt told Berlin little of value. Modern claims that the Germans broke into Russian higher ciphers deserve to be treated with caution: certainly from 1942 onwards, there is no evidence that Hitler’s generals profited from any such insights; if they had, they would have been less often deluded by Soviet deceptions.

Most German senior officers – though by no means all their cryptographers – were confident that Enigma ciphering machines, which scrambled messages by means of shifting rotors and a plugboard, and rendered them comprehensible only by a matching machine with identical settings, were immune to the attention of any enemy, and indeed to the workings of the human brain. It is unsurprising that in 1939 they discounted the possibility that electro-mechanical technology might dramatically accelerate exposure of the Enigma’s secrets, because it did not then exist. It is extraordinary, however, that such serene confidence persisted through six years that followed, even following the discovery that the Poles had broken some pre-war Enigma traffic, and several warnings from their own experts. Amazing hubris was expressed by the Wehrmacht’s last signals chief, Lt. Gen. Albert Praun, who preened himself before his Allied captors after the war ended: ‘The achievements of German communications intelligence … may speak in favour of the German type of intelligence organisation.’ His organisation, he said, ‘gave German commanders a hitherto unattained degree of [signal] security’.

The British breaking of the Enigma, then subsequently and separately of German teleprinter traffic, was a progressive, incremental operation which attained maturity only between 1943 and 1945, and was never uninterrupted or comprehensive: even at peaks, only about half of all intercepts were read, many of them too late to provide practical assistance ‘at the sharp end’. What was done at Bletchley Park was indeed miraculous, but the codebreakers were never able to walk on all of the water, all of the time.

The 1939–40 Phoney War conferred few benefits on Britain, but it granted GC&CS precious time to bolster its strength and refine its methods. Without mechanical aids Bletchley’s brainstormers made modest and delayed breaches in a small number of enemy ciphers. The Germans employed acronyms and codenames which took weeks or months for their enemies to interpret. The importance of what happened at Bletchley in the first two years of war was not that it enabled Britain’s generals to avert or arrest a disastrous run of defeats, which it certainly did not, but that it lit a candle of hope about what the codebreakers and their embryo technology might accomplish in the future. It enabled the directors of the war effort to lay upon the board a few scattered pieces of a vast jigsaw, which would be filled only during the Allies’ years of victory.

Bletchley Park – Station X, Box 111 c/o The Foreign Office – was a notably ugly Victorian pile of bastard architectural origins surrounded by fifty-five acres of trees and grassland, located fifty miles from London. It was purchased in 1938 to house GC&CS at a safe distance from German bombs by Admiral Sir Hugh Sinclair, then head of MI6; as legend has it, he used £7,500 out of his own pocket, but more plausibly he paid with secret funds under his control. Whatever MI6’s humint weaknesses, the service’s chiefs, especially Sinclair, deserve full credit for backing the establishment of Bletchley at a time when resources were desperately constrained. Work began at once on laying direct phone and teleprinter lines to London, and in the following year MI6’s skeleton team of cryptanalysts moved from Broadway to the Park, where they came under the orders of Alastair Denniston. One of his colleagues from the old Admiralty days, Dillwyn Knox, an expert on ancient Egyptian papyri, became an early Bletchley stalwart. The most prominent of the younger recruits were Gordon Welchman of Sidney Sussex College, Cambridge, Hugh Alexander, Stuart Milner-Barry, John Jeffreys – and Alan Turing.

This last, the twenty-seven-year-old son of an Indian civil servant and the product of an austere and emotionally arid childhood, had just returned from a stint at Princeton clutching one of his own creations, a so-called electric multiplier machine mounted on a breadboard. His headmaster at Sherborne had once written: ‘If he is to stay at a public school, he must aim at becoming educated. If he is to be solely a Scientific Specialist, then he is wasting his time.’ In the headmaster’s terms, Turing had indeed been ‘wasting his time’: he had evolved into a shy, narrow, obsessive. Noel Annan wrote: ‘I liked his sly, secret humour … His inner life was more real to him than actuality. He disliked authority wherever he was … [and] enjoyed games and treasure hunts and silliness … Turing was the purest type of homosexual, longing for affection and love that lasted.’ More even than by his sexuality and his often childlike immaturity, however, his tragedy was to be afflicted by the exquisitely painful loneliness of genius.

Other drafts of young academics followed, variously codebreakers and linguists, together with the first of what became successive waves of young women, who would play a vital role in the operations of ‘BP’. The first two of these were daughters of golfing partners of Denniston, reflecting the importance of personal connections in Bletchley’s recruitment process in the early days, before industrialisation became inescapable. Indeed, the whole wartime intelligence machine emphasised the cosiness of the upper reaches of British life. Oxford University Press was entrusted with responsibility for printing vast quantities of codes, maps and reports, because of its pre-war experience producing examination papers under secure conditions. The Admiralty’s liaison with OUP was handled by Margaret Godfrey, wife of the director of naval intelligence. The Royal Navy’s Topographical Photographic Library was housed in the basement of the nearby Bodleian Library, which eventually dispatched 300,000 images a month to operational areas. The World War I intelligence veteran Admiral Sir William ‘Blinker’ Hall introduced Godfrey, his modern successor, to the City of London banking giants Montagu Norman, Olaf Hambro and the Rothschilds, who helped to identify suitable recruits for the NID.

Candidates being scrutinised for Bletchley were often asked: ‘Do you have religious scruples about reading other people’s correspondence?’ Twenty-year-old Harry Hinsley was interviewed at St John’s College, Cambridge by Alastair Denniston and Col. John Tiltman, the senior codebreaker. They said: ‘You’ve travelled a bit, we understand. You’ve done quite well in your Tripos. What do you think of government service? Would you rather have that than be conscripted?’ Hinsley would indeed, and joined the Naval Section located in Bletchley’s Hut 4. Through the icy winter of 1939–40, such men and women wrestled with Enigma traffic. Working conditions were dismal, with staff muffled in overcoats and mittens. The first break into a Luftwaffe Enigma key – designated ‘Green’ – is thought to have been made on 25 October 1939. In December, by unaided intellectual effort Alan Turing is believed to have broken five days’ worth of old naval messages. By the end of March, the French – or rather, the Poles working at France’s Station Bruno – had broken twenty days’ worth of old signals and BP about thirty, all Luftwaffe traffic.

Turing was much more importantly engaged. He compiled a 150-page treatise on Enigma, studded with schoolboyish blots, deletions and illegibilities. While most codebreakers addressed each other by first names or nicknames, heedless of age and status, almost everyone knew Turing as ‘Prof’ rather than as Alan. When his Enigma study was circulated later in 1940, it became known as ‘the Prof’s book’. He also set about fulfilling his concept for a ‘bombe’, a primitive but revolutionary electro-mechanical device for exploring multiple mathematical combinations. This borrowed its name, though not its design, from the Polish ‘bomby’, and would be capable of examining the 17,576 possible wheel deployments for a three-rotor Enigma in about twenty minutes: the order for the first machine was placed in October 1939, and the prototype became operational six months later. Meanwhile, outside in the park, workmen sawed and hammered at an ever-widening array of low wooden buildings which housed the growing staff. Eventually, only administrators worked in the main building, where the telephone switchboard was established in the ballroom. In the huts, signals were shifted from one section to another on a small trolley pushed along a makeshift wooden tunnel.

Hut 8 attacked German naval traffic, which was then passed to Hut 4 for translation and processing. Hut 3 performed the same function for Wehrmacht and Luftwaffe traffic decrypted by Hut 6. The former would eventually play a pivotal role in Allied wartime intelligence, but in its early incarnation it had a staff of just four. Frank Lucas, who was one of them, wrote: ‘On a snowy January morning of 1940, in a small bleak wooden room with nothing but a table and three chairs, the first bundle of Enigma decodes appeared. [We] had no idea what they were about to disclose.’ A few score yards away, Hut 6 run by Gordon Welchman wrestled with army-Luftwaffe ‘Red’ key traffic, which was the first to be broken in bulk.

From the outset, pains were taken to disguise from all but the most senior operational commanders the fact that information was being gained from codebreaking. This gave an unintended boost to the prestige of MI6, and to that of Stewart Menzies in particular. When Reg Jones gave a disguised report based on an Ultra decrypt to the RAF’s director of signals, Air Commodore Nutting, the airman professed astonished admiration for the courage of the presumed spies who had provided the information, saying, ‘By Jove, you’ve got some brave chaps working for you!’ The ever-growing scale of the enemy traffic to be trawled was intimidating. It is a measure of the expansion of communications as a branch of warfare that by August 1943, 305,000 personnel among the Luftwaffe’s total strength of 2.3 million were employed on signals duties – transmitting, receiving or processing – and the same was true on both sides of the war, and of all armed forces.

At GC&CS there were inevitable personality clashes. Gordon Welchman, whose creative contribution became second only to that of Turing, and whose organisational skills were also priceless, found it hard to work with the highly-strung and fractious Dillwyn Knox, a contender for the hotly contested title of Bletchley’s star eccentric. A notoriously awful driver, Knox giggled: ‘It’s amazing how people smile, and apologise to you, when you knock them over.’ He sustained a stream of intemperate complaints and demands to Denniston, his old comrade from Room 40, about staff shortages, working conditions, low pay, together with the intrusion of and excessive authority conceded to non-cryptanalysts: service intelligence officers ‘who maul and conceal our results’. Knox was seriously ill with the cancer that would kill him in February 1943, but meanwhile he and Welchman bickered: the older man accused the younger of exceeding his narrow initial brief, and was also impatient of Turing, writing, ‘He is very difficult to anchor down. He is very clever but quite irresponsible and throws out a mass of suggestions of all degrees of merit. I have just, but only just, enough authority and ability to keep him and his ideas in some sort of order and discipline. But he is very nice about it all.’ Turing prompted mirth by joining Bletchley’s Home Guard because he was seized by an impulse to learn to shoot, then provoked the apoplectic rage of its colonel by absenting himself from parades once he had fulfilled this private purpose. His unworldliness could provoke real exasperation among those under relentless pressure to produce results. A colleague spoke of Turing’s ‘almost total inability to make himself understood’.

There were plenty of minor tantrums lower down the hierarchy, unsurprising when staff were performing stressful tasks through long hours in bleak working conditions. Angus Wilson, the later novelist, once vented such a storm of rage that a colleague said wearily, ‘Do stop it, Angus, otherwise we’ll put you in the lake!’ Wilson retorted defiantly, ‘Don’t worry, I’ll do it myself,’ and duly plunged into the water in front of the house; on another such occasion he hurled a bottle of ink at a Wren. Many wartime codebreakers suffered temporary or permanent physical or mental collapses, brought on by their work: William Friedman, one of America’s pioneer practitioners, underwent a nervous breakdown in January 1941 which incapacitated him for three months. Hugh Trevor-Roper languished for several months at about the same time, and others regularly succumbed.

German signals were at first intercepted by a battery of army wireless-operators stationed in an old naval fort at Chatham, a role later assumed by Gambier-Parry’s organisation at Whaddon Hall. In the early war years there were never enough operators, and both the RAF and the army were reluctant to acknowledge the priority of meeting GC&CS’s demands. Signals were brought in batches to Bletchley’s guardroom by motorcycle couriers, at all hours and often in dreadful weather, then distributed between the relevant huts. From an early stage, the codebreakers learned to identify German senders by the unencrypted preamble to their message texts, none of which was longer than 250 words. It was then a matter of sitting hunched over a deal table through the hours of a shift, pondering a jumble of numbers and letters from which only men – and Bletchley’s handful of women – with remarkable logical or mathematical powers might hope to extract fragments of meaning. ‘The ideal cryptanalyst,’ Stephen Budiansky has written, ‘was Beethoven with the soul of an accountant.’ When Christopher Morris was a new recruit to Bletchley he heard one of his senior colleagues, asked the requirements for the job, respond laconically, ‘Oh, I suppose a sharp pencil and a piece of squared paper.’ Morris himself thought that the main requisites – except at the exalted level of Knox, Turing, Welchman and later Max Newman – were ‘patience, accuracy, stamina, a reasonably clear head, some experience and an ability to work with others’.

They opened what became vast card indexes, stacked in shoeboxes along the sides of the huts, cataloguing enemy units, personnel, codenames, locations, abbreviations and military hardware; different Enigma keys were distinguished by colour – for instance, yellow, green, red and blue, respectively indicating Norway, Wehrmacht, army-Luftwaffe and air training codes. ‘When a new word came up in the message you were translating,’ wrote Hugh Skillen, ‘a new type of jet fuel, or machine part – you looked for it, and if it was not there, the indexer put it in with a reference time and date stamp.’ Bletchley’s meticulous record-keeping became a critical element in its triumphs.

For security reasons, the Park’s operations were rigidly compartmentalised, and there was little exchange of information or gossip between sections. Even Welchman remarked years later how little he knew about what colleagues were doing a few yards away from his own Hut. As the staff grew from hundreds into thousands, facilities lagged behind: a section head complained that two hundred men and women enjoyed the use of just one lavatory. The food in the Bletchley canteen was poor even by wartime standards. Former debutante Sarah Norton one night found a cooked cockroach in her meat: ‘I was about to return it to the catering manageress when my friend Osla, who had the appetite of a lioness with cubs, snatched the plate and said: “What a waste – I’ll eat it!”’ The nearest available delicacies outside the wire were to be found at the Station Inn in Bletchley town, which offered ox heart. Welchman recalled having to provide his own newspaper to wrap fish and chips at the local shop. The codebreakers worked around the clock in three rotating shifts, starting with 8 a.m. to 4 p.m. When the exhausted men and women cycled or took a bus through the blackout to their lodgings, they found few comforts: dim lighting, hot baths often rationed to one a week, draconian rules about inter-sex visiting.

It is deeply impressive that those who worked at Bletchley sustained such dedication while working day after day, month after month, in a drab world devoid of glamour, excitement, variety, glory and decorations. In Hut 3, the watch sat around a horseshoe table, translating deciphered signals, each one resembling a telegraph flimsy, forwarded from Hut 6. Ideally, decrypts were composed of complete German words, but often there were interruptions and corruptions in the texts, which demanded leaps of imagination from the linguists. William Millward recalled with shame a night when he invented a place named ‘Senke’, near Qatara in the North African desert – having forgotten that Senke was the German word for a geographical depression. Schoolmasters proved ideal as watch chiefs, wrote Peter Calvocoressi, because they were naturally meticulous: ‘If not satisfied, they would throw back a translation at even an eminent professor. It reminded me of Chief Examiners at “A” Level who would send back scripts to an Assistant Examiner to re-mark.’ No one could work at GC&CS who did not love brain games. There were dreary, idle yet sleepless lulls, when no traffic arrived for the watches to work on. Peter Hilton once used such a doodle time to compose a palindrome: ‘DOC NOTE, I DISSENT, A FAST NEVER PREVENTS A FATNESS. I DIET ON COD’.

Although Alan Turing was acknowledged as the highest intellect at Bletchley, its achievement was supremely a team effort; the creative input of some others, Welchman prominent among them, was almost as important as that of Turing. One night in February 1940, several months before the arrival of the first bombe, twenty-one-year-old Cambridge mathematician John Herivel was smoking a pipe before the fire in his billet, and concentrating furiously on encoded messages even as he drifted in and out of a doze. An inspiration struck him as he gazed with the mind’s eye at a German Enigma operator. He perceived such a man starting his morning’s work bored or weary or hungover, and thus not troubling to change the previous day’s ring setting on his machine before starting to cipher messages. Herivel scarcely slept that night, as he went on to deduce how such an act of carelessness might be detected, then exploited to break a message.

Welchman, who had supervised him at Cambridge, immediately saw the importance of this flash of insight, a marriage of mathematical brilliance to a grasp of human weakness. He told the young man fervently that he ‘would not be forgotten’, and indeed his inspiration became known as ‘the Herivel tip’. Dillwyn Knox had already identified another entry point to messages, rooted in operator errors and text settings – what the codebreakers christened ‘Cillis’ or ‘Sillies’. Welchman wrote later that Bletchley remained ‘entirely dependent on Herivel tips and Cillis from the invasion of France to the end of the Battle of Britain’. In other words, until the arrival of the bombes, codebreaking was being done by raw brainpower, without significant mechanical assistance: at this early stage, the British lagged behind their American counterparts in exploiting technology – both the US Army and US Navy codebreaking teams used Hollerith punched-card sorters, of a kind that only began to appear at Bletchley in May 1940, because chief codebreaker Col. John Tiltman had been sceptical about them. Ultra provided no important material during the summer of 1940, but several indications about the postponement of ‘Sealion’, the Nazi invasion of Britain, notable among them a September Luftwaffe message ordering the dismantling of air transport equipment at Dutch airfields.

Fred Winterbotham, the MI6 officer who eventually became overseer of the ‘Special Liaison Unit’ network which fed Ultra decrypts to commanders in the field, described the first bombe – christened ‘Agnus’, corrupted to ‘Agnes’ – as resembling ‘some Eastern goddess who was destined to become the oracle of Bletchley’. It was installed in Hut 11 on 14 March 1940, but suffered substantial teething troubles. Gordon Welchman made an important contribution to Turing’s creation by devising a ‘diagonal board’, an element introduced into the first really effective model, which came into service in mid-August. Agnes and its many successors were not computers, because they had no memory. They were instead electro-mechanical key-finding aids, six and a half feet high and seven wide, mounted in bronze-coloured cabinets, and containing thirty-six banks of high-speed electrically-driven Enigma replicas. Each bombe contained eleven miles of wiring and a million soldered contacts. Built by the British Tabulating Machine Company at Letchworth, they depended partly on components assembled in scores of local village halls, by casual workers who had no clue of the importance to the war effort of the twenty-six-way cables and other small electrical parts they contributed.

With the assistance of a clue or ‘crib’ – a vital identifying link, usually a codebreaker’s guess about the nature of part of a given signal – a bombe could test millions of mathematical possibilities for the settings of three Enigma rotors. Figuratively, Agnes and her kin were bloodhounds needing a slipper or handkerchief to take up a scent. If there was no ‘crib’, the bombe could not solve the key – but mercifully often, there was one. Subsequent machines, miracles of reliability given their continuous operation, were given their own names by the Wrens who manned them around the clock, usually those of warships – Warspite, Victory and suchlike. The bombes did not take in enemy cipher messages and disgorge them in fluent German. They were instead priceless accelerators, once the codebreakers secured an insight into the nature of a given signal or traffic stream. Also useful was a battery of British Type-X cipher machines, modified to match the behaviour of Enigmas, on which Wren operators tested speculative message solutions. One of the principal constraints on codebreaking, especially between 1940 and 1942, was that access to the scarce technology had to be apportioned between competing claimants of the three services, and there was never enough ‘bombe time’ to go around.

Throughout 1940, human brainpower remained the dominant element in Bletchley’s successes, which increased with every passing week. It was ‘the Herivel tip’, not bombes, that enabled the team to crack the army-Luftwaffe ‘Red’ key in May. The overwhelming bulk of enemy traffic read through the rest of the year – around a thousand messages a day – was that of the Luftwaffe, and until the end of the war air force material was accessed more swiftly than that of the other services. An important requirement for success was what the codebreakers called ‘depth’ – possession of sufficient messages in a given key to give them playing space for calculations and speculations.

Luftwaffe signals provided many clues to the Wehrmacht’s parallel activities, but in the early days interpretation was impeded by lack of understanding of German terminology and abbreviations. In September 1940, Bletchley broke some traffic from Göring’s pathfinder unit, KGr100, which enabled it to forecast the targets of several bomber raids. But warning was of little practical usefulness to the defenders when hard power, in the form of radar-guided RAF night-fighters, was lacking in both numbers and effectiveness.

As more bombes were built – by 1945 there were 211 of them – they were dispersed around the London suburbs as a precaution against an enemy air attack on Bletchley. The operators, chiefly young women of the WRNS, found long hours beside the hot, smelly, clattering machines extraordinarily gruelling, especially when they were obliged to use tweezers to adjust the delicate electrical wiring. Some girls were unnerved by the monotonous racket. One of them said: ‘It was like a lot of knitting machines working – a kind of tickety-clickety noise.’ They went home with their shirt cuffs blackened by a fine spray of oil from the bombes’ revolving drums.

Naval Enigma remained for many months impenetrable. This was partly because its system of eight rotors, of which three were used at any one time, posed greater difficulties than the army’s five, and partly also because the Kriegsmarine’s operators were more disciplined than their Luftwaffe counterparts, and committed fewer errors to provide openings for Bletchley. There was a brief spasm of success in late April 1940, when five days’ traffic was read, but thereafter more than a year elapsed – an eternity in the minds of those who wrestled vainly with the problem day after day, week upon week – before the big breakthrough. Denniston said gloomily to Frank Birch, a 1918 veteran of Room 40 who now headed the naval section: ‘You know, the Germans don’t mean you to read their stuff, and I don’t suppose you ever will.’ Alan Turing himself had been dallying with the Kriegsmarine’s traffic almost since his arrival at Bletchley. A colleague, Hugh Alexander, observed that he became engaged because nobody else seemed to be making headway, and in his remote fashion he was fascinated by the abstract challenge.

It was Turing who devised a new method christened ‘Banburismus’, employing long punched paper sheets manufactured in the town of Banbury, which assisted the first important breaks into Kriegsmarine messages by reducing the number of possible Enigma rotor orders to be tested from 336 to around eighteen. This system was introduced in the spring of 1941, just as British losses to U-boats began to become alarming. On land, the British in those days lacked power and opportunity to do much with such knowledge of the Wehrmacht’s movements as they secured, and there was a large element of luck about what messages were broken. In North Africa in early 1941, the British Army profited from some good sigint derived from eavesdropping on the Italians, but few Enigma messages were broken quickly enough to assist decision-making on the battlefield. At sea, by contrast, there was an early golden prize for GC&CS’s labours.

The impetus towards success was provided by a series of captures far out on the ocean, which dramatically increased Bletchley’s knowledge of the enemy’s naval communications. On 23 February 1941, British commandos raiding the Lofoten Islands seized the German armed trawler Krebs, from which spare rotors for a naval Enigma were recovered, though the machine itself had been thrown overboard. This ‘pinch’ prompted the Royal Navy to launch an operation explicitly designed to capture more Enigma material, targeting German weather-reporting trawlers between Iceland and Jan Mayen Island. On 7 May, a sweep by three cruisers located and seized the München – but too late to save its Enigma and associated coding data from the Arctic deep. On 25 June the navy caught its sister vessel the Lauenburg, again minus its Enigma, but with a useful haul of cipher material.

Hut 8 now had enough information to read some U-boat signals, but the seizure which opened the traffic to fluent decryption was the fruit of chance and high courage, rather than of design. On 9 May 1941 a convoy escort group attacked and forced to the surface Julius Lempe’s U-110. A boarding party from HMS Bulldog commanded by Sub-Lieutenant David Balme secured the submarine, prevented its sinking, and brought back to his destroyer pearls beyond price: documentation for current Enigma. Though U-110 later sank under tow – fortunately so, from a security viewpoint – the short signal book, officer ciphering instructions and other material reached Bletchley safely, and the secret of the submarine’s capture was preserved beyond the war’s end. An Enigma machine was also recovered, but perversely this was the least useful element of the booty, because Bletchley had one already, together with assorted rotors seized in other ‘pinches’. Within days, Hut 8 was reading a steady stream of German naval messages. Ralph Erskine, one of the foremost experts on codebreaking at Bletchley, believes that the Park was already close to reading the Kriegsmarine traffic, even without the U-110 haul. What is for certain, however, is that it was impossible to break the U-boat ciphers without the assistance of captured material, which would again become a vital issue later in the war.

The breakthrough into the Kriegsmarine ciphers came just too late to influence the pursuit of the Bismarck in the latter days of May 1941. Conventional direction-finding on the behemoth’s wireless transmissions, supported by air reconnaissance, were the key factors in enabling the Royal Navy to intercept and sink it on the morning of the 27th, though assisted in the last stage by decryption of a Luftwaffe signal revealing Brest as Bismarck’s destination. Thereafter, Bletchley produced a steady stream of messages that revealed U-boat positions and intended courses. The so-called Hydra cipher was laid bare, and other keys were progressively broken: the more the Park knew, the more it was able to discover. The flow of decrypts was never assured, however, and disturbing delays sometimes took place. ‘Huff-Duff’ – High-Frequency Direction-Finding – played an important secondary role in the location of U-boats. The outcome was a relentless shift in the balance of advantage in the Battle of the Atlantic through 1941 and into 1942. Here was a case where intelligence indisputably and importantly influenced events.

Bletchley was also reading a significant portion of Italian naval traffic. On 25 March 1941, one of the small number of women decrypters, nineteen-year-old mathematician Mavis Lever in Dillwyn Knox’s team – he was famously supportive of talented girls in a male-dominated institution – played a critical role in breaking a message which revealed that the Italian fleet would shortly put to sea to attack British convoys. The warning enabled the Mediterranean C-in-C Admiral Sir Andrew Cunningham to contrive an encounter off Cape Matapan during the afternoon and night of 28 March which ended in a striking victory for the Royal Navy. By dawn on the 29th, three cruisers and two destroyers had been sunk, while the battleship Vittorio Veneto was damaged, an outcome that deterred the Italian surface fleet from making any further attempt to interdict British troop movements to Greece.

Spring brought an increasing flow of decrypts about Wehrmacht operations in the Eastern Mediterranean. Senior officers strove to streamline the transfer of information from Bletchley to battlefields, so that material reached commanders in real time. One of the most significant intercepts, detailing German plans for the May 1941 invasion of Crete, reported ‘probable date of ending preparations: 17/5. Proposed course of operation … Sharp attack against enemy air force, military camps and A/A positions … Troops of Fliegerkorps XI: parachute landing to occupy Maleme, Candia and Retiomo; transfer of dive-bombers and fighters to Maleme and Candia; air-landing operations by remainder of Fliegerkorps XI; sea-transport of flak units, further army elements and supplies.’ Churchill personally annotated the flimsy: ‘In view of the gt importance of this I shd like the actual text transmitted by MOST SECRET together with warnings about absolute secrecy.’ This information was passed to Wavell and Freyberg, the relevant commanders, at 2340 on 6 May. The loss of the subsequent Battle of Crete, following the German invasion which began on the morning of the 20th, emphasised a fundamental reality about Enigma decrypts: they could change outcomes only when British commanders and troops on the ground were sufficiently strong, competent and courageous effectively to exploit them. Stuart Milner-Barry of Hut 6 said that he and his colleagues looked back on Crete as ‘the greatest disappointment of the war. It seemed a near certainty that, with … every detail of the operation spelt out for us in advance … the attack would be ignominiously thrown back.’

The Cretan signal, informing British generals of German intentions in time to respond, was an exception rather than a commonplace in 1941. Bletchley was able to provide an ever-growing flow of information about the deployments of the enemy, not least in Eastern Europe, most of it derived from Luftwaffe and army-Luftwaffe decrypts. Wehrmacht traffic stubbornly resisted penetration, but German railway codes provided information about – for instance – troop movements to Yugoslavia, Greece and Eastern Europe in the summer of 1941. Hitler’s looming invasion of the Soviet Union, the towering event of the war, was also the first great strategic development for which Ultra intelligence provided explicit warning. While Britain had no power to influence or impede Hitler’s Operation ‘Barbarossa’, it was clearly of the highest importance to Churchill and his generals to be able to monitor its unfolding.

It became a source of increasing frustration to the prime minister that British troops in North Africa failed to frustrate or defeat Rommel when they had not only superiority in men, tanks and guns, but also an ever-growing stream of information about German deployments and movements, for instance at Halfaya Pass in May. Churchill pored intently over his own daily file of Ultra material. When he read a decrypt reporting petrol stocks at various Luftwaffe airfields in Libya, he scrawled on it in his red ink: ‘CAS [Chief of Air Staff] How many hours flying can their a/c do on this – about? WSC.’ Air Chief Marshal Sir Charles Portal responded testily: ‘Unfortunately it is not possible to make any general deduction since the figures only relate to the stock at Benghazi. We do not possess complete figures for the supply and consumption of oil and petrol throughout Libya. All we know is that there are indications of an overall shortage which is limiting operations in the forward area.’ This problem was endemic when decrypts were fragmentary. Stewart Menzies performed an important service by dissuading the prime minister from fulfilling his frequent desire to dispatch raw Ultra direct to commanders-in-chief in the field, as he had done in the case of Crete. ‘C’ was surely correct, on security grounds, and also because decrypts that lacked the context of other intelligence could be highly misleading to untrained eyes.

On land, in 1941 Bletchley provided more guidance to strategy than to tactics: it gave Churchill’s high command an authoritative, though never comprehensive, picture of German deployments in every theatre of war. Ultra could do little to assist the RAF’s ongoing struggle with the Luftwaffe for mastery of the skies. Only the Royal Navy gained immediate advantage, both in the Mediterranean and the Atlantic. Nothing altered the fact that, until the worldwide balance of strength began to shift in the Allies’ favour in the latter part of 1942, the operational superiority of German and Japanese forces enabled them to keep winning victories. Bletchley was an increasingly important weapon, but it was not a magic sword.

The practices and disciplines of GC&CS evolved progressively, with many wrangles and turf wars along the way. Deputy director Nigel de Grey complained about the ‘very low standards of military behaviour’ prevailing in what was supposed to be a military establishment. But how could it be otherwise? Noel Annan wrote: ‘Many of the cryptanalysts who produced Ultra were agnostic, heterodox dons who did not set much store by the normal interpretations of patriotism and democracy.’ It was not easy to combine the discipline essential to the operation’s smooth functioning with sensitivity to the wayward and frankly eccentric character of some of its resident geniuses. Col. Tiltman wrote ruefully on 2 March 1941: ‘Cryptanalysts have to be handled delicately and do not take kindly to service methods of control, which are essential to the good working of signals.’ When the director of the Royal Navy’s women personnel visited the Park, she demanded indignantly: ‘Why are my Wrens working with civilians?’ WAAFs in the teleprinter room expressed resentment about taking orders from civilians. In December 1940 the War Office’s director of military intelligence staged a grab for Bletchley’s military output. Until 1941, the Admiralty tried to continue some cryptographic work under its own roof. In Hut 3, rows erupted between representatives of the three armed services. Stewart Menzies received a constant stream of complaints from rival interests, while Bletchley staff referred to Broadway without enthusiasm as ‘the other side’. One of the most durable criticisms of ‘C’ is that he was ever eager to accept credit for the achievements of the Park, while declining to engage with its chronic resource problems, which eventually prompted the October 1941 letter to Churchill signed by Turing and his colleagues pleading for more staff, that caused Churchill to send his famous ‘Action This Day’ message: ‘Make sure they have all they want on extreme priority.’ It is a serious charge against Menzies, that he was an absentee landlord of GC&CS.

Yet all this made mercifully little impact on the work of the codebreakers. Edward Thomas, a naval officer who worked at the Park, was impressed by the absence of hierarchical distinctions: ‘Despite the high tension of much of the work … anyone of whatever rank or degree could approach anyone else, however venerable, with any idea or suggestion, however crazy.’ Few people of any rank or status felt denied a voice – an unusually rare and privileged state of affairs in the wartime institutions of any nation. From 1941, the Cambridge scientist and novelist C.P. Snow became a key Whitehall intermediary, responsible for channelling suitable mathematicians and other scholars to Bletchley. GC&CS also employed thousands of humbler folk, recruited chiefly for their language skills. Its files record details of some RAF personnel interviewed, such as Leading Aircraftsman Berry, aged twenty-three, who had started training as a pilot but re-mustered owing to his conscientious objections to dropping bombs. His German language skills were graded only ‘B’, and the recruiters noted: ‘if interested in work might do well, but needs careful handling’.

LAC Gray was also ex-aircrew, ‘grounded as result of crash’, had ‘B’ grade Spanish. Cpl Hodges, aged twenty-six, was unfit for aircrew, ‘anxious to use his German “A”, in civil life worked in architect’s office’. AC1 Tew, a twenty-eight-year-old clerk, had German ‘A’, as well as some Spanish, French and Danish, acquired while working in his father’s leather-trading business. There was much snapping between Bletchley and the Air Ministry about the latter’s reluctance to grant commissioned rank to RAF men seconded to cipher or wireless interception duties. Group-Captain Blandy of the Y Service complained that such people were ‘picked individuals having considerable linguistic qualifications and a high standard of education … [Mere Aircraftsmen] and NCOs lack the necessary authority required to carry out their duties efficiently.’

Not all the personnel posted to Bletchley proved suited to its demands. A March 1941 report on an RAF officer returned to general duties after a spell at BP noted: ‘Although an excellent linguist, he does not appear to me to have any aptitude or inclination for the research side of the work. He had been relegated to clerical tasks, but did not seem thus to justify his pay.’ There were equally bleak verdicts on the performance of some women staffers lower down the hierarchy: ‘Wren Kenwick is inaccurate, very slow and not a bit keen on her work, not very intelligent. Wrens Buchanan and Ford are unintelligent and slow and seem unable to learn. Wren Rogers suffers from mild claustrophobia and cannot work in a windowless room.’ The report concluded: ‘The remainder … are doing most excellent work,’ but the selectors were urged to recognise the importance of the jobs the women were required to fulfil, ‘and not to send us too many of the Cook and Messenger type’.

Enfolded within their oppressive security blanket, Bletchley’s people lived, loved and largely played within their own community. Almost all were paid a pittance: nineteen-year-old mathematician Mavis Lever, one of ‘Dilly’ Knox’s team, initially received thirty shillings a week, of which she paid twenty-one shillings for her lodgings. When staff did escape into the world beyond the perimeter fence, the civilian status of the young men incurred dark suspicions among the uninitiated about their absence from any battlefield. The dramas and pantomimes performed by the Park’s amateur dramatic society became high spots in the annual calendar: Frank Birch, formerly of King’s College, Cambridge and now head of Hut 4, was celebrated for his appearances as the Widow Twankey in productions of the pantomime Aladdin.

By 1942, common sense had achieved some important successes in the Park’s management. Each section worked to its appointed head, irrespective of rank or lack of it. Cryptanalysis for all Britain’s armed forces was handled entirely at Bletchley and its Indian out-stations, a concentration of effort that neither Germany nor the United States ever matched. Gordon Welchman emerged as the foremost lubricator, curbing feuding; several notoriously stupid service officers were transferred out; the popular Eric Jones was appointed to head Hut 3. It was acknowledged that the civilian codebreakers must be ridden on the lightest possible rein, though the director was prone to occasional surges of authoritarianism.

On 1 February 1942, Admiral Karl Dönitz introduced a reflector or fourth rotor into the Atlantic U-boat service’s Enigma, with immediate and calamitous results for Allied fortunes in the Battle of the Atlantic: this imposed a twenty-six-fold increase in the range of possible settings, and blinded Bletchley. Sinkings soared. At sea, the Royal Navy was obliged to rely upon ‘Huff-Duff’ to locate enemy submarines until these approached within range of underwater detection by the Asdics of convoy escorts, which were impotent against night surface attackers. Breaking what was now designated the ‘Shark’ submarine key became the Park’s foremost priority, a challenge unresolved for nine frightening months, by far the most stressful period of the war for those engaged in the task. They knew, as they sat hunched over their labours in those austere huts, that at sea men were dying every day because of their failure – though no rational person would have called it such.

Also on 1 February, coincidentally, Alastair Denniston was pushed aside into a subordinate London role, to be replaced by his deputy, Edward Travis. In some measure this development reflected a clash of personalities – Denniston and Stewart Menzies disliked each other – together with the infighting characteristic of any large bureaucracy. But it was widely felt at Bletchley that its operational head was being overwhelmed by the strains of running an establishment that since the outbreak of war had increased fourfold in size, and many times that much in its importance to the war effort. Power struggles were unavoidable. Denniston was a good and kind man who had done many things well, but Bletchley had outgrown him. Travis, whose edicts were issued in a curious trademark brown ink, was generally considered a success in his new role, not least by such influential creative figures as Welchman. When another codebreaker, Ralph Bennett, returned that summer from detached duty in the Middle East, he found that the atmosphere had changed markedly: ‘I had left as one of a group of enthusiastic amateurs. I returned to a professional organisation with standards and an acknowledged reputation to maintain. Success was no longer an occasional prize, but the natural reward of relentless attention to detail.’

Throughout 1942, Bletchley’s activities were hampered by a desperate shortage of bombes, and thus by argument about their best employment. In January the army-Luftwaffe Hut 6 was receiving 1,400 intercepts a day, of which an average of 580 were broken, a proportion that slowly increased, reaching about 50 per cent by May 1943. Often no more than one three-wheel bombe was available at any given time to work on the Shark U-boat cipher, because the others of what was still only a handful of machines were committed to breaking army and air traffic. The codebreakers said later that they would have needed ten four-wheel bombes – which did not then exist – significantly to accelerate their progress. By November, a note of desperation had entered the Admiralty’s pleas to the Park about Shark. The Battle of the Atlantic, said the navy’s Operational Intelligence Centre, was ‘the one campaign which BP are not at present influencing to any marked extent – and it is the only one in which the war can be lost unless BP do help’. A critical breakthrough was imminent, however. On 30 October in the Eastern Mediterranean U-559 was attacked by an escort group, and forced to the surface by depth-charging. Tony Fasson, thirty-year-old first lieutenant of the destroyer Petard, along with Able Seaman Colin Grazier, hastily stripped naked and swam sixty yards to the stricken submarine, then hauled themselves into the conning tower. The crew had opened the seacocks before abandoning their boat, and the sea was flooding in even as the two men searched the control room with desperate urgency.