by Max Hastings
Piece by piece, though painfully slowly, the codebreakers at Bletchley groped towards a solution to a riddle even more dense than that of Enigma, partly because they lacked a physical example of the transmitting machine. In the early months of 1942, by sheer intellectual endeavour the Park’s research section created a theoretical model of the Lorenz SZ40/42. Most credit for this went to a chemistry student turned mathematician named Bill Tutte, who deserves to be almost as well-known as Turing and Welchman. He was born in 1917, the son of a gardener and a cook-housekeeper at a Newmarket racing stable. He won a scholarship to the Cambridge and County Day School, then progressed to Trinity College. In October 1941 he was assigned to study the Tunny traffic, and spent the months that followed performing the extraordinary cerebral feat of deducing what kind of machine the Germans must be employing to produce the noises recorded by the interceptors. Tutte established that the teleprinter would have two sets of five wheels, one of these ‘stepping’ irregularly, with 501 settable pins and a further two motor wheels, between them creating a range of combinations much greater than that produced by Enigma. His astounding feat, a triumph of intellectual effort unassisted by technology, caused his senior colleagues to support his later successful application for a prize fellowship at Trinity, based entirely on his doings at Bletchley, though their nature was not disclosed to the college. Nigel de Grey hailed his contribution as ‘one of the outstanding successes of the war’, and so it was.
Establishing the character of the machine was an important beginning, but one that brought the British only a little closer to reading its traffic. Edward Travis, BP’s chief, observed that the German teleprinter’s output was ‘as analogous to the other machine ciphers as a Maori and an Eskimo’. In May 1942, Tiltman acknowledged that ‘the Geheimschreiber [Siemens teleprinter] is a great worry to us’. A young Oxford mathematician, Michael Crum, modelled the T-52, and his findings led the codebreakers to conclude that it presented too great a challenge to be pursued. Instead, they must concentrate all possible resources on the Lorenz – and quickly. The more the Germans used their fast-expanding WANDA-Netz Continental teleprinter system for top secret communications, the less they would use Enigma to encrypt them. It is remarkable that Berlin’s enthusiasm for the teleprinter was so great, because its vulnerability had already been exposed – back in 1940, by the Swedish codebreaker Arne Beurling, who tapped into the Stockholm exchange link connecting Berlin to its forces in Norway, and broke T-52 messages by a methodology never disclosed. The Swedish company Ericsson built a machine it called ‘the App’, to assist Beurling’s operations, and he read considerable traffic until May 1943, when the Germans, warned by the Finns of what was happening, introduced new keys and security measures. Berlin did not, however, question the integrity of the whole system, and the British knew nothing of Beurling’s activities.
Between July and October 1942, by endeavours which owed little to mechanical assistance, a group working within Major Ralph Tester’s new section, ‘the Testery’, read some Lorenz SZ40 traffic, using a higher mathematical method known at Bletchley as ‘Turingery’, after its inventor. Those responsible included eighteen-year-old Donald Michie, who later taught the Baudot code, through which messages were transmitted after Vernam encryption, to the future statesman Roy Jenkins; Peter Hilton, a twenty-one-year-old Oxford mathematician; and Peter Berenson, who much later founded Amnesty International. Increasingly disciplined German procedures made the flow of decrypts irregular: different keys were allocated names of fish and marine creatures – ‘Bream’, ‘Grilse’, ‘Octopus’ and so on; ‘Jellyfish’ later proved to include some of the most momentous German high command messages.
Human brainpower remained the main engine of the teleprinter codebreakers until the middle of 1943. In June the Park broke 114 Lorenz signals out of 575 dispatched to Berlin by the German high command in Italy. Bletchley reported in August: ‘the quality of the intelligence derived from Fish is of the highest order’. Though Lorenz traffic was never broken in anything like the same volume as Enigma, it was of exceptional importance, because it addressed the enemy’s most sensitive exchanges. Moreover, Bletchley’s difficulties and delays in breaking German army Enigma persisted until the end of the war, and Tunny offered a priceless alternative route into military traffic.
The transformative development for the codebreakers came, inevitably, from the enlistment of machines. These were even more innovative than Turing’s bombes, and were created by other minds and hands. Max Newman was born in 1897. His father was a German named Neumann, and like the Saxe-Coburgs his son changed his name during the First World War, in which he served briefly and reluctantly as a British Army paymaster. Between the wars Newman gained a formidable reputation as a mathematician at Cambridge, where he came to know Turing. Professor Pat Blackett drew him to Bletchley’s attention, describing him as a fine chess-player and musician. Newman was initially unwilling to join the Park, because he feared that the work would be insufficiently interesting. When he grudgingly accepted an appointment at the end of 1942, he did so on condition that he retained an option to leave after a year if he became unhappy. Few men, however distinguished, dared to make such a stipulation in the midst of a world war – and even fewer found it accepted.
Newman’s first months at Bletchley proved so frustrating that it looked as if he would indeed quit; he was not a success as a codebreaker. But he initiated a critical breakthrough by studying Tutte’s analysis of the teleprinter’s workings, and urged that a machine could and should be constructed to test the 1.6 x 1,019 possible start positions for its wheel settings. Alan Turing, newly returned from a long trip to the US, was now exploring the science of electronic circuitry, as was Charles Wynn-Williams, a circuit specialist transferred to Bletchley from radar research at Malvern. Turing urged Newman to discuss his project with Tommy Flowers, a senior engineer at the Post Office’s Dollis Hill research station in north-west London, who had played a modest role in the creation of the bombes.
Newman was a gifted organiser of considerable diplomatic as well as intellectual skills. An American who served at Bletchley, Sgt George Vergine, described him as ‘a marvellous fellow’, always open to new ideas: ‘We used to have tea parties which were mathematical discussions of problems, developments, techniques … a topic would be written on the blackboard and all of the analysts, including Newman, would come tea in hand and chew it around, and see whether it would be useful in cracking codes.’ He assumed direction of a new Bletchley section, dubbed ‘the Newmanry’, charged with identifying more advanced mechanical and electronic aids to codebreaking. He could claim credit for recognising the practicability of a machine to assist in breaking Fish messages, and for securing approval and resources for the first relatively primitive such device to be built. This version, dubbed ‘the Robinson’, was inspired by the design of Wynn-Williams, and built at Dollis Hill under the direction of engineer Francis Morell. The collaboration of Newman and Wynn-Williams – assisted by others of almost equal gifts such as mathematician Jack Good, who had worked with Turing on the bombes – produced a succession of technological marvels which outclassed any other codebreaking aid created on either side of the Atlantic in the course of the war. The first Robinson was delivered to Bletchley in June 1943, followed by a dozen stablemates by the end of the year, and more thereafter. The Robinson operated as a super-fast bombe, attacking the output of the German teleprinters by exploring punched tapes photo-electrically at the then fantastic speed of a thousand characters a second. It enabled the Park to read some Lorenz messages in the autumn of 1943, and hundreds by the following spring. Its limitation was mechanical – the difficulty of synchronising two tapes which had to run simultaneously; preventing breaks; dealing with repeated valve failures.
Tommy Flowers was impatient with the Robinson and its weaknesses. This senior telephone engineer nursed a far more ambitious, all-electronic vision. He was a builder’s son from the East End of London, born in 1905, who won a scholarship
to a technical college where he displayed a precocious talent for mechanics and science. Having joined the Post Office, he spent a decade working on the evolution of automated telephone systems. For much of the war, though he held a title as head of the telephone switching department, he played a leading role in manufacturing technology for Bletchley. But while he forged a close working relationship with Turing, who often visited Dollis Hill, the formidable and influential Gordon Welchman took against the engineer, who was certainly no ‘gentleman’ in the parlance of those days. Welchman treated him with disdain, as a mere artisan with ideas above his station. Flowers is nonetheless considered to have made a brilliant contribution to realising and improving upon the concepts of Newman and Wynn-Williams, by creating the new wonder of ‘Colossus’, which may be considered the first computer in the world.
It is a measure of Bletchley’s difficulties with the German teleprinters that while the output of ‘Fish’ messages from the enemy’s high command doubled in the course of 1943, the volume of decrypts fell, from 330 in January to 244 in December, albeit many of high value to Allied intelligence officers. Flowers grasped the improvement in reliability that could be achieved if hard valves were used in place of gas-filled ones, and were never switched off. He initiated production of the first Colossus without an explicit directive from the Bletchley authorities, who urged focusing on delivery of Robinsons, which required only a hundred relatively scarce valves, while Colossus used 1,500, and a more advanced 1945 version 2,500. Max Newman was always supportive of Flowers, but others were less so. The engineer, in his own field an obsessive like so many of those involved with Bletchley, was obliged to use some of his own money to purchase scarce components. Within ten months Dollis Hill, using fifteen of its own engineers and forty technicians at a Post Office factory in Birmingham, had brought into being a huge machine which processed data at five times the speed of the Robinson. It was first tested on 25 November 1943, and entered service at Bletchley in January 1944. Flowers after the war received an ex-gratia payment of £1,000, together with an MBE, a shamefully condescending recognition of his role as begetter of Colossus. Most Bletchley hands testify that he was the practical brain who played a pivotal role in translating the concepts of Turing, Newman and Wynn-Williams into reality, and indeed for advancing them to a new level of sophistication. The most advanced codebreaking technology of the war was devised at Bletchley, though when the Americans built their own variants they often improved upon the originals, as with the US Navy’s bombes. Arlington Hall’s ‘autoscritchers’ performed some of the same functions as Colossus, though they operated against Enigma traffic, and each cost as much as a fighter.
An unnamed Bletchley staffer penned a description of his own fascination with the spectacle of Colossus at work, a machine of a complexity and energy such as no previous generation had ever seen: ‘the fantastic speed of thin paper tape round the glittering pulleys; the childish pleasure of not-not, span, print main headings and other gadgets; the wizardry of purely mechanical decoding letter by letter (one novice thought she was being hoaxed); the uncanny action of the typewriter in printing scores without and beyond human aid … periods of eager expectation culminating in the sudden appearance of the longed-for score … the frantic chatter of a motor run, even the ludicrous frenzy of hosts of bogus scores’.
By 1945 six hundred staff were working around the clock at the Knockholt interception station to record Germany’s teleprinter traffic. Bletchley Park expanded from 3,800 personnel in 1943 to 5,600 in 1944, then to 9,000 in 1945. Though Enigma remained by far the most productive source of intelligence by volume, the teleprinter decrypts conveyed the most important messages. The Newmanry alone employed twenty-six cryptanalysts, twenty-eight engineers and 273 Wrens to service ten Colossi, three Robinsons and scores of lesser machines. Dollis Hill was building one new Colossus a month. A total of 476 Tunny teleprinter decrypts were recorded in May 1944, 339 in July, 404 in August, then a flood in the last months of the war, for a grand total of 13,500 Tunny decrypts out of 168,000 German transmissions intercepted since November 1942. The cryptographic historian Ralph Erskine has described the breaking of the teleprinter traffic as ‘the greatest code-breaking feat of the war … finding Tunny’s wheel patterns and settings required the highest cryptanalytic skills and involved advanced statistical techniques and some of the most complex electronic equipment of the war’.
Although it is right to marvel at Bletchley’s achievement, it is also essential to recognise its limitations, even in the last eighteen months of the war. In February 1944, just 17 per cent of the German army’s traffic was being broken. Around half the US Navy’s high-powered bombes at Mount Vernon were handling Kriegsmarine traffic for Hut 6, because Bletchley’s resources could not alone bear the strain, and sometimes American aid was also needed to address German army-Luftwaffe material. A significant portion of Enigma traffic was being read within ten to twenty hours, but teleprinted Fish messages often took a week to read. Each Robinson was able to process an average of one signal a day, Colossus fifteen. The Lorenz Tunny was the goldmine in quality terms, providing a stream of intelligence about German high command thinking, most importantly in advance of D-Day.
Consumers at Allied military headquarters received all such material without distinction as ‘Ultra’. Only those serving at the Park were aware of the infinite variations and nuances in its cryptographic operations. Even within the Enigma traffic, there was never enough bombe capacity to attempt the breaking of all the enemy’s messages, and thus daily choices had to be made about allocation of resources. In the last year of the war American and British air intelligence officers grew progressively less interested in Luftwaffe decrypts, because what the enemy was doing had little influence on their operational decisions. At sea, Ultra had played its critical part in securing Allied dominance both in the Pacific and in the West, and by 1944 this had been achieved. Bletchley once again lost much of the U-boat traffic in November that year, when the Germans introduced ‘one boat’ ciphers – unique encryption of a message for a single recipient. Moreover, a significant number of Kriegsmarine Enigma keys, such as ‘Pike’, were never broken.
By 1944–45, however, the Allies were so strong, and Dönitz’s force so weak, that this no longer mattered. Moreover, the US Navy’s Op-20-G had assumed the lead role in handling U-boat intercepts, because of its greater resources. On land, Ultra’s contribution was far more often strategic than tactical, because so much material reached commanders out of real time: Ralph Bennett of Hut 3 wrote of ‘the frequently recalcitrant army keys’. Even more fundamental, while knowledge of the enemy’s strength and deployments was immensely reassuring to commanders, it provided no assured passport to victory. From the late summer of 1943 onwards, for instance, Ultra thoroughly informed the Allies about the opinions and intentions of Kesselring, commanding Hitler’s considerably inferior forces in Italy. But it told them nothing that enabled British and American troops to defeat Kesselring’s army until the very last weeks of the war.
4 INFERNAL MACHINES
Intelligence about enemy weapons systems often yielded more practical value to the Allied war effort than insights into Hitler’s thinking. The five-year grapple in the sky between the Allies and the Luftwaffe – the challenge being to overcome Germany’s invisible electronic defences as well as its fighters – persisted until 1945, and prompted endeavours almost as dramatic as the Bruneval raid. A thirty-year-old Belgian doctor named André Mathe of ‘Service Marc’ was one of several Resistance workers who took extraordinary risks to explore local German night-fighter direction stations. In the summer of 1942 he dispatched a moving message to London pleading for more effective support, and better guidance about what the British would like to know. Mathe – whose real identity was then unknown to MI6 – suggested that the seriousness with which the Germans guarded certain installations in his area implied their importance. He and his comrades had several times been fired upon by sentries while reconnoitring them,
‘fortunately with more zeal than accuracy … As far as our work is concerned, it would be helpful if we knew to what extent you and the British services are interested. We have been working so long in the dark that any reaction from London about our work would be welcome to such obscure workers as ourselves. We hope this will not be resented since, whatever may happen you can rely on our entire devotion and on the sacrifice of our lives.’ One of Mathe’s network, a local jeweller named Willi Badart, made detailed sketches of a Seeburg bomber-plotting table, after bribing a Belgian SS guard to give him access to the fighter direction tower while its Luftwaffe controllers were off-duty, which gave useful aid to Reg Jones and his colleagues at the Air Ministry. Mathe proved that there was nothing vacuously histrionic about his message pointing out the risks of such men’s efforts for the Allied cause: he himself was arrested on 31 March 1943 and executed in the following year.
Some such intelligence could only be assembled by men and women who could physically examine relevant German equipment. Other information, however, had to be gathered by probing the enemy’s defences in the air, at mortal peril. One of the most heroic flying missions of the war took place on the night of 2–3 December 1943. The British established the key role of the Germans’ Lichtenstein airborne radar in guiding their night-fighters. For some time thereafter, efforts were made to provoke an encounter with a night-fighter by sending lone Ferret aircraft to roam the night skies over France, Belgium and Holland – comparatively close to home for the RAF – but the enemy ignored these. It became plain that data could only be obtained by dispatching an investigatory mission deep into enemy airspace. Thus, the decision was made to send a twin-engined Wellington of Bomber Command’s 1473 Wireless Investigation Flight to join the ‘stream’ on a night raid over Germany.