A naval operation it might have been, but Room 40 also had an atmosphere of academic informality. This was deepened with the arrival, in 1916, of the ferociously intelligent – and in some ways, simply ferocious – King’s College scholar called Alfred Dillwyn ‘Dilly’ Knox.
Knox was a classicist, but of an extremely unusual calibre: he was an expert on ancient papyri. This, ironically, provided him with the perfect flair and skill for codework. It would serve him especially well when squaring up to the challenge of Enigma.
Intriguingly, the First World War demands of Room 40 had dragged this irritable scholar away from deciphering one particularly beautiful and breathtakingly valuable papyrus found in southern Egypt: the 2,000-year-old Mimiambi of Herodas. Consisting of satiric dialogues only previously known by virtue of being mentioned in other Greek works, the discovery was wildly exciting to the academic world, much as if Aristotle’s Second Book of Poetics had been found.
Over the space of many years, Knox had travelled between Cambridge and the British Museum in Bloomsbury, there to study the intensely complicated strips of papyrus. The nature of Herodas’s dialogues was earthy, involving delinquents, brothels, slaves, sex-shops, flagellation and other such salty topics. But complications arose in deciphering such matters as where the speech breaks came, and indeed what was speech and what was not, and also in identifying errors of copying, since the papyrus may have been inscribed as a copy by an insufficiently attentive servant.
Then there was the question of how the crumbling text should be reassembled – how to ensure that the order was correct and that the pieces of the jigsaw were not out of place. This was a matter not merely of great classical learning or ability with language, but something of a cryptographical problem too. So when Knox was pulled into Room 40, the match seemed appropriate.
As soon as he arrived at work at the Admiralty in 1916, Knox bagged a room at the end of a long, untidy, undusted corridor; the room, arrestingly, also had a bath within it. This suited him extremely well; Knox was inordinately fond of hot baths.
And his department had an enormous early triumph: the decrypting of the so-called ‘Zimmerman Telegram’ – a message from the German foreign minister to the German ambassador in Mexico, urging that Mexico be encouraged into an alliance against the United States. It was this intelligence that brought America decisively into the First World War.
There was love (and laughter) in those dusty Room 40 corridors too; Alistair Denniston met his wife-to-be in the department. ‘The camaraderie of the members of Room 40,’ wrote Denniston’s son Robin, ‘all of whose names are inscribed on a silver salver which was given to Denniston and his bride on the occasion of their wedding in 1917, was borne out at the end of the war by … a pantomime, sung by all present. It was written by Frank Birch, himself one of the original cryptographers who left the secret service for the stage and for King’s College.’3
In the interwar period, the Government Code and Cypher School (as Room 40, reduced to a small number of codebreakers, was now known) had moved to Broadway Buildings and devoted itself largely to dealing with Soviet codes. Dilly Knox was especially adept in this area. Bolshevism, together with Stalin’s colossal ambition, was understood to be the most pervasive threat to the national interests. With Hitler’s seizure of power in Germany in 1933, however, those geopolitical tectonic plates shifted very rapidly.
The German navy had been using Enigma since 1926. The machine itself – the basic model of rotating letter wheels with electric contact studs, keyboard and lampboard with illuminated letters, all looking a little like a typewriter – was adapted by German electrical engineer Arthur Scherbius from an earlier, simpler design.
Enigma had been on the market since 1923, when it was used by a few commercial banks to make sure their communications were kept secret. Too few commercial banks, though: the machine was a commercial failure. Curiously, in 1926 the British government purchased one model after the machine was demonstrated at the Foreign Office. The War Office, felt, however, that it would be too ungainly for use in the field.
Once the German navy acquired the system, Enigma was completely taken off the open market, both military and commercial; the Germans then set about making a series of modifications that would make the machine’s security very much tighter. Soon afterwards, the system was adopted by the German air force, and then by the army. The British War Office had been perfectly wrong about it. The machine was brilliantly portable; thousands were manufactured.
The principle of Enigma was that the machines both enciphered messages and, at the other end, deciphered them. The operator would type a letter on the normal-looking keyboard; a couple of seconds later, via an electric current sent through the rotating code-letter wheels, another letter on the adjacent lampboard would be illuminated. This substitute letter would be noted. And so on through all the letters of a message. The enciphered version would then be radioed in Morse to its intended recipient.
The recipient, with his Enigma machine set up in exactly the same way, would tap these encoded letters in, one by one – and one by one, the real letters would be illuminated on the lampboard.
‘Although it would have been possible for one cipher clerk to carry out all the tasks of the enciphering procedure himself,’ noted codebreaker Alan Stripp, ‘this would have been a lengthy and confusing process; normally it called for a team of two.’ Even with two, it was a time-consuming business. On top of this, the machine settings would be changed every twenty-four hours.
In 1927, GC&CS took the wise precaution of studying their basic, wholly unmodified Enigma machine. Hugh Foss – eventually to become a Bletchley leading light, brilliant at Japanese decrypts – was the man assigned to the job.
John Herivel later noted, in a tone of obvious admiration, the intricate innards of this aesthetically intriguing machine: ‘the function of the [letter] wheels with their studs, pins, rings and serrated flanges, how they could be taken in and out of the scrambler, the function of the left-hand reflector drum’ and ‘the wonderfully ingenious way each of the three wheels was forced into the nearest of 26 equally spaced “allowable” positions where they were firmly held, yet not so firmly that they could not be turned by finger pressure on the flanges to any one of the 25 other allowable positions’.4
And so this neat machine of bakelite and brass was an irresistibly beguiling prospect for any mathematician or logician. Even if the machine settings were being changed daily, there surely had to be some means by which the device could be defeated?
But there was an extra difficulty. The new German version had what was known as a stecker-board, or plug board, which made the machine’s wiring vastly more complex, and allowed for literally millions more potential encoding combinations. ‘The Germans regarded the Enigma as a perfectly secure machine,’ noted Stuart Milner-Barry, who was at Bletchley in the early days of Hut 6. ‘Proof against cryptanalysts however talented and ingenious they may be.’
Right from the start, though, Hugh Foss wasn’t so sure about that. He wrote a paper saying that if enough could be gleaned about Enigma’s internal wiring, then the machine just might be broken with the use of cribs – basically using guessed words or phrases to give a starting point into all the other letters. One crucial key to the machine was that any letter – say, ‘W’ – could never be encrypted as itself. In other words, no matter how often one typed in ‘W’, the letter would never be encrypted as ‘W’. But this didn’t make the task of breaking the machine much easier to contemplate. There were still millions upon millions of potential combinations.
By 1936, when it was increasingly obvious within Whitehall that Hitler’s aggression would not be contained, GC&CS was applying renewed vigour to the Enigma problem. And in 1937, at the time of the Spanish Civil War, Dilly Knox devised a way into the earlier, unmodified version of the machine that was being used by Italy. He did this partly by means of ‘rodding’; Knox’s ‘rods’ were, in the most basic terms, a painstakingly calculated slide
-rule style representation of the wiring and rotor position of the machine upon which cipher-text letters could be moved and rearranged.
And the British were not alone in these efforts. There was also invaluable aid from another source, for an early, slightly simpler German military version of the Enigma machine had been cracked as far back as 1932 by several gifted mathematicians in Poland.
The Polish triumph was extinguished a little later in the 1930s when the German army – now regenerated and strengthening, and keen to tighten its security – increased the number of code-letter wheels on the machine from three to five, thus increasing the huge number of potential combinations another tenfold. However, this setback was in part countered thanks to a valiant Frenchman called Major (later Colonel) Gustave Bertrand, who was working in close contact with the Polish mathematicians. ‘I think we should acknowledge what the French did in the field of Enigma,’ says Mavis Batey. ‘And indeed how they really did work with us until the fall of France.’
In the early 1930s, Gustave Bertrand had been carefully monitoring Germany’s development of Enigma and the machine’s uses. Some years earlier he had made contact with a German spy (or traitor, as the Germans themselves would have put it) called Hans Thilo Schmidt. Schmidt supplied Bertrand with crucial Enigma documentation. He had come by such paperwork because he worked in the German Ministry of War.
This was not Bertrand’s only success. ‘When the Germans improved the plugboard of Enigma,’ says Keith Batey, ‘they sent out a manual. And the idiots actually gave a plaintext telling how one set up the machine and this manual gave you the answer. The Germans realised, and recalled the manual right away – but Major Bertrand got hold of one none the less. That’s what gave the Poles the entry they needed.’
The Poles devised two cipher-checking methods. One was a manual method, using ‘Zygalski sheets’, named after their inventor, mathematician Henryk Zygalski. These, in essence, were a series of twenty-six thick sheets, one for each of the Enigma’s possible sequences for the insertion of the machine’s rotors. The sheets had specially prepared grids printed upon them, twenty-six by twenty-six, letters of the alphabet on the outer edges with holes punched or cut through the squares in certain combinations. The principle was based on what were termed ‘females’ – letter positions that would be repeated in an enciphered message. The sheets would be placed on top of one another above an illuminated surface, and moved and rearranged in carefully calculated sequences until the number of lights shining through the holes was reduced to one light shining through one section or square; this in turn would reveal the particular Enigma ring-setting. As a method, it was both wildly cumbersome and impossibly time-consuming. But before the Germans made further adjustments to Enigma, it worked. The principle was later to be expanded at Bletchley by John Jeffreys.
By the summer of 1939, as the Polish nation faced certain invasion, these Enigma experts, together with a small French contingent led by Bertrand, decided to share their knowledge with the British, in the hope that they might be able to help further. Conversely, the Poles had information that the British side needed very badly indeed. On 24 July 1939, British and French cryptographers went to meet their Polish counterparts at Kabackie Woods near Pyry, a few miles south of Warsaw. Among the British members of this party was Dilly Knox. With him was Alistair Denniston.
The meeting was vital. As Knox and Denniston would have been painfully aware, they had to get a serious head start before Britain and Germany were at war. In particular, there had been an unthinking assumption among many, before 1939, that Britain still enjoyed unchallenged global naval supremacy. It would soon become clear that this was no longer the case. Moreover, the German navy was even more security conscious than the army. While the army could use cables to transmit messages, ocean-going battleships were forced to use radio signals, all of which could be picked up by others. These signals had to be encrypted with real cunning.
By 1939, Knox had run into a dead end simply because of the internal wiring of the military Enigma – an extra dimension of difficulty, distinct from the code wheels themselves. The trouble was that the Germans could have used countless different combinations of wiring on the keyboard. However, on that day outside Warsaw, the Poles told Knox that the Germans had in fact followed the most obvious, alphabetic pattern: A to A, B to B – with, as Jack Copeland explained, ‘the A-socket of the plug-board connected to the first terminal inside the entry plate, the B-socket to the second, and so on.’5 This was by no means the solution to the Enigma problem – but it did provide a valuable chink of light.
To be told, after months of worrying away at the wiring problem, that the solution was in fact the most obvious one, apparently proved a little too much for the habitually unpredictable Knox. Initially, according to Denniston himself, Knox ‘raged and raved’ when back in the car to Warsaw, shouting that ‘the whole thing was a fraud’. As Penelope Fitzgerald noted, however, for Knox ‘it was a swindle, not because he had failed to solve it, but because it was too easy. Games should be worth playing.’6
In a letter written some years later, Denniston said: ‘our position became increasingly difficult as even Bertrand, who knew no English, was aware that Knox had a grudge against the Poles who, so far as Bertrand knew, had only been successful where Knox had failed …’ According to Hut 6 veteran John Herivel, too, Knox’s temper could easily have had the most terrible knock-on effect. As he later wrote:
If Knox had continued to be in such a bloody-minded and intransigent mood, the conference would have been wound up, the French and British delegations would have returned home empty-handed, the further breaking of Enigma by the method of Zygalski sheets would never have taken place, and the Red Luftwaffe code would have remained unbroken, so that the Allied High Command would have been deprived of what Nigel de Grey termed ‘the prime source of intelligence’ for the most of the time from May 1940 until the end of the war.7
But the wild storm that Denniston seemed to recall must have passed very quickly. In a taxi on the way back from that forest rendezvous on the second day, Knox started cheerily chanting: ‘Nous avons le QWERTZU, nous marchons ensemble.’ And in a letter written at the time, he stated crisply: ‘I think we may hand some bouquets to the Poles for their lucky shot.’
Luck or skill aside, the information about the wiring was vital. Knox telephoned the information through to Peter Twinn. It is said that by the time Knox returned to Bletchley, Twinn had worked out the wheel wiring from this information alone, and had set to work on a few messages sent and intercepted the year before.
‘I was the first British cryptographer to have read a German services Enigma message,’ recalled Twinn lightly, adding, ‘I hasten to say that this did me little if any credit, since with the information Dilly had brought back from Poland, the job was little more than a routine operation.’ And, he pointed out, ‘of course, reading a few scattered messages [from] a single day in 1938 was a whole universe away from the problems that lay ahead.’8
The Poles also presented the British with a replica of the Enigma machine that they had built. ‘Dilly always said that we owed a huge amount to the Poles,’ says Mavis Batey, though she is equally adamant that the work of Colonel Bertrand should be properly celebrated. ‘Bertrand really did a good deal, with his Pimpernel pinches [the acquisition of coding information from the Germans]. Until the fall, Bertrand had his own cipher bureau in Paris and we had constant traffic and all the correspondence, and whoever got the key out that day shared it. That went on right up until the fall of France.’
Elsewhere, Alan Turing had been quietly busy upon his own researches. And by December 1939, quite independently of Knox and his new Polish friends, he managed to break into five days’ worth of Enigma material. Though this was, by itself, hugely encouraging, the messages Turing had worked upon were old – pre-war in fact. Neither he nor any other codebreaker had yet managed to crowbar their way into current German traffic.
In the deceptive quiet of thos
e early weeks and months of the Phoney War, there was still time for theorising and experimentation. ‘Some days it was actually very slack,’ says John Herivel of his workload at that time. ‘You wouldn’t get that many intercepts in at all.’ But as the season grew darker, even with all the clues and the help given by the Poles, the codebreakers knew they were facing an increasingly fearsome proposition.
It was not just the mentally exhausting prospect of facing, day after day, these groups of random-looking letters, trying to think from every conceivable angle of some logical formula that would bring order to the chaos and make the letters resolve into language. It was also the knowledge that they simply had to succeed.
The Italians provided a little succour. It was discovered that they were still using the earlier commercial version of the Enigma system, which although ingeniously complex, was known to be breakable. Dilly, with his rods and his fillies, was kept furiously busy in the Cottage, formulating means by which the Italian codes could be cracked by hand. At about the same time – the first Christmas at Bletchley Park, the ducks flapping in the freezing waters of the lake – Knox was also looking over Turing’s plans for new ‘bombe’ machines – which were to prove revolutionary in their potential.
But another obstacle for Bletchley was the fact that the different arms of the German military used subtly different versions of the Enigma system. The army Enigma machine was already fearsomely complex; the naval Enigma, as the cryptographers knew, was quite a different proposition – more complex, with extra code-wheels and more disciplined settings, using strict tables. Denniston doubted it could ever be beaten. For some of the new recruits making their way in blithe ignorance to Bletchley Park in the early months of the war, this would be the overriding priority.
6 1939–40: The Enigma Initiation
From the start, one of the disorientating elements about Bletchley for a new recruit was the ambiguity of its status. It seemed to be neither a military nor a civilian operation, but – especially in its earliest days – a curious blend of the two.
The Secret Life of Bletchley Park Page 5
