Chef de Codage

Brian Rotman

  • Between Silk and Cyanide: The Story of SOE’s Code War by Leo Marks
    HarperCollins, 614 pp, £19.99, November 1998, ISBN 0 00 255944 7

In 1940, Winston Churchill gave the fledgling Special Operations Executive its sabotage and resistance mission: Set Europe Ablaze – an encroachment on its turf not to the liking of the espionage establishment, which used its more powerful ministerial presence in the wartime Cabinet to work against SOE whenever it could. Born thus under sufferance, occupying a lowly level in the Signals hierarchy, run by an establishment-respectful general and composed of independent units – one for each occupied country except France, which had two, one for de Gaulle’s Free French and another for those loyal to Giraud – each recruiting and running its own partisans, SOE was an intelligence professional’s bad dream.

June 1942: enter young Leo Marks. Having failed to convince the cryptographic services that he was suitable for the coding unit at Bletchley, but evincing a wayward brilliance that indicated he was too smart to do routine work, Marks was assigned to keep an eye on the coding activities of SOE’s agents: a job more or less invented for the occasion. Over the next three years, he transformed the coding operation, from the training and briefing of agents to the way code keys and incoming messages were handled. He rose from cryptographic quality-checker to admired chef de codage of a vastly expanded, efficient and vital SOE unit which by the end of the war had handled some forty million messages from agents in the field. Between Silk and Cyanide is his account (or as much of it as was allowed by the Official Secrete Act) of how he did it, of how, with him leading from below, SOE fought its code war.

Codes are alluring. Most of us respond to the challenge of licensed deception and subterfuge that secret writing and cyphers, or for that matter the alphabetic trickery of crosswords and anagrams, presents. With Marks, the challenge of cyphers seems to have been an overwhelming passion, discovered in himself at the age of eight, when he broke the price code his father pencilled inside the covers of the books sold in his famous secondhand bookstore at 84 Charing Cross Road. A dozen years later, as if by pre-established harmony, the cryptographic needs of SOE gave him a perfect opportunity to exercise his passion and fight Hitler from a tobacco-fumed office in SOE’s Baker Street headquarters.

Being excluded from Bletchley probably worked to Marks’s (and the war effort’s) advantage. He seems to have been daunted by the abstract abilities of the mathematicians originally tested with him, and no doubt botched his performance accordingly. At Bletchley he would have worked on the codes used by the German Enigma machines, a mathematics-driven task compared to which SOE’s cryptographic problems were decidedly low-tech, quaint even. Nowadays, in the light of the open-key cryptography made possible by computers, even machine codes look like archaic relics of the electro-mechanical era. In any event, it’s hard to see how Marks’s coding skill and social acumen would have cut any ice with the mathematical guns at Bletchley – though they were later to see him as the ‘one who got away’. At SOE he didn’t do anything as dashing or groundbreaking as help break the Enigma code or introduce a new cypher: the double transposition cypher in use by agents was quite secure enough for what they did. Marks’s code war was more down-to-earth and urgent. His scrutiny of SOE’s coding practices pained his practical intelligence and, when he thought of the dangers it posed to agents, filled him with horror. He brooded over it; in particular, over the large number of indecipherable messages it engendered. Why so many? What was causing them? And, when they did occur, wasn’t it dangerous to instruct agents to resend them immediately?

The problem wasn’t double transposition itself (an old, widely used coding method) but the way the agents created the keys with which to encode their messages. In training, agents were required to learn a poem by heart. To send a message, they would then choose five words from it to construct a numerical key. For example, having chosen the five words THAT SWEET BONDAGE FREEDOM’S SELF, an agent would assign a number to each letter according to its alphabetical position and its place within the sentence. Thus, A in THAT gets I written under it, A in BONDAGE 2, B in BONDAGE 3, D in BONDAGE 4, D in FREEDOM 5, the first E in SWEET 6 and the second 7, the E in BONDAGE 8; and so it goes until finally the W in SWEET gets 28 written under it to obtain the complete key 25-15-1-26-22-28-6-7-27-3-19-18-4-2-14-8-12-21-9-10-5-20-17-23-24-11-16-13. The five words of the poem now become irrelevant; it is the permutation of the numbers 1 to 28 provided by the key that is used to encrypt a message by rearranging its first 28 letters, then its second 28, and so on. For example, the message SET EUROPE ABLAZE SAYS CHURCHILL (which happens to have exactly 28 letters) will be reordered by putting S in the 25th place, E in the 15th, T in the first, E in the 26th, U in the 22nd and so on, to produce TZAAH OPSSC IALEE LRLBU YUCHS EER, (I simplify: SOE used two keys and a more complicated transpositioning, but the principle is the same). If all went well, the coded message would be a perfect anagram of the original text, which the SOE de-coders – who of course knew the agent’s poem and chosen words, via an indicator group adjoined to the message – would be able to unjumble, and any German cryptoanalysts who intercepted it would not. That was the theory.

In practice the system was full of holes. Messages had to be at least two hundred letters long, which increased the transmission time so that German direction-finding vans had longer to pinpoint wireless operators; the same poem couldn’t bear a lot of traffic without its messages becoming vulnerable to decryption; the use of well-known poems made it possible to recover the rest of the poem from the decryption of a single message and thereby break the agent’s entire traffic; the indicator keys were sensitive to Morse mutilation; agents would make transcription errors when deriving their keys; they often used the same keys, sometimes with the same length messages (known as a ‘depth two’ in the trade and relatively easy to crack). In addition, like anything memorised, poems could be tortured out of agents (for whom cyanide pills were standard issue).

Marks’s solution was simple, eccentric and watertight. Scrap the poems and give agents a list of pairs of already worked-out keys (WOKs), printed on silk squares flimsy enough to avoid detection in random searches and still legible after many foldings and unfoldings. A pair of keys would be used for one message only: after use, the agent would cut them out of the silk and burn them. That way, even if captured, the silk could give no clues to deciphering an agent’s past traffic.

Much of Marks’s story describes in vivid and mock-heroic detail his battle to institute WOKs in the face of blocking from higher up, interdepartmental obstruction, bureaucratic inertia and a refusal to recognise the dangers to which the poem system exposed agents. Aside from convincing people that there was more to silk than knickers, he had to overcome the intransigence of the head of SOE (whose casual anti-semitism he rebutted with wit); educate everybody about the shortcomings of the poems and the dangers they posed, and about the risk of indecipherability; secure promises of photographic services and supplies of silk from hard-nosed industrialists who had their time cut out making sure there was enough of the stuff for parachutes; and generally charm and cheat his way through the labyrinth of security services and wartime production inside and outside Baker Street.

That he was able to accomplish this testifies to an uncommon mix of qualities. Only a generation away from the Whitechapel ghetto (his bookshop-owning father started life as a barrow boy), Marks had been thoroughly gentrified by attending prep school and then St Paul’s. Quick wits and Jewish survival skills were allied to a charm that could talk the birds off the trees. He was, on his own account, a small, cheekily brilliant loner; an ill-disciplined Jew-with-an-education who was ‘a bit of a cunt’, a description his headmaster at St Paul’s made sure found its way (suitably coded in Latin) into his report. He was also a doted-on only child whose mother made him smoked salmon sandwiches and supplied him with black-market coffee and other goodies. He used these to ease his various dealings with others (not to mention to confirm a racial sterotype which, one suspects, he turned to his advantage whenever possible).

It took months of campaigning and missionary work before the idea of WOKs was accepted and many more before they were used. In the meantime, Marks did his best with the poem system. He made up poems for the agents to learn rather than letting them use ones the Germans might know. He found ways to boost the morale and improve the success rate of the army of young (wonderfully brazen-tongued) women whose mind-numbing job it was to crack the indecipherables: a process that entailed trying thousands of possible keys (five thousand was not uncommon, often many more) and hunting for anagrams. At the same time, he was busy poking about in the corners and picking at the edges of SOE’s coding activities. He quickly figured out, for example, that the special code de Gaulle had insisted his Free French use was the standard SOE code with the indicator groups disguised. He then secretly intercepted the French traffic and eliminated all its indecipherables to protect the agents from the dangers that resending them posed. This operation and his subsequent spilling of it to an initially angry but ultimately admiring crowd of French agents is one of Marks’s numerous tales of mischief and triumph.

Between Silk and Cyanide is an almost totally enclosed, 600-page series of encounters and incidents in unnamed, unlocated, undescribed rooms, punctuated by a small number of out-of-school trips. One such has Marks going to Cairo as SOE’s representative. Outside the entrance of his hotel, his trousers fall down, exposing his ‘kosher rump’ to the amusement of all and sundry; one of the sundry turns out to be his favourite kosher idol, the comedian Jack Benny, who grants him an hour of his presence. Another, more intimate trip has Marks going late at night to 84 Charing Cross Road in order to get inspiration from sitting in the chair Freud had sat in a few years before when he was writing Moses and Monotheism and had had all the books on Moses brought to him. But these trips out of the office are few and carry little sense of everyday life outside – of air raids, bombing, sirens, blackouts, rationing, last-minute marriages, evacuation, or any other aspect of wartime London. And, in the body of the text at least, there is no reference to the future, no indication that the events we’re reading about have any reality other than their blow-by-blow unfolding. One emerges at the end of a three-year tunnel, blinking in the grey light of peace. The significant exception to Marks’s careful exclusion of the future is a photograph taken in 1995, in which he and another, older man are sitting together talking. On the opposite page is a cutting from the Daily Telegraph, with the headline: ‘The code master flies out to clear Ebenezer.’

Ebenezer was the code name of an agent working for the SOE in Holland, and we’re introduced to him with a sense of foreboding when Marks’s intuition starts sending him an indecipherable message: ‘I had a gut feeling about the Dutch but it was a bride unwilling to be carried over the threshold of consciousness.’ What he was reacting to was a major (and unusual) snarl-up in the traffic between Holland and London in the second half of 1942: irregularities in Ebenezer and his fellow agents’ security checks back to March of that year (three months before Marks joined SOE) had been followed by a six-month uninterrupted stretch of messages with the unheard-of record of not a single indecipherable. Marks became convinced that Ebenezer had been turned by the Gestapo and the entire Dutch operation was blown, but he had no way of proving it. The Dutch section was, he felt, criminally complacent but the SOE set-up forbade him from officially criticising them, let alone interfering in their affairs. Much later it was decided at a level beyond his reach that the Dutch operation was indeed compromised; from then on, only suitably doctored messages were transmitted to Holland. The Germans, seeing the game was up, added a final insult to the injury, sending a message in plaintext regretting the cessation of their business with London.

In The Codebreakers, a massive account of codes and cyphers that dominates the field, the American historian David Kahn devotes only a few pages to SOE, most of which are taken up with its notorious mistake. He talks of its ‘stupidity and incompetence’ and, with a hint of schadenfreude, describes the Dutch affair as ‘the worst Allied defeat in the espionage war’, detailing exactly how the Ordnungspolizei captured Ebenezer, how he was turned by the Abwehr, who, hardly believing their good fortune, then ran the whole Dutch SOE radio network for the longest period of deception (Funkspiel, as it is playfully called in German) ever. The cost in agents’ lives and matériel, though substantial, was perhaps small compared to the military consequences of there being no Dutch resistance at work when the Allied forces liberated Northern Europe. Evidently, Ebenezer had tried to warn SOE from the moment he was caught by repeatedly sending a transparently false security check, but, given the situation within the Dutch section, it was to no avail. When Marks travelled to Holland fifty years later it was to appear on TV and, by giving the SOE version of events, to remove the slur of collaboration from Ebenezer’s name. If his book is similarly intended to give the SOE and its many agents a memorial more complex and sympathetic than the sort of withering dismissal accorded it by historians such as Kahn, it does so very successfully.