Between Silk and Cyanide
When codes saved networks
Description
In 1942, a young man barely into his twenties walked into a nondescript building in London and was handed one of the strangest jobs of the Second World War: making sure the coded messages of British agents behind German lines could not be broken. Leo Marks was the son of an antiquarian bookseller, a boy who had grown up cracking the price codes his father used on rare volumes. He had failed the official cryptography course. He was hired anyway, and sent to the Special Operations Executive — the sabotage-and-subversion outfit Churchill had told to set Europe ablaze. What Marks found there, in his own telling, was a system that was quietly getting its own people killed.
The agents who parachuted into occupied France carried their codes in their heads. When they were caught — and in certain networks they were caught with terrible regularity — the Gestapo could torture the key out of them and then keep sending messages as if nothing had changed. London, reading traffic that looked authentic, would drop fresh agents straight into waiting arms. Marks spent the war staring at indecipherable messages, arguing with superiors, and trying to build a way of writing in secret that a captured man could not betray, because he would no longer be carrying the secret at all.
His memoir, written decades later, is a book about codes that turns out to be a book about people — about the courage of agents he coded for and often never heard from again, and about the private arithmetic of sending strangers into danger with tools he had made. It is one of the few accounts of the intelligence war told from the desk of the man who checked whether the dead were really dead.
The question we’re asking : How does a wartime code stop being a technical puzzle and become a matter of who lives and who dies?What we’ll see : An insider's account of the codes the SOE trusted, why they failed in the field, and the method that finally made a captured agent's silence unbreakable.
Table of contents
01Chapter 1 — The agents who never came back
The pattern was the thing that haunted the SOE's French sections. An agent would be trained for months in the craft of sabotage, wireless work, and coding, then flown out on a moonlit night and dropped into a field where a reception committee was supposed to be waiting. The agent would make contact, signal London that all was well, and begin transmitting. For a while the traffic would come in. Then something would shift — a message slightly off, a security check missing — or nothing would shift at all, and only later did the truth surface: the agent had been taken, sometimes within days of landing.
What made it worse was that the capture did not always stop the transmissions. The Germans ran what they called a Funkspiel, a radio game, in which they operated a captured agent's set themselves, either forcing the prisoner to send under duress or using the codes extracted from him. To London, the messages looked genuine. Requests for supplies, for money, for more agents came through in the familiar hand, and London answered them. Fresh recruits dropped into the Netherlands and into parts of France floated down into the arms of men who knew exactly when and where they were coming.
02Chapter 2 — Poems, and why they killed people
The SOE's standard method when Marks arrived was the poem code. An agent memorised a poem — sometimes a famous one, sometimes verses written for the occasion — and used a set of its words as the key to a transposition cipher. To encode a message, the agent picked several words from the poem, numbered their letters, and shuffled his text according to that grid. It was portable, needing no incriminating codebook, and it could be carried entirely in the mind. That was its appeal, and it was also its undoing.
A poem is a terrible secret because it is a shared one. Famous verses could be guessed; a codebreaker who suspected an agent was using Keats or the Lord's Prayer had a finite field to search. Worse, once the Germans had extracted an agent's poem under interrogation, they had not just that message but every message that agent would ever send, past and future, because the key never changed. A single broken man handed over his entire correspondence.
03Chapter 3 — The pad you used once and burned
The solution Marks championed was the one-time pad, and its genius lies in a single ruthless rule: each key is used exactly once and then destroyed. Instead of a memorised poem, an agent carried a pad printed with columns of random letters or numbers. To send a message he used the first unused key, enciphered his text against it, and then physically removed and burned that portion. The next message used the next key. No key was ever reused, and none had to be remembered.
The mathematics behind it is severe and beautiful. If the key is truly random, as long as the message, and never reused, the resulting cipher is unbreakable in principle — there is nothing in the encrypted text to analyse, no pattern, no repetition, no favourite poem lurking underneath. Cryptographers had understood the theory for years, but the practical obstacle had always been distributing enough random key material to the people who needed it. For an agent in the field, that meant printing the pads small enough to hide and getting them into his hands before he left.
04Chapter 4 — The quiet arithmetic of trust
What lifts Marks's account above a technical history is that he never lets the reader forget who was on the other end of the codes. Every pad he designed, every indecipherable he stayed up all night to break, was attached to a named person standing in a field in France, or in a cell, or already dead. He writes about learning to read the traffic the way a doctor reads symptoms, and about the specific dread of recognising, from a missing security check, that the friendly-sounding message in front of him was being typed by a German with a gun to someone's head. The code was never abstract. It was the last thread connecting London to a person it had sent into the dark.
This is why the one-time pad matters in his telling for more than its cleverness. A cipher, in the end, is a promise about what a system will not ask of the people inside it. The poem code, for all its convenience, made a quiet demand that no one had thought through: it asked an agent to hold in his own body a secret that could doom others, and then relied on his ability to die before revealing it. The pad withdrew that demand. It was designed so that a captured agent's silence would be unbreakable not because he was brave, but because the system had made sure he had nothing left to break.
05Conclusion
Leo Marks kept his silence for half a century before publishing the memoir, and when it appeared it read less like a spy thriller than like a long act of remembrance. The codes are all there — the poems, the indecipherables, the pads printed on silk — but the book keeps returning to the agents themselves, to the ones whose garbled messages he rescued and the ones whose transmissions he came to understand were being sent by their captors. The one-time pad was his answer to a problem that was never only technical: how to send a person into danger with a secret that could not be turned against them.