Between Silk and Cyanide (80 page)

To be fair to the Foreign Office, it has retained its sense of humour, whatever else it may have lost, and I was informed by the then curator that 'Ciphers, Signals and Sex' had been graced with a label 'to be PRESERVED AS A DOCUMENT OF HISTORICAL IMPORTANCE', but someone had crossed out 'historical' and written 'hysterical'. Having re-read it, I agree with him.

A film was made about Violette Szabo called Carve Her Name with Pride, and I allowed its producer, Daniel Angel, to use the poem in his film providing that its author's name wasn't disclosed. Thousands of letters poured in asking who'd written it and the Rank Organization professed not to know but felt they should send me a letter they'd received from the father of an eight-year-old boy. He said that his son was desperately ill, and could someone please answer the enclosed letter, which was written in code.

I managed to break his baby code, and the clear-text read: 'Dear code-master. She was very brave. Please how does the poem work. I'm going to be a spy when I grow up.'

I replied to him in his code (this was essential), saying that as soon as he was better of course I'd show him how it worked. And as soon as he was better he might like to come to the Special Forces Club and meet some of the other agents he might have read about. In the meantime I was sending him a chess-set which Violette once gave me because I knew that she'd like him to have it.

Six weeks later I received a letter from his father saying that his son had rallied for a month, and had died with the chess-set and the poem on his bed.

In 1949 the Dutch government instituted a Commission of Inquiry to establish the truth about Holland. It was particularly anxious to discover whether the disasters had been caused by a traitor in SOE or whether Dutch lives had deliberately been sacrificed as part of a British deception scheme (a theory as prevalent then as it is today). The tribunal sat for almost a year and, as a gesture of goodwill, and to avoid any suggestion of a cover-up, the Foreign Office gave its chairman a list of SOE officers, from Gubbins downwards, who'd been 'responsible' for the conduct of clandestine operations in the Netherlands, and suggested that he should invite them to have frank discussions with him.

Every one of them, including Gubbins, agreed to meet him, and their conversations took place between 3 and 10 October.

It is doubtful whether a single participant in these 'frank discussions' could have shown the chairman how to break an indecipherable, or assess the significance of a total lack of them.

No member of the Signals directorate was invited to attend, though Nick and I were available at the time.

It is possible that we'd have had something useful to contribute.

A W.T. operator's touch on the keyboard was as individual as a fingerprint. This didn't deter the enemy's radio experts from trying to simulate it, and from the summer of '43 until S.O.E. stopped passing traffic, detailed recordings were made of every operator's 'fist' before he or she left for the field.

Method. The operators were instructed to transmit every letter of the alphabet at varying speeds, followed by every numeral, but were given no warning that they were being 'fingerprinted' to avoid selfconscious transmissions. Their dots, dashes and morse hesitations were then transferred to a paper tape which moved at 16 feet per minute. This magnified even the smallest morse dots by quarter of an inch, allowing an in-depth study to be made of every operator's style. The details were recorded on square-ruled paper, and lodged with the Chief Signalmasters at the W.T. stations. When the operators reached the field, their 'fists' varied from message to message, responding to the tensions of the moment, but the basic characteristics were always present. Yet the Signalmasters needed no 'fingerprinting' charts to identify these characteristics: their morse-trained ears were attuned to every nuance of an operator's touch, and they recognised it at once. The Funk-Horchdienst (the German interception service) was equally adept. Its radio experts were able to counterfeit our operators' style to perfection if a sufficient number of messages had been intercepted.

The only real value of 'fingerprinting' was that it gave the operators confidence—which in itself was priceless. The concept of 'fingerprinting' had been brought to Nick's attention by a fair-haired WAAF Officer named Kay Cameron whose father had invented it. Nick immediately took her into S.O.E. as she was otherwise homeless. Determined that she should be based in London and posted to the stations only when necessary, he seconded her to the code-department and asked me 'to keep an eye on her'. As a contra-account (not that I needed one) he authorised me to head-hunt young Captain Appleby, a camouflage expert from the Thatched Barn who was a genius at devising new hiding places for WOKs and LOPs.

Shortage of space meant that Kay and Appleby had to share an office. They also shared an urge to help agents and each other in every way they could, and their enthusiasm for their near-impossible tasks turned their tiny office into a suite.

Those close to them sometimes wondered if they indulged in mutual fingerprinting, and though this was a pointless speculation as they were experts at camouflage, it allowed us to forget for a few moments that the average life-expectancy of a W.T. operator in France was at best six weeks.

Owing to the very simple construction of the letter one-time pad code it is most improbable that many indecipherable messages will be received. However, when these indecipherables occur, the following are the lines of attack which should be attempted.

Assume that the agent has written his message beneath his indicator group. Consequently, all groups will have to be moved one group to the left. If this does prove to be the case, in outward messages the home station should also write beneath the indicator group. There are two points to bear in mind however.

Firstly, if the agent sends a series of messages correctly, and then suddenly makes a mistake by writing his message under the indicator group, the home station should not follow suit, but assume that it is an error on the part of the agent. If the agent sends his first message with the clear text written beneath the indicator group, the home station should reciprocate for his outward traffic until further notice.

The outstation may be confused over prefixes, therefore the second, as well as the first group of the message should be eliminated. This gives two attempts, i.e. eliminating the first two groups of the message and writing the third group as it should be written underneath the first group of the pad; and secondly, eliminating the first two groups and writing the third group underneath the indicated group of the one-time pad.

Should a message commence by reading sense, and suddenly breaks off into gibberish, the first group from which the gibberish appears should be moved immediately to the right, in case:

1. a group has been omitted in wireless transmission, and
   
2. the agent has slipped a group of his one-time pad when encoding his message.
   

The Home Station should now try to assume that the agent has written the message on top of the one-time pad groups instead of beneath them. Therefore, in order to decode, they should take the column of large capital letters on the extreme left of the substitution square, and this column should be regarded as the letters of the one-time pad. The small letters running along this line should be regarded as the letters of the cipher message. Letters at the head of the column in which the little letter stands must be regarded as the en dair group. For example, assume that the one-time pad group is ZVRBI, and that the agent has encoded the word 'house'. Write the letters HOUSE above ZVRBI and the result is BIPJB.

To encode to agents who are making this error, the first pair of letters for the home station to examine will be B over Z. They will go to B in the column at the extreme left of their substitution square, and glance along until they find the little letter Z. When they have found Z, they must glance up and see in which column it stands. They will find it to stand in column H.

The next pair of letters to examine will be I over V. They will go to column I at the extreme left of their substitution square, and glance along until they find little letter V. When they have found V, they must glance up and see in which column it stands. It will be found to be in column 0.

Take P over R, J over B, B over I in the same manner, and the word HOUSE will be decoded.

It must be assumed now that the agent glances to the wrong side of his substitution square when enciphering, i.e. he will always look to the left of the capital letter instead of the right.

Assume that the agent, instead of taking the little letter as the cipher group and the large letter as the en clair, reverses the process and takes the little letter as the en clair group, and the large capital as the cipher group.

In all cases where any great difficulty is experienced and the above methods fail, the home station must concentrate on the end of the message and try to work backwards, as it must never be forgotten that an agent may be passing a message for someone else, and begin by Playfairing without giving us any warning. Caution must be taken, however, with the last group, which may consist of dead letters. The penultimate group is really the key group at this stage of the attack.

When you find the agent's indicator, assume that all the onetime pad groups he has used consist of the groups immediately beneath this indicator group, i.e. in the same column instead of running along the same parallel.

It must also be assumed that an agent may omit a line, start on the wrong line, or use the same line twice. He may also use the wrong page.

The breaking of this code will depend very greatly on individual observation, as sometimes an agent will be merely a letter out, which means a sliding along to the left of one letter instead of five. The first attempt of all will therefore be a "fanning out" in both directions, firstly group by group, secondly letter by letter. If this fails, the "fan" must consist of two letters on either side being left out, then three, then four.

In extreme cases, it must be assumed that up to ten groups of the pad may be omitted.

[1]
84 Charing Cross Road (Andre Deutsch, 1971).

[2]
Warning from the author. The curse should be used only in emergencies, and in the hands of the inexperienced has been known to backfire.

[3]
Ultimately used as a reserve code by an American wireless operator.

[4]
Security-jargon for screening.

[5]
Information known at the time only to CD, Gubbins and Wilson.

[6]
Freud completed
Moses and Monotheism
(1937-9) in London.

[7]
Trade jargon for misaligned.

[8]
A few years after the war ended (I shan't give away who won it—for those who may not know) the BBC asked Sir Colin Gubbins, as he had then become, and me, to broadcast a tribute to the FANYs on a programme called 'Now It Can Be Told'. Tom Waldron, the producer of this programme, wanted me to contribute one of the talks I'd given to the FANYs, using as nearly as possible the same words. A signals technician who'd seen an announcement of the broadcast sent me a copy of the original tape. I'm sure he meant it kindly. Not to include it in this book would justify today's equivalent of a white feather, if there is one, so out of obligation to my former colleagues and as a further tribute to all that the FANYs had to endure, it has been quoted in full.

[9]
The outcome of 'fingerprinting' is dealt with in the appendix, and is my sole justification for having one.

[10]
C's headquarters were in a street called Broadway, which to our regret was in the City of Westminster instead of New York.

[11]
Frank's concept of personal service, which was altogether different from Doris's, took on a new dimension when Helene Hanff discovered him in 1949.

[12]
So was the Foreign Office from time to time.

[13]
It's still there. Tommy. Hope you know it.

[14]
The letters ING were the start of Skinnarland's next line.

[15]
He transmitted it in December '41 and when I joined SOE in June '42 I broke it as a matter of interest.

[16]
Sunsequently included in a batch of microfilmed poems sent to Jugoslavia.

[17]
It took the Germans six months to restore the plant to even partial capacity.

[18]
Even in 1998, when so much has been written about SOE that only its secrets remain I am still credited with inventing the letter one-time pad. WOKs, yes—but with LOPs I was pre-empted, and I wish I knew by whom! Whoever you are, and wherever you may be, my apologies and thanks. L.M.

[19]
He broke his leg in a practice jump and never left for the field.

[20]
'Innocent letters' were usually sent to neutral territories for onward transmission to SOE. The only innocent thing about them was the code they used, which was a form of Playfair at its most vulnerable.