Misread Signals E-Book

Dermot Turing

• Herausgeber: The History Press
• Kategorie: Geisteswissenschaft
• Sprache: Englisch

Seitenzahl: 461

Veröffentlichungsjahr: 2025

Beschreibung

'It is inspirational.' – Helen Fry, author of Women in Intelligence 'An important and exciting contribution to the history.' – Clare Mulley, author of Agent Zo Bletchley Park is remembered as a land of male intellectuals who were supported by a staff of women in menial roles, with figures such as Alan Turing, William Tutte and John Tiltman taking centre stage. These are the men who worked on the fearsome Enigma and Lorenz ciphering systems – the men who helped sway the course of the war in the Allies' favour. But, as is often the case in the historical record, this is not the whole story. Women were not just secretaries and assistants: they had serious full-on codebreaking roles. And this was not just at Bletchley, or in the UK, or even only in the Second World War. These were women like Margaret Rock, who solved Enigma and other machine problems; Agnes Driscoll, the first US Navy codebreaker; and Asta Friedrichs, who postwar became a prime source for information on German Foreign Office codebreaking. Yet, when the histories were written, these women – and many more besides – somehow got left out. Who were they? What did they achieve? How did they 'vanish'? In Misread Signals, expert codebreaking historian Dermot Turing turns his attention to these long-ignored women and puts their contributions back in the spotlight where they belong.

Leseprobe

All due diligence has been made to trace original copyright owners and to give appropriate credit. If any error or omission has been made, please contact the author via the publishers.

Cover illustrations: Front: Japanese Army Section at Arlington Hall. (NSA/National Cryptologic Museum) Back, top: Madge Dale; middle: Mavis Lever; bottom: Elizebeth Friedman. (NSA/National Cryptologic Museum) Author photograph: © KT Bruce.

First published 2025

The History Press

97 St George’s Place, Cheltenham,

Gloucestershire, Gl50 3QB

www.thehistorypress.co.uk

© Dermot Turing, 2025

The right of Dermot Turing to be identified as the Author of this work has been asserted in accordance with the Copyright, Designs and Patents Act 1988.

All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without the permission in writing from the Publishers.

British Library Cataloguing in Publication Data.

A catalogue record for this book is available from the British Library.

ISBN 978 1 80399 794 0

Typesetting and origination by The History Press

Printed and bound in Great Britain by TJ Books, Padstow, Cornwall

The History Press proudly supports

www.treesforlife.org.uk

EU Authorised Representative: Easy Access System Europe

Mustamäe tee 50, 10621 Tallinn, Estonia

[email protected]

eBook converted by Geethik Technologies

CONTENTS

Introduction

Prologue: Codes and Ciphers Orientation

Glossary

Who’s Who

Part I Unknown

1   ‘Facing the prospect of women’

2   ‘Not quite so intelligent’

3   ‘A typist’

4   ‘Approximately equal’

5   ‘Difficult woman’

6   ‘Worst conceivable security risk’

Part II Unfound

7   ‘To give up this work’

8   ‘Stumbling upon what I do’

9   ‘Written off’

10  ‘Not Enigma’

11  ‘One of the only’

Notes

Further Reading

Acknowledgements

INTRODUCTION

The British people are very proud of Bletchley Park. It’s where a war-winning weapon was created, relying on brains rather than brawn. The balance of power was tipped because the codes broken there gave hard-pressed generals and admirals an extra edge, a bonus card in a fight where every advantage was needed. What is more, the most secure German ciphers, created by machines, ought to have defeated the cryptanalysts, but somehow, through genius and new technology, the most complex secure communications systems were brought to their knees.

It’s not just what was done there that makes us proud of Bletchley Park. The fact that continues to amaze is that we did not know about it until thirty years after it closed its gates. Not a whisper was breathed by any of its 10,000 employees. The revelation that codes had been broken – that the intelligence history of the Second World War needed to be rewritten from scratch – was explosive, but the effectiveness of the secrecy wall is astonishing. And those 10,000 employees grab our attention, too: so many of them, 75 per cent of them indeed, were women. Bletchley Park was a place of diversity as well as intellect. For many reasons, these included, what happened there in the Second World War seems to have anticipated the present century, our concerns and our values.

Probe a little, beneath the veneer of national pride, and the picture of Bletchley as a paragon of diversity begins to crumble. The women of Bletchley Park are celebrated as bombe machine operators, punched-card feeders, decoders using Typex machines to convert Enigma messages into plain German, indexers and drivers and messengers. By contrast, the ‘famous names’ of Bletchley Park: Alan Turing, Dilly Knox, Gordon Welchman, Tommy Flowers, John Tiltman … all seem to be men. The men, it seems, did the thing we are most proud of, winning the war by force of intellect. The women were there to help.

Go back to the earliest literature about Bletchley Park, and the gender divide seems to be written in from the outset. Take Peter Calvocoressi, who as a young RAF officer spent his war there, and as a retired barrister living not far from Bletchley, wrote one of the first ‘I was there’ accounts of the place when the secrecy restrictions were eased in the late 1970s.

There were at BP Chiefs and Indians. The Chiefs in both areas – cryptography and intelligence – were distinguished from the Indians because they were fewer and preponderantly male and had the better jobs – better because they were more responsible and closer to the brush of real events.1

It rings true; it’s what we expect from the Britain of the 1940s. And when we see the numbers – 5,623 women known to have served at Bletchley or its outstations in the women’s branches of the armed services, 6,640 women2 at Bletchley and its outstations in the last week of the war in Europe (75 per cent of the workforce at that point) – it reinforces the same picture.3 So do the memoirs of those who served. So does the business of the Women’s Committee.

Women’s Committee

… This Committee exists for the promotion of the well-being of all women working at the War Station. It does not deal with questions concerning work or pay. The Chairman of the Committee will be glad to receive for the consideration of the Committee suggestions from any women serving at the War Station. Any civilian woman who is in difficulty of any kind should ask to see Miss Wickham, Hut 9 … Service women should refer to their own officers.4

The coded message in Commander Travis’s Serial Order No. 118 meant that unmarried women who became pregnant had somewhere to turn. So the business of the Committee could focus on hair-washing facilities, thefts, bathrooms and blood donations. Women, being a species apart, had different concerns from the ‘chiefs’. If Bletchley Park was an intelligence factory, the blue collars were worn by women.

That much is incontrovertible. But something is missing. Bletchley Park was a place for codebreakers. Having identified the ‘chiefs’, those who managed the huge organisation, and the ‘indians’, where did the codebreakers fit in? Who were they, and were they men or women? It is very tempting to take the narrative of women in support roles at Bletchley Park, and conclude that the codebreakers were men. The roll call of famous names implies that conclusion is right. So does the archival material, which begins the Second World War with a list of men, and a letter by Commander Alastair Denniston, the head of the Government Code & Cypher School, that says: ‘For some days now we have been obliged to recruit from our emergency list men of the Professor type …’5

Historians have not resisted the temptation. It is assumed that women were not recruited into codebreaking roles; rather, the women who were recruited – even women with degrees – were subordinated to the codebreaker professors, doing translations and working in intelligence analysis. Vital roles, but not codebreaking.

It makes sense, because it fits with post-war experience. As John Ferris, the official historian of GCHQ, explains in his massive history of the place:

In 1970, across GCHQ, all divisional heads were men and few women were branch or section heads. In administrative divisions, women headed all sections dominated by females, such as typing and libraries … Women had more status in intelligence-producing organisations, however. In divisions H, J, K and V, men and women attacked problems as teams, side by side, in different ways. J and K stemmed straight from GC&CS, though their members analysed intelligence much more than their predecessors had done … Women were marginalised in mathematics, and thus also among the cryppies of H division …6

If women were sidelined from technical cryptanalytical and management roles in the 1970s, how much more likely it would be for that to have been the case in the darker days of the Second World War?

Of course, there are one or two exceptions. Among the famous names is Joan Clarke. She is famous not because she was deputy head of Hut 8, where the codebreakers of Bletchley Park carried out their attack on the German Navy’s Enigma-encrypted signals, the signals that directed U-boats to their prey in the North Atlantic. She is famous because she was briefly the fiancée of Alan Turing, a fact known to the public through a Hollywood movie. But having discovered Joan Clarke, we have stumbled upon a rarity: a woman codebreaker.

According to an extended obituary of Joan Clarke in the Annals of the Institute of Electrical and Electronics Engineers:

She was congenial but shy, kind, gentle, truthful, nonaggressive, agreeable to all, and always subordinate to the men in her life, except in Hut 8, where she was treated as an equal …7

Evidently – at least, it was evident to the obituarists, writing in 2001 – codebreaking at Bletchley Park was the province of men, to whom any women were self-evidently subordinate. Even as deputy head of Hut 8, Joan Clarke only claws her reputation up from subordinate to ‘an equal’. The wonderful Roll of Honour, compiled by the staff of the current Bletchley Park Trust about the people who served there in the Second World War, lists no fewer than 171 people who worked in Hut 8 at various stages; thirty-four were men, not all of whom can have been Miss Clarke’s superiors or equals. Something doesn’t add up.8

And then there is the matter of the awkward photographic evidence. GCHQ has a small collection of photographs taken at Bletchley in the second half of the war, when most of the staff had moved into the ugly brick block buildings specially built for them (ugly, but better than the wooden huts they moved out of). The photos show women. Sure, some of them are minding bombe machines, the invention that helped find the daily ‘key’ used by the German military to change the set-up of their Enigma machines. Some are operating Typex machines, the British cipher machine rigged to work like an Enigma: these women are decoding intercepted signals, which they can do because the cryptanalysts have found the daily key. But a good proportion of the photos show women doing desk work. In some shots an Enigma machine is prominent. The places where these photos were taken are the codebreaking rooms of Bletchley Park. There is hardly a man in sight. In these photos, the codebreaking was being done by women. The same is true in the much more extensive library of photos of the US Army’s Second World War codebreaking centre at Arlington Hall in Washington. These show roomfuls of women breaking codes – not typing or filing or minding machinery – with a scattering of men interspersed. On VE Day, there were 10,500 staff at Arlington Hall, around half of whom were women.9

Perhaps, we might wonder if we have been misled. Perhaps the standard narrative, of boffins breaking codes and Wrens (Women’s Royal Naval Service) minding machines, was too glib. Perhaps there is evidence, beyond the photos and the special case of Joan Clarke, that there were women codebreakers; documents now available to researchers that might cast new light.

We might expect that America would find it easier to embrace women into a technical workforce than Britain; if so, our expectation would also be that Nazi Germany would have excluded women altogether. In 1945, American soldiers captured a huge volume of archival material belonging to the Reich’s various codebreaking services. Among those papers is the telephone list of the Chiffrierabteilung of the Oberkommando der Wehrmacht – the cipher division of the Armed Forces Supreme Command.10 There is not a single woman’s name on the list. And yet, in the German Foreign Office, at the same time, a different team of codebreakers included numerous women. Not everything is exactly as we expect it to be.

Finding the German women codebreakers presents a different challenge from the Anglo-American search. In Britain and America, the women’s names are there in the record, even if their roles and careers need to be teased out. In Germany, it is different. For sure, there is an archival record, and there is much less fuss about declassification of cryptological material than one finds with Germany’s former enemies. But the lode to be mined is thin. By contrast with the hundreds of women in codebreaking roles at Bletchley Park, and at America’s two huge codebreaking centres at Arlington Hall (US Army) and Nebraska Avenue (US Navy) during the Second World War, women doing codebreaking in Berlin were far less numerous.

Codebreaking in Germany was dispersed across the armed forces, which employed men, mostly, while smaller codebreaking groups served the Foreign Office and a personal intelligence service set up by Hermann Göring for his own private machinations. Only the Foreign Office group employed women as codebreakers, but in a gender mix that would look good by today’s critical standards. To stumble upon these women without making further enquiry would have been wrong; and to omit them from the book because the information available is less voluminous a worse error. It is a shame that the type of material that has survived allows us fewer insights into the German women codebreakers, but they provide a perspective from a different cultural context on themes common to both sides of the mid-century conflict.

What is common, and central to this book, is an apparent contradiction, or at least an anomaly to be explored. It is the paradox of the women codebreakers, present in photographs, but missing from history. How is that so? Perhaps women couldn’t be hired as codebreakers because the profession was a male one, so women like Joan Clarke and the women of the German Foreign Office were indeed exceptions. Perhaps women didn’t qualify as codebreakers because the skills and training required for the role were in subjects that girls’ and women’s education of a previous age did not cover. That would fit with our present-day picture of under-representation of women in computer science and girls not taking advanced courses in mathematics; codebreaking, after all, is presumably a scientific discipline.

If women did break into this secret world, was that exceptional? Did they last, or were the tables weighted against them, so that the few who made it were waved goodbye once hostilities ceased? Has the archival material been misread? And if they did exist, contrary to our expectation, why is it that we didn’t know before? How were the women codebreakers kept out of the history books?

This book tries to answer those questions. To begin the search for women codebreakers it is right to begin at the start of modern codebreaking. That is, at the beginning of the twentieth century, when the invention of long-distance radio signalling revolutionised military and diplomatic communications. No longer was it necessary to tap cables or open letters to discover secrets – but broadcast signals in the open ether could be listened to by anyone. So long-distance wireless telegraphy equated to much more cryptography, and the sudden need for codebreakers, especially after the onset of the First World War. Codebreakers of both sexes.

Dun EaglaisMarch 2025

PROLOGUE

CODES AND CIPHERS ORIENTATION

This prologue provides some background information on codes and ciphers for readers who might otherwise find the jargon of codebreaking, and processes involved in codebreaking, unfamiliar territory. The book does not cover the technicalities of codebreaking, so to avoid breaking up the text with explanations or footnotes, a brief and high-level overview of some of the concepts that women codebreakers of the twentieth century were dealing with is presented here. For those who are familiar with the science of codes and ciphers (cryptology) this will be rather elementary stuff, and those readers may prefer to skip this prologue. For those less familiar, fear not: the most difficult piece of mathematics we will encounter is the addition of single digits, and it is more likely that confusion will creep in because of linguistic oddities. The British called codebreakers cryptographers until after the Second World War, when the American usage, cryptanalysts, was adopted. Semi-technical terms that are underlined in this prologue, and a few more that are encountered in the book, are explained in a reference glossary table on pages 23–24.

To begin: encryption of signals can be done in one of two basic ways, by using a code, or by using a cipher. Most of the women codebreakers featuring in this book handled codes.

Codes are what General Marcel Givierge, an early French textbook-writer on cryptology, called ‘dictionary systems’.1 As with a dictionary, you look up the word or phrase you would like to encrypt, and the dictionary will give you one or more code words. Typically code words are in numerical form: General Givierge gives the following example.

Encoding table

Decoding table

piège

4367

1020

madame

pierre

1025

1021

convoi

pierrerie

9872

1022

accord

piété

0013

1023

marne

pile

1421

1024

heure

pillard

5718

1025

pierre

The number of words or phrases in the codebook will be limited by the number of digits in each code word. General Givierge’s example is a four-digit code, so there are only 10,000 possible words that could be encoded, which might be limiting. A five-digit code has a lot more scope and allows for redundancy.

Redundancy: some dictionary codes will have more than one code word for the same plain text word or phrase. This is a trick to add security, since commonly encountered words that, in the coded version of signals, recur frequently, might be giveaways: readily guessed by dint of their recurrence. Already, variants of the word ‘code’ have appeared six times in this prologue and ‘dint’ only once – so, if there is only one code word for ‘code’ it might become overused and its meaning guessable.

Three-digit codes are used, but you get a lot more words if, instead of using numbers for the code words, you use letters. So, instead of 4367, 1025, 9872 and so forth, the code words could be RTB, QWM, OKX etc, giving a book of 17,576 words. The German Air Force in the Second World War used three-letter codes like this for some operational purposes.

Dictionary codes, like the one in General Givierge’s example, usually randomise the way the numerical code words are assigned to their plain text equivalents. Guessing the meanings of words in unrandomised dictionaries would be too easy: if in a naval code 3446 means ‘convoy’ and 3448 means ‘course’, there can be only so many possibilities for 3447. ‘Corvette’ might, in this instance, be a good candidate. Or it might be ‘Corsica’. Once the context is known, the codebreaker’s task becomes easier.

Interpolating in order to find the plain text meaning of unknown code words is part of the codebreaker’s art, requiring linguistic skills and an insight into the way that the message’s author is likely to express something. Diplomats write differently from generals. The code-maker’s business is to make it all much harder for outsiders to reconstruct the code, regardless of who is sending the message.

One way to make it harder is to superencipher the coded text. Superencipherment works by adding a new random sequence – found in a superencipherment table – to the coded text in order to disguise it. A superencipherment table is nothing more than a long sequence of random numbers (for a numerical code; or random letters for an alphabetical code) broken into groups whose length matches the length of the numerical code words.

Example of a Second World War superencipherment table.2 (The National Archives, Kew)

To use a superencipherment table, the coding clerk tasked with encrypting the message selects a starting point in the table and lines up the random number groups from the table under the code words in the encoded text. Then the groups are added together, thus:

If you look again at the final row, where the numbers have been added together, it will be apparent that the arithmetic looks awry.

In normal primary school arithmetic, adding 95044 and 29879 gives 124923. Here the process differs: non-carrying addition is used so that 9+2=1, not 11. You can envisage this as logical if you imagine the numbers are printed on the rim of a wheel, which clicks as you move it round to the next figure. When you reach 9, the next click on gives the number 0, and after that 1, 2, 3 and so on. ‘Addition’ just means turning the wheel by the requisite number of clicks.

When the signal is received, the decoding clerk carries out the process in reverse:

The same superencipherment groups are taken from the table, a copy of which must be in the recipient’s possession along with the codebook. The groups are lined up underneath the signal, and this time the superencipherment groups are subtracted from the signal. (This is why the British often referred to superencipherment tables as ‘subtractors’.) Taking the first group, 8–8=0 is straightforward but the other four digits of the group involve subtracting a larger number from a smaller one. But again, if you think of the numbers as being on the rim of a wheel, then 1–2 just means move the wheel backwards two clicks from position 1, which gives 9; and 4–9 means move the wheel backwards nine clicks from position 4.

Having stripped off the superencipherment through this process of modified subtraction, the decoding clerk can look up the code groups in the dictionary and reveal the plain text of the message.

This type of encryption – a codebook, plus superencipherment – was used by all the major powers for the first half of the twentieth century. The British, Americans, Germans, Russians and Japanese – all of them used systems like this, with variations, for diplomatic or armed forces signals. Not exclusively, but not reliably either. The story of codebreaking during the two world wars shows that the British, Americans and Germans could all break this type of code, and it may be that the other powers could do so as well.

The codebreaker’s task in this sort of case is threefold: to reconstruct the codebook; to reconstruct the superencipherment table; and to understand the indicator system. The ‘indicator’ is a piece of information included in the transmitted message that tells the recipient which place, in the superencipherment table, was the starting point from which the superencipherment groups were chosen. It wouldn’t do to use the same sequence of superencipherment groups for every signal, since that would make the codebreaker’s job too easy. But sometimes, particularly with over-used or short superencipherment tables, the run of superencipherment groups used in one signal might overlap with the run used in another. Such cases, where the same cipher sequence is repeated in different messages, were called depths.

Finding depths isn’t easy because the codebreaker only has the intercepted signal to go on: even if ‘94921’ was a reused cipher group it might not be apparent if, in two different signals, it was used to conceal different code words (say 22189 in one message and 00766 in another). Statistical analysis can, however, help. And statistical analysis of the type needed can be computerised.

Before the 1950s, computers were people doing repetitious work with adding machines. Still, some of the work done by computers could be carried out by machinery ordinarily used for doing business tasks like invoicing, accounting and payroll. Businesses used punched cards to record data and sorters, tabulators and collators to compare, segregate and do basic arithmetic with them. During the Second World War, all the leading powers made great use of punched-card processing to take the drudgery out of routine and repetitious codebreaking tasks, such as looking for depths in superenciphered coded signals.

Just before we leave the subject of dictionary codes – systems where whole words or phrases are replaced with coded sequences of letters or numbers – we should take another look at the superencipherment process. What is going on here is a substitution: each digit in the sequence of code is replaced, in the superencipherment stage, with a different digit. This characteristic, swapping each letter or number for an apparently random substitute, is what defines a cipher.

Ciphers are almost as old as writing itself. The most famous type of cipher is the ‘Caesar cipher’, where each letter of the alphabet is changed into the letter three places on: A becomes D, B becomes E, and so forth. Caesar ciphers could as easily involve five- or eight-letter shifts, and might switch between three, five or whatever number from one signal to another, as long as the recipient knows what to expect. However, the simplicity of Caesar ciphers means that the art of codebreaking is about as old as encryption.

More complicated ciphers can be created by changing the way that substitution is done: with a new substitute-alphabet for each new letter in the text. In theory, with a twenty-six-letter language like English, there are 403,291,461,126,605,635,584,000,000 possible alphabets (meaning different ways to order the twenty-six letters), which ought to rule out the ability of codebreakers to guess, or test by brute force, what substitution scheme is being used. But cipher systems like this tend to put a limit on the number of substitute alphabets or the books that tell the coding clerks how to carry out the encipherment would become too long and too complicated to distribute and use. So substitution ciphers tend to reuse the same randomised alphabets after a certain point, called the period. Long periods tend to increase security, whereas short periods allow codebreakers to find the pattern more easily.

The following charts show how a substitution cipher might work. To begin with, here are some substitute alphabets we might use to encrypt a message:

To encipher the phrase ‘feed caged face’ we look up the substitute for F in the first cipher alphabet, for E in the second and third cipher alphabets, D in the fourth, and so on; when we run out of cipher alphabets, at the seventh letter (G) in the message we have to start again. The result is QLXZ LURLQ MCYM. The decryption process is just the same thing in reverse.

Long cipher systems like this are very error prone if used by human beings, especially if the cipher system requires the code clerks to change the order in which the substitute alphabets are used from one day to the next or from one message to the next. A better and more secure method is to have a single-use system. Imagine that the code clerk at either end of the transmission has a pad of enciphering tables, each page of which consists of random groups of cipher. After transmitting the signal, the page is ripped off the pad and destroyed at both ends. With true randomness, the cipher could not be guessed at or analysed into order, so the one-time pad method offers complete security as well as relative ease of use. It is, however, expensive and difficult, because vast quantities of randomised cipher are needed.

Simplicity, as well as reliability and affordability, was offered by systems that mechanised the encipherment process, albeit at the price of diminished security. Machine ciphers came into being just after the First World War ended and continued to be the preferred secure method of encipherment until well into the 1950s and beyond. The most famous of the machine ciphers was the one called Enigma, the history of which has been written extensively. There are also many accounts of how Enigma machines work, and how the codebreakers (notably at Bletchley Park) managed to read signals encrypted with what was thought at one point to be an unbreakable system. This prologue, and this book, do not go into Enigma in any detail, and anyone whose interest has been sparked can follow up the story elsewhere. For our purposes, it doesn’t really matter how the Enigma machine worked: all that is significant is that it used a machine to carry out the encipherment, changing the substitute alphabet with every keystroke as the operator typed out the message. Only after 16,900 letters would the sequence of substitute alphabets repeat itself. One clever thing about Enigma, among many, was that the machine was reciprocal. The sender used one Enigma machine to encrypt the message, and if the receiver set up their machine the same way, the receiver just had to type in the received encrypted message and the machine would render it back into plain text.

What made Enigma so secure was the vast number of ways in which the machine could be set up. There were rotors to pick, plugs to insert into a patch panel, rotor positions to set and so forth, amounting to over 150 quintillion possible set-ups. The German armed forces typically changed the machine set-up every day at midnight on every network, in accordance with a prearranged scheme called the key. Only if the message recipient had the key, and could set up their own Enigma machine the same way as the sender, could the encrypted signal be decrypted properly.

Once the mechanism of the Enigma machine had been understood by Germany’s enemies, the challenge for the codebreakers was to figure out the key system. If the key of the day was known for a given network, anyone with an Enigma machine (or a rip-off that worked the same way) could set up their machine and translate the encrypted messages into plain text.

One technique for breaking cipher systems is to exploit a ‘probable word’ or phrase the codebreaker imagines the message’s author will use. Military signals are rather fertile ground for repeated words and phrases, such as honorifics respectful of the receiver’s rank or stereotyped formatting of reports. Matching up the probable word against the cipher text sometimes reveals patterns that uncover the structure of the cipher. At Bletchley Park, the probable words were called cribs, and exploitation of cribs drove the design of the famous bombe machines used by Bletchley Park to find the Enigma keys every day.

Enigma was not the only type of cipher machine in use during the Second World War. All sorts of mechanical and electrical devices were invented to do a similar job. Although they are less well known, some types of cipher machine were actually more numerous than Enigmas, and other types of cipher machine yielded higher-quality intelligence than Enigma. None of this really matters for present purposes, because the principle is the same in each case. The machine does the letter-by-letter substitution, and the machine operators at both ends need to know what key to use to carry out encryption and decryption correctly. Codebreakers who have intercepted the signals from the radio waves need to know how the machine works, and to figure out the key.

GLOSSARY

Book-building

Reconstruction of a code dictionary from clues in the coded text.

Brute force

Trying each possible setting of a cipher system in turn.

Cipher (older British English spelling ‘cypher’)

A system by which each letter or number in a portion of text is converted into something else. Further explanation above in this prologue.

Code

A system by which each word or phrase in a portion of text is converted into something else. Further explanation above in this prologue.

Crib

A word or phrase expected by the codebreaker to appear in the original plain text of an enciphered message.

Cryptanalysis

Codebreaking. The British did not use this term until after the Second World War. Nowadays, even British codebreakers are called ‘cryptanalysts’ or in slang usage ‘cryppies’.

Cryptography

In modern usage, and American mid-century usage, this means code-making. The British used the term in a more general sense, to encompass all code and cipher work, in particular codebreaking. This book does not deal with code-making, so in passages quoted from archival sources ‘cryptographer’ means ‘codebreaker’.

Cryptology

The study of codes and ciphers. Codebreakers are one subspecies of cryptologists.

Decrypt

To reveal plain text by reversing a code or cipher.

Depth

More than one enciphered message enciphered using the same, or part of the same, cipher sequence.

Encrypt

To impose secrecy by converting to a code or using a cipher.

Indicator

A piece of information included in a message transmission which notifies the recipient of some otherwise unknown component of the key.

Key

Specification of the variable features of a cipher system. For example, in a modified Caesar cipher (see description above in this prologue) that shifts all letters down the alphabet by a different number each day, the key would be the number of letters specifying the shift. Further explanation above in this prologue.

One-time pad

A cipher that is only used once, then thrown away.

Plain text

The unencoded (or unenciphered) text that it is intended to disguise before sending, and to be read by a legitimate recipient.

Superencipherment

A group of numbers added to a code group from a book, or dictionary, code to disguise the code group. Further explanation above in the prologue.

ACRONYMS

CBME

Britain’s Combined [Intelligence] Bureau, Middle East, based in Cairo

GC&CS

Government Code & Cypher School, later GCHQ

ISOS

‘Illicit Signals Oliver Strachey’ – Bletchley Park’s section dealing with manual codes and ciphers of the German intelligence service

OP-20-G

US Naval Intelligence codebreaking subdivision

SIS

US Army’s Signal Intelligence Service

TJAO/TSAO

Temporary Junior/Senior Administrative Officer, grades for wartime staff that matched the interwar junior assistant and senior assistant ranks. Below TJAOs were TAs (Temporary Assistants)

WHO’S WHO

Listed here are a selection of the characters who feature in the story. In the book, they are called by the name by which they went at the time. Boldface type indicates the name, or in a couple of cases names, by which they were then known – that is to say, for most of them, when they were breaking codes.

Emily Anderson

GC&CS’s first woman senior assistant

Margaret (Madge) Anderson (Mrs Winterbotham)

British head of Spanish diplomatic codebreaking

Patricia Bartley (Mrs Brown)

British head of German diplomatic codebreaking

(Mrs) Wilma Berryman (later Mrs Davis)

Early recruit to SIS

Frank Birch

Head of Naval Section at Bletchley Park

Margarethe Bitter

Lawyer and diplomat who broke codes for Germany

‘C’

(to Nov 1939) Sir Hugh Sinclair

(from Jan 1940) (Sir) Stewart Menzies

The head of MI6 (or Secret Intelligence Service), until 1944, also ‘director’ of the GC&CS

Alan Bradshaw

Bletchley Park’s head of admin

Dorothy Brooks (Mrs Carr)

Junior codebreaker at CBME

Violet Cane

Statistician who broke codes in Bletchley’s Naval Section

Ann Caracristi

NSA’s first woman deputy director

Joan Clarke (Mrs Murray)

Britain’s most famous woman codebreaker

W.F. Clarke

Interwar head of GC&CS Naval Section and sardonic observer

Josh Cooper

Head of Air Section at Bletchley Park

Myra Curtis

Senior Treasury official, later principal of Newnham College

Marjorie (Madge) Dale (Mrs Webster)

Classical scholar on the 1939 list of potential wartime codebreakers

Nigel de Grey

First World War codebreaker, second-in-command at Bletchley Park

Alastair Denniston

Head of the GC&CS until 1942, then ‘Deputy Director (Civil)’

Agnes Meyer (Mrs Driscoll)

US Navy’s first and principal codebreaker

Fiona Ede

British head of South American diplomatic codebreaking

Marie Rose Egan (Mrs Palmer)

CBME Air Section chief cryptanalyst

William Filby

British head of German diplomatic codebreaking, after Patricia Bartley

Elizebeth Smith (Mrs Friedman)

Head cryptanalyst for US Coast Guard

William F. Friedman

Chief cryptanalyst of SIS and later NSA

Asta Friedrichs

German Second World War codebreaker

Gene Grabeel

First American codebreaker to attack VENONA messages

Genevieve Grotjan (Mrs Feinstein)

SIS codebreaker whose breakthrough led to mastery of PURPLE

Ursula Hagen

Head of British diplomatic codebreaking at German Foreign Office

N.M. (Natasha) Harris

British interwar and Second World War GC&CS codebreaker

Florence Hayllar

British First World War and briefly GC&CS codebreaker

Annaliese Hünke

German Second World War codebreaker

Freddie Jacob

Director of CBME

Dilly Knox

First World War codebreaker, later led British attack on Enigma

Mary Lane

SIS codebreaker and NSA staffer

Mavis Lever (Mrs Batey)

Senior codebreaker in Bletchley’s Enigma team

H.C. (Nellie) Lunn

Interwar and Second World War GC&CS codebreaker

Alda Milner-Barry

British First World War codebreaker, later talent spotter at Newnham College

Maude Moore

Foreign Office ‘Chief Woman Officer’

Juanita Morris (Mrs Moody)

SIS codebreaker and NSA staffer

Erika Pannwitz

Mathematician and German Second World War codebreaker

Adolf Paschke

Principal figure in German Foreign Office codebreaking unit

Margaret Rock

Senior codebreaker in Bletchley’s Enigma team

Frank Rowlett

Senior SIS codebreaker and leader of PURPLE team

Laurance Safford

Head of US Naval codebreaking until 1942

Evelyn Sinclair

British First and Second World War codebreaker

Delia Taylor (Mrs Sinkov)

SIS codebreaker

Claribel Spurling

British First World War codebreaker

(Mrs) Ray Strachey

Women’s rights campaigner

John Tiltman

GC&CS’s Chief Cryptographer after Dilly Knox

(Sir) Edward Travis

Head of Bletchley Park and then GCHQ from 1942

Alan Turing

Britain’s most famous male codebreaker

Gwen Davies (Mrs Watkins)

British Second World War codebreaker

Constance Webb

British Second World War codebreaker

Rhoda Welsford

British First and Second World War codebreaker

Winifred (Wendy) White

British interwar and Second World War codebreaker

Ann Williamson (Mrs Mitchell)

British Second World War codebreaker

Eunice Willson (Mrs Rice)

OP-20-G codebreaker

1

‘FACING THEPROSPECT OF WOMEN’

Elizebeth Smith wasn’t so much inducted into the obscure arts of codebreaking as abducted:

After graduating from college I started forth to seek a congenial way of earning my living … I was told by the agency to go to see the reference librarian at the Newberry Library in Chicago where there was a First Folio of the Shakespeare Plays, and that this librarian knew of a possible position in which I might be interested … It seems that there was a very wealthy man who lived on an estate outside of Chicago, who had become interested in the disputed authorship of the Shakespeare plays and was carrying on an investigation of some kind at his estate and that he wished a college graduate who had a background of study in English literature … This hypothetical person, preferably female, was to be young, personable, attractive and a good talker … Instead of requesting her to send me to his office, he decided to come to the library.

He arrived. He was a large man, bearded, which was very unusual in those days, not too well dressed, but with a dashing, imperious manner. He wasted no time; he didn’t wish to talk to me at the library but at once invited me to go with him to his estate at Geneva, Illinois and spend the night. I was a bit taken back, because in my youth and innocence and small-town origin and habits, this kind of thing seemed a bit startling to me. But he was the kind of man who did not take no for an answer. His town car, with chauffeur, was waiting outside and I proceeded with him to the Chicago Northwestern railroad station where we took a train for Geneva, Illinois. He talked about everything under the sun except the possibly forthcoming position, until we got on the train and sat down. Then after a very few moments he turned to me, sitting next to the window, and shouted, ‘WHAT DO YOU KNOW?’

I was realizing by this time that I probably appeared a demure little nobody to him … So when he thundered at me, ‘Well, what do you know?’ I leaned as far away from him against the window as possible, and looked at him quizzically out of my half turned eyes and said, in a firm but low tone, ‘That remains, Sir, for you to find out.’ My answer apparently pleased him, for he burst forth in a loud guffaw which could be heard all over the car.1

It was 1916. In this manner began the career of a woman destined to be not only America’s best-known codebreaker but also one of the finest, with a string of achievements ranging from gang-busting in the era of Al Capone to the solving of Enigma machines. Not to mention the social code of dealing with the oppressive, obsessive personality of Colonel George Fabyan, who was not only her abductor and her employer for the next four years, but also the person who introduced her to her husband-to-be.

But the one problem Elizebeth Smith was never going to crack was the mystery of secrets hidden within Shakespeare’s First Folio.

Fabyan had been convinced by a certain Mrs Elizabeth Gallup that the First Folio printing used a cipher, which hid in plain sight the conclusion that Francis Bacon had written the plays we attribute to Britain’s greatest bard. The cipher had been described by none other than Bacon himself, in a work called De Augmentis Scientarium, which was conveniently published in the same year as the First Folio. The idea was simple enough: each ordinary letter could be represented in five-digit binary form, say 00000 for A going through to 10111 for Z. The 1s and 0s were concealed in the printed text, using differences in the typeface to distinguish them. So a simple phrase like ‘Now is the winter of our discontent’ could have a double meaning, if printed like this: Now is the winter of our discontent. Because, if you cut the phrase into sections of five letters, and take the letters in italics to be 1s and the rest to be 0s, you get:

Nowis│thewi│ntero│fourd│iscon│tent

or 00001 00000 00010 01101 01100, which happily spells out BACON in the binary code. Mrs Gallup’s plan was to spot the typographical oddities that indicated which letters were zeros and which were ones in the First Folio.

It was never going to be. Handcrafted type of the Shakespeare-Bacon era conspired with the muddy ink and artisanal paper of the time to produce random spots, dots, broken and bent glyphs, inconsistent serifs, and all manner of inconsistency that you don’t find in modern printed works. On the one hand, there was enough variation in the First Folio to keep Mrs Gallup, Miss Smith and a host of others gainlessly employed for a long time at the expense of Colonel Fabyan, but on the other hand, teasing out the secret signals concealed by the late Mr Bacon was a matter more of imagination and self-deceit than of codebreaking.

Still, the pointless work had its benefits. One was a ‘dark-haired young man, who was in charge of the genetics experiments’ at Fabyan’s neo-enlightenment research park, which Fabyan called the Riverbank Laboratories. Fabyan’s interests were many, and among them was the emerging study of fruit fly characteristics. Fruit flies breed rapidly and exhibit mutations that are easily seen under a microscope, making them ideal for exploration of Mendel’s notions of genes as atomic components of life. The man in charge of the genetics experiments:

was found to be a gifted person in the realm of drawing and photography. So in his spare time he was pressed into service by Colonel Fabyan, who got some idea that, by enlarging the type forms in the Elizabethan books, the differences which Mrs. Gallup claimed were there would show up … This order on the part of Colonel Fabyan for the young man to go into this type of affirmation of Mrs. Gallup’s work, threw him and me together a very great deal, in and out of social hours. His name was William F. Friedman, and we were married within a year.2

‘When asked how he happened to become the father of American cryptology, Mr. Friedman smiles and says, “I was seduced.”’3 That remark was made in 1954, by which point Mrs Friedman’s husband had established himself as the premier figure in military codebreaking. The path he followed started at Riverbank, when the ever-boastful Colonel Fabyan notified the US Army that he could help them with any codebreaking problems they might have.

Fortunately for Colonel Fabyan (whose colonelcy was honorific, rather than military), and very fortunately for the Friedmans, the army accepted the offer. In the run-up to America’s involvement in the First World War, the US Army had no codebreaking organisation, although two officers tried to do the codebreaking work in their spare time. So the limited experience of the Friedmans in attacking an unencrypted seventeenth-century text in search of an undetectable Baconian cryptosystem was about as good as it got. One of the army’s part-time codebreakers, who worked on unsolved messages in partnership with his wife Genevieve, was Captain Parker Hitt. Captain Hitt was not only a real army officer, but he had written a manual on code-making and codebreaking, and Fabyan got hold of a copy of this manual for the Friedmans.4 Now they had the chance to get stuck into genuine encoded material, and some basic precepts to apply. And, fortunately for the reputation of honorary colonel Fabyan, both the Friedmans turned out to be very good at it – very good indeed.

They began with Mexican signals. The United States entered the war in 1917 because of a threat from Mexico: knowing what the Mexicans were thinking was crucial intelligence. There followed a flood of encrypted material sent over by the War Department, the Navy Department, the State Department and others. Then there was the security of America’s own enciphered signals. For the Friedmans, a lifetime in codebreaking had now begun in earnest.

By the 1950s, the Friedmans were the established authority in American codebreaking. But in the years following the First World War, as befitted people with a secret occupation, they were largely unknown. Instead, America’s best-known codebreaker was Herbert O. Yardley. His fame came from a book he wrote in 1931 called The American Black Chamber. This described, in detail embarrassing to the authorities, the codebreaking activities he and others had engaged in during and after the war in the State Department. His book begins with an explanation, of sorts: his operation had been closed down in 1929 on the orders of Henry L. Stimson, then the Secretary of State, who ‘was dealing as a gentleman with the gentlemen sent as ambassadors and ministers from friendly nations, and, as he later said, “Gentlemen do not read each other’s mail”’.5 So Mr Yardley was out on his ear, and – so it seemed – there wasn’t going to be any codebreaking in the United States any more.

That might have held true in the State Department, but other agencies and arms of the US government were less inclined to be gentlemanly. Discretion would have to prevail, in case Mr Stimson were to find out, but the US Army needed to ensure that its communications methods were secure, and security meant imperviousness to enemy codebreakers. Or, to finesse it, friendly codebreakers were needed to check up on one’s own codes, and to keep pace with developments.

And there were enemies about, even in times of peace. After an interlude of working for the army, the navy and some freelancing, Mrs Friedman had taken on a new role in December 1925. While people in Britain were wrapped up in a continuing debate about extending the voting franchise to women (until 1928, it was limited to over-30s graduates, homeowners, renters and wives of home-owners), America was wrestling with a different piece of post-war policy: prohibition. Whether you were on the side of temperance or not, there was a law enforcement problem that has resonances with today’s war on drugs. Booze smugglers were not just supplying drinkers with their daily tot but organised in crime syndicates, and the burden of enforcement fell on the US Coast Guard, whose duties had previously been about search and rescue for shipwreck victims. Importers of banned spirits would meet suppliers in offshore locations to transfer the illicit goods – and the signals detailing the rendezvous would be sent in code. To deal with the smugglers, the Coast Guard needed a codebreaker.

Codebreaking for the Coast Guard was ideal for Mrs Friedman. Her daughter Barbara was born in 1924, and a son John followed in 1926; she could work from home, so combining family life and professional duties in a way that is very familiar to the 2020s, even if we might imagine it to be a recent innovation in workplace culture. And the work was interesting, challenging and even dangerous. The Reader’s Digest reported on it, somewhat breathlessly:

So in 1923 the Coast Guard chiefs sent for Mrs. Friedman and appointed her ‘key woman’ of codebreaking activities on the crime front. She is entrusted with more secrets of the crime world and of federal detection activities than any woman in history. Gang bosses have gone to penitentiaries as a result of her expert efforts. Syndicates have been broken up and millions of dollars’ worth of outlaw business has been interrupted.6

Breaking up the syndicates and putting the gang bosses into penitentiaries involved court appearances, for without the crucial evidence from decrypted signals the guilt of the participants could not be proved. Mrs Friedman had to have a bodyguard for one of the cases, in which a central figure had been gunned down for ratting on his mates. Undaunted, she decided to ask for an expansion of her team in 1930, reckoning that the amount saved on fruitless Coast Guard expeditions through better targeting would more than justify the hiring of further codebreakers.

Codebreaking reached maturity in Britain during the First World War, just as it did in America. Of course, that war began earlier for Britain, and its codebreaking correspondingly developed earlier. The story has become rather well known, through the stories of the Admiralty’s Room 40, which housed several codebreaking characters who became senior and influential figures at Bletchley Park in a later war. Room 40 was good at self-publicity, too, with carefully managed leaks after the war was over, explaining how Germany’s naval codes had been mastered, allowing the Royal Navy a critical tactical advantage and revealing that America’s entry into the war was in part attributable to Room 40’s break into the notorious Zimmermann telegram – a German diplomatic signal that fed straight into America’s paranoia about the Mexicans. What happened in Room 40 and the men who broke codes for Britain became, in due time, the definitive account of secret intelligence in the First World War.

Codebreaking wasn’t really a business before the war. What changed was the way people sent messages to each other. The telegraph had allowed for telegrams (or ‘cables’) to go via land and sea lines, but the first step taken by the British on the outbreak of war had been to cut the lines. Communication, at least for long-distance signals, was now on the airwaves, where anyone with a receiver tuned to the right frequency could pick it up. Hence the growth in cryptography. And more: battlefields were bigger, so command and control could rely less on couriers and landlines than before. Navies could be commanded by radio. Military and naval signalling, by now, was a matter of cryptology as much as diplomatic communication.

And after the hostilities? Afterwards is when the historical records really begin. For it was in late 1919 that a new peacetime organisation was set up for the British codebreakers – the creation of the obscurely named Government Code & Cypher School, headed by Commander Alastair Denniston from Room 40. The ‘school’ was no such thing, because there was officially no such thing as codebreaking.

The records show that, beside Commander Denniston himself, there were six senior assistants (all men) and twenty others in the new GC&CS. One, a naval lieutenant, was the ‘wireless expert’. Sixteen junior assistants included two women, Miss E. Anderson and Miss F.H. Hayllar. There were also three ‘lady translators’, Miss H.C. Lunn, Miss J.F. Carleton and Miss C. Spurling. Translators were apparently doing the same work as junior assistants, as appears from a letter of 23 August 1919 in which the Treasury asks that the class of junior assistants be ‘subdivided into (a) Junior Assistants at £200–500 as proposed, and (b) Translators on a scale of (say) £150–15–300 plus war bonus’.7

The records are as cryptic as the signals the codebreakers were supposed to decipher. The senior and junior assistants were assisting the head of the school, Commander Denniston, and despite the lowly sounding job titles they had a high degree of seniority, as measured against other Civil Service grades. The junior assistants, in fact, ranked equally with assistant principals on the regular Civil Service scale, which is definitely not a junior grade. One can easily be misled when assistant principals are described as ‘junior’, and the error can be compounded by looking at those salaries. ‘£200–£500’ was not generous, possibly because real salary values had halved due to wartime inflation. ‘£150–15–300’ means an annual salary of £150, rising by annual increments of £15 to a ceiling of £300. While not generous, these were professional salaries, comparing with about £40 a year for the average working man.

The job titles are not just lowly sounding: they were a mechanism for saving money in the cash-strapped post-war era. Translators were not only a sub-category of junior assistants, they were women, and women were paid less than men. Junior assistants could be men or women, but only women could be lady translators. So the money-saving Treasury created the lower pay grade for translators, into which the women codebreakers were largely placed. It helped the Treasury further that, when the GC&CS was finally brought into being, lady translators were not permanent, pensionable staff.

In 1922 responsibility for the GC&CS was transferred from the Admiralty to the Foreign Office, which had to take over the payroll. Ninety-one staff were taken on to the Foreign Office budget. Who was employed at what level is quite interesting:8

Women

Men

Salaried

5

25

Weekly paid

52

6

There is a clear gender line, with women taking the majority of the lower-grade roles and men in the preponderance at higher grades; a gender pay gap analysis is familiar enough a century later, and might make for similar reading in some organisations. What is also interesting is that all but four of the fifty-two weekly paid women were classified as temporary woman clerks. ‘Temporary’ denotes a staff member who does not qualify for an occupational pension, and in their cases, the top of the pay scale would be reached after only three years.

What might be more surprising is that women were accepted at the codebreaker level, with one in six positions being filled by females. By 1922, Miss Hayllar, Miss Carleton and Miss Spurling had left. Miss Hayllar, who retired aged 52 to focus on writing poetry, was replaced by a man. But Miss Anderson stayed on. For several years, she would be the only woman to enjoy the dizzying seniority of a junior assistant at the GC&CS.9