Dilly

by Batey, Mavis;

  It was decided that we must recruit some newly graduated mathematicians to join the team, at least one of whom would become a professor in later life. Dilly insisted that I should be present at the interviews, not to check their capabilities, as that would be taken care of, but to see if these males were the sort of people the girls could get along with. By then I had confessed that I had become engaged to the ‘wranglercove’ Keith Batey and although Dilly congratulated me heartily, he asked me if I knew that mathematicians as a breed were not usually very imaginative. I reassured him that this one was all right. When I showed Olive my engagement ring and it transpired that I had chosen it myself, she told me that Dilly had bought hers himself and given it to her in Room 40 as he thought that was what the bridegroom was meant to do. By the tone of her voice I rather detected that she too would have preferred to have been in on the choice.

  Dilly wanted to hear about the visit of the Americans to see how our Abwehr machine worked. I told him that two service types came, one large and bullying with two ribbons and the other lean and acquiescent with one. After five minutes, the leader said: ‘We’ve sure got your machine taped now,’ which of course he hadn’t. The other one made what we thought was quite an intelligent observation to be crushed by ‘Don’t be ridiculous’ and he said meekly: ‘Sorry, it was only an idea.’ He awoke our maternal instincts and as we had managed to get the ingredients for a chocolate cake together, which we baked in the kitchen for their visit, we made sure he was given a much larger piece. When they left they said that they had never had such hospitality, nor met such charming people. My note ended: ‘I hope we are not going to have a whole string of Dough-boys to drive dizzy.’ I think it was the lack of hierarchy that amazed all the Americans. Anybody, however junior, with a bright idea was listened to in ISK and often it worked. William Friedman, America’s top codebreaker, on a visit to Bletchley Park in 1943 was astounded at what he regarded as a bunch of amateurs could produce. ‘In a technical sense, we are ahead,’ he said. ‘But in a practical sense, judged by accomplishments, these amateurs have very largely surpassed us in detail, attention to minutiae, digging up every bit of intelligence possible and applying high-class thinking, originality and brains to the task.’

  Admiral Godfrey made a point of visiting Courns Wood as often as he could to tell Dilly how strategic deception was being planned as a result of his Abwehr break, now that Torch, the first Allied operation, was scheduled for the autumn; this was something which was not known to the staff in ISK itself of course, where to us it would just be action stations jumbo rush when it happened. On one occasion, Dilly told Godfrey that he was sure he would get better if he had a cruise and asked him to try and get him on an Arctic convoy; the astounded admiral could only say that these were not vessels to recuperate on. We knew Dilly had said he wanted some sea air and thought he just meant a yachting trip on the Fal with Frank Birch as in the old days, but alas not even this could take place. Peter Twinn and I continued to visit Courns Wood regularly to keep Dilly informed of ISK progress and were taken by official car. There was one driver who seemed to think we were out for a joy ride and was determined that we should see all the sights of Buckinghamshire – Chequers, Ishbel Macdonald’s pub and some speed trials, but we told him not to waste time and petrol and take the shortest route.

  At this time, Margaret Rock had for several months been working with Dilly at Courns Wood on an ‘isolated problem’ which was not connected with Abwehr. She was able to drive Dilly over, sometimes bearing lovely fruit, which we ate in the garden at Elmer’s School. She was very secretive about the work but as it had no bearing on ours we knew better than to ask her. It has been suggested that Dilly was working on Soviet ciphers before he died and this may well be true. Russia was our ally and officially we had ceased to read their traffic, but in March 1942 ‘C’ and the director general of the security services decided that some Russian systems should be worked. Dilly had of course worked on Comintern and with the great Russian codebreaker, Ernst Fetterlein, and it might have been the answer to let him work on it away from mainstream codebreaking. Margaret and I went on to work on Russian at Eastcote after the war. But ‘need to know’ was just as strict then and I did not ask her if that was what she and Dilly were doing in 1942. Margaret Rock did come back to hold the fort for the jumbo rush before Operation Torch when Keith and I were married in London with Peter Twinn as best man. Dilly sent us a lovely wedding present and said that if the Muse hadn’t left him he would have composed my wedding hymn.

  Soon afterwards Dilly was admitted to University College Hospital and when I visited him his brother Evoe was by his bedside and they were roaring with laughter composing Dilly’s last words. They were reading from a book of collected famous last words called The Art of Dying, which Dilly handed to me when I left. I still have it. Dilly’s last words were in fact: ‘Is that Ronnie outside in the corridor bothering God about me?’ He wrote a pentelope for his epitaph:

  A wanderer on the path [A]

  That leads through life to death, [E]

  I was acquainted with [I]

  The tales they tell of both [O]

  But found in them no truth [U].

  Dilly died at home on 27 February 1943, aged fifty-eight, and was buried in his woodlands, where Olive’s ashes would later be scattered. John Maynard Keynes, who was present, described his old friend as ‘sceptical of most things except those that really matter, that is affection and reason’.

  Just before he died, Dilly was appointed as a Companion of the Order of St Michael and St George ‘for services to his country’. It was explained to Olive that security conditions precluded a more illustrious honour. Dilly insisted on getting dressed to receive the emissary from Buckingham Palace and, typically, he sent the decoration to ISK with a touching note saying it was really meant for them. It was heartening to know that he regarded ISK and what he called the ‘Cottage tradition’ as the fulfilment of his career. The official farewell letter shows that, until the very last moment, he was insisting that codebreaking and intelligence work were inseparable and that the secrecy of ‘need to know’ and ‘working in blinkers’ was counter-productive. His letter ended: ‘In bidding farewell and closing down the continuity but not, I hope, the traditions of the Cottage, I thank once more the section for their unswerving loyalty. Theirs I remain

  Affectionately

  A. D. Knox.’

  Courns Wood

  Hughenden

  Bucks

  Jan 3, 1943

  Dear Margaret [Rock]

  Mavis [Batey]

  Peter [Twinn]

  Rachel etc.

  [Denys] Page

  Peggy [Taylor]

  Very many thanks for your, and the whole section’s, very kind messages of congratulations. It is of course, a fact that the congratulations are due the other way and that awards of this sort depend entirely on the support from colleagues and associates to the Head of the Section. May I, before proceeding, refer them back.

  It is, I fear, incumbent upon me at the same time to bid farewell. For more than ten years, I have taken up with A.G.D. the position that (a) there is no proper distinction between research cryptography and cipher and intelligence work, (b) that it is as improper to ask a person of any degree of education to run a key-setting bureau, as to ask him or her to run a typewriting section, (c) that it is impossible to edit or translate satisfactorily without a precise knowledge of the cipher in question, (d) latterly we have recruited during and shortly before the war from the Universities; and Academic tradition does not understand the idea that a half-fledged result should be removed from the scholar who obtains it and handed over to another. The discoverer loses all interest in further discovery and the recipient has no interest in the offspring of another’s brain. Still more doubtful is the case of ‘Research’. Until we know who will handle and circulate any result we get, the irrepertum aurum of our search will very probably be sic melius situm.*

  I have recently arranged wi
th the authorities to attempt a line which should give wider scope though it will be of far less importance. In bidding farewell and in closing down the continuity but not, I hope, the traditions of the Cottage, I thank once more the section for their unswerving loyalty. Theirs I remain

  Affectionately

  A. D. Knox

  The original letter, written in green ink on blue paper, an old naval tradition, can be found in The National Archives (Public Record Office) HW 25/12

  * The Latin is a quote from Horace, Odes III, 3, 49, irrepertum aurum et sic melius situm – ‘the gold unfound and so the better placed’.

  Introduction to the appendices

  Dilly began to investigate the characteristics of the Enigma machine as soon as he had access to one, probably as early as 1925. With an actual machine in front of him, he could tell how the keyboard letters were connected to the entry plate, and by battery testing could determine the internal wiring of the wheels. By tapping messages out to himself he could then begin to experiment on ways in which the wiring could be ascertained, using his own ‘crib’ for the cipher text. A method for breaking the wheel wiring was essential as the manufacturers had indicated that to increase security the wiring of the original wheels could be changed.

  The instructions given to ordinary German soldiers, the Schlüsselanleitung für die Chiffriermaschine Enigma, included an example using a real machine which would have given Dilly some of the answers if only SIS had agreed to pay for it when it was first offered to them by the French in 1931. Appendix 1 is a translation of those instructions, which explain how the machine was set up to encipher a message. The technique Dilly invented to break the machine was by ‘rodding’ and ‘buttoning-up’, which he first put into practice with operational messages in the Spanish Civil War in 1937, when Mussolini and Franco obtained from Hitler improved ‘K’ models of the commercial Enigma machine with rewired wheels.

  Throughout the chapters in this book, there is frequent mention of Dilly’s system of rodding for breaking messages. Appendix 2, written by Frank Carter for the Bletchley Park Trust, explains how the rods were made specifically to match each machine. Although it refers to the rods’ original use for the un-Steckered Enigma machine, there were occasions when rodding could be applied to the Steckered German service machine, such as the ‘routine operation’ put in place in confirming the authenticity of the ‘crib message’ provided in the 1930 operator’s manual when Dilly stripped off the Stecker as a reciprocal substitution super-encipherment. His fundamental methods for breaking Scherbius’s original machine had been a vital step in cryptanalysis, allowing later complications, such as the multi-turnover nature of the Abwehr machine, to be dealt with as they arose. As stated in the appendix, the rods used are those made for the German Railways ‘K’ model machine, broken on Dilly’s methods in 1940. The term ‘rotor’, as is now normal in modern cryptographic texts, has replaced the word ‘wheel’, which was always used at Bletchley Park, being a near translation of Walze, the word used in the German description of the Enigma machine.

  Appendix 3, also written by Frank Carter for the Bletchley Park Trust, explains ‘buttoning-up’. The author is very grateful to Frank for providing these detailed explanations of Dilly’s techniques. Frank in turn would like to thank Keith Batey, the author’s husband, for his assistance in their compilation. Appendix 4 is Dilly’s wonderfully idiosyncratic description for Alastair Denniston of the ‘lobster hunt’ that preceded the breaking of the Abwehr Enigma, and Appendix 5, compiled by Ralph Erskine, gives a complete list of the machines broken by Dilly’s ISK section.

  APPENDIX 1

  SECRET

  Directions for Use of Keys for the Enigma I Cipher Machine

  I. Distribution of Keys

  1. Within a given district (e.g. a Division, or on manoeuvres one side, etc.) one ‘district key’ is in force.

  2. For communication beyond the limit of this district ‘transverse key’ is employed.

  II. Keys

  3. Each ‘key’ contains all necessary directions for setting the cipher machine for the period of one month, as follows:

  (a) Directions for the position of the wheels, changed every three months.

  (b) Directions for the setting of the tyres (changed daily).

  (c) Directions for the basic setting (changed daily).

  (d) Directions for plug connections (changed daily).

  4. While the above directions are similar and hold good for all users of the cipher machine, the choice of the message key is left (with one or two restrictions) to the encipherer. Hence the ‘key’ contains no directions for setting the ‘message key’.

  III. Setting the Key

  5. The directions for the position of the wheels give the order, from left to right, in which the individual wheels are to be inserted, e.g.: II I III.

  6. The directions for setting the tyres give the setting of the tyres of each individual wheel, e.g.: II:24 I:13 III:22.

  7. The directions for the basic setting give the figures which must be set in the three windows from left to right, e.g.: 06 15 12.

  8. Directions for plug connections indicate which twin sockets are to be connected to each other by means of the twin plug leads, e.g.: 1/13 6/9 14/22 16/19 20/21 23/26. (Each of these figures denotes a particular twin socket and the stroke separates the numbers of the twin sockets that are to be connected together, thus socket 1 has to be connected to socket 13, 6 to 9, 14 to 22, etc.)

  9. Every message is enciphered using a particular message key. This is chosen at will from the range of figures 01 01 01 to 26 26 26. For every message on the same day a new message key must be used for the starting point. It is forbidden to encipher two or more messages on the same day using the same message key.

  10. The recipient of the message is informed secretly of the chosen message key. The figures of the message key are replaced by the corresponding letters on the plugboard (e.g.: 1 is replaced by A, 13 by M, etc.), and using the basic setting the keys are pressed twice in succession. The resulting six letters are prefixed to the message which is being enciphered with this message key.

  IV. Example

  A. Enciphering

  11. Key in force for the day:

  Position of the wheels: II I III; Setting of the tyres: II:24 I:13 III:22; Basic setting: 06 15 12; Plug connections: 1/13 6/9 14/22 16/19 20/21 23/26

  12. A message to be enciphered at 1035 reads: ‘Feindliche Infanteriekolonne beobachtet. Anfang Südausgang Bärwalde. Ende 3 Kilometer ostwärts Neustadt.’ [Hostile column of infantry observed. Extends from the south exit of Bear Woods to position 3 kilometres east of Neustadt.] As ‘ch’ is replaced by ‘q’, ‘ä’ by ‘ae’, and ‘ü’ by ‘ue’, and punctuations are replaced by Xs, the message is enciphered in the following form: FEINDLIQE INFANTERIEKOLONNE BEOBAQTET X ANFANG SUEDAUSGANG BAERWALDE X ENDE DREI KM OSTWAERTS NEUSTA DT.

  13. The sender selects the [arbitrary] message key 01 02 12 and converts it into letters on the plugboard, this reads A B L. In accordance with Para III, 10, the key is pressed twice in succession (with a basic setting 06 15 12), and the letters PKPJXI will result. As the message in five-figure groups follows immediately after these letters, they are written down in the following form: PKPJX I.

  14. Then, after setting the message key 01 02 12, the clear text is enciphered by pressing the keys. The resulting letters are written down after the above six letters. Thus the first four letters of the message are added to the I in Para IV, 13 above and the remainder are likewise divided into five-letter groups: PKPJX IGCDS EAHUG WTQGR KVLFG XUCAL XVYMI GMMNM FDXTG NVHVR MMEVO UYFZS LRHDR RXFJW CFHUH MUNZE FRDIS IKBGP MYVXU Z

  15. The following precede the enciphered text:

  (a) Time group (four figures)

  (b) Number of letters (two or three figures)

  (c) To be able to see at once that the message was enciphered using the machine, an arbitrary three-figure number. The enciphered message therefore reads as follows: 1035-96-341 PKPJX IGCDS EA
HUG etc., etc.

  Should a message be longer than 180 letters, then it is to be divided. The second and third portions are always to be enciphered with a different message key, which is to be disguised in the prescribed way.

  16. On conclusion of enciphering, the final figures appearing in the windows are to be altered by casual rotation of the wheels.

  B. Deciphering

  17. The enciphered message reads: 1035-96-341 PKPJX IGCDS EAHUG WTQGR KVLFG XUCAL XVYMI GMMNM FDXTG NVHVR MMEVO UYFZS LRHDR RXFJW CFHUH MUNZE FRDIS IKBGP MYVXU Z.

  18. The directions preceding the message which give the key for the day also show the position of the wheels, tyre settings and plug connections for that day. The machine is set according to these. Using the basic setting, the first six letters PKPJXI are typed. The resultant letters are ABLABL.