by Alastair Denniston- Code-breaking From Room 40 to Berkeley Street
In general, he would like to employ liaison on the intelligence side of those cases where liaison on the crypt side does not seem to him desirable. He is willing to let Taylor and another properly vested officer look at all material which is circulated by Berkeley Street, including all material into and out from London which is thus circulated. There is a limited class of material that his office sends to C instead of circulating, just as in the case of Park. He illustrates these by the case of some British diplomat ‘dropping a brick’ but very likely any case where, from the British standpoint, a delicate subject was involved would fall into this reserve class, and with an American officer reading the material it would only be natural that a new class of reserve material would arise, to wit, those that it might be unwise to let an American officer read. However, Denniston’s whole approach is very reasonable and he says he has great hopes that all mutual misunderstandings will be cleared up and the line of liaison straightened out. While the basis for his feeling is not yet entirely clear to us, it is plain that he has been annoyed by the way Bailey and Maidment have handled the exchange problem, without consulting him as he thinks they should, and without giving his people any clear idea of just what the exchange is supposed to cover. Taylor found that Waterfield, the man in charge of traffic, had no list to guide him on material to be interchanged and that he was sending what the various geographical sections gave him. Taylor then went around to the geographical sections and found that they had no very precise idea of the problem and that the section heads were deciding what to send on their guess as to what might be wanted.
This situation needs clarification, and Denniston has said on several occasions that he did not like being unable to deal directly with Arlington on traffic exchange problems or on how exchanges should be handled. He also is very anxious to obtain good liaison with G-2 and has been going on the assumption that Taylor is to be quartered with him to function in that capacity. While there would be no point in trying to work out any revised traffic deal now, certainly not for us to try it, it appears to me that there is some value in pursuing discussions with Denniston along the above lines, so as to test his general ideas by specific cases, in order to carry back a fairly good idea of what sort of deals his authority permits him to make and what the general viewpoint here is on these various problems. His attitude, in my opinion, will permit all important intelligence problems along his alley to be solved satisfactorily, one way or another. Note also that Denniston, more than anybody else here, has turned his people over to us for questioning and given us a free run of his place.
TOP SECRET
ULTRA
Appendix 12
Denniston/Friedman Correspondence
In this correspondence, William Friedman questioned AGD’s memory over the Zimmermann Telegram and how it was transmitted.
On 4 May 1958 AGD claimed that ‘there were 4 (not 3) ways of transmission &, as you say, 2 bases’.1 Friedman wrote on 26 May, still uncertain about how many routes were used:
I note that you feel sure that there were four and not three ways of transmission. This would make it appear that our deduction with regards to the non-use of the Nauen-Sayville route is incorrect. Unless, of course, there is one additional route which we don’t mention.2
In June, Currier was able to give AGD a copy of the telegram as well as a new book dealing exclusively with it by Barbara Tuchman. Both AGD and Friedman subsequently questioned claims made in the book that code No. 13040 was found by Hall amongst papers belonging to Wilhelm Wassmuss, German Vice-Consul at Bushire on the Persian Gulf. As well as the James and Tuchman books, a BBC radio broadcast on 26 May talked about the Zimmermann telegram and gave a ‘very full & accurate details of 40 OB & the methods employed and even the names of the actual performers’. AGD assumed that James has advised on the programme, and both wondered how James had gotten clearance for his book. In July, Friedman and his wife attended a summer symposium for a specially-selected group of mathematicians, including two from GCHQ, and he has given talks on the telegram and other ‘certain classified matters’. In a lecture on 27 June 1958, Friedman had read from one of his letters from AGD.3
On 23 August, AGD sent his recollections about Room 40 to Friedman:
But do remember also the origins of ‘40 OB’ – a collection of amateurs with a good knowledge of German and no experience of cyphers collected by Sir A. Ewing in Aug 1914 to study the vast amount of W/T material which was coming into the Admiralty. Within a few weeks, NAVAL material was sorted out & the First Lord (Churchill) instructed us to make a profound study of the methods of the German Admiralty. We carried this out successfully & the staff grew & by the middle of 1915 we began to seek fresh fields where we could tackle the Germans. But we all had to learn the technical side of our job! Not easy work even for enthusiastic amateurs. Out of that small body & a similar party in the W.O. studying the German Army, & you know as well or better than I what has grown up from these sections!4
On 30 August, Friedman wrote to AGD and queried his assertion that the Zimmermann Telegram had been sent via the ‘main line’ Nauen-Sayville route. AGD replied on 5 September:
I know that we received the Z.T. by 4 routes, one by W/T Nauen-Sayville & three by cable of which one was procured by M. W/T interception was then in infancy & results often garbled. Cables were only available to us after delays as they were controlled by censors & not directly by us. I only remember that we did receive a lot of W/T traffic on this ‘main line’ route but I cannot confirm that we did hear the Z telegram accurately on this route.5
Friedman replied on 19 September:
You make a categorical statement that the ‘main line’, the Nauen-Sayville route was used. Permit me to say, for your information, that I am troubled by your statement, because in a good deal of research I find no evidence that this route was used. Whenever that radio channel was employed, the messages were carefully examined and decoded by means of a copy of the ‘Englischer Chiffre’ which was deposited with the State Department by Ambassador Bernstorff. I do not see how a message such as the Zimmermann Telegram would have escaped the censorship which was imposed on the radio route. I know, of course, that a radio route from Nauen to Mexico City was being established but there was so much difficulty in communication via that channel that it was not until months or perhaps years after the Zimmermann Telegram episode that it could be used in a practical manner. Of course, it is possible that Nauen may have transmitted the message in 13040 via that channel to Mexico City and this was what was intercepted in London. Do you remember anything to this effect? Could this serve as an explanation of your recollection that the radio route was used? Do you recall whether German Government official messages, transmitted via the ‘main line’, were at any time disguised by ‘phoney’ addresses and signatures intended to make it appear that they were strictly business messages between business men in Germany and in the United States.
I know that three other (cable) routes were used but it is important for my official story about the Zimmermann Telegram (a revision of the brochure you had and which you returned) to be accurate with regard to whether or not the radio route was actually used.6
AGD replied on 5 November:
As to the routing of the telegram, I said in the final paragraph of my long note of recollections that you were in a position to disprove any of my views if you have access to the records either in the State Department or in the Cable & Wireless companies though I greatly doubt if Sayville has preserved copies of all the traffic they received from Nauen. But I am surprised that you consider that all traffic on this route was in the Englisher Chiffre. You say that traffic between Bernstorff & Berlin was in 7500 & you thought that possibly Z Tel. was so sent. Now I make another suggestion which I will develop if you wish – if the German F.O. had the ‘nerve’ to use the good-will of your Ambassador in Berlin to send dispatches by his bag why should they hesitate to include ‘unreadable’ cyphers amid the telegrams or radiograms sent in the Englisher Chiffre.
Again my memory fails me but I think they did.7
Friedman sent his final letter on the subject to AGD on 29 December:
About the Zimmermann Telegram and whether it was sent by radio in addition to having been sent by cable over more than one route: I now have a Photostat of the records of that episode as they appear in documents of the German Foreign Office. A soon as I get an opportunity I’m going to read a long memorandum in those records, dealing specifically with how the telegram was sent, speculations as to how the plain text fell into American or British hands, and so on.
In answer to AGD’s ‘other suggestion’, Friedman goes on to say:
The answer to your question is that every message sent via the radio route Nauen-Sayville or Nauen-Tuckerton was carefully scrutinized by our communications censorship imposed long before the date the Zimmermann telegram was sent. No, I’m pretty sure the telegram couldn’t have gotten through that way. I think I know how the theory that the telegram was sent by radio came to be held but it would take too long to explain it in a letter. When and if we come to England next autumn I’ll hope to visit you again and tell you.8
Notes
Chapter 1: A life in signals intelligence
1. Attributed to a comment by Garnett Wolseley in discussion about a lecture by Major C.F. Beresford, ‘Tactics as Affected by Field Telegraphy’, Journal of the Royal United Service Institute, Vol. 31 (1887), p. 591.
2. Memorandum by G-2, GHQ, undated but c. 1945, ‘The Use of Ultra by the Army’, WO 208/3575.
3. National Army Museum (NAM), Leith-Ross papers, 8312-69, ‘The Strategical Side of 1(a)’, undated and unsigned but Spring 1919 according to internal evidence and presumably by Leith-Ross.
4. Intelligence Corps Museum (CM), Kirke papers, Accession Volume 58, notes for lecture by Kirke, 27 November 1925.
5. Friedman’s lectures were also published by the NSA for the first time in 1963 and reprinted in 1965 to help provide an authoritative history of the subject. The latest version was published in 2006 by the NSA’s Center for Cryptologic History – see bibliography.
6. James Thurloe had been granted control of both Inland and Continental postal offices through a so-called farming system at an annual rent of £10,000 per year in 1655 and retained control after his appointment. He was Secretary of the Council of State from 1652 to 1659 and MP for Ely from 1654 to 1655 and 1656 to 1658.
7. Isaac Dorislaus was the son of the ambassador to Holland and served in the post from 1653 to 1681 and possible until his death in 1688. He was paid £200 per year as Solicitor to the Admiralty from 1653 to 1660 and afterwards, £220 per year, issued to the Secretaries of State from the Post Office revenue.
8. John Wallis was born in 1616 and started his career in cryptography by deciphering the King’s dispatches for parliament in 1643. He was appointed Secretary to the Westminster Assembly in 1644 and Savilian Professor of Geometry at Oxford in 1649. Together with his work on conic sections, Wallis published the book on which his fame as a mathematician is based, Arithmetica infinitorum, in 1656.
9. Commons Debates lxxv (1844), 1291. Similar sentiments were expressed by the US Secretary of State, Henry Stimson, who closed down the US Army’s codebreaking section, the Black Chamber. Years later in his memoirs, Stimson made the frequently quoted comment, ‘Gentlemen do not read each other’s mail.’
10. In the latter part of WW2, Hitler’s cryptography experts wanted to replace the Enigma en masse with the more secure SG-41 cipher machine. However, by that time there were tens of thousands of Enigmas in service. In mid-1944, the German Supreme Command ordered a procurement of 11,000 SG-41 machines for the Armed Forces. Despite these large orders, only about 500 units of the SG-41 were delivered, mainly due to the production problems experienced during the last two years of the war and operators finding the machine difficult to use. See Klaus Schmeh, http://www.heise.de/tp/artikel/
11. ADM 137/4701, Intelligence E(C) to G.S.I. e., 10 October 1918.
12. See Ferris, ‘The British Army and Signals Intelligence in the Field During the First World War’.
13. The British Red Cross is part of the International Red Cross and Red Crescent Movement, the world’s largest independent humanitarian network. The Red Cross and Red Crescent Movement is made up of three parts: the International Committee of the Red Cross (ICRC); the International Federation of Red Cross and Red Crescent Societies; 190 National Red Cross and Red Crescent Societies around the world, including the British Red Cross.
14. He was awarded the Star of the Third Order of the Osmauli and the Star of the Third Order of the Medjidie and with the war medal for the Serbian campaign.
15. Merchiston Castle was the former home of John Napier, the inventor of logarithms.
16. James Alfred Ewing was born in Dundee in 1855 and went from Dundee High School on a scholarship to Edinburgh University in 1872. He studied engineering under Fleeming Jenkin, Professor of Engineering. Jenkin was remarkable for the versatility of his talent. Known to the world as the inventor of Telpherage (a transportation system in which cars [telphers] are suspended from cables and operated on electricity), he was an electrician and cable engineer of the first rank, a lucid lecturer, and a good linguist, a skilful critic, a writer of and actor in plays, and a clever sketcher. Jenkin sent Ewing to South America in 1874 to work on a cable to Montevideo and while doing this work, Ewing became interested in devising telegraphic codes. He returned to Edinburgh in 1876 to take his degree and two years later accepted the new Chair of Mechanical Engineering in the University of Tokyo. In 1887 he returned to Dundee as Professor of Engineering at the new University College and in 1890 went to Cambridge as Professor of Mechanism and Applied Mechanics.
17. Fisher was appointed First Sea Lord at the end of October 1914 even though he was in his seventies. He had previously served in the role from 21 October 1904 until 25 January 1910. He served until 15 May 1915 when he was replaced by Admiral Sir Henry Jackson who served until 30 November 1916. Jackson in turn was replaced by Admiral of the Fleet Sir John Jellicoe who served until 10 January 1918. His successor, Admiral of the Fleet Sir Rosslyn Wemyss, served until 1 November 1919.
18. By the end of 1919 the Admiralty had decided to abolish Osborne, take in boys at 13, give them three years at Dartmouth, then a year in a cruiser. The winding down of the college began at the end of 1920 and it finally closed on 20 May 1921.
19. See Partridge, p. 17.
20. The Osborne Lists are a record of staff at the College and some of these are held in the Archive of the Britannia Royal Naval College in Dartmouth. The Navy Lists are of naval personnel including those at educational establishments. The latter are also held in Dartmouth.
21. A selection of the Osborne College magazines are held in the Archive of the Britannia Royal Naval College in Dartmouth.
22. See Partridge, p. 53.
23. Custance Committee, Fourth Report, 20 September 2012, Inclosure 9, The National Archives (TNA) ADM116/1288.
24. Godfrey’s brother John would serve as DNI from 1939 to 1942.
25. A series of figures, numbers or letters which are used in the encryption/encipherment and decryption/decipherment of messages in a given cipher system.
26. A Watch usually consisted of a group of men (watchkeepers) tasked with translating and analysing incoming enemy traffic. The term originated in the Royal Navy system which linked a group of personnel to a period of time.
27. Treasury file T1/11937.
28. DENN 1/2, Churchill College Archive, Cambridge.
29. A group usually of three or more letters and/or figures, sent either in clear or in cipher, either in the preamble or in the body of a message, and serving to identify the sender and/or the recipient.
30. DENN 1/2, Churchill College Archive, Cambridge.
Chapter 2: British sigint in World War one
1. Two hundred and fifteen wireless stations have been documented in an English Heritage report as follows: eighty-seven coastal wireless (W/T) and/or intercep
t sites; twenty-two Royal Flying Corps (RFC) home defence sites; forty-one Royal Naval Air Service( RNAS) aerodrome sites; fifty-two RFC aerodrome sites; seven lightship sites; six experimental/portable sites. See First World War Wireless Stations in England. English Heritage Report, January 2015.
2. This is generally considered to be a fable of Churchill’s.
3. See Hammant.
4. All three German systems: the naval code proper (the three-letter Signalbuch der Kaiserlichen Marine); the four-letter Handelsverkehrsbuch, used by both naval and merchant vessels; a five-figure naval attaché cipher (the Verkehrsbuch) fell into British hands early in the war but their successors were broken through cryptanalysis. All code books were alphabetical until 1917 when the Signalbuch was replaced by the three-letter ‘hatted’ Flottenfunkspruchbuch (FFB). While other complications were introduced, Room 40 gradually mastered most German naval cryptographic methods. The main limitation was the volume of messages which could be intercepted at the Y stations at Hunstanton (HQ), Mercar in Aberdeen, Stockton-on-Tees and Cambridge.
