Hinsley’s account of Bletchley Park activities during the first year of the war is wrong in almost every detail, and I do not know why. I believe that he was originally commissioned to write the history as GCHQ wanted it written. He was not at BP himself in the early days, so he must have been told that the first bombes arrived in May 1940. I can only suggest that GCHQ wanted to conceal the methods by which we managed to keep on breaking from the German change of procedure in May 1940 to the actual arrival of the first bombes [i.e. the first production examples] many months later.
In any event, the development of the British bombe involved four new ideas that Welchman described in The Hut Six Story: loops derived from a crib, the double-ended Enigma scrambler (a column of three drums on the bombe which simulated three wheels on the Enigma machine), the diagonal board and taking advantage of the ‘filling up’ of the test-register.
Tadeusz Lisicki confirmed that this was the case:
None of the ideas which you listed were Rejewski’s. He was happy with the sheets and discarded the idea of improving his bomba. The loops, the double-ended Enigma, the diagonal board, and the filling up of the test register were all British ideas, and Rejewski in his letters to me several times mentioned that he had no idea how to mechanize the search for the keys and thought that the British mechanized the sheets, but that would be useless after May 1940. In Bruno he had absolutely no time for creative work. The running of a number of Enigma scramblers was the only idea which the Poles first used and perhaps was born from the Cyclometer.15
A later version of the Kozaczuk book included an article called ‘The Polish Success with Enigma in British Literature’ by Zdzisław Jan Kapera which was much more complimentary about Welchman’s book:
By giving in his book the first complete and accurate description of the significance of Enigma’s intricacies and of the extraordinary methods needed to break its signals he created a proper standard by which to evaluate the Polish breakthrough in 1932 and the subsequent successes until the end of the 1930s.16
Lisicki replied to one of Welchman’s letters in early September 1984. He had access to some unpublished Polish documents and papers written during the war by Colonel Guido Langer, head of the Polish Cipher Bureau in 1930–42, and Marian Rejewski. He also had over sixty letters from Rejewski written after the war. He said that Bruno (codename for the French cryptological unit during the campaign in France in 1940) had seventy cryptanalysts and staff working on codes and ciphers on 12 May 1940. He went on to say that:
How Bruno collaborated with Bletchley was agreed on a conference held in London on December 1939. From France came Braquenie and Langer. I do not know who was on the British side. The main points agreed were: interception, exchange of solved keys and decrypted messages. All this collaboration was super secret, Mcfarlan17 acted in Bruno as a cut-off and had a direct line (teleprinter) to BP; all messages passed through him and were enciphered on an Enigma.
By 18 September, Welchman had been able to read a paper sent to him by Professor Jean Stengers and published by Macmillan in The Missing Dimension in 1984. As already mentioned, Stenger had also sent him an earlier paper, in 1981, published in L’Histoire. Welchman had received this paper, along with Jozef Garlinski’s The Enigma War, too late to include details in his book. He had also read the English translation of Kozaczuk’s Enigma – How the German Machine Cipher Was Broken and How it was Read by the Allies in World War Two. This edition had been published in 1984 and in his review for Parameters, the Journal of the US Army War College, the cryptanalyst Cipher Deavours had written:
The book’s chief flaw consists of its notably anti-British approach. In particular, Welchman (whose book is quoted at excessive length) comes in for a lot of undeserved criticism. It is the thesis of the book that ‘virtually all major cryptologic techniques that the British used to break Enigma in the Second World War had been thought of and used by the Poles earlier’. This statement is simply not true.
In fact, the Poles never did break the actual Enigma machine, but only the ill-conceived German keying methods used with the device. This is, in fact, the true lesson to be learned from the book. It is generally the circumstances of use that determine the security of a cryptographic system and not the actual security inherent in the machinery itself. The use of electronics in cryptography has not changed this situation.
Not until September 1940 when the first British Bombes (codebreaking machines) started recovering daily Enigma keys using only short plaintext cribs was the Enigma obsolete in its then current form.
Without the Polish work, the British would likely never have gotten started in the first place, but once they did get started, British codebreaking was as dazzling as the earlier Polish accomplishments. The British bombes were in no way related to or derived from the earlier Polish bomby (codebreaking machines) nor were Polish methods of cryptanalysis particularly useful after the Germans changed to a better message keying system in May 1940.
At the end of October, Lisicki sent Welchman a list of Enigma keys broken at Bruno. He was also critical of the Kozaczuk book and noted that the translator was not a cryptanalyst and the censor was probably looking over his shoulder. By early December, Welchman was able to send Lisicki a breakdown of his new paper and he promised to ask Denniston to send him a copy of the preliminary draft. Lisicki duly received the manuscript in March 1985 and made a number of comments.
Welchman had heard from Dennis Babbage at the beginning of August 1982 (see Appendix 3) and was keen to speak to him and David Rees as both had worked with Knox in The Cottage. He said that he learned nothing from Knox himself but both Peter Twinn and Tony Kendrick had been helpful to him. Babbage passed on Harry Hinsley’s greetings and advice to get GCHQ approval before publishing anything else. Hinsley warned against relying on memory but Welchman’s view was that Hinsley may have been too much influenced by the history that Frank Birch wrote. In fact, he believed that Peter Twinn’s memory was quite likely to be more trustworthy than the Bletchley archives.
Welchman had still not heard from Rees or Parker. In desperation, he asked Fletcher to deliver letters to them personally. Parker had devised the Parkerismus method to show up repeats of keys and Rees was one of Welchman’s Hut 6 ‘wizards’. Alas, not all former BP colleagues were keen to speak about their wartime activities and Welchman never heard from either Parker or Rees.
Harold Fletcher was still willing to help where he could but, as he had not been well, he hoped that Houston Wallace could answer Welchman’s questions. Fletcher confirmed that, at the end of 1942, the overall organization of Hut 6 was still based on Welchman’s original proposals and that indeed the basic framework remained for the whole war. He also suggested that Peter Marychurch would be a good GCHQ contact if Welchman had further publication plans. This would prove to be a less than successful recommendation! Regretfully, Wallace confirmed that Fletcher was seriously ill with a growth on his pancreas and the outlook was gloomy.
Welchman was hoping that, if nothing else, former colleagues would be able to help him to pin down names and dates. He heard from John Herivel, who provided some helpful information.18 The difficulty with remembering names was brought home when Malcolm Chamberlain wrote to Welchman. Much to his embarrassment, Welchman had referred to him as John in The Hut Six Story. A possible explanation for this was that Chamberlain and Herivel were inseparable. They had both arrived at BP on 27 January 1940, two days after Welchman had recruited them from Cambridge. They had even attended Welchman’s tutorials together. Chamberlain confirmed that the ‘sheets’ were already in use by the ‘girls’ when he and Herivel arrived in Hut 6. A bit later, they had two days’ training in The Cottage.
In July 1984, Welchman received a letter from Jean Howard who, as Jean Arlington, had worked in Hut 3. She had been asked to research BP preparations for D-Day for a programme on the subject. She worked in 3L in Hut 3. One of her problems was that in trying to research both the ‘Y’ Service and its traffic analysis activit
ies, she had discovered that GCHQ had decreed that no book could be written about the ‘Y’ Service. Welchman was amazed by how much wrong information she had been given. She had heard one so-called expert who had worked in signals give a lecture at an Anglo/Yugoslav Symposium at the Imperial War Museum at which he said blandly ‘I suppose we just covered frequencies by luck.’ It seemed that all those in Hut 3 worked in ignorance of each other’s work and the programme never went ahead. Howard also quoted General A. L. Gadd who eventually became head of Sixta:
The control of interception was never properly managed, in my view, and was effective only because a few people like Gordon and, later, Oscar and 3L, studied the problems. There were both technical and security aspects which added to the difficulties. Hamish B-C [Blair-Cunynghame] and Neil Webster and Philip Lewis, whom you don’t mention, were the key figures on the T/A side. SIXTA was a belated effort to rationalise matters and was my particular baby.19
Webster had played a key role in liaising with Hut 6’s research sections, the Watch and Control. He was very much liked by all of his colleagues and, as Philip Lewis remembered in a letter to Welchman:
To your question about Neil Webster who was appreciably older than I was. He was indeed one of my officers – a supremely brilliant evaluator living an almost trance-like existence – totally unmilitary but immensely capable. He seemed at times utterly unaware of his surroundings and we would find him in profound and totally immobile meditation halfway up the stairs working out the significance of some message we had received. His thinking was superb and we owed a great deal to him.
Like Welchman, Webster had been scrupulous about not speaking about the war but when the books about BP and Ultra started to appear, he followed developments closely. After he retired in 1976 he began working on a book about his intelligence activities as he felt existing books were short on crucial elements of his work. For example, Welchman had not said much about re-encryption work in his book. As Webster recalled:
It was my job to see that wireless intelligence gave the cryptographers all possible help in finding cribs … this aspect has hardly been mentioned in accounts so far … Yet Bletchley’s rapid and regular breaking [of Enigma], a major factor in Allied victory, was based on cribs, and accounts which omit this are simply ‘Hamlet without the prince’.
In September 1980 Webster received formal written permission from GCHQ to publish his draft text. Having lined up a publisher and sent drafts to Ronald Lewin, Ralph Bennett and Jean Howard, he suddenly learned that his permission to publish had been withdrawn by the Director of GCHQ, Sir Peter Marychurch. Webster consulted a solicitor and wrote to GCHQ in protest but in August was told that it should not have been cleared and that no amount of editing would make it publishable. He then withdrew graciously and even turned over copyright, voluntarily, to GCHQ. He received a cash payment to cover expenses, disappointment and loss of face with his publisher.20
Between 5 August 1984 and 26 February 1985, Welchman had an exchange of letters with Peter Twinn, the first mathematician to be recruited by GC&CS and, arguably, the first man to break a message encrypted by an Enigma machine which included a plugboard. Twinn was also the only living member of Knox’s team from the period leading up to the war before the arrival of GC&CS at BP. Welchman was keen to get Twinn’s take on Knox and establish once and for all in his paper what the Poles did and the extent of Knox’s contribution
As Peter Twinn would not publicly comment on his time at BP until 1993, and even then in the most general of terms, some of his comments to Welchman in a letter dated 25 September 1984 appear below verbatim:
How was the liaison with Vignolles handled? Very cloak & dagger. Clearly Dilly was involved but he never told me much about it. He was extremely secretive even with me & I was for some time of course before the war, his only assistant!
General discussion of how to cope with Enigma was widespread in the then tiny circle of people concerned (except for Dilly who was always aloof). Indeed I think all the writings about Enigma tend to try to attribute the ideas to specific people in a way that hardly represents the atmosphere correctly. So much was discussed that I think it is not at all easy in all cases, to say where ideas sprung from – often I think they developed in general conversation until someone crystallized them in final form.
I am sorry that I cannot see eye to eye with you on the subject of Denniston. I do not regard him as a success. I think he failed between the wars to get GCHQ (GCCS as it was) the status & facilities it needed. And he was on a pretty good wicket in the years just before 1939. The organization was having stunning success in reading the Spanish Civil War codes & had clearly demonstrated its potential at trifling expenditure. Denniston’s posting to the U.S.A. was clearly demotion. Indeed Denniston said to one of my colleagues, when his posting was arranged & Travis took over ‘I am not jealous of Travis – what grieves me is the realization that I didn’t prove man enough for the job.’ These are my own views – naturally not for publication & on no account whatsoever to get back to Robin Denniston.
My own impression is that the drive to set up BP came much more from people like Tiltman & Menzies.
In response to Welchman’s question about whether Dilly and the others would have been able to read Enigma traffic without the Poles if they had guessed the connection pattern between keyboard and entry drum – ‘Yes indeed – a simple guess. But not only Dilly & I failed to make it – so did Kendrick & Turing – neither of them remarkable for lack of imagination!’
Also in September 1984, Welchman received a rather different assessment of Alastair Denniston than that of Twinn’s harsh account. Thomas Parrish was an author who had interviewed him for a book he was writing on the American involvement in Ultra. Parrish attached a letter from William Friedman to Denniston’s daughter following her father’s death. The letter could not have been more effusive in praising Denniston’s contribution. A further testimonial came from Frank Rowlett in a telephone call to Welchman. Rowlett was the first of William Friedman’s original employees, hired for the Army’s Signal Intelligence Service in 1930. As well as their vital work on Purple, he and Friedman also helped design the Sigaba cipher machine, which was never solved by the Germans during the war. As Welchman wrote to Robin Denniston:
[Rowlett] well remembers your father’s visit before America entered the war and says that both he and Friedman very much admired him both as a person and as an outstanding cryptologist. He also says that the impression your father made on the Army cryptological organization undoubtedly helped to establish the close relations with Bletchley Park that were to develop later.
I asked him about relations with Travis. He said that he and Friedman, and others at Arlington, appreciated Travis as a sincere person whose word could be trusted. But, because he had no feel for cryptology, Travis never won their affection and admiration to the extent that your father did.
On 28 January 1985 Welchman sent a copy of his paper to Twinn to make sure he was happy with it. Twinn made the following further comments in a letter dated 25 February 1985:
As late as July 1939, GC&CS could not see how to capture the Enigma problem with the information they possessed. An inspired guess as it turned out, would have made the available evidence adequate for a solution.
I can remember reading in the Cottage an account of the construction of the Polish Bomba. That Hinsley was ignorant of this is irrelevant. He was, naturally as it wasn’t his job, ignorant of all the technical problems.
As regards the question of whether Dilly gave us a complete picture of what the Poles had achieved, I think almost certainly he didn’t because everything paled into insignificance in comparison with the knowledge that the diagonal was A, B, C, D, … But we all appreciated the Polish achievement. We did because knowledge of what they had done arrived at Bletchley as the war progressed (by way of the French). Remember that we had the knowledge about the diagonal from the Poles just before the war began & Poland was overrun. Information about the su
ccess the Poles achieved came later.
*
Meanwhile, the news from McGraw-Hill was that they intended to declare The Hut Six Story out of print and dump an unsold 1,125 copies on the remainder market. Up to 1985, Welchman had received advances totalling $6,855.74, an amount which author royalty earnings barely covered. Penguin accounts also made depressing reading with royalties just paying off Welchman’s initial advance. The untold wealth forecast by Winterbotham had certainly not materialized. And there was also to be one final twist in his ongoing battle with the NSA as he related to Denniston on 5 September 1984:
Another helpful fact cropped up in my long talk with Frank Rowlett yesterday morning. He told me about the troubles he has been having with Mr MJ Levin, Chief of Information Security Division, NSA. Apparently Rowlett had been asked by the NSA historians to write an article that in their view could be published. But Levin maintains that it must be classified ‘confidential’ and not published. Rowlett challenged this and attended a meeting in Washington, I think only a few days later. At the meeting, the subject of The Hut Six Story came up, and Levin said that I had really gone too far. At this Rowlett blew up. He told Levin that in his opinion nothing that I said in my book could do any harm today, and that it is shocking that a man to whom both Britain and the USA owe so much, should have been so badly treated. From what he said to me, Rowlett thinks that Levin is the principal, if not the only source of my troubles with NSA.
In early November, Welchman received a letter from Christopher Andrew, Senior Tutor at Corpus Christi College, Cambridge, suggesting that his article might appear in the first issue of a new publication, Intelligence and National Security, of which he was editor. Welchman was encouraged that Andrew and David Dilks had been able to publish Stengers’s article in The Missing Dimension. By now he had sent final corrections to Denniston, who was acting as an intermediary, for his ‘From Polish Bomba to British Bombe’ paper. He also liked one of the new journal’s stated objectives: ‘While recognizing that current intelligence operations usually require secrecy as a condition of success, Intelligence and National Security will attempt to lift the veils which still pointlessly conceal the past history of intelligence.’
Gordon Welchman Page 25