Alastair Denniston
Page 22
Before their departure, the Americans had been briefed on the importance of security in all aspects of their engagement with the British:
All members of the Delegation were asked to pledge themselves to secrecy to the extent that the secret should only be told in the case of the Navy Department to the Director of their Cryptographic Bureau and the case of the War Department to their DMI, Director of Sigs., and their Chief Cryptographer, and moreover we disclosed them only our cryptographic methods of solution and asked them not to request to be shown the resulting intelligence.70
After arriving at BP, the American party was taken to AGD’s office, where they were greeted by him along with Travis and Tiltman. AGD’s personal assistant, Barbara Abernethy had been tasked with providing sherry for the guests and she managed to carry a large wooden cask from the Army and Navy Stores to the Mansion House at a spot adjacent to AGD’s office,71 from which she duly served sherry to all assembled there. They were then driven to their billet. According to Currier: ‘Now our billet was the country home of the chairman of the board of the, at the time, Anglo-Persian Oil Company, an extremely wealthy man. He had turned over his entire house to these four travelling Americans, fully staffed, the larder full, with a butler and three upstairs maids and a cook to take care of us for the time that we were there.’
The Navy and Army groups were split up and never travelled together while visiting BP’s outstations or parts of BP itself. Currier and Weeks spent much of their time in the Naval Section and visited naval intercept stations such as Scarborough. All four were given details of the Enigma machine and shown the Bombe machines in operation.72 On 3 March, AGD informed Menzies that, ‘Our American colleagues have been informed of the progress made on the Enigma machine.’ Weeks wrote to AGD to assure him that all information about the Bombe machines would be treated with the utmost secrecy:
For: Commander Denniston
3/3/41
We are in accord with the purport of your memorandum of today.
We undertake to vary out all instructions for the preservation of the secrecy of the work mentioned, informing by word of mouth only the head of our section, Commander L.F. Safford, USN.
In connection with the naval aspects of the above work we deem it advisable to obtain the wiring of interest to us (i.e. the device on which Turing is working), and to disclose that only when it is decided to work on the problem. In such an event we shall observe all precautions and keep you informed of our actions.
As far as is practicable we shall make arrangements for the forwarding of communications through our naval attaché, as mentioned by you.
Respectfully,
R.H. Weeks’
The Americans boarded HMS Revenge at Helensborough in the Clyde several weeks later for the journey home. On 19 March, Weeks wrote to AGD with a list of the materials that they had received from GC&CS.73
AGD had overseen the visit by the Americans despite serious health problems. On 27 February, he was X-rayed by a consultant called Shanks at 68 Harley Street and a gall stone was found in his bladder. Yet two days later, he invited the Americans to lunch where his family were billeted at Stapleford Mill Farm. On 10 March, he went for a further consultation in Harley Street and subsequently underwent surgery at Luton General Hospital on 13 March. He was not discharged until 14 April but shortly afterwards he had an attack of orchitis and he was again hospitalised on 20 April, this time at Ashridge Military Hospital at Ashridge Park, Berkhampstead, Hertfordshire. He did not return to BP until 27 May by which time his hospital bills had taken up a third of his monthly salary of £98.
One of the outcomes of the visit of the Americans to BP was that AGD agreed to visit the US and Canada to discuss further collaboration between the three countries. He cabled Washington on 28 July, informing them of his plans, which were to travel from Scotland on 11 August using American transport. He hoped to arrive in Washington on the 12th or 13th and following meetings there and in New York, he planned to visit Ottawa.74 AGD duly departed for the US as planned and after stopping briefly in Newfoundland on the 12th, he eventually arrived in Washington. On 14 August 1941, Travis telephoned the Denniston house (telephone number – Soulbury 7) to see if Mrs Denniston was okay. He subsequently cabled AGD that same night with a message that all was okay at Soulbury 7.
AGD met with William Friedman and members of his SIS team75 in Room 3341 of the Munitions Building on 16 August at 9.00 am. The purpose of the meeting was to prepare a schedule for his visits to the various sections of SIS from 19 to 22 August. Apart from Friedman’s own team, which consisted of Sinkov, Rosen, Rowlett and Kullback, also present were SIS Executives Lieutenant Colonel R.W. Winckler, Captain H.G. Hayes and Captain E.F. Cook. The schedule that was agreed included visits to SIS’s German, Italian, Central and South American, Japanese Diplomatic and Army Sections. The final day would be spent visiting SIS’s tabulating and machinery, intercept and school sections. AGD must have been reassured that SIS was keen to cooperate fully with him. During a three-hour meeting, he reciprocated by giving a brief report on the probable status of cryptanalytic work engaged in by certain other governments, in so far as the facts were known to him. He also reported that cooperation between GC&CS and the recently established Canadian cryptanalytic section in Ottawa, the Examination Unit, was wholly dependent on the removal of Herbert Yardley from that organisation. Yardley had written a sensational exposé of American and British cryptography in WW1, The American Black Chamber, in 1931. Despite Yardley’s endorsement by General Mauborgne, the US Army’s chief signals officer, his views were not shared by others officials in US intelligence services.76 They also agreed on a more safe and direct forwarding and exchange of documents.
AGD hoped to convince Friedman that there were no spare Bombes to give to the Americans and that he should not waste valuable resources by commissioning one from IBM. He also wanted to persuade SIS that it should leave efforts against European military work to the British. When the American Army began to be involved directly in the European theatre, the British would reveal more about its methods and supply information where appropriate. AGD also offered to have some of SIS’s mathematicians visit BP, despite his security concerns. One of AGD’s key objectives was to persuade Friedman’s team to concentrate on Japanese systems, given that he assumed they had vast technological resources to deploy against them. When he toured SIS facilities, however, he was disappointed to see how little space and equipment they actually had. But Friedman was receptive to AGD’s advice, and accepted that a British Bombe was not available. Furthermore, the meetings between AGD and Friedman established what would become a lifelong friendship. In a letter to AGD’s son Robin years later, Friedman’s close colleague in SIS, Frank Rowlett, said of the visit: ‘I remember well your father’s visit to Washington in 1941. It was his visit which laid the foundations for the collaboration between the cryptographic activities of the US and the UK which produced intelligence vital to the successful prosecution of WW2. We spent considerable time together discussing the technical activities undertaken by both countries and worked out some of the details of our early collaborative efforts.’77
AGD’s diary shows that he dined with senior staff of the US Navy’s OP-20-G group at 20.30 on 18 August and SIS on the 19th. While his dinner with OP-20-G was cordial, his reception when he arrived at their offices was anything but. He was shown around the OP-20-G offices by Currier before meeting with the leaders of the naval communications section. AGD was agreeable to providing more information about French and Italian systems and improving the security of naval communications with their British counterparts.78 Several days later he had a more difficult meeting with OP-20-G’s head Laurance Safford who had a history of Anglophobia. In 1937, the State Department had requested approval for a British engineer to spend four months with Bell Telephone Laboratories and the Radio Corporation of New York. Safford’s reply on behalf of OP-20-G was: ‘The Director of Communications can raise no specific objections to Mr. Gee’s ex
tended visit to the Bell Laboratories. However, it is considered an undesirable practice for this country to disclose its technical secrets to foreigners while getting nothing in return.’79 Despite the fact that the American Navy was virtually on a war footing in the Atlantic, AGD urged Safford to restrict his team’s efforts to nothing more than research when it came to European naval systems.80
When AGD met one of OP-20-G’s top codebreakers, Agnes Driscoll,81 it was clear that she was privy to some of the information that Currier and Weeks had brought back from their visit to Britain. It had been agreed that only Safford would have sight of British cryptographic secrets. One can only imagine AGD’s thoughts when he met Driscoll, aged 53 and a 20-year veteran of cryptanalysis. Her attitude and approach to cryptanalysis bore startling similarities to that of Dilly Knox! She rejected BP’s automated solutions based on mathematics and claimed to have developed a much better hand method based on intuition. Driscoll told him that she was evolving a method which would enable her to solve keys on a small amount of traffic. She went on to say that the American Navy did not want or need a Bombe machine and thought little of the British methods for attacking Enigma. She showed AGD a sample solution based on a short eight-letter crib and claimed that it would require only two dozen people a few days to find the Enigma settings using her soon-to-be completed ‘catalogue’. AGD maintained his composure despite the insulting nature of Driscoll’s comments. She could well have known of British successes against Naval Enigma in 1941. He offered to provide more information about the Bombe machines and Turing’s methods and even offered to supply OP-20-G with a machine when one became available. Furthermore he invited her to BP, but a serious automobile accident she had suffered in 1937 prevented her from travelling long distances.82 However, she did not offer to send a member of her team in her place, nor did she show any interest in a visit from a British expert.83 AGD explained to Driscoll that the British and the Poles before them had explored the catalogue approach which she was proposing to use. It had had limited success against earlier versions of the Enigma system but would be of little value against the more formidable German naval Enigma system.84 Finally, after further discussion, Driscoll conceded that she did not fully understand the Enigma system and she rather forcefully demanded that BP provide further information. AGD asked her to provide a list of questions which he would take back to his colleagues and he promised to provide answers to all of them. AGD, with only a limited knowledge of the Enigma system, must have been startled by how basic some of the questions were. She was clearly after just enough information to pursue her own methods.
AGD returned to England on 23August. The family had had no news of him and as the Battle of the Atlantic was in full swing, all they knew was that he had flown rather than gone by sea. BP was worried as was the family. By 8 pm on the 23rd they were waiting, worrying and wondering. To quote his son Robin:
There was a heart-stopping moment when we heard the crunch of car tyres on the farm track. Could it be? It was too much to hope. But it was. It had taken him 15 hours flying from New York to Gander, Newfoundland, across the North Atlantic to Prestwick in the bomb bay and thence to Hendon where he was collected and driven home. It was the best moment of all our lives. He had brought us presents, for me a green pullover with black and white borders. He had had an amazing week but he could not tell us anything about it. That did not matter, what mattered was that he was alive and okay.
AGD’s health was still causing him problems and his diary for 26 August records ‘neuritis returns’, but around 30 August he was on his way back to North America by air. He flew to Washington and then on to Ottawa the following day to brief the Canadians about British cryptographic achievements against Japanese codes. In November 1939, Canada had offered to help the Allied intelligence effort by setting up its own cryptanalysis unit, specifically to attack German, Italian and Spanish traffic.85. AGD had told the Canadians that to get into the business of breaking ciphers, they needed a ‘high-grade cryptographer of long experience’ who would need at least three months of training in Britain.86 They took his advice but, surprisingly, hired Yardley to set up and run their new cryptographic department, the Examination Unit. AGD arrived in Ottawa on 3 September to meet Canadian cryptographers and he was keen to press home the point that Britain wanted the US and Canada to concentrate on Japanese military ciphers while Britain looked after diplomatic material.87 Having ensured that he would not have to meet Yardley, AGD told Lester Pearson, assistant to Norman Robertson, acting Undersecretary of State for External Affairs, that Yardley had to go or there would be no cooperation between Canada and Britain. AGD offered to provide one of Britain’s best experts to replace Yardley, Oliver Strachey, a veteran of MI1(b) during WW1 and a member of GC&CS since its inception. Under pressure from Washington, Yardley’s contract was not renewed and Strachey arrived in early 1942, having refused to go to Ottawa until Yardley left the city. He brought with him his personal assistant, Miss Rogerson and keys to high-level Vichy and Japanese diplomatic codes, which initiated close cooperation with Washington and London. Although he did not speak or read Japanese, he helped break the Japanese code, which was very complex, since it used variations of kanji, hiragana and Romanization
AGD flew back from Ottawa via Gander. He arrived in Ayr, Scotland, at 11.30 am on 13 September, went on to Hendon and then home to Bletchley. As soon as he arrived back in Britain, AGD sought from his experts the answers to Driscoll’s questions. He also established detailed procedures to register and track all communications with OP-20-G and SIS. By October, GC&CS had provided answers to all of OP-20-G’s questions and also dispatched the specifications of Enigma settings, details of wheel wirings and all relevant intercepts for 1941.88 AGD also made enquiries about obtaining an Enigma machine for Driscoll and he then waited for her to honour her commitment to provide GC&CS with a description of her methods. While GC&CS cryptanalysts were extremely sceptical about these methods, AGD’s view was that ‘We are on a good wicket at present but can’t afford to neglect any side lines.’89 In the end, it was left to Turing to write a scathing critique in which he concluded that even if the majority of the settings of an Enigma machine were known, Driscoll’s method would take far too long to find a solution.90
In mid-December 1941, Driscoll finally sent partial details of her special method to AGD. By now, GC&CS had concluded that her method had failed and AGD was reluctant to provide her with further information. Perhaps to hide its own failures, the US Navy suddenly claimed that AGD had never sent the material requested by Driscoll. Safford’s superior, Leigh Noyes, sent a series of critical telegrams and seemed to support Driscoll and the power of her methods. AGD responded by saying that he had indeed provided the information and if the US had not received it, why had it taken so long to report this to GC&CS. Suddenly, in the days after the Japanese attack on Pearl Harbor, Noyes apologised and miraculously, the missing documents from AGD were discovered in the OP-20-G offices. However, friction between British and US intelligence agencies persisted and eventually reached the ear of higher authorities on both sides of the Atlantic. This would certainly play a part in events which would unfold at BP in the early days of 1942. Similarly, by February 1942, Safford had been replaced by Joseph Wenger and a new breed of US cryptanalysts began working in OP-20-G. It seemed that Anglo-American cooperation would at last be restored.91
***
***
Throughout 1941, administrative matters continued to occupy much of AGD’s time. At the first meeting of the JCC on 6 April, Captain Ridley stated that GC&CS had asked for £500 for two full-size hard tennis courts on existing grass courts.92 On 12 April, the JCC considered and approved a proposal from AGD to build a cafeteria to feed 1,000 at one sitting.93 On 7 June, AGD was reporting to the JCC that:
Owing to the course of events in the Mediterranean we have been asked to combine the German and Italian Naval Sections. Need has also arisen for a combined German-Italian Research Intelligence party. Birch to take
charge of the Section and Clarke Research Intelligence. It will not be possible to amalgamate fully or to the best advantage until the new premises are ready.94
Accommodation issues continued to be a pressing concern and on 4 October, AGD commented at a meeting of the JCC:
Priority must be given to the accommodation of the staff at Elmers and Wavendon within the perimeter. Nothing can happen until the new large huts (Naval and Military) are complete. Progress on these appears to be very slow. Elmers School are at present roughly 120 strong and Wavendon 60. These will therefore occupy at least Huts 10 and 5, leaving 4 and 9 vacant. When the Dining Room is completed two large rooms in the house will be available and suitable for offices. Is it possible that the Recreation Hut could be linked up with Hut 9 and that this joint building might be used for various amenities such as an enlarged coffee room which could be used as the general common room, a Senior common room, quiet rooms, etc?95
Interestingly, Travis, who has always been characterised as the ‘hard man’ at BP, responded to the suggestion that the rooms in the Mansion should be used for offices as follows: ‘This has never been the view up to now and, although conversation on the subject has been desultory and indecisive, it has been generally hoped and assumed that those large and commodious rooms, which are not really suitable for offices, will be used for the social amenities of our community.’96