by Alastair Denniston- Code-breaking From Room 40 to Berkeley Street
In 1939 (and for some years before) the Army at their Chatham Y Station were intercepting traffic from a very large group of stations in Germany, which were thought to be serving the German Army.
In January 1940, thanks mainly to the S.I.S. and their association with the G.C. & C.S. , this traffic became legible and turned out to be point to point station traffic of the German Air Force. From that moment until fairly recently the Army (MI8) have endeavoured to give up taking this traffic, well knowing that there is no one else to take it and despite the fact that during Spring and Summer it provided Army intelligence of the highest value in Norway and France, but because it emanated from the G.A.F. MI8 have tried to drop it.
The R.A.F. had always maintained that nothing of operational value would come out of point to point traffic and therefore, had made no plans for taking it and did not commence doing so until and are not even now in a position to take it over completely.
In April, it soon became apparent that the Army had insufficient sets and operators to cover the traffic satisfactorily and 18 sets were eventually diverted to this work at the Foreign Office Y stations at Sandridge and Denmark Hill, whose proper job is the interception of traffic in commercial routes. Without these additional sets much less of the intelligence would have been obtained.
I maintain, therefore, that if it had not been for our very close association with S.I.S. in the first place, and the fact that G.C. & C.S. were able to pool available W/T and cryptographic resources, much of this very valuable intelligence would have been lost and would not now be obtainable.
This example of the benefits of unification of cryptography in one centre is just additional argument to the many that can be advanced for the advantages cryptographers desire from working in the closest association.39
GC&CS continued to grow and by the beginning of 1941, numbers had more than trebled to around 685. The teams working on Enigma had also grown quickly and Huts 6 and 3, which worked on German Army and Air Force Enigma, had ninety-three and sixty staff respectively. Huts 8 and 4 worked on German Naval Enigma and had thirty-seven and forty staff respectively. Knox’s Research Section in The Cottage numbered nine staff! Responsibility for recruiting additional scientists and mathematicians was given to C.P. Snow of Christ’s College, Cambridge.
***
AGD had moved his family into Stapleford Mill Farm near Bletchley on 13 September 1939, and they would remain there for two years. His son Robin was sent to boarding school at Downsend School between Ashtead and Leatherhead. They regularly entertained friends and GC&CS colleagues and de Grey, Adcock and Birch were frequent visitors. The Dennistons would visit the Knoxs’ at their home at Courns Wood, a large ten-bedroom house set in ten acres of private woodland near High Wycombe. Dorothy got on well with Knox’s wife Olive but the two men were not close friends. Their battles over control of the Enigma problem at BP had no doubt taken its toll. De Grey would usually appear at Christmas with presents for Robin and ‘Y’. These would inevitably be tokens to buy framed Medici prints as he had run the Medici gallery before the war. AGD’s circle of friends included Rhoda Welsford whose mother rented Newton Longueville manor house, a mile from BP for the duration of the war. Rhoda worked in the Air Section from September 1939 until November 1944. Christmas 1940 found the family at the farm with a gardener and a cook. Christmas dinner was at the Welsfords at Newton Longueville. The Manor housed ‘the Profs’ – ERP Vincent, Frank Adcock, Tom Boase and Hugh Last from Kings, Magdalen and Brasenose respectively. All took part in paper games with the Denniston children. Tiltman was also a close friend of AGD, as were their respective wives. On 12 June 1941, AGD was appointed a Companion of the Order of St Michael and St George (CMG), as with his previous honour, in his capacity as head of a department in the Foreign Office.
Ernst Fetterlein and his wife shared at least two Christmas dinners with the Dennistons. Conversation was difficult, as Mrs Fetterlein did not speak English and her husband’s only common interest with AGD was work. Robin Denniston remembers Mrs Fetterlein being large and Mr Fetterlein small, formally dressed with boots and a monocle. They always came bearing gifts for the Denniston children.
Ever the frugal Scotsman, AGD kept immaculate records of his finances throughout the year and compiled a month ledger. October showed £155-9-5 received and £155-10-5 paid. November showed his account balanced at £80-7-9.
***
By the beginning of 1941, GC&CS had developed functions other than cryptanalysis and some of these seemed more appropriate for the Signals or Intelligence Departments of the Ministries. At the same time, the War Office and Air Ministry virtually controlled the activities of their respective Service Sections at GC&CS, while the CSS had control of policy over the output of Enigma cryptanalysis. Menzies presided over the Y Board and in effect sat between GC&CS and the Services. In November 1940, Colonel Butler of MI8 approached AGD about moving the seventy members of CIS to BP. The driver for this was to place W/T intelligence (WTI) and cryptography work in one location.40 AGD’s reluctance to take them all was based simply on a lack of space but this was seen as a lack of vision on his part as well as administration inefficiency. It was thought that AGD wanted to preserve GC&CS as a purely cryptanalytical bureau and stop it evolving into a Sigint Centre. The prevailing view of some such as Group-Captain Blandy was: ‘I think it is perhaps as well to emphasize that Cryptographers are not intelligence officers, but only exist for providing the material from which Intelligence is produced, and it is as well to keep the intelligence side as far divorced from the cryptanalytical side as possible.’41
The Admiralty was quite content with the arrangements at GC&CS, and Captain Sandwith wrote that naval ‘cryptography and W.T.I. are both carried out at B.P.’.42 As far as the War Office was concerned, there should be greater Services control of GC&CS and the Y stations. GC&CS’s output had ‘greatly exceeded our expectations and pre-war forecasts’, and Travis went on to say: ‘From 1919-39, the interception services were very definitely run by G.C. & C.S., and it is only since the appointment of senior officers to A.I.1 E and MI8 on outbreak of war that any change has been desired, it is now urged that cryptography is subservient to interception. It is quite obvious that cryptographers will always know more of interception than the interceptors can possibly know about cryptography’.43 The issue came down to whether or not Y included cryptanalysis. DMI and D’s of I said it did, DNI and CSS said not. In January 1941, the latter view prevailed at a meeting of the Y Board: ‘The Board discussed at length the definition of “Y” Services and it was ultimately agreed that “Y” indicated interception and the development of all means of interception which might produce intelligence but did not include cryptography and its fruits.’44
There still remained the task of defining the relationship between Y and cryptography. On 2 July 1941, a sub-committee of the Y Committee reviewed the issue and concluded that:
There do not exist two mutually exclusive subjects that can be defined by these names. The two are so inter-related if not very consistent or even, perhaps, legally sound, a compromise which may be expressed (necessarily loosely) thus:
The ‘B.P. Committee’ was set up and had its first meeting on 25 February 1941, with AGD’s objective being ‘to bring forward matters concerned with cryptography and anything that might arise there from’.45 It quickly turned into a GC&CS all-purpose committee and petered out after a few meetings.
At the suggestion of Group Captain Blandy, the Y Committee appointed a special ‘E’ Committee on 12 March, consisting of military and air force members and Y station commanders with Travis as chair. They met at BP so Travis could pull in staff as required.46 It confined its work to the requirements for Army and Air Force Enigma traffic and, at its final meeting on 18 November 1941, decided upon an ‘E’ interception plan for the future, with interception carried out on an Inter-Service basis around four stations – Beaumanor (Army), Chicksands (Air Force) with Whitchurch and Harpenden in support.47 This was approved by
the Y Committee several weeks later.
Y Interception was controlled by a structure overseen by the Chiefs of Staff. AGD sat on the Y Board chaired by CSS and Travis sat on the Y Committee reporting to it, chaired by Air Marshall Sir Philip Joubert. It in turn had Sub-Committees: Technical, Administrative, ‘E’ and a Noise Investigation Bureau for non-Morse traffic. MI8 still wanted some kind of ‘Operational Intelligence Section’, but unlike the Admiralty that had OIC, it, unlike MI8, was a Command. All military units engaged on aspects of army-air WTI work were combined into a single group on 25 March 1941 called Intelligence School No. VI or 6 I.S. The exception to this was a group called the Special Liaison Party (SLP) which worked in Hut 3 at GC&CS. 6 I.S. eventually moved to Beaumanor in July along with the War Office Y Group (WOYG) in October. This was compatible with the Board’s official ruling that Y did not include cryptography and that the two were indivisible was acknowledged by the appointment of a cryptanalyst from 4 I.S. to command 6 I.S. An MI8 officer was sent to GC&CS to represent MI8(a), the intelligence branch of MI8, and to act as ‘the channel through which intelligence from traffic analysis reached the War Office’. He produced a weekly report of the ‘changes in the German W/T picture as shown through decodes, and aiming at adding to the fund of intelligence produced by G.C. & C.S’.48
AGD had recommended the creation of an ‘Interservice Distribution and Reference [D & R] Office’ at BP to counter Service dissatisfaction with the apparent lack of uniformity and standardisation in the reporting of Service Sigint products. While diplomatic sections had a uniform and tidy routine through a D & R, naval, military and air Sigint was subdivided among a large number of units, Service, inter-service and partially inter-service with separate outlets to different customers. A Service D & R with a civil head and Service intelligence officers attached to it would ‘assure the Directors that their Intelligence Divisions were fully represented at the centre and their interests in the circulation of the results completely covered’.49
The proposal was approved by DMI50 and DNI51 and then by the Y Board on 14 February. However, opposing factions within the Ministries and GC&CS caused the proposal to collapse. Part of the problem was that Army and Air Force decrypts were disguised as agents’ reports, with all indication of wireless origin removed. According to de Grey ‘It would have been difficult to contrive a system more likely to prevent signal intelligence from ever being of the fullest use to its recipients.’52 The system had worked for Diplomatic Sections because they had little interest in the problems of interception or W/T in general.
In April 1941, de Grey suggested a D & R with two main functions:
To receive the cryptographic output of the various sections of G.C. & C.S. and ensure that such intelligence as they produced was passed to other sections who might be interested, with a view to aiding cryptography;
To ensure that all information likely to interest the three Service Ministries was passed to the Intelligence Officer of the Service Sections of G.C. & C.S., so that such information may reach the Ministries in as intelligible a form as possible.53
AGD reported to the JCC on 14 July that he had set up a new section called the Intelligence Exchange (IE). It was to be headed by de Grey and include an officer from each service. Its remit was to receive and analyse the complete daily output from each Section at BP and include Diplomatic, Naval, Military, Air, Hut 3 (Joint), Russian, French Naval, ISOS (Intelligence Services, Oliver Strachey) and Commercial. They would then be in a position to see that all the Ministries received what might interest them. The exception to this would be Hut 3 as it already served the three Service Ministries and there was no question of reissuing any of its product. This was not the ‘D & R’ that the Service Ministries were expecting and,
only the Admiralty complied, sending an officer from N.I.D., who had no familiarity with and no adherence to Naval Section or to I.E. He proceeded to institute direct reporting from I.E. to N.I.D., so upsetting the principle for which I.E. was then fighting, namely, the unification of the channels of reporting, and the principle for which Naval Section was contending, that Naval Section should be the only channel to the Admiralty. His attachment was of short duration.54
The appointment could be seen as an attempt by DNI to safeguard his interests at GC&CS, as unlike the Army and Air Force Section’s subordination to their respective Ministries, Naval Section was not subordinated to NID. In any event, DNI had already appointed Captain J.R.S. Haines, RN, as his liaison officer with the title Assistant Director of the Operational Intelligence Centre (ADIC), with the following remit:
He will visit B.P. frequently to facilitate cooperation between N.I.D. and B.P. and will be responsible to D.N.I. for the coordination of the results of the work of the Naval Section at B.P. and for the action taken by N.I.D. on the material provided by B.P.55
This arrangement, according to Birch, ‘enduring for the rest of the war and proved an unqualified success’. I.E. became a ‘comprehensive up to date library of G.C. & C.S. output’ with Naval Section remaining ‘its most constant supporter’.56 With a flood of GAF Sigint pouring in from the Y stations, the Air Ministry’s Directorate of Intelligence was upgraded in the spring of 1941.
***
Intelligence obtained from non-military sources was predominant in GC&CS before the war and in wartime continued to be overseen directly by AGD. This part of his organisation operated purely as a cryptanalytical bureau, limited to the decryption and reporting of enemy and neutral diplomatic and commercial communications. The Section had moved to BP from Broadway Buildings (part of Commercial Section) on 26 August 1939, with around ninety staff comprising some thirty on a Senior or Junior Assistant grade (cryptanalysts with varying degrees of experience, in some cases gained in WW1) and some thirty support staff (Higher Clerical Officers, Clerical Officers and Clerical Assistants); the remainder were typists, traffic-handling staff and office keepers. The structure remained as it had been before the war, with AGD as Head with a Registry; the Distribution and Reference Section with its own typists; the Traffic Section; a large number of country sections, each separately responsible to AGD, and some cryptanalysts engaged in specific research problems outside the framework of sections. This last group, consisting of three experienced men and several juniors, was engaged in advance problems such as Enigma (in support of the Services Sections) and the Japanese ‘Purple Machine’. The move went smoothly, with no interruption to its output. AGD was based in the Mansion, while the Section occupied Elmers School which provided working accommodation for some sixty people. The Senior Assistant of Distribution and Reference had day to day administration responsibility for sections at Elmers School, while AGD remained in charge of policy and progress of sections whose heads continued to report directly to him.
The Distribution and Reference Section maintained contact with the users but according to Birch ‘the producers were not concerned with traffic analysis intelligence or the use of their products and were in general ignorant of even the sources of supply of their raw material’. The key ‘country’ sections were Italian, Japanese, Near Eastern, Chinese, Balkans, Portuguese and French, which was disbanded in May and reconstituted after the fall of France to deal with Vichy government traffic. The traffic of twenty-six countries had been studied, 70,000 out of 100,000 telegrams received, with 8,495 translated and circulated. At the same time, 135,000 coded or plain text messages had been circulated by the Commercial Section (with thirty-nine staff). There had not been a German section since 1919, because following the publication of the Zimmermann Telegram, German diplomatic ciphers used two very high-grade systems, one of which was an OTP. No breaks had been made and only one cryptanalyst was assigned on a ‘care and maintenance’ basis.
Seven-day-a-week working was introduced shortly after arrival at BP and a two-shift system was later introduced due to a shortage of space. Adjustments had to be made to the work of the section to take account of wartime priorities. The diplomatic and consular communications of Axis c
ountries could reveal military plans as well as their military and industrial resources. Those of neutral countries who had representatives in enemy countries might reveal similar information as well as their own intentions. It was felt that Allied countries and safe friends did not need to be covered. Work on German diplomatic traffic was treated like a research problem, with German Service problems being a priority. However, a veteran of Room 40 took up the work and when one of the systems in use was identified (the OTP) in 1941, collaborative work with the Americans could begin. The Italian, Near Eastern, Balkan and Chinese Sections were reinforced as was the Japanese Section when linguists were found. The French Section had been disbanded in April 1940.
The Diplomatic Section benefited from the recruitment of the fiftysix seniors and thirty women linguists upon the outbreak of war. While the majority went to the Service Sections, some were sent to Elmers School and were immediately useful. Working space was barely adequate to house the peacetime staff, but this was solved temporarily by moving the Commercial Section back to London. This freed up an outbuilding but it was demolished by a stray bomb on 20 November 1940.
Italy’s entry into the war allowed the Diplomatic Section to demonstrate that more than political intelligence could be obtained from the enemy’s diplomatic and consular communications. It could reveal their military strength, intentions and plans for espionage and economic war. GC&CS had access to traffic from the Italian embassies in Moscow, Madrid, Lisbon, Washington and Tokyo. Legations in the Balkan capitals had military attachés who talked openly to Rome of German military movements until the Germans subsequently stopped their telegrams. Their representatives in the Near East, South America and Ireland along with their consuls at Lourenço Marques (now Maputo, the capital of Mozambique) and other strategic points, frequently reported on Allied convoys. Much of this traffic was now readable and by July 1942, twenty-five communications per day were being translated and issued. The Italian Section was placed on a 15-hour watch when Italy entered the war and on occasion this was extended to a 24-hour watch. The Section was reinforced to nineteen staff so that even low-grade traffic could be covered. Further traffic was obtained when the Admiralty cut the undersea cable running from Malaga to Genoa, forcing the Italian Embassy in Madrid to use W/T. This provided access to invaluable intelligence about the Spanish Government’s relations with the Axis countries.
