by Alastair Denniston- Code-breaking From Room 40 to Berkeley Street
In September 1940, the Commercial Section moved to Wavendon House in the village of Wavendon near BP to avoid London bombing. By the end of December 1940, it had thirty-nine staff. In 1940 it circulated 135,000 messages either in code or as plain text to the Ministry of Economic Warfare (MEW) and to the Information and Records Branch of censorship. Many of these were also circulated to the Admiralty, War Office, Air Ministry and the Ministry of Shipping.
The success of GC&CS’s diplomatic and commercial sections, under AGD’s direct control, would prove to pave the way for events which played out later in the war. The academic David Alvarez has provided a comparison between the records of GC&CS and the US’s SIS in reading foreign diplomatic codes.57 It provides reasons why the US would stand to benefit significantly from any collaboration with Britain:
British codebreakers for example had solved high-grade Italian ciphers while the Americans were struggling to master low-grade versions. GC&CS was reading several Vichy French diplomatic ciphers at a time when the SIS was hoping that the next staff increase might free up one or two officers to open a French section. GC&CS routinely solved Balkan and Near Eastern systems that did not even appear on the cryptologic horizon of the SIS. Even in the matter of Japanese codes and ciphers (an American speciality) the British were reading more systems (including Red) than the Americans, although GC&CS had had no more luck with Purple than their cousins across the Atlantic. Only in the area of German and Russian communications were the prospective partners equal. In their efforts against Berlin’s diplomatic systems neither had been able to advance beyond the reconstruction of the DESAB code. As for Russian systems, neither was studying Moscow’s diplomatic ciphers, although GC&CS was reading a few Red Army and Comintern systems.
Throughout 1941 BP’s Service Sections continued to expand. There were six separate naval sections four of which covered German, Italian, Spanish and French traffic. Hut 8 dealt with German Naval Enigma and Knox’s section handled Italian Naval Enigma. Japanese naval Sigint had been passed to FECB and the staff had moved to the Japanese Diplomatic Section. Most of the French traffic was being read, but very little of the Italian traffic apart from Italian Naval Enigma, solved by Knox. For German naval work, there was an organisational division between Hut 8, which was tackling the Enigma problem and the German Naval Section, which was dealing with everything else. However, Hut 8 supplied the decrypts and German Naval Section, which became known as Hut 4, provided W/T cover, traffic analysis and relevant intelligence. At this time only two Bombe machines were operational although eighteen had been ordered. According to Birch: ‘The sum total of cryptanalytic success on German naval systems of all grades and kinds cannot be said to have as yet achieved much operational significance.’ Furthermore:
The most significant achievement of German Naval Section in 1940 was the control it had over its own domain. At the beginning of the war, it did no cryptanalysis, because it had no cryptanalysts; no traffic analysis, because ‘operational intelligence’ was the monopoly of 8G; no general intelligence because there was none from Sigint sources. By the end of 1940 it had established a monopoly of all naval Sigint undertaken on German sources at G.C. & C.S. – even the raw products of Hut 8, the only exception, passed directly to it for processing and issue – and was already the champion of the ‘one Service, one Section’ principle as opposed to the interservice and the functional theories, for the proper organisation of G.C. & C.S.
While there was a considerable amount of operational intelligence coming from traffic analysis, the Admiralty was not convinced of its reliability. Birch, in his capacity as Head of the Naval Section at BP, had written to AGD on 20 October 1940:
Apart from questions of efficiency and convenience here, there cannot, from the point of view of Admiralty, be several separate entities at B.P. independently responsible for supplying them with intelligence and its interpretation. The German Naval Section is the responsible body and Admiralty rightly holds it so. To them it’s no excuse at all for me to say: ‘Oh, yes, we’re the German Naval Section, but this bit of intelligence is done elsewhere, so I’m not responsible.’ It merely looks like a phoney alibi; it brings the Section into disrepute and makes B.P. look like a badly organised muddle.
The Military Section differed considerably from the Naval Section. It was part of 4 I.S. and controlled by MI8 with its head, Tiltman, responsible to DMI in the War Office as well as AGD. At this stage it had no Enigma traffic to work on as it had handed over Japanese work to FECB and Italian work to CBME. It devoted most of its staff time to the cryptanalytical problems of other Services. Tiltman was the leading cryptanalyst at GC&CS, although his role was not formalised until the following year and Military Section tended to become synonymous with his functions. Examples of this were the Russian Section created in March 1940, the setting up of an inter-service school for the training of cryptanalysts called Interservice Special Intelligence School (ISSIS) and a Research Section to tackle tough cryptanalytical problems.58
Air Section, like the Military Section was separated from Enigma traffic and under the control of a Ministry Section, AI1(e). However, it had had great success and could claim to read all GAF non-machine ciphers, including four German, five Russian and two Italian meteorological ciphers. It did not report or interpret the decrypts, as this was done in conformity with the Ministry’s policy of decentralisation at Cheadle from where it was sent on to AI(e). Air Section was subordinate to AI1(e) whose view was that their function was purely cryptanalysis. Any effort to fill the many gaps in RAF intelligence from Sigint sources was discouraged and direct contact between them and AI3(b), Fighter Command and Bomber Command was denied as long as possible. A Fighter Intelligence Subsection had been formed in Air Section to study W/T traffic in December 1940, and the following month, a member of Hut 3 and a computer59 from Cheadle formed a subsection called SALU to tie up information between low and high grade sources. The work of both sections was ignored by the Air Ministry. Air Section also housed a meteorological service which provided weather data for British forecasting services in preparing weather forecasts for the RAF and, later in the war, the US Air Force.60
Another valuable source of intelligence for AGD’s organisation was from the German military intelligence organisation, the Abwehr, headed by Admiral Wilhelm Canaris. They used a version of the Enigma machine to encrypt their communications with agents abroad. The system was broken at BP in 1940 by Oliver Strachey and his ISOS team. The intelligence was sent to SIS’s counterespionage branch, Section V, which was located in St. Albans, about 26 miles from BP and the same distance from Whitehall in London. The intelligence war station in St Albans was headed by Captain Felix Henry Cowgill, who spoke about their work after the war:
At first they were little more than fragments of German intelligence signals, in German, and either meaningless or almost so. But in a surprisingly short time, certainly less than a year, we were reading Canaris’s traffic and that of the Sicherheitsdienst [the intelligence agency of the SS and the Nazi Party] too. From that basis we could, therefore, work towards first the control of the German intelligence service wherever and whenever it appeared, and then the liquidation of that service.61
Secret service communications were intercepted by the Radio Security Service (RSS) and worked on by the ISOS at GC&CS, although the cryptanalysts were working for SIS. By the end of the year, RSS had become the unit responsible for the breaking-down of the global communications of the enemy secret services. They relied on Post Office Y stations and amateur wireless operators throughout the country. They grew from 150 at the beginning of the war to 1,000 strong.
The internal organisation of GC&CS juxtaposed individual Service sections. This facilitated the sharing of experience, interchange of personnel, collaboration on common problems, pooling of research and machinery. The exception was the complete fusion of German Army and Air Force Enigma in Hut 6. They handed decrypted messages to Hut 3, who camouflaged the decrypts as agents’ reports and by-pass
ed the official Sigint sections MI8 and AI1(e), going directly to the Country Sections, MI14 and AI3. MI8 and AI1(e) did place staff in Hut 3 with a section called 3M advising on Military Intelligence and another section called 3A advising on Air Intelligence. Both of these sections disseminated material to ministries and commands.
Travis took charge of mechanical aids to cryptanalysis, and as the Bombes were used by the cryptanalysts of all Services, his command included Huts 6 and 8. Therefore, Hut 8 was divorced from Naval Section and Huts 3 and 6 became separate entities. Other variants of Service ciphers were attacked by a small and separate group led by Knox.
Knox persisted with his complaints throughout 1941 and AGD made one final attempt to explain the situation to him on 11 November 1941:
My dear Dilly,
Thank you for your letter. I am glad that you are frank and open with me. I know we disagree fundamentally as to how this show should be run but I am still convinced that my way is better than yours and likely to have wider and more effective results.
If you do design a super Rolls Royce that is no reason why you should yourself drive the thing up to the house of a possible buyer, more especially if you are not a very good driver. I lost any confidence you had in me when I disagreed with you in Dec. 1939 and said that you could not exploit your own success and run huts 6 and 3. I was right – you broke new ground while the building in your foundation was carried on by Travis etc. who, I say, were better adapted to this process than you.
our next big show was K. You alone among us found the way but the full value of your work could only be obtained by fitting results into the full picture in the Naval Section.
And now comes your latest effort which only proves again that you are the right man in the right place. You told me of a side-line in Intelligence that you wanted to develop. I agreed but begged you to remember your real metier. So you produced this result which none other of our party could have done.
You say you did it because you are a scholar who proceeds from his raw material to his finished text, well – who is preventing you – you have access to all past material and copies of all new.
What are my grocer’s window dressing. Eric Smith offers all productions in neat form to those who need them. Birch ties up information from every naval source and tells the story. Hut 3 collects and reports accurate information derived from the source you invented.
Do you want to be the inventor and the car driver? Do you want to be Eric Smith, Strachey, Birch and Wing-Commander Humphreys and De Grey wrapt into one which will include Knox who is the source? If so I don’t agree and I don’t mind at all what steps you take.
You are Knox, a scholar with a European reputation who knows more about the inside of an Enigma machine than anyone else. The exigencies of war need that latter gift of yours though few people are aware of it.
The exploitation of your results can be left to others as long as there are new fields for you to explore.
I do disagree with you.
Yours ever,
A.G.D.62
Alas, it was to no avail and, demonstrating an astonishing amount of arrogance, Knox wrote directly to Menzies. It is remarkable that AGD continued to treat Knox in a professional manner given the apparent contempt which Knox seemed to hold him:
Though it was not my business, since I was ordered to do something else, I devised a theory which reduced, often to a twentieth part, the time necessary to be expended on the solution of a day’s messages, a method which, for at least nine months had escaped the Polish bureau. Having earned thereby still greater unpopularity with Commander Denniston, I was powerless to insist that the whole affair was still in an experimental stage, that watches should be constituted, that the best brains should be employed, and that we should be prepared when ‘the war should begin’. The hut had, since Commander Denniston laid it down, to remain in charge of one officer physically unfit for night work, a very able worker of no marked push or originality. In consequence when the push came watches often possessed workers of the highest ability but unused to certain minutiae, and some early and vital keys were lost. For this Commander Denniston’s mandate was responsible.
When a cipher is out Commander Denniston is willing to parade superiors round sections of whose work he understands literally nothing and to assume credit for achievements his mismanagement nearly ruined.
Two things remain to be said. As to my right to criticise I need only remind you that I am a Senior Cryptographer. At the end of the Great War Commander Denniston (with a staff of about 30) was administering one of the German Fleet Cyphers and I (with a staff of three) another. If memory serves me at the end of the war the smaller unit was supplying copious and accurate information, while the larger remained silent. Secondly, Sir, if you criticise me on the score that I have accepted money in peace time and desert during hostilities I need only say that neither Commander Denniston’s friends, if any, expected, nor his many enemies feared that, on the outbreak of war such responsibilities should be left in hands so incapable.63
***
On 8 February 1941 a meeting took place in AGD’s office at BP which many historians believe heralded the start of the ‘Special Relationship’ between Britain and the US. However, intelligence cooperation between the two countries can be traced back to WW1.64 In July 1917, General John Pershing, commander of the American Expeditionary Force (AEF), decided that his intelligence staff (G-2) should follow the model of their British counterparts in the BEF. Pershing kept his distance from the British Army and its commander, Field Marshal Sir Douglas Haig, but did establish a good rapport with the French commander, General Philippe Petain. Major Dennis Nolan (later Brigadier General) was head of Pershing’s intelligence staff. He established a good working relationship with the head of the BEF’s intelligence staff, Brigadier General John Charteris. Charteris revealed details of the British intelligence system, including Sigint, and, after a presentation to the AEF by Charteris, Pershing approved the subsequent intelligence regulations on the British model without amendment. The British agreed to provide both document-based and training support.
The AEF’s decision to copy from the British brought junior intelligence officers from both armies into close contact. In 1917 Major Stewart Menzies was responsible for counter-espionage at BEF GHQ, where he established close links with his opposite numbers. By 1941, he had become Major General Sir Stewart Menzies, head of the British SIS. In 1918, William Friedman was a cryptographer working for the AEF G-2. By 1941, he was head of the American Signals Intelligence Service and had established close links to AGD. As for AGD himself, while not directly involved in Room 40’s work on the Zimmermann Telegram, he had witnessed first-hand how Admiral Hall and the American, Edward Bell, had worked closely together. The outcome of their collaboration had, in effect, brought the US into WW1. So AGD understood full well the potential benefits both to Britain and his own organisation, of a close working relationship with US intelligence services.
The 1941 meeting in AGD’s office had its origins in an informal visit to London in July 1940 by William Donovan, a special envoy of President Roosevelt. His visit had been prompted by SIS’s new representative in the US, William Stephenson, and his task was twofold: to assess Britain’s determination and ability to continue with the war, and see whether closer collaboration could be arranged with the Admiralty on intelligence. At the end of August, the British Government invited the American Military Observer Mission to attend a meeting in London with the British Chiefs of Staff. While the purpose of the meeting was to discuss the ‘Standardisation of Arms’,65 Brigadier General George Strong told the meeting about ‘the progress his Service was making against the Japanese and Italian ciphers and formally proposed to the Chiefs of Staff that the time had arrived for a free exchange of intelligence’.66 Interestingly, the US naval representative at the meeting, Admiral Ghormley, said nothing about cryptanalysis and the British offered nothing in return. Donovan returned to Washington and recommended a full and
direct exchange of intelligence between the two countries’ naval intelligence departments.
Within several months, the US Navy’s position had changed dramatically and Ghormley attended a meeting in London on 22 October with DNI John Godfrey and Menzies. The following day in Washington, Secretary of War, Henry Stimson, met with Secretary of the Navy, Frank Knox, to discuss sharing of information with the British on ‘the code methods of the Germans’.67 By the end of December, an Anglo-American pact on cryptanalysis had been agreed with the intention of sharing with each other, information about the codes and ciphers of Germany, Italy and Japan.
On 7 February 1941, four passengers disembarked from HMS King George V after she had anchored in Scapa Flow, the Home Fleet’s remote base in the Orkneys. Abraham Sinkov and Leo Rosen worked for the US Army Signal Intelligence Service, and Robert W. Weeks and Prescott H. Currier were US Navy communications officers. Sinkov was a late replacement for William Friedman who was unable to travel due to illness. They were greeted on the quay by a naval officer68 and several drivers for the two staff cars and a lorry which were at their disposal. The equipment that they brought with them would not fit on the aircraft sent to meet them, so they were transferred to another dock where they boarded a Royal Navy destroyer. It took them south, along the coast of Scotland and England to the Thames and London, docking at Tilbury. They travelled through London and late on the evening of 8 February arrived at BP, becoming the first Americans to do so. Remarkably, they brought with them perhaps the greatest achievement of US cryptanalysis during WW2. Japan had developed a high-level cipher machine to encrypt its diplomatic traffic and, like the Enigma system, it was also used to decrypt messages. The American SIS team under William Friedman and Frank Rowlett had broken the system, angoo-ki taipu b (Type B Machine – codenamed ‘Purple’ by the SIS) and subsequently were able to build replica machines. Their gift to the British was one of their replica machines,69 along with other codebreaking material
