In his Churchill encounter in 1915, Hall went on to add, ‘If we were to get peace, or if we were to get a peaceful passage for that amount [up to £4,000,000] I imagine . . . [the cabinet] would be glad enough to pay.’ This money paid into a Turkish bank would have split off the Turks from the Germans and allowed the Royal Navy a free passage through the Dardanelles, thus averting the catastrophe that threatened. Churchill refused to sanction the bribe, the campaign duly started, and this may have accounted for his subsequent unavailing pursuit of a Turkish alliance in the Second World War.15
Between the Wars
The armistice in 1918, followed by an acute cash crisis in Whitehall, threatened the continuity of Room 40 and its military counterpart, MI 1B, but not for long. Unlike the State Department in Washington, the FO continued to monitor peacetime diplomatic messages into and out of London and as many European capitals as could legitimately be targeted by cable scrutiny as well as by intercept stations. Funds were made available, procedures and priorities were established, and the Government Code and Cipher School (GCCS), was formally set up on 1 November 1919.
The key to the whole operation was, and is, interception. Without interception there can be no intercepts and no clandestine eavesdropping. The British seem to have been alone in adding cable interference as an additional means of reading messages – though telephone tapping became standard international practice in the 1930s, and tapping into other peoples’ landlines may also have been practised. The operation of censorship is impossible to track, for lack of any references in the files. But the building of intercept stations in the interwar period in England, Scotland, Iraq and India was on a substantial scale, whose funding has only recently been identified.16
GCCS was the responsibility of the Head of the Secret Intelligence Services, but for reasons of security became the ‘adopted child’ of the FO, of which it was in fact a department. Both parents valued this unacknowledgeable child. In the case of the head of SIS, Adm Hugh Sinclair, the evidence for this is largely negative. Sinclair allowed GCCS’s distribution of Russian commercial and diplomatic intercepts to be used politically to expose the horrors of the revolution, thus compromising the work with the result that the Soviets resorted to the OTP which made their traffic unreadable for many years to come.17 Sinclair has been almost entirely weeded out of the FO files but, despite his Russian blunder, those who worked for him (and his sister) held him in such respect and admiration that the group must have forgiven him; and under him GCCS continued to function effectively up to the time that he bought Bletchley Park for the nation in 1938, and his death in November 1939.18
The role of the FO as foster-father is easier to trace from the record but more difficult to evaluate. The cultural climate of the prewar FO was many-layered but at the top lay, like thick cream on a trifle, an Eton, Winchester and Oxbridge elite. For these people integration would have diluted their privileges by association with the Consular Service and the refugee specialists, linguists and ex-servicemen of GCCS. The crossover between the FO and GCCS may also have been limited by the absence of any career structure in the latter, as well as by its formidable linguistic requirements. But the FO could recognise value when they saw it, as they did as early as October 1922, when the distribution of Turkish diplomatic intercepts within Whitehall kept the authorities au fait with the dangerous Turkish and Greek military build-up in Thrace which resulted in the Chanak crisis.19
These 1922 intercepts, shared by those in power in Britain, revealed the extent of Turkish anger at the unprovoked bellicosity of the Greeks, from whom first Italy and then France withdrew support, leaving only Britain to back an untimely attempt to keep Turkey out of Europe. A year later GCCS’s successful reading of the Soviet diplomatic cipher brought about a diplomatic crisis which nearly precipitated a confrontation with the Bolshevik leaders.20 Both these crises were managed by the cabinet on the basis of the diplomatic decrypts of GCCS. In 1922 Churchill and Lloyd George regularly discussed their content and implications, while Sinclair, who as Head of the Secret Intelligence Services saw them routinely, encouraged their use as a means of expelling two leading Bolshevik diplomats then in London, though he later regretted this.21 Fifteen years later when the Italians uncharacteristically erupted south-east into Abyssinia and west into participation in the Spanish Civil War, increased volume of diplomatic and naval traffic, mostly Italian, enabled GCCS not only to break machine and other ciphers but to contribute to foreign policy. Anthony Eden, to whom historians have not been kind, then personally drafted a successful request to the Treasury for funds to cover GCCS’s enhanced activities in 1936–37.22 The War Office lent its support to the expansion: ‘This form of intelligence must be thoroughly organised in peace and . . . the cost entailed merely represents a small insurance premium.’23
At the end of 1919 GCCS employed 66 staff. This number went up to 94 in 1924, of whom 65 were support staff. By 1935 there were 104 on the payroll, of which 67 were support staff. The Abyssinian War produced a temporary need for Italian specialists, some of whom were laid off when the emergency receded, but the remainder supplied a firm basis for Bletchley’s outstanding Italian section. A few were allowed to stay on because other and larger shadows were looming. In 1939 there were 125, of whom 88 were support staff. Of the 37 senior and junior assistants recruited between the wars and employed in 1939, many served throughout the war and into the Cold War period – testimony to the success of the informal and even eccentric method of recruiting practised by the head of GCCS.
Little is said about formal training, which perhaps tells its own story. About training the War Office early on took the view that ‘the only way a man can learn to be a cryptographer is by devilling for an expert. A training programme would be an impossibility.’24 The senior assistants were all proven cryptographers with a track-record of achievement and expertise stretching back in some cases to 1915. Cryptography (or cryptology, to use the American word) has a history which goes back to ancient Egypt, and in the USA formal cryptanalytical training, including the history of the subject, was promoted by the Head of Army Signals Intelligence, William Friedman.25 He had studied the subject academically as well as being an outstanding practitioner. Other than Oliver Strachey’s notes of his lectures to GCCS recruits in Churchill College, Cambridge, there is little evidence of similar training in the 1920s, which suggests that training was mostly ‘on the job’, and this is confirmed by the memoirs of recruits to Bletchley Park between 1940 and 1943.26 By then the techniques were changing so fast, and the work to be done was so urgent, that formal training may well have been impossible. In the prewar GCCS the head and his deputy and the senior assistants were considerable cryptographers, but regarded their job as too hush-hush even to give it a name, much less a pedigree and academic respectability. They referred to ‘special work’ as opposed to ‘administrative duties’.
The age and previous experience of most recruits, until the arrival of Oxbridge graduates in the mid-1930s, would indicate that a career structure within GCCS would have been impossible anyhow. Within the FO and Diplomatic Service as a whole there was, of course, a clear structure. GCCS’s staff, both senior and junior, was explicitly bound out of this. Any crossover between the Diplomatic Service and GCCS, was minimal.27 Yet to achieve what GCCS undoubtedly did required intelligence, dedication, discretion, flair, self-discipline and self-motivation of an uncommonly high order. The dichotomy between what was needed and the lack of incentives, both for recruits and experienced staff, was acknowledged by the head of GCCS:
It must be remembered that beyond a salary and accommodation vote GCCS had no financial status; it became in fact an adopted child of the Foreign Office with no family rights, and the poor relation of SIS whose peacetime activities left little cash to spare.28
GCCS’s relations with other parts of government changed when Sinclair died on 8 November 1939. He held the trust, loyalty and affection of all the staff, and his death at such a crucial moment was the worst possibl
e news for them. Denniston himself never recovered professionally, and found it difficult to share problems with his successor. ‘The Admiral’ was, indeed, irreplaceable: he understood the high degree of autonomy essential to cryptography. He ensured that the specialists worked in conditions appropriate to the intellectual strains of their work. He moved confidently in the highest Whitehall circles of power. Menzies’s role as Director of Bletchley Park is more difficult to assess, and outside the scope of this chapter. He certainly kept Churchill supplied with intercepts, and regularly annotated them in green ink right through the war. He successfully protected their security.
By whatever means and to whatever effect, GCCS had developed a capacity for diplomatic eavesdropping on every major country except Germany, enhancing traditional practices in the light of experience and the intelligent application of critical analysis. The recruiting of cipher-brains destined to play a vital part at Bletchley might have appeared haphazard and restricted but the results must with hindsight be said to have been remarkable. Prof Peter Hennessy, the historian of Whitehall wrote:
Commander Denniston, head of GCCS and one of the best-informed people in Whitehall when it came to events in Europe, was . . . at his desk in Broadway that Sunday [3 September 1939], drafting a letter to the Clerk’s Department in the Foreign Office which controlled his manpower. Referring with sublime understatement to the activities of Hitler, he wrote: ‘For some days now we have been obliged to recruit from our emergency list men of the professor type who the Treasury agreed to pay at the rate of £600 a year. I attached herewith a list of these gentlemen already called up together with the dates of their joining. I will keep you informed at intervals of further recruitment.’
The first list included Nigel de Grey, one of the Zimmermann telegram codebreakers, and Professors E.R.P. ‘Vinca’ Vincent and Tom Boase, both Italian specialists who worked at BP throughout the war. The second list ‘was even more glittering with some of the greatest names ever to work in the trade’.29 They included E.R. Norman, John Jeffries, Gordon Welchman, Frank Adcock, Hugh Last and Alan Turing. Of these at least three were to become crucially important to the work of Bletchley Park. The belief of the seniors at GCCS that trawling Oxbridge for bright graduates was not just the only but the best way to solve the new cryptographic problems led to the assembly at Bletchley by the end of 1939 of ‘the ablest team of cryptographers and intelligence analysts in British history’.30 But even before they joined in late 1939 GCCS had made significant strides towards providing a total cryptographic service to those in government able and willing to use it. How this took place is the subject of the section that follows.
While the relationship of GCCS to the FO is documented, it also had a totally secret relationship to the Secret Intelligence Services (SIS) whose interwar head, Adm Hugh Sinclair, had acquired responsibility not only for GCCS but also Interception and Direction Finding.31 This enabled GCCS to develop a number of intelligent enhancements of their strictly cryptographic work. The first of these was Traffic Analysis (TA).32 This was the scanning of a targeted subject’s wireless activity when the actual messages could not be read but the volume and direction of them might indicate future activity. TA was acceptable when the cryptographer was temporarily unproductive, because the scale, intensity and direction of signals emanating from a targeted station could yield trustworthy information. TA played an important part in tactical planning, and indeed led in 1942 to a whole world of Y (or low-grade tactical) sigint.
The second was Direction Finding (DF), whereby the origin of a stream of signals and the location of a targeted enemy signals station could be worked out mathematically, by tracking the volume and direction of messages and applying co-ordinates. DF and TA were not useful in relation to diplomatic as distinct from service traffic. Diplomatic signals all came from fixed sources and the ups and downs in its volume did not reveal much. However, the work of GCCS and its outstations between the wars on TA and DF developed into the successful and secret Radio Security Service, which provided continuous useful traffic for all three branches of the British armed forces from 1941–45.
Perhaps even more significant was the evaluation of what should be fully processed and circulated, and what should not.33 Eighty-five per cent of traffic processed was never distributed. The credibility of what was circulated depended on eliminating the inessential, the marginal, the boring and the irrelevant. Keeping the customer happy was an essential part of this assessment, and required an intimate knowledge of the workings of Whitehall minds.
A conviction arising from GCCS’s response to interwar diplomatic eavesdropping was that apart from secure machine encipherment only the OTP would stay impervious to cryptanalytical attack. OTPs are essentially codes only available for one message, and known only to sender and recipient. The sender indicates the page, the column and the line where the message is to start in the first group of the signal. Other groups of numbers are added, and once used the whole page is torn off the pad and destroyed. GCCS between the wars promoted OTP, which became standard procedure for British diplomatic and Most Secret Source traffic throughout the war.34 One disadvantage of OTP is the scale of the printing requirement. Oxford University Press was the chief supplier of OTPs to the government, through HMSO, in the Second World War. Since it had unfilled capacity and a deserved reputation for discretion it proved a valued supplier.35
Some of the classical cryptographers and their support staff, with their different backgrounds and skills, working in close proximity with each other over the years, formed a group, became friends, played golf together and got to know the wives and children of other members. GCCS became known humorously as the Golf Club and Chess Society, a soubriquet bestowed on it by Dilwyn Knox. To what extent is the word ‘group’ applicable? Nigel West takes the view that for more than twenty years ‘Denniston essentially was GCCS’.36 An alternative theory emerges from the foregoing that under his leadership a group of like-minded but differentially talented civil servants evolved together a strategy which, while dealing adequately with the needs of the 1920s and 1930s, successfully anticipated the wartime requirements of a greatly enhanced cryptological bureau, and thus enabled Churchill and his government to read German intentions on a comprehensive basis from January 1940 till May 1945.
There is sufficient evidence that group-thinking prevailed from 1933 till 1940 to evolve the systems and practices which led to the successful breaking of Luftwaffe Enigma and its distribution on a continuous basis from January 1940. That in turn led to other Enigma successes, and to Ultra, which got the messages at speed to Churchill and the other relevant users. And that played a significant part in the Allied victory against the U-boats, and in Europe and the Far East. There is a strong microcosmic resemblance between the cryptographic structures and working practices of the First World War and the interwar period on the one hand, and the greatly expanded but still recognisably similar work of the ‘huts’ at Bletchley Park 1939–45.
Diplomatic Intercepts in the 1930s
By 1936 the substitution of German aggression for the Bolshevik threat was high on the FO agenda and GCCS was correspondingly required to reallocate its resources accordingly. While the German diplomatic cipher ‘Floradora’ remained unrecovered, Italian and Japanese traffic, together with that of Turkey and other countries, enabled GCCS to keep its masters fully informed of the approach of war. But who read the intercepts? How did the GCCS product figure in the minds of the Secretary of State, the mandarins at the FO, politicians in the know and service chiefs who preferred their secret intelligence brought to them by their own people? The evidence for the prewar period is slight, but in practice anyone above the grade of Third Secretary had access to them.37
What is even more difficult to discover is what the FO made of its intercepts. The head of GCCS does not mention what happened to BJs after they left his office. The distribution lists of wartime BJs indicate that the head of the Civil Service, Sir Edward Bridges, received a copy of everything circ
ulated.38 Under the rules of its establishment GCCS delivered selected material to the FO and later to the service ministries. Sir Robert Vansittart had his own sources of secret intelligence and must have read BJs. Eligibility to handle them was a function of FO bureaucracy. Ordinary outgoing and incoming telegraphic correspondence were treated differently from the receipt of intelligence from SIS and GCCS. The former was dealt with in the Communications Department, ciphered or deciphered, typed and circulated. There was a distinction between the ‘circulation’ and the work on the telegrams in the separate departments. The circulation of all but telegrams classified Secret or Most (later Top) Secret was very wide throughout the FO. But incoming telegrams (as other incoming correspondence) would be ‘entered’ in the appropriate registry and submitted by it to the department which it served. In the department all papers would go to the junior in the ‘Third Room’ (who could be a second or even a first secretary) responsible for the subject in question. Material received from SIS and GCCS came direct to the office of the permanent under-secretary and distributed from there under the direction of the latter’s private secretary, who was responsible for the FO’s relations with GCCS and MI6. BJs relevant to Turkey, for instance, would then go to the Southern Department which had been carved out of the old Central Department in 1932. The increased Italian traffic produced important information in the mid-1930s which would have been evaluated in the Southern Department, and would ‘be taken into account in forming judgments on situations or making recommendations for action. Their contribution could, of course, be very important.’39
It is impossible to be more specific about prewar BJ reading, and this makes difficult the assessment of the direct value of intercepts in the formation of government foreign policy before the advent of Churchill. The question is indeed unanswerable because they were but a part of the information-gathering service on which the FO based its advice to ministers. Moreover GCCS was making only slow progress on the German diplomatic machine cipher, of vital importance to the FO, until 1943. FO officials, like most other civil servants, read what was brought to them without too many questions about what could not be decrypted. They would have had preconceived ideas of what was important. They might reject whatever relevant information conflicted with these ideas. In general, their classical education made them philhellenic, which may by the same token have disposed them to be anti-Turk. Moreover, they had other sources of information on international diplomacy, privileged news and views, both clandestine and semi-official, in addition to BJs.
Churchill's Secret War Page 6