Or, the Battle for Little Bletchley
During my researches into the history of the Corps, particularly relating to the world of signals intelligence and the Government Code and Cypher School (GC&CS), I have been amazed how often personal and service ambitions have played what might have been decisive roles in decision-making. Following the Venlo Incident in which Captain Payne Best played an unfortunately vital role the SIS found itself with virtually no agents and no sources of intelligence apart from that garnered from SIGINT. As a result, control of this intelligence would be a source of controversy between the services and GC&CS for several years, even at the outbreak of war. Gordon Welchman, equally responsible for the breaking of Enigma as Turing, but now virtually unknown, wrote that Dilly Knox, one of the most senior men in GC&CS ‘did not seem to like me’. After a week or so, Knox gave him some sort of test and was annoyed, it seemed, when he passed. Knox saw the breaking of Enigma as an academic problem akin to solving the Rosetta stone and resented any outside intrusion into his little world – particularly from a mathematician. He saw it as a role for a classicist! When in 1940 the Corps was founded and the call went out for regular officers, particularly with language skills, to transfer, there was great reluctance to do so by many seeing the Corps as a promotion dead-end. These are not isolated examples! With the rush for integration of the various suppliers of intelligence, it is perhaps worth looking at what happened in the North African campaign from 1940.
Partly as a result of the special features of the Middle East theatre and partly because military SIGINT had been plentiful in the Middle East for several years, the services had organised the control of SIGINT locally through all its stages including interception and cryptanalysis, interpretation and its use all on a single service basis. From 1924 the Army had centred its SIGINT organisation on No. 2 W/T Company at Sarafand. From 1937 the RAF, which had previously concentrated its intelligence at Baghdad, had built up Heliopolis (Cairo) into what soon became known as the ‘Cheadle of the Middle East’. By the late spring of 1939 the Navy was preparing to undertake Italian traffic analysis and to exploit Italian low-grade codes and ciphers at its OIC in Malta. The Italians were making liberal use of ‘plain language’ and Italian low-grade ciphers were easy to read, and even Italian high-grade ciphers could be exploited locally, once the initial cryptographic work had been done at GC&CS. It seemed obvious to the three services – if not to GC&CS – that the whole SIGINT process should as far as possible be handled in close proximity to their operational headquarters. On the other hand there was a drawback to this, in that with those HQs and their intelligence staffs widely dispersed and physically remote from their subordinate operational authorities and their intercept and Y stations, they naturally gave priority to improving their own SIGINT organisations. Thus, in August 1939 that
part of the Sarafand unit which dealt with Italian Army SIGINT was moved to Mersa Matruh to begin work on material intercepted in the field, and in the first half of 1940 the RAF were undertaking similar work in Cairo as well as setting up additional intercept stations in Aden, Khartoum, Malta and the south of France. Under these circumstances, not surprisingly, they resisted suggestions from GC&CS that, in the Middle East as in the UK. the exploitation of SIGINT should be centralised as far as possible at a single inter-service centre – and which one.
In the UK, there was a conflict between GC&CS on the one hand, with its view that all high-grade cryptanalysis and at least the direction of all intercept programmes should be concentrated at one inter-service centre and the Service Intelligence Directorates in Whitehall on the other, with their insistence on their own responsibility for evaluating intelligence and their fear that this responsibility would be eroded if they did not resume control of their own cryptanalysis on the outbreak of war. By 1939, it was coming to be settled by a compromise, which largely but not entirely suited GC&CS. In 1938 and again at the end of 1939 and the beginning of 1940, GC&CS were advocating the creation of a Combined Signal Bureau in the Middle East in the hope that their views, and possibly their responsibility, would be extended also to that theatre as it was elsewhere. They hoped that whatever arrangements the three services might make to preserve their right to assess for themselves all intelligence affecting them, at least they would consent to pooling their cryptanalytical resources in the one place, modelled on GC&CS at Bletchley Park. This would assist them with the growing problem of getting intercepted traffic back to London with considerably less delay and to agree with them a satisfactory division of labour between the U.K and the Theatre. But the intelligence staffs in the Middle East rejected this proposal; even to the extent of attempting to persuade GC&CS that there were grave obstacles to a centralised SIGINT centre. Nor did these obstacles disappear when Italy entered the war. This attitude of the Middle East intelligence staffs rested not only on their concern for the interests of their own service, but also on their jealousy of each other. The RAF feared that in a combined bureau it would be dominated by the larger and longer established army intelligence organisation.
These mutual jealousies gave way to a common front when it came to arguing about how the work on the Italian ciphers and their limited number of cryptanalysts trained in Italian work should be divided between the UK and the Middle East. Until the war extended to the Middle East the division of labour had worked well enough. With the outbreak of war with Italy in June 1940, however, the communications links between the UK and the Middle East proved to be quite inadequate to meet the need to exchange a growing body of intercepts and decrypts without delay by W/T; not unnaturally, the complaints by the Middle East authorities that in return, the product of GC&CS was reaching them too slowly were accompanied by urgent calls for the transfer of cryptanalytical staff from GC&CS to the Middle East.
For most of the period between September 1939 and Italy’s entry into the war, British authorities were more concerned about the threat of a German move through the Balkans together with a Russian advance through the Anglo-Iranian oilfields. In July 1940, in response to an urgent appeal from C-in-C Middle East and with the hope that it might lead to the development of a combined Bureau, GC&CS sent out a small party consisting of cryptanalysts from all three services as well as civilians. They were to be joined later by a small party of Italian specialists. Despite this there were still increasingly bitter complaints and suggestions that GC&CS were giving too little priority to Italian ciphers.
During most of 1940 the GAF (Red) Enigma was as yet the only important SIGINT source about Germany’s activities and, as for Russia, SIGINT threw no light on her intentions. In the pre-war years, work on her service codes and ciphers had been confined to a small unit in India and to No. 2 W/T Company, the Army’s SIGINT unit at Sarafand, until GC&CS took it up after the conclusion of the German–Soviet pact in August 1939. Since then, GC&CS had broken the Russian meteorological cipher, read a considerable number of naval signals and decoded about a quarter of some 4,000 army and police messages. But like much of what had long been exploited in India and Sarafand, it was local stuff and although of some use tactically it yielded nothing of any strategic value.1
Because of the inadequacy and insecurity of communications, GC&CS found it difficult to keep Cairo abreast of its progress against the Italian cyphers, although the army authorities in the Middle East were better served by their field SIGINT organisations than the other two services. No. 2 W/T Company at Sarafand had long experience with the Italian Army’s medium and low-grade codes and cyphers. By September 1940, when No. 2 moved its Italian section to Egypt it had organised four sections for work in forward areas. One of these went to Mersa Matruh in August and during the Italian’s brief advance in September was able to gain a fairly comprehensive picture of the Italian forces from work on its cyphers. Helped by this experience the forward section attached to the Western Desert HQ during the British counter-offensive in December 1940 was even more useful. Despite the windfalls brought by field SIGINT during the advance against the Italians, its very character and liability to interruption prevented it from being treated by the operational commands as being on a par with regular, orthodox, non-SIGINT intelligence sources. Rather, its products were treated as a bonus. This reacted on the non- SIGINT sources in that their results could not reveal their true value, unless integrated with the depth of reliable information that SIGINT could characteristically provide. So, intelligence despite its occasional big bonuses remained during the first years of the desert campaign as the Cinderella of the staff, and information gained about the enemy was frequently treated as ‘interesting’ rather than valuable.2
1 All work on Russian codes and ciphers was stopped from 22 June 1941, the day on which Germany attacked the USSR.
2 Williams. The use of Ultra by the Army, p.3.
Despite the advent of better field SIGINT and a better supply of high-grade German SIGINT to the Middle East on a direct service from GC&CS from March 1941, this attitude towards SIGINT was to change only slowly. Partly because stricter security conditions and precautions were applied to German rather than to Italian SIGINT, and partly because of the blunted impact of its fragmentary nature, the commanders and regular officers who held the senior intelligence posts at forward HQ’s could not grasp quickly the relevance and value of SIGINT information. It would not be until the summer of 1942 that intelligence came to be fully regarded as an integral part of the conduct of the war, and that field SIGINT came to be fully integrated with other intelligence sources at army level and to be a major source of tactical information. The foundation for this development was laid by the pioneering work of the forward sections which had worked in the desert from the autumn of 1940.
The disputes that had continued to simmer between GC&CS and the Middle East intelligence staffs came to a head in November 1940. The major conflict was about the arrangements that should be made for producing or procuring SIGINT rather than for interpreting and using the product. There were also other issues like complaints from the Middle East about the delays involved in the system by which GC&CS decrypts were despatched to the Whitehall intelligence departments for onward transmission to the theatre; however, these were subsidiary to the main bones of contention which were, first, how much cryptanalysis should be done at GC&CS and how much in Cairo and, secondly, whether GC&CS were devoting to the Italian cyphers as much effort as it should. GC&CS emphasised that the cypher changes that had taken place since the entry of Italy into the war necessitated the concentration of high-grade cryptanalysis in the UK and used its considerable success against the new cyphers to justify its position. The Middle East intelligence authorities, on the other hand, stressing the losses that had been inflicted upon them by the cypher changes, became more and more convinced that the new cyphers would be broken with less delay, and their product used with more effect if the main cryptanalytical effort was moved to the Middle East and undertaken within their own separate Service intelligence organisations. So strongly did they put their case that the chiefs of staff were prevailed upon to demand London’s agreement to the transfer of this work from GC&CS to Cairo.
On 31 October the Cs-in-C rejected this demand and, even worse, insisted upon the establishment at the RAF station at Heliopolis of a Combined Bureau Middle East (CBME), and defined the functions of the bureau and GC&CS in such a way as to preserve GC&CS control of cryptanalysis. There was some little consolation. The bureau was to be administered by the army and was to comprise the cryptanalytical sections of the separate service intelligence staffs in Egypt, together with the cryptanalylists who had been sent out in July and August from GC&CS. GC&CS was to remain responsible for the basic research and the initial attack on the high-grade Italian cyphers while the bureau was to work on the lesser cyphers and, with the aid of the GC&CS results, to be responsible for exploiting readable high-grade cyphers for the benefit of the service Intelligence staffs in the theatre.
Why, it might fairly be asked, had the chiefs of staff reached this decision which seemed to fly in the face of what the service intelligence chiefs wanted?
The chiefs of staff took these decisions on the advice of the directors of intelligence in the Whitehall service departments. Up until then, the departments had supported the resistance of the Middle East commands to a combined bureau. But now they had powerful reasons for supporting GC&CS’s demand – against the wishes of the commands – to control the work on Italian cyphers and thus for accepting GC&CS’s claim that in the Middle East, as in the United Kingdom, cryptanalysis should be concentrated in a single bureau. For one thing, they needed Italian SIGINT in the Middle East as much as did the Middle East commands, if not quite so urgently! As GC&CS pointed out, the Middle East was not the only area in which the Italian forces might operate. For another, the time was rapidly approaching when German forces might be operating in the Mediterranean and North Africa. Although it might have been arguable that the Italian high-grade book cyphers could just as well be handled in the Middle East, it was unarguable that work on the machine cyphers (Enigma) was beyond the experience of the cryptanalysts in the Middle East. This was reinforced when GC&CS broke an Italian machine cypher in September adding weight to the argument that work on the cyphers had to be concentrated. For all these reasons, the struggle over the production of SIGINT locally, lost out in favour of the principle of concentration.
The Service departments had conceded the argument reluctantly in favour of GC&CS in the U.K. They then had to accept – in the form of CBME – a miniature version of GC&CS, which would also act as an outpost of GC&CS. Just as the Whitehall intelligence directorates had, equally unwillingly, yielded control of cryptanalysis and the management of interception to GC&CS, they had insisted on their right and need to remain responsible for interpreting SIGINT products for their own services, so did the Middle East commands, remaining determined to limit CBME strictly to analytical work.
The arrangements reached turned out less tidily than hoped for and expected. In practice CBME comprised only an army and RAF section – no naval section – due to the failure of GC&CS to break the new naval book-cyphers and the continued concentration of other naval SIGINT in Alexandria.
On the whole, however, except that GC&CS and the CBME were constantly to criticise each other for failing to interchange captured code and documents, recriminations between the Middle East theatre and GC&CS about the problems connected with the procurement of SIGINT had come to an end.
Was this the end of the quarrelling in the Middle East? Far from it! Relations between CBME and the army and RAF intelligence staff were still bedevilled by the
determination of the two services to resist encroachment by CBME upon their individual control of their own intelligence components. And they, rather than GC&CS, had succeeded in recovering the IAF high-grade cypher when it was changed in November 1940. This success was not entirely welcomed by GC&CS. Relations were also bedevilled by prolonged demarcation disputes concerning such things as security, the exploitation of low-grade SIGINT and, above all the allocation of intercept priorities. Nor was this the only factor which was preventing CBME from developing into a full-blown inter-service SIGINT centre both producing and appreciating SIGINT for all the Middle East commands. Whilst the service intelligence HQ’s were anxious to restrict CBME to cryptanalysis, they were equally insistent that the other SIGINT processes such as interception, the exploitation of low-grade SIGINT and its interpretation should be carried out on a single-service basis under their own control as near as possible to their own operational commands. This was soon to lead to further strife! But this is another story!
10 December 2016