The Defence of the Realm

by Christopher Andrew

  After ‘lengthy and intricate investigations’, David Clarke reported in October 1943 that fifty-seven Communist Party members working in government departments, the armed forces and scientific research had ‘access to secret information and in some cases to information of the highest secrecy’. Three Communists were employed on the TUBE ALLOYS project, Britain’s top-secret atomic research programme. Clarke urged that all should be moved to non-classified work:

  The whole experience of the Security Service shows that members of the Communist Party place their loyalty to the Party above their loyalty to their Service and that their signature of the Official Secrets Act always carries a mental reservation in favour of the Party. The fact that the Communist Party is at present supporting the war would not prevent them from using any secret information in their possession irresponsibly and without regard to the true interests of the country.88

  Most of the fifty-seven Communists identified by Clarke had gained access to classified information because the departments concerned had no coherent vetting procedures for Party members and had failed to follow Security Service guidelines.89 The Home Secretary proposed that all Communists identified by the Security Service should be moved from secret work. Churchill was initially inclined to agree. He seems to have been persuaded otherwise by his intelligence adviser, Desmond Morton, a former SIS officer who had become hostile to the Security Service. A secret Whitehall panel which contained no Security Service representative was given responsibility for examining all cases of Communists in government departments submitted to it by the Service. The Security Service believed, probably correctly, that the panel was not up to its job. Before the panel was wound up in July 1945, it referred only one case to it.90

  As well as being seriously dissatisfied with Whitehall’s approach to protective security, the leadership of the Security Service was also well aware that it was failing to keep track of Soviet espionage. With at least partial justice, it blamed its failure on the severe restrictions placed by the Foreign Office on investigation of the Soviet embassy and Trade Delegation, and therefore of the intelligence residencies for which they provided cover. In March 1943 Liddell noted in his diary:

  I had a talk this morning with Hollis about Soviet espionage. There is no doubt to my mind that it is going on and sooner or later we shall be expected to know all about it. On the other hand if we take action and get found out there will be an appalling stink.91

  MI5 decided not to risk ‘an appalling stink’. Blunt reported to his case officer that the Security Service had no agents inside the Soviet embassy and that even surveillance of callers at the embassy had been suspended. Only telephone calls were monitored. MI5 energies were overwhelmingly directed against Nazi Germany. Instead of welcoming this news, the Centre was incredulous. Convinced that intelligence operations against Soviet targets must be as high a priority for British intelligence as operations in Britain were for the Centre, it concluded that Blunt (like others of the Five) must be deceiving them. ‘Our task’, the Centre instructed its London resident, Anatoli Gorsky, ‘is to understand what disinformation our rivals are planting on us.’ Modrzhinskaya concluded in October 1943 that ‘all the data’ indicated the Cambridge Five were part of an organized deception mounted by British intelligence. In reality, much of the data indicated the opposite. Before the great victory of the Red Army at Kursk in June 1943, the GRU had reported that ULTRA intelligence on German operations forwarded by Blunt from Leo Long, his sub-agent in military intelligence, was ‘very valuable’. Most of it was later ‘confirmed by other sources’. Further ULTRA intelligence from Long, the GRU concluded, was ‘highly desirable’.92 The Centre’s conclusion that the Five were trying to deceive it derived not from a rational assessment of the intelligence they supplied but from its own paranoid tendencies.

  To try to discover the exact nature of the British deception, the Centre sent an eight-man surveillance team to London to trail the Five and other supposedly bogus Soviet agents in a vain attempt to discover their contacts with their non-existent British controllers. The team, all conspicuously dressed in Russian clothes and unable to speak English, were inevitably and hilariously unsuccessful.93 The Centre’s suspicions of the Five did not disappear until Operation OVERLORD, the D-Day landings of 6 June 1944.94 On 26 May that year Blunt passed on a complete copy of the entire deception plan devised as part of OVERLORD. On 7 July he provided a comprehensive account of B Division’s role in the deception and, in particular, its use of double agents.95 The stress of his own double life took a greater toll on Blunt than on the rest of the Five. He was under such visible strain that the Centre did not object to his decision at the end of the war to return to his career as an art historian and accept appointment as Surveyor of the King’s Pictures.96

  F Division was acutely aware that both its successful wartime prosecutions of Soviet spies – the Green and Springhall cases – were the result of outside leads, and that there must be a series of others it had failed to detect. It was worried also by the commitment shown by both men and by their sub-agents. The head of the Security Executive, now Duff Cooper, wrote to Churchill in October 1943, reflecting the MI5 view: ‘These agents differ fundamentally from the very poor type employed by the Germans who belong to the dregs of civilisation. Communist agents are intelligent and are inspired by altruistic, idealistic motives. Their first duty is to the Soviet Union.’97 The senior MI5 officer most alert to the continued threat from Soviet espionage was Roger Hollis, who as F2 was responsible for monitoring Communism and other left-wing subversion. Hollis had regarded the main SIS Communist expert, Valentine Vivian, head of Section V until January 1941, ‘with the veneration of a pupil for a master’,98 but disliked his successor, Felix Cowgill. Philby, who worked under Cowgill, told the Centre that he was shy but combative, had ‘few social graces’ and was unable to delegate. In one respect, however, Philby likened Cowgill to Karl Marx: both smoked pipe tobacco ‘in prodigious quantities’ (an irreverent comparison which cannot have amused the NKVD).99 In April 1942 the Security Service, which had hitherto believed it received all ISOS material, discovered that Cowgill had been withholding Abwehr decrypts containing references to SIS agents, thus depriving the Service of a large amount of relevant intelligence. Philby ingratiated himself with MI5 by attacking Cowgill’s failure to share all Abwehr decrypts with it.100 Hollis told Petrie in April 1942 that he probably saw more of the product of Section V than Vivian (then Deputy Chief of SIS). But ‘To be honest even I do not see much.’ Hollis was highly critical of what he did see:

  I can think of no document produced by SIS which makes it appear that international communism had been seriously studied by them since the outbreak of war. We, on the other hand, have started, rather belatedly, to follow the activities of the Comintern wherever it appears, and it was we who drew the attention of SIS to the growth of a settlement of refugee Comintern leaders in Mexico, and not SIS who told us . . .

  It may be said that our job is internal security, and that we are going outside our charter. To that I should misquote Litvinoff and say ‘Communism is indivisible.’ You cannot study the CPGB and the refugee communists in this country without knowing what the Comintern is doing.101

  After the signing of the 1942 British–Soviet Treaty, which bound both powers not to make a separate peace with Germany, Petrie circulated a memorandum by Hollis warning that the Soviet Union and international Communism had not changed their spots: ‘Once Russia is safe and out of the battle, our Communists will seize every chance to make trouble, as they did in the past.’102

  Within the Security Service the main pressure for keeping record cards on all Communist Party members came from Hollis, who argued that Soviet intelligence often selected rank-and-file Party members with ‘clear records’ for espionage.103 The more than threefold wartime expansion of the Party from the 15,000 members at the end of the era of ‘revolutionary defeatism’ in June 1941, however, made it impossible for the Registry to cope, and Hollis had to set
tle for selective registration.104 In the final stages of the war Petrie had a series of meetings with Hollis to discuss the post-war threat from Soviet espionage. On the morning of 5 September 1945, they discussed, at length, the ‘leakage of information through members of the Communist Party’.105 Their meeting turned out to be remarkably well timed. Later the same day a GRU cipher clerk in Ottawa, Igor Gouzenko, defected with dramatic evidence of the ‘leakage of information’.106

  Hollis’s most remarkable insight into Soviet intelligence penetration was that, probably alone within the Security Service, he became suspicious of Blunt. Philby later recalled that ‘Hollis was always vaguely unhappy about him.’107 After Blunt eventually confessed to working as a Soviet agent, he told his interrogators, Peter Wright and Arthur Martin, ‘I believe [Hollis] disliked me – I believe he slightly suspected me.’ Blunt recalled one particularly dramatic example of Hollis’s suspicions. After Gouzenko had revealed the existence of an unidentified Soviet agent codenamed ELLI, Hollis turned to Blunt and said, ‘Isn’t that so, ELLI?’108 It was sadly ironic that Wright and Martin, the most damaging conspiracy theorists in the history of the Security Service, should later persuade themselves that the unidentified Soviet agent was Hollis himself.109

  3

  Victory

  In the middle of the Second World War, the Double-Cross System, like the British war effort as a whole, moved up a gear. Masterman later listed seven main aims which the double agents were intended to serve:

  (1) To control the enemy [espionage] system, or as much of it as we could get our hands on

  (2) To catch fresh spies when they appeared

  (3) To gain knowledge of personalities and methods of the German Secret Service

  (4) To obtain information about the code and cypher work of the German Service

  (5) To get evidence of enemy plans and intentions from the questions asked by them

  (6) To influence enemy plans by the answers sent to the enemy [by the double agents]

  (7) To deceive the enemy about our plans and intentions.1

  It was not until the summer of 1942 that the Twenty Committee began to make its priority strategic deception (the deception of the enemy high command, not merely of its forces in the field). Tar Robertson informed the W Board on 15 July:

  It is reasonably certain that the only network of agents possessed by the Germans in this country is that which is now under the control of the Security Service . . . The combined General Staff in this country have, in MI5 double agents, a powerful means of exercising influence over the OKW German High Command.2

  Strategic deception began not in the European theatre but in the Middle East, where early in the war the British Commander in Chief, General Sir Archibald Wavell, appointed an intelligence officer, Lieutenant Colonel Dudley Clarke, to devise deception plans.3 Clarke’s ‘A Force’ set the tone for deception campaigns throughout the Middle East, and ultimately for all other theatres during the war. The official historian of British wartime deception, Sir Michael Howard, concludes:

  A small acorn planted by Dudley Clarke in December 1940 in the shape of a few bogus units in the Western Desert was to grow into a massive oak tree whose branches included the non-existent British Twelfth Army in Egypt (and the barely existent Ninth and Tenth Armies in Syria and Iraq) and the First United States Army Group (FUSAG) in the United Kingdom.4

  Dudley Clarke’s capacity for deception extended even to his still mysterious private life. In November 1941 Kim Philby, who enjoyed retailing personal scandals to Soviet intelligence, reported that Clarke had been arrested in Madrid, dressed in women’s clothing.5 Philby occasionally perplexed Moscow, probably after heavy drinking sessions, by sending, along with high-grade intelligence, bizarrely improbable scandal such as a report that Germany was infiltrating cocaine and other hard drugs, probably by parachute, into the Irish Republic, whence they were smuggled into Britain by Welsh fishermen in motor launches and supplied to London clubs where RAF officers ‘under the influence of drugs, alcohol, sexual orgies or Black Mass are induced to part with information’.6 By contrast, Philby’s report of Clarke’s transvestite tendencies, which led to his temporary imprisonment in Spain, was quite correct.7 Despite his personal eccentricities, Clarke’s deception operations in the Middle East inspired the creation of the London Controlling Section (LCS), headed from May 1942 by Lieutenant Colonel J. H. Bevan, in order to ‘prepare deception plans on a worldwide basis with the object of causing the enemy to waste his military resources’. Though given the grand title of ‘Controlling Officer’, Bevan lacked executive authority; his role was to plan, co-ordinate and supervise.8

  Strategic deception co-ordinated by the LCS was central to the first major Allied offensive of the war: Operation TORCH, the invasion of French North Africa, for which planning began in July 1942. The two main deception plans devised by Bevan, OVERTHROW and SOLO 1, successfully persuaded the Germans that Allied preparations for landings in, respectively, northern France and Norway were at an advanced stage.9 Throughout the deceptions, the LCS maintained close contact with the Twenty Committee and B1a. Eight double agents were used to pass disinformation to the enemy.10 The most inventive disinformation came from the Spanish double agent GARBO and his full-time case officer, Tomás ‘Tommy’ Harris,11 the bilingual son of an English father and Spanish mother, who formed one of the most creative and successful agent–case-officer partnerships in MI5 history.12 Harris had established himself before the war as a wealthy London art dealer, artist and socialite, and had been recommended to MI5 early in 1941 by Anthony Blunt, with whom he shared artistic interests.13 Throughout the war Harris and his wife kept open house at their Mayfair home, with generous supplies (despite wartime rationing) of champagne and canapés for friends in the intelligence and art worlds: among them Blunt, Guy Burgess, Kim Philby, Guy Liddell, Dick White, Victor Rothschild, Bond Street art dealers and Sotheby’s auctioneers. Harris’s friendship with three leading Soviet agents did not impair, though it adds piquancy to, his operational effectiveness in MI5.14

  Before Operation TORCH, GARBO sent the Abwehr numerous reports from fictional sub-agents invented by Harris and himself. Sub-agents in Scotland reported on mountain warfare training for Canadian, Scottish and Norwegian troops which pointed to preparations for an invasion of Norway. GARBO’s contacts in the Ministry of Information were said to have revealed that officially inspired rumours of an expedition to Dakar were intended to distract attention from an attack elsewhere, possibly Norway or France. On 29 October GARBO reported, correctly, that a convoy had just set sail from the Clyde. Three days later, on 1 November, he informed the Abwehr, also correctly, that intelligence from the Ministry of Information pointed to an Allied invasion of French North Africa. B1a, however, arranged for the letters containing these reports to be delayed in the post. They did not reach GARBO’s case officer until 7 November, a few hours before the Allied landings and after the invasion force had already been spotted by the Germans. It did not occur to the Abwehr to blame GARBO for the delay or to suspect the involvement of British intelligence. ‘Your last reports are all magnificent,’ they told him, ‘but we are very sorry they arrived late.’15 General Alfred Jodl, Chief of the Operations Staff of the German High Command (OKW) and Hitler’s closest military adviser, told Allied interrogators after the war that the landings in North Africa had come as ‘a complete surprise’.16

  No agency in British history, probably none in the history of intelligence, had ever devised such a wide range of ingenious deceptions with such a high success rate as B1a during the Second World War. Most new recruits to the section were so enthused by the ethos created by Tar Robertson that they learned the art of deception with unusual speed. Among them was the twenty-five-year-old Oxford-educated Flight Lieutenant Charles Cholmondeley, who joined MI5 from the Intelligence Directorate at the Air Ministry in 194017 and was later described by Tar as ‘a most extraordinary and delightful man who worked in my section largely as an ideas man’.18 Roberts
on’s less enthusiastic deputy, John Marriott, found him ‘an incurable romantic of the old cloak and dagger school’.19 The most ingenious cloak-and-dagger deception devised by Cholmondeley, soon after the beginning of TORCH, was a scheme to plant bogus documents on the enemy designed to mislead them about the target of a forthcoming Allied operation in the Mediterranean:

  A body is obtained from one of the London hospitals (normal peacetime price £10). It is then dressed in uniform of suitable rank. The lungs are filled with water and the documents disposed of in an inside pocket. The body is then dropped by a Coastal Command aircraft at a suitable position where the set of the currents will probably carry the body ashore in enemy territory . . . Information in the form of the documents can be of a far more secret nature than it would be possible to introduce through any other normal B1a channel.20

  Cholmondeley quickly won the support of Lieutenant Commander Ewen Montagu, the naval representative on the Twenty Committee (whose pro-Russian brother Ivor was later identified as a Soviet agent). On 4 February they informed the Committee that a suitable corpse had been found (that of Glyndwr Michael, a homeless Welshman who had died by ingesting rat poison), and won approval for what became known as Operation MINCEMEAT. Michael was given the fictitious identity of Major William ‘Bill’ Martin of the Royal Marines, an officer on the staff of the Chief of Combined Operations, Vice Admiral Lord Louis Mountbatten.21

  Some modifications were made to Cholmondeley’s original plan. In order to ensure that the bogus official documents on the body were not overlooked, they were placed in a briefcase attached to the belt of ‘Martin’s’ trenchcoat rather than in his pockets, which contained personal items and letters from his fiancée. It was also decided that, instead of being dropped into the sea from an aircraft, the corpse should be taken by submarine to a point near Huelva on the Spanish coast where the local currents could be relied on to wash it ashore.22 As controlling officer of the LCS, Colonel Bevan called on Churchill at 10 a.m. on 15 April to gain his consent to MINCEMEAT. He found the Prime Minister sitting up in bed, smoking acigar, ‘surrounded by papers and black and red Cabinet boxes’. Churchill quickly gave his enthusiastic support for the deception (subject to the agreement of General Dwight D. Eisenhower, Allied commander in the Mediterranean theatre, which was also obtained). If the operation did not succeed at the first attempt, he said, not entirely seriously, ‘we shall have to get the body back and give it another swim.’23