The Defence of the Realm (16 page)

German intelligence seems not to have realized, however, that
all
Müller's reports since the end of February had been bogus. Hinchley Cooke later recalled that MO5(g)'s attempt to continue the deception was finally abandoned when the Kriegsnachrichtenstelle summoned Müller back to Rotterdam to receive further instructions. The money sent from Antwerp to Müller enabled MO5(g) to buy a much needed second car, a two-seater Morris which was promptly christened the ‘Müller'.
⁶⁶
Maurice (later Lord) Hankey, secretary of the Committee of Imperial Defence, who in December 1916 became Britain's first (and so far longest-serving) cabinet secretary, later recalled, while carrying out a review of British intelligence in 1940: ‘During the last war I myself was driving an automobile bought by the MI5 and lived on a salary paid by the Germans for . . . imaginary services.' He also recalled that MO5(g) had ‘troubles with the Treasury' over its unauthorized expenditure of German money.
⁶⁷
According to Hinchley Cooke, ‘the Treasury sent a full-dress letter to the Army Council pointing out the irregular use of funds which should have been paid into the Exchequer.'
⁶⁸

The Müller deception anticipated in some respects the Double-Cross System, MI5's greatest success in the Second World War. Kell's inability to keep the arrest and trial of German spies secret, however, meant that the deception was bound to be short lived. The government was too anxious to prove to a sometimes sceptical public that it was catching German spies to suppress news of its triumphs. Even had the Müller deception continued for longer, however, its success would have been limited. Unlike the turned German agents of the Second World War, Müller was executed and therefore unable to add personal credibility to the disinformation sent to German intelligence in his name. The Double-Cross System also depended not on one individual but on a series of turned German agents as well as on the co-operation of the whole intelligence community to provide a mixture of information and disinformation capable of both impressing and deceiving the enemy for the remainder of the war. Unlike MI5 in the Second
World War, MO5(g) lacked the regular flow of SIGINT (signals intelligence) which enabled it to monitor the impact of the deception.
⁶⁹

The discovery by German intelligence of the Müller deception made it difficult for MO5(g) to attempt a similar deception without arousing German suspicions. In the summer of 1915, however, before it realized that its deception had been detected, Kell had one further opportunity. Early in May the postal censors had discovered in a mailbag from Denmark a letter addressed to Berlin which had been wrongly included in the London post. The letter was from one Robert Rosenthal in Copenhagen to ‘Franz Kulbe' at an address in Berlin which MO5(g) knew to be used by German naval intelligence. It also knew that ‘Franz Kulbe' was an alias employed by Captain von Prieger of the German Admiralty. Though the letter purported to be a business communication, the Censorship flat-iron revealed writing in secret ink which disclosed that he was about to leave for England disguised as a travelling salesman of cigar lighters. When Rosenthal was arrested in Newcastle, travelling on a US passport, no incriminating evidence was found on him. But when confronted with his intercepted letter to Berlin, he admitted he was German and had been sent by Prieger to spy on the Royal Navy. Though he denied he had sent Prieger information of any value, he also admitted that he had been on two previous wartime espionage missions. Like some of the German spies sent to Britain before the war, Rosenthal had a criminal record, having been imprisoned for forgery in Germany and subsequently spending five years in the United States.
⁷⁰

After Rosenthal's confession he quickly offered to become a double agent and, to prove his willingness to do so, provided information on secret inks, secret codes and the methods of passport forgery employed by German intelligence.
⁷¹
Kell rejected the offer, probably influenced by memories of the pre-war Graves case, when his first double agent (who, like Rosenthal, had a criminal record) had successfully deceived him. As the Müller deception shows, Kell thought it safer for MO5(g) to impersonate a dead German agent than to turn a living one. Rosenthal tried desperately to save his life by claiming that his loyalties were to the United States, not to Germany: ‘I indeed have never cared for any German, never liked Germans, always wanted to be a real Yankee.'
⁷²
He went to the scaffold at Wandsworth military prison on 15 July, pleading for his life and, according to an officially authorized account, ‘gave unutterable disgust to the authorities by his lack of common courage'. The commandant called him a ‘cur'.
⁷³

The use of invisible ink by German spies prompted Kell to make his first move into forensic science by seeking the assistance in 1915 of
Henry Vincent Aird Briscoe, an inorganic chemist at Imperial College, London. He and his laboratory devised a series of techniques for the detection of invisible ink in letters and the increasingly sophisticated chemicals used in secret writing.
⁷⁴

During interrogation Rosenthal had revealed that, instead of recruiting German nationals as spies, German intelligence intended to make more use of agents from neutral countries disguised, like himself, as commercial travellers. Kell responded by ordering detailed censorship of letters and cables from Britain to neutral countries. This ‘super-censorship', as it was known in MO5(g), revealed a telegram dated 25 May sent from Southampton by a Dutchman Haicke Janssen, posing as a travelling cigar salesman, to Dierks & Co. in The Hague (an address already known to be used by German intelligence), ordering 4,000 Sumatra cigars ‘mark A.G.K.', which MO5(g)'s existing knowledge of German codes enabled it to identify as a reference to four
Alte Grosse Kreuzer
(old-model large cruisers), which inquiry showed were indeed docked at Southampton. A further reference to 4,000 of a brand of Sumatra Havanas was thought to refer to four torpedo boats, also present in Southampton Harbour. A request from Kell to the Chief Censor for copies of all other cables sent to Dierks & Co. revealed two more sent from Edinburgh by another purported Dutch cigar salesman, Willem Roos.
⁷⁵

Though later to become essential to Security Service operations, intelligence liaison during the First World War was limited by the determination of the French and British armed forces to prevent intrusion by their allies into their own spheres of operations. French intelligence, however, made a major contribution to the Janssen case.
⁷⁶
On 15 June Colonel Cartier, the head of French military SIGINT, handed to Major (later Major General Sir) Walter Kirke, an intelligence officer at British GHQ in France, decrypted German telegrams giving some details of agents in British ports. One of the decrypts clearly identified Janssen, then – like Roos – under arrest, and Kirke noted with satisfaction that ‘it should ensure his being shot'. Cartier travelled to London soon afterwards for discussions with Kell.
⁷⁷

After Janssen and Roos had been sentenced to death on 16 July, both tried, like Rosenthal, to save their lives by offering to change sides. Janssen claimed that his sympathies had really been with Britain all the time and provided intelligence on the German espionage network in Holland which assisted the discovery of subsequent German agents. Roos feigned madness in a further attempt to avoid execution. Both were shot by firing squad in the Tower of London on 30 July. Kell seems never to have contemplated using either as a double agent. His wife's judgement probably reflected his
own: ‘It was all done for money and therefore Janssen and Roos were despicable men ready to do any dirty work merely for gain.'
⁷⁸

While Janssen and Roos were awaiting trial, the head of Cumming's Rotterdam station, Richard Tinsley (codenamed ‘T'), a tough former naval officer and shipping manager described by one of his staff as ‘a combination of sea captain and prize fighter',
⁷⁹
reported that Josef Marks, a German agent travelling on a US passport, was shortly to leave for England. Marks was arrested on his arrival at Tilbury on 18 July, admitted he had been sent to gather intelligence on munitions production and revealed the codes he had been told to use, but claimed that he had been put under heavy pressure to undertake the mission. On 28 September he was sentenced by court martial to five years' penal servitude.
⁸⁰
Intelligence from ‘T' led to the detection of four other German agents in 1915
⁸¹
and one in 1916.
⁸²

The high point of MO5(g)'s counter-espionage operations came in June 1915, when seven German spies were rounded up in little more than a fortnight.
⁸³
Thereafter, German intelligence operations in Britain were drastically scaled down. German archives show that the number of agents in Britain declined from twenty-two in January 1915 to only four at the end of the year.
⁸⁴
For the remainder of the war Germany relied largely on brief visits by bogus commercial travellers from neutral countries who carried as much information as possible in their heads rather than on paper. They had little success. Kell's Bureau (renamed MI5 in January 1916) concluded that most had neither the aptitude nor the training required for successful espionage. The last to be executed was Ludovico Hurwitz y Zender, a Peruvian of Scandinavian descent who aroused the suspicions of the censors by sending bogus orders for large quantities of Norwegian sardines at the wrong season, and was shot in the Tower on 11 April 1916.
⁸⁵

Shortly after Hurwitz y Zender's execution, MI5 recruited its most successful double agent of the war, an American in Holland codenamed COMO, described in what remains of his file as ‘working in Holland as a German Agent, but double-crossing for us'. In May 1916 COMO was sent to Britain by the Kriegsnachrichtenstelle to look up agents who had been out of contact for some time. He reported to MI5 that the first of those he contacted, Fritz Haas, was ‘harmless and probably never an agent, if so unknowingly'. In July COMO was sent on a second mission to obtain intelligence on Canadian troops in England and when they were likely to be sent to France, as well as details of British naval losses at the battle of Jutland. According to an MI5 note:

We gave ‘COMO' information about our Canadian Divisions – their numbers and positions, etc – at places not obviously incorrect, and gave particulars of their numbers etc, as would make them plausible. Also, fictitious information of Naval, Industrial and Political situation, most convincingly written. Also sent information regarding the effects of the Zeppelin raids on London.

In September, on MI5 instructions, COMO sent further disinformation on the military situation in England and likely movements on the Western Front. He told the Germans that the British were planning an attack on the Belgian coast on the 15th. COMO continued to pass on at irregular intervals details of the intelligence missions entrusted to him by the Kriegsnachrichtenstelle, and MI5 continued to channel disinformation through him. According to a post-war summary, COMO ‘was always quite honest and trustworthy and did some v[ery] valuable work for us, at the same time holding the German's [
sic
] confidence'. War Office nervousness, however, prevented the development of a full-blown Double-Cross System. On at least one occasion, the Director of Military Intelligence described the questions put to COMO by the Kriegsnachrichtenstelle as ‘extremely dangerous', and instructed that ‘no answers could be given'.
⁸⁶

The final attempts by German intelligence to establish a significant agent network in Britain were based on the recruitment of Americans. On 3 June 1916 Tinsley in Rotterdam supplied MI5 with a cover address in The Hague used by German intelligence, which Kell immediately ordered to be put under censorship.
⁸⁷
This was reckoned by MI5 to be ‘one of “T'”s best efforts',
⁸⁸
since it led to the discovery of an ‘important gang' of spies run from New York by Karl Wünnenberg, a German naval reserve officer, and the journalist Albert Sander.
⁸⁹
To strengthen their cover in New York, Wünnenberg and Sander had no contact with German officials in the United States and were run by the Antwerp Kriegsnachrichtenstelle. They succeeded in recruiting six US journalists to work as agents, paying each up to $1,000 in advance and promised another $125 per week.
⁹⁰
The large sums offered are an indication of how short of agents in Britain the Kriegsnachrichtenstelle had become.

On 29 September 1916 the censors opened a letter written from England a week before by one of the US journalists recruited by German intelligence in New York, George Vaux Bacon, to the German intelligence address in The Hague identified by Tinsley. There was a further delay before Major Carter of MI5 examined the letter on 9 October, by which time Bacon had been in Holland for over a fortnight and realized his letter had been delayed by censorship. Bacon's suspicions were further aroused when he was
approached, ‘somewhat clumsily' in MI5's view, by one of Tinsley's agents. Though MI5 believed that Bacon ‘suspected that the British authorities were on his tracks', he none the less returned to Britain on an espionage mission early in November. MI5 noted that he had a ‘loose' (promiscuous) lifestyle, but ‘nothing suspicious' was found in his possession on his arrival and he managed to ‘elude his watchers' while visiting naval bases in Ireland. On 9 December 1916 Bacon was interviewed at Scotland Yard, not with the expectation of persuading him to confess to espionage but in the hope that he could be ‘frightened out of the country'. However, this time enough incriminating material was found in his possession for him to be kept in custody. In addition to equipment for secret writing, ‘two of his socks were found to produce invisible ink when soaked in water'.
⁹¹