Fighting to Lose

by John Bryden

  Meanwhile, on August 19, Felix Cowgill of MI6(V) sent the full German-language version of the Popov questionnaire to MI5’s Major Robertson. He did not mention there was an English-language version, so MI5 went ahead with its own translation.23 This translation was reproduced and published for the first time in Masterman’s Double-Cross System.

  On August 28, the XX Committee discussed the questionnaire. The minutes of the meeting make no mention of its Hawaii/Pearl Harbor content.

  On September 6, Menzies asked for a meeting of the Wireless Board. Without mentioning Popov specifically, he said he wanted to discuss how to handle the questionnaires of British double agents operating from the United States, proposing that questions relevant to Britain be dealt with by British officials in Washington rather than in London. The board met in Room 206 of the War Office on September 10.24

  It will be helpful to recall at this point that the Wireless Board was an ad hoc committee of the three service directors of intelligence, plus Menzies of MI6, Liddell of MI5, and Sir Findlater Stewart, representing the civilian authorities. It did not officially exist within the chain of command, its members normally took no paper to or from meetings, only the chairman retained a copy of the minutes, and there was no formal reporting mechanism to the chiefs of staff, the War Office, or the Foreign Office. Each committee member was individually responsible for telling (or not telling) those above him what transpired in the meetings.

  Three of the men who were there had definitely seen the whole questionnaire: Menzies, Cowgill, and Robertson.

  The discussion turned to Menzies’s letter. Members approved his proposal. Cowgill then read out the answers to those parts of Popov’s questionnaire that were pertinent to the British. Liddell suggested that Popov had much promise as a double agent in America. Presumably Admiral Godfrey, the director of Naval Intelligence, was told about the information being sought on the Pacific Fleet. It would have been highly improper not to have disclosed that detail. Nevertheless, the minutes taken by Ewen Montagu make no mention of Hawaii or Pearl Harbor.25

  Nothing went up to the chiefs of staff, or to the War Office, or to the Joint Intelligence Committee, or to the Foreign Office.26 The British did not forward the Hawaii portion of Popov’s questionnaire to the Americans because it never got outside the circle of MI5/MI6, the XX Committee, and the Wireless Board.

  On December 17, when the oily water of the devastated Pacific Fleet still lapped the Hawaiian shore, this entry appeared in Liddell’s diary:

  TRICYCLE’s questionnaire is now in our possession. It shows quite clearly that in August last the Germans were very anxious to get as full particulars as possible about Pearl Harbor.

  There is something wrong here. The original copy of the German-language version of the questionnaire had been sent to Major Robertson on August 19, then translated by MI5 and discussed at the XX Committee on August 28. Its implications were debated by the Wireless Board on September 10, with Robertson, Montagu, Menzies, Cowgill, and Liddell present. The German original is still to be found in an MI5 file beside MI5’s original translation. There is no way the head of MI5’s B Division could honestly suggest that the questionnaire only “came into our possession” after the Pearl Harbor attack.27

  Liddell’s diary was kept in a series of ring binders. It still is. Ring binders make it possible to take pages out and put pages in, anywhere, anytime, without a trace. The entry for December 17, 1941, cannot be authentic.

  But who, then, wrote it? And why?

  14

  September–December 1941

  Contrary to British accusations during and after the war, the FBI made a sincere effort to run Dusko Popov as a double agent. Within the first week of the FBI taking him over, Charles Ellis was asked to prepare his first invisible-ink letter reporting to the Germans that he had arrived safely.1 It was probably thought that Ellis had a better chance of composing it in proper Englishman’s English, even though he was an Aussie. Popov’s ink was ammonium chloride, easily developed by passing a warm iron over the paper.

  For some time there was no response. Then, on October 25 and October 30, Popov received a letter and then a cable in open code indicating that he should proceed to Rio de Janeiro. This was in response to a suggestion by the FBI that he propose to meet with German agents in South America to turn over some photographs and notes he had made. So far, no spy in the United States had come forward to collect the questionnaire, and no one had offered to be his radio operator. On the other hand, the FBI had been tracking the activities of German agents in Brazil since the spring through decrypts of their wireless traffic supplied by the small cryptanalysis unit attached to the U.S. Coast Guard and by Canada’s Examination Unit.2

  Next, a lengthy letter arrived from “Mady,” Popov’s supposed fifteen-year-old girlfriend in Portugal. Much of it seemed innocent chit-chat, but then:

  My dear uncle … has been travelling abroad, but has returned now. He too sends you a lot of greetings because, as you well know, he is very fond of you. I was very anxious about Dicky, but he is really a nice chap. I got a letter from him some days ago. I would be glad if we could arrange to meet all of us together in some nice place.…3

  When the FBI asked Popov who this “Dicky” might be, he quickly replied that it was “the name of one of the British intelligence officers in London who was acting as a German spy for that organization.”4 And so there was; except another person nicknamed “Dickie” was standing right there at Popov’s elbow. One can imagine Charles Ellis’s forehead beading with perspiration.

  It may have been a threat, of course, but it also may have been an open-code instruction for Ellis to attend a meeting set up by his Abwehr contacts. It is known that Ellis flew to London on November 2.5 It is not known whether he went on to Portugal.

  On November 16, Popov left Miami for Rio. Ten days later, the FBI had the satisfaction of reading the following intercept:

  24 November 1941 CEL to ALD

  No. 46. ivan ten thousand deciphered. Receiving his news this evening. Regarding cable via B. Firstly, I can give ivan apparatus 50 Watt. Secondly, in view of lost mail, shall ivan send mail here in future? Can give him other cover addresses.

  alfredo6

  There were other similar messages intercepted. The FBI could be forgiven for thinking that Popov’s trip to South America was working out just fine.

  The FBI had another German agent in its net late that autumn of 1941. Even as Popov was making arrangements to go to Brazil, another potential double agent was coming the other way.

  It began the preceding July, just after the Spaniard, Juan Pujol Garcia, offered his questionnaire to the British in Lisbon and just before Popov set out for the United States with his. An Argentinian named Jorge Mosquera presented himself to the American embassy in Montevideo, Uruguay, to declare that he had been recruited by the German secret service to go to the United States to spy. He had letters of reference from high Nazi officials, three microdot photographs of instructions, the names of individuals he was to contact in New York, and an earnest desire to betray his German masters.

  According to his identity papers, he was born in Rosaria, Santa Fe Province, Argentina, in 1895. He said he was the youngest of ten children and that his father had been born in Spain. In 1924 he had moved to Germany where he ran a small import-export business until it was closed down due to the war.7 He spoke Spanish and German only. No English. He had been recruited by a man named Hans Blum of Ast Hamburg. On hearing all this, the FBI said he could carry on to the United States. Two FBI men were waiting for him when his ship docked in New York. It was November 18.

  He was put up in a hotel, placed under blanket surveillance, and questioned intensely. The report on him that went to Hoover ran to forty-eight closely typed pages. The profile that emerged was typical of the double agents already acquired by the British.

  As with Popov, and Arthur Owens before him, Mosquera’s cipher was of the simple transposition type, full of complications but easy
to break. Like them, the keyword was derived from a book, in his case a Spanish one, Los majos de Càdiz; and like Owens, he had brought it with him. Also like the others, someone else was needed to operate his transmitter. He had failed the wireless telegraphy course at spy school, he explained.

  His invisible ink was phony. Based on zinc and sulphur, it was more complicated than the one given Popov, but still practically useless. The description of how it was to be used filled two pages of the FBI’s report, which concluded by noting that only linen paper could be used with it, because the chemical solution reacted with wood pulp. It took no time for the lab to find a developer.

  The contacts Mosquera was to make in the United States also turned out to be dead ends. Despite extensive investigation over many months, hidden cameras, microphones, and staged meetings with Mosquera, the persons whose names he had been given proved to be nothing more than naturalized German-Americans with normal sympathy for their homeland. They showed no sign that they were involved in espionage and appeared horrified when Mosquera hinted that he was.

  Much, much later, the FBI was to learn that a Spanish-speaking Abwehr spy originally on loan to Franco was “in America” and usually travelled on an Argentinian passport.8 If this was Mosquera, the connection with the Spanish dictator would have made him one of Canaris’s especially trusted agents. At the time, however, the Bureau had no reason to be suspicious. Up to that point, the British had not shared the details of their own experiences in running double agents, and they had given the Americans the false impression that Popov had led to the uncovering of whole networks of spies.

  The highlight of Mosquera’s recruitment was his microdots. The FBI had seen Popov’s only two months earlier; now here was another set, containing an even longer questionnaire — one focused on American aircraft production and new weapons technology. It included some novel items, including questions on poison gas, and this puzzling line: “Deduce artificially the uranium or other alloy which may be substituted therefore as an atom destructor.” The awkward English was due to Mosquera having translated the original German into Spanish for the microdots.

  When he was asked to elaborate on this item, he reported having a conversation with his Hamburg spymaster (blum). The following is taken from a report by Foxworthy to his director in December:

  blum further stated to mosquera that he should not limit himself to the instructions on the microphotographs but should give attention to other details, especially details which pertain to experiments performed in the United States relative to the shattering of the atom.

  blum stated that the German military authorities believed that a great future lies in the developing of high explosives derived through experiments pertaining to the shattering of the atom.

  According to blum, if any success is gained in the shattering of the atom and high explosives produced as a result thereof, the future high explosive bombs would not have a gross weight in excess of one and one-half pounds. blum also stated that the nation which will be victorious in this war will be the one which accomplishes the task of shattering atoms and applying the results thereof.9

  There is good evidence as to what at this time was prompting the Abwehr’s interest in things nuclear. While most of the files at the Abwehr’s Asts and branch offices were deliberately burned toward the end of the war, the card index of Nest Bremen survived. The card for R-2232, identified as the spy Hans Dahlhaus, describes him as having submitted a report on U.S. atomic experiments on July 29, 1941, not long after he had returned to Germany from an extended espionage tour of the United States. He had posed as a tobacco salesman and had developed a considerable network of sub-agents.10

  His report is lost, but one can make a good guess as to some of its content. Up until the fall of 1940, before a publication ban took effect, the scientists working on sustainable nuclear fission published their findings. The March and April 1940 issues of Physical Review carried articles that identified the lighter isotope of uranium, U-235, as being most likely to split in an ongoing chain reaction leading to a massive burst of energy and, theoretically, to an explosion of unprecedented magnitude. The problems were how to separate enough of the isotope from natural uranium and how to prove the chain-reaction effect without blowing oneself up. This much was accessible to Dahlhaus even without having an agent inside the relevant scientific circles.11

  What the FBI made of what blum said is not recorded. Uranium was featured in the popular press as a kind of miracle super-fuel, but the concept of releasing energy by splitting atoms in lightning-speed chain reactions leading to colossal explosions had not yet got much beyond a tiny circle of mathematicians and physicists. One would think that the FBI was still well out in the wilderness as to what blum was talking about.

  Others were not. Roosevelt had been encouraging research on uranium since 1939, and on October 9, 1941, received a briefing on the prospects of developing a uranium super-bomb from his scientific adviser, Dr. Vannevar Bush. He was told of British enthusiasm for U-235, that ten kilograms should be enough to flatten a city if the predicted runaway nuclear chain reaction took place, and the technology to separate that much of the isotope from regular uranium was likely to be hugely expensive. The discussion touched on how little was known of what the Germans might be doing. The president told Bush to do a cost analysis of what it would take in scientific and industrial organization to prove that a U-235 chain reaction was feasible. Roosevelt would then decide what to do next.

  In the meantime, strict secrecy would prevail. Knowledge of Bush’s assignment would be restricted to the vice-president, Henry Wallace, to the secretary of war, Henry Stimson, to the army chief of staff, General Marshall, and to James R. Conant, chairman of the National Defence Research Council. On November 27, Bush submitted his report.12

  On December 6, two days after the FBI’s report on Mosquera was received by the White House13 and the day before the Japanese attacked Pearl Harbor, Dr. Bush secretly met with a small group of senior scientists. The president had given the order, he told them. There was to be an all-out effort to determine whether concentrated uranium-235 would produce a nuclear chain reaction fast enough to explode. Money was no object. “If atomic bombs can be made, then we must make them first,” Dr. Bush quoted the president as saying.14

  And so it was. The day before the Japanese dropped the first bombs that opened the war with the United States, the Americans began the process toward the two bombs that would end it — one for Hiroshima, the other for Nagasaki.

  Back in England, remote as if on the dark side of the moon from these happenings in America, MI5 was fussing about its entitlements. By taking over the Radio Security Service earlier in the year, MI6 had acquired an absolute monopoly on the distribution of ISOS (decrypts of Abwehr wireless traffic), and Felix Cowgill had clammed up about anything else to do with TRICYCLE. He was unmoved by requests from the XX Committee and deaf to Major Robertson’s plea that he at least give “some small indication” as to how Popov was doing. Liddell had no luck either, Cowgill archly declaring that he was disappointed that his MI5 colleagues did not seem to think he was competent to run a double agent on his own.15

  There was good reason for MI5 to be anxious. It had sent Popov to the United States with the question still unresolved as to whether Arthur Owens had really confessed to the Germans, which would have blown nearly all of MI5’s double agents, including Popov. The folly of notionally linking them together through the payments of Plan Midas had sunk in. As one MI5 officer was to write:

  J.H.M. has advanced the theory that if snow on his last visit to Lisbon blew his traffic as he said he did, it follows as a natural consequence that the Germans realize TATE and RAINBOW to be under control, and further that they regard TRICYCLE as blown and may also believe that balloon and gelatine are controlled agents. The logical consequence is perfectly clear, if the Germans believe that snow’s traffic for two and a half months before his visit to Lisbon was controlled by us they must assume since snow paid TAT
E in that period that TATE has been a controlled agent. They must also assume that RAINBOW, who TATE subsequently paid, is also controlled …16

  B1A staff and members of the XX Committee had been agonizing over the problem for months, chasing the faint hope that somehow it would turn out that Owens had lied about giving everything away.

  In fairness, at the beginning of the summer Robertson had received spectacular confirmation that all of his double agents were okay, including Popov. When snow went off the air for good in April, the Germans appeared to accept s A-3504’s explanation that he was on the verge of a nervous breakdown, and sent back proposing that A-3725 — Schmidt — be the Plan Midas paymaster instead.17 Then, when he supposedly received the £20,000 of German money Popov brought to the United States, the only spies the Germans named for him to pay were RAINBOW, the dance-band spy, and mutt and jeff, the inept pair of Norwegian saboteurs captured in Scotland earlier in the year. Astoundingly, MI5 took this to mean that there were no other spies in Britain to be paid — in other words, that there were no unknown German spies at large.18

  This was very wrong. Most of the Abwehr’s files, at headquarters in Berlin and in all the Asts, Leitstellen and KOs of Europe, disappeared at the end of the war, probably systematically burned as happened at Ast Hamburg.19 Nest Bremen was the exception. Some of its records were salvaged, and they show that beginning in August 1943, and speeding up to September 1944, Bremen officials burned the reports of most of their agents in England and the United States. All that is to be learned of the spies Bremen employed against England — eight, at least, in 1943–44 alone — is the code numbers on their file cards. Occasionally a name emerged, but this was only due to the diligence of the U.S. Naval Intelligence officers who combed through the unburnt files matching fragment to fragment.