Secret Alliances Read online

  The Germans gradually built up extremely effective direction finding organisations throughout occupied Europe. There were two: one military (which covered Denmark among other areas) and the other run by the civil police, whose responsibilities included Norway. Its work was coordinated by a central plotting office in Berlin. Long-range direction finding stations based in Europe often picked up the first signs of transmissions, and were able to provide rough cross-bearings which they passed on to local services for further more detailed investigation. The Germans employed more than 140 people on direction finding in Norway by the end of 1944. They used a range of stations across the country, as well as mobile vehicles and also individual body-borne systems to provide the specific locations of transmitting sets once an investigation was in its final phase. They also sometimes deployed direction finding aircraft and boats, though a post-war report noted that atmospheric and geographical conditions meant that the one Fieseler Storch light aircraft available in Norway for this purpose was never able to get a single cross-bearing for mobile direction finding.25 Poor atmospheric conditions also made direction finding work more difficult in northern Norway – though it also caused problems for transmitting stations, with the unwelcome consequence that they had to come up on the air more often in order to ensure that their messages were received in London. (In the worst of circumstances, it could take days or even weeks before a message was able to reach its destination.)

  It was never easy for either service to find out about the security problems affecting their agents in Norway. The Germans tried, where possible, to conceal the arrests they made, in the hope that a quick and successful interrogation of captured resistance members would enable them to capture more people linked to a particular network who had not yet learned of the danger they were running. When losses and setbacks were reported to London, the information was often fragmentary and incomplete, so it could be difficult to work out the likely cause of the loss of a station. There might be plenty of other possible explanations for what had gone wrong, ranging from betrayal to loose talk, inadvisable contact with another resistance network or sheer bad luck. But as early as the spring of 1941, SIS stations such as Oldell in Oslo and Skylark B were reporting that they had detected signs of German direction finding activity shortly before the Germans made arrests. Cheese reported similar concerns to SOE during his first deployment in June 1941, though he was able to escape in time. There was a steady increase in the number of SIS stations which were lost as a result of direction finding. By 1944 it had risen to twelve, and it is estimated that most of the losses in 1945 were due to the same cause. There is little to show that SIS or their Norwegian colleagues were concerned about the dangers which direction finding posed. In December 1944, General Hansteen and Ragnvald Alfred Roscher Lund wrote to Stockholm that while there was a possibility that any station could be located through direction finding, it was unlikely that such interception could actually locate it precisely. Although it was possible that some stations were lost as a result of direction finding and for no other reason, they were not aware of any such cases.26 In his post-war report, Roscher Lund acknowledged:

  From Norway, there were sometimes reports that our radio agents had been D/Fed. The information available in England, however, indicated that these stations were taken for other reasons. However, after the end of the war, it was found that a larger number of our stations than we knew about had been located by the German D/Fing organisation. Thus, we learned that twelve Norwegian radio stations were taken in 1944 by German D/Fing.27

  One of the disadvantages for SIS was that their coast-watching stations were generally static, needing to cover fixed vantage points, and it was difficult for them to move very far. SOE generally did not suffer from the same problem and their wireless operators were able to be more mobile. They were also more aware of the dangers. For example Knut Haugland, who was dropped on the Hardanger plateau with the Grouse party in October 1942 to prepare for the FRESHMAN attack on Vemork and who later worked in Oslo, commented on German use of direction finding. He wrote in September 1943 that he did not believe that the Germans were using small aircraft in the mountains for direction finding purposes. However, Oslo was another matter. There the Germans had mobile vans, which was more dangerous as they could approach quite close to the house or site from which transmissions were being made.28

  Moreover, when stations were lost, SOE regularly assessed the likelihood that the cause had been direction finding, concluding for example that Crow was located by this means in May 1942, while SOE in Stockholm reported that the same had happened to Lark in December 1942 – indeed the station had been warned about it earlier. SOE advised its radio operators to restrict their transmission times as much as possible, and to move their radios to new locations every three weeks. Although making such a move could be risky, it was considered less dangerous than the risk involved in continuing to operate from the same place for a long period. While this was a counsel of perfection – which could not always be followed – it helped to emphasise the need to take suitable precautions. And, while it did not prevent SOE losses from direction finding, it must certainly have helped to reduce them.

  The way in which German direction finding tracked down an SIS station is illustrated by the capture of the Sabor station, which operated near Stavanger between January and April 1945, providing intelligence about the movements of ships from Stavanger and Egersund and activities at Sola airfield. Before their departure the two agents, Ernst Askildsen and Magne Bakka, had been told by SIS that the Germans would not be able to identify the frequency which they would be using. In fact, after it started transmitting on 9 January, sometimes several times a day, it was picked up in mid-February by a long-range direction finding station in Konstanz, southern Germany. Further bearings were obtained shortly afterwards by German stations in Vienna and Berlin. On 21 February, local fixed stations in Norway began work to narrow down its location, and then started to deploy mobile and body-borne equipment, using bearings from seven different points, to establish that, by then, it was based on an island in Lutsi Lake. A large number of Germans raided the base on 4 April, in the course of which Askildsen was killed and Bakka captured.¶

  German map showing the final stage of the direction finding operation which located Sabor, giving details of the bearings used to find it. © NHM

  Gestapo conference before the operation to capture SIS station Sabor in April 1945. SS-Obersturmführer Friedrich Wilkens, second from the right, was killed during the attack. © NHM

  Magne Bakka, captured and handcuffed after the raid on Sabor. He survived the war. © NHM

  While it has not been possible to establish precisely how many stations were lost to direction finding, there are good grounds for concluding that SIS losses were much greater than those of SOE. SOE’s post-war Norwegian section history concludes rather optimistically that the Germans were never really successful with their direction finding of SOE stations and that although some stations were troubled, there were only a few located and captured.|| 29 But it is certainly clear that SOE were well aware of the threat from German direction finding from some of their own agents. For example, Gunvald Tomstad, who worked closely with Cheese and was for a time his radio operator, joined Quisling’s NS especially on SOE instructions. He became highly regarded and was well placed to be able to ingratiate himself so thoroughly with the police and Gestapo that they confided in him many of their suspicions about local Norwegians. They finally even informed him of their attempts to locate the radio station from which he was transmitting. After months of unsuccessful direction finding, they involved Tomstad himself more directly in their investigations. As a condition for his cooperation, he requested that they should brief him on the various direction finding systems they were using. He continued to operate his set, while taking appropriate precautions. On one occasion, when they got quite close to locating it, he persuaded the Germans to strip the panelling from the whole house next door to his, which was occupie
d by a German official, thereby distracting them and also causing them extra work.30 After his arrival in England in April 1943, Tomstad provided a more detailed report on the direction finding systems used by the Germans.31

  In August 1944, Norwegians working for SOE produced a paper on German direction finding, and ways of countering it, which was based on information obtained mainly from Norway, but also from other occupied countries. While we now know that it was not completely accurate, it described quite precisely in outline how German direction finding worked, and how quickly investigators might be able to detect an illicit signal. It considered that if a station transmitted three times a week, it would not take long before a direction finding service would be able to identify it. (In the case of Sabor, it was about five weeks.) It also drew attention to some German investigative techniques when they were close to a transmitting station, but not yet sure of its precise location – such as cutting off the electricity to certain houses but not others, to see whether a station was still receiving power and continuing to transmit.32

  There is no evidence to show that SOE passed on any of the information about German direction finding capabilities in Norway to SIS, or to explain why it failed to do so. Moreover, despite prompts from some of their Norwegian agents,33 there is nothing which shows either that SIS ever considered this threat seriously or took steps to deal with it. Why they did not, and how much of a difference some preventive action might have made to the safety and survival of SIS stations in Norway, are now likely to continue to remain unresolved mysteries.

  Notes

  1 TNA, HS 7/174.

  2 For an entertaining account of these manoeuvrings, see Jebb’s biography by Sean Greenwood, Titan at the Foreign Office (Leiden: Nijhoff, 2008), pp. 85–101.

  3 Mackenzie, Secret History of SOE, p. 95.

  4 Ibid.

  5 TNA, paper by Godfrey, 3 April 1941, ADM 223/851.

  6 Jeffery, MI6, p. 374.

  7 TNA, Menzies to Nelson, 5 February 1942, HS 8/321.

  8 Ibid., pp. 354–356.

  9 TNA, HS 2/150.

  10 TNA, Scott minute, 15 May 1942, FO 1093/155.

  11 TNA, HS 8/321.

  12 NHM, FO II.8.2 F3.

  13 NHM, SIS progress report late March 1943.

  14 HS 9/892/7.

  15 NHM, Wilson history, p. 102.

  16 TNA, correspondence between Cordeaux and Campbell, DDNI, September 1944, ADM 223/481.

  17 NHM, SIS progress report. October 1944.

  18 Kristian Ottosen, Theta, Theta (Oslo: Universitetsforlag 1983), p. 56, quoted by Ian Herrington, The SIS and SOE in Norway 1940–1945: Conflict or Cooperation? War in History (2002), Vol. 9, No. 1, p. 101.

  19 TNA, HS 7/182.

  20 TNA, HS 2/184.

  21 TNA, HS 2/161.

  22 NHM, FO.II 8.2 E5.4.

  23 RA, Nagell box 3.

  24 TNA, ADM 223/884.

  25 NHM, report on German direction finding and interception system against illicit transmitters, August 1945, FO.IV box 21.

  26 Ragnar Ulstein, Etteretningstjenesten i Norge (Oslo: NHM, 1994), p. 24.

  27 NHM, Roscher Lund report, Hovedtrekk i etteretningstjenestens utvikling, p. 10, NHM 283.

  28 TNA, report from Haugland, 22 September 1943, HS 2/173.

  29 TNA, HS 7/175, appendix M: W/T communications.

  30 TNA, report by Wilson to the SOE security section, 12 April 1943, HS 2/151.

  31 NHM, SOE box 25.

  32 NHM, FO.IV box 40, report on German direction finding and preventive measures.

  33 See for example Nøkleby p. 147, giving details of comments from Einar Johansen (Upsilon) and Bjørn Rørholt (Lark).

  * Two SIS officers were abducted by the Germans on the German-Dutch border at Venlo on 9 November 1939, when participating in what they thought was a negotiation with German officers planning a coup against Hitler. The Germans exploited this for propaganda purposes.

  † Larsen and the crew were rescued two weeks later by a boat which was sent from Shetland.

  ‡ See Chapter 7 for further discussion of these actions.

  § A more detailed account of this operation is given in Chapter 7.

  ¶ Bakka was harshly interrogated and beaten after his capture. Rørholt writes that in the last days of the war the German responsible, Sturmscharführer Hölscher, promised Bakka that he would shoot him before killing himself. On 7 May, when very drunk, he started to go to the jail to carry out this undertaking, but became distracted and forgot, before he committed suicide. So Bakka survived. (Rørholt, Usynlige soldater, p. 321.) Bakka was also awarded a DSC after the war.

  || Though Berit Nøkleby points out that the head of the Gestapo in Norway, Siegfried Fehmer, stated that all the SIS and SOE stations in the Oslo region captured by the Germans after 1943 were taken as a result of direction finding. (Berit Nøkleby, Pass godt på Tirpitz! (Oslo: Gyldendal, 1988), p. 145.)

  CHAPTER 4

  CRACKING ABWEHR CODES

  HOW BLETCHLEY PARK MADE THE BREAKTHROUGH

  Gentlemen don’t read each other’s mail

  Writing in his memoirs,1 Henry Stimson, who had been appointed Secretary of State by Herbert Hoover in 1929, used this aphorism to explain his decision, shortly after he took office, to withdraw funding from the Black Chamber,* America’s leading cryptographic agency.2 The Black Chamber closed soon afterwards, and it took years for American cryptographic work to recover from this setback. European cryptographers were not constrained by any similar concerns about ungentlemanly behaviour and, while limited by their budgets, were able to develop their activities during the interwar period, reading a wide variety of cyphers, without necessarily discriminating too much about whose cyphers they broke.

  The extent of codebreaking in the interwar period

  In Britain, the Government Code and Cypher School (GC&CS) was formed in November 1919 out of the remnants of the wartime Admiralty and War Office cryptographic branches. After some bureaucratic wrangling about where it was to be housed (the Foreign Office staked a claim), it was agreed that it should be based in the Admiralty, but the Foreign Office soon assumed administrative responsibility for it because of the extent of the diplomatic decrypts which it was producing. This division of responsibility proved unsatisfactory and in 1923 it was agreed that GC&CS should be placed under the authority of Sir Hugh Sinclair, the new head of SIS, an arrangement which lasted for more than twenty years. GC&CS was responsible for cryptography, while SIS circulated the intelligence which was thereby produced. Financial stringency limited the activities of GC&CS as much as it restricted those of SIS in the interwar years, but growing awareness of the threat from Germany enabled this obstacle to be reduced from early 1938 onwards as budgets expanded. By the beginning of the war, the range of diplomatic telegrams which GC&CS was able to read included American, Japanese, French, Italian, Ethiopian, Spanish, Saudi, Turkish – and Norwegian. GC&CS began to move to its wartime location at Bletchley Park in August 1939, and at the same time Alastair Denniston, its head, started the process of recruiting the remarkable range of very talented academics who were to play key roles in the successes of GC&CS in the next few years.

  But it was not only GC&CS which was successful in cryptography during this period. A conversation in November 1941 between Roscher Lund, who possessed some knowledge about Swedish cryptographic work and who had just returned from Stockholm, and David Rees of GC&CS, provides a good illustration of the extent to which it had developed in Scandinavia. Roscher Lund told Rees that the Swedes had enjoyed much success in their attacks on French cyphers (reading about 95 per cent of them), Italian military cyphers (85 per cent) and had made fair progress on German cyphers. The Germans were sending many messages by teleprinter to their troops in northern Norway, and the telephone manufacturing firm of L. M. Ericsson had developed a machine which could decypher them. He said nothing about any Swedish success against British cyphers, but added that the Finns had broken Russian military cyphers and were reading about
95 per cent of their traffic.3

  GC&CS seems to have had fairly extensive coverage of Norwegian diplomatic and commercial traffic during the interwar period. Most of the diplomatic material relates to Norwegian relations with other countries, and there was little on bilateral relations with Britain. GC&CS was already aware that Roscher Lund had started a cypher bureau in Oslo in 1936, knew plenty about Swedish cryptography and was concerned about Norwegian cypher security. He explained to Rees that he was most anxious to have the opinion of British experts on the security of the Hagelin cypher machine, as he had heard rumours from British sources that it was unsafe and the traffic it generated was being read. When he asked Rees if he knew anything about this, Rees was less than honest in his reply, stating that it was news to him and that in his twenty years of cryptographic work, to his knowledge no work had been done on Scandinavian cyphers. This was far from the truth. In further discussion, Roscher Lund told Rees that he had been interested in cryptography for some years but that he had found it hard to get the Norwegian authorities interested in their security. He quoted an example of a code compiled in 1922 which was still being used in 1938. In his report, and revealing the extent of his knowledge, Rees commented on this statement: ‘That is correct, but the code was used only for outlying posts’!4

  Steps were subsequently taken to provide advice to the Norwegians on the security of their cypher machines.5 GC&CS continued to collect Norwegian traffic, though without registering or analysing it, and later circulated material from the Quisling government. No further attempts were made to read traffic from their ally, though a minute to Denniston in January 1942 noted that a Norwegian commercial cypher was available and asked whether it should be provided. There is nothing to show that this offer was taken up.6 GC&CS also circulated decyphered traffic from wireless stations operated by members of the Norwegian communist resistance in the north of Norway. Many of the messages were terse, and related to the need for supplies by both submarine and aircraft, with occasional requests for reports on the outcome of operations. There were also messages marking the celebration of significant revolutionary anniversaries, but in general they were quite didactic and impersonal. ‘Nr 92 received. Your decision is right. Be alive and energetic’. ‘Investigate the results of our air attack on a German convoy on 21 April in Kongsfjord’. ‘Greetings on the occasion of the 1 May to the unconquerable and persevering reconnaissance troops and tanks for their work. We wish them good health and excellent results in their operations’. They were circulated quite widely within SIS, but it is not clear what other use was made of them.7