This is an Accepted Manuscript of an article published by Taylor & Francis in Intelligence and National Security

The BRUSA/UKUSA procedures created an interlocking system for controlling and referring to Comint and the combined markings and different terms can be seen on issues of the CIA’s Current Intelligence Weekly Review. The review was published by the CIA’s Office of Current Intelligence every Friday and issued to American officials in Comint or non-Comint versions.⁶⁵ A Comint version of the weekly review of 30 August 1956 had the security classification and the Category III Comint code word, ‘TOP SECRET EIDER’, printed on the top of its interior pages.⁶⁶ Its cover page was emblazoned with the customary phrase ‘THIS DOCUMENT CONTAINS CODE WORD MATERIAL’, followed by the stern instructions that the document

is to be seen only by US PERSONNEL especially indoctrinated and authorised to receive Special Intelligence information. The security of this document must be maintained in accordance with SPECIAL INTELLIGENCE SECURITY REGULATIONS. No action may be taken by any person on the Special Intelligence presented herein, regardless of any advantage which may be gained, unless such action is first approved by the Director of Central Intelligence.

Editions of the Weekly Review from the late 1950s and 1960s had a similar format but substituted the phrase ‘communications intelligence’ for ‘special intelligence’ in the front page instructions.⁶⁷

The influence of the BRUSA/UKUSA security and dissemination system spread beyond these formal markings on documents. Indoctrinated American and British policymakers sometimes used the different control terms as euphemisms for Comint in their internal correspondence. Rather than write ‘Comint’ or ‘Sigint’ they wrote ‘special intelligence’, ‘code word material’ or the current high level Comint code word. Examples of this usage can be seen in British documents from the Suez War of 1956. In a minute to the JIC in November 1956 a British official explained how intelligence would be coordinated during the conflict with Egypt. He noted that the staff would ‘check for the exclusion of codeword material from the daily report prepared...for circulation to the Cabinet.’⁶⁸ In another minute, an official described the early morning daily intelligence briefings in the War Room in Whitehall as being only for ‘those officers who are cleared to receive Eider information’.⁶⁹ Despite the introduction of the term ‘Crypt Intelligence’ and the triple category system, the phrase ‘special intelligence’ seems to have been the euphemism most commonly used in correspondence. This practice may even have been an approved part of UKUSA security procedure, for the NRO review and redaction guide defined ‘Special Intelligence’ as ‘the unclassified term which is used to identify COMINT in the unclassified environment.’⁷⁰

Understanding the BRUSA/UKUSA security system and identifying the various code words and euphemisms is important because they can help pierce some of the secrecy surrounding Comint. Although the Special Intelligence and Category III code words and Comint euphemisms were originally intended to protect and conceal intelligence, once their meanings are known they can signal the presence of Comint in a document and reveal information about it. They can indicate NSA and GCHQ Comint targets, show what intelligence the agencies were providing American and British policymakers and give clues to code breaking successes. Yet there are limits to using the code words in this way. In the past American government reviewers often redacted the high level Comint code words on documents prior to their public release; on CIA intelligence reports it is common to see the classification ‘top secret’ with the following word deleted. Indeed, reviewers sometimes overlooked a code word and inadvertently revealed what they were trying to hide. On a 106 page long CIA report about Soviet defence spending from 1955, a reviewer has carefully deleted the word following ‘TOP SECRET’ on top and bottom of every page, except for the top of page 23, where the tell-tale phrase ‘TOP SECRET EIDER’ can be seen.⁷¹

It should also be borne in mind that the Special Intelligence and Category III code words were not exclusively for intelligence derived from code breaking. The guidance in BRUSA Appendix B suggests that the high level code words were sometimes used for Comint based on traffic analysis and perhaps plain text and speech intercepts. They could also protect documents which contained information on Comint collection methods or the handling of Comint instead of actual Comint. There is an element of uncertainly as well, because parts of Appendix B relating to Special Intelligence and Category III Comint are redacted and procedures may have been amended after 1956. Nonetheless, if a specific intelligence report was marked with a Special Intelligence or Category III code word it does indicate that some of the information in that report came from Comint. Furthermore, it will have come from Comint requiring the highest degree of security protection. On most occasions this would be decrypted or deciphered communications rather than traffic analysis or plain text.

Despite the best efforts of the reviewers, high level code words and Comint euphemisms appear on declassified American intelligence reports relating to many different countries during the Cold War: Bolivia, China, the Congo, France, Greece-Turkey, India-Pakistan, Indonesia, Laos, North Vietnam, South Africa, South Vietnam and the Soviet Union.⁷² This article will take the examples of the Congo, Bolivia, Indonesia, South Vietnam and China to demonstrate how code words and euphemisms can be used with other evidence to pick out NSA and GCHQ intelligence targets and uncover Comint. For simplicity’s sake, the article will treat the NSA as being the producer of Comint in American intelligence reports but because of the BRUSA/UKUSA cooperation arrangements, GCHQ may actually have been the source for some countries in Africa and Asia. For example, Comint on China in a CIA memorandum might have originated in a GCHQ intercept from Britain’s listening post in Hong Kong. Equally, NSA intercepts, solutions to foreign codes and decrypted Comint would have been shared with GCHQ.

Firstly, Category III code words and Comint euphemisms on documents can identify little known targets of the NSA, such as the Congo in the 1960s. After independence in 1960, the Congo went through a prolonged and complex crisis, with multiple secessions, civil war, military coups and intervention by Belgian and United Nations’ troops. Amidst the chaos and instability, the United States was determined to deny the USSR a foothold in the Congo and it opposed Soviet-learning Congolese leaders, such as Patrice Lumumba, the prime minister, and Antoine Gizenga, Lumumba’s deputy, who later established a rebel regime based in the city of Stanleyville.

It is clear from code word documents that the NSA supplied Comint on the Congo to American policymakers. On 11 August 1960 the Congo featured in a ‘Synopsis of Intelligence Material Reported to the President’ written by John Eisenhower, Assistant Staff Secretary to President Dwight Eisenhower.⁷³ The short synopsis carried the classification ‘TOP SECRET DAUNT’, marking it out as containing Category III Comint, but frustratingly the text on the Congo has been completely redacted. The following month, the CIA sent the Director for Intelligence for the Joint Chiefs of Staff an analysis of Soviet Bloc aid to the Congo that was also stamped with the words ‘TOP SECRET DAUNT’.⁷⁴ The paper, which is partially redacted, gave details of Bloc supply flights into the Congo and the numbers of Bloc personnel in the country. Other evidence points to the NSA being able to read the communications of Gizenga’s rebel government in Stanleyville. Gizenga was in contact with the President of Ghana, Kwame Nkrumah, who had recognised the Stanleyville regime and secretly agreed to funnel it Soviet supplied arms.⁷⁵ The Comint edition of the CIA’s Current Intelligence Weekly Review reported on 25 May 1961 that

Although there is no firm information that Gizenga plans any major offensive action, intercepted messages from Gizenga to Nkrumah have urged that the Ghanaian President expedite the airlifting of Soviet arms to Stanleyville.⁷⁶

A veiled reference to Comint implies that Bolivia was another target of the NSA in the 1960s. Bolivia was governed by the pro-American President Rene Barrientos and during 1966-67 he successfully defeated a Marxist guerrilla insurgency led by Ernesto ‘Che’ Guevara. Yet in 1968, NSA Comint on Bolivia was connected to more mundane matters of taxation and American financial aid. In July Barrientos was due to visit the United States for talks, including a meeting with President Johnson at his ranch in Texas. Ahead of the visit Barrientos delayed a surtax on imports into Bolivia which was an agreed condition for an American loan. Walt Rostow, the Special Advisor for National Security Affairs, advised Johnson that

From our special intelligence, we have the distinct impression the delay is related to Barrientos’ visit to the Ranch. We suspect his advisors have told him that by talking to you, he can probably get the budget support money without imposing the surtax.⁷⁷
These documents do not conclusively prove that the NSA could read encrypted Bolivian and Congolese communications - the Comint on Bolivia and the Congo could have come from intercepted plain text or voice - but it does seem likely. In the 1950s and 1960s non-communist developing countries like Bolivia and the Congo frequently used codes and ciphers which were vulnerable to attack by the NSA and GCHQ.⁷⁸

Indonesia and South Vietnam are already known to have been targeted by the NSA during the Cold War and there is evidence that the agency broke their codes and read their communications. In these two cases, code word documents substantiate the existing evidence and illustrate what kind of information the NSA was passing on to the White House. Indonesia was a country of interest to the United States because of American hostility to the radical, left wing Indonesian leader, Sukarno. In 1958 the CIA covertly supported rebellions against Sukarno on the Indonesian islands of Sumatra and Celebes.⁷⁹ The United States soon had to stop aiding the rebels but code word documents suggest that in 1959 the NSA was monitoring the Indonesian authorities’ response to the rebellions. According to two ‘Synopsis of Intelligence material reported to the President’, President Eisenhower was informed in March 1959 that rebel action had forced the Indonesian government to send air and ground reinforcements from Java to North Celebes.⁸⁰ He was also told that an Indonesian army mission trying to procure arms abroad was authorised to investigate Czech and Polish offers to build an arms factory in Indonesia. These two very brief synopses on Indonesia were marked ‘TOP SECRET EIDER’, implying that the information for the president had come from Comint. This Comint was probably obtained through reading encrypted Indonesian traffic for in September 1960, two defectors from the NSA announced at a press conference in Moscow that the agency collected intelligence and decoded messages from Indonesia, along with six other countries.⁸¹

Claims that the NSA was spying on South Vietnamese communications first surfaced during the Vietnam War.⁸² In 1971, the well informed Washington Post columnist Jack Anderson alleged that the agency was intercepting and decoding South Vietnamese messages.⁸³ This operation had in fact been going on for several years and in 1968 it had become entangled with American domestic politics.⁸⁴ In the closing stages of the American presidential election campaign NSA intercepts showed that Richard Nixon’s camp was secretly encouraging the South Vietnamese government not to commit to peace talks in Paris.⁸⁵ Lyndon Johnson confided to a Democrat senator in November 1968 that ‘intercepted telegrams revealed that the South Vietnamese were being told “that Nixon is going to win: therefore they ought to wait on Nixon.”’⁸⁶ Defense Secretary Clark Clifford, complaining about the Republican Party’s efforts, referred to ‘the intercepts of messages from Bui Diem [the South Vietnamese Ambassador in Washington] to Saigon’.⁸⁷

Checking for Category III code words on documents uncovers direct evidence of this NSA eavesdropping on South Vietnam. On 21 October 1968 the Director of the NSA sent the White House a telegram which contained a report about a meeting in Paris on 19 October between American and South Vietnamese officials.⁸⁸ The content of the report was unremarkable: the Americans and South Vietnamese negotiators had disagreed about the involvement of the communist National Liberation Front at the Paris peace talks. What is interesting is that the NSA telegram was marked ‘TOP SECRET TRINE’, indicating that the report was high level Comint. The report has been partially redacted to hide the author’s identity but the phrasing of the text and the point of view of the author suggests that it was written by a South Vietnamese rather than an American participant at the meeting. Taking the Comint code word into account, it appears to be an intercepted South Vietnamese message (presumably this was before these intercepts were given the additional Gamma Gout code words).

Finally, Category III code words and Comint euphemisms can offer insights into Anglo-American Comint on China, demonstrating the sort of intelligence the NSA and GCHQ could produce on this primary Cold War target. Close reading of an American document suggests that during the 1960s the NSA was able to read some Chinese military traffic. A CIA intelligence memorandum from 1965 on Taiwan’s capability to attack the mainland included a map of the communist order of battle in southern China.⁸⁹ The map gave the locations of Chinese army divisions, border defence divisions and air and naval bases and provided figures for the local and national strength of the Chinese army, navy and air force. It was marked ‘TOP SECRET DINAR’, indicating that some of the information on the map had come from Comint (by contrast, an accompanying map of the Taiwanese order of battle was only classified ‘secret’ and had no Comint code word). This implies that the CIA was using intercepted Chinese military communications, along with other forms of intelligence, to work out the location, organisation and strength of the Chinese armed forces. Some of the Comint might have come from plain text and speech transmissions and traffic analysis but it is possible that the NSA could solve at least low level Chinese military codes

Comint markers and other evidence show that the NSA could use this intelligence to track the movements of the Chinese military. For example, in June 1962 Beijing secretly sent six infantry divisions and one artillery division into Fukien province, raising the battle strength opposite Taiwan from 194,000 to 311,000 men.⁹⁰ Although the Chinese government initially made this move without any publicity, Washington was aware of it, partly through Comint. From 1962 until 1966, the noted Asia scholar Allen Whiting was head of Far East division of the State Department’s Bureau of Intelligence and Research (INR) and he was responsible for collating and assessing information on the Chinese build up.⁹¹ Whiting had access to intelligence from the NSA and he subsequently wrote that American ‘signal intelligence [was] monitoring [Chinese] headquarters deployment in advance of the total force.’⁹² Roger Hilsman, the Director of the INR, assessed the intentions behind the deployment by examining ‘the limited collateral evidence and some aspects of the special intelligence on Chinese Communist activities in the East China area’.⁹³ Another map in the CIA’s 1965 memorandum plotted the build up of Chinese forces in Fukien in 1962 and was also classified ‘TOP SECRET DINAR’, betraying the presence of Category III Comint.

In a similar way, the NSA eavesdropped on China’s military response to the escalating war in Vietnam. Whiting described how after the Tonkin Gulf crisis in August 1964, the US followed Chinese deployments through the use of ‘signals intelligence, drone reconnaissance planes and Chinese Nationalist piloted U-2s’.⁹⁴ American ‘communication intercepts monitored the movement of [Chinese] air units from one point to another.’⁹⁵ Code word material supports Whiting’s claims. In February 1965 the CIA produced a special report on China’s post Tonkin Gulf military deployments.⁹⁶ The report has been partially redacted but the phrase ‘TOP SECRET DINAR’ is still visible on one section entitled ‘Military Preparations in Communist China’. This section stated that since August 1964, Beijing had transferred the headquarters of its South Sea Fleet to Fort Bayard (now Zhanjiang), moved the 7th Air Army nearer the Vietnamese border area and established Sino-Vietnamese air warning liaison facilities. The Chinese air force had boosted its strength in southern China, building new airfields, improving existing ones and bringing in an extra 200 fighter planes as reinforcements. The CIA paper also reported a secret meeting between Chinese and North Vietnamese military leaders in June 1964. Books based on Chinese sources have since confirmed that this meeting and many of the military deployments did in fact take place.⁹⁷

These are only a few examples but they illustrate how Comint code words and euphemisms can help interpret documents and extract more information about Anglo-American Comint. During the Cold War the BRUSA/UKUSA states developed a comprehensive system to secure and manage Comint. The security categories and code words were an integral part of the system, with the code words always meant to mark out the documents that contained Comint. Security officers could then ensure that these papers were kept under lock and key and only seen by staff indoctrinated into Comint. Now that the code words have been declassified and identified, they can expose Comint in American and British (and Canadian, Australian and New Zealand) documents. Euphemisms were also intended to protect Comint by disguising the source of information in minutes and letters. Yet euphemisms are very weak form of security because once the meaning of the euphemism is known, the information immediately becomes apparent to the reader. This works to the advantage of researchers, for government reviewers seem to let the euphemisms ‘special intelligence’ or ‘codeword material’ pass unaltered on documents while assiduously redacting mentions of ‘Comint’ or ‘Sigint’.

The net result is that the code words and euphemisms act as Comint markers and enable researchers to discover more about Anglo-American Comint, even when documents have been heavily redacted before public release. These markers can flag up papers containing Comint and thereby reveal little known targets of the NSA and GCHQ, like the Congo and Bolivia in the 1960s. They can reveal what type of intelligence the NSA and GCHQ was able to supply on countries such as Indonesia, South Vietnam and China during the Cold War. Comint markers on documents also hint at cryptanalytic successes. For example, they suggest that the NSA and GCHQ could read Congolese and Bolivian codes and cyphers and perhaps some Chinese military codes. In short, they lift some of the veil of secrecy over Anglo-American Cold War Comint and provide a useful research tool which can be used more widely in the future.

1^ Matthew Aid, The Secret Sentry: the untold history of the National Security Agency, (New York: Bloomsbury Press 2009); Richard J. Aldrich, GCHQ: The Uncensored Story of Britain’s Most Secret Intelligence Agency, (London: Harper Press 2010).

2^ Richard J. Aldrich, The Hidden Hand: Britain, America and Cold War Secret Intelligence (London: John Murray 2001), pp. 541.

3^ British National Archive (TNA), HW 80/4, British-US Communication Intelligence Agreement, 5 March 1946; HW 80/6, Technical Conference for the implementation of the US-British Communications Intelligence Agreement, 11th-27th March, 1946, Introduction, not dated.

4^ TNA, HW 80/6, Technical Conference for the implementation of the US-British Communications Intelligence Agreement, 11th-27th March, 1946, Appendix B, not dated; HW 80/7, Technical Conference for the implementation of the US-British Comint Agreement, 15th July – 26th July, 1948, Appendix B, not dated.

5^ TNA, HW 80/9, Introduction to the Appendices to the US-British Comint Agreement, Appendix B, Annexure B1, 1 June 1951.

6^ TNA, HW 80/6, Technical Conference for the implementation of the US-British Communications Intelligence Agreement, 11th-27th March, 1946, Appendix B, not dated

7^ Ibid; National Security Agency (NSA), Joint Meeting of State-Army-Navy Communication Intelligence Board and State-Army-Navy Communication Intelligence Coordinating Committee, 15 February 1946. http://www.nsa.gov/public_info/_files/communication_intel_board/army_navy_19460215_mtg.pdf, (Accessed 11 January 2011).

8^ TNA, HW 80/6, Technical Conference for the implementation of the US-British Communications Intelligence Agreement, 11th-27 March, 1946, Appendix B, not dated.

9^ TNA, HW 80/7, Technical Conference for the implementation of the US-British Comint Agreement, 15th July – 26th July, 1948, Appendix B, not dated.

10^ TNA, HW 80/6, Technical Conference for the implementation of the US-British Communications Intelligence Agreement, 11th-27 March, 1946, Appendix B, not dated.

11^ TNA, HW 80/10, BRUSA Planning Conference 1953, Final Report, Appendix B, BPC53/B/Final, 19 March 1953.

12^ The wording of Appendix B suggests the categories may already have been in use when the Appendix was written in March 1953.

13^ TNA, HW 80/10, BRUSA Planning Conference 1953, Final Report, Notes to Appendix B, BPC53/B/Final, 19 March 1953.

14^ TNA, HW 80/11 UK-US Communications Intelligence Agreement, Appendix B, Not dated.

15^ Jeffrey Richelson, The US Intelligence Community, (Boulder: Westview Press 2008) pp. 512-513; National Security Agency (NSA), Cryptologic Almanac, 50th Anniversary Series, ‘Quiz’, July/August 2002. http://www.nsa.gov/public_info/_files/crypto_almanac_50th/Quiz.pdf, (Accessed 3 February 2011).

16^ TNA, HW 80/10, BRUSA Planning Conference 1953, Final Report, Notes to Appendix B, BPC53/B/Final, 19 March 1953.

17^ Richelson, op. cit., pp. 511.

18^ Foreign Relations of the United States (FRUS), 1964-68, Volume XXXIII, Organization and Management of Foreign Policy: United Nations, Document 270, 16th Report of the Intelligence Organization of the Department of State, October 1967, pp, 592; David Wise, The Politics of Lying: Government Deception, Secrecy and Power, (New York: Random House 1973) pp. 55-56, 61.

19^ Richard Aldrich, The Hidden Hand: Britain, America and Cold War Secret Intelligence, (London: John Murray 2001) pp. 541.