Chapter 5
1986/7 ZIRCON

Throughout the latter part of 1986 and the first weeks of 1987, the investigative journalist Duncan Campbell was working on a BBC television series to be called Secret Society. Campbell had been a thorn in the side of the intelligence establishment for years. His articles, usually in the New Statesman magazine, uncovered matters ranging from which office blocks in London were used by the agencies to allegations of dirty tricks in Northern Ireland. An earlier attempt to convict Campbell and one of his sources under the Official Secrets Act had failed; this was one of several cases which had prompted the Home Office to draw up a new secrets law.

Campbell discovered that the government planned to build a new satellite which would enable GCHQ to eavesdrop on the Soviet Union. He believed the plan violated a 1982 government agreement to inform the House of Commons Public Accounts Committee of any military project costing more than a certain amount. During the course of his filming for the Secret Society programme, Campbell questioned former Ministry of Defence officials and Robert Sheldon MP, chairman of the Public Accounts Committee. Inevitably, word of his inquiries reached government, and the Prime Minister decided to act. Nigel Lawson, then Chancellor of the Exchequer, recounts in his memoirs that 'the government managed to lean on the BBC to ban the programme'. Faced with the prospect of an injunction, Campbell rushed his research into print in the New Statesman of 23 January 1987. The spy satellite, hitherto one of Whitehall's most cherished secrets, had become a major story, as had its code-name: ZIRCON.

The Falklands war of 1982 was regarded by most people in the defence and intelligence establishments as a textbook example of Britain's 'special relationship' with the USA in action. The Americans had made certain advanced weapons available to the British and had shared vital intelligence about the location of Argentine ground and naval forces.

In Cheltenham, though, there were people who knew that this assistance had sometimes required special pleading. The National Security Agency, GCHQ's US counterpart, had not achieved global coverage with its sigint satellites by 1982. The craft which was in a position to help Britain monitor Argentine communications was being used by the Reagan administration to eavesdrop on central America, principally El Salvador. One of the GCHQ officers who liaised with NSA recalls, 'We had to negotiate very hard to get it moved, and then only for limited periods.' During these spells of a few hours each, the satellite's listening dish was reorientated towards the south Atlantic in order to help Cheltenham. The NSA did not monitor the downlinked take during these periods, asking GCHQ to alert them if there was anything of US interest in the transmissions.

Although GCHQ was grateful for NSA's help, and learnt a good deal from the occasional use of its satellite, the senior officers in Cheltenham, notably the then Director, Brian Tovey, drew certain conclusions. As one former GCHQ officer says, 'We can ask the Americans to do things, but we cannot compel them. There may be targets they don't want to cover. The Falklands was a factor here. It brought going it alone back into fashion.' It was already apparent to GCHQ management that space represented the future of sigint and that gaining a British foothold in such technology might be possible, given the Prime Minister's largesse towards the agencies.

In 1968 the National Security Agency had launched the first of seven satellites code-named CANYON. This programme had alerted the US agencies to the extraordinary possibilities of communications intelligence-gathering by satellite. CANYON was able to pick off various types of voice and data traffic from space. It was followed by a type of satellite initially code-named RHYOLITE and later AQUACADE. RHYOLITE marked a breakthrough in the sigint world. It could pick up various types of transmission, but the most important take came from the microwave telecommunications links which by the 1970s had been installed across the Soviet Union. Microwave circuits had been considered highly secure by the Russians because they use a narrow beam of energy between a transmitter and receiver which have to be within line of sight. Trying to pick up the microwave transmission from even a few miles away is pointless. But the parts of the microwave beam which shoot past the receiver - 'spillage' in sigint jargon - continue in a straight line up into space. Microwave beams may also strike the ground in places, bouncing signals straight upwards. RHYOLITE was placed in a geosynchronous orbit - positioned 24,000 miles from earth with its speed exactly matching the turning of the globe - and so was able to 'hover' over the Soviet Union. It was equipped with a large parabolic dish so that the feeble fragments of microwave energy could be refocused on its receiver. Each microwave circuit could carry hundreds of conversations. The possibilities of RHYOLITE were, says one sigint insider, 'mind-blowing'.

Under the Anglo-Saxon sigint arrangements, GCHQ was a full party to the product of this satellite. Owen Lewis, then an Army officer working in sigint and now a communications security consultant in industry, remembers, 'When RHYOLITE came in, the take was so enormous that there was no way of handling it. Years of development and billions of dollars then went into developing systems capable of handling it.' NSA's response to the explosion of information coming from space included passing large amounts of it over to GCHQ for transcription and analysis. The USA developed two types of geostationary sigint payloads: one descending from RHYOLITE was used mainly to gather interesting UHF signals such as missile telemetry and various forms of communications; the other specifically targeted microwave traffic. Each type required three satellites continuously in geosynchronous orbit over the Equator to provide global coverage. The NSA found that the amount of information being picked up from microwave circuits was so large that it had to be immediately beamed down to an earth station within line of sight. For two of the three satellites, this required ground stations outside the USA.

From 1970 the NSA had built a new constellation of eavesdroppers initially code-named CHALET and later (after a US newspaper had published this code word) VORTEX. Two large ground stations were built to downlink the traffic from the two payloads that could not be run from the USA. At Menwith Hill in north Yorkshire, hundreds of NSA staff were involved in an expanded operation believed to be taking down the product of the CHALET over the Soviet Bloc. At Pine Gap in Australia they downlinked the product of another 'bird', probably the one covering China, south-east Asia and parts of the Pacific rim. Even by the mid-1990s, with a third generation of comint satellites in orbit in these slots, the NSA has still not found another way to relay this information; imaging craft, on the other hand, can send pictures back to the USA via a network of relay satellites. Menwith and Pine Gap had thus revitalized the UK-USA sigint alliance (to which Canada, Australia and New Zealand are also partners).

In January 1985 the space shuttle Discovery is believed to have deployed the first of a new, improved class of satellite code-named MAGNUM. This enormous satellite, thought to weigh 2.7 tonnes, was reportedly a descendant of the RHYOLITE series, designed to intercept Soviet missile test signals (telemetry) and data-links as well as microwaves. The cost ($300 million each even in 1985) and complexity of the project is such that few have been launched.

Despite Menwith's usefulness to the NSA, GCHQ officers felt that the quantity of material coming from these new satellites was tilting the sigint relationship so far in the USA's favour that there was a danger of the British contribution becoming insignificant. The Director and senior managers therefore frequently used the need to ensure continued access to US sigint as an argument in seeking funding for their own projects. In the 1960s it had been used to get money for GCHQ's listening post at Bude in Cornwall, placed to allow the interception of transatlantic telephone calls - which was of great interest to the NSA.

One civil servant who sat on the budgetary sessions of the Permanent Secretaries' committee on the Intelligence Services (PSIS) recalls, in connection with this GCHQ tactic, that 'the American card was played quite often. There was always this awkward question of how far can you pare away the commitment.' For the mandarins, it was often difficult to know how far the representations of a GCHQ Director in support of major new projects could be taken at face value. The one-time PSIS member reveals, 'One didn't feel very confident with technical matters, but that's always the case. It was like nuclear weapons, there were no neutral specialists. That is true of intelligence. The best we could do was prod and see how loud they squealed.'

In its daily tasking, GCHQ did the maximum to help the USA. As one officer remembers, 'The requirements from our friends across the water often had to be met first under the special relationship. They were quite clearly the Big Brother. If a suitably-worded request came in and the only way to meet it was to divert resources from a low-priority UK target to a "flash" US one, then it was pretty obvious what would happen. The special relationship was regarded with the highest possible esteem.'

Britain also tried to pass on any technological advances to the USA. During 1986-7 GCHQ installed in some of its signal interception systems a new software package which provided superior recognition of key words. It had been developed by a private company, PA Technology, at its centre near Cambridge, under a substantial GCHQ contract. The software marked a significant step forward in processing sigint, being based on a phonetic system of sound recognition superior to anything previously available. It is believed to have been offered to the NSA, which said it was impressed but had substantial projects of its own to develop similar packages.

By the mid-1980s the inequalities in the GCHQ-NSA relationship led some to believe that the USA was less than committed to it. Martin Morland, who left the Chief of Assessments post at the Joint Intelligence Committee in 1986, says, 'Everything is meant to be completely shared, but even then the Americans were gradually holding back a bit. It didn't happen on the central area of Soviet Bloc, but more where they had particular interests, like Cuba, or where commercial matters were concerned.'

One former GCHQ officer suggests that the ratio of US sigint intercepts of Soviet Bloc traffic to GCHQ ones was running at about five to one during the mid-1980s. The Director of the NSA between 1985 and 1989 was a tough-talking lieutenant-general, William Odom. Despite professing himself an Anglophile descended from seventeenth-century English founding fathers, Odom is brutally frank about GCHQ and the nature of sigint ties. He told me, 'It's a very uneven relationship, to put it mildly . . . the name of the British game is to show up with one card and expect to call all the shots.'

Lieutenant-General Odom suggests that the claims frequently heard in Whitehall (and often repeated by British interviewees during the writing of this book) that GCHQ remains a world-class player were by the 1980s self-delusion. The former NSA director notes, 'Technology has changed so much that what the British brought in World War Two, they do not bring any more. They had a great tradition for a kind of eccentric cryptanalysis. Well, today that and seventy-nine cents will only buy you a cup of coffee at Seven Eleven. Today, this business requires huge investment and Britain doesn't have that.'

ZIRCON had been conceived by Brian Tovey to keep the special relationship sweet and to take his organization into space, the next logical area of sigint development. Tovey, says a former GCHQ officer, 'was gung-ho, a real expansionist'. The Director's background, the sensitivity about the relationship with NSA, and a perception of the gains to be made from a sigint spacecraft all seem to have ensured that, from an early stage, Tovey's plan was for an eavesdropping satellite rather than one able to send back pictures.

The British satellite was conceived as a geosynchronous one that would sit over the Soviet Union. Tovey's vision of ZIRCON was to survive from 1983, when the Cabinet Office approved initial studies, to the autumn of 1986. A member of the Defence Intelligence Staff at the time says, 'It was held at an incredibly tight level. We knew that there was something called ZIRCON and we knew it would be incredibly expensive.' Just how expensive would be the subject of heated argument following the publication of Campbell's article, but given the costs of the US systems, it is unlikely that ZIRCON and its associated ground station would have cost Britain less than 500 million. Much of this enormous investment would have to be repeated after five years or so, the expected life of the satellite.

There was some debate within Whitehall as to how this expenditure could be justified, given that by this time the NSA had near-global coverage with its constellations of two different types of sigint satellite. ZIRCON's backers at GCHQ argued that it would allow the USA not to fill a certain slot, or give the British complete independence if they required it. However, the managers from Cheltenham conceded that three geosynchronous satellites would be needed for a complete stand-alone UK constellation - and nobody thought Whitehall would pay for that.

The publication of Duncan Campbell's article in January 1987 brought the hitherto 'black' project into the glare of publicity. Following publication, says Nigel Lawson in his book, 'Margaret instructed Michael Havers, the Attorney General, to issue an injunction against [Campbell] . . . in a somewhat unfortunate blaze of publicity, the police raided offices both of the BBC in Glasgow and of the New Statesman in London.' The sight of Special Branch detectives carrying off videotapes and papers antagonized liberal opinion, and was evidence of how deeply entrenched the Prime Minister had become in her desire to protect the intelligence services from journalistic scrutiny, even if the political cost was high.

Nigel Lawson says in his memoirs, 'Well before all this blew up, I had succeeded in getting the ZIRCON project cancelled on grounds of cost.' He says the satellite 'did not in any sense leave the ground'.

In August of that year Michael Evans, defence correspondent of The Times, wrote a lead story in which he revealed that ZIRCON had been cancelled in February 1987, but that 'the Prime Minister and key Cabinet colleagues have decided to keep alive the idea of Britain having its own spy satellite by going ahead with a programme that will rely instead on American technology'. Evans's leak was authoritative and accurate, whereas the Lawson version was incorrect in saying that the British satellite did not leave the ground 'in any sense' - and probably also incorrect about the timing of the cancellation decision. What happened after the furore over the Campbell article died down has not been revealed before, but was discussed by key figures involved in the decision in their interviews for this book.

The Prime Minister's extreme sensitivity may have been connected with the fact that she was close to having to make a decision about ZIRCON at the time that Duncan Campbell's programme was being made. Despite Thatcher's generous attitude towards the intelligence services, she and the senior Cabinet colleagues who knew about the plan were coming to the conclusion that ZIRCON was simply too expensive. In 1987 GCHQ's entire annual budget was about 350 million. The cost to the UK of owning and maintaining a single ZIRCON satellite would have added about 100 million a year to GCHQ's budget in perpetuity. This was about twice the annual cost of the DIS, which had 800 staff at the time, and approached the budget of SIS.

A series of meetings between heads of agencies and senior officials in PSIS in late 1986 had been moving towards the view that Britain could not continue with the project in its original form. They sought an alternative arrangement that would allow GCHQ to enter the space sigint game. Lieutenant-General Derek Boorman, the Chief of Defence Intelligence, was party to those meetings and sums up their outcome: 'Getting that capability was essential. There was no divergence of opinion about it. We may have studied going it alone, but the UK simply isn't able to afford that geographic coverage on its own, so we subscribed to their system.' The 'subscription' that the spymasters had in mind was a cash payment to the NSA to cover part of the cost of one of their new sigint satellite constellations.

In February 1987 a small group of ministers, including Mrs Thatcher and Geoffrey Howe, met to discuss the issue. The intelligence committees had also put two other options to them: continuing with a 'made in Britain' satellite, or scrapping all UK involvement in the field. In the time-honoured Whitehall fashion, the ministers took the middle way and opted to buy into the US system. It was the right choice, says Geoffrey Howe: 'A decision to maintain access to that facility can be justified. Beggars can't be choosers. If you can't afford a wholly independent operation then you have to put in a share.'

At his headquarters at Fort Meade in Maryland, the Director of the NSA was relieved by Britain's decision to abandon the 'go it alone' national project. Lieutenant-General Odom says, 'I never thought they should even have tried ZIRCON. They thought about it for their industrial base, but it didn't make sense.' Not surprisingly, the general preferred the idea of a large GCHQ cheque being paid into his satellite programme rather than the money going to British industry.

The amount spent on ZIRCON up to that point was 70 million, according to the Times report of August. The size of Britain's contribution to the NSA was around 500 million. It was, says one senior figure in British intelligence, 'part of the way we kept up our subscription to the US country club'. Britain had paid a price equivalent to the cost of a single satellite, part of a three-craft American constellation. The complex arrangements were agreed in a super-secret memorandum of understanding between the US and UK governments which I believe was signed in the latter part of 1988. One of the satellites would, to borrow Ernest Bevin's phrase about the British nuclear bomb, have a Union Jack on the side, but Britain could also consider itself part-owner of all of them, sharing the take of the entire constellation. The UK would also have the right to 'task' any of the three satellites for up to one third of the time. On the other hand, the 'British' satellite would never actually be delivered to the UK and the highly-sensitive technology within it remained firmly within the NSA's security system. Furthermore, it is said that the NSA can override GCHQ even in the tasking of the 'British' craft.

Britain's decision to join this US scheme was a one-off financial transaction. Someone party to the deliberations says, 'There was a strong presumption that we'd never replace it.'

In 1994 the first satellite in the second generation MAGNUM series was launched. The payload fairing on the rocket which carried it was around eighty feet long, indicating that the listening dish mounted on the craft may possibly have a diameter as large as 160 feet. Britain appears to have invested in the largest sigint satellite yet built. A second satellite in the series was launched in 1995. It is possible one of these was the British 'bird'.

The decision was a sensible one in several ways. It ensured that British money was going on the best possible technology; there was no duplication of what NSA was doing; and the UK was insulated from the possibility that the satellite might be blown up during launch or fail in orbit. The disadvantages were equally apparent: it transformed the UK from being a virtual client of the USA to being a literal one, reinforcing national dependence in intelligence; it left open the question of what happens when the 'subscription' runs out, particularly if, as some suggest, it was a once-only deal; and it opened up the possibility that a system Britain had partly paid for might be used by the NSA to spy on the UK's allies, further blurring the frontiers of the nation's sigint sovereignty.

These negative aspects made the February 1987 decision a watershed for British intelligence. There was, however, a choice that was not one of the three options presented to ministers: co-operation with France. There were reasons why such a decision would have been timely.

After the shock of Reykjavik, the Cabinet decided to start talks with France on another area central to the UK-US special relationship: nuclear weapons. In fact, the month after the ZIRCON decision, there were discussions about collaboration with the French in the development of an air-launched nuclear-armed missile that could have been used on Royal Air Force Tornado bombers. These talks did not pay off; the Ministry of Defence eventually stated its preference for a US missile, before having to cancel the project altogether owing to financial pressure.

France's intelligence services had also thought through the intelligence implications of the space age, and during the mid-1980s had come to very different conclusions. France had decided that its national commitment to space, exemplified by the Ariane rocket and the extensive complex from which it was launched at Korou in French Guiana, should be considerable. Spending heavily on military space projects did not frighten French ministers - in fact, it appealed to them. France committed itself to buying two photographic satellites code-named HELIOS. In 1986 funds were agreed for the development of a radar imaging craft called OSIRIS and later renamed HORUS. Research was also started into the only other significant area of space intelligence-gathering - the area that interested GCHQ - with the sigint project ZENON.

The price of developing three different kinds of space-based intelligence systems is obviously high. In 1995 the cost of the HELIOS project alone is estimated at about 950 million. France's annual spending on its military space programme grew from about 200 million in 1990 to around 390 million in 1994. Legislators in the National Assembly considered these plans too ambitious and tried to stop some of them. Paris tried to attract Spanish and Italian investment in HELIOS and to persuade Germany to contribute to HELIOS 2 - so helping to pay for a second generation of satellites - in return for allowing these allies to share in the tasking and product of the systems.

A British offer to take a large share - perhaps exploiting GCHQ's know-how to take the lead in ZENON - would probably have been welcomed by France, and would have had at least some benefits for British industry. To speculate along these lines misses the point, however; even in its 'go it alone' form, ZIRCON was at least partly conceived as a tribute to the NSA - a way of paying them back. Joining traditional rivals France in such a venture would have touched deep chords of national insecurity. Furthermore, taking even a one-third share in France's array of projects would, by the 1990s, have been costing Britain more than ZIRCON. One senior civil servant argues that the French programmes were not a real alternative: 'Investing anywhere else would have bought far less capability. The French don't even know how far behind they are.' By 1987 Britain had taken the decision, to borrow Geoffrey Howe's words, to play the role of beggar rather than chooser in the world of high-technology intelligence-gathering.

The saga of Britain's spy satellite, complete with court injunctions and police raids, coincided with episodes in two other difficult and long-running public dramas involving Thatcher and the intelligence services: banning trade unions at GCHQ and trying to prevent the publication of Peter Wright's book Spycatcher. In these three matters, the Prime Minister's determination to pursue her policies won her the admiration of many, but also the antipathy of others in the intelligence world who hated the fuss and public attention she had brought them.

In November 1986 the British government found itself in an Australian court, trying to stop Wright's publishers sidestepping a UK injunction against publication by bringing the book out on the other side of the world. The encounters between Malcolm Turnbull, acting for Wright's publishers, and Robert Armstrong, the Civil Service chief sent from Britain to defend the government line that the ex-MI5 man owed a lifelong duty of confidentiality, provided a daily drama in the British press. Armstrong, trying to deal with Turnbull's references to MI6 - which the Thatcher government did not want to admit existed - agreed to refer to it as 'the other place'. In one session, struggling under cross-examination, the Cabinet Secretary admitted, in a phrase which was to enter the language, that he had been 'economical with the truth'.

In March 1987 the Australian court rejected the UK government's request for an injunction. An appeal failed six months later and the matter then went to the House of Lords, where the government also lost. The Prime Minister had underestimated both the anti-English sentiment in Australia, which found its expression in the courts, and the practical impossibility of getting the Spycatcher genie back in its bottle once copies had begun to circulate.

In the Security Service itself, Wright was widely despised by the staff, who dedicated themselves to the principle of keeping secret what they knew, and who also regarded much of what he said as questionable or untrue. In a subsequent BBC television interview with Panorama's John Ware, Wright's most disturbing allegation - that there had been a plot against Harold Wilson's government - fell apart on screen, with the ageing writer admitting that it had not involved thirty MI5 officers, as stated in his book, but had consisted of little more than idle chatter between Wright himself and a small number of his colleagues. What Security Service officers resented was the fact that the government attempt to ban publication had invested Wright's allegations with credibility. Anthony Duff, Director General of MI5 at the time the decision was made to proceed against Wright, told me, 'The whole thing was a disaster in terms of (a) making a lot of money for Peter Wright and (b) holding up the British state to ridicule. I went along with it. I should have tried to stop it.' Duff had deferred to his Legal Adviser, Bernard Sheldon, and other government law officers who had originally advocated prosecution.

Armstrong's grilling in the witness box had disturbed some civil servants. One Whitehall mandarin recalls, 'I felt very sorry for Robert. I'm absolutely certain he did it as a civil servant doing his duty, not being influenced by considerations of personal discomfort - which you might say is the understatement of the year.'

The banning of the trade unions at GCHQ was effectively accomplished by 1987. Thatcher had launched this initiative in January 1984; it married her instinctive dislike of union power with the desire on the part of some Cheltenham and Whitehall mandarins to remove such organisations from this field of intelligence work. Thatcher and her advisors believed banning unions was vital to the UK-US relationship, which they argued had been put under strain by industrial action at Cheltenham in 1979 and 1981. The union ban became a rallying issue for the left, later producing a Labour Party commitment to reverse the process once in power.

Subsequent verdicts on the affair from two figures who were closely involved and still believe in the principle of de-unionization, but who felt deep reservations about its 'handling' - in other words, the politics of what was done - are instructive. Michael Herman, a senior GCHQ officer until 1987, wrote after his retirement, 'The likely verdict on de-unionization will be that the consequences were not thought through - perhaps as was repeated with the Poll Tax . . . consultation with the Opposition seems to have been no part of the plan. It was not the Thatcher style.' Geoffrey Howe, who as Foreign Secretary had to carry the Parliamentary can for the de-unionization exercise, felt the GCHQ story said something deeper about the government's attitudes to secrecy, and shares Herman's view that the lack of consultation 'was our fundamental mistake. We made it because of our failure to appreciate the difficulties, the subtleties indeed, of moving . . . from darkness to light.'

Thatcher had spelt out her 'say nothing' approach in an interview in the summer of 1984, following calls for an inquiry into the Libyan People's Bureau affair. She referred to remarks made at the beginning of the Falklands crisis by 'someone who knew a bit about intelligence which was totally and utterly devastating in the amount which it gave away', and added, 'The moment you say too much, the sources dry up.' Her remarks were directed against Ted Rowlands, a Labour MP who had served as Foreign Office Minister of State from 1976-9. During the emergency debate on 3 April 1982 that followed the Argentine capture of the islands, he told the Commons that a possible invasion in 1977 had been deterred because of the quality of intelligence available, and said, 'I shall make disclosure. As well as trying to read the mind of the enemy, we have been reading its telegrams for many years.'

During the research for this book, more than one of the intelligence chiefs interviewed cited Rowlands's remarks as a textbook example of the sensitivity of signal intelligence methods and a reason for limiting the role of MPs in scrutinizing sigint activities. One suggested that Argentine intercepts had dried up within hours of the parliamentary disclosure. The Rowlands anecdote does, however, shed more light on the attitude of Thatcher and those mandarins to democratic oversight than it does on the real nature of the south Atlantic intelligence war. While writing this book, I put those allegations to someone with an intimate awareness of GCHQ's product during those months; they said, 'The Argentines were doing everything they could do within their knowledge and apparatus to protect themselves. Even if they were told a channel was unreliable they would not have been able to do anything differently.' In other words, the Labour MP's remarks had had no noticeable effect. They had, however, provided Number 10 with useful political ammunition.

The Prime Minister felt that matters of secret policy were the exclusive preserve of leaders and their intelligence agencies, and could not be understood by others, particularly the media. In her memoirs she notes her sympathy for Ronald and Nancy Reagan, who were 'hurt and bemused' by public comments on the Irangate affair, with much 'cruelty and contempt . . . pouring out from the liberal media'. Regarding Reagan as a friend, she could not help sympathizing with him when his honesty was being questioned by those outside the Administration.

The Irangate crisis had been growing since the autumn of 1986, as revelations emerged about the work of a secret group within the National Security Council (NSC) first in trading arms for US hostages in Beirut and later in using the profits from the sale of weapons to Iran to back the Contra rebels in Nicaragua. Congress had forbidden the use of US government funds to back the rebels, one of the President's right-wing causes célèbres. John Poindexter, the National Security Adviser, had resigned, as had Colonel Oliver North, the principal NSC aide organizing the project. The issues at stake in Irangate were serious enough, but it seems that the British Prime Minister still found it difficult to accept that a leader should come under media and legislative scrutiny for acts committed in their name.

During my research, I learned that SIS picked up information about Admiral Poindexter and Colonel North's activities before they became public. Washington had kept Britain in the dark about its covert hostage strategy, but MI6 learned about it through a Middle East agent. I asked one spymaster whether Whitehall told the Americans it knew what they were up to. He replied, 'All we could do was tuck it away in a box, we couldn't have discussed it with them. This was UK Eyes Alpha, after all!'

At one point during the long hostage saga, Britain received sufficient intelligence to plan a rescue mission. Interviewees suggest it was provided by the Americans. In Immediate Action, published in 1996, a former SAS sergeant writing under the pseudonym Andy McNab suggested that he had been part of such an operation. McNab claimed dozens of troops had been standing by in the Middle East; a small number of members of G Squadron were actually on the ground in Beirut as an advanced party when the operation was stood down. However close the SAS may have got, when the highly reliable information needed to launch such an operation was obtained there was not enough time. In fairness to SIS, which spent years trying to locate the British hostages, the captives were moved every few months (Brian Keenan suggests seventeen times in all) and held in areas where the kidnappers could ensure the discretion of local people.

Thatcher's desire to keep all intelligence matters shrouded in darkness was central to the union problems at GCHQ, the ZIRCON police raids and the Spycatcher court battle. Some senior figures, such as Anthony Duff at the Security Service, were already undermining this strategy by discreet briefings of newspaper editors. Others felt that the publicity attending all three cases had been truly damaging and that the Prime Minister lacked the sophistication to know what was in the best interests of her 'secret services'. One senior figure remarks, 'I did not have great respect for her mind - a 2.2 was about right.' These occasional tensions between the intelligence chiefs and the Prime Minister were eventually to become focused on the central issue of the day: Gorbachev and his sincerity.