F.J. blasted me when he heard that I had talked about the penetration issue in Canada, but I told him that it was impossible to avoid discussion after Hollis' abortive visit. To ignore the problem only made it worse in the eyes of our allies.
In Washington interest was just as acute. I remember a house party at Michael McCaul's, the man who in 1964 became the SLO in Washington in succession to Harry Stone. Angleton and I detached ourselves, and he quizzed me hard on the state of affairs inside MI5.
"What the hell's got into you guys," he kept saying, "Hollis comes out there with some cock-and-bull story about Mitchell. He doesn't seem to know the first thing about the case. There's been no interrogation, and now he says there's nothing in it!..."
I tried to talk him through the case. Mitchell, I told him, was in the clear, we thought, but I stated that as far as Arthur and I were concerned, Hollis was our next suspect. I asked him if he had any information which might help us break open the case. He said he would see what he could do. It was a difficult time for the CIA. Kennedy had just been assassinated and the Warren Commission of inquiry was sitting, and Angleton had pressing problems of his own.
In 1965, British security seemed once again catastrophically bad to American eyes. In just four years a succession of spy scandals and disasters had engulfed both M15 and MI6. First Houghton was unmasked, having betrayed vital parts of NATO's underwater-detection systems. Although the Houghton case was a triumph for MI5's new counterespionage capability, it caused outrage in the U.S. Navy, which had long fostered hostility to its British counterpart. The enmity surfaced at a National Security Council meeting, soon after the Houghton trial, at which the U.S. Navy had sought a complete break in the British-American intelligence and secrets exchange. Jim Angleton and Al Belmont from the FBI nipped the Navy ploy in the bud.
"The only difference between us and them," said Belmont dryly, "is they catch spies, and we don't."
But nothing Belmont said could possibly mitigate the string of disasters which followed. Blake was tried and convicted in 1961, casting doubt on virtually every European CIA operation including the Berlin Tunnel. Vassall was caught the next year, 1962; once again, valuable NATO Naval secrets had gone East because of a British spy. In January 1963 Philby defected, with the British authorities apparently mute and impotent. There were the security implications of the Profumo affair in the same year, with its suggestions, taken seriously at the time in the FBI, that the Russians were obtaining nuclear secrets from Profumo via Christine Keeler - Blunt, Long, and Cairncross confessed in 1964, while other cases simply collapsed humiliatingly in court. The Kodak case in 1964 was one, but far worse in American eyes was the Martelli case in early 1965.
The Martelli case had started in 1963 with an allegation by Fedora that the KGB had a foreign ideological source inside an English nuclear research establishment. He had been operational only in the last one to two years. While this meant that the defector, Golitsin, did not know about him, it severely limited the field of likely candidates. After a few false steps, the investigation closed in on Giuseppe Martelli, who had come to the Culham Laboratory in the autumn of 1962 from Euratom. But Martelli was not cleared for secret atomic material. Despite this, the investigation went ahead. It was possible that, like Houghton in the Lonsdale case, when he was at Portland, Martelli got his secrets from a girlfriend who did have access. When it was found that Martelli HAD a girlfriend who had access to secrets it became CERTAINLY possible that Martelli also had access to secrets that he should not have had.
Further investigation did not produce any evidence that Martelli was able to acquire secrets. A search of his office at Culham produced rendezvous information from a locked drawer in his desk. At this time Martelli was away in Europe on holiday. When he returned he was picked up at Southend airport. He was questioned by Special Branch and identified Karpekov as a Russian he knew. He also had a map in his possession which indicated rendezvous arrangements. As a result his house at Abingdon was searched and a concealment device was found which contained miniature one-time pads a la Lonsdale. Part of a page of one pad had apparently been used. A diary was found which had the details of a grid for transforming letters, and therefore words, into numbers for the one-time pad to be used to encipher a message.
A long meeting was held by Hollis, with Mitchell present, to decide what to do. The crucial factor was that no evidence had been found that Martelli had access to secrets or was passing them to a foreign power. The Official Secrets Act (OSA) did have a clause which made it a crime to prepare to commit espionage. But it would be very difficult to prove that Martelli was doing this. There was no proof that he had been in clandestine communication with a foreign power. GCHQ could attest that the cipher pads were similar to those used by spies to communicate with their Russian masters but, unlike in the Lonsdale case, they could not prove that Martelli had done so. It is not often realized that it was the GCHQ testimony in the Lonsdale case that ensured the defendants were convicted. Without this evidence Lonsdale and his associates would have got off either scot-free or with a minor sentence.
I, as the MI5 SIGINT expert, pointed out to the management at the meeting that the evidence MI5 had was not sufficient to prove even the intention to communicate secrets to a foreign power. The Legal Branch of MI5 were keen to try to get Martelli on "the act preparatory" clause of the OSA to establish it as a valid reason to prosecute under the OSA. To the astonishment of the professional counterespionage officers present, Hollis and Mitchell pressed for the prosecution of Martelli to go ahead. The result was that the Attorney-General did go ahead and MI5 suffered the damage.
Even today I find it very difficult to understand why the Martelli case went ahead, unless one remembers the date of the trial - July 2, 1963. This was at the height of the Mitchell case. It is obvious that at this juncture it would have suited the Russians and Hollis for the CE side of MI5 to be knocked down.
The other case to be considered here is that of Frank Bossard. Early in 1965 Top Hat, the FBI-GRU agent, produced photocopies of documents from the Ministry of Supply of the highest security grading in the guided weapons field, involving high-level secrets of the USA. It was relatively easy to narrow the field of suspects to a few people. The suspects were put under all kinds of surveillance. It was discovered that Bossard, one of the suspects, occasionally during his lunch hour would collect a suitcase from the Left Luggage Office at Waterloo Station. He would go to a hotel in Bloomsbury and book himself into a room under a false name. He would stay there alone for about half an hour. On leaving he would take the suitcase back to the Left Luggage Office and return to work. MI5 in due course removed the case from Waterloo. In it were found document-copying cameras, cassettes of film, and two gramophone records on which there were about eight Russian songs. The details of the Russian songs were copied. The entire contents of the case were photographed and restored to the case, which was then returned to Waterloo. I rang up GCHQ and gave them the details of the records. It took GCHQ less than an hour to identify five of the tunes as having been transmitted by a Russian transmitter, found to be in the Moscow area by direction-finding. This transmitter was known to be a GRU Russian Intelligence Service transmitter.
We decided to arrest Bossard next time he took his case from Waterloo and went to a hotel with it. This took place on March 15, 1965. He was caught in the act of photographing Top Secret documents. When confronted with the fact that MI5 knew all about the tunes on the records, he admitted that he had been supplying photographs of secret documents to the Russians for money through dead letter boxes, i.e. secret caches. He received his money the same way. After his initial recruitment, he had met a Russian only once in nearly five years. He said that the individual tunes broadcast indicated which dead letter box to fill, or not to fill any. MI5 had all they wanted for a Section One prosecution. On May 10, 1965, Bossard was sentenced to twenty-one years in jail.