I began my analysis of the Lonsdale case by asking GCHQ to provide me with files on any known Soviet espionage case, like the Lonsdale case, which had involved clandestine radio broadcasts. They produced leaflets, around a hundred in all, listing first the details of the agent under consideration - when he started and finished, what his targets were, which service he worked for, and so forth; they then produced a detailed summary of the agent's signal plans, and finally lists of the traffic that he received from the Soviet Union, including the numbers of messages, their group counts, details of the cipher systems used, and dates when they were changed.
I organized this mass of material into KGB and GRU categories, and secondly into types of agent-singletons, sleepers, illegal spies actively running one or more sources, illegal residents running a group of other illegals, and so forth. I found, to my astonishment, that changes in the radio traffic mirrored the different types of agent. For instance, by looking at the operational radio procedures, such as the types of call signs used, it was possible to tell whether the spy was of KGB or GRU origin. Similarly, by analyzing the group counts and the lengths of the messages, it was possible to tell what type of spy was receiving the traffic. For instance, the singleton sleeper received very little traffic, the GRU singleton not much more, while the KGB singleton received a quite considerable volume. The KGB illegal resident, the most important spy of all, always took the greatest amount of traffic - generally between five hundred and a thousand groups a month.
I soon began to realize that the Lonsdale case was utterly different from any other single case among the hundreds I had studied. No other case had so many different forms of communications, some duplicated, and some even triplicated. Yet there was apparently only one spy Houghton serviced by the whole Lonsdale/Kroger apparatus. He was an important spy, it was true, with access to vital details about British and American submarine-detection systems. But why involve the Krogers? Why not just use Lonsdale?
Even at face value, it seemed unlikely that other spies would not have been involved in the ring. The Krogers were located in Ruislip, close to American Air Force installations, while Lonsdale, we discovered, had earlier studied at the School of Oriental Studies, on a course commonly used by British military officers and MI6 trainees.
Lonsdale was certainly the illegal resident in Britain, and I carefully tabulated the communications he received from Moscow after he returned to Britain in October. He averaged 300-350 groups per month. Yet in each of the other illegal cases I studied, the resident received 500-1000 groups a month, and generally closer to the higher figure. Where was Lonsdale's missing traffic? Lonsdale had a three-character call sign which included a figure 1 if the broadcast carried traffic, and omitted the figure 1 if it was a dummy stream. I asked GCHQ whether they could find any messages similar in length to those we knew Lonsdale received after October, for the period preceding his departure in August. After considerable search GCHQ found what is called a "continuity," which went back six years, to roughly the time when Lonsdale entered Britain.
The average group count of this continuity was in the correct range of 500-1000 groups per month, and it ceased suddenly in August 1960, at the same time Lonsdale returned to Moscow. Of course, without the pads, we could not read any of the messages, but if, as seemed likely, this was Lonsdale's original traffic, the question remained: why did it suddenly diminish when he came back?
I turned my attention to the Krogers' communications. These were the most baffling of all. Most of them were for their use, yet they appeared not to be running any spies at all - merely acting as support for Lonsdale. But some of the communications were clearly being stored by the Krogers for Lonsdale. The pads, for instance, hidden like Lonsdale's in a cigarette lighter, were almost certainly his. I calculated up the group counts on the pads. The total was equivalent to the groups missing from Lonsdale's traffic after his return in October. The Russians, it seemed, had split Lonsdale's traffic when he came back, leaving the Shah (Houghton) on the channel we could read, and placing his other communications, perhaps containing his other spies, onto a secure channel with the Krogers, and using their high-speed transmitter, which we could not detect, to send any messages he needed.
This apparent alteration in radio procedures suggested that Lonsdale knew, in some way or other, that the messages he was receiving from Moscow in the White House, using the pads in the cigarette lighter, were compromised. But why, if he feared that, not just use new pads? And why, if the Russians feared he was compromised, was he sent back at all?
I began to analyze the sequence of events over the weekend of the arrests. I had arranged a continuous coverage of the Russian Embassy diplomatic transmitters from the Friday before the arrests until midday Monday. The last Embassy transmission took place at 11 A.M. on the Saturday morning, well before the arrests, and the next was not until 9 A.M. on the Monday morning. So, although a major espionage ring had been smashed, the Russians apparently made no contact at all with Moscow. This beggared belief, unless, of course, the Russians already knew we were about to lift them.
I checked what we knew about the movements of known Soviet intelligence officers in London over that weekend. On the Sunday night, when news of the arrests first broke on the television bulletins, an illegal KGB resident, Korovin, and Karpekov, the KGB legal deputy resident, had dinner together. Our probe microphones picked up every part of their conversation. We heard them listen to the news. They made no comment, and no move to contact the Embassy.
I then looked at the beginnings of the case, and made a shattering discovery which convinced me that the case must have been blown to the Russians. In its early stages, the case was handled by D2 when it was suspected that Lonsdale was Polish. Checking the records, I discovered that D2 were not indoctrinated into RAFTER. They had no knowledge of the fact that the Russians were listening to our Watcher radios, and therefore, prior to the case's being handed over to Arthur, they used Watchers on each of the seventeen occasions they followed Lonsdale in July and August.
Ever since the beginning of LIONSBEARD, all Watcher communications were recorded by MI5 and retained, so I organized a test. I gave Evelyn McBarnet, who worked with Arthur as a research officer, the tape of the Watcher communications during the day that D2 followed Lonsdale to the bank for the first time. I also gave her a London street map book, similar to that used by the Watchers, and asked her to mark out the route she thought the Watchers were following, based solely on listening to their radio communications. Evelyn McBarnet was not experienced in traffic analysis, and had no previous access to the case, but within three and a half hours, she reconstructed the movements flawlessly. If she could do it, the Russians, who had been analyzing our Watcher communications for years, were certainly capable of it too. They must have known from the beginning that we were onto Lonsdale.
By the time I was writing my report, Sniper was safely in a CIA safe house near Washington, where he identified himself as an officer in the Polish Intelligence Service named Michael Goleniewski. One fragment of his story seemed devastating in the context of the thread of ambiguity which ran through the Lonsdale case. He told the CIA that in the last week of July a senior officer in the UB told him the Russians knew there was a "pig" (a spy) in the organization. Goleniewski said that initially he was deputed to assist in the search for the spy, but eventually, by Christmas, realized that he himself was falling under suspicion, so he defected.