Code Warriors
Page 22
Verkuijl had shared with William Friedman information he had gathered on Japanese diplomatic codes and had been welcomed at Arlington Hall, at first as an official liaison of the Dutch government-in-exile. But by the following year Verkuijl was already in bad odor with the American authorities; many of his coworkers at Arlington Hall found him arrogant, not particularly hardworking or competent, contemptuous of Americans, and much more interested in collecting information than sharing it. In August 1943, the War Department suggested to the Dutch government that Verkuijl’s skills might be better employed in the Southwest Pacific theater; when that tactful attempt at diplomacy failed to produce results, the U.S. Army’s G-2 bluntly informed the Dutch ambassador two months later that “Verkuijl had seen in Arlington Hall everything he could see” and it was time for the arrangement to end.29
But one of his American coworkers who had taken to the Dutchman was Petersen, who for a while was assigned as his assistant. Before leaving Arlington Hall, Verkuijl put his new American friend and colleague in touch with the head of the coding section at the Netherlands’ Washington embassy, Giacomo Stuyt: explaining that the Netherlands was trying to set up its own signals intelligence organization and needed American help, Verkuijl suggested that Petersen could continue to act as a liaison between the two countries. Over the next decade, Petersen supplied his Dutch contacts with a steady stream of documents, meeting several times a month in a restaurant or car to hand over material he had removed from Arlington Hall.
Stuyt was particularly interested to learn of the Americans’ considerable efforts directed at reading the diplomatic traffic of neutral nations, the Netherlands included. As early as 1945, Stuyt was warning his government that the Hagelin cipher machines that Dutch diplomats and the Foreign Ministry relied upon for high-level communications had been broken by Arlington Hall: one of the many documents Petersen turned over was Friedman’s 1939 paper on the cryptanalytic solution of the Hagelin B-211. Petersen’s roving commission gave him access to a considerable array of materials, but what the Dutch were most interested in by the late 1940s, and frequently pressed Petersen to supply, were documents relating to the cryptanalysis of Belgian, French, British, Italian, and Norwegian diplomatic code systems.30
It was not until 1953 that suspicion fell on Petersen, and then only because a naval officer dismissed for homosexuality named him to investigators as a fellow homosexual. When the FBI began questioning Petersen, he acknowledged his contacts with the Dutch and consented to a search of his apartment, which uncovered a large cache of classified documents and letters from Verkuijl and other Dutch officials requesting information and thanking him for his help. Against the objections of USCIB, Canine insisted that Petersen be prosecuted, and specifically wanted him charged under the recent communications intelligence amendments to the Espionage Act.31
Petersen’s indictment on October 20, 1954, was front-page news in the New York Times, which identified the National Security Agency as “the closely guarded codebreaking organization of the Pentagon.” An Associated Press report carried across the country further explained that NSA
is essentially a radio monitoring service. It has a network of radio receiving stations and other equipment, some of which is based overseas. It listens to the world’s radio traffic, both conventional messages and coded material….Secrecy even tighter than that shrouding the Central Intelligence Agency surrounds the National Security Agency. It is not listed by name either in the Washington directory or in the Pentagon phone directory.32
Though a subsequent Top Secret damage assessment by USCIB concluded that Petersen had practically given away the store (“the catholic scope of the information supplied to the Dutch was enough to keep them, at least, well informed as to the level of competence of the U.S. Comint effort up through 1951”), both governments quickly conspired to keep the true extent of Petersen’s spying a secret. The U.S. State Department said it had “no reason to question the good faith” of the Netherlands government—whose ambassador in Washington, Jan Herman van Roijen, had hastened to issue a statement that his government thought Petersen was acting all along “in accordance with an authorized arrangement between the two countries.” Meeting privately with CIA director Dulles, Van Roijen made a show of fully cooperating with the U.S. investigation, providing a precise accounting of the 538.50 Dutch guilders the embassy had spent on dinners, lunches, and drinks entertaining Petersen. He omitted to mention the $5,000 a year they had directly paid him for his services beginning in 1947, later increasing that to $7,500 a year, which almost doubled his $7,700 NSA salary. A classified internal history of the Dutch signals intelligence service a decade later candidly acknowledged that were it not for the trove of information supplied by Petersen, their cryptanalytic bureau could never have attained its current status and capabilities, which far exceeded those that might be expected of such a small country. The author wryly concluded, “Petersen never even received a thank-you note.”33
Dulles for his part strongly pressed the Justice Department to avoid introducing any evidence at Petersen’s trial that would reveal just how much he had handed over. Prosecutors agreed to limit their case to a few of the more innocuous documents he took, including a copy of the Chinese telegraphic code—a standard system used to render Chinese characters as numbers in commercial cablegrams that NSA, ridiculously, classified as Secret. Expecting a lenient sentence after agreeing to plead guilty to one count of misusing classified communications intelligence information, Petersen was stunned when the judge threw the book at him, ordering him to prison for seven years on the grounds that the offense was “not what the defendant withdrew, but that he withdrew, records from the National Security Agency.”
If Canine was hoping to make an example of Petersen as a deterrent to other NSA employees, that backfired as well; most of his colleagues, aware only of the charges that had been publicly revealed at his trial, thought he had been railroaded by overzealous prosecutors. Petersen was paroled after four years, and was subsequently hired by a former Arlington Hall colleague from the Japanese section, David H. Shepard, whose company, Intelligent Machines Research Corporation, was working to commercialize optical character recognition technology; at Petersen’s trial, Shepard had testified on his behalf and vowed to give him a job as testimony of his belief in his innocence.34
Though U.S. officials briefly worried that some of the information Petersen passed on about American work on Soviet codes might have subsequently leaked from the Dutch to the Russians, hastening the Black Friday code changes, USCIB never could find any evidence to substantiate that. Far more damaging than any direct loss of information to the Dutch—who, however duplicitous they were in spying on their ally, were after all members of NATO—was NSA’s own handling of the case, the board concluded:
Last, but by no means least, the publicity attending the arrest and prosecution of Petersen has very materially increased the already damaging aggregate of information in the public file on the nature and extent of the U.S. COMINT effort in general….It must be assumed further that all nations of the world have once more been alerted to the fact that if their communications can be read they very likely will be read. This can hardly fail to reduce the extent of cryptologic naiveté which has been so helpful to the U.S. COMINT effort in the past.35
USCIB’s damage assessment also insisted that “the polygraph screening procedure now used by NSA (and CIA) would have detected one major weakness of Petersen’s and caused his rejection at the time.” That was a not very veiled allusion to his homosexuality, but it was also a logical absurdity. Although it was not far-fetched to worry in the social milieu of the 1950s that homosexuals might be subject to blackmail by foreign agents, Petersen’s homosexuality had nothing whatever to do with his espionage. Petersen told his NSA and FBI interrogators that “he did not know why he did it.” Those who knew him thought it had simply been the sense of importance Verkuijl’s friendship gave him. At Arlington Hall, Frank Lewis and his wife had tried to bef
riend Petersen, too, recalling him as “such a strange guy,” socially awkward and lonely.36 But that was hardly a particular sign of anything out of the ordinary at a place like NSA or GCHQ. As a British security commission would later conclude, “because of the nature of GCHQ’s work and their need for staff with esoteric specialisms, they attracted many odd and eccentric characters,” and trying to spot a security risk by keeping an eye out for unconventional behavior at a signals intelligence agency was like trying to find a KGB agent by looking for someone with a Russian accent.37 The very nature of the work, the burgeoning growth in the size of the staff, and NSA’s unquestioning embrace of the mission to “get everything” had created a situation ripe for exploitation by even the friendliest foreign power.
—
Simply trying not to drown under the torrential flow of data from its field collection sites was becoming a major, and self-justifying, driving force behind NSA’s relentless expansion since the start of the Korean War.
By 1955, two thousand intercept positions were sending in thirty-seven tons of printouts and paper tape a month, most of it shipped by air; more urgent traffic was forwarded by radio teleprinter at the rate of thirty million words a month. One single Soviet system was producing two million messages a month, mostly plaintext. It was not unusual for the machine section at Arlington Hall to punch a million IBM cards a month on just one problem. In June 1955, Canine noted that even with the additional 1,065 civilian positions he had recently been authorized to fill, “we…are still unable to process fully all the material we receive,” and current expansion plans for field intercept were going to increase the amount of incoming traffic by 60 percent when completed in 1958.38
Canine was appalled upon his arrival as director to discover what an internal study had just described as the “deplorable and deteriorating state” of the agency’s communications systems and its inability to handle the flow of incoming traffic. It could take five to six days to deliver a routine message. Even the most urgent messages took at least five to six hours to get from the field to the cryptologists at Arlington Hall: basic security demanded that intercepted traffic had to be reenciphered using NSA’s own codes before being sent over the airwaves to avoid letting a target know which of his messages were being collected and studied, and if it was coming from a distant field site it might have to be sent through as many as six relay centers on the way to Washington. At each of those centers a teleprinter would punch a tape from the incoming signal; it would then have to be decoded by a cleared operator, reencoded, and a new tape punched and placed in a box stacked with other messages awaiting transmission on an outgoing circuit to the next radio station down the chain. By the mid-1950s, NSA was sending or receiving 70 percent of all encrypted traffic flowing in and out of Washington. Canine pushed for a system of dedicated cryptologic radio channels manned only by cleared personnel and equipped with automated switches at relay stations that would seamlessly handle the forwarding of the traffic, and work had begun on the system, known as the COMINT Comnet. But funding was slow to come and technical disputes with the services kept delaying basic decisions.39
Innovations in radio technology were meanwhile filling the airwaves with new types of signals to be intercepted. During the Korean War, in early 1952, the voice channels connecting Russian and Chinese ground controllers with MiG pilots in the air suddenly went silent; after several months of searching, U.S. intercept stations discovered that the communications had all shifted from HF to VHF channels. Very high frequency radios, operating at frequencies of 30 to 300 MHz, had been experimented with during World War II, but this was their first major operational appearance. VHF could carry more information in a smaller bandwidth, but it also was limited to line of sight, which meant that it was no longer possible to place an intercept station hundreds or even thousands of miles away and pick up a signal that had “skipped” off the earth’s atmosphere, as was the case with HF. Radars, and radio signals in airborne IFF systems—used to identify friendly aircraft to one another and to antiaircraft units to prevent accidental shootdowns of one’s own planes—also used these much higher-frequency, line-of-sight signals in the VHF or even higher UHF bands.
So, too, did the new microwave links that communications companies were beginning to use to handle large volumes of short-range, point-to-point telephone and data traffic as an alternative to laying expensive landline cables. In the mid-1950s, Western listening stations in Berlin discovered that the Soviet military had installed a spiderweb of crisscrossing microwave links in East Berlin, carrying multichannel signals. Covering all of these new kinds of transmissions required not only entirely new types of receiving equipment but a proliferation of intercept stations located considerably closer to their targets than in the past.40
This was one area where the British retained an important advantage over their American cousins: that was a chief reason NSA remained willing to continue the BRUSA agreement into the postwar period. West Berlin offered an incomparable outpost right in the heart of the enemy camp, but elsewhere in the world, thanks to its foothold in the former empire, Britannia still ruled the airwaves. Even in many former colonies that had since achieved independence, Britain had negotiated an ongoing military presence. British intercept posts located at military bases in Ceylon, Singapore, Malta, and Cyprus as well as naval bases at Kiel in Germany and in Turkey and RAF bases in Europe and the Middle East (from which the RAF also operated its own ferret flights) were well positioned to collect Soviet signals. Britain also turned over to the Americans the World War II–era radio intercept station at Chicksands, an RAF base in southeast England, which gave NSA a major outpost of its own and would later house one of the elephant-cage antenna arrays.41
Although the British dominion countries, which included Canada and Australia, were not parties to the 1946 BRUSA intelligence-sharing agreement, their signals intelligence bureaus were heavily controlled by London, and GCHQ’s governing body, the London SIGINT Board, had the authority under BRUSA to disseminate any shared communications intelligence to the dominions without specific U.S. approval. Gouzenko’s 1946 revelations of Soviet spies in Canada and the discovery the following year of the NKGB Moscow–Canberra messages pointing to Communist infiltration of the Australian government prompted the United States to cut off intelligence cooperation with those countries. But after the Conservatives won Australia’s elections in 1949 and replaced the previous Labour government, concerns diminished and Australia, along with New Zealand and Canada, were ultimately accepted as full members of BRUSA, whose official name was changed in 1954 to UKUSA to reflect the inclusion of the United Kingdom’s dominions. Australia’s signals intelligence service subsequently helped staff a listening post in Hong Kong that added to the coverage of the periphery of the Communist countries.42
That still did not necessarily solve the problem of intercepting short-range, line-of-sight signals, but here the British with their tradition of naval daring and reputation for eccentric pursuits were also able to make a significant contribution. Apparently no one thought much of it when yet another team of British archaeologists showed up and set up camp in some remote and inhospitable location, and more than once a British undercover team using that eminently believable story carried out missions to intercept Soviet radar and missile-test signals from spots such as northern Iran along the borders of the Soviet Caucasus. The Royal Navy also began conducting signals-gathering missions with submarines that could slip in undetected near major Soviet naval bases. Although the Soviets had already demonstrated their willingness to confront “unfriendly air intrusion,” noted the director of British naval intelligence, Rear Admiral John Inglis, “no difficulties were placed in the way of submarine visitors.” Submarines operating off Murmansk had collected “considerable VHF voice, IFF, and radar” traffic from Soviet air and coastal defense units in the area. The Royal Navy was preparing to expand these operations when a bungled spy mission executed with almost unbelievable ineptness by the British SIS broug
ht it all to a screeching halt.43
The facts still almost seem to defy belief, the storied British spies acting more like Inspector Clouseau than James Bond. On April 17, 1956, the Soviet leaders Nikita Khrushchev and Nikolai Bulganin arrived in Britain for a state visit. It was the first-ever visit to the West by any Soviet leader, and Prime Minister Anthony Eden viewed it as affirmation that Britain still counted as one of the Big Three world powers. The Russians traveled on the latest cruiser of the Soviet navy, the Ordzhonikidze, accompanied by two destroyers. The day before the warships were to arrive at Portsmouth harbor, two men checked into the city’s Sally Port Hotel. The first gave his name as “Smith” in the hotel register, an alias that might have been criticized for a certain lack of imagination had the point not been rendered moot from a tradecraft standpoint when he added in the address column, “Attached to the Foreign Office.”
The second man registered under his own name. Lionel “Buster” Crabb was a minor celebrity in Britain. A World War II “frogman” who had carried out a number of courageous diving exploits, Crabb had earned the George Medal for heroism. He was now forty-seven, retired from the navy, and not the man he had been after years of heavy smoking and drinking. The following night he rang up an old navy diving buddy in the area, and when they met at a pub, Crabb asked his friend if he could help him with his gear in the morning, as he was going to carry out a “secret” dive in the harbor to “take a dekko at the Russian bottoms.” Crabb then proceeded to down five double whiskeys, each followed by a beer chaser. At seven the next morning he met his friend, donned his gear, slipped into the harbor, and was never seen alive again.44