The Secret in Building 26

by Jim DeBrosse

  Others at OP20G may have taken the lead in making use of the tabulators, but Driscoll’s analytical acumen and doggedness eventually broke the code. Thanks to her work, the Americans learned in 1936 that Japan’s newly refitted battleship, the Nagato, had reached speeds of more than twenty-six knots, a significant leap over its projected twenty-four knots. As a result, the Navy quickly redesigned the newest American battleships to eclipse the Japanese by reaching twenty-eight knots. Laurance Safford ventured that this coup alone “paid for our peacetime RI [radio intelligence] a thousand times over.”

  Driscoll continued on her roll of breakthroughs. In the mid-1930s, she led the Navy’s attack against a new automatic enciphering machine for Japan’s naval attachés, the M-1. Cracking it was a major accomplishment that gave her and her team the confidence to tackle more sophisticated mechanical ciphers, including the Enigma.

  Like many cryptanalysts, Driscoll paid a personal price for her successes. The stress of grappling with the Japanese codes led to periodic bouts of weight loss, and often she returned to her family in Westerville and the slower pace of life there to regain her stamina.

  Still, there was nothing frail about Driscoll’s dealings with her male counterparts in the Navy. Those men who tried to patronize “Madame X,” as her colleagues sometimes called her, seldom did a second time, according to several Navy veterans who knew her after the war. They recounted a famous confrontation between Driscoll and a gung-ho Navy lieutenant who had dared to question her application to renew her security clearance after the war. The young officer barged into her office, approached her in an overbearing manner, and said, “Ma’am, in this section where you are required to list five references, you have listed the names of five admirals. You are supposed to list the names of people who know you well.”

  Driscoll looked up at the young man with a squint in her eye and said, “Sonny, I knew all of them when they were ensigns and lieutenants, and if you don’t straighten up, I’m going to tell them to never promote you.”

  In his book, And I Was There, Admiral Edwin T. Layton wrote that Driscoll

  was sensitive to her role as a woman in a man’s world. Because of this she kept to herself as much as possible and none of us was invited to socialize with her and her lawyer husband. While she could be warm and friendly, she usually affected an air of intense detachment, which was heightened by her tailored clothes and shunning of makeup. It was surprising to hear Miss Aggie curse, which she frequently did—as fluently as any sailor whom I have ever heard.

  Driscoll had many contacts among Washington’s tightly knit cryptanalytic circles, in all branches of the military, including Elizebeth Friedman, the Coast Guard’s chief codebreaker and wife of the famous Army cryptanalyst William Friedman. Driscoll may have been party to the exchanges between the Army and Navy in the 1930s that, some in OP20G claim, eventually helped William Friedman and his colleagues use statistical techniques, as well as cribs, to conquer Japan’s most secure diplomatic cipher machine, known as “Purple,” in late 1940.

  After piling success upon success, Driscoll’s private life took a sudden, tragic turn in October 1937 when she was seriously injured in an auto accident. Two other passengers were killed. She took nearly a year to recover fully from her badly broken leg and jaw. When she finally returned to full-time duty in September 1938, she needed a cane to walk and, by some accounts, had grown even more reticent and withdrawn.

  Regardless, within two years, she was leading the Navy’s attack on JN-25, the newest Japanese naval-operations code. She and her team were able to make some inroads, but by late 1940 only a fraction of JN-25 was readable, even though Driscoll and her group had solved the logic of the system. The remaining grunt work was turned over to a new team of relatively inexperienced reserve officers who arrived in Washington as the Navy quietly prepared for war. As the team labored slowly and tediously to reconstruct the new codebook, the Japanese introduced a new, fifty-thousand-group list of additives on August 1, 1941, forcing the codebreakers to start their arduous task all over again.

  In the meantime, Driscoll had been reassigned to what the Navy apparently felt was a more pressing matter: the German U-boat codes and ciphers.

  THE EXACT WHEN, who, and why of the Navy’s decision to divert its most experienced codebreaker from the critical Japanese problem—a year before Pearl Harbor—are unknown. But surely the growing involvement of the American Navy in the Atlantic and the fears that England, under Nazi siege by air, might be invaded must have been factors in redirecting Driscoll and her team, as well as much of America’s radio-intercept equipment, to the Atlantic.

  Because the Americans couldn’t read the German messages, they concentrated much of their effort on radio-direction finding and traffic analysis—that is, studying the number, origins, and types of radio transmissions rather than the content of the messages themselves to guess at enemy activities. OP20G had begun intercepting and logging German naval communications as early as 1938 and was gaining at least some knowledge into the nature of the U-boat communications network.

  Driscoll decided to attack the German systems with a labor-intensive approach, despite the Navy’s scarcity of experienced people. She started with former Navy radioman Milton Gaschk and at least two women. In late 1940, she added to her team four young Navy officers, led by a new Annapolis graduate, Robert Weeks. Those four officers, Gaschk, and one of the new civilian employees, Eunice Rice, were assigned a near impossible task: define the inner workings of the mystery machine that was generating the German naval messages without a copy of the machine or any of its decrypted messages. She asked the young group to perform “pure” cryptanalysis against a foe they couldn’t even identify.

  Driscoll provided them with copies of 113 intercepted messages, and that was it, except their own ingenuity. The team put the messages through a variety of analyses, ranging from the statistical Index of Coincidence techniques of William Friedman to the use of cribs. Amazingly, they came close to understanding how the naval Enigma worked but could not define enough of the machine’s parameters to allow decryption to begin. They were able to report that the machine, one they had never seen, generated a cipher that corresponded letter for letter to the plain text, used all twenty-six letters of the alphabet, never enciphered a letter to itself, and used X and Y as special characters, such as punctuation. They also spotted that the machine had two basic and changeable components: one that generated encrypted letters and another that was some sort of means to link thirteen pairs of letters together (the plugboard). The group’s findings were a codebreaking tour de force, but they had to admit they still did not fully understand the workings of the machine.

  That the Navy would expend its top talent on a blind attack on the Enigma raises some interesting questions about the motivations of its decision makers. Safford, who may have played a key role in Driscoll’s reassignment, was to make it clear over the next two years that he feared America’s dependence on Britain for tackling the Atlantic problem, especially as more of the Navy’s own men and ships entered harm’s way. On the other hand, Driscoll’s reassignment orders might just as well have come directly from the White House, where President Roosevelt was determined to give the European war the highest priority. There are no documents to support either theory.

  Certainly, as bombs rained on London during the Battle of Britain, its leaders were pressing the United States for more supplies and technical aid; by the summer of 1940, Britain was asking to exchange cryptanalytic as well as scientific secrets. Roosevelt and his advisers, including William Donovan, who would become his chief spy, were receptive and began putting the pressure on Army and Navy signals-intelligence divisions to come up with plans for collaborating with the British.

  The sharing of intelligence is fraught with risks even between the closest of allies, and, as the two countries moved by necessity toward partnership, both employed ruses and smoke screens to hide their intentions.

  While it’s hard to imagine today,
with the Internet, satellite TV, and scores of daily transatlantic flights, the communication and cultural barriers between the British and the Americans ran deep sixty years ago. Most citizens of either country had seen the other only in movies, where stereotypes were reinforced: class-conscious, snobby Brits talked down their noses, and uncultured, boorish Yanks couldn’t keep a secret.

  But while the Americans sometimes gave the British ample justification for their distrust, the British presumed that their colonial offspring were naïve, undisciplined, and childlike. The British Army’s Captain Geoffrey Stevens, Bletchley’s liaison to the U.S. Army’s codebreaking unit in Washington, sent a letter to Bletchley in September 1942 with a scathing report on the Army operations that included this observation:

  Sometimes I think they are just a lot of kids playing at ‘Office.’ You must have noticed yourself how very many childish qualities the American male has: his taste in women, motor-cars, and drink, his demonstrative patriotism, his bullying assertion of his Rights, his complete pig-selfishness in public manners and his incredible friendliness and generosity when he likes you—Hell! Anyone would think I didn’t like them. But perhaps it is as well I’m fond of children.

  The British sometimes affectionately teased the Americans about their stereotyped traits, including their heavy reliance on machinery. Joe Eachus, the first American codebreaker to work at Bletchley Park, recalled what a British officer had had to say when he found Eachus trying to assist a clerk whose skirt had been caught in the wheel mechanism of a Bombe and was being twisted and pulled from her waist: “Don’t you Yanks ever do anything by hand?” The American military, and especially the Navy, had its own caricatures of the British and, like younger brothers out to prove themselves, were sensitive to any signs of being treated as a junior partner.

  When Driscoll’s group began its work on the Enigma in late 1940, U.S. Army and Navy codebreakers were feuding over how much information to reveal to the British. Both services were willing to exchange radio intercepts, but Friedman’s Army codebreakers, who dealt mainly with diplomatic systems rather than military ones, were more willing to share their successes and failures. The Navy, on the other hand, was more concerned about its dependence on Britain and may have still been smarting from the domineering treatment it received from the Royal Navy during World War I, when the British tried to dictate U.S. naval standards down to the types of weapons and ammunition they should use. Safford and his boss, Leigh Noyes, the director of naval communications, stubbornly refused to reveal any of the Navy’s cryptanalytic conquests, including one the Army later considered its own, the then imminent break into Japan’s Purple enciphering machine. After all, in a world still hungry for empire, Britain had always been among the Navy’s list of potential enemies at sea.

  In November, a final top-level promise was made in Washington to send American codebreakers on an exchange mission to England. Friedman’s Army group quickly devised a complete “gift list” for the British, including copies of the Purple machine, then under construction at Navy facilities using Navy funds. The Army felt it had every reason to expect full reciprocity from the British, but to be on the careful side they drew up a detailed list of what they wanted in the exchange.

  The head of naval intelligence, Rear Admiral Walter S. Anderson, informed the British that both he and the Army demanded an unambiguous answer to what British codebreakers had meant by a “pretty free exchange” of secrets in their latest offer. He insisted that unless there was a guarantee of a complete exchange on Italian, German, Japanese, and Russian systems, the Army and Navy might withdraw. Even if the guarantee was made, he said, the Navy was not going to send any of its codebreakers on the exchange trip—on the excuse that it had none available.

  Meanwhile, the British were grappling with their own reservations about the Americans: “We are entitled to recall that America sent over at the end of the last war the now notorious Colonel Yardley for purposes of cooperation. He went so far as to publish the story of his cooperation in book form.” Indeed, Herbert O. Yardley—ladies’ man, high roller, and head of America’s codebreaking effort during World War I—had penned a memoir in 1931, The American Black Chamber, that became a bestseller. In it, he detailed his successes against enemy codes; later, he supplied the background for a Hollywood spy film.

  Thus, the British codebreakers were not planning to reveal all, especially about the Enigma. If an American expert did arrive at Bletchley, “steps will be taken to steer him away from our most secret subjects,” an internal memo read. But to avoid alienating America’s leaders, the British sent an ambiguous response to the proposed exchange.

  It worked. The top echelons of the Army and Navy told their codebreakers the British were to be trusted, the Americans were to share all their secret machines, and both Army and Navy were to send representatives. By mid-December, with the planned trip to England just weeks away, OP20G had to race to catch up with the well-prepared Army team. Laurance Safford, however, remained convinced that the Navy had little to learn from Bletchley Park, and his last-minute selection of rather junior people reflected that attitude. In contrast, the Army quickly promoted its two top civilian codebreakers to high military rank to try to ensure that they would be respected by the British. *10

  In February 1941, two young Navy communications officers selected for the exchange with Bletchley Park—Robert Weeks and Ensign Prescott Currier—boarded Britain’s new, showpiece battleship, the King George V, for what would become a stay of several weeks in England. The final leg of their trip, from the Orkney Islands to the Thames, proved to be an adventure when German planes attacked their British destroyer, but with little effect. Lieutenant Weeks, from his work in trying to identify the German cipher machine, should have known what information Driscoll sorely needed. Currier was far less experienced. He had been a noncommissioned officer and then a civilian employee at OP20G before being recalled to active duty in December 1940.

  Neither Currier nor Weeks was mathematically trained, and neither considered himself a polished cryptanalyst. Originally, Robert B. Ely, a Navy reserve officer with the background to grasp the intricacies of the British method, had been assigned to the exchange, but Weeks took his place at the last minute.

  Likewise, William Friedman did not make the trip, claiming “illness” as a cover for what his subordinates knew was a nervous breakdown suffered just after New Year’s Day. Instead of going to Bletchley, he checked into Walter Reed Hospital. Friedman was replaced by an electrical engineer, Leo Rosen, who had recently joined the Army’s codebreaking team as a civilian employee. The head of the Army’s mission, Abe Sinkov, was one of the young civilians Friedman had hired when he created the Army’s cryptologic branch in 1931. But neither Sinkov nor Rosen had significant experience with German military systems. In fact, none of the Army’s codebreakers had spent much time on even the earliest Enigma machines.

  The Americans weren’t the only ones nervous about the visit. Barbara Abernethy, Alastair Denniston’s personal assistant, remembered her excitement and her fears when the commander told her to prepare for the arrival of Bletchley’s first American visitors and to serve them sherry in his office: “So at twelve o’clock the bell rang and I went in and I somehow managed to pour glasses of sherry for these poor Americans, who I kept looking at. I’d never seen Americans before, except in the movies. ‘What were they doing here?’ we said to ourselves.”

  With the United States still not officially at war and with opposition at home from isolationists, the American visitors to Bletchley wore civilian clothes to hide their mission. They were housed just a few miles from Bletchley in the mansion of Lord Cadman, president of the Anglo-Persian Oil Company. A stenographer was made available to the team, and food rationing was suspended for their hearty meals. Two large cars with drivers—one for the Army, one for the Navy—were provided for their travels, perhaps because the British had made separate promises to the two branches. The generosity was all the more notable conside
ring the privations the codebreakers faced at Bletchley: a pencil sharpener was rare, paper scarce, copying equipment only to be dreamed of. Their working quarters, called huts, were near-primitive, sectional-built wooden structures made of Canadian pine and covered in either shiplap boarding or asbestos sheets.

  The four American codebreakers brought with them nearly one ton of documents and machinery. Among the items delivered to the British was America’s crowning intelligence achievement at that time: two working copies of Purple, the Japanese diplomatic service’s cipher machine. The Americans handed the British two copies of nearly every item, including Japanese consular codes, JN-25, and Japanese naval attaché ciphers. The double copies were a generous offering in the days when highly secret material had to be reproduced by hand.

  In return, the Americans were given a tour of Bletchley’s decrypting room and shown the techniques and machines the British were using against Germany’s Enigma system, including the only device capable of quickly penetrating the Enigma: the Bombe.

  The Americans never knew how close they had come to leaving Bletchley without having seen its prime treasures. Permission to show them Britain’s Enigma capabilities had been withheld until literally the last days of their scheduled stay. In seeking that permission, Bletchley’s leaders argued that England would probably have to depend on American help in the future, especially for Japanese problems, and that the Americans would be bitter if they learned of the Bombe and its methods after their aid had been secured. Brigadier Stuart Menzies, the head of British intelligence who was then known only as “C,” passed the request to Prime Minister Churchill on February 26. Churchill gave his approval the next day, less than a week before the Americans were scheduled to leave.