by Kahn, David
The mystery was resolved when they were sent to the intercept station near Winchester, a large Royal Navy shore installation named as if it were a ship: H.M.S. Flowerdown. The main intercept building had places for about seventy-five monitors, all facing the controllers at one end of a big room. Work went on around the clock. The intercept operators lived in rows of one-story bungalows.
Each monitor listened through earphones to the peepings brought in by the radio set on a high bench in front of him or her. Each took down all the traffic—every schedule, in the intercept operators’ term—on a single frequency. Axon, for example, copied primarily a frequency used by station DAN at Norddeich, which broadcast weather reports, and secondarily a frequency used by station LLA-bar, whose location she did not know. Some stations transmitted at regular times, others only when they had messages. When nothing was being transmitted, many of the women knitted.
But when a message began, the monitor immediately started to write it down in pencil on the topmost sheet of a pad of intercept forms. No carbon copies were made. He or she noted the time, the frequency, the signal strength, the addressee, and the sender, and wrote down the code groups in the little rectangles on the form’s grid. The monitor was careful to write down only letters of which he or she was certain, omitting the others and indicating the omissions; guesses were not allowed, since a wrong letter was worse than a blank. But the German transmissions were very precise, very uniform; they reminded Axon of the goose step. When a transmission was completed, the monitor ripped the intercept form off the pad and held it up high. Often he or she was holding up one message while taking down another. A controller would take it to the teleprinter to be sent to the codebreakers.
The breaks into SHARK, the Atlantic U-boat key, of December 13, 1942, were followed by others. Not all keys could be recovered: in January 1943 no settings were recovered for a ten-day period, and none were found between February 10 and 17. Except for difficulties like these, however, SHARK solutions, early in 1943, rarely took longer than seventy-two hours and often less than twenty-four, making them of operational value.
Then disaster loomed. Hints appeared in the traffic that the Germans were going to replace the Short Weather Cipher that had just been recovered from the U-559. The new, unknown edition—the third—would cost Hut 8 its kisses. The first sea lord messaged the vice chief of the naval staff, then in Washington, that “U-boat Special Intelligence has received a severe setback. After 10th March it is unlikely that we shall obtain more than 2 to 3 pairs of days per month and these will not be current. After 2 to 3 months the situation should improve considerably.”
The first sea lord was too pessimistic. When the new weather cipher went into effect, Hut 8 used the Short Signal Book taken from the U-559 to find cribs among the numerous short signals being emitted by the U-boats during the heavy convoy battles then going on. Like the weather code, this book used the Enigma in three-rotor mode, which speeded up the work of the bombes. Hut 8 fitted cribs to cryptograms by identifying U-boats through direction-finding, radio fingerprinting, and transmitter identification by TINA as well as through an index of the behavior and procedures of individual submarines and their skippers. As a consequence, of the 112 days from March 10 to the end of June, Hut 8 solved the SHARK keys for 90.
On July 1, however, yet another complication appeared: the Germans brought into service an alternative fourth rotor. Called gamma, it sometimes served in place of the original, called beta (no alpha rotor was ever recovered). Hut 8 reconstructed its wiring crypt-analytically within a few days. The cryptanalysts learned that the choice of rotors was made on the first of the month. This regularity simplified Hut 8’s task: it was only when making the first break of a month that they had the extra task of determining which fourth rotor was being used.
As the convoy battles subsided and the supply of short signals dwindled, Hut 8 exploited the kisses between texts, such as general orders that were transmitted in three-rotor Enigma to surface forces and shore stations and in M4 to U-boats. These provided longer and more reliable cribs, but they took longer to test on the bombes than the short signals.
Bletchley’s first bombe, AGNES, was installed on the grounds of the park itself. But to minimize damage from air raids, such as the one during the night of November 20–21, in which a direct hit destroyed a telephone exchange and a typists’ room and a near miss burst six paces from Hut 4, the next bombes were sited elsewhere. A low, one-story brick utility building was constructed for them at Wavendon Manor, about 3 miles northeast of Bletchley. Still, by the spring of 1941, the codebreakers had only eight bombes and, by the end of 1941, only twelve. Huts 6 and 8 shared the bombes in those days of paucity. The question of who should have them was often difficult but only rarely resulted in fights—usually when Hut 6 was being pressed strongly to break an Afrika Korps key. One reason for this relative harmony was that the deputy heads of the huts at that time—Stuart Milner-Barry for 6 and Hugh Alexander for 8—were good friends. Another was that Milner-Barry always believed, correctly, that losing the Battle of the Atlantic would mean losing the war; he often yielded his bombe time to the naval cryptanalysts, without, however, telling those who were pressing him for breaks.
Gradually the British Tabulating Machine Company increased the tempo of production at its main factory at Ignield Way in Letchworth, some 30 miles north of London. While the first machines took five weeks to construct, the later ones, built in batches of six, came off the assembly line at the rate of one a week. The bombes, about the size of three standing cupboards, were loaded onto ordinary trucks and, to maintain secrecy, given no special escort as they were driven to Crawley Grange and Gayhurst Manor, both stately homes near Bletchley, and Adstock, a village in Buckinghamshire. By August 1942, G.C.&C.S. had thirty bombes and had established a bombe station on London Road in Stanmore, a neighborhood in northwest London. The two parallel one-story brick buildings were divided into several big rooms, each with an office area and ten machines in two rows down the sides. The rooms smelled of oil, metal, and cement. The arrangement proved so successful that a similar facility was built 5 miles away between Eastcote Road and Lime Grove in Eastcote. To these were brought some of the bombes from the country. By March of 1943, sixty bombes were in operation and, by the end of the war, two hundred.
The bombes were black iron monsters 6 to 8 feet high, 10 to 12 feet wide, 3 feet deep. Each one was named. Those at Adstock were named after fighter planes, such as SPITFIRE. Those at Stanmore and Eastcote carried the names of towns in the United Kingdom and the British Commonwealth and of cities in Allied countries. On the face of the machine were rows of wheels, the analogues of the Enigma rotors, each 1½ inches thick and 5 inches in diameter, bearing a large number in the middle and the twenty-six letters of the alphabet around its circumference.
The bombes were tended by Wrens. Bletchley sent them a menu—the instructions for setting up a bombe, based on a crib and the loops derived from it. A Wren put on the rotors in the order specified by the menu, turned them to the menu starting positions, plugged in cables in the rear according to instructions, and turned on the machine. Then, sleeves rolled up, jackets and ties off, they would stand by, chatting over the whirring of the bombe and keeping an eye on it as it ran.
One bombe tender, Diane Payne, had joined the Wrens dreaming of the sea and with the romantic idea of marrying a sailor. At recruitment she was asked if she could keep a secret; she replied that she really didn’t know because she had never tried. Despite this unsatisfactory answer, she was assigned to special duties. After basic training, she and the others were taken to a hut at Bletchley Park and told that their jobs would consist of shift work, with little hope of promotion, and the need for complete secrecy. Unreflecting patriotism led Payne and nearly all the others to accept—although the necessary silence about her work led some of her family and friends to consider her something of a failure in her war work. Another bombe tender, Marjory Mitchell, a Scott who, like all young women ove
r eighteen had been called up for war work, was apparently assigned to the bombes because she had studied in Switzerland and learned four languages, although that had nothing to do with the work she eventually did. Unlike the Oxbridge cryptanalysts, who were never checked for possible subversion, she was carefully vetted, with police in her home towns of Fraserborough and Aberdeen asking friends and neighbors whether she had any connections with Germany. At Stanmore in 1942, after signing the Official Secrets Act, she was put into the D Watch of fifty young women and began working a three-shift system, 8 A.M. to 4 P.M. one week, 4 to midnight the next, then midnight to 8 A.M. the third week, with two days off between each shift change.
The change in hours didn’t bother the women, who sometimes violated regulations by sneaking off to central London after a mid-night-to-8 A.M. tour. But the 2,000 Wrens who eventually tended bombes took their work very seriously. From time to time B.P. would tell them that their work had led, say, to the sinking of a U-boat or the successful diversion of a convoy. “We felt the weight of responsibility that any mistake or time wasted could mean lives lost,” Diane Payne said. Once a coworker turned white while taking tea with Payne and some relatives, remembering what she thought was a mistake while on duty. But she and Payne worked out the problem on a bit of paper and decided her work had been correct; then, frantic about security, they burned the paper in the sink. For some, the strain showed in upset stomachs, collapses, nightmares, going berserk on duty. For others the work was just boring. Marjory Mitchell, perhaps representative of others, did not think the work tedious; it was a job she had to do, it wasn’t dirty, and she enjoyed the company of the other women.
The bombes broke down frequently; on average, one of the eight machines in D Watch’s room would fail on a shift. The soldering joints on the cables in back were a particular source of trouble, frequently coming loose, and the wire brush contacts needed constant attention. Royal Air Force personnel would open the machine and brave its labyrinth of cables and gears to repair it. When a bombe functioned properly, however, the wheels in the top row all spun fast, those in the second row turned at one twenty-sixth the rate, and those in the bottom row seemed hardly to move at all. When the circuitry sensed a possible solution position, or “drop,” the position of the spinning top rotors at that point was shown on an indicator; they were then slowed and stopped. The middle and bottom rotors were frozen in their positions. The Wren would report the settings of all the rotors, as well as the rotor order being run, to Hut 8. Because bombe time was so precious, she restarted the bombe while the cryptanalysts checked to see if the drop represented an actual key or merely a chance concatenation—what the cryptanalysts called a “legal contradiction.” If the latter, they let the bombe run. If it was a key, they called to stop it. It took a bombe about 15 minutes to run one rotor order and so, with shifting rotors from one position to another and dealing with drops, some 100 to 150 days would be needed on one bombe to exhaust all 15,360 rotor orders possible on the menu from a single crib. The Wrens had little idea of what all this meant, but it was a thrill when the correct stop came from one’s own machine.
When the bombes found a successful U-boat key, a half-dozen Hut 8 women assistants, using converted British rotor cipher machines resting on a shelf that ran around their room, transformed the intercepted messages into German on strips of gummed paper, which they glued onto full-sized sheets.
Messengers took them to the Hut 4 watch. Hut 4, like the other huts, had expanded and moved into a larger brick building, still separate from Hut 8. The translators sat around a long table in a double-sized room, and the messengers put the solved intercepts in a basket in front of the No. 2 on the watch. He sorted them according to urgency—an extremely responsible job—and distributed them to the four or five translators on the watch. They handwrote the English version of the German plaintext—usually only a line or two—on a separate sheet of paper. The abbreviations, which were numerous, often puzzled them. One translator, Leonard Forster, considered that they were specialists in the expansion of abbreviations. A separate team ascertained the meaning of the technical terms for newly invented devices—new mines, new torpedoes—and recorded these in a kind of dictionary for the translators.
They initialed their translations and gave them to the head of the watch, who vetted them and decided which should be sent to the Operational Intelligence Centre. One watch head, the prematurely balding A. A. Ernest E. Ettinghausen, considered that his job was chiefly to get the material out quickly and accurately and in the right order of priority. He did not have to motivate his men in this: there was a war on, and they sensed, if they did not know, that “to save an hour … was to gain an hour in which a U-boat gained six miles upon a convoy.” He often picked up the direct telephone to the O.I.C. to alert them that an important message was coming through. The “teleprincesses” addressed the messages to O.I.C. to “I D 8 G,” the ID for (Naval) Intelligence Division, the 8 for its Section 8, the O.I.C, the G for O.I.C.’s liaison with B.P. The sending teleprinters at B.P. made six carbon copies. These went to various sections of Hut 4, such as the unit dealing with U-boat technology. The teleprinters in the O.I.C. likewise made six copies, which were distributed to Denning, Winn, and others who needed that intelligence.
In Hut 4, a copy went as well to Harry Hinsley, who, seated in his smallish office, scanned them all. He could handle the volume of hundreds a day because most of the messages were so short and routine and took only a few seconds to read. He was looking for any variation from the ordinary patterns that might foretell some German activity. Once, for example, he realized that the U-boat addressed was under special orders because the frequency on which the message was sent was one not normally used by U-boats. Another message ordered a sub to rendezvous in an unusual position. Still another revealed that the group of subs was called GOEBEN, the name of a World War I German battle cruiser that achieved fame in the Mediterranean. All these clues gradually made clear to Hinsley in 1941 that submarines were to be sent from the Atlantic into the Mediterranean. And they were.
At the end of 1942, a U.S. Army contingent was sent to Britain to help B.P. Its three units—the 6811th, 6812th, and 6813th Signal Security Detachments—dealt respectively with interception, bombes, and crypt-analysis. The 6812th, with some two hundred men under the command of Captain Mortimer Stewart, a Texan and former I.B.M. employee, worked first at Stanmore and then at a bombe complex in the northern London suburb of Ruislip. Its work area was called UNITED STATES and each of its twelve bombes was named for an American city—among them Houston, for Stewart’s hometown. The Americans were pleased to be working with the Wrens. One amazed lieutenant noted that the buildings at Stanmore “contained marvelous machines and many attractive ladies. The machines were made by the British Tabulating Machine Company and the ladies by God.” The feeling was sometimes reciprocated: one tall, dark-haired Wren officer fell in love with blond Lieutenant Rolf Christiensen at first sight, to the envy of several of his compatriots.
Some of Stewart’s men were also I.B.M.ers, and their practice of keeping records on the machines’ performance delighted bombe maker Doc Keen, who used this information to make improvements. Stewart took pride in his unit’s producing two to three times as many solutions as a comparable Wren unit, not because it ran the machines faster but because the men changed setups much more quickly. The 6812th began work on the U-boat cipher as soon as it arrived in Britain and continued to work a great deal on that prime target.
The role of the United States in breaking the naval Enigma eventually equaled Great Britain’s. The cooperation between the two nations had started long before the American contingents arrived at Bletchley. Though British and American cryptanalysts had been in contact during World War I, a far more intimate relationship began in 1940 as part of a general exchange of scientific information. The Americans agreed to give the British their reconstruction of the main Japanese diplomatic cipher machine, which they called PURPLE, in return for British info
rmation about Enigma cryptanalysis. Four Americans were selected to accompany PURPLE to Bletchley Park and to learn about Enigma: two army reserve officers, Abraham Sinkov, a mathematical cryptanalyst, and Leo Rosen, who had made an important breakthrough in solving PURPLE, and two navy officers, both communications intelligence specialists, lieutenants Prescott Currier and Robert Weeks. Flying was considered too risky, so the men were to travel on Britain’s newest battleship, the King George V, which had brought over the new British ambassador.
The battlewagon anchored in Annapolis Roads on Friday, January 24, 1941. That same day government station wagons brought the four Americans, plus four wooden crates, each about 2 by 2 by 3 feet and containing the PURPLE machine and accompanying papers, from Washington to Annapolis. In a steady downpour the crates were manhandled into a liberty boat to be ferried with the men to the King George V. Rosen, in an open boat, oversaw the crates as they were lifted aboard in a cargo net and stowed in a strongroom below. The KGV sailed the next morning.
The trip across was uneventful except for occasional calls to action stations. The Royal Navy officers in the wardroom kept asking the Americans to explain the cartoons in The New Yorker and Esquire. At 2:32 P.M. on February 6 the ship anchored in Scapa Flow in a heavy snowstorm. The Americans and the crates transferred to a cruiser going to the Thames estuary. Steaming down the east coast of England, they saw the signs of war: ship after ship sunk in the shallow water, their masts and funnels visible. And they experienced the war as well: a German bomber attacked, its bombs exploding on the sea bottom and lifting the ship gently, as if by a hand. The crates, on deck, were peppered with machine-gun bullet ricochets but were not damaged.