Seizing the Enigma: The Race to Break the German U-Boat Codes, 1933-1945

by Kahn, David

  “If it’s locked,” he said to himself, “there must be something there.”

  He pulled out his pistol to shoot the lock off, as he had seen it done in movies. He cocked the gun, aimed it at the lock, looked away, and pulled the trigger. The gun fired; the lock fell apart. Warmington pulled open the drawer. Inside was a wooden box about 9 by 5 by 3 inches. He took it out and opened it. There, unharmed, lay two black disks about the size of hockey pucks, with electrical contacts and letters around the circumference and an indented flange. He had never seen anything like that before—the Royal Navy used codebooks and he knew nothing of the Enigma machine—but he recognized that they had to be for ciphering. He closed the box and looked about the cabin a bit more.

  He did not have time to search the wireless office before the Somali flashed for them to return. They had been aboard the Krebs for forty-five minutes, and Caslon may have been worried about air attacks. Warmington handed the charts and other papers to Aslett but kept the rotor box as the three officers got the five prisoners onto the Norwegian vessel that had ferried them over. The Somali rewarded the Norwegian fishermen with cigarettes and food. Once back on the Somali, Warmington gave the box to Caslon, who glanced at the rotors, understood what they were, and sent the box directly to his sea cabin, the room just beneath the bridge where destroyer captains spent most of their time at sea when not on the bridge. Warmington never saw the box again: Caslon was very security-minded. Aslett gave Caslon the papers the three had collected.

  As the action on land continued, the Somali sought to sink the Krebs. It was not easy. Whalers were built very sturdily to withstand the ice of the northern seas. Twice the destroyer steamed past the Krebs and dropped depth charges near her; the charges shook the British warship and sent her cutlery flying but left the Krebs undisturbed. Then the Somali fired a broadside: the Krebs rocked but stayed afloat. The gunfire finally took effect, and at 10:30 the Krebs sank, taking with her the eighteen bodies of her twenty-four-man crew.

  By this time the commandos were reembarking. They had sunk 10 ships, destroyed 18 fish-meal and oil factories, and set afire 800,000 gallons of oil. They took with them 12 quislings and 213 German servicemen, civilians, policemen, and merchant mariners. A Norwegian trawler escaped to the Faeroes. At 1:30 P.M., the destroyers and the troop carriers assembled and, in day cruising order, set course at best speed down Vestfjord, leaving a pall of smoke over the scene of the day’s operations. At 4 P.M., Caslon radioed the success of the mission to the Admiralty, but said nothing about the cipher wheels or the other documents.

  As it turned out, these included nothing less than the Enigma key tables for February: the inner and outer settings and the plugboard setting. Upon his return to Scapa, Caslon forwarded everything to the director of naval intelligence. From there the cipher material went to Bletchley Park.

  11

  KISSES

  DESPITE THE DRAMA OF THEIR RECOVERY, THE ROTORS THAT Lieutenant Warmington had gotten did not help Hut 8; it already had them. What did help were the February Enigma keys, registry number 1566, indicator HAU, that had been scooped up from the Krebs. When the keys arrived on Wednesday, March 12, Alan Turing receipted for them and immediately put them to work.

  The volume of solutions jumped at once. On March 12, Naval Section teletyped ten messages to O.I.C.; on the thirteenth, thirty-four. The volume remained at that level or higher for almost a week, as Hut 8 read much of the Home Waters back traffic for February. The number of messages teletyped to O.I.C. declined thereafter, but within five weeks Hut 8 had read the back traffic for all of February. The knowledge gained from these solutions helped the cryptanalysts crack, between April 22 and May 26, all of the naval Enigma traffic for April. Later they read, with delays reduced to only about a week, much of the May traffic. Though this did not have much immediate effect on the sea war, it provided information for a series of events that did. And by showing that the naval Enigma could be broken, it led G.C.&C.S. to bring in more people as cryptanalysts to work on that problem.

  Turing recruited his friend Shaun Wylie, whom he knew from Princeton, where both had obtained doctorates in mathematics. Wylie, another member of Hut 8 said, “had a very exact logical mind. When he understood anything, he seemed to understand it completely.… Shaun was a perfect gentleman who never lost his temper except on purpose, and was an extremely good listener. I used to believe that he wouldn’t interrupt a conversation even to mention that the war was over.” An international hockey player and a first-rate teacher, Wylie eventually became president of the Bletchley Park dramatic club and a winner of the unarmed combat competition of the local battalion of the Home Guard. He never mentioned any of his accomplishments.

  The coworker who made these observations came to Bletchley six months after Wylie. He was Irving John Good, a short, wiry Cambridge mathematician who had been born Isidor Jacob Gudak, the son of a London shopkeeper. At Cambridge he had won the prestigious Smith mathematics prize and, placed on the draft-deferred list for scientific work, was doing postdoctoral work in his field, unhappy that it was not war work. One day the university employment office summoned him for an interview. As he was walking to the appointment, a friend told him to wear his scarf inside his coat, not outside, so he would look less like an undergraduate. The scarf may have helped, but more important was that one of the three Bletchley men who interviewed him was Hugh Alexander, with whom he had played chess. Good always suspected that he was taken on because he played chess well. When he reported to Bletchley, Alexander met him at the railroad station. As they walked, he told Good about the Enigma in a sensational conversation that Good would never forget. He was pleased and excited to realize that his work would be of immense importance to the war effort.

  In Hut 8, his first assignment was to read a badly typed, very technical, very hard-to-follow description of the Enigma and how the cryptanalysts were attacking it. Presently he asked Turing, “How on earth did we get the wiring of the rotors?” Turing, who was seated at a table with his friend Joan Clarke, replied briefly, “Perhaps the Poles.”

  “Perhaps a pinch?” Good suggested.

  “Something like that,” said Turing, and Good saw that he should not ask any more questions.

  G.C.&C.S. was wise—or desperate—enough to enlist women as cryptanalysts as well as clerks. Before the war Hilary Brett-Smith had been teaching Anglo-Saxon at various Oxford colleges; the students thought the subject was boring, and she enjoyed converting them by explaining such things as why English speakers today say “foot” and “feet” but not “book” and “beek.” She was casting about for some war work when one day, just after the evacuation from Dunkirk, a mathematician whom she had known from the Dragon School, a private grammar school in Oxford, who was now at Bletchley, urged her to go there and work, though he couldn’t tell her what the work was. Since the Foreign Office paid more than the War Office and was regarded as rather grand to work for, Brett-Smith accepted. And because of the gender ambiguousness of “Hilary,” some clerk somewhere assumed she was a man and allotted her a man’s salary. When she was recruited it seemed likely, because of her linguistic background, that she would work on making the many card listings of foreign technical terms and names; but in fact she began in Hut 8, working with Turing and Twinn and chess expert Harry Golombek.

  One of the most intuitive of the cryptanalysts was Mavis Lever, who was eighteen when she arrived at Bletchley. She was sent there because she had studied German at London University. Assigned to work with Knox at the Cottage, she received her introduction to his arcane labors with the greeting, “Hello! We’re breaking machines! Have you a pencil? Well, here you are.” It was Lever who discovered that the baffling hexagram STGOCH did not stand for “St. Goch,” a place or sanctified soul that none of the pundits had been able to find in any gazetteer or catalogue of religious names; nor was it an error for “St. Roch,” a town in France, but instead was an abbreviation for “Santiago, Chile.” It was also she who sensed something s
trange in an Italian Enigma intercept, realized after a moment that the message had not a single l in it and, knowing that the Enigma never replaced a plaintext letter with the same letter, concluded that the message was a dummy whose plaintext consisted entirely of l’s. She used this to reconstruct the key. Lever was a member of the team whose solution of Italian naval Enigma messages led the British to their victory over the Italian fleet at Cape Matapan—a victory that made the British masters of the eastern Mediterranean. After the battle Rear Admiral Godfrey, the director of naval intelligence, rang Naval Section and asked to speak to Knox. When told he was at home, Godfrey said, “Tell Duly we have won a great victory in the Med and it is entirely due to him and his girls.”

  Few of the women at Bletchley reached such high levels. Most served in indispensable clerical positions as typists, filers, or decipherers of messages once the cryptanalysts had found the key. Many were subdebutantes or postdebutantes—daughters of earls who wanted to do their bit to win the war. Though the work was sometimes boring, they did not complain: many of them would never have had jobs in peacetime, and they loved working. They were extremely security-conscious and did not discuss what they were doing with chums outside B.P. One day the king visited. He saw a young woman who was a friend of his daughter, Princess Elizabeth, and asked her what she was doing there. “I can’t tell you, sir,” she replied. Like the others, she took her responsibilities very seriously indeed.

  Bletchley Park was by now working at peak energy, if not yet at peak strength. Its workers, both the old-timers and the newcomers, were billeted in pubs and cottages in a radius of a dozen miles or so around the Park. Harry Hinsley shared half a cottage with two other Oxbridge types; a farmer and his wife and a cleaning girl lived in the other half and kept them tidy. Alan Turing lived at the Crown Inn at Shenley Brook End, a hamlet 3 miles to the north. He bicycled to work each day, but many of the thousand-odd staffers were brought to their jobs by buses, which made three circuits a day to cover all shifts. The workers ate together in a mess hall. Bletchley was not military and hence rather democratic. Uniforms were worn rarely; rank was sought mainly for greater pay or recognition, but it seldom affected relations among individuals because they knew the war work would not determine their futures. At a time when beards were rare in the general population, they were not uncommon at B.P. The inmates recognized what an odd collection they were. As one worker observed, “To work there, you needn’t be mad, but it does help!”

  In good weather the Bletchleyites sat on the lawn, walked, or played tennis or rounders or croquet. During these games, the journalist Malcolm Muggeridge, temporarily assigned there, observed that the players assumed

  the quasi-serious manner dons affect when engaged in activities likely to be regarded as frivolous or insignificant in comparison with their weightier studies. Thus they would dispute some point about the game with the same fervour as they might the question of free will or determinism, or whether the world began with a big bang or a process of continuing creation. Shaking their heads ponderously, sucking air noisily into their noses between words [they would say], “I thought mine was the surer stroke,” or: “I can assert without fear of contradiction that my right foot was already …”

  In the evening, for those not working the night shift, the Bletchley Park Recreational Club offered rooms for coffee, newspapers, the radio, and quiet, as well as a library, a drama group, a musical society, bridge, chess, dances twice a week, badminton, and fencing. Groups, often quite talented, put on plays, revues, concerts, even operas. Romances flourished. Muggeridge wrote: “The females, too, were mostly donnish; either dons’ molls, with solemnly pretty faces, studiously amorous or amorously studious, according to their temperaments or the exigencies of the occasion; or themselves academics, grey, untidy, rough and hairy and spluttering.” Many women met men who became their husbands. Hilary Brett-Smith moved into the other half of the cottage Harry Hinsley was living in and came to feel that he needed a woman to buy him socks when they wore out. They married, and so did a great many others.

  Hut 8’s cryptanalytic section was divided into two groups. One group worked in an oblong room with three wooden tables, two chairs at each, and a rack for wide sheets of paper that were prepared in Banbury, a town just north of Oxford. The cryptanalysts called the procedure performed on intercepts in this room using the sheets “banburismus.” It attacked intercepts for which no cribs could be envisioned.

  The sheets were of heavy white paper about 10 inches from top to bottom. Printed on each were dozens of vertical A-to-Z alphabets side by side. The sheets varied in width from 2 to 5 feet, the narrower sheets being used for the shorter German intercepts. The young civilian women assigned to Hut 8, working from the “red forms,” or intercepts on pink paper, punched out, in the successive alphabets of a sheet, the successive letters of an intercept. They did this twice on a sheet for a single intercept, then repeated the procedure with other intercepts of the day. Next they placed a sheet on one of the tables, with its dark surface, and laid another sheet on top of it, aligning the first alphabets above one another. Here and there dark spots appeared where two punched holes, representing the same letter, coincided. They counted the number of spots, then slid the top sheet one place to the right and counted the number of spots that appeared then. The women continued this procedure, later shifting the top sheet to the left, until all the statistically valid positions had been tabulated.

  Few spots at a position represented only the coincidences that chance might produce. This was the case with most positions for most pairs of messages. Many spots, however, could mean that the two messages had been enciphered with the same rotors at the same positions. If the messages were long enough, and if several were found to have key overlaps, statistical analysis could sometimes show, on the basis of the number of spots, where this synchronization of the rotors had been ended by the middle rotor’s being turned one place by the rightmost rotor. Analysis could sometimes determine the ring position of this rotor. And since each of the rotors I through V had the notch in its alphabet ring in a distinctive place—at R in rotor I, at F in rotor II, and so on (Royal Flags Wave Kings Above)—counting from the starting ring position could show where the notch was and so identify the rotor. (This did not work, of course, for rotors VI, VII, and VIII, because the navy had notched all of those rotors at H and U.) Knowing which rotor was in the rightmost position cut down on the number of trials the bombes would have to make, thereby saving precious bombe time and so speeding solutions.

  The statistical determination of where the synchronization ended used a method developed by the eighteenth-century English mathematician Thomas Bayes. Turing took it a step further by inventing units in which the Bayesian calculation could be made. Basing his terminology on the cryptanalytic procedure, he defined a “ban” as a bit of evidence that would make a hypothesis ten times more likely than otherwise. He considered that a deciban, or a tenth of a ban, was about the smallest change in weight of evidence that is “directly perceptible to human intuition.” So “deciban” became another of the singular terms used at Bletchley. A later refinement into half decibans, or HDBs, led to the cryptanalysts’ talking to one another about “hudubs.” Another technique, for improving the scoring of long repeats, was sarcastically called “romsing,” for “resources of modern science.”

  Hugh Alexander was the champ at banburismus. He had a better feel than the others for the overlaps and the end of the synchronization and the figuring of the ring positions. But others sometimes had their lucky shots: once Good found a twenty-two-letter repeat, making the others rather jealous. Good liked the work. As he pondered letter chains, alphabets, probabilities, he was concentrating as if he were playing chess. The problems were hard enough to keep him absorbed but not so hard as to make him want to give up. Above all, the importance of the task motivated him.

  The banburismus room was mostly quiet. There was little talk. And the feeling in it was good. The members got along well
with one another; quarrels were rare. Workers often arrived ten to fifteen minutes early for their shifts so their colleagues would not have to work overtime. Those who were leaving would explain that they had found some chains of letters or some alphabets to enable those arriving to pick up the work, primarily the identification of ring positions. The cryptanalysts slid strips of paper with letter chains under alphabets to see if anything matched or looked good. In view of the Enigma’s reciprocal substitution, they would reject an alphabet or a chain that produced an R, say, against an R—a “crash.” They examined sheets of paper with the banburismus scores on them. Curiously enough, they rarely looked at the actual intercepts.

  Banburismus was not used in cases where cribs were available. This was the job of the other part of Hut 8. Assumed plaintexts were set against cryptograms that were thought to be their ciphertexts, chains of letters were derived to generate possible bombe starting positions, called menus, and the possible keys were tested to see whether they revealed plaintext. The technique called for linguistic skills: Wylie, who had had a classical education before becoming a mathematician, and Brett-Smith, the Anglo-Saxonist, worked in the crib room.

  By early 1941, most cribs came from “kisses”—identical messages transmitted in two cryptosystems, one of which G.C.&C.S. could read. Hut 4, which handled non-Enigma naval systems, was then reading the German navy’s Dockyard Cipher currently. Among the more important messages it carried were those dealing with new mines discovered in a channel that had previously been swept clear. This information had to be reported to all vessels using the channel. But since auxiliary support vessels, smaller patrol boats, and minesweepers carried Dockyard but not Enigma, while combat U-boats carried Enigma but not Dockyard, the same information was transmitted in two cryptosystems. Often the two plaintexts were identical, since the drafter of the message did not know of the danger of this situation, and the communicators either did not care about it or refused to change the wording. When one of the Dockyard cryptanalysts in Hut 4, such as Christopher Morris, saw a solution that looked as if it might serve as a crib for an Enigma message, he carried it next door to Hut 8, where the codebreakers there sought to fit it to a cryptogram. And not infrequently it worked.