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

by Kahn, David

  In the face of this situation, Poland, in February 1921, signed political and military agreements of mutual assistance with France. Poland gained the larger nation’s support; France used Poland to make Germany worry about a two-front war. Poland clung tightly to the agreements through the 1920s as Germany rejected all attempts to accept the postwar borders with Poland, raised tariffs against that country, and thundered out ceaseless propaganda against her.

  Self-preservation thus compelled Poland to keep her two dangerous neighbors under observation. One way was to read their messages. This work devolved upon the unit founded by Kowalewski and headed in the late 1920s by Franciszek Pokorny. It continued to solve Soviet cryptosystems, most of them hand ciphers or simple codes. And it cracked as well the German army field cipher, also a pencil-and-paper system. Then in 1926 the German cryptograms began to change.

  In the naval messages, the indicators—the groups of letters that told the receiver’s cipher clerk which keying variables were used—were different from the old ones. The frequency distribution of letters bore no resemblance to the older messages. Repeated groups of letters all but vanished. On July 15, 1928, Pokorny’s cryptanalysts noticed similar changes in many messages of the German army. Perhaps the two German armed forces had converted to machine cryptography.

  Pokorny assigned the analysis of these new messages to three German specialists: Captain Maksymilian Ciȩżki, who had had to serve in the German army in World War I, Lieutenant Wiktor Michalowski, and a civilian, a Mr. Czajsner. They made little progress. They observed that the indicators for the new army messages consisted of six letters and that in all cryptograms of a single day in which the first indicator letter was, say, R, all the fourth indicator letters were, say, M. The second and fifth and the third and sixth letters likewise seemed to be related. But it was not clear what this meant. The cryptanalysts confirmed, perhaps from spies, perhaps from radiomen’s chatter, that the cryptograms had been machine-enciphered. When they discovered that the machine was the Enigma, they purchased a commercial model. Experiment soon showed them that it could not decipher military messages. Beyond this, however, they were unable to learn anything. Statistical tests showed that there were not enough identically keyed messages for the standard superimposition form of solution to work. It appeared impossible to take even the first step in reconstructing the machine.

  This was the situation in January 1929. But Pokorny and Ciȩżki were beginning to understand that cryptology was changing. For centuries nearly all cryptosystems had been linguistically based: elements of language, such as words or syllables or phrases, were replaced with codewords or codenumbers. Cryptanalysts thus had to be linguistically oriented: Dillwyn Knox was an archetype. After World War I, however, cryptography began to be mechanized. Increasingly, armed forces adopted cipher machines. And their basis is literal, or letters, not linguistic: a cipher machine will divide the t from the h in the, for example. While codes are books, cipher machines are like typewriters. Breaking their ciphers calls mainly for mathematical or mathematical-like knowledge. The generation of keys, the production of cipher alphabets, and other elements are ascertained through logical analysis, sometimes without recovering a single word of the plaintext.

  To solve the Enigma, Pokorny and Ciȩżki sought mathematicians, particularly among students at the university at Poznán. Even though that city was not one of the centers of Polish mathematics, then perhaps the finest in the world, it was in the part of western Poland that had been German territory from 1793, the date of the second partition of Poland, to 1918; the Germans called the city Posen. Many mathematics students there had grown up in the area, had attended German schools, and knew German. They thus would possess the mathematical skills necessary and the linguistic skills possibly helpful for attacking the Enigma. An instructor assembled a group with those qualifications for Pokorny and Ciȩżki, who invited them to join a class in cryptology. Some twenty accepted. Pledged to secrecy, they attended a night course once a week at the university’s Mathematics Institute in the fake-medieval castle built by Kaiser Wilhelm II. The instructors—Ciȩżki, Pokorny, and a civilian cryptanalyst, engineer, and radio ham named Antoni Palluth—came from Warsaw, 200 miles away.

  Palluth lectured first, on the basics of cryptology. Then Ciȩżki spoke on the German army field cipher that Warsaw had solved. It was not the Enigma but a pencil-and-paper system called a double transposition. It mixed the letters of the plaintext message rather than replacing them with other letters. The cryptanalyst’s task was to unscramble them, to restore their original order. Ciȩżki assigned the students some actual intercepts to break. To help them, he told them that the messages dealt with winter quarters and bivouacs on training grounds.

  Within a few hours, three students had solved the cryptograms. Gradually, as the test cryptograms became harder, more and more students dropped out of the course. And then one of the three who had solved the double transposition, Marian Rejewski, left—but not for lack of ability or interest. He had received his degree in mathematics and wanted to pursue studies in actuarial mathematics at one of the world centers for mathematics, the university at Göttingen.

  Rejewski, a short, unprepossessing twenty-three-year-old, did not impress the other Polish mathematicians at Göttingen by his mind or his manner, by his looks or his personality. He had no close friends at the university, but he tagged along on the long walks that one of the Polish mathematicians, Henry Schaerf, liked to take. Rejewski’s political views, in particular his opinion that the Jews should be expelled from Poland, seemed derived from newspaper articles on the program of the National Democratic Party. But he was not so rigid as not to listen to contrary positions. Schaerf thought him relatively immature in his mathematical work and saw in him no extraordinary ability, no flashes of brilliance.

  For a year Rejewski studied applied mathematics, specializing in actuarial questions. He expected eventually to work in a relative’s insurance firm. But upon his return home for the 1930 summer vacation, he found a letter offering him a teaching assistantship at Poznán. He accepted it and, with the Depression rapidly making job prospects scarce, kept the position instead of returning to Göttingen. He wondered what had become of the cryptology course and soon learned that the other two students who had cracked the double transposition were now solving German cryptograms twelve hours a week in a basement office of the Poznán military command post in St. Martin’s Street. Rejewski told one of them that he wouldn’t mind working there as well. After an interview, he was hired. In mid-1931 the unit formally became an outpost of Warsaw’s Biuro Szyfrów, or BS, which had been expanded by putting together several intercept and codebreaking units.

  Ciȩżki, short, corpulent, jovial, now head of the bureau’s fourth, or German, desk—BS-4—had progressed no further in the solution of the Enigma than he had when Pokorny first presented the problem to him. In desperation, Ciȩżki called in a noted clairvoyant, but even his crystal ball could not reveal the mind of the machine. Ciȩżki’s long-range plan, however, seemed hopeful; his three part-time mathematician cryptanalysts were showing promise. He offered them full-time jobs as cryptanalysts in Warsaw, and on September 1, 1932, Rejewski and his two younger colleagues, Jerzy Różycki and Henryk Zygalski, who had only recently graduated from the university, began work in a wing of the Saxon Palace, the general staff building on Saxon Square.

  Ciȩżki did not think they were ready yet to attack the Enigma. He gave them instead as their first assignment the solution of a four-letter Reichsmarine code—a step up in difficulty from army hand ciphers but a big step away from the Enigma. The three began by making frequency counts of the code groups. Here Rejewski’s actuarial studies found application, since much cryptanalysis rests on statistics. The trio noticed that many codegroups began with Y. Perhaps these groups represented the series of interrogatory words that begin in German (as in English), with w, a letter that, like y, stands at the end of the alphabet: wer, was, wann (who, what, when) and so on.
One day, mulling this possibility, they noticed a short, six-group radiogram beginning YOPY that was answered with a four-group message. It appeared to be an exercise in which one operator had put a message into code and transmitted it to another, who had replied. Perhaps the first message was a question, the second its answer—probably, in view of its brevity, a year, with each codegroup representing a digit. The question would then be a query as to when something happened. Was it a battle? The birthdate of a famous man? The three quickly reasoned their way to the supposition that the six-letter message was Wann wurde Friedrich der Grosse geboren? (When was Frederick the Great born?) This guess proved correct and yielded as well the meanings of the four reply codegroups: 1, 7, 1, and 2. After making this first break, the three cryptanalysts merely expanded the solution. Their apprenticeship had ended.

  The achievements of this young and relatively inexperienced cryptanalytic bureau equaled, curiously enough, those of one of the world’s oldest. France’s military cryptanalysts, whose work dated back decades before World War I, had had a remarkable history of success in cracking German military ciphers during the war. Some of their solutions had aided generals at crucial moments during the fighting. But by 1928 France had reduced the number of her army cryptanalysts to eight, and their capabilities were limited. They did not have the techniques needed to solve rotor machines. They dealt only with simple systems: the German army double transposition that Rejewski and the others had solved in their cryptology class, some German codes, a British code. The cryptanalysts and their superiors seemed content with that. The anti-German revanche that had spurred France to the forefront of cryptology after the defeat of 1871 had evaporated in the victory of 1918; the need for intelligence from code-breaking had declined now that France had shackled Germany with the restrictions of Versailles and possessed an army widely regarded as the best in the world. The attitude was the cryptologic equivalent of the Maginot Line.

  But one French cryptologist, at least, was not content with his army’s inadequate results. Gustave Bertrand had enlisted as a private in 1914, was wounded the next year in the Dardanelles, and was assigned after the war to the cipher section of the staff of French forces in Constantinople.

  Cryptology attracted him. He served during the 1920s in the cipher sections of various headquarters and in 1929 was summoned to that of the army general staff. The poverty of the codebreakers’ results may have led him to conclude that the coming generation of cipher machines would be solved not by pure cryptanalysis but only with the help of bought or stolen keys or descriptions. He proposed establishing a unit to purchase such information from traitors or to burglarize offices and examine and photograph the needed papers or mechanisms. On October 30, 1930, Bertrand, by now a captain, instituted the new Section D, for Décryptement et Interceptions, of the Service de Renseignements (Intelligence Service). Section D and the separate cryptanalytic section were both part of the army general staff’s famous Deuxième Bureau (Second Bureau) for intelligence evaluation.

  For almost two years, Section D yielded nothing of great value. With the approval of his chiefs, Bertrand contacted the intelligence services of countries keeping an anxious eye on a restive Germany: Poland and Czechoslovakia, with both of which France had military alliances, and Great Britain. They exchanged intercepts and direction-finding results, but no cryptanalytic results. Then Bertrand got a break.

  In the summer of 1931, there arrived at the squalid offices of the Service de Renseignements, in a Ministry of War annex at 75 rue de l’Université on Paris’s Left Bank, a letter dated July 1 and mailed from Prague. It stated that the writer had contacted the intelligence representative at the French embassy in Berlin on June 8 and had offered to sell documents of the highest importance. If the French were interested, they should contact him at Kaufhausgasse 2 in Basel, Switzerland, by October 1. If he had not heard from them by that date, he would go elsewhere. He listed two documents that he could deliver: the instructions for the use of the German army Enigma cipher machine and the instructions for setting its keys. The letter was signed Hans-Thilo Schmidt.

  Paris contacted the Berlin embassy, and the Deuxième Bureau contact there, Maurice Dejean, confirmed that Schmidt had visited him. He added that Schmidt and his older brother, a lieutenant colonel, were both listed as being in the German Defense Ministry. This improved the chances that the approach was real and not a provocation.

  The Service de Renseignements assigned the task of making the first contacts with Schmidt to its man Friday, who handled all sorts of details for the service—he could get train reservations that Cook’s couldn’t—but who specialized in recruitment and in the puchase of secret codes.

  His codename was REX, he claimed to have been named von Koenig (“king” in his native German), his legal name was now Rodolphe Lemoine, but he had been born Rudolf Stallmann in Berlin on April 14, 1871. The son of a wealthy Berlin jeweler, he preferred travel—in France, Italy, England, Africa, Chile, and Argentina—to going into his father’s business. Somewhere along the way he met and married a Frenchwoman, whose name he adopted; in 1900, he was naturalized as a French citizen. During World War I, in Spain, he developed a taste for spying, and in 1920 he came to Paris to work full-time for France’s intelligence service. His pay came in the form of protection from the police in his shady dealings and of business concessions abroad that the French government awarded him. These activities he ran out of an office at 27 rue de Madrid in Paris’ fancy eighth arrondissement.

  Writing to the Basel address, REX arranged to meet Schmidt on November 1, 1931, at the Grand Hotel in Verviers, a town in eastern Belgium some 15 miles from the German border. There he learned much of the would-be spy’s history.

  Schmidt, then forty-four, had been born May 13, 1888, in Berlin, the second son of Professor Dr. Rudolf Schmidt and his wife, Johanna. The father, who was thirty-seven when Hans-Thilo was born, taught at the Charlotten school in Berlin. The first son, Rudolf, two years older than Hans-Thilo, had brought honor to the family when at twenty he was accepted into the army as a cadet. Hans-Thilo had had the standard college-preparatory classics education but then had studied business, with an emphasis on chemistry and technology. Both brothers had served in World War I. Rudolf distinguished himself in various signal corps posts, winning the Iron Cross, rising to captain, and ending the war in the general staff of the Fourth Army. Hans-Thilo, a lieutenant, likewise won the Iron Cross but had the bad luck to be gassed.

  Rudolf was retained in the 100,000-man army that the Versailles treaty allowed Germany. Hans-Thilo started a soap factory, but in Depression- and inflation-ridden postwar Germany, the business failed. By the mid-1920s, married and with two children, Hans-Thilo was desperate. He turned to his brother for help. Rudolf, who had no children and was close to his younger brother, prevailed upon a fellow signals officer to give Hans-Thilo a job. The officer was head of the Chiffrierstelle (Cipher Center), known as the ChiStelle, and he hired Hans-Thilo in part because Rudolf had been his predecessor. In fact, in one of the most exquisite ironies of intelligence history, Rudolf Schmidt, as head of the ChiStelle, had approved for army use the Enigma cipher machine that his brother was now proposing to betray! Hans-Thilo became a civilian clerk who distributed cipher material and supervised its destruction when it expired. He did his work well.

  But he was embittered and rapidly grew more so. Despite his war record, his job paid poorly. His life seemed to be going nowhere. He was living in a furnished room at Lorenzstrasse 17 in west central Berlin, having sent his family to less expensive Bavaria to live with his in-laws. The contrast with his past was striking. His family had had enough money and high status: his mother was a baroness, his father a Professor Doktor—probably the highest nongovernmental standing a civilian could achieve in Germany. And the contrast with his brother’s position was no better. Rudolf had not only been retained in the army, Germany’s most prestigious institution, but had been promoted to lieutenant colonel and, after two assignments in division
al staffs, had been named chief of staff of the signal corps. Like thousands of others similarly disaffected, Hans-Thilo applied for membership in the Nazi party. But this did not immediately ameliorate his situation. And though a desire for more money and a sense of relative failure must trouble many men who hold state secrets, most do not betray their country. Hans-Thilo, from whatever motives, did.

  REX learned much of this at Verviers. He sized up his man and examined the documents the spy brought with him. REX concluded that both the documents and the offer were genuine and directed Schmidt to return to the same place one week hence.

  The Service de Renseignements, now persuaded that Schmidt would really spy for them, gave him, as they gave their other agents, a designation consisting of a group of letters. His was HE, which seems not to have borne any particular significance. In French, this pair is spoken ahsh-AY. This somewhat resembles the German word Asche (ash), which is sounded AH-shuh and which Schmidt himself later chose—with a premonitory shiver about the fate of most spies?—as his codename.

  The rendezvous was set for November 8, 1931. The day before, REX, accompanied by the stout and sometimes difficult Gustave Bertrand and the Service de Renseignement’s photographer, Bintz, arrived at the Grand Hotel, where REX had installed himself in princely fashion. Bertrand and Bintz had adjoining rooms with a bathroom between them: if it served as the camera studio, the shutter noise would not be heard in the hall. They spent part of the night adjusting the camera and their portable lights. The next morning they waited in their room for the call from REX. It came at 10 o’clock.

  In a room filled with cigar smoke and with a radio playing music to deter eavesdropping, REX presented Schmidt to Bertrand. Schmidt, a whisky in his hand, smiled and bowed, heels together. Bertrand saw a man who wore a dark gray suit and down-at-the-heel shoes but who seemed to carry himself with assurance and to display the manners of a good upbringing. Bertrand was introduced as Monsieur Barsac. After some pleasantries, he examined the documents that Schmidt produced. They included an organization chart of the ChiStelle, an army hand cipher with its keys, a memorandum on poison gas, which Schmidt claimed to have obtained from his brother, and—most important—the two Enigma documents. The first was Gebrauchsanweisung für die Chiffriermaschine Enigma, the instructions for the use of the machine, numbered H.Dv.g.13 (for Heeresdienstvorschrift geheim 13, secret army service regulation 13). The other was the Schlüsselanleitung für die Chiffriermaschine Enigma, the directions for setting the keys on the machine, numbered H.Dv.g.14.