THE CODEBREAKERS

by DAVID KAHN

  Hebern refused to give up. Pinning his hopes on the Navy, he incorporated the International Code Machine Company in Reno, Nevada. Things started to look up in 1928 when he sold four five-rotor machines to the Navy at $750 for each machine and $20 for each rotor. Hebern and a handful of employees had built them by hand, and he himself then drove them to the 12th Naval District Office in San Francisco. One machine stayed there; the others were sent to the Navy Department and to the commanders in chief of the United States Fleet and the Battle Fleet for field tests. The Navy wanted to determine their mechanical reliability rather than their cryptographic capabilities, which were regarded as satisfactory, even though Friedman had made a crypt-analytic breakthrough and solved the first rotor system. During 1929 and 1930 these machines handled a considerable portion of the Navy’s official high-command communications. Things looked even better for Hebern in 1931, when the Navy purchased 31 machines for $54,480. These were not experimental machines, but were issued to the more important flag officers as the top cryptographic system in the United States Navy. In 1934, Hebern, who was continually trying to improve his machines, submitted one that proved a complete failure. The officer who had dealt most with him, Safford, was on sea duty, and some Navy man who did not know Hebern sent him an abrupt and discourteous letter, discontinuing business with him. As Safford later put it, “They pulled the rug out from under Hebern and were not even polite about it.”

  That virtually ended Hebern’s chances, for although his machines were still in service, when they wore out in 1936 after carrying heavy loads of traffic they were replaced by another, non-Hebern cryptographic system. Interestingly, the Hebern machines themselves were renovated and sent to shore stations, where some remained in use until 1942. Two were, in fact, captured by the Japanese during World War II.

  During this time, Hebern was living on income from properties left by his wife’s sister. He continued to improve his machines and to take out patents, despite the setback of losing a patent interference case against International Business Machines in 1941. In 1947, convinced that the armed forces had used his basic ideas throughout the war without compensating him for them, he filed a claim of $50,000,000 against the three services. In the six-year period that this remained entangled in bureaucratic red tape, Hebern died. He was 82, and had suffered a heart attack on February 10, 1952, while trying to lift a box that was too heavy for him.

  Early in 1953, the departments of the Army, Navy, and Air Force rejected his claims, and a few months later his estate sued the government for the $50,000,000. On the basis of legal technicalities, the United States Court of Claims limited the period of recovery to 1947-1953 and the infringement question to the exceedingly narrow one of a particular dog arrangement for turning the rotors. Ignored was the basic question of whether the armed forces had adopted the rotor principle from Hebern and used it without just compensation in hundreds of thousands of high-security machines in World War II and in the cold war—which they had unquestionably done. Ignored were the ethics of having obtained Hebern’s best developmental efforts on the implied promise of large production contracts, which were awarded instead to the Teletype Corporation.

  The government, taking refuge from the spirit of justice in the letter of the law, fought to keep from giving him a penny. In 1958, it finally settled for the pittance of $30,000—and not out of a sense of fair play, but because it feared that the court’s sense of right would compel it to bare some cryptographic secrets. The payment was disproportionate to Hebern’s contribution, which was worth, not $50,000,000, to be sure, but $1,000,000 at the least. Hebern deserved better. His story, tragic, unjust, and pathetic, does his country no honor.

  At 2:55 p.m., Tuesday, October 7, 1919, the man who viewed the rotor most comprehensively filed what was to become Netherlands patent No. 10,700 for a “Geheimschrijfmachine” (“secret writing machine”). Hugo Alexander Koch, then 49, a native of Delft, had apparently devised it as an outgrowth of his engineering hobbies. He foresaw some commercial value for the system, for he set up a corporation, the Naamlooze Venootschap Ingenieursbureau “Securitas,” in whose name the patent was issued. Koch pointed out in this patent that steel wires on pulleys, levers, rays of light, or air, water, or oil flowing through tubes could transmit the enciphering impulse as well as electricity did. He also observed that this impulse did not have to flow through a rotor, but could move through tubes drilled through bars that slid between plates, or from an interior disk to a circumjacent ring. He favored the rotor mechanism, but no machine ensued in any of the forms. In 1927, he assigned the patent rights to the German inventor of a rotor device, and the following year he died in Düsseldorf.

  The German was Arthur Scherbius. Little more is known of him than that he was an engineer, had a doctorate, held a number of patents involving such far-from-cryptologic materials as ceramics, and lived in Wilmersdorf, a suburb of Berlin. His first cryptologic device enciphered codenumbers into pronounceable codewords, which were then favored by international telegraph conventions. It did this by feeding the placode numbers alternately to vowel and consonant encicode equivalents. The device included “multiple switch boards which connect each arriving lead with one of the outgoing leads and which are adapted to interchange this connection with great facility of variation.” Though Scherbius did not describe this device further, it was the basis of a rotor system. Such appeared full-blown in his next patent. The rotors served only for numeral encipherment in this device, but in subsequent ones they expanded their contacts from 10 to 26 and so could be used for standard literal encipherment.

  He called his machine the Enigma. Model A, a monster about the size and shape of a cash register, was soon discarded for Model B, which stuck the enciphering mechanism on the right side of an ordinary typewriter. Model C was a portable, nonprinting device in which the letters were indicated—as in the early Hebern models—by lamps. All models had typewriterlike keyboards. The Enigma differed in two important ways from the other rotor conceptions. Its final rotor was a half-rotor: it had contacts on one face only and these were interconnected. An impulse coming to this rotor would thus be reflected back through the rotors through which it had just come. This doubly enciphered each letter, but it also made the encipherment reciprocal (if plaintext e became X, then plaintext x had to become ciphertext E), which is a weakness. The second difference was that the rotor progression was governed by gears to make it irregular. Unfortunately, the gears had so low a pitch that their period came to only 53,295 letters. Later machines improved this.

  Scherbius seems to have formed a little company of his own, Gewerkschaft Securitas, to promote his machine. Evidently some businessmen saw possibilities in the mechanism, for in July of 1923 a corporation was set up to manufacture and sell it. Chiffriermaschinen Aktiengesellschaft (“Cipher Machines Corporation”) was capitalized during Germany’s disastrous postwar inflation at 500,000,000 marks, distributed in 50,000 shares of 10,000 marks par value. It paid Gewerkschaft Securitas 300,000,000 marks for a controlling share of the company and its patents, models, drawings, and tools. Scherbius sat on the six-man board of directors.

  Chiffriermaschinen Aktiengesellschaft began operating on August 24, 1923, at 2 Steglitzerstrasse, Berlin, and worked hard to create a demand for its product. It exhibited the Enigma before the 1923 congress of the International Postal Union, and the following year got the German post office to exchange Enigma-enciphered greetings with the congress. It got some publicity in Radio News and extensive coverage in a book on cipher machines by Dr. Siegfried Türkel, scientific director of the Criminological Institute of the Vienna police. It printed flyers in German and illustrated brochures in English: “The natural inquisitiveness of competitors is at once checkmated by a machine which enables you to keep all your documents, or at least their important parts, entirely secret without occasioning any expenses worth mentioning. One secret, well protected, may pay the whole cost of the machine….”

  But nothing helped. A few machines
were bought for study purposes by the armed forces of various nations and communications companies, but mass sales never materialized. Production constantly declined. In 1924, the first full year of operation, Chiffriermaschinen Aktiengesellschaft spent 102,812 reichmarks (the new monetary unit) for expenses, salaries, and wages. In 1929 it disbursed only 56,345. By then Scherbius’name was no longer listed among the directors, most probably because he had died. After ten full years of operation, the firm had still failed to pay a dividend. So on July 5, 1934, it dissolved and transferred its assets to Chiffriermaschinen Gesellschaft Heimsoeth und Rinke, a new cipher machine firm organized by Dr. Rudolf Heimsoeth and Frau Elsbeth Rinke, both directors of the old firm.

  Soon Hitler began rearming Germany, and the cryptologic experts of the Wehrmacht, deciding that the Enigma offered satisfactory guarantees of security, began supplying their expanding forces with it. Whether Heimsoeth and Rinke enjoyed this new prosperity, or whether the Nazis nationalized their business or merged it into others, is unknown. During World War II, the portable glowlamp Enigma, battery-powered, and, in its wooden box, about the size and weight of a standard typewriter, served as the top German Army, Navy, Air Force system. Signal officers regarded it as very dependable and believed it to be secure. Its only disadvantage was that it did not print, and speedy operation required three men—one to read the incoming text and press the keys, one to call out the letters in a loud voice as they lit up, one to write down the text.

  Oddly enough, the fuzziest of the four original rotor conceptions reached a patent office only three days after the clearest. Koch had filed for his patent in Holland on a Tuesday, and on Friday of that same week in October, 1919, Arvid Gerhard Damm applied in Stockholm for what was to become Swedish patent No. 52,279.

  Damm’s device employed a kind of double rotor arrangement. Two circular flat plates, both wired like rotors, turned above and below a horizontal intermediate plate. Gears moved the rotors an irregular number of spaces at each plaintext letter. Damm, however, regarded the rotor as only an auxiliary feature of the enciphering mechanism, which was so extraordinarily clumsy and complicated that it seems never to have been built. And though Damm’s conception of the rotor accords him a listing in the honor roll of cryptologic inventors, his real impact on cryptology comes from his having established a cipher machine company that eventually became the only commercially successful one in the world.

  Damm is one of cryptology’s “characters.” He was originally a textile engineer. While engineering manager of a cloth-making factory in Finland, he became enamored of a Hungarian equestrienne in a traveling circus. Unable to overcome her virtue, he had a chum dress up like a clergyman and “marry” them in a fake ceremony in a chapel, thus achieving his goal. He had a flair for mechanics, and in his villa in Rönninge, a suburb of Stockholm, he had chairs whose armrests and footrests could be adjusted at the touch of a button, controls at his desk that would flick lights on and open doors, and other gimmicks to astonish guests.

  Arthur Scherbius’“Engima” U.S. Patent 1,657,411. Rotors 6, 7, 8, 9 are set to key NIAG. Figures 2 and 3 schematically show rotor connections

  He had several inventions relating to the Jacquard pattern-weaving loom to his credit when, just as World War I was breaking out, he and an English cloth manufacturer, George Lorimer Craig of Huddersfield, filed three applications for a cipher machine in the German patent office. Damm’s interest may have been awakened by his brother, Ivar, a mathematics teacher in the high school at Gävle, Sweden, who dabbled in cryptanalysis. Damm brought his machine to an acquaintance at the Swedish Embassy in Berlin, who urged a meeting with his brother, Commander Captain Olof Gyldén, who was commandant of the Royal Naval School in Stockholm and took an interest in all kinds of new ideas. In 1916, Gyldén and Damm were instrumental in founding Aktiebolaget Cryptograph (“Cryptograph, Inc.”). Among the investors were Emanuel Nobel, nephew of Alfred Nobel, inventor of dynamite and donor of the Nobel prizes, and K. W. Hagelin, manager of the Nobel brothers’oil production company in Russia, a close friend of Emanuel and at one time Swedish consul-general in St. Petersburg. In 1921, the firm had its three-room offices at 19 Karduansmakaregatan in Stockholm. It seemed to employ considerably more bosses than workers: excluding Damm himself, there were altogether a managing director, a technical director, a draftsman, and a bookkeeper.

  Damm designed quite a few machines. One included his ingenious “influence letter.” This was a plaintext: letter whose key on the keyboard was disconnected from the mechanism that advanced the cipher elements, though it itself was enciphered; this letter thus interrupted that advance at the wholly irregular intervals of its appearance in the plaintext. Another machine enciphered numerical codegroups into pronounceable codewords in which vowels and consonants alternated. The firm concentrated most on Damm’s Mecano-Cryptographer Model A 1, the “Cryptotyper,” an ugly apparatus that printed one copy of the plaintext and two of the ciphertext (one to be sent, one to be filed) on three tapes. Intercommunicable with it was the portable Model A 2, which displayed the ciphertext letter in an aperture. Their key consisted of a chain, assembled by the user, some of whose links moved the so-called key-body forward and some backward. The Electro-Crypto Model B 1, handsome but massive, was later installed at the main office of Sweden’s telegraph bureau.

  Meanwhile, Damm had fallen in love with a girl in her early twenties whom he had met on the commuter train. He decided to jettison his “wife” in a divorce proceeding that he thought would be no more valid than the marriage and, taking no chances, by a denunciation of her as a spy, to get her out of the country. He was severely embarrassed in court, however, when his partner Gyldén revealed the phony wedding as well as the spy gimmick. Damm rewarded Gyldén for this bit of candor by giving the managing directorship to someone else when it later became available. But he got what he wanted, and his new fiancée, Miss Spång, accompanied him to Paris on his business trips, living with him at the Hotel Périgord. Unfortunately, this saga ended sadly for him when she jilted him—doubly sadly, since he had given her his villa, though it was mortgaged, as the Swedes say, to the chimneys.

  Damm had won orders for a single test Model B 1 from several major wireless companies. He hoped they would adopt it to safeguard the world’s commercial wireless traffic. But the machines proved erratic. In tests in France in 1925, sometimes 1,000 letters could be deciphered correctly, and sometimes none at all. The following year the firm demonstrated its machines at The Hague for the Japanese military attaché. Difficulties abounded. Direct current was used where alternating current was needed; parts were too heavy to respond quickly; numerous malfunctions and cryptographic errors occurred. Sales were not spectacular.

  By then, however, a new personality was asserting himself in the company. This was Boris Caesar Wilhelm Hagelin, son of the consul-general and investor in the firm. Born on July 2, 1892, in the Caucasus, where his father was working, he studied for three or four years in St. Petersburg, then returned to Sweden and was graduated from the Royal Institute of Technology in Stockholm in 1914 with a degree in mechanical engineering. After six years of working for ASEA, Sweden’s General Electric, and one in the United States for the Standard Oil Company (New Jersey), in the expectation of returning to Russia for the Nobel interests, he and the Nobels realized that the Communist regime was not going to fall as they had hoped. Accordingly, his father and Emanuel Nobel put him into the Damm firm in 1922 to represent their investment.

  Three years later, while Damm was in Paris, young Hagelin learned that the Swedish military was considering buying the Enigma. He simplified one of the Damm mechanisms, giving it a keyboard and indicating lamps like the Enigma’s, and making it more suitable for field use. Its operation was based on the checkerboard. It electrically altered the row and column assignments to convert a plaintext letter to ciphertext. These alterations were controlled by a group of keywheels, each with pins near its rim that could be made active or inactive by extending or retract
ing them. Each wheel had a different number of pins. The machine produced a polyalphabetic substitution whose period was the product of all the pin numbers. Hagelin offered this machine, the B-21, to the Swedish Army. Damm criticized it but the Army liked it, and, in 1926, placed a large order.

  On the verge of success, Damm, early in 1927, died. Aktiebolaget Cryptograph, which was in poor financial shape but which had a big order in its pocket, was purchased at a good price by the Hagelin interests and reorganized as Aktiebolaget Cryptoteknik, 14 Luntmakaregatan, Stockholm. Boris Hagelin ran the firm. He saw that printing cipher machines were faster, more accurate, and more economical in terms of manpower than indicating mechanisms like the Enigma. He first hooked up the Type B-21 to an electric typewriter and found it unacceptably bulky. So he merged the printing mechanism with the cryptographic in a single unit, producing the Type B-211. It weighed 37 pounds, operated at 200 characters a minute, and could be carried inside a case about the size of an attaché case.

  This was the most compact printing cipher machine available in 1934, when the French general staff asked Hagelin for the impossible: a pocket-sized cipher machine that would print the ciphertext and so permit one-man operation. He first whittled a piece of wood that would fit into a pocket to mark the limits of his dimension. While trying to concoct a mechanism that would fit inside such space and also produce an effective cipher, he bethought himself one day of a construction that he had conceived three years before for the inventors of a vending machine. It was an adding device that would accept different amounts of money, and it consisted of bars arranged in a cylindrical cage with lugs projecting from them in rows. There were 10 lugs in one row, 8 in the second, 4 in the third, 2 in the next, and 1 in the last; by combining these rows in various ways any number from 1 to 25 could be produced. This was just what he needed. The inventors had given him the rights to it when they could not pay for the prototype that he fabricated. He now adapted it so that the rows would shift a cipher alphabet to any one of 25 positions, thus giving a plaintext letter any one of 25 ciphertext equivalents. And to produce the combinations of numbers for these shifts, he could employ the key wheels with the variable number of projecting pins that he had used in his B-21.