Transcription
The story of Bletchley Park breaking Enigma has been told in many books and films. In Berlin during those same years, the mathematicians of the Cipher Department of the German High Command, known as OKW/Chi, were trying to do the reverse. For three years, under the direction of Doctor Erich Hüttenhain, they attacked the strategic cable traffic between General George Marshall in Washington and General Dwight Eisenhower in London, and they never found a way in.
The campaign ended in September 1944 with a thirteen-word entry in a German war diary, and Allied interrogators from the Target Intelligence Committee recovered that diary after the war. Translated from the German, the entry read: "U.S. five-letter traffic. Work discontinued as unprofitable at this time."
The agency that filed that entry had been in operation under various names since the previous war. The Reichswehr Cipher Bureau founded in 1919 became the Cipher Department of the German High Command, or OKW/Chi, after a Wehrmacht reorganization in 1939. It occupied offices at the Tirpitzufer in central Berlin, sharing the building with the Abwehr military intelligence service until Allied bombing in November 1943 forced it to disperse. At its peak the cryptanalysis core grew to between three and four hundred mathematicians and linguists, according to TICOM order-of-battle estimates. The director was Wilhelm Fenner, who had served as an intelligence officer with the Tenth Army in the First World War and had run the German Cipher Bureau continuously since 1923. The section that handled American and British traffic was the Anglo-American Section, and the man who ran it was Oberregierungsrat Nikolai Rohen. While Rohen ran the intake and reporting, the analytical work on the American five-letter traffic was the responsibility of a separate unit called Group IV, and the head of Group IV was Erich Hüttenhain. He was born in Siegen in 1905, took his doctorate in mathematics at the University of Münster in 1934, and joined OKW/Chi in 1937 after passing an entrance test that Fenner had personally designed. From 1939 onward Hüttenhain ran Group IV, the analytical cryptanalysis section, and his first major assignment came in 1940 at Frankfurt, where his team contributed to the breaking of the French Army cipher system shortly before the fall of France.
The American cipher that Hüttenhain's group would never break was officially called SIGABA in the United States Army and ECM Mark II in the United States Navy. William F. Friedman and Frank B. Rowlett had designed it in the late 1930s. The Germans, after failing to break it, gave it their own name: the American Big Machine. Teletype Corporation and the National Cash Register Company together produced roughly eleven thousand machines during the war. Each one weighed about ninety-five pounds and ran on standard electric current. Inside the steel casing were fifteen small wheels, called rotors, that scrambled the letters of a message before it went out over the wire. The critical difference between SIGABA and Enigma lay in how those wheels turned. Enigma's wheels advanced in a regular, predictable pattern, and with enough intercepted messages the cryptanalysts at Bletchley Park could work that pattern out and back-solve to the original German text. SIGABA's wheels were driven by a second set of wheels in a way that no German analyst could reduce to any predictable pattern, and the number of possible settings ran into the trillions of trillions, far beyond anything OKW/Chi could test by hand or by punch-card machine.
Because of its weight and its dependence on electric power, SIGABA could not travel with troops in the field, and the American military therefore divided its cipher work into different levels, with different machines for each. SIGABA itself stayed at the top, handling strategic messages between the Joint Chiefs of Staff in Washington, theater commanders abroad, and Allied governments. Below that, American diplomats overseas used a simpler system, the M-138-A, made of paper cipher strips that needed no power and fit in a briefcase. For use in the field, the Army issued a small hand-cranked machine called the M-209 Hagelin, made under contract by the Smith-Corona Typewriter Company. The Hagelin became the workhorse of American tactical signals, and Smith-Corona produced about one hundred and forty thousand of them during the war, almost thirteen times the SIGABA production. Battalion and company-level traffic used simpler tools still: code books and a printed-card system called Slidex. The strip cipher could fit in a briefcase, but SIGABA depended on its generator and stayed at theater headquarters.
At the strategic tier, where SIGABA stayed, a great deal of American traffic by the summer of 1942 was running between two desks. George Marshall served as Army Chief of Staff in Washington, and Dwight Eisenhower took up his post at Grosvenor Square in London on June 24 of that year as Commanding General of the European Theater of Operations. From that point onward the cables between the two men were operational. Operation Torch in November of 1942 was planned through this channel, and Marshall confirmed Eisenhower's appointment as Allied Commander-in-Chief for Torch over the wire in August. Through 1943 the same channel coordinated the Sicily landings and the Italian campaign, and in December Marshall conveyed to Eisenhower in Tunis the decision that he would command Operation Overlord. Every one of those cables was encrypted at SIGABA terminals in the United States and in Britain, and every one of them was the kind of high-value strategic traffic that OKW/Chi was assigned to break.
The German attempt to read that traffic had begun a year before the Marshall-Eisenhower channel was even operational. In June of 1941, OKW/Chi began intercepting American diplomatic and military cable traffic that arrived in groups of five letters, and the Cryptanalytic Machinery Section, designated IVb, took the new traffic as a fresh project. Two analysts received the initial assignment: Hans-Peter Jensen and a colleague identified in the TICOM interrogations only as Boehm. Jensen and Boehm applied the techniques that had worked on Enigma and on French and Polish ciphers: frequency analysis of letter distributions, and probable-word cribs built from common diplomatic phrases. By July of 1941 the internal report was finished, and according to the postwar interrogation of Hans Peter Weber the conclusion was simply that the traffic could not be read. The report blamed external factors. Jensen suspected the intercepts themselves were too few in number, or that American operators had introduced randomization at the keyboard. No one in the July file suggested the traffic might be coming from a rotor machine of a different design from Enigma.
In January of 1942, six months after the failure of the Jensen and Boehm effort, the Japanese naval attaché in Berlin sent a cable to Tokyo summarizing the state of joint Japanese-German cryptographic cooperation. The cable went out on a Japanese naval attaché cipher designated JN-A-20, which the United States Navy code-breaking organization Op-20-G had already broken. The cable's American readers therefore received its contents in plain language within hours of its transmission. The Japanese attaché reported that German efforts were "highly satisfactory" and that the Germans had recently "experienced some success on English Navy systems," but the cable continued in less reassuring terms. The Germans, the attaché wrote, were "encountering difficulty in establishing successful techniques of attack on enemy code setup," and Japanese efforts in parallel had "made no real progress against the American cipher system."
One year after OKW/Chi had begun the attack on American five-letter traffic, the German admission of failure went directly to American cryptanalysts at Op-20-G. While the high-level attack failed, the lower tiers began to give way. In early 1942 the cipher branch of the German Foreign Office, called Pers Z S, broke into the M-138-A strip cipher used by the State Department, and the way in came through operator errors rather than through any structural weakness. Hüttenhain recorded in his postwar TICOM interrogation that on one occasion his superior Kempf ordered him to break a strip cipher before the coming Sunday, and his team succeeded through a combination of long hours and luck. The episode created an internal problem because other officers began to expect that cryptanalysis could be commissioned on schedule, like industrial production.
In the autumn of 1942, during the fighting in North Africa, Hüttenhain's group also broke into the M-209 Hagelin. American battalions sent repetitive tactical messages such as contact reports and supply requests, and the German analysts used these as cribs in a known-plaintext attack against several keylists. The lower-tier successes continued the following summer at a larger scale. The Sicilian landing on July 9, 1943, codenamed Operation Husky, produced a German cryptographic windfall. In the confusion of the Allied beach push, an American unit on the southern coast left behind a complete M-209 cipher set including the current keylists. German forward signal units recovered the documents and delivered them to OKW/Chi. Between July and September of 1943 the Germans read one hundred and twenty-six M-209 messages and issued six intelligence reports through the VN system, and the reports went to KONA 7, the signal intelligence command at the Italian front.
The American discovery did not come through signal intercept but through pattern analysis: American analysts noticed that German tactical decisions in Italy were anticipating American moves that had been communicated only on M-209 networks. The American response was a complete keylist change across all M-209 networks rather than a withdrawal of the machines, and the new keylists shut the Germans out within weeks. Throughout this same period the parallel German attack on the higher five-letter traffic had produced no break-in.
Within weeks of the M-209 keylist change, a different development in Italy opened an indirect channel into Allied cryptographic intelligence. The Italian armistice on September 8, 1943, and the German seizure of central and northern Italy under Operation Achse, opened that channel from a direction the Germans had not anticipated. German troops captured Italian cipher offices before staff could destroy the materials, and among the captured papers were what the Italians had been reading: decrypted Allied cables from systems the Italians had broken before the armistice. Some of these Italian decrypts contained internal Allied discussions of the security of American cipher machines. In his postwar interrogations under TICOM I-2 and the supplementary session designated I-84, Hüttenhain acknowledged that after he reviewed the Italian materials he never had another major success against any high-level American system.
Beginning in late 1943, OKW/Chi started reducing the resources assigned to the American five-letter project. The reduction in resources continued through the next eight months. By June 6 of 1944, the Allied landings in Normandy had begun, and through July and August the Allied armies broke out and drove east across France. In the first week of September, American units reached the German frontier near Aachen. Inside OKW/Chi, somewhere among the dispersed offices that had replaced the bombed-out Tirpitzufer building, a clerk recorded the closure of the file in a single line of office shorthand in the German Signal Intelligence Group war diary, and no extended report and no internal review accompanied it. Resources that had supported the American attack moved over to Soviet and British targets. The Anglo-American Section under Nikolai Rohen continued at reduced scale until May of 1945, but the attack on the American strategic cipher had ended. The closure coincided with the week Hitler issued the preparatory orders for Operation Wacht am Rhein, the offensive that would open in the Ardennes in December and be remembered in English as the Battle of the Bulge.
After September of 1944 the SIGABA traffic between Washington and Eisenhower's headquarters at Versailles did not slow. Marshall and Eisenhower exchanged cables daily on supply, command, and strategy, and the volume rose sharply when the German Ardennes offensive opened on December 16, 1944. In the cables Eisenhower sent to Marshall that week, he described his decisions as he made them: committing the SHAEF reserve and ordering the 101st Airborne Division into Bastogne, dividing the battlefront along the line of the German penetration, placing Field Marshal Montgomery in temporary command of all Allied forces north of the breakthrough, and turning Patton's Third Army out of the Saar to drive ninety degrees north on Bastogne. The correspondence was later published in The Papers of Dwight David Eisenhower under Alfred D. Chandler's editorship, and the German cipher service could not read any of it. The offensive that depended on surprise was being met by command decisions in cables that OKW/Chi had stopped trying to penetrate three months earlier.
On the night of February 3 to 4 of 1945, in the Alsatian town of Colmar five months after the Berlin clerk had closed the file, a truck belonging to the United States 28th Infantry Division disappeared with a SIGABA machine inside it. Eisenhower personally ordered the search, and Counter Intelligence Corps officers and military police worked with local French forces for weeks looking for the truck and its contents. They eventually found the safe containing the SIGABA in a river bed near the town, and German forces had not recovered it. The machine returned to the United States for inspection. Three months after the Colmar safe came out of the river bed, the war in Europe ended, and the principal figures of this story went their separate ways.
Erich Hüttenhain was captured by Allied forces in May of 1945 and underwent interrogation at Flensburg and afterward at the British CSDIC facility, where the questions for the supplementary session TICOM I-84 were written personally by William Friedman. Hüttenhain worked as a cryptographic consultant for the Gehlen Organization after the war and died at Brühl on December 1, 1990, at the age of eighty-five. Wilhelm Fenner also joined the Gehlen Organization and died in 1961. William Friedman served as a senior cryptologic advisor at the National Security Agency from its founding in 1952 and died in 1969. George Marshall became Secretary of State from 1947 to 1949, received the Nobel Peace Prize in 1953, and died in 1959. Dwight Eisenhower served two terms as the thirty-fourth President of the United States and died in 1969.
The machine itself remained on the classified list for more than half a century. SIGABA was declassified by the National Security Agency in 2001, and the Friedman patent for the design appeared in public form on January 16 of that year, seventy-one years after he had filed the original application. One of the machines now stands at the National Cryptologic Museum at Fort Meade in Maryland. A second ECM Mark II is preserved aboard the submarine USS Pampanito at Fisherman's Wharf in San Francisco, where the museum also displays a captured German Enigma in the same exhibit area. Many thanks for watching. Please subscribe and share.