by Kahn, David
The Enigma remained the navy’s basic and most widespread cryptosystem. And the navy proved itself not stiff-necked in continually evaluating it. In particular, it accepted the uncomfortable conclusions of a study by Lieutenant Henno Lucan, second radio officer of the battleship Elsass, that in neither physical nor cryptologic security did the Enigma meet modern requirements. At about the same time, the army proposed that the navy adopt the army version of the Enigma, with its twenty-six-contact rotors and with a plugboard, which the navy’s lacked. The army gave two reasons: the plugboard had greatly improved security, and a single machine would make the services cryptographically compatible.
In February 1930, the chief of the Naval Command requested that the B-Dienst look into the army’s proposal. The B-Dienst, successor to the codebreaking B- und E-Dienst of World War I, replied that the main questions to be answered were, did the army machine meet navy requirements and did the use of the same model in larger numbers imperil the navy’s cryptographic security?
On June 21, the B-Dienst offered a positive judgment of the army machine, chiefly because “it offers considerably greater security.” The improvement stemmed from the plugboard, which, the B-Dienst said, raised the number of possible enciphering circuits by 2 to 3 billion. This greatly outweighed the loss of 5,213 starting positions that the army machine’s smaller rotor would entail. Several years passed before the navy’s bureaucracy accepted this argument and before enough of the new machines and rotors were produced and distributed throughout the navy. Finally, in August 1934, the navy began using the army Enigma machine, with its plugboard and typewriter keyboard. But it sought an extra measure of security: each machine had a set, not of three rotors, like the army model, nor of five, like the earlier naval machine, but of seven. The instructions for the new machine—the Funkschlüssel M (Radio Cipher M, the M perhaps for Marine)—required that rotors I, II, and III serve for talking with the army, that IV and V stay in reserve, and that VI and VII be used when the navy wanted to send messages to itself. The navy improved the system early in 1939. It recalled rotors VI and VII of all its Enigmas and cut a second notch in the alphabet rings. Notches now stood next to H and U. Each notch caused the rotor to the left to move one space when the notch reached a certain point in its rotor’s revolution. Rotors I to V and later the added rotor VIII each did this once in a revolution. When rotor VI or VII was used in the machine, its extra notch caused the rotor to the left to step twice in each revolution of VI or VII. Though this shortened the period, it also reduced the probability of a successful superimposition.
A few months later, the navy divided its cipher communications into two nets, one using a home waters key, one a foreign waters key. This lowered the volume of messages in each key and so lessened the chances for solution. In addition, the navy continued its three grades of messages. The general keys were widely held; enlisted men could handle them. Officer-grade messages used the same inner settings of the Enigma as the general grade but different plugboard settings, and if officer-grade messages were to be radioed, they were reenciphered with the general key. Staff-grade messages had their own inner settings and plugboard arrangement.
Paralleling these improvements in cryptographic security were those in physical and personnel security. On January 24, 1930, the Naval Command notified its four major units that “a surveillance of the machine and the rotor box [which held the rotors not in the machine], more comprehensive than before, especially on smaller vessels, is planned…. A sharp supervision of the personnel who have access to the machine is necessary.” The Naval Command noted that it intended (as Lieutenant Lucan had proposed) to secure the machine with a lock instead of just a lead seal.
Subordinate units reported, as ordered, on the measures they were taking. The commander of the battleship Hessen said the machine was “housed in a specially prepared, lockable compartment of the cipher desk in the radio shack…. The rotor boxes are kept under a secret lock in the office of the radio officer.” Captain Wilhelm Canaris of the battleship Schlesien urged keeping the lead seal. “If an unauthorized person succeeded despite all security measures in obtaining the key to this [proposed new] lock,” Canaris wrote, “it would in that case be difficult, often impossible, to detect this unauthorized intervention.” The commander of the Baltic Naval Station reiterated the regulation that only officers should change and set the rotors.
Personnel security was reinforced by the Defense Ministry’s memoranda on espionage. Officers used these to instruct enlisted men on the need for secrecy and the dangers from foreign intelligence agencies. The memo of October 15, 1934, included a monitory tale about the Enigma.
During a change of position in a field exercise, a noncommissioned officer and a private of a communications battalion left, through negligence, the cipher machine and a hand cipher in the field. The loss was noticed by them upon arrival in their new position, but the cryptographic material was not found in the place where it had been lost. A civilian had taken it and delivered it to the mayor of a neighboring village within half an hour.
Both soldiers were sentenced to confinement in a fortress for negligent betrayal of military secrets, thereby endangering the security of the Reich.
The same memorandum included a frightening summation of persons sentenced in Nazi Germany for treason or betrayal of military secrets: 148 in all of 1933, 155 for just the first seven months of 1934.
One of those people may have been Radioman Second Class Egon Bress of the Fourth Torpedo Boat Half-Flotilla in Wilhelmshaven. His careless lifestyle raised suspicions, and an investigation showed that, though his correspondence consisted largely of erotic exchanges with young women, he had taken several hundred photographs of cryptographic documents—the Allgemeines Funkspruchbuch (AFB) code, several superencipherments, and a keying procedure for the Enigma, among others. The photos, which he had had developed at a drugstore, were found in the home of a barmaid. Bress was arrested on February 16, 1934.
At the same time that it was watching over these obvious aspects of security, the navy was also seeking to prevent cryptanalytic losses. It recognized that using the cryptosystems correctly was as important as having good systems—perhaps even more important. Navy cryptologists may not have remembered the time in 1916 when the British transmitted virtually the same message in three different codes, enabling the German codebreakers at Neumünster to expand its solution of all of them. But the lesson had sunk in: improper usage could jeopardize even the best system. So cryptologists monitored the navy’s cryptographic operations. In some cases they merely warned of poor practices. After a radio exercise in 1932, for example, a reviewer cautioned that messages in a particular system “should not have time groups ending in 0 or 5” and that “the multiple representations [of geographical positions] were not sufficiently utilized.” In other cases violators of the rules were punished.
On the evening of January 8, 1932, Radioman Kunert of the Baltic Naval Station at Kiel was encrypting a message. He first encoded it in a codebook, then enciphered the resultant groups with the Enigma. The cipher form of the time had three columns: one for the plaintext words, one for their four-letter codegroups, and one for the codegroups’ four-letter Enigma encipherments. Regulations called for the twenty-line column of unenciphered codewords to be crossed out so that they would not be transmitted. But Kunert erroneously ran his pencil instead through the column of enciphered codewords. Around 10:30, he began to transmit his message, consisting of the unenciphered codewords, to the T-151, the command vessel of the Second Torpedo Boat Half-Flotilla, then at sea. When Kunert reached the twenty-ninth group, he spotted his mistake and stopped transmitting. He reported the situation to Radioman First Class Schmaland aboard the T-151.
On the assumption that the urgent message should suffer no delay, Schmaland directed Kunert to retransmit the message, giving the correct groups. This would give anyone listening both the unenciphered and the enciphered forms of an Enigma message. The monitors raged. Through this “ent
irely incomprehensible order” arose a radio violation “of the grossest sort. A better handhold for tackling the Funkschlüssel C [the 1926 naval Enigma] cannot be given to foreign cryptanalytic services. Thirty-six positions of the letter sequence were given to foreign cryptanalysis, which under certain circumstances could lead to further solutions.” This was rather exaggerated, as the cryptologists half admitted later on. To avoid a repetition of this blunder, they proposed that “all radio personnel should be forcefully instructed … about such errors and how their consequences can endanger the security of national defense.” Then, taking no chances, the navy changed keys.
In addition to moving against theft, betrayal, and accidental compromise, the navy sought to nullify the possibility of seizure of cryptographic material by enemy action. It did not fear capture of an Enigma alone—indeed, one memorandum stated specifically that “it is assumed in judging the security of the cryptosystem that the enemy has at his disposition the machine including all rotors”—because it believed that the great number of keys would not allow the correct key to be recovered in time to be useful. But it worried that seizure at the same time of the list of current machine settings and the booklet of indicators for the message keys would expose its messages to alien eyes for at least the duration of validity of the list and the booklet.
So the navy began to print its cryptographic documents in water-soluble ink. The ink was red; the paper, pink, was like blotting paper, intended to soak up water even if pressed tightly between book covers on a shelf. Of course, on a ship the danger existed that spray, rain, condensation, or spilled liquids would make the ink run and the document illegible. So the navy issued two copies of each document and urged that each one be enclosed in a protective envelope. The envelope, however, was not to be airtight but was to allow water to come in when submerged.
Finally, the navy provided a fallback system in case the enemy captured all four elements needed for a current reading of German cryptograms: a machine, a list of current settings, the booklet that listed the prescribed rotor starting positions and their indicators, and the bigram tables for enciphering the indicators. The method, to be memorized or written in an innocuous form, changed the rotor order and the ring positions, thus invalidating the captured documents. In the system’s first implementation, the cipher clerk would add 3 to the key list’s rotor number, so that when the key list said the rotor II was to go into the left-hand position, rotor V would be inserted instead. And the clerk would set the alphabet rings on the successive rotors respectively 4, 5, and 6 places farther down than the key list called for. Thus the key list’s ring positions KYD would be set at ODJ. A cue word would institute these changes. The first was ALDEBARAN, the name of a star. The clerk was allowed to set this down in his notebook as “Aldebaran, R 3, L 456,” in which R stood for “rotor” and L for “letter.” All these measures complemented the basic complex keying system (see Appendix).
Thus, by the start of the war, the Kriegsmarine had in place an extremely well thought out cipher system. It sought to preclude the dangers of operators’ stupidity and laziness, the capture of some documents, betrayal, and superimposition. It was content with its cryptographic systems. When the Monitoring Center that it had established at the start of 1939 to watch over its radio traffic submitted a report stating that communications should be held down as much as possible “because our cipher systems are to be viewed as not 100 percent secure,” the Naval War Command agreed with the recommendation to limit communications but said that such conclusions as “not 100 percent secure” should not be drawn.
The excellence of the navy’s cipher system played an important role in the ideas of one officer who had commanded a U-boat in 1918. He had adapted some World War I ideas on U-boat tactics that he thought would enable German submarines to choke off Britain’s imports in the event of a new war. The attacks of isolated submarines on convoys had not succeeded in doing this in World War I; like others before him, the officer concluded that group attacks would be necessary. As he worked it out, the first U-boat to spot a convoy would radio its position and then maintain contact with the target. U-boat headquarters would centralize information about its own submarines and enemy vessels and instruct other U-boats to hasten to the convoy’s expected position and join in a combined assault. These concerted tactics alone promised success, the officer believed.
He knew that this tactic would compel the U-boats to break a vital rule of the invisible arm: radio silence. But he felt that the value of coordination among U-boats outweighed the dangers of radio direction-finding by the enemy, and he further expected that the enemy’s inability to solve either the coded sighting reports or his coded instructions to the U-boats would deprive the foe of the advance information needed to thwart the subs. The secrecy of these messages would result from the unbreakability of the German naval cipher machine. On the strength of the Enigma, then, turned the success of the wolfpack strategy of that officer, Commander Karl Dönitz.
4
THE CODEBREAKER AND THE SPY
THE BLACK ART OF CODEBREAKING GLEAMED GOLDEN IN THE 1920s and 1930s. Bespectacled men in guarded offices in Paris, London, Washington, Berlin—in every capital—read the intercepted secret messages of other powders. Hunched over their desks, they puzzled over the mumbo jumbo of five-letter codegroups in cablegrams, hesitantly wrote words in languages not their own in colored pencil between the lines of gibberish, erased them, tried others. They filled in the blanks of a partial solution with silent jubilation and had the completed text typed up for distribution to policy makers.
Everybody seemed to be reading everybody else’s codes. France cracked the German and English codes. Germany read French, English, and Polish codes. Britain, whose postwar codebreaking agency was descended from Room 40, solved Soviet, French, and American codes. The United States broke the Japanese code; Italy, the British—among others in all of these cases. Usually, these were the minor codes of major powers or the major codes of minor powers. Less often were the major codes of the major powers broken—always with the exception of those of the United States, whose cryptograms were as transparent as a fish tank for any competent cryptanalyst.
Many of these countries had established cryptanalytic agencies to provide in peace the benefactions discovered in war. Yet one nation that had not learned codebreaking in World War I gained experience in its postwar struggle for existence, created an instant tradition of excellence, and advanced to an achievement that none of the great powers ever matched. This was Poland.
Shortly after her restoration as a state in November 1918, Poland had first resisted Soviet advances into what had been Polish territory before the first partition of the country in 1772, then had herself greedily moved into Soviet lands. In the seesaw struggle that raged from Kiev to Warsaw, codebreakers seemed to emerge in the Polish army fully formed. Their chief was Lieutenant Jan Kowalewski, a former chemical engineer who had always scored high on his mathematics examinations. At twenty-seven, he was a bear of a man whose good looks and sense of humor attracted many women. In 1919, when Kowalewski was on the army general staff, a friend asked him to take over the friend’s duties for a fortnight while he went on leave to get married. The friend’s work dealt with translating and evaluating intercepted telegrams. Kowalewski, who knew many European languages, had no difficulty with this. And when one day he found a Russian intercept in code on his desk, he determined to solve it. Within a few days he had done so. The message revealed information about the movement of some White Russian forces fighting the Reds in the Russian civil war—information of great interest to the Poles.
Kowalewski was directed to organize a radio intelligence service. Reluctantly, for he did not want to work in intelligence, he did so. He recruited mathematicians and brought in army officers to form a unit of a dozen men. He himself became passionate about the work. One day he ran about headquarters asking everybody for the words to a Russian folk song: it was evidently the key to a cipher. His unit very rapidly
achieved remarkable successes. Many Russian cryptograms were solved the very day they were intercepted; most of the others were solved the next day.
In August 1920, with the Soviets at the gates of Warsaw, more than 400 solutions revealed to the Poles the enemy’s organization, strength, locations, and plans. On August 20, for example, Kowalewski’s unit read an operational order of Soviet General Mikhail N. Tukhachevsky, setting out the assignments of all his armies. Such intercepts were sent to the Polish commander, General Jozef Pilsudski. He used them to stop the Russians outside of the Polish capital and then to drive them back, preserving Poland’s freedom and ending the Communists’ dream of marching to Berlin to start turning Europe Red. Kowalewski received the high Polish order of Virtuti Militari. And though he soon left cryptology, the unit he had founded, the Biuro Szyfrów, or Cipher Bureau, remained in the Second Department (Intelligence) of the army general staff.
Among the bureau’s functions was helping Poland move with surer knowledge in her difficult international situation. Punching the Soviet Union in the nose and settling the eastern boundary did not end the country’s difficulties with her huge neighbor.
But the clamor from the west was louder. Germany, though defeated in World War I, claimed parts of Poland that had been German since the partitions. She was enraged by the Polish Corridor, which divided East Prussia from the rest of the nation and caused unending difficulties in trade, transportation, and communications. Some German leaders, inspired in part by historical precedent, urged the obliteration of the new Polish nation. Since the days of the Teutonic knights seven centuries earlier, Germans had pressed for a Drang nach Osten. And though Germany’s army was only a third the size of Poland’s, Germany believed herself to be, acted like, and was treated by all as a great power.