The Emperor's Codes

by Michael Smith

  The Japanese cipher clerk operating the Type B machine used a specific letter sequence listed in a book of 1,000 daily changing sequences to plug the connections from the input and output typewriters into the cipher machine's plugboard. He then selected a five-figure indicator group at random from a list of 120, each of which had different settings that should be applied to the banks of stepping switches and to the main switch which controlled the encryption of the ‘sixes’. This five-figure indicator group was enciphered using an additive and sent immediately before the enciphered text in order to tell the receiving station what settings to use.

  To encipher the message, the cipher clerk first encoded it using a basic commercially available substitution cipher known as the Phillips Code. He then typed the enciphered text into one of the electric typewriters, generating an electrical current which ran though the machine's internal wiring and typed out the enciphered text, letter by letter, on the second keyboard.

  The eighteen-month operation to break the Type B machine, which the Americans designated Purple, was based in the SIS headquarters in rooms 3416 and 3418 of the Munitions Building opposite Seventh Avenue on Constitution Avenue, Washington. It was led by William F. Friedman, America's leading codebreaker.

  Friedman was born in Kishinev, the capital of Moldavia, in 1891. When he was still a baby, his parents had emigrated to the United States where he studied genetics and was offered a job in the laboratories of the wealthy textile manufacturer George Fabyan. It was here that he first began to take an interest in ciphers, as a result of Fabyan's obsession with proving that Shakespeare's plays were written by Francis Bacon. After doing some codebreaking work for the US Government during the First World War, Friedman became head of the US Army's small cryptography unit. He set up a special team to break the new Japanese cipher machine. The ‘Purple Section’, as it was known, was led by Frank Rowlett, a small, bespectacled mathematics teacher from Rocky Mount, Virginia, who had been one of Friedman's earliest recruits. ‘The salary offered with the position was two thousand dollars a year, which was more than both my wife and I were making as schoolteachers and we decided we would give it a try,’ Rowlett recalled. He was called to Washington to be interviewed for the job.

  It was during this first exchange with Friedman, when he got through inquiring if I had a place to live and had I been to Washington before, just sort of simple conversation, that I found an opportunity to ask him what a cryptanalyst was supposed to do and he said: ‘You mean you don't know what a cryptanalyst is?’ and I said: ‘I never heard the word before.’ Then he looked out the window and said: ‘Well, that's not strange. I just invented it.’ We were collaborating with the navy at that time and both of us were very anxious to get an insight into this new machine, which we early on named Purple, and we had some success with it in the first weeks but not enough to allow us a close solution of it. All we had was pen and paper and a calculator, one of those desktop things with nine keys and ten columns wide.

  Rowlett's section found that, just as in the traffic enciphered with the Type A Red machine, six of the romaji letters stood out as appearing either more or less frequently than the other twenty. It was not difficult to isolate and recover these six letters, Rowlett recalled.

  While we had no idea of what sort of enciphering mechanism the Japanese cryptographers were using, we were able to design a ‘pen and paper’ analogue which enabled us to decipher the sixes wherever they appeared in the intercepts. This analogue was essentially a deciphering chart six columns wide by twenty-five rows deep representing a polyalphabetic substitution system of twenty-five differently mixed alphabets composed of the same six letters. Recovery of this chart enabled us to decipher the sixes of any key period by the straightforward process of relating each of the six letters to its proper position in the chart.

  The ability to identify six letters wherever they occurred in a message meant that a number of other letters could sometimes be guessed. The Japanese had not only made the mistake of enciphering the message numbers but had incorporated them into the same series used for traffic sent on the Red machine. The message numbers, spelled out in letters, were therefore instantly recoverable. Once again the Japanese tendency towards flowery, long-winded, stereotyped introductions such as the phrase ‘I have the honour to inform your excellency’ assisted the codebreakers, Friedman recalled.

  When the ‘sixes’ in a given message were deciphered, the plain text value of cipher letters scattered here and there throughout the text became available, so that the skeletons of words and phrases offered themselves for completion by the ingenuity and the imagination of the cryptanalyst. For example, suppose that on a given day the six letters forming the ‘sixes’ were E Q A D R H and the following text was at hand:

  Cipher: B R A X E F Q C E V Q O O X H E C F

  Plain: – H E – A – A – E – E – – – E R – –

  Cipher: D L N H Q R V Q P P L C E R P

  Plain: E – – R E Q – E – – – – H A –

  It is not difficult to imagine that the missing letters are those shown below:

  Cipher: B R A X E F Q C E V Q O O X H E C F

  Plain: T H E J A P A N E S E G O V E R N M

  Cipher: D L N H Q R V Q P P L C E R P

  Plain: E N T R E Q U E S T S T H A T

  The codebreakers also had access to a number of helpful ‘cribs’ of the plain text. Messages sent on the Purple machine had often also been sent on the Red machine. Japanese diplomats were usefully unwilling to cut anything from US State Department communiqués, copies of which the codebreakers had no difficulty in obtaining.

  But while all this could help to break a number of individual messages, it did not provide any continuity of decryption. The American codebreakers were only able to take one message at a time with only the position of the deciphered letters from the day's ‘sixes’ and the context in which the message was being sent to assist them, Friedman recalled.

  It speedily became apparent that any cryptographic relationship between the plain text and the constantly shifting cipher text values in the case of the letters constituting the group of ‘twenties’ had been most carefully eliminated, disguised, or suppressed. In several cases, after a few words had thus been obtained by pure ‘guessing’, a clue was afforded as to the general nature of the message. This led to a frantic search for a complete document which might be available either in our own files or in the files of other government agencies.

  In all the plain texts for parts of some fifteen fairly lengthy messages were obtained by the methods indicated above, and these were subjected to most intensive and exhaustive cryptanalytic studies. To the consternation of the cryptanalysts, not only was there a complete and absolute absence of any causal repetitions within any single message, no matter how long, or between two messages with different indicators on the same day, but when repetitions of three, or occasionally four, cipher letters were found, these never represented the same plain text.

  In fact, a statistical calculation gave the astonishing result that the number of repetitions actually present in these cryptograms was less than the number to be expected had the letters comprising them been drawn at random out of a hat. Apparently, the machine had with malicious intent – but brilliantly – been constructed to suppress all plain-text repetition. Nevertheless, the cryptanalysts had a feeling that this very circumstance would, in the final analysis, prove to be the undoing of the system and mechanism. And so it turned out.

  The US codebreakers soon realized that what they needed was an immense ‘depth’ of enciphered text sent on the same day and using the same indicator. But although by now they had managed to reconstruct, at least in part, around one thousand messages, only in a very few cases were there any messages sent on the same day with the same indicator and even then there were only ever two messages that coincided in this way. There were, however, a number of messages sent on different days but using the same indicator and, ingeniously, by working out how the machine treat
ed each of those different messages, the codebreakers managed to deduce how it would have enciphered them on one single given day. They now had a depth of six different messages all notionally sent on one day's keys.

  Their attempts to find any links between the way the plain text in the messages was enciphered in the various messages that would let them into the machine were hampered by the need to find extra space in the crowded Munitions Building. Workmen had begun adding a fourth floor directly above their office, Rowlett recalled.

  Our ears and minds were filled with the distracting sounds of hammering, banging and shouting. When the construction cranes hoisted heavy materials to the workmen above us, we could feel the building shake as the loads were dropped on the roof. Worst of all was the incessant vibration from the jackhammers overhead which started early in the morning and continued all day long. This took place while the heat of the late summer was still on us, long before the days of air-conditioning. When we closed the windows to cut down on the noise, we sweltered. When we opened them to get relief from the heat, the racket was unendurable. At times it was impossible to communicate.

  The codebreakers were understandably jubilant when on 20 September 1940 Genevieve Grotjan, one of the key members of Purple Section, found similar sequences of enciphered and plain-text letters in a number of the messages she was studying. When she called out to her colleagues to tell them what she had found, one of them became so excited that he started dancing around Grotjan's desk holding his hands in the air like a victorious boxer. Another codebreaker, normally among the most studious members of the team, was yelling, ‘Hurrah, Hurrah’ at the top of his voice, Rowlett recalled. ‘I could not resist jumping up and down and waving my arms above my head and exclaiming: “That's it! That's it! Gene has found what we have been looking for.” ‘

  The Purple Section celebrated its success rather modestly by sending out for bottles of Coca-Cola. But the strain of the eighteen-month operation had been too much for Friedman, who suffered a nervous breakdown and as a result was off work for several months.

  The US Army codebreakers used the relationships between the like sequences of enciphered and plain text to work out the wiring of the machine, and Leo Rosen, a young electronics engineer, began to construct a customized machine that would simulate its operations.

  The most urgent priority, once the machine was built and the Purple traffic could be broken regularly, was to go back over the messages sent before the signing of the Tripartite Pact to uncover the fine detail of the negotiations. The diplomatic signals intelligence produced by deciphering the Purple traffic, codenamed Magic by the Americans, was to be of inestimable value to the Allies as the war progressed and none was to prove more valuable than the messages passing between Tokyo and its Ambassador in Berlin, Oshima Hiroshi.

  After being promoted to Ambassador in 1938, Oshima had been recalled to Tokyo in the wake of the German invasion of Poland in September 1939. Joachim von Ribbentrop, now German Foreign Minister, had made strenuous efforts to persuade the Japanese Government to leave him in place, but to no avail. Nevertheless, Oshima maintained his close contacts with the Germans and was one of the leading advocates of the Axis alliance. In the wake of the signing of the Tripartite Pact, he was sent back to Berlin as Ambassador.

  In a speech given at Oshima's farewell party in Tokyo, Matsuoka Yosuke, the Japanese Foreign Minister, praised the newly appointed Ambassador. ‘His thorough knowledge of German affairs far exceeds one's imagination,’ Matsuoka said. ‘He has developed the highest personal trust among the leading members of the German Government, thus he can have heart-to-heart talks with them.’

  It was the product of those heart-to-heart conversations that the Allies were anxious to hear and it was not just the Americans who were to have access to them. For the breaking of the Purple cipher was just the bargaining chip the US Army needed as part of the tentative negotiations that were already in place for a codebreaking alliance with the British.

  A few months earlier, the Bletchley Park codebreakers had succeeded in breaking the Enigma machine cipher used for all high-level communications throughout the German armed forces. Armed with their own bargaining chip, they had approached the US Navy in June 1940, offering an exchange of cryptographic information, but they were rebuffed.

  However, a direct approach to President Franklin D. Roosevelt succeeded in winning his backing for an exchange of technical information. At a meeting in London with the British armed forces chiefs of staff, a month before Friedman's team broke Purple, the Americans suddenly suggested ‘a free exchange of intelligence’ on the breaking of the Japanese diplomatic ciphers.

  Although Captain Laurance Safford, the commander of the US Navy's codebreaking operation OP-20-G, was still very much opposed to any major exchange of information, Friedman and Colonel Spencer B. Akin, the civilian and military heads of SIS, were both very keen. Despite the US Navy's reservations, Roosevelt approved a deal under which Britain and America would ‘exchange complete technical information re Japanese, German and Italian codes and cipher systems’ but which excluded an exchange of actual intercepts.

  The cryptographic exchange accord was agreed by senior US and British representatives in Washington in December 1940. The following month, nearly a year before the Japanese attack on Pearl Harbor brought the Americans into the war, a four-man American delegation – comprising two US Army officers, Captain Abraham Sinkov and Lieutenant Leo Rosen, together with two US Navy officers, Lieutenant Robert Weeks and Ensign Prescott Currier – set sail for Britain carrying ‘certain packages’. The presence in the party of Rosen, the technical expert who had constructed the so-called Purple machine, was significant. At least one of the ‘packages’ the Americans brought with them to Bletchley was a Purple machine.

  ‘It was early in 1941,’ said Barbara Abernethy, who was then working as Denniston's personal assistant. ‘Commander Denniston told me he had something important to tell me. “There are going to be four Americans who are coming to see me at twelve o'clock tonight,” he said. “I require you to come in with the sherry. You are not to tell anybody who they are or what they will be doing.” ’

  Currier described landing at Sheerness dockyard on the afternoon of 8 February, and being met by a small delegation from Bletchley Park which included Tiltman. The crates containing the precious top-secret ‘packages’ were loaded on to lorries and the convoy headed west towards London en route for Bletchley, Currier recalled.

  It soon became dark and the countryside was pitch black with rarely a light showing except for the faint glow emanating from a small hole scraped in the blacked-out headlight lens of the cars. When we arrived at BP, the large brick mansion was barely visible; not a glimmer of light showed through the blackout curtains. We were led through the main doors, and after passing through a blacked-out vestibule, into a dimly lit hallway, then into the office of Commander Denniston RN, chief of GC&CS. Denniston and his senior staff were standing in a semicircle around his desk and we were introduced to and greeted by each in turn. It was truly a memorable moment for me.

  Barbara Abernethy served each of the American guests with a glass of sherry.

  It came from the Army & Navy Stores and was in a great big cask which I could hardly lift. But Denniston rang the bell and I struggled in and somehow managed to pour glasses of sherry for these poor Americans, who I kept looking at. I'd never seen Americans before, except in the films. I just plied them with sherry. I hadn't the faintest idea what they were doing there; I wasn't told. But it was very exciting and hushed voices. I couldn't hear anything of what was said but I was told not to tell anybody about it. I guess it wasn't general knowledge that the Americans had got any liaison with Bletchley. It was before Pearl Harbor, you see, and presumably Roosevelt was not telling everybody there was going to be any liaison at that stage.

  The British kept to the precise letter of the agreement, providing detailed information on how they had broken the Enigma cipher and on their work on a numbe
r of other codes and ciphers, including Tiltman's studies of the main Japanese Army system. But in line with the Washington discussions, no details were provided of any of the actual messages they had intercepted. Even if this had not been an American condition, it seems likely that the British would have raised it since they were concerned over the Americans’ lack of a secure system for the dissemination of the ‘Special Intelligence’.

  Denniston told Stewart Menzies, who had taken over as ‘C’ in September 1939 when Sinclair died, that Currier and his colleagues had been ‘informed of the progress made on the Enigma machine’. The Americans were given a ‘paper model of the Enigma machine’, detailing its wiring and how it worked, together with details of the Bombes, the primitive computers designed by the British code-breaker Alan Turing to break the Enigma keys. This was as much as, if not more than, the Americans provided.

  Without a shadow of doubt, the most significant contribution on the American side had been the ability to break Purple, provided generously from the outset by the US Army codebreakers. The British were again able to read all of the ‘State Secret’ communiqués passing between the main Japanese missions and Tokyo, just at the point when Oshima was returning to Berlin to continue his close relationship with the Nazi leadership and the German High Command.

  Safford complained at what the Americans received in return, horrifying the British, and doing nothing to assuage their concerns over US security, by writing an unclassified letter to demand that the Americans be given an Enigma machine. Safford later claimed that the British reneged on their side of the deal and had ‘double-crossed us’.

  The false perception that the British were holding back on the exchange deal, largely the result of the US Navy codebreakers’ initial failure to understand the ‘paper Enigma machine’ the British had handed over, was to become endemic among a number of senior US officers. Yet, at the cutting edge, US codebreakers said there was nothing the British held back. Currier recalled an atmosphere of ‘complete co-operation’ and said the members of the American delegation were shown everything they wanted to see.