Dilly

by Batey, Mavis;

  I couldn’t understand what Penelope meant when she said: ‘Half a moment, you know what you are saying, don’t you? That is exactly what Alice said to Dilly the Dodo when she met him for the first time in Room 40.’ She then produced a copy of Alice in ID25, which I have treasured ever since. His appearance, as I remembered him, was still as in Frank Birch’s skit, when ‘Alice thought that he was the queerest bird she had ever seen. He was so long and lean and he had outgrown his clothes and his face was like a pang of hunger.’ Alice also noted that he was very shortsighted and that his spectacles were kept in his tobacco pouch. I was reminded of the amount of time we spent searching for his spectacles and his tobacco tin buried under stacks of messages and how often when preoccupied he mistook the cupboard door for the exit. By one of those inexplicable coincidences, Penelope and I were sitting in my husband’s rooms at Christ Church overlooking Alice’s garden and I was writing a book on Lewis Carroll. I have always felt a bond with Dilly, who was for me Alice’s White Knight, endearingly eccentric and concerned about my welfare.

  There were only about eight girls in the Cottage when I arrived. Their work was devoted to ‘retrieving the misses by Hut 6 and to discovering new lines of attack’, according to a Dilly memorandum. For new girls, this meant endlessly putting dots in squares for frequency counts. There were catalogues called ‘corsets’ attempting to locate the word EINS in the text. There was also something in Hut 6 called Dillyismus, being a method for determining Stecker when the rest of the message setting was known, but it meant nothing to us then. Claire Harding was in charge of the allocation of work and for relief we sometimes made snake alphabets for cillis. It was hardly stimulating but everything changed when the phoney war ended in May, followed by Dunkirk. It was brought home to us starkly when the trains taking exhausted troops north stopped at Bletchley. A cry went up for help at the forces canteen and we rushed down in between shifts. The fierce woman in charge ordained that the young girls should be confined to the kitchen out of sight and that only the godly matrons would take out the tea and chips to the troops on the train. We thought they would have preferred it the other way round; we certainly would have done. Dilly answered the call for the Home Guard, which practised in the stable-yard. When I told Penelope that we were glad that they only had broomsticks and not rifles she said that we need not have worried as her uncle was rather a good shot in spite of his poor eyesight.

  It was now imperative, with U-boats on the rampage in the Atlantic, to break the German wartime naval Enigma traffic, as Alan Turing had not yet solved the additional indicator bigram problem. Admiral Godfrey was very conscious of how much Room 40 had done to beat the U-boats and he wrote to Frank Birch, now head of the naval section, to say that he was now setting up an organisation to arrange ‘pinches’, thefts of original German codebooks, or cipher material. ‘I think the solution will be found in a combined committee of talent in your department and mine who can think up cunning schemes,’ Godfrey said. Who better to dream up such schemes than his personal assistant, Ian Fleming, and who better to send him to than Dilly Knox, the ace cryptographer who had broken the First World War U-boat code? Dilly was not lost for bright ideas and Birch records: ‘When talking to Lt-Cdr Fleming the other day Mr Knox put forward the following suggestion. The Enigma key for one day might be obtained by asking for it in a bogus signal.’ Dilly would have explained to Fleming that it was not the actual Enigma machine that was needed, as was often thought. We already had a working model of it, but that was useless by itself as there were millions of possible settings; it was the key to these that was needed. James Bond’s creator had a much more bloodthirsty suggestion: to obtain the keys from a German naval vessel in the Channel. He sketched out Operation Ruthless, involving a captured German bomber and a participant who could have been the model for his own fictional hero and ‘should be tough, a bachelor, able to swim’. The tough bachelor would be part of a crew of five who would dress up as Luftwaffe crew trying to fly a damaged German bomber back from a raid on the UK. They would ditch in the Channel near a German ship and once rescued kill the crew and sail the German ship with the vital Enigma keys back to Britain. Fleming pencilled his name in as a possible participant although he knew full well that anyone who had knowledge of what went on at Bletchley Park was forbidden to put himself in a dangerous situation where he might be taken a prisoner and interrogated. The circumstances were never right for Operation Ruthless to be activated, causing immense disappointment at Bletchley, and Birch, who had given enthusiastic backing to the ‘very ingenious plot’, gave a flavour of the dismay of Dilly and the other codebreakers over the lack of ‘pinches’ in a memo to Fleming. ‘Turing and Twinn came to me like undertakers cheated of a nice corpse yesterday, all in a stew about the cancellation of Ruthless,’ Birch said. ‘The burden of their song was the importance of the pinch. Did the authorities realise that there was very little hope, if any, of deciphering current, or even approximately current, [naval] Enigma at all.’ During the Fleming centenary year in 2008, I was frequently asked if I had met the creator of James Bond. I certainly met Godfrey, on whom Fleming modelled Bond’s ‘M’, but I feel sure that if Dilly had become aware of the Fleming lifestyle he would not have let him near his girls.

  Life had changed for us in the Cottage when Italy declared war in June and instead of being at the bottom of the pile doing unconnected things for Dilly to pass on to Hut 6 we were now put on to breaking the Italian naval Enigma in the Cottage. Denniston’s ‘new fields to explore’ revived Dilly’s spirits. He had of course already broken the machine the Italians had used in the Spanish Civil War, which only ended in March 1939. There was great excitement when the first wartime Italian naval messages were intercepted but there was only a slim chance that their naval high command (Supermarina as we would learn it was called) was using the ‘K’ Enigma for which Dilly had worked out the rodding solution. Naturally the first thing to do was to try out the existing crib charts, such as PERXCOMANDANTE [‘For Commandant’, the X representing the space between the two words], but this yielded nothing as the traffic we had was from Rome to stations in the colonies, which did not use any known cribs. Dilly’s girls then had to resort to rodding likely beginners, such as ‘destinario’ or ‘per’, where the guessed letter and the cipher text did not ‘crash’; that is to say when hunting for ‘per’ we would disregard a message beginning with P as P cannot be enciphered as P and likewise with the other crib letters.

  Rodding was a great improvement on dots in squares. The important thing about it was that Dilly had made it into a kind of word game which even a beginner could do without knowing how the machine worked, rather like driving a car without knowing what goes on under the bonnet. Few of Dilly’s girls ever grasped the mechanics of the machine but became skilful rodders. It was, however, a very laborious task requiring infinite patience as there were seventy-eight different rod trials to make in order to cover all twenty-six positions of each of the three wheels. Finally in September 1940, after many frustrating weeks of rodding, when it was feared that the wheel wiring had indeed been changed, success came. When I rodded PERX for the right-hand wheel, as instructed, in one position the first three rod pairings necessary for the cipher and clear texts produced good letters in other positions on the rod but instead of X in the required place the letter S appeared of its own accord from the last rod coupling. My lucky guess was that the message might begin PERSONALE, rather than PERX, and I then had five other couplings to work on and my Italian crossword puzzling was looking good. Of course it all depended where the turnover would be and gibberish could then be expected. When that happened I was meant to hand the message over to Dilly, but it was the evening shift and he was at home and I was determined to get over the turnover on the right-hand wheel and on to the middle wheel.

  Dilly’s advice was always to use pencil for guesses in rodding and ink for ‘clicks’, confirmations of probable cribs of individual letters. There were two kinds of clicks. The most import
ant were what Dilly called ‘beetles’. These were direct clicks in which both letters of the crib occurred side by side on the same rod. Dilly called the other kind ‘starfish’. This was a cross-click where one of the crib letters was on one rod and the other on the second rod. A starfish click came as an effortless bonus, as mine did in PERSONALE, but beetles were very important. If one occurred in, say, positions 24 and 25 on the rod, then one of the remaining bigrams in the same column on the rod square would have to be used and most of them would be impossible combinations that could not be parts of real words such as WJ, QB, KP and so on. When I came to my beetle in the text there was a limited range to choose from as I had already used up so many of the rods. I tried NO, which produced the making of SIGNOR, and then, having put up the rods a gobbledegook man’s name appeared, which caused me a lot of trouble. However, I had got enough couplings on the other side of the turnover to know that I must try and find somewhere on one of the other rod squares where these would not only comfortably lie side by side but produce sense when the missing rod couplings were supplied for the previous stretch of the message before the turnover. I then struggled with the rest of the text with my pocket Italian dictionary and it was well after midnight when the task was completed.

  Dilly couldn’t believe his eyes when I handed him the deciphered message the next morning. He took me out to dinner in the evening to the Fountain Inn at Stony Stratford. It was my first experience of being driven in Dilly’s Baby Austin. As Denniston said, he was not a good driver, especially in the blackout. Preparations for a possible German invasion were already in place but he somehow managed to evade the tank traps along Watling Street. We had a very relaxed dinner and he was anxious to know what I wanted to do when the war was over. I hadn’t given it much thought, with the Germans poised at the other side of the Channel, and, having abandoned all idea of an academic career with the German Romantics, I said I might try for journalism. He promised to introduce me to his brother Evoe, who would get me started. More helpfully, he managed to get me put up a grade, which was a great relief as I was only earning 35 shillings a week and having to pay a guinea for my billet. I was also promoted from a backroom girl to the front room as a machine cryptographer.

  By this time, Alan Turing had written a manual for newcomers to Huts 6 and 8, which became known as ‘Prof’s Treatise on Enigma’; it covered Dilly’s breaking of the un-Steckered Enigma, leading on to the Steckered German armed forces machine they would be dealing with. Examples were given of the breaking of the ‘K’ model railway Enigma machine by Turing and Twinn in 1940, using Dilly’s rodding methods and depth-reading of messages by John Tiltman. There was no necessity for providing the history of Enigma-breaking so that Dilly’s name does not appear and consequently it has often been assumed that the methods given for solving un-Steckered Enigma were Turing’s own work, although he never claimed them to be so himself. Needless to say, Prof’s manual was never used in the Cottage and the first time I heard of it was when it was released as a national archive in 1998. Dilly’s rodding technique looked strangely different without his whimsical nomenclature. Beetles and starfish had become direct and cross-clicks and buttoning-up adding-up. Probability had replaced serendipity and it all seemed very heavy going compared with our treatment of Enigma linguistic puzzles.

  The good thing was the psychological effect of knowing that we were no longer working in the dark and that we were dealing with the ‘K’ machine, whose wheel wirings Dilly had broken in the Spanish Civil War. Each message had to be broken separately, however, as the indicating system was never solved, presumably having been taken from a codebook. Moreover, since no real values for the settings were ever available, the wheel turnovers were never known. Consequently each message had to be broken independently of any other on the same key. To standardise the volume of traffic the Italians used to send out dummy messages, usually of two kinds. Fortunately these were easily recognisable, otherwise it would have been infuriating to have spent days working on a dummy or worse still to have broken a real language message only to find it was a passage from Dante’s Inferno.

  One dummy occurred when a series of unenciphered groups were tapped out and transmitted with the usual type indicator, where the groups always formed patterns from the German keyboard such as QWEAS PYXCF. The second kind was extremely useful, since it was thirty or more groups of solid crib and the first time one was sent enabled a new wheel wiring to be broken. As soon as I picked the long message up, I saw it had no Ls in it and realised that this operator must just have pressed the last letter of his keyboard, probably relaxing with a fag in his mouth. Dilly’s usual method of ‘buttoning-up’ on the QWERTZU diagonal was proving very difficult owing to the repetitive nature of the crib. Once again I was alone in the Cottage on the evening shift and this time I sought the help of someone on Hut 6 watch. As luck would have it, it was Keith Batey. We put our heads together and in the calmer light of logic and much ersatz coffee, the problem was solved.

  After the Italian break, the good thing was we could not expect many L duds but the biggest gift the Italians gave us was that they insisted on their operators spelling out full stops as XALTX, and in order to keep the message in five-letter groups, the end one would have Xs added, so that there might be an eight-letter crib XALTXXXX right at the end of the message. Dilly’s XALTX charts, also known as ‘click charts’, were not confined to locating adjacent starfish and beetle clicks, they showed all possible clicks of the X to A, L, T and X; of the A to L, T, X and so on; the clicks were all reduced to the rod position of the original X, so that a click between X and T in positions 1 and 4 and A and L in positions 2 and 3 would appear in the chart as a double click in 1. Little did Mussolini know that a humble full stop would be responsible for the defeat of his navy in the Mediterranean.

  NINE

  The Battle of Matapan

  Tell Dilly we have had a great victory in the Mediterranean.

  The Battle of Matapan was Britain’s first major naval victory during the Second World War and coming at a bleak time in 1941 it was a real morale booster. A naval victory was traditionally encouraging to the British public, especially as it could now be seen in newsreels with guns blazing and the dashing Admiral Sir Andrew Cunningham on the bridge of his flagship HMS Warspite. Prime Minister Winston Churchill hailed it as ‘the greatest fight since Trafalgar’ and Cunningham was acclaimed as a second Nelson. It was a fortnight before the Admiralty received the news of Nelson’s victory but almost before the last shot was fired at Matapan, in the early hours of 29 March, Admiral Godfrey rang through to Bletchley Park with the message ‘Tell Dilly we have had a great victory in the Mediterranean and it is entirely due to him and his girls’. All very heady stuff for Dilly’s girls and we only wished we could have told our families that we had had a hand in it.

  For Dilly it was a twofold triumph, not only for breaking the Matapan messages in the Cottage but also for the special procedure by which Cunningham had received the intelligence derived from them, known as ULTRA. This was the brainchild of Wing Commander Frederick Winterbotham, the head of the SIS air intelligence section, who superintended the setting up of Hut 3 in 1939 in readiness for Hut 6’s production line of German air force and army traffic. Winterbotham was an old hand, having been sent to Germany in the early 1930s to monitor the expansion of the Luftwaffe and appeared (misspelt) on the Gestapo wanted list. He put up a plan to Stewart Menzies to ensure that the intelligence derived from Hut 3 messages would be disseminated in a way which was both operationally effective and secure.

  In the normal course of events, information from a secret source would be passed only to the directors of intelligence of the service ministries and it was up to them to distribute it as they saw fit. This would work quite well with the small number of messages then in circulation but when it came to thousands, not only might the translations and interpretation differ with so many people handling them, but information might be transmitted to commanders in the field in seve
ral different ciphers, which Winterbotham had learned from the codebreakers was a dangerous practice. He also knew that those commanders in the field could be lax about security and had been known to leave secret documents in their shaving mugs. He set up Special Liaison Units to deliver ULTRA messages to commanders under constant surveillance. By 1942, after America had come into the war, ULTRA became the standard designation for intelligence derived from all decrypts of high-level messages, but at the time of Matapan it was only used as Winterbotham originally defined it.

  Since the red Luftwaffe signals were the first to be broken in Hut 6, Winterbotham got permission to send three or four German-speaking RAF officers to Hut 3. The Hut 6 codebreakers would pass the decoded signals to them for translation and assessing for priority, whether for distribution to the Prime Minister, the chiefs of staff, directors of intelligence or commanders in the field. Very soon a joint RAF and Army section was put in place, but the sea lords were adamant that decoded naval signals should be sent direct to the operation intelligence centre at the Admiralty verbatim for them to take the necessary action. When it came to Dilly’s turn to learn how GC&CS was to deal with the naval intelligence following the break of the Italian Enigma machine messages in September 1940, he was determined that the material we were working on in the Cottage should receive the same treatment as the ULTRA intelligence put out by Hut 3. He angrily claimed that he had been tricked into allowing the Cottage to become ‘an obscure subsidiary’ of ‘Nobby’ Clarke’s Italian naval section. ‘This Italian business has now reached large proportions and takes up almost all our time in the cottage and very much work is done in the hut,’ Dilly told Alastair Denniston, while ‘the Enigma results are of an order of certainty differing wholly from the products of most other intelligence sections. On personal grounds, I find that I have been tricked by your arrangement of 5 December. Had I appreciated at the time the sense in which you now take it I should have gone to any lengths to oppose it. I must ask you to deal at once with the question. I have no intention of continuing to work as an obscure subsidiary of Commander Clarke.’