Most Secret War

by R. V. Jones

  I could also on these occasions tell 80 Wing the frequencies of all the beams that would be used on the particular night. It may help at this stage if I explain a few details of the X-beam set-up.

  In principle, the aircraft had to fly along a beam (Figure 3) that was laid directly over the target (the Director Beam) and release its bombs at a point rather short of the target. The information needed to compute the release point involved the height and speed of the aircraft, and where it was at any instant relative to the target and the type of bombs. The way this information was derived in the X-beam system was to lay two beams across the Director Beam, crossing it at pre-determined distances before the target. In general one, the ‘Main Signal’, crossed the Director Beam five kilometres before the target and the other, the ‘Fore Signal’ at twenty kilometres.

  While the pilot flew along the Director Beam, either by listening to it or watching a direction indicator, the bomb aimer listened for the cross beams. The time interval between crossing the two beams would be the time taken to cover fifteen kilometres, which gave the aircraft’s speed, and the main signal also told him that he was five kilometres away. The problem of determining the release point was simplified by a small mechanical computer involving a stop clock that was started by the bomb aimer as he crossed the Fore Signal and stopped as he crossed the Main Signal; and then, if he had fed in the correct height information from his altimeter, the mechanism would work out by itself when the bombs ought to be released.

  Thus, strictly, only three beams were absolutely necessary for the operation of the system, but this was made easier by providing additional beams, one to give a rough indication of where the fine Director Beam lay and the other an additional cross beam to be laid 50 kilometres or so in front of the Fore Signal. A typical layout is shown in Figure 3, which indicates the layout for a raid on Coventry, with the Director Beam coming from near Cherbourg, and the cross beams from near Calais. To provide an insurance against the failure of the Director Beam, which was usually from the station known as Weser, a reserve Director Beam was provided by the nearby station, Spree. Thus on a typical night, with five beams operating (the Main Signal might also be duplicated in addition to the Director Beam) we needed to knock out at least three beams, preferably the two Directors and one other, if we were to deprive the system of its accuracy.

  Fig. 3. The X-Beam dispositions for the bombing of Coventry 14/15 November 1940. Also shown are the two Knickebein beams found by Flight Lieutenant H. E. Bufton and Corporal Mackie for the night of 21/22 June 1940

  Incidentally, the potential accuracy of the system was so great that in calculating the paths of the beams, it was necessary to take into account the fact that the earth is not a simple sphere, but is somewhat flattened towards the Poles. This made a difference of three hundred yards or so in where a beam starting from Cherbourg would actually cross London, compared with where one would calculate it to be on the assumption that the earth was a true sphere. It was necessary to be able to perform rapid calculations, working with a spheroid that more accurately approached the true figure of the earth, and I enlisted the help of Colonel C. J. Willis, who was then in charge of Maps in the Air Ministry. He put me in touch with Dr. L. J. Comrie of the Scientific Computing Service, who did most of the work that ultimately led us to establishing the exact positions of the beam stations in the Calais area, as well as enabling me to deduce the nightly targets accurately.

  As for the accuracy which KGr100 realized in practice, I doubt whether they ever did as well as they might have hoped, although frequently a few bombs actually fell on the pin-point target. On one occasion, we were able to obtain a clear ‘signature’ of their efforts, when they attacked a factory in Birmingham on 26th/27th October. We were able to plot all the bombs, with the result shown in Figure 4. The bombs lay in three pencils running roughly south to north, with the central and heavier pencil 150—200 yards east of where we had calculated. The length of the pencils was presumably unintentional, and due to the difficulty of establishing the exact instants of flying through the cross beams. The two side pencils, lying roughly half a mile on either side of the main, have never been explained, either by me or the Germans, who were surprised when I told them after the war. I thought that they must have included a deliberate offsetting device in the X-Gerät so as to be able to mark an area, but this turns out not to have been the case.

  If my narrative gives the impression that my work was done in a state of unruffled calm, this would be false. We were bombed, or at least alerted for more than seventy consecutive nights with the exception of 3rd November, and slept on the fifth floor. We did indeed sleep, despite all the noise and risks, because we knew that not to sleep would result in exhaustion; and this—as far as my unravelling of the German systems was concerned—would have been fatal. With the disruption to trains, I had often to use my car, and even then getting to the office was not easy. On one morning after I had crossed northwards over Putney Bridge I found Kings Road blocked, where I would normally have turned east-wards towards Westminster, and indeed all alternative routes blocked right up to Notting Hill Gate, where I finally got through. One of the advantages of the Blitz was that if trains were running people went home so early you could get a seat in the normal rush hour.

  Fig. 4. Plot of bombs dropped on Birmingham on 26/27 October 1940 by Kampfgruppe 100 using X-Beams

  Despite the fact that I was a rather vital cog in the defence machine, I had no special privileges. There was indeed talk of giving me a Class C.C. commission in the Royal Air Force at the height of the invasion alarm—but although I would have accepted it if it had materialized I would have felt it was a masquerade, because I had been brought up to regard anyone in uniform who was not in the Front Line as bogus. I had no official transport, either at this time or throughout the war. Indeed, I had no help, apart from what Daisy Mowat was able to give me as secretary when she was not occupied with Fred Winterbotham’s work. His Second-in-Command, John Perkins, knowing that I was beginning to predict nightly targets, suggested that one of his officers, Harold Blyth, should be attached to me. Harold, an old Harrovian who thought his greatest achievement was to have clean-bowled Victor Rothschild, was a concrete specialist in private life, and a good draughtsman. He established the format of all my subsequent wartime reports, and introduced me to the only type of filing system that I have ever found to work. This consists simply of box files into which papers of any size relevant to a particular subject can be hurriedly deposited. I ended the war with more than four hundred files built up on his system, and I have continued to use it ever since.

  Helpful though Harold was, there was obviously going to be a gap if I myself were killed during the Blitz, for nobody but another scientist could know the details of the work, and Fighter Command would therefore have been deprived of some basically very useful information. And so at last I persuaded the Director of Scientific Research, D. R. Pye, to let me have the assistant he had denied me earlier in the year. Naturally I wanted Charles Frank, who by this time had moved to the Chemical Defence Establishment at Porton; he managed to extricate himself and it was arranged that he should join me around 5th November. As recognition of the importance of my work I was promoted as from 11th November to the grade of Senior Scientific Officer, which meant a rise in salary from £575 to £680 per annum.

  Before Charles Frank joined me, I accepted an invitation from A. P. Rowe to visit the Telecommunications Research Establishment at Swanage, partly to tell his people about what I had so far found, and partly to be briefed regarding the newer developments, including what was happening in generating and using waves of around 10 centimetres wavelength. Several of us before the war, including Watson-Watt and myself, had been pressing for new generators of centimetric waves because this was obviously the way to improve the sharpness of the information that radar provided, and several laboratories had been put on to this work at the outbreak of war. A great breakthrough had been made in Birmingham by Randall and Boot, and
this was now being applied to airborne radar at Swanage. It was to give us a great advantage in the radio war.

  CHAPTER EIGHTEEN

  Coventry

  BY EARLY November 1940 the X-Beam stations had become so adept at setting and resetting their beams that KGr100 could mount attacks on two targets in succession on the same night: on 4th November Birmingham and Coventry, on 5th November Coventry and Birmingham, on 8th November Liverpool and Birmingham, and on 12th November Liverpool and Coventry. On some of these raids the X-beam flyers may well have been acting as pathfinders for fairly small forces drawn from other bomber units. One homeward bound KGr100 aircraft, call sign 6N+AH, suffered a compass failure on the night of 5th/6th November, and was then misled by our ‘masking’ of German radio beacons (this was achieved by sending out signals on a British beacon identical to those being sent by the German beacon) into thinking it was over France when it was still over England. Running short of fuel, it landed on the beach near Bridport in Dorset in the early hours of 6th November. It was potentially a unique prize, for it carried a complete X-Gerät; but there then followed an episode as grimly humorous as the Porter’s performance in Macbeth, and equally incongruous as a prelude to high tragedy.

  There are various accounts of what happened but, as I first heard it, the aircraft was found by an Army unit stationed there for coastal defence. The Officer in command left two Other Ranks as guard; he did not, of course, recognize the unique importance of this particular aircraft, with its complete X-Gerät, but he was well aware that it ought to be properly guarded. He left the soldiers with the instruction that they were to let no one touch it until further orders and, unfortunately, added the phrase, ‘I don’t care if even an Admiral comes along. You are not to allow him near it!’

  Now the aircraft had come to rest between high and low water and, as the soldiers watched, the tide inexorably came in. A local naval detachment offered to help the soldiers drag it above high water mark; but the latter, remembering their instructions, stopped the Navy from doing anything, with the result that the whole aircraft was awash. By the time Technical Intelligence officers were available to examine it, all the radio gear was full of sand, and all the light alloy components corroded. This failure to save the aircraft intact may have contributed to the disaster eight days later at Coventry.

  Versions of the story of the Coventry raid have recently appeared, in which Churchill is said to have been presented with the dilemma of evacuating Coventry because of our prior knowledge of the raid, or of doing nothing and so preserving the security of our Enigma work. To the best of my knowledge, this is not true, and I will give my account of what happened.

  By 6th November I was back in London after the visit to Swanage, and on the afternoon of, I think, Sunday 10th November I received a teleprint of a decoded Enigma signal to the X-beam stations. It had been sent the previous day, and it was most unusual in telling the stations to prepare for operations against three targets numbers 51, 52 and 53. It gave the beam settings for the three targets, and it was only a few minutes’ work to find that 51 was Wolverhampton, 52 Birmingham, and 53 Coventry. Instructions for three targets had never been sent out at the same time before: it might be that KGr100 was getting even more ambitious, and intending to attack three targets in one night, but never before had any target instructions been sent out more than a few hours in advance of the attack. Moreover, whereas for the previous month or so all the beam settings had been specified to the nearest second of arc, for the three new targets the settings were merely specified to the nearest minute, implying a coarsening of the accuracy by a factor of 60. Were the seconds to be specified later, or was it that whoever had drawn up the instructions had been as well trained as I had myself in not specifying a higher degree of accuracy than the operation required? I inclined to the latter alternative: remembering that we had already found KGr100 dropping flares, it seemed likely that unless the beam orders were to be made more precise subsequently, the Gruppe was expecting to drop something less accurate than ordinary bombs, which could be either flares or incendiaries. I alerted the proper authorities, for it seemed that the foreshadowed change in German policy was now imminent, but I could not say when or in what order the three targets would be attacked.

  There were other signs that a change was in the offing. On 3rd November we had some indication that preparations were being made for an abnormal amount of signals traffic between Luftflotte 2, commanded by Kesselring, and its subordinate formations. And on 11th November a long Enigma decode of a signal sent on 9th November contained orders for what was evidently to be a very major operation under the code name ‘Moonlight Sonata’. KGr100 was to be involved, and among its tasks was to check the positions of the Knickebein beams. Four target areas, A, B, C and D, were mentioned, and there was one inexplicable word which, had we been able to interpret it, could have given us a clue: ‘KORN’. It meant ‘Corn’, and I wondered whether it might be a code name for the appearance of radar screens when jamming was present, which we ourselves often called ‘Grass’ or when spurious radar reflectors were to be dropped, which although we were subsequently to call them ‘Window’ received the American code name ‘Chaff’. What we did not guess was that it was an alliterative code name for Coventry, which the Germans spelt with a ‘K’.

  Where were target areas A, B, C and D? Since a map from a crashed aircraft had shown some larger areas thus lettered in the south of England, Wing Commander ‘Tubby’ Grant of Air Intelligence made a dubious correlation, for the map areas were too large to be suitable as bombing targets; they would have made much more sense as dropping areas for a large airborne invasion. Grant’s interpretation was circulated by teleprinter, but it was hard to believe. I would have much more likely believed something that Felkin reported, had I seen it. This was that a prisoner from KG1 had said that the heaviest possible attacks were to be made between 15th and 20th November on Birmingham and Coventry. Grant’s interpretation, however, tended to hold the day as far as the Air Staff was concerned, and a series of countersteps was planned on the assumption that the main objectives of Moonlight Sonata were somewhere in the south of England. For some of our proposed counters, these misjudgements hardly mattered: these were the offensive measures against the German bomber bases, and against the beam stations, which were to be attacked by flying specialist bombing aircraft down their own beams, as I had previously suggested.

  My own part in the operational story was the detailed forecasting of KGr100’s targets on a night-to-night basis; and although I had pointed to the unusual threat to Birmingham, Coventry and Wolver-hampton, and to KGr100’s pathfinding role, I had no standing in the making of the general Intelligence assessment. If someone had correlated my warning with that of Felkin, and especially if the KG1 prisoner had mentioned Wolverhampton as well (which he did not) thus making the correlation complete, the Air Staff appreciation might have been less in error. Whether or not this would have made much difference to what happened to Coventry is nevertheless doubtful, because we did not yet know the exact order in which the towns were to be attacked, and we might reasonably have expected Wolverhampton to be first; and even if we did know the correct order, both previous and subsequent experience showed that neither our nightfighters nor our guns could seriously damage the Luftwaffe.

  On 11th November, while we were speculating about Moonlight Sonata, KGr100 was to have attacked Liverpool (target No. 34) and Coventry (target No. 49, different from 53) but these operations were cancelled on account of weather. The Gruppe came out against these two targets on the following night, and then on 13th November it had no operations. So far there was nothing unusual—the target directions were specified with the old accuracy. By this time Charles Frank was with me, and I had introduced him to the Enigma situation, commenting that—thanks to Bletchley—if we could hold out through the winter we had a chance of winning in the end.

  Together we braced ourselves for the following night, and for whatever ‘Moonlight Sonata’ might me
an. It happened to be one of those afternoons when the Enigma signals to the X-beam stations were not broken in time, and we were therefore left guessing. There was, of course, the evidence that could be gathered by flying our own aircraft along the German beams and establishing where they were pointing. Moreover we could listen and obtain the beam frequencies, so that we could set our jammers appropriately. Somewhere between half-past-five and six o’clock Addison telephoned me from his head-quarters at 80 Wing asking for my help in deciding the frequencies on to which to set his jammers. He was fairly sure that the target was somewhere in the Midlands, and his problem was to decide which beam was which, and therefore which to try to obliterate with the three or four jammers that he had available. He then read out to me the list of radio frequencies as determined by our listening aircraft. I could see at once that the measurements must be wrong, in that they did not match up with the figures that I knew from the ‘Anna’ code. I therefore made a mental correction of the measurements as far as I could—for example, 68.6 should probably have been 68.5, if our receivers had been properly calibrated, or 70.9 should have been 71.0. But deciding what, for example, 66.8 meant, was more of a lottery.