My servant or ‘bearer’ was Abdul Hamid, a cheerful and resourceful man who looked after me and my possessions devotedly. He must have been born about the time his notorious namesake the Sultan of Turkey died. I settled in and walked up the stony track to the security area, where I was issued with a pass and introduced to a roomful of people informally presided over by Hugh Lloyd-Jones. The room was square and high, and had a huge old-fashioned fan in the ceiling which revolved just fast enough to give a gentle breeze without swishing our papers on the floor or disturbing the fat pigeon who sat on the blade and spent most of his life placidly going round and round. When the fan stopped he stayed there, giddy or not.
Stripp was initially put to work translating decoded messages in a code designated 6633, a variant of the Army Air General Purpose Code 3366 which was in widespread use within the Burma area.
I soon moved from Lloyd-Jones's room to Robin Gibson's, where we were concerned not so much with routine translation as with tougher problems: messages which could not be dealt with because the key-book or substitution table had been changed, or the indicator lost, or because either the Japanese operator or our own had slipped up. Such messages could play havoc with shift-working, especially when a badly mangled message seemed to contain urgent information; it continued to go round and round in your head long after you had given it up as hopeless. Sometimes light would dawn, or at least a promising line of attack would suggest itself, some hours later, and mean an un-expected return in someone else's shift to try it out.
The 6633 and other signals broken by us and by other Allied centres, and worked up into joint intelligence summaries, produced by late 1944 a virtually complete order of battle for the JAAF. It showed not only where almost every hikoo-sentai (air regiment, nominally of forty-eight fighters or thirty-six bombers, plus one aircraft for the commanding officer) was stationed, but how many of those aircraft were serviceable, what type they were, where they had served since 1942, where their base unit in Japan was, and often what the full name of the commanding officer was.
Japanese Navy Air Force transmissions were covered in Ceylon. Two RAF wireless units, 357 WU and 370 WU, were attached to the Royal Navy's codebreaking centre in Colombo – where they were known collectively as RAF Anderson – and a separate naval air code-breaking section was set up within the navy codebreaking base under Commander E. H. M. Colegrave.
The navy air operational codes, JN165 and JN166, were broken by Eric Nave's section in Central Bureau, allowing advanced warning of air raids and providing important intelligence on Japanese naval convoys. With the Allies now firmly in the ascendant, the Japanese Navy was forced to provide air cover for any convoy in reach of Allied bases. The escort aircraft provided daily positions of the convoys. As the dangers the convoys were exposed to became greater, it became normal practice to shelter overnight in safe anchorages, and these too were always faithfully reported by the escort aircraft, Nave recalled.
I found a few messages from planes to base which fell into no known pattern and were unreadable at first sight. However, experience in this field was a most valuable thing, and identifying first a few numerals, I found the aircraft was reporting to base its position. It was a valuable discovery to read: ‘Have taken up position over convoy,’ with latitude, longitude, course and speed given. Almost invariably I could get sufficient of the whole message to be able to give this to our US Navy contact. This situation provided a rich field of operation for the US submarines of which they made good use.
The naval air transport codes were also easily read. Security on these frequencies was very slack with large amounts of operator chatter often referring in clear to facts that had been encoded elsewhere. Frequent cribs were also provided by the common use of different systems; for example, two aircraft might be flying to Truk at around the same time, one from Japan using the latest code and the other from some more remote location such as Rabaul which was still having to use the old code. The main Japanese meteorological station would then have to provide the same weather details for one aircraft in the old code, the details of which were familiar to the Allied codebreakers, and for the other aircraft in the new code, which until this point had not been broken.
The codebreakers working on army air systems at Bletchley Park were part of the military section. Bob Biggs was one of those transferred across to Japanese codes as the war in Europe began to wind down.
I was called up into the Royal Army Service Corps in December 1941 and eventually found myself sent to Bletchley Park. I had been to grammar school but I had never gone to university. I certainly wasn't one of the geniuses you always hear about. But I think at that stage they were just looking for people with a reasonable amount of intelligence, common sense and a good memory, and that I did have. I worked first in the Italian section, then on the Balkans and from mid-1944 I was in the Japanese army air section. We were dealing with two codes – 3366 and 6633 – both four-figure reciphered codes using the same book but different methods of recipherment.
The only way first to decipher and then to decode these messages was to collect a batch of messages each of which had the same indicator number. These were then written out, one underneath the other, on large sheets of graph paper to form a depth. It was essential that the code groups of every message were lined up in the columns because the break would be done vertically.
There were six of us, the others from the Intelligence Corps, and we were known jokingly as ‘strippers’ because our main task was to find the cipher and then strip it off the codebook groups column by column. I knew no Japanese at all when I started but it is possible to do the sort of work I was doing without knowing how to read, write or speak any foreign language. My previous experience as an assistant in the Italian and Yugoslav sections had taught me to look for a proforma message such as a strength report, stock returns, rations incident, casualties or sick report, or even a weather forecast.
The Japanese codebooks had code groups for special instructions such as open and close brackets, full stops, figures 0–9, and Latin alphabet romaji letters to spell out proper or place names. We would start off by deducting the code group for, say, an open bracket from each of the four-figure groups of one of the messages in turn to produce a potential cipher additive figure. By deducting that additive figure from all of the groups in the same column, you might be lucky enough to find another common code group and expose further clues as to what might precede or follow that particular group. In any event, if you found the position of an open bracket then it is obvious you must look for the close bracket. It was known that brackets would enclose figures, which meant you had ten chances of finding out what each of those figures was. By perseverance and patience you would now be creating a jigsaw-like pattern across the sheet, which allowed you to predict more code or additive groups.
If individual letters appeared in any of the messages you might be fortunate enough to recognize what they were spelling out and so be able to complete a run of breaks. Although much of this work could be done without any knowledge of the language, it was, of course, essential for a linguist to be available to interpret words and phrases or guess at the meanings of the Japanese text. The completed graph sheets were then translated into plain English by the linguists and were passed to the officers of the watch to be sorted into priority and distributed for any action deemed necessary. The cipher groups were recorded on a master sheet and newly identified code groups were put into the master codebook.
Bernard Keeffe was one of the Japanese linguists working in the army air section, having been one of the bright young schoolboys recruited when John Tiltman set up his language course.
I had won a scholarship in Classics to Clare, Cambridge, and was invited to take a test to see if I was apt to take the Japanese course. We were given crosswords to solve and some basic words in Japanese to see if we could recognize them and deduce their meaning in different contexts. There was also some simple substitution ciphering. I was accepted and called up at t
he age of eighteen to undergo six weeks’ basic training in Bury St Edmunds before being transferred to Bedford – a brief but painful and revealing experience of the idiocy of army life. The NCOs knew nothing of our destination but assumed these potential officers (POs) were something special. If we did anything wrong – didn't get the body-locking pin, ‘the ladies’ delight’, on the Bren gun in the right place – ‘Call yourself an f***ing PO’ – we found ourselves on a charge. Unfortunately, I developed bronchial pneumonia from lying in the snow on an exercise and missed the delight of the passing-out parade.
The Japanese course was held in a large house in De Parys Avenue, Bedford. We lived in digs. As you entered the house the Japanese course was on the left and a codebreaking course on the right. We knew each other, of course; one of the code-boys was Robert Pitman, a passionate lefty who turned into a right-wing columnist for the Daily Express. But even at that stage it was remarkable that we didn't discuss each other's activities. I was interested in music and had started solo singing as a baritone at school. Bedford was like musical heaven because the BBC Symphony Orchestra had been evacuated there, giving broadcasts from Bedford School Hall or from the Corn Exchange, and added to that there were shows by Glenn Miller and his US Army Air Force Band.
At the end of his course, Keeffe was sent to the army air section, where his boss was Alexis Vlasto. By the time he arrived at Bletchley the capture of a codebook had eased the difficulties faced by the codebreakers, who nevertheless still had to identify the indicator and strip off the additive cipher.
We were working on the order of battle of the Japanese naval air force, collating tedious reports of aircraft damage, how many were transferred, say, from ‘Kuara Rumpuru’ [Kuala Lumpur] to ‘Shingaporu’ [Singapore], how many sorties had been flown – rarely anything dramatic in itself, but giving the analysts material to keep in touch with the changing situation. We worked from clumsily photographed copies of the four-figure codebook, which I believe had been captured by the Americans. They were nothing like modern copies but stiff photos in a ring binder.
Like all other members of the British Army, Keeffe was quartered in the Shenley Road Barracks.
This was commanded by Colonel Fillingham – quite mad, probably too much even for the Durham Light Infantry, who no doubt gladly shot him off to what the rest of the army regarded as a nuthouse. He was not allowed into BP and wasn't told what was going on. He took out his frustration on us; he hated the sight of long-haired intellectuals and used to stop them and give them sixpence to get a haircut. He organized boxing – I was put into the ring with someone six inches taller and just about survived. Then he started cross-country runs before breakfast. I shall never forget the sight of Staff Sergeant Asa Briggs, future professor and historian of the BBC, trying to keep up with less portly young blades.
BP was an astonishing community. I was born to a poorish family in Woolwich. My father was a clerk in the local Co-op society, my family descendants of illiterate Irish immigrants who fled the famine in 1849. I shall never forget the impact of arriving in BP; it was a microcosm of the highest intellectual life. I discovered there was a lively opera group run by James Robertson, later Music Director of Sadlers Wells. I sang with an orchestra for the first time as the Gardener in The Marriage of Figaro and the Constable in Vaughan Williams's Hugh the Drover. Soon after I arrived I organized lunch-time concerts in the Assembly Hall outside the main gate. There were many professional musicians: Captain Daniel Jones, the doyen of Welsh composers; Lieutenant Ludovic Stewart, violinist; Jill Medway, a singer who later married Vlasto; Captain Douglas Jones (later Craig), singer and later company manager at Glyndebourne. There was a choir conducted by Sergeant Herbert Murrill, future Head of Music at the BBC. Working with me on the army air codes was Lieutenant Michael Whewell, bassoonist, and later producer in charge of the BBC Symphony Orchestra. There was a great deal of bed-hopping, the odd pregnancy and post-war divorces. All that was much easier for the civilians who lived in outlying villages; we had to make do with the Wrens in whatever nest we could find.
Despite the presence of the Japanese Army air sub-section in the Bletchley Park military section, a separate Japanese air sub-section was set up as part of Hut 10, the RAF section, in October 1943, following the Italian surrender. It dealt not with the high-grade four-figure codes like 3366 but with lower-grade three-figure material and collation of both army and naval air intelligence. The section was made up largely of the old Italian air sub-section and headed by Joe Hooper, one of the young graduates who had been recruited by GC&CS in the years immediately prior to the war and the move to Bletchley. Hooper had been given a watching brief over Japanese air codes and ciphers, in addition to his responsibilities as head of the Italian air section, ten months earlier.
Margaret Daly, a twenty-year-old WAAF sergeant, was moved from the German air section to be his PA. ‘I typed his reports out for him. I remember typing out this great order of battle chart of the Japanese Army Air Force. Joe used to walk up and down dictating. He had tremendous mental control and never hesitated. He was a marvellous boss, very considerate and very good with people. He must have been around thirty and married but he was very popular with all the girls. There were quite a few of them swooning over him.’
Hooper's deputy was Brin Newton-John, an RAF officer whose daughter Olivia would later become a famous pop star. Newton-John was transferred to the Japanese operation from Hut 3, the section dealing with intelligence gathered from the German Enigma ciphers. The German influence led to the intelligence-reporting process being known as ‘melding’, from melden, the German verb ‘to report’. The intelligence reporters were known as ‘melders’.
The melders were split into two separate watches. The ‘J’ Watch handled the current material decoded at Bletchley Park while the ‘M’ Watch processed the less immediate material coming in from the out-stations and Washington, carefully checking it against an index of previously produced material to extract every scrap of new intelligence.
Margaret Robertson was just twenty and two years into a history course at Aberdeen University when she was called up, sent as a Foreign Office civilian to Bletchley Park and put to work on the ‘M’ Watch.
I was called a melder, which meant I was given a pile of already decoded messages, scraps of this and that. If they showed any Japanese troops had moved, I had to write it up and hand it over. I was straight from university and for a while I had a WAAF officer called Margaret Pellow looking after me, checking I had got it right. We did it day after day after day. We didn't ever feel there was anything of moment because the messages were very much out of date. They were just trying to keep up with where units were.
The melders researched the background of any information that came up in their reports in a large index which contained details extracted from all previous messages, recalled Ann Lavell, a WAAF indexer.
We used to get the decrypted messages and index every minute detail therein so that a higher form of life than us, the melders, could consult the cards and decide with their help whether to include any given message they had to deal with in their intelligence reports. There were about nine or ten of us, some of us WAAFs and some civilians. We lived in Block F, right down at the bottom of the main corridor – the Burma Road – and right down at the bottom of the last spur on the left – so we were in the very corner of the whole BP perimeter, the boundary fence dividing us from what was then a stud farm. Thereby hangs a tale, quite a drama in our humdrum lives.
One evening one of the WAAFs put her in-tray (choc-a-bloc with top-secret documents waiting to be indexed) on the window ledge and went off to dinner. While she was away somebody did the blackout but failed to close the window, and a considerable wind sprang up and whisked away most of the contents of the basket, whirling some of them off over the perimeter fence, and some of them all over the place inside the fence. Major panic ensued – we were all conscripted to charge outside and retrieve what we could (lucky it wasn't raining), some go
ing right outside into the stud farm and combing the field there.
When we had retrieved what appeared to be all there were, the question remained – had everything been found? The lost documents were all items in three or four series of reports, all of course numbered; so the only way to check that all had come safely home was to locate all the numbers in all the series. That was a huge task because items could be almost anywhere in the section – in various baskets, in the index, and in the melders’ room – and it took Barbara de Grey, Nigel de Grey's daughter who was head of the section – together with the unfortunate WAAF who was the main cause of it (tearfully fearing that she would be posted away from BP in disgrace), and a handful of helpers – the whole night to satisfy the demands of security. I'm afraid I didn't join that stage of the fun, but slunk off home to bed in the RAF camp which adjoined the park.
20
DEFEAT INTO VICTORY
The increased value of the codebreakers’ efforts to the campaign in Burma led to serious concerns at Bletchley Park over the security of the material emanating from its outposts on the Indian sub-continent. Signals intelligence had produced little of any interest to the military in India until the middle of 1943 and the need for total security was therefore not properly understood among some of those receiving the Ultra reports. There were also concerns over the large numbers of Indian staff employed. Many of the Indian servicemen and women working as intercept operators were excellent. But substantial numbers of their frontline colleagues had deserted to the Japanese-backed Indian National Army and some of the locally employed staff clearly had no understanding of the need for secrecy.
‘There was a fruit stall outside the secure perimeter at the WEC,’ recalled Michael Kerry. ‘One day to our amazement his fruit was all wrapped in pieces of secret documents. The Indian who was supposed to destroy the confidential waste had decided it would make more sense to give it to the fruit wallah to wrap his fruit.’
The Emperor's Codes Page 29