Slim's Special Wireless Group now moved down to Rangoon by road with his headquarters, Underwood recalled.
We had an overnight stop when a small party of escaping Japanese went through the middle of our temporary camp. I suppose they came upon us by accident and decided their safest way was through as we couldn't fire on them because our own people were just the other side.
In Rangoon we occupied houses that would have been the homes of Westerners before the war and life was much better. We even had running water and crude showers. But with the arrival of the monsoon season life became much grimmer; constant damp, always sweaty, plagued with prickly heat, message pads sticking to hands and wrists. The lightning would blow the fuses in the antennae feeds so that we lost our stations – better than having the sets burned out, though.
Standard dress was more or less anything we fancied. Our warrant officer preferred drawers cellular, jungle green, fastened with a safety pin at the front, with sandals made locally from old vehicle tyres. Most of us wore something similar. The Quartermaster ran out of plimsolls as we wore them out so quickly playing badminton on an old tennis court and we had to play barefoot. We swam in a proper pool at what had been a club for Europeans, called, I think, the Konkine Club. Here the water was rarely changed and was a pea-green colour, as well as being very warm, so we found another natural pool and used that. Leeches were a problem but the free cigarette ration came in very handy again for getting rid of them by touching them with the hot end to make them release their grip.
We obtained a lot of our food locally by exchanging army rations for chickens, ducks etc. (no doubt strictly illegal) and we also purchased salad items from a nearby Chinese market gardener. Some time after we reached Rangoon we managed to travel on leave to Ootacamund in the Nilgiri Hills in southern India. This was accomplished by a typical service wangle as strictly speaking no leave was being allowed. The Adjutant arranged a posting to a sister unit outside Calcutta and on reporting there we were given our leave passes and travel documents.
As there was no normal route out of Burma we left by going to the airstrip and hitching a lift out. We had hoped that someone might be going to Calcutta but we could only manage to find an aircraft returning empty to Chittagong after bringing in supplies to Rangoon. On arrival we reported to the local railway transport officer and moved into a transit camp. It was two or three days before we got places on a train to Calcutta and I remember that one evening we went to a sort of circus performance (in a ring but with no animals): there were acrobats, jugglers, dancers and a clown.
Eventually we reached Calcutta and another transit camp to await a place on a train to Madras and then on to Ootacamund. This was a tented camp on what was (and is again, I think) the racecourse. We had to conserve our money but managed a visit to the Lighthouse Cinema (popular because it was air-conditioned) and declined the attentions of young Indians who wanted to provide us with their sisters. Not that we weren't interested, but the fear of disease kept us on the straight and narrow.
21
MACARTHUR RETURNS
By the beginning of 1944, army and air force codebreakers at Bletchley Park, Brisbane and Delhi were able to read the mainline army air and water transport systems without difficulty. But, despite David Mead's break into the indicator system of the Japanese Army Administrative Code, Arlington Hall had suffered a setback when the Japanese introduced a new codebook and false addition squares like those for the Japanese military attaché code.
But help was at hand. The Japanese forces in northern New Guinea were now pulling troops back to defend Madang, the next major base facing MacArthur. When the 20th Division withdrew from Sio, it was unable to take all its codebooks with it so they had to be destroyed. The division's chief signals officer should have ensured they were burned. But they were dumped in a metal chest in a deep water-filled pit. When US Army engineers began searching the area with mine detectors, they found the metal crate and demolitions experts were called in, only to find that it was full of codebooks.
The captured codebooks were for the Army General Administrative System, 7890, and its derivative systems: the Army Ordnance System, 2345; the North Pacific System, 5555; the Southwest Pacific System, 7777; and the old Philippines System, 6666, which was now used by all cut-off units in an attempt to protect the other systems against being compromised by capture. The codebooks were sent immediately to Central Bureau, Joe Richard recalled.
I remember when the chest arrived at our hut in Ascot Park, Brisbane. We had just moved from the Henry Street House to the fenced-in small park, just across the suburban railway from the Eagle Farm racetrack where our enlisted men were now in wooden-floored tents. Our hut, Hut 7, had two new kitchen stoves for making tea, an important ritual in Australia. Colonel Sinkov and some other officers immediately took the captured codebooks and started to dry them out in the ovens of these stoves. Everyone seemed to be busy drying additive books. I had one. It was sopping wet. I had it on the edge of a table with a desk fan blowing on it. I and others were taking turns holding a small one-bar electric heater close to it. As I recall there was more than one copy of each book so we could package one copy of everything and send them by officer courier to Arlington Hall.
The Allied codebreakers were now able to read all the mainline Japanese Army systems – in the case of the 6666 code until the end of the war – and the Japanese seemed unaware of the ‘pinch’. The codebooks were minus their covers which under Japanese signals security regulations were supposed to be handed in as ‘proof of burning’ and Central Bureau even monitored the 20th Division's chief signals officer insisting that all its codebooks had been completely destroyed. The bureau's translators were overwhelmed with deciphered messages and MacArthur appealed personally to Washington on their behalf for assistance. Amid the beginnings of a rapprochement between the US Army and Navy codebreakers, two of FRUMEL's top Japanese linguists, Lieutenant-Commanders Forrest R. ‘Tex’ Biard and Tom Mackie, were sent to Brisbane to help translate the backlog. Biard said:
The first morning at work remains one of the most memorable events of my entire life. As I remember it, the Central Bureau was on the outskirts of Brisbane in a wooded section into which temporary screened wood huts had been placed. According to one post-war author, it had been the location of a rather famous bordello that was run by an equally famous madam. As to the authenticity of this tale I cannot comment. But it was nice to have our hut and the surrounding area well shaded by the heavy growth of eucalyptus trees, for the summer sun in Brisbane is quite hot. The bureau was well fenced and quite adequately guarded. We were escorted to the codebreaking section in what I will call the ‘hut’, which was one of only three buildings I entered while working at the bureau. The other two were the mess hall and the latrine. We were not introduced to anyone outside the group of fifteen or so working directly with us. All our new cohorts were very pleasant and most co-operative and we truly liked them, but they needed help very quickly and badly.
The US Army codebreakers had been working their way through the pile of decoded messages in date–time order. Biard and Mackie decided it would be more productive to work from the top, looking for the most interesting and current messages.
Tom and I took the top two messages and started to read. Neither of us said a word for a minute or two but very soon I leaned over to see what Tom had, and he leaned my way to see what I was reading. Both of us had eyes as big as mill wheels. It was almost unbelievable. Mackie and I stumbled all over the words in our haste to tell our hosts that we had just read two parts of a thirteen-part message that was hotter and more explosive than Vesuvius. The message was the work of one ‘Staff Officer Izumi’. He was reporting to his area commander on the decisions being made at a conference of high-level officers of both the Imperial Army and the Imperial Navy on the general situation in the New Guinea, Bismarck Archipelago, Admiralty Islands and Solomon Islands areas. The purpose of this conference was to decide what could be done about the ever-worseni
ng situation for them in these regions to the north and north-east of Australia.
The thirteen-part message, dated 19 January 1944, gave precise details of the Japanese plans for the region. They were reinforcing Wewak and the Madang–Hansa Bay area in anticipation of MacArthur's next move forward but defences at Hollandia further west were weak. Other messages and traffic analysis of Japanese radio communications allowed Central Bureau to provide MacArthur with detailed information on the Japanese order of battle along the northern New Guinea coast. He decided on the strength of this new information to drop plans to work his way along the coast and to concentrate his attack on the defensively weak Hollandia, isolating the Japanese at Wewak and Madang and allowing them to be picked off later.
Meanwhile, Nimitz was leading a far-reaching campaign to recapture the islands of the central Pacific. Operation Galvanic had begun in November 1943 with the seizure of the Gilberts and was followed by the invasion of the Marshall Islands in late January and Eniwetok three weeks later. Informed by the US Navy codebreakers of a buildup of ships and aircraft of the Combined Fleet at Truk, the main Japanese naval base in the Caroline Islands, Admiral Marc Mitscher's Fast Carrier Force swept ahead of the advance to carry out a Pearl Harbor-style attack. Admiral Koga Mineichi, Yamamoto's successor, was warned of the impending attack by his own signals intelligence units and managed to get most of his ships out of Truk. But fifty merchant vessels and more than three hundred aircraft remained when the US force arrived on 17 February. It destroyed 200,000 tons of Japanese shipping; some 200 aircraft; plus 2 cruisers and 3 destroyers that had failed to flee in time. But more importantly, Truk was no longer a safe haven for the Japanese Navy which withdrew behind the protection of the Marianas to the Philippines Sea.
As a result of the codebooks found at Sio, Central Bureau had built up a good picture of the Japanese defences along the New Guinea coast from their own forward base at Saidor to Wakde. The RAAF's No. 1 Wireless Unit had moved forward from Port Moresby to Nadzab in the Markham Valley in February 1944 from where it monitored a steady build-up of enemy aircraft being ferried into Hollandia. Although the Japanese air base was relatively poorly defended, the Japanese were keeping the bulk of their aircraft there because it was outside the known range of the US P-38 Lightning fighters that would have to escort any bombing raid. But General Kenney's Lightnings, unbeknown to the Japanese, had been modified with the addition of wing tanks that would extend their range beyond Hollandia. Beginning at the end of March and continuing for the next three weeks, his 5th Air Force carried out a series of raids that destroyed the vast bulk of the Japanese aircraft, many of them on the ground.
Despite the sustained bombardment of Hollandia, the messages intercepted by Central Bureau showed that the Japanese still saw the Hansa Bay–Madang area as the most likely focus of MacArthur's next ground attack. Secure in this knowledge, he landed two divisions at Hollandia, at the same time seizing the airstrip at Aitape in order to provide continued air support. The move was a brilliant use of Ultra that left an entire Japanese army totally encircled at the cost of very few casualties.
Signs that the Japanese were attempting to reinforce their beleaguered New Guinea garrisons came initially in naval intercepts referring to the imminent formation of a large convoy and the anticipated arrival of an infantry division. These were hardened up by water transport messages decoded at Central Bureau which announced that two infantry divisions, the 32nd and 35th, were on their way to New Guinea via the Philippines aboard the ‘TAKE’ convoy.
The submarine USS Jack intercepted the convoy off Luzon in late April, sinking one transport vessel. As the convoy moved towards New Guinea there was a plethora of messages in both naval and army codes describing its route and make-up. At the beginning of May, Central Bureau decoded water transport messages indicating that there were 12,784 troops in nine transport vessels, escorted by seven other ships. The messages gave precise speeds and directions for the route, revealed that the convoy was due to split into two parts on 7 May north of Halmahera, and provided the scheduled positions for noon on 2 and 9 May. The day before the convoy was due to separate, the submarine USS Gurnard waited for it in the Celebes Sea 100 miles off Menado. In the space of ten minutes three troop transports were sunk and, despite the rescue by other ships of a large number of troops, 4,000 had drowned.
All the time the codebreakers were looking for pinches that would help them in their work. When Central Bureau intercepted a message reporting that the Yoshimo maru, a Japanese merchantman sunk in Aitape Harbour, had an additive book for the Army General Purpose Systems on board, General Akin sent down divers to recover the cipher book. It was found in a metal box lodged behind a ladder in the burned-out ship. The whole operation was supervised personally by Akin, Joe Richard recalled.
He requisitioned an aircraft and had the box flown back to Brisbane. When opened it was found to contain an additive book we didn't have but in a very charred condition and with the number groups un-readable. Our photographic officer, a graduate chemist called Robert Holmes, devised a method for reading the numbers. He took a surgeon's scalpel and flaked off a page. Then he laid it on some cotton padding on a table and sat two soldiers alongside each other. One dipped a cotton bud in neat alcohol and then gently touched two groups on the additive page with it, making them instantly readable for long enough for the other soldier to copy them down. Since there were 50,000 additives on 500 pages, plus the page co-ordinates, you can see this was a lot of work.
In the wake of his capture of Hollandia, the islands of Wakde, Biak and Noemfor soon fell to MacArthur's troops. Meanwhile, 55 Australian Wireless Section, which had moved to Finschhafen in February 1944, followed the fighting to Hollandia, from where it intercepted a large number of reports about the 18th Army's preparations for a counter-attack. Using the newly captured additive groups, Central Bureau deciphered detailed messages about enemy intentions. The 18th Army was marching through the jungle from Madang towards Aitape, intent on recapturing it. The codebreakers read operational orders in the mainline Army General Administrative Code giving full details of the Japanese plans. The attack was repelled and the remnants of the Japanese forces were pushed back into the jungle where with neither supplies nor reinforcements their numbers steadily dwindled.
During the course of 1944 Bletchley Park reinforced the British contingent at Central Bureau with a number of young army and RAF officers who had passed through Captain Tuck's six-month Japanese course. Hugh Melinsky, who had been recruited on Tiltman's behalf by his tutor at Christ's College, Cambridge, arrived in April 1944 and was put to work on the naval air desk, part of Eric Nave's air–ground section
What Captain Nave did not know about codebreaking was not worth knowing. He had a sixth sense which enabled him to sniff out a meaning in what looked to me like a jumble of letters or numbers. He gave me a collection of messages with some words already translated and told me to do the rest. I learned the hard way, for six weeks. Captain Nave received messages from several wireless units. This was useful because if one version of a message contained gaps or faults another version might put these right. The greater the quantity of material available the better was the chance of breaking the code.
Apart from the cribs, the most helpful tool to the army code-breakers was the use of stereotyped reports, Melinsky recalled.
Many messages were routine and followed the same pattern; for example, weather reports gave the place and time of origin, the general weather, the temperature, the amount of cloud, the wind direction and speed and perhaps the further outlook. In addition, the uncoded messages or traffic from or to the aircraft might very well give a clue to one or several groups in a coded message.
Brisbane was not all work. There was a club for servicemen where the lady in charge arranged for my friend Donald Fletcher and me to meet an Australian family at the cinema on a Saturday evening, and they took us home with them to spend the night. We slept on the veranda under mosquito-nets. The next morning their
two daughters and son, about our age, took us for a long walk through the bush despite a temperature of ninety degrees. In the afternoon we went for a car ride to some beautiful spots and tasted strange fruit like avocados, custard apples and passion fruit. In the evening we were joined by two American friends of theirs and had a grand party. The parents took us back to the camp rather late.
There were so many kind and hospitable families that I could have gone out almost every evening. They loved to hear about England, the bombing and the rationing, and they offered to send food parcels to our parents by post. There were also concerts, a ballet and dances, though army boots were unkind to one's partner's toes. I had to go and buy a pair of shoes, which were not rationed as in England.
Claude Lanacaster Jr, a US Army codebreaker from North Carolina, flew out on the same flight from San Francisco as another group of British Army codebreakers.
I was twenty years old when I was called to active duty. I had enlisted in the Signal Corps in 1942 after a year and a half at University of North Carolina at Chapel Hill. I took my training at Vint Hill Farms, about thirty-five miles from Washington, DC. We were there for about six months and then one group was sent to England and the other group went to Australia.
At Central Bureau we worked in a frame building about 40 foot × 40 foot, like one of the huts at Bletchley Park. It was divided into three sections; one section had Captain Nave and the Australian Army's Captain Jury, who was the oldest in the building and a former teacher at the University of Adelaide. There were also two young British Japanese translators, Geoff Spencer and Hugh Melinsky.
The Emperor's Codes Page 31