Headquarters of the joint U.S.-Australian-British Central Bureau codebreaking organization at 21 Henry Street, Brisbane. Australian War Memorial negative number P0125/05/01
Members of the American contingent to Central Bureau in September 1943. Seated in the front row from right to left are Joe Richard, who broke the Water Transport Code, known as 2468; Zach Halpin, the head of the IBM tabulating machine operation; Larry Clark, the veteran of Rowlett's Purple Section who advised Richard during the break into 2468; Colonel Abe Sinkov, the head of the bureau's codebreakers; and Hugh S. Erskine, who was in charge of the translation branch. Joseph E. Richard
Joe Richard after receiving the Legion of Merit for his work breaking the Water Transport Code. Joseph E. Richard
Wilfrid Noyce, the British cryptographer based at the Wireless Experimental Centre, Delhi, who broke the Water Transport Code at the same time as Richard. Noyce, a leading mountaineer, is pictured during the preparations for the first successful ascent of Everest in 1953.
Central Bureau tents on the Eagle Farm racecourse at Ascot Park, Brisbane.
The Monterey flats in Melbourne, which were originally acquisitioned to house Eric Nave's Special Intelligence Bureau and were eventually taken over by Joe Fabian's U.S. Navy FRUMEL operation. National Archives, Washington, DC
Captain Joe Rochefort, the eccentric U.S. codebreaker sacked as head of the U.S. Navy's Pearl Harbor cryptanalysis centre for proving his superiors wrong. National Security Agency
Members of the Women's Royal Australian Naval Service (WRANS) working in the General Index at FRUMEL. National Archives, Washington, DC
The WRANS barracks at Fleet Radio Unit, Melbourne (FRUMEL). National Archives, Washington, DC
Three WAAAF Japanese intercept operators at the rear of the Central Bureau, Henry Street headquarters. Sergeant Joy Linnane is centre. Australian War Memorial, negative number P0123/09/08
(Front row, right to left) Lieutenant-Colonel Norman Webb, senior British officer at Central Bureau; Colonel Mik Sandford, head of the Australian contingent; and Wing-Commander Roy Booth, senior RAAF officer, pictured during a liaison visit to WEC Delhi. Australian War Memorial,
The headquarters of No. 1 Wireless Unit, RAAF, at Stuart Creek, ten miles west of Townsville, Queensland. The building looks like a farmhouse, but only the tanks and steps are real; the rest of the features are painted on to the reinforced concrete walls. Australian War Memorial, negative number P1443/71/03
The entrance to the HMS Anderson Royal Navy intercept and codebreaking base near Colombo, Ceylon. © Public Record Office, HW4/3
Peter Budd and other Royal Navy wireless operators inside their ‘banda’ at HMS Anderson. Peter Budd
The makeshift canteen for ‘C’ (Indian) Special Wireless Group in Meiktila, Burma. Dennis Underwood
The Central Bureau forward operations base at San Miguel, in the Philippines. Australian War Memorial, negative number P1443/71/58
The first twelve British army codebreakers sent to Central Bureau, April 1944. (Back row, left to right) Brian Warmington, Cyril James, Rupert Fenn, Michael Webster, Bernard Billingham, Bennie Polack. (Seated, left to right) Donald Fletcher, Peter Hall, Barry Smallman, John Smart, Hugh Melinsky. (Front) Ray Eddols. Hugh Melinsky
Royal Navy wireless operator Peter Budd controlling a direction-finding site. Peter Budd
Vice-Admiral Lord Louis Mountbatten, Supreme Allied Commander South-East Asia, celebrating with members of the Special Liaison Unit in Kandy after the victory over the Japanese. Ken Kelsey is holding a glass, to Mountbatten's immediate left. Ken Kelsey
The Bletchley Park liaison team in Washington, 1945. (Left to right): Peter Laslett, Barron Chalkley, William Bodsworth, Philip Howse, Kevin O'Neil and Molly Darby. National Archives, Washington, DC
The next thing the codebreakers would look for would be the inevitable code group indicating that the use of the Chinese Telegraphic Code, the maru code or figures had come to an end. This would give them another additive group to be subtracted from all the enciphered groups in each column, providing yet more basic code groups, and so the process continued.
Figures were often easy to dig out, particularly in sighting reports which gave complete longitude and latitude for each ship spotted. Since the reporting station was routinely located by the direction-finders, the codebreakers knew the first two figures of the longitude and the latitude would have to be the same as that of the reporting vessel, while the last figure was invariably a zero. This recovered further additive groups that could be subtracted up the column, filling in more gaps.
By now the worksheet would have begun to resemble a massive crossword and a Japanese linguist could begin to suggest other potential basic code groups, providing more additive groups to be subtracted from other five-figure groups in the columns. This process of uncovering the meanings of new JN25 code groups was known as ‘penetration’, or rebuilding the book. ‘The production of unreciphered groups leads to the last stage of cryptography which consists of “penetration” into the underlying book,’ said Barham. ‘Each group is listed and the messages in which it occurred noted. From the varied contexts, guesses are made at its significance. It may easily be seen that the process of “book-building” is long and laboured and the speed with which it can be done varies with knowledge of types of messages and the number of complete messages which can be stripped from the reciphering table.’
As the codebreakers laboured away trying to break into JN25b, everything depended on the Japanese not changing the codebook or the cipher table, John MacInnes recalled. ‘As the life of the cipher table [introduced at the beginning of December 1941] was extended, so more and more readable messages became available. The table remained in force for nearly six months. The book-building was delayed at first by much new jargon unknown in peacetime but, as regards units, was much helped by the possession of a library of messages going back to early 1941, so covering a period when the call signs were well identified.’
Among the first messages the codebreakers were able to read, if only in part, were a number which indicated a direct Japanese threat to British assets in the Indian Ocean. The first, on 3 March 1942, was the news that the Japanese Navy was to base five I-boat submarines at Penang in Malaya to operate throughout the Indian Ocean. A few weeks later they began intercepting messages indicating a threat even closer to home.
The decoded JN25 messages had revealed that the Japanese were using letters as code for key areas and places. By mid-March, the codebreakers began noting repeated references to an operation by a Japanese carrier force, accompanied by another force, in the area of ‘D’ with an air raid planned for 2 April on something coded ‘DG’.
George Curnock, one of the most experienced of those working on JN25, would later recount how on the afternoon of 28 March they were working on messages that were talking about plans for a major attack on ‘DG’. Amid confusion over the location, one of the Japanese operators spelled out the name of the target in kana phonetics: KO-RO-N-BO. Curnock said an electric shock ran through what had been until that moment a very relaxed office. The Japanese were planning a major raid on Ceylon, with Colombo itself as the main target.
Admiral Sir James Somerville, the Commander-in-Chief of the Royal Navy's Eastern Fleet, summoned Shaw and asked him if the codebreakers were certain that ‘DG’ was Colombo. Told that they had not a shadow of doubt, he withdrew his fleet to ‘Port T’, its secret Indian Ocean hideaway at Addu Atoll in the Maldives, and sent the merchant shipping in Colombo Harbour to the southwestern Indian port of Cochin.
When 2 and then 3 April came without any sign of a Japanese attack, the codebreakers’ stock slumped, Shaw recalled. On Saturday 4 April a lone Catalina flying boat on patrol over the Bay of Bengal spotted ships of the Japanese Navy's 1st Air Fleet, commanded by Admiral Nagumo Chuichi, who had led the attack on Pearl Harbor. The Catalina got off a short radio message before being shot down. Then the codebreakers began to pick up Japanese air-to-ground messages, indic
ating aircraft were within 500 miles range. The plain-language messages made it clear that Nagumo had merely delayed his attack until Easter Sunday, when the Royal Navy might be less alert. But Somerville had already decided the codebreakers were wrong. The cruisers HMS Cornwall and HMS Dorsetshire, which were to escort an Australian troop convoy on its way back home, had been sent back to Colombo, while the aircraft carrier HMS Hermes and her destroyer HMAS Vampire were dispatched to Trincomalee. As the realization dawned that the codebreakers were right, the Cornwall and the Dorsetshire were ordered to leave Colombo immediately and Somerville set sail with his fastest ships immediately in the vain hope of intercepting the Japanese fleet.
Lillie Feeney was having her hair permed when the Japanese aircraft attacked. ‘I was in the hairdresser's and someone said the Japs were bombing and I thought: Well, yes. I knew all about that. I remember being quite pleased with myself at the time because we had been reading all the Japanese signals about it.’
The codebreakers’ warning had prevented British losses from being as heavy as they might have been. There was an exceptionally heavy mist with visibility restricted to about 200 or 300 yards, but the Japanese naval bombers still managed to sink the destroyer HMS Tenedos and the armed merchant cruiser HMS Hector while two other ships were damaged. RAF fighters shot down nineteen enemy aircraft and a further five were brought down by anti-aircraft fire. However, British air losses were just as high.
Somerville's scepticism that the codebreakers’ prediction would be fulfilled was punished a few hours later when the Dorsetshire and the Cornwall were caught out in open seas as they raced to link up with the rest of the Eastern Fleet and were sunk.
That evening, a JN25 message gave the movements of the Japanese carrier force for the next day. There were no additive recoveries covering the page given in the message indicator, but within a few hours the British codebreakers had worked them out and were able to pass the details of Nagumo's plans to HQ Eastern Fleet.
Unfortunately, the intelligence report passed on to Somerville was garbled and he failed to find Nagumo's force. Three days later, a signal from the Japanese flagship Akagi was located by the code-breakers’ direction-finding network. It appeared to be preparing an attack on Trincomalee. The Hermes and the Vampire immediately put to sea, leaving very little for the Japanese to attack in the harbour. Damage was slight and the Japanese lost a further twenty-four aircraft. But as the remnants turned away they spotted Hermes and her escort hugging the coastline and heading south.
The codebreakers heard the Japanese aircraft report the sighting and passed it immediately via scrambler to Eastern Fleet Headquarters but it was too late. They listened helplessly to the Japanese aircraft reporting the sinking of Hermes and Vampire. Had the codebreakers been fully trusted, the Hermes, the Vampire, the Cornwall and the Dorsetshire would probably not have been sunk. But the belated acceptance that they knew what they were talking about had dramatically cut British losses and Nagumo had received his first setback, however minor, in four months of rampage throughout the Pacific and Indian oceans.
For the first time since moving from Singapore, the wireless operators had been close enough to their target Japanese stations to pick up the messages without difficulty, said MacInnes. ‘They intercepted every transmission from the Commander-in-Chief of the raiding force. The messages were immediately available to the translators and translations were made after a very short interval from first transmission. That no great victory, but only avoidance of defeat, could be claimed by sigint was due to the material inadequacy of the British forces, by sea and air alike.’
11
MIDWAY: THE BATTLE THAT TURNED THE TIDE
The Japanese raid on Ceylon and the growing concern that India itself might come under attack persuaded the British to withdraw the Eastern Fleet out of range of the Japanese. On 25 April Somerville sailed for Mombasa, in Kenya, taking the codebreakers with him.
‘It was one of the biggest thrills of our service to be part of this fleet,’ said Joan Sprinks.
We embarked in the AMC Alaunia, surrounded by many ships of the Eastern Fleet: HMS Warspite, the flagship; the carriers HMS Indomitable and HMS Formidable; the cruisers HMS Emerald and HMS Newcastle; and many others. And as the voyage proceeded we took part in many naval manoeuvres. On deck one day, when the rum ration was being issued, we asked: ‘What about us, don't we get any?’ ‘No love,’ the Master-at-Arms said. ‘You are on the ship's books as boys and while you're in the tropics you'll get your share of lime juice, same as the young lads.’ We arrived in Kilindini Harbour, Mombasa, on 3 May 1942 to an accompanying welcome of wolf-whistles from HMS Royal Sovereign and were quartered in a small hotel in Mombasa – the Lotus.
The local authorities insisted that those members of the female clerical staff who were accompanied by their children, most of whom were codebreakers’ wives, must be sent up-country. Harry Shaw said the women were essential to the operation but the authorities, led by the District Commissioner, demanded that they be removed. The problem was solved by the women themselves, who refused to leave the Lotus Hotel, despite all attempts by the local authorities to have them evicted.
Meanwhile, Shaw and his officers had selected an Indian boys’ school at Alidina, overlooking the Indian Ocean about a mile outside Mombasa. The formalities of requisitioning the school and arranging for an armed guard from the King's African Rifles took some time, Shaw recalled. ‘It was a week before the Temporary Women Assistants could be brought in, during which time their all-day presence in the centre of town did nothing towards alleviating local resentment.’
It was scarcely a promising start, made worse by poor reception conditions and lack of equipment. One antenna mast had to be borrowed from a local RAF station, Shaw said. ‘Frequently there were periods of “black-out” and no messages were received for hours.’
The civilian wireless operators had been left behind in Colombo with two codebreakers to attempt to keep continuity on the JN25 operation. With the reduced reception and half the number of operators, the codebreakers were receiving only a quarter of the JN25 messages they had done previously. ‘Our watchroom was constantly plagued by bats and masses of flying insects, including praying mantises,’ recalled Joan Sprinks. ‘We cursed the Arab dhows that came in from the sea beating drums so loud that we couldn't read signals at times.’
For the British codebreakers, their work on JN25 already badly affected by the original move from Singapore, the transfer to Mombasa was a disaster, recalled John MacInnes. ‘The moves from Singapore to Colombo, and Colombo to Kilindini, followed by the miserable volume of traffic which was intercepted there, caused an almost complete collapse in this field of work. When efforts were resumed the leeway was too great. Signals frequently took up to a fortnight to be enciphered, transmitted and deciphered. There were several periods when even the task of absorbing additives from elsewhere was almost overwhelming.’
Despite the setbacks suffered by the British, the codebreakers left behind in Colombo were still able to make a major contribution to what was to become a vital stage of the war. A JN25 message decrypted at Colombo in early March had revealed that the Japanese had sent two aircraft carriers, the Soryu and the Hiryu, from the Indian Ocean to the naval base at Truk in the Caroline Islands. Over the following weeks the British and American codebreakers began to pick up repeated references to Japanese naval movements indicating a drive southwards towards Australia. As early as 3 April Hawaii intercepted a message indicating an offensive to be mounted from Rabaul, on the northern tip of New Britain, and a few days later OP-20-G reported that air-search patterns indicated an interest in the Coral Sea, off the north-east coast of Australia.
At the US codebreaking base in Pearl Harbor, Joe Rochefort became convinced that the Japanese were planning an operation against Port Moresby, on the south-eastern coast of New Guinea, and the neighbouring Solomon Islands. This would give them a base from which to mount attacks on Australia and threaten America's links w
ith MacArthur's headquarters in Melbourne. His predictions were confirmed on 9 April when a JN25 decrypt revealed the existence of an ‘Operation MO’ strike force, based at Truk, and an RZP occupation force, to be launched from Rabaul.
The Emperor's Codes Page 16