Battle in Asia and the Pacific, 1941–42
As the Japanese swept south, US personnel and resources were starting to arrive in Australia, with the first 4600 US soldiers landing in Brisbane on 22 December 1941. In early 1942, as the situation turned from bad to worse, the Australian government began to panic and even considered evacuation of Canberra to Wagga Wagga. Military ‘refugees’ began to pour in from Singapore, the NEI and the Philippines, and among those refugees were the men from the SIGINT intercept sites in Singapore and Corregidor, in the Philippines.
On 20 January, the prime ministers of Australia, Canada and New Zealand received a telegram from the Secretary of State for Dominion Affairs in London, Lord Cranborne, informing them that the entire Naval Intelligence Staff, SIGINT organisation and supporting elements at Singapore had been moved to Colombo in Ceylon.1 Admiral Sir Geoffrey Layton, RN, the C-in-C, Eastern Fleet had ordered this action on 1 January in what was a clear signal of impending defeat in Malaya.
On the same day as the news of FECB’s evacuation from Singapore was communicated, news arrived from the Australian naval attaché at the Legation in Washington concerning another evacuation. This was the US Navy’s CAST unit, which had been evacuated to Java in the NEI. For Australia, this was bad news with a silver lining. The US Navy was looking for a secure base in Australia or New Zealand from which CAST could continue to operate.
The ACNB jumped at the chance and immediately invited CAST to establish operations in Australia. It did not pass the offer on to New Zealand as requested by the naval attaché, but kept the news secret, ensuring CAST would come to Australia.
The US Navy’s offer had come about as the alternative positions for the CAST unit rapidly became untenable. First the Philippines and then Malaya, Singapore and now the NEI all became insecure. Even Colombo would soon be judged too insecure for FECB and its SIGINT elements, and it would be relocated to Kilindini in East Africa, from where it would operate for most of the next four years.
The final act in the movement of CAST to Australia came about as the unit was attempting to establish operations in support of the US Navy in Java. Vice Admiral William Glassford, the US Navy Commander in Java, did not believe the NEI could be held, and advised the Navy Department in Washington to remove CAST to the greater security of Australia.2 The advice from Glassford, and the rapid destruction of the NEI defences in the string of naval battles over January and February 1942, convinced the Navy Department to open discussions with both the British and the Australian naval attachés. In due course, Australia’s attaché informed Australia and CAST’s new home was decided.
The impact of CAST on Australian SIGINT cannot be understated. This single US unit had eight officers and 68 other ranks, comprising Japanese linguists, cryptographers and experienced intercept operators trained in Japanese Morse, along with their equipment.3 In early 1942, it was larger than all of Australia’s SIGINT organisations combined. On 23 February, the CNS quickly cabled the attaché that the US personnel ‘would be a valuable addition to W/T intelligence and the Special Intelligence Organisation that now exists in Australia’.4 This was fast work for Australia.
Although it would not be immediately apparent, CAST was not coming to Australia to supplement Australian SIGINT but to serve the needs of the US Navy and OP-20-G in the Office of Chief of Naval Operations in Washington. Neither was it joining in with the RAN’s existing SIGINT organisation; it was in fact going to launch a hostile takeover, one so complete that it very soon became clear that the ACNB was left with little control over the new entity, called Fleet Radio Unit Melbourne (FRUMEL).5
FRUMEL made use of many RAN personnel, but at the lowest possible levels, and they worked for OP-20-G, not the ACNB. In some ways, it is difficult to say that FRUMEL was an Australian operation; it was Australian only in that it was located in Melbourne and used Australian personnel for much of its grunt work. Admiral Joseph R. Redman at OP-20-G in Washington exercised almost complete control over policy decisions, unit procedures, and all command and control, even down to deciding whether the SIGINT produced by FRUMEL should be shown to anyone outside the US Navy. At best, Australia was FRUMEL’s landlord, but its operations did involve Australians, particularly in the first part of 1942, so it is part of our story.
One of the major mistakes made in historical accounts of the war in the Pacific is to overlook the US response and focus on the losses suffered in Malaya, the Philippines and in South-East Asia more generally. In fact, when it is examined carefully, the US response was extraordinary. It was certainly beyond anything the Japanese thought possible.6 It is somewhat sobering to consider that on 10 May 1942, the very same day General J.M. Wainwright and General W.F. Sharp finalised the surrender of all Filipino and American troops in the Philippines, the US Navy permanently wrested the strategic initiative from Japan in the Battle of the Coral Sea. Within five months, the US Navy had put the Japanese permanently onto the defensive. In less than another month, by the end of 7 June 1942, the US Navy would destroy the strategic reach of the IJN at the Battle of Midway. This was a startling turnaround and one that is too often neglected.
As CAST arrived in Melbourne over March and April 1942, the military situation was critical. Singapore had fallen, the British were in full retreat through Burma to India, and Rear Admiral Karel Doorman’s English, Dutch and American fleet was being systematically destroyed. The extent of the IJN’s victory is shown by the losses inflicted on the Allied navies between 27 February and 1 March 1942—ten ships, including two heavy cruisers, three light cruisers and five destroyers—in the two battles of the Java Sea and the Battle of Sunda Strait.
The CAST advance party led by Lieutenant Rudolph (Rudy) Fabian quickly established operations in Melbourne.7 Intercept operations were commenced immediately and IJN messages were soon being couriered by motorcycle every two hours to the cryptanalytical section based at the Monterey Apartments, in Queens Road, South Melbourne, which the Americans had taken over rather than try to squeeze into Victoria Barracks.8
The Americans were not impressed by the state of the facilities at the RAN’s Moorabbin intercept site, which was slowly being built by Australian workers.9 The pace of work annoyed the Americans so much they built the antenna farm and accommodation facilities themselves.10 They also quickly took control of the inexperienced sailors and members of the Women’s Royal Australian Naval Service (WRANS), who constituted the bulk of the workforce at Moorabbin, and began training them in the use of the American equipment, particularly the KANA typewriters, and in US Navy processing techniques.
The contrast between the attitude of the Australians in Melbourne and the aggressive impatience of Lieutenant Fabian and his unit must have been stark, but polite American pushiness got things moving without ruining relationships. Given their extensive experience of working Japanese naval codes and their recent exploits in the Philippines and Java, the Americans held the upper hand and were easily able to impress the less experienced Australians.
The equipment suite salvaged from Corregidor was, by Australian standards, quite extensive, and included portable DF equipment and the valuable KANA typewriters. They had no IBM card-punching equipment, however, and the PURPLE machine that OP-20-G had sent them needed multiple faults fixed before it could be made to operate with any reliability.11 The CAST unit was lucky to have procured the services of Ralph E. Cook, an electrical engineer and officer in the US Navy Reserve. Cook had been a field engineer for IBM until he was called up for duty in April 1941, and he played a significant role in getting the equipment operational.
Cook began automating as much of the cryptanalytic process as he could, and he kept the IBM equipment running. His main problem was obtaining more IBM equipment, and it would be late 1942 before any more arrived from Washington. Even then, it was missing the card feeder and the printing mechanism for the tabulator. The tabulating section at FRUMEL had to hand-punch cards for three months until the missing equipment arrived.12
Despite all the political, organisational and logistical difficulties affecting FRUMEL at its creation, the men and women, American, British and Australian, continued to carry out their duties and to learn the art and science of SIGINT. It was during these first four months of FRUMEL’s life that it made its most significant contribution to the war effort by providing SIGINT on the IJN’s Indian Ocean raid of March, the Battle of the Coral Sea in May and then the Battle of Midway in June.
As soon as it had arrived in Melbourne, FRUMEL began receiving intercept from the site at Moorabbin and sites at HMAS Coonawarra outside Darwin, Adelaide River, Exmouth in Western Australia and HMAS Harman in Canberra, as well as from SIGINT teams afloat on both US and RAN ships. All of this was supplemented by SIGINT flowing in from US and British sites around the world.13
One of the first SIGINT wins FRUMEL and other SIGINT sites achieved was the preservation of Admiral James Somerville’s Eastern Fleet in April 1942. SIGINT enabled Somerville’s unpractised and unready fleet to avoid Admiral Nagumo’s force, sparing them the destruction that befell Karel Doorman’s fleet.14
Initial indications of the IJN’s intentions to extend operations into the Indian Ocean came from intercepted IJN signals on 3 March, which were orders for setting up a submarine base at Penang in Malaya. On 7 March, further IJN traffic referred to a timetable for an operation against a location identified by the digraph ZL. The British cryptanalysts at Colombo had tentatively identified ZL as the Andaman Islands, as it had been used in IJN traffic relating to setting up a seaplane base there. This indicated the IJN was planning a move into the Indian Ocean. The reasons for the IJN’s move west were not known, but the traffic analysis showed a build-up was occurring. Further evidence was slowly collected as IJN units moved to Malaya, Singapore and the archipelago, and traffic analysis at FRUMEL, the Fleet Radio Unit, Pacific (FRUPAC) in Hawaii and Colombo produced a picture of a substantial concentration of IJN forces.
On 21 March, FECB at Colombo informed Somerville that the IJN was planning a major attack on ‘DG’. By 26 March, Admiral Nimitz, C-in-C, Pacific Fleet (CINCPACFLT) at Pearl Harbor, believed the IJN was going to break out into the Indian Ocean, and he warned Somerville, confirming the intelligence Somerville had been given by FECB. Nagumo’s target, DG, was finally identified from a crib from a confused IJN radio operator who identified DG as KO-RO-N-BO when he spelled it out in KANA.15 Somerville ordered Allied merchant ships to run to India for safety and withdrew his own fleet to Port T, a secret base established at Addu Atoll in the Maldives. This action left the whole Bay of Bengal and Colombo unprotected, but the Admiralty had ordered Somerville to keep his fleet intact, even if it meant the loss of Ceylon.16
The IJN’s Indian Ocean raid resulted in unfortunate losses, including the aircraft carrier HMS Hermes; the two cruisers HMS Cornwall and HMS Derbyshire; and HMAS Vampire, HMS Tenedos and 26 merchant ships. Some of these losses were unnecessary. The Hermes, Cornwall, Derbyshire and Vampire were lost because Somerville thought the SIGINT was wrong when Nagumo’s force did not appear precisely on 1 April as predicted.17 Despite this error, Allied SIGINT now had a firm grip on IJN communications, and this allowed commanders and governments to act on the basis of reliable information. One of the actions that SIGINT made possible was the Doolittle Raid on Tokyo.
This attack has often been described as having been carried out by sixteen B-25 (Mitchell) medium bombers of the US Army’s 17th Bomb Group launched from the aircraft carrier USS Hornet. Although it is true that the actual attacks on Tokyo, Yokohama, Nagoya, Kobe and Osaka were to be carried out by the heavy land-based aircraft launched from Hornet, the attack was made by Task Force 16 (TF16), commanded by Admiral Halsey. TF16 comprised the carriers Hornet and Enterprise, the two heavy cruisers Northampton and Salt Lake City, four destroyers and a re-fuelling tanker.18 An IJN picket boat, the 90-foot long fishing junk, Nitta Maru, sighted and reported the presence of TF16 at approximately 0744 hrs on 18 April and Halsey decided to launch the B-25s immediately. They launched approximately 650 nautical miles from the Japanese coast, around 150 nautical miles further out than planned.
The B-25s reached Japan and took the Japanese government and military by surprise. A Japanese aircraft carrying General Tojo was forced to take evasive action when one of the B-25s passed so close the occupants saw that the pilot was Caucasian. After completing their mission, the B-25s flew on to crash-land in China, or in one case, Russia. The Nitta Maru did not go down without a fight, shooting down one US aircraft that attacked her and exchanging fire with the destroyer USS Nashville. It took 928 6-inch rounds and 1200 rounds of .50-calibre machine gun ammunition to sink her.19
This operation was only possible because the SIGINT flowing into Washington from FRUPAC, FRUMEL and the other SIGINT sites provided the US Navy and President Roosevelt with accurate positional information for the IJN’s fleet units. The Americans knew that neither Nagumo nor the Combined Fleet could intercept the Hornet, and this made the operation feasible.
The military destruction caused by the Doolittle Raid was minuscule, but its political and psychological impact on Japan’s military and government was massive. Admiral Yamamoto vomited when he was told Tokyo had just been bombed.20 It was a stroke of genius. Lieutenant Colonel James (Jimmy) Doolittle’s raid of 18 April 1942 completely negated Japan’s strategy and shattered the confidence of Japan’s senior military leadership that they could hold the Americans at bay.
Doolittle’s raid made it imperative that the IJN regain face by inflicting a serious loss on the Allies. The RN had refused battle in the Indian Ocean; seizing Port Moresby was the next best thing. As the IJN’s plans progressed, FRUMEL in Melbourne and FRUPAC in Hawaii were busy intercepting and breaking out the IJN’s communications. The product from FECB had fallen away due to its withdrawal from Colombo to Kilindini, but FRUMEL and FRUPAC, supported by other SIGINT entities, kept the IJN well covered, and the Japanese plans to seize Port Moresby and to reach further into the Pacific to prevent more Doolittle-type raids were being unwrapped.
Indications of the IJN’s intentions towards Port Moresby began to accumulate as the work of revictualling, rearming and repairing the ships assigned to the task force created more and more traffic. Even the target of the intended operation was easily identified from the intercept of weather messages for the Coral Sea and the area around Papua, New Guinea. In early March, a message in JN25 (the main IJN code), directing the carriers HIJMS Soryu and HIJMS Hiryu to move from the Indian Ocean back to Truk to take part in Operation MO, was intercepted and broken out.21 On 3 April, FRUPAC broke out traffic indicating the IJN planned to launch an operation from Rabaul, and OP-20-G in Washington added an assessment that the IJN had increased reconnaissance of the Coral Sea.22
On the Japanese side, they had no idea what they were up against.23 They were expecting 200 combat aircraft operating from airfields between Townsville and Darwin and the remnants of Somerville’s Fleet, possibly a battleship and up to three cruisers, with some submarines as a screen and maybe even a carrier.24 Their assessment of ground forces was more accurate; Port Moresby was defended by a small garrison made up of the militia of the Australian 30th Brigade, and an artillery regiment armed with field guns last used in 1918. At the time of the Doolittle Raid, the IJN’s Combined Fleet Staff thought the Americans had only two carriers left in the Pacific. Their appreciation of American naval strength was consistently poor and they were having great difficulty identifying US naval strength and locations.25
There were, however, 225 land-based combat aircraft concentrated around the Coral Sea, and the US carriers Lexington and Yorktown in Admiral Frank Fletcher’s Task Force 17, would prove capable of bringing more. In addition, there were strong US submarine forces operating within the intended area of operations, as well as an RAN force.26 In the face of these forces, conducting a successful transit to and amphibious assault on Port Moresby would not be easy. Again, the IJN and IJA were implementing plans that made little use of the limited intelligence that they had available and the IJN was about to pay the price.
On 9 April 1942, FRUMEL intercepted IJN traffic indicating that a major Japanese task force was moving to capture Port Moresby. The evidence for this assessment was found in a JN25 message from Admiral Yamamoto Isoroku requesting a report on the progress of repairs to the carrier HIJMS Kaga. FRUMEL believed Kaga had been damaged off Lombok and was now in dock for repairs. This message also compromised the Port Moresby operation when Yamamoto wanted to know Kaga’s status because she was needed for the attack on RZP,27 which FRUMEL had already identified as Port Moresby.28 The analysts at FRUMEL now increased the intensity of their operations and began identifying the other IJN elements assigned to the RZP campaign, including the new light carrier Shoho, and the 5th Cruiser Division consisting of Nachi, Huguro and Myko.29
On 20 April, FRUMEL and other sites showed the 5th Carrier Division, consisting of Shokaku, Zuikaku and possibly Shoho, en route to Truk carrying approximately 357 combat aircraft.30 By 27 April, FRUMEL was reporting that the 5th Carrier Division, plus Shoho, the 5th Cruiser Division less Nachi, the 29th Destroyer Division and half the 8th Submarine Squadron were all now at Truk, and that two Occupation (assault) Forces, RZP, Moresby and RXB, Tulagi had been formed and allocated call signs in the Truk Area Network. FRUMEL also established that the 5th Carrier Division was now passing messages across the operational net for New Britain Command while apparently sailing to operations in the waters around the Solomon Islands.31
The next day, 28 April, FRUMEL was reporting that the 5th Air Attack Force was conducting offensive air patrols between 80 and 120 degrees up to 700 miles (1200 kilometres) from Rabaul. By 30 April, the existence of the Port Moresby Occupation Force aboard the Fumi Maru and another merchant ship had been confirmed, and their orders to depart Rabaul and rendezvous with the Saipan Base Force off Deboyne Island on ‘X-minus 5’ were reported by FRUMEL.32
If there is a single element in SIGINT more important than the others it is cribs. On the eve of the Battle of the Coral Sea, FRUMEL, FRUPAC and the other sites working against the IJN benefited enormously from the IJN’s inability to distribute new codebook and additive tables in time for their introduction on 1 May 1942. These delays led to the new additive tables being used with the old JN25B codebook, and meant Allied SIGINT could read IJN traffic until the table expired on 25 May. It allowed FRUPAC, FRUMEL and FECB to verify the IJN’s preparations for the assault on Port Moresby and, even more importantly, to decrypt the IJN’s battle plans for the capture of Midway Island.33
The Battle of the Coral Sea occurred over five days, from 4 to 8 May, and its twists and turns are clearly shown in FRUMEL’s records from 0640 hours on 4 May onwards. The first Japanese contact with the opposing naval force was relayed to the IJN Task Force by a Japanese aircraft that reported one Allied battleship, two cruisers, one carrier and seven destroyers. At 1250 hours, the Japanese reporting mentions one enemy battleship being sunk. Further intercepts of air–ground networks on New Britain on 6 May show Chichijima Air Group sighting an enemy task force of one battleship, three carriers and five destroyers 195 degrees and 200 miles (300 kilometres) from Tulagi. A large volume of operational traffic was subsequently intercepted over 7 May, and the battle report on the same day from the 5th Carrier Division’s flag officer described the failure of one attack by dive bombers and fifteen bombers because they ran low on fuel.
The following day, FRUMEL’s record shows the Japanese reporting the sinking of the Saratoga and the striking of Yorktown with three torpedoes and eight bombs over the two days. They also reported that Shokaku had been set on fire but that this was being managed, and that her aircraft had flown over to Zuikaku.34 Most importantly, FRUMEL was able to report that Rabaul had sent a broadcast ordering that the attack on Port Moresby be postponed.35
The battle assessment from FRUMEL on 10 May indicated that Yamamoto had postponed the seizure of Port Moresby by five months and had ordered the 5th Carrier Division back to home waters because of the damage it had suffered. FRUMEL also provided comment on intercept to underline the fact that the Shoho had not been intercepted sending traffic and that her captain was now having his messages routed via Iro, a fleet oiler. This intelligence indicated to the traffic analysts and cryptanalysts that Shoho had been sunk. They had derived this assessment from the previous habit of the Japanese, first seen during the Battle of the Java Sea, to continue sending official messages to the captains of sunken warships using the original ship’s call sign but routed to the ship the surviving captain was now on.36
FRUMEL was also able to report that 22 of Shoho’s aircraft had crash-landed, and that Shokaku had been hit by three torpedoes and eight bombs, and was listing and restricted in speed to alternating between 12 and 16 knots.37
Records show that the combined US and Australian cryptanalytical organisation in Melbourne was contributing its fair share to the intelligence picture of the IJN. It is hard to say how much use the US commanders made of this intelligence, but the fact is they had it at their fingertips.
Despite the indecisive nature of the Battle of the Coral Sea, it provided the Allied SIGINT system with a vast amount of operational traffic that was available for considered analysis and assessment once the battle was over. This provided deep insights into the IJN’s procedures and cryptographic systems, as individual operators, working under the stress of combat, made numerous errors that provided cribs. This follow-up analysis led to the recovery of many code values in the five-digit code and provided a vast trove of information for the traffic analysts at FRUPAC, FRUMEL, FECB, OP-20-G and GC&CS. The battle may not have resulted in a decisive victory for the US Navy, but it blew open the IJN’s communications.38
In the lead-up to and during the Battle of the Coral Sea FRUMEL, FRUPAC and FECB were collecting and reading IJN traffic on another, larger, operation that the IJN intended to launch against US forces in the Pacific. The designator for this target was AF in the Western Pacific. A number of the intercept and cryptanalytical units working the Japanese problem formed the view that AF was most likely Midway Island. Others, particularly OP-20-G in Washington, believed AF was the Aleutian Islands.39 Others thought it could be Port Moresby and some thought it designated a target for a revenge raid on Hawaii or, even more implausibly, a target on the west coast of the continental United States.40 At FRUPAC, the analysts continued to believe that the IJN was planning to attack and seize Midway Island.
It was at this point that FRUMEL helped clear up some of the confusion and thus played a role in saving Midway Island.41 FRUMEL had already provided SIGINT showing that Yamamoto had postponed the attack of Port Moresby until the outcome of another operation had been decided. FRUMEL and other SIGINT sites were also intercepting Japanese traffic that indicated the Japanese were now considering a rapid land assault across the Owen Stanley Range against Port Moresby as the most suitable strategy at this time.42 With a naval assault on Port Moresby moving out of consideration, this left Midway, Hawaii, the Aleutians and the west coast of the continental United States as the potential objectives.
The main evidence that Midway was the target was the continued use of the place designator AF in IJN traffic. Three messages dated 18 May referred to AF as the target. The first message stated that an IJN unit intended to be ‘fifty miles NW of AF’ on the day of the attack.43 The second originated from the Chiba Air Group requesting ammunition for AF to be sent via Imieji.44 The third message, from Combined Fleet, clearly mentioned Midway as a destination for five torpedo craft taken from Jaluit and Wake aboard an unidentified vessel.45 All of these messages were obtained and broken out two weeks ahead of the attack on Midway.
On 18 May, FRUMEL was reporting that the 5th Carrier Division—Shokaku and Zuikaku—were getting replacement aircraft, including four fighters, fifteen bombers and ten torpedo planes in preparation for the campaign.46 On 21 May, FRUMEL intercepted a message from Naval Intelligence in Tokyo to Yamamoto, reporting that ‘AF informed Pearl Harbor that they had only enough water for 2 weeks and asked for immediate resupply’.47 The Americans at FRUPAC were waiting for this message, which they had generated in the hope the Japanese would report it alongside the designator AF. FRUPAC had called the garrison on Midway using the secure telephone cable and asked them to radio in clear for a water resupply. The IJN swallowed it hook, line and sinker.
The following day, FRUMEL added to the growing picture by intercepting and reporting information that Yamamoto wanted the aircraft transport ship Hishin Maru to load twelve or more carrier bombers for the forthcoming campaign, indicating that it might be starting before Zuikaku had completed her refit. On 23 May, FRUMEL reported the IJN had changed its geographic designators for phase one of the intended operation, and in so doing avoided the designators YU, FU, MA and MI.48 They probably thought MI was too obvious and retained AF as the designator for Midway.
By 24 May, 14th Air Group orders to move to Midway Island following its capture were reported by FRUMEL, as were orders to the 2nd Combined Landing Party and 7th Cruiser Division to concentrate at Saipan with 2nd Destroyer Division, one carrier division, and the 6th and 7th Air Squadrons, plus additional 4th Fleet units that were now joining the escorting force.49
The volume of intelligence traffic originated by Jaluit was also noted by FRUMEL as was the fact that, on 25 May, the call signs for the C-in-C, Combined Fleet, and the Commander 2nd Fleet were added to the action address list of Jaluit’s messages, indicating these commanders were now present within Jaluit’s area of responsibility.50 This was traffic analysis at its best.
On 27 May, FRUMEL reported the IJN attack force would arrive off Midway at 1900 hours on 6 June. This information came from a message sent from Saipan informing Yamamoto of the schedules of the 4th, 1st and 11th Air Fleets. Now the Americans had the place, date and time of the IJN’s attack. All they needed was the composition of the attacking force, and this arrived on 29 May, when FRUMEL intercepted and reported that the C-in-C, 2nd Fleet, was at sea in the Truk communications zone and that the Midway invasion force consisting of 1st Carrier Division, Kaga and Akagi; 2nd Carrier Division, Soryu and Hiryu; 7th Cruiser Division, Kumano, Suzuya, Mikuma and Mogami; 18th Cruiser Division, Tenryu and Tatsuta; three or four battleships; 2nd Destroyer Division; part of 4th Destroyer Squadron; 6th Destroyer Division and associated transports; and 3rd Submarine Squadron were underway.51 All of these vessels had put to sea and then gone silent, while a large volume of traffic was being broadcast from shore stations to them, clearly showing they were maintaining radio silence. This silence was noted by FRUMEL on 2 June as the ‘calm before the storm’.52
The storm broke on 3 June, with each side’s aircraft finding the other and launching preliminary air attacks. The battle started in earnest on 4 June and raged until 6 June, with the IJN losing the carriers Akagi, Kaga, Hiryu and Soryu, as well as the heavy cruiser Mikuma. FRUMEL’s record of the battle mentions these losses and notes the exceptionally heavy volume of operational traffic using secret call signs in the Midway and Aleutian areas. FRUMEL also noted that Admiral Yamamoto was not at Midway, but located on board a ship near Jaluit, from which he was overseeing the Midway battle. On 7 June, FRUMEL suggested that the carrier Akagi had been sunk because messages to the 1st Air Fleet’s commander were being passed to another vessel’s call sign, indicating he was on board that vessel and not on Akagi. FRUMEL also reported that the carrier Ryujo might have joined the Midway force.53
By 8 June, FRUMEL had DF fixes on Nagara and 1st Air Fleet; C-in-C, 2nd Fleet; C-in-C, Combined Fleet; Commander 7th Cruiser Division; the tanker Genyo Maru, and five submarines. It was able to say, ‘The Midway Occupation Force appears to be retiring to Saipan,’ and it was. The losses suffered by the IJN were quickly detected by SIGINT. By 18 June, FRUMEL was emphasising the fact that none of the carriers of the 1st and 2nd carrier divisions had been intercepted since the battle, and that the Chief of Staff, 1st Air Fleet, was now communicating with 3rd and 5th Carrier Divisions. As the FRUMEL report puts it, this ‘accentuates the fact that none of the vessels of the 1st and 2nd Carrier divisions have been heard on net’ because they are gone.54
Administrative traffic from personnel offices to the commanding officers of Akagi, Kaga, Soryu and Hiryu confirmed their loss on 2 July. All of this was confirmed in message 566 from Commander, First Air Fleet, on 6 June that was eventually broken out and read on 28 September 1942.
Midway was the deathblow for all the IJN’s ambitions. It killed the romantic military idea that a large decisive battle could win a war against a major power, and it destroyed any hope of delaying the expected US counteroffensives. After Midway, Japan was on the defensive and war, real war, total war was coming to the Japanese Home Islands.
On 14 June, Yamamoto Isoroku, the C-in-C, Combined Fleet, signalled all IJN units and personnel that the carriers Kaga and Soryu and the cruiser Mikuma had been badly damaged. Fittingly, the epitaph for Combined Fleet was written by a cryptanalyst at FRUMEL:55 ‘C-in-C, Combined Fleet is apparently feeding false information to his own subordinates, presumably for purposes of morale and propaganda.’56
With the defeat of the IJN at Midway, the defensive phase of the naval war in the Pacific and Indian oceans was now over and the offensive phase began in earnest. Within two months of Midway, the US Navy forced the pace of the war in the Pacific by launching Operation WATCHTOWER, the US Navy invasion of Tulagi, Gavutu-Tanambogo and Guadalcanal.
Guadalcanal would see the integration of HUMINT from FERDINAND, SIGINT from FRUMEL and FRUPAC, and imagery from aerial photo-reconnaissance missions pioneered by Vice Admiral Robert L. Ghormley of the US Navy. Ghormley, having seen how the British were using aerial reconnaissance in their operations against German-occupied Europe,57 believed that the integration of aerial reconnaissance with SIGINT provided a more comprehensive picture, and he was right.
On the Japanese side, the intelligence picture being built was good. On 5 August, the Special Duty Group (Radio Intelligence) (SPD) at Imperial HQ in Tokyo issued a warning to the IJA’s 8th Fleet that there was a marked increase in enemy communications in the South Seas Area. The Japanese were using DF and traffic analysis to produce this intelligence and, to be frank, it was good tactical intelligence. Fortunately for the Allies, the operations staff of 8th Fleet chose to ignore the possibility of an attack on Guadalcanal or any of the islands, preferring to assign the increase to the overland advance on Port Moresby.58 A further piece of luck occurred on 6 August, when Imperial HQ ignored a report from the garrison on Guadalcanal that all of the indigenous workers on the airfield had fled into the jungle on the evening of 5 August. The disappearance of the locals from the airfield was most likely prompted by Lieutenant F.A. Rhodes’s party passing warnings to the locals working on the airfield.59
All of this intelligence was relayed to Vice Admiral Mikawa Gunichi’s 8th Fleet Command at Rabaul. Mikawa discounted the threat of a US attack on Guadalcanal, and in the early morning of 7 August, Admiral Richmond Kelly Turner began an unopposed landing there of 13,500 marines of General Alexander Vandegrift’s 1st Marine Division.
Admiral Turner was not unopposed for long. Mikawa responded aggressively and obtained permission from Yamamoto to take the heavy cruiser, HIJMS Chokai; four other heavy cruisers of Admiral Goto Aritomo’s Cruiser Division 6; two light cruisers of Division 18; and a single destroyer HIJMS Yunagi, south to Savo Island.60
The presence of this IJN fleet around Rabaul was known to FRUMEL, FRUPAC, OP-20-G and naval intelligence, and it had been repeatedly reported to US naval commanders. When the IJN changed its code settings on 1 June from JN25B to JN25C, this did not stop the reading of the messages, but it delayed them significantly. This meant that although FRUMEL and FRUPAC eventually broke out Mikawa’s concept of operations for Guadalcanal, it took until 23 August, three weeks after the Battle of Savo Island, to report the findings.61
Mikawa attacked the Allied warships in the south and, through a mix of superb torpedo technology and superior watch-keeping and night-fighting capabilities, his squadron burst upon the Allies and fatally damaged the cruiser HMAS Canberra in the first few minutes. They then took the bow off the USS Chicago. The Japanese turned north and sank the US cruisers Astoria, Vincennes and Quincy.62 The US Navy was lucky that Mikawa’s aggressive streak stopped at sinking warships. If he had continued and engaged the undefended US transports lying off Guadalcanal, he would have inflicted a major strategic defeat on the US Navy. Those amphibious landing ships and transports were the only such vessels then available in the Pacific, and their loss would have slowed the US operations considerably.63
The Allied defeat at Savo Island should not have happened. The US Task Force under Admiral Turner had ample information from 14th Naval District’s Combat Intelligence Unit on the IJN elements in the waters around Rabaul. He and his officers, including Captain Howard Bode of USS Chicago, the former head of foreign intelligence at the Office of Naval Intelligence (ONI), were well aware that the Japanese knew the location of the US and Australian vessels because they had been attacking them from the air. On top of this, a reconnaissance report, supposedly from an RAAF aircraft shadowing Mikawa’s task force, was received in the US Task Force one hour before Mikawa’s attack was launched.64
It is at this point in our story that we can dispense with the excuse that Turner was let down by SIGINT. Turner and his command were receiving SIGINT support from FRUMEL and FRUPAC, and we know that SIGINT from FRUMEL indicated a force of eight IJN warships had been detected at 1125 hours on 8 August, 50 miles (80 kilometres) north of the Bougainville Strait and steering 120 degrees at 15 knots.65 The official Admiralty report on the Battle of Savo Island attributes the sighting and identification of this force to an RAAF aircraft, but there is no record of an RAAF squadron operating anywhere in the area at the time of the sighting, and the weather was so bad that Rear Admiral John S. McCain’s air patrols could not take off, a fact of which Admiral Turner seems to have been unaware.66 It is likely that the Admiralty was working off a sanitised report that used an RAAF overflight and sighting as plausible cover to disguise the true source, which was probably SIGINT.67
Admiral Turner appears to have been adequately warned by FRUMEL and FRUPAC that an IJN Task Force of eight warships was heading his way, and that they appeared intent on engaging his force. Turner also missed the significance of the two IJN aerial reconnaissance missions flown over his task force that evening. Even if he was untrusting of the SIGINT, reconnaissance overflights should have alerted him that his command was being targeted for attack. The idea that a change of codes prevented SIGINT from providing sufficient warning is wrong. SIGINT is not just cryptanalysis, it is also DF, TA and the other technical elements, and all of these were available and being reported to Admiral Turner. The hard truth is he was caught unprepared because he did not take the intelligence picture seriously.68
The next phase of operations around Guadalcanal was well covered by SIGINT coming out of FRUMEL, FRUPAC, FECB, India and OP-20-G in Washington. Once again, attempting to force a main fleet action, the IJN assembled another force based on Admiral Goto’s Cruiser Division 6 and Admiral Nagumo’s carrier Shokaku. This force was to sail south and engage the US Navy in yet another encounter battle. This time, SIGINT kept the US commanders informed, or they decided not to ignore it, and by 22 August, Admiral Ghormley and the other US commanders knew there was another fleet action brewing.69
To meet this threat the US Navy assembled a task force comprising the carriers Enterprise and Saratoga; one battleship, North Carolina; four cruisers, Portland, New Orleans, Minneapolis and HMAS Australia; and eleven destroyers, supported by 101 land-based aircraft. The Allied task force faced an IJN force comprising the fleet carriers Shokaku and Zuikaku; the light carrier Ryujo; the seaplane carrier Chitose; the battleships Hiyei and Kirishima; nine heavy cruisers, Suzuya, Kumano, Chikuma, Tone, Ataga, Maya, Takao, Myoko and Haguro; 27 destroyers and five transports, supported by 100 land-based aircraft.70
The US Navy commanders understood good SIGINT was essential in these encounter battles, and a four-man mobile SIGINT detachment commanded by Lieutenant Gilvin Slonim was on board the USS Enterprise to provide tactical SIGINT directly to Admiral Thomas Kinkaid.71 Another team was on Saratoga providing tactical SIGINT to its commander and air commander. In this encounter, the carrier Ryujo was so severely damaged that it later sank, and the Japanese troop convoy coming down the slot off Guadalcanal’s west coast was subjected to heavy air attack, resulting in the loss of the reinforcing troops and their equipment. More significantly, the IJN lost another 90 aircraft along with their aircrew.72
The IJN was not defeated yet, though. As September arrived and operations continued, IJN submarines began inflicting losses. They severely damaged the carrier Saratoga and sank the carrier Wasp on 15 September, and seriously damaged USS North Carolina and sank the destroyer USS O’Brien.73 The IJN efforts to defeat the US invasion of Guadalcanal continued with a remarkable ferocity culminating in the Battle of Tassafaronga between 30 November and 1 December 1942. During the Guadalcanal campaign, the IJN lost two battleships, Hiyei and Kirishima, and the light aircraft carrier Ryujo; and the fleet carriers Zuiko and Shokaku and the seaplane carrier Chitose were damaged. Other losses included five cruisers sunk and eight damaged, and thirteen destroyers sunk and 39 damaged.74
FRUMEL, with its Australian contingent now firmly led by the communicator Commander Newman, continued its work supporting the US Navy in the Pacific. Its relevance was slowly diminishing, however, as the US Navy’s island-hopping campaign pushed closer to the Japanese Home Islands. Eventually, the US Navy began removing US personnel and equipment from FRUMEL so that by January 1945, 90 per cent of FRUMEL’s SIGINT function was gone. Unfortunately, the tenor of the poor relationship between FRUMEL and Central Bureau transferred to Commander Newman’s residual RAN SIGINT operation, which Lieutenant Colonel Alastair Sandford was more than happy to describe as ‘a sort of glorified communications channel’.75 Newman’s organisation now had no friends in Australia, and despite the contribution its Australian personnel had made to Allied victories in the Pacific, they were completely sidelined by Central Bureau.
Central Bureau had not adopted FRUMEL’s closed approach to SIGINT, but had developed itself along British lines, which emphasised cooperation. The result was that when Rudy Fabian and the US Navy organisation left FRUMEL, Commander Newman was left friendless. Everything he now did was subject to criticism.76 The animus was so significant that Alastair Sandford, who was not known for narrowness or bitterness, wanted Newman and the remnants of FRUMEL cut out of all SIGINT, and the last Australian cryptanalyst at FRUMEL, Lieutenant Colonel Treweek, removed and posted to Central Bureau.77 He got both. FRUMEL was now completely isolated from Australia’s ‘Y’ service and, thus, from Britain’s. The ACNB had backed the wrong horse.
For Australia, FRUMEL provided a wake-up call. In 1941, Britain may have failed to live up to the expectations created in Australia, but she never let Australia down. Despite the attitude of the Curtin government and despite the failings of the Australian government in protecting the secrets of her allies, the British argued with the United States on behalf of Australia in many of the lower-level policy arenas, particularly SIGINT. FRUMEL provided a stark lesson for Australia as to how a real international relationship worked. The United States did not regard Australia as a member of an extended family. It was, as the ACNB found out, just another player at the card table, and it had better ante up or lose its seat at the table.