Home / WWII / Under Water, No one can hear you scream
Under Water, No one can hear you scream – The loss of the submarines Argonaut, Amberjack, Grampus, and Triton January-March 1943.
By Jeffrey Cox

In space, no one can hear you scream.

That was the tag line for the 1979 movie Alien. The story of the interstellar commercial freighter Nostromo and its crew of seven, who unwittingly bring on board an eighth. The unfortunate crew ends up isolated, trapped in the metal box of their ship in the vacuum of deep space while being stalked by this creature that is, in the words of the Nostromo’s treacherous science officer, “The perfect organism” whose “structural perfection is matched only by its hostility.” One by one the crew members are picked off by this perfect organism, later christened a “xenomorph.” Some of the crew disappear without even a sound. Science fiction, but a necessary reminder to those who hope to meet other intelligent life in this universe that there is no guarantee, no reason to even presume, that such life would be friendly.

In space, no one can hear you scream.

As it appears in the movie, the interior of the Nostromo is not your neat, shiny, state-of-the-art, everything-in-the-future-is-perfect science fiction space ship. It is cramped, dark, claustrophobic. With pipes, wires, gauges, warning lights, and support beams throughout. Worn, dirty, and dripping water in places. Functional, mostly, with few if any luxuries for the crew. Much like a World War II submarine, on which the design of the Nostromo’s interior is said to be based.

Under water, no one can hear you scream.

Warships in general are cramped, even today. But as anyone knows who has visited one of the preserved World War II submarines like the Cod in Cleveland, the Requin in Pittsburgh, and the U-505 in Chicago, these boats ramped up the cramped by at least an order of magnitude. As if the designers took their cues from a passenger jet’s lavatory. Almost every square inch available was used for something related to the operation of the boat, even to the point of requiring the crewmen to share beds with torpedoes. Which was perfectly safe, provided the men still respected them in the morning.

Under water, no one can hear you scream.

It takes a special kind of person to serve in a submarine. Warships and support ships are cramped enough, isolated enough when in the middle of the ocean and no land in sight, but one can at least go on deck and see the ocean, the sky, the sun. Maybe other ships. Maybe planes flying around. There is still a visible connection to the world outside the ship. On a submarine, once that boat submerges, that’s it. Your world is limited to the submarine. The control room, the torpedo rooms, the engine room, the wardroom. That’s it. There is nothing outside that metal box. Nothing you can see, unless you’re one of the lucky few to have access to the periscope.

But you know there’s a mysterious undersea world outside that metal box. You can’t see it, but the sound guy can hear it, and if those sounds are loud enough, you can, too. The eerie, mournful calls of various sea creatures outside the hull. Calls that can be unnerving, but represent almost no threat, unless a giant baleen whale tries to mate with the submarine or a great white shark tries to eat it.

Other sounds come from things that are not quite as harmless to a submarine. The churn of propellers of a ship. Is it friendly? A friendly ship can still be dangerous – ask the submarine USS Seawolf, sunk by a US escort in a case of mistaken identity and poor communications. An enemy ship is, of course, far worse, especially if it’s a destroyer, the natural enemy of a submarine.

The submarine force was and is called the “Silent Service” for very good reasons. For starters, the crewmen could tell nothing – absolutely nothing – about their service to their families. “Loose lips sink ships” might sound exaggerated, but it was true in the case of submarines.

Additionally, submarines like to approach their targets silently, attack silently, and escape silently. In between, they prefer the noise to be limited to their targets exploding and sinking under the force of submarine torpedoes and deck guns. That’s not always possible, however, as the target’s escorts often decide to make noise as well, using the pings of their sonars and the bangs of their depth charges.

The submarine goes deep. Hiding under the sea’s thermal layer, where the abrupt change in temperature helps to reflect the sonar pulses. The skipper orders “Rig for silent running.” All unnecessary machinery is turned off – including the air conditioning. The air becomes hot and stuffy. Crewmen strip down, even to their underwear, anything to cool off in the oppressive atmosphere. They speak only in whispers, many walk barefoot, just to keep down the noise, anything the enemy’s sound gear might pick up. The submarine’s unfortunate crew ends up isolated, trapped in the metal box of their boat in deep water while being stalked by, say, a destroyer.

The submariners cannot see their stalker but they can hear it. Can hear the pinging of the destroyer’s active sonar and its reflection. It’s that loud. And it gets louder and faster as the destroyer homes in on the boat’s location. Bad enough, but what comes next is far, far worse. The muffled explosions of depth charges. Then they have to feel – feel the shock waves, known as a “bubble pulse,” strike the submarine’s unarmored pressure hull.

And when those shock waves strike the hull, they can dent and rupture the steel plates, shatter gauges, crack batteries, short electrical circuits, open ballast tanks, warp hatches, rock the boat, the whole boat and everything and everyone in it.

Under the pummeling of the destroyer’s depth charges, the men of the submarine try to keep their wits about them, try to keep their bearings and rely on their training, as they are shaken and stirred by someone and something they can’t even see. Trying to control themselves as well as their boat.

Maybe the shock waves split the hull open, sending seawater rushing in with catastrophic force, dooming them to drown.

Or, perhaps even worse, they knock out the submarine’s controls, sending it plunging to depths where the water pressure will crush and crumple and crumble the hull and everyone in it like a cookie in a closing fist.[1]

Or, perhaps worse still, they knock out the submarine’s controls, leaving it to drift underwater with no way to surface, no way for the crew to escape, no way to call for help. Trapped inside their metal box, their metal coffin, as the air inside slowly runs out.

None of which registers outside the submarine. Maybe the destroyer sees an oil slick and some debris. Maybe the destroyer hears a muffled explosion or two from under the waves. But other than that, not a sound.

Silence.

As far as the destroyer is concerned, the submarine died in silence.

Under water, no one can hear you scream.

The early months of 1943 were a dark time for the submarines of the US Navy in the Pacific. Within a little over two months, four American submarines – Argonaut, Amberjack, Grampus, and Triton – would, as far as the US Navy could tell, just vanish. No warning. No word. Nothing. Only one, the Argonaut, would ever have anything resembling a definite explanation as to what happened. The circumstances surrounding the disappearances of the Amberjack, Grampus, and Triton confounded naval investigators post-war and continue to confound historians today.

The common denominator in these losses was one Captain James Fife, Jr. On December 22, 1942, Fife took command of Task Force 42, the submarine unit of the Southwest Pacific Fleet – that is, General Douglas MacArthur’s fleet. Its area of operations included the middle and upper Solomon Islands as well as the Bismarck Sea west of Rabaul. Fife was an experienced submariner, had started the war in Manila on the staff of the Asiatic Fleet’s submarines, where he had seen the excessive caution on the part of submarine skippers and the torpedo issues on the part of the Mark 14 that crippled American undersea operations during that dark time. In commanding Task Force 42, Fife had taken those lessons to heart. Maybe too much.

What normally happened was that a submarine would head out – usually from Brisbane on the east coast of Australia, where Captain Fife was building a giant submarine base – to patrol a designated area. Headquarters would send out updates and orders. Perhaps too many. As the respected submarine historian Clay Blair, Jr., put it:

Fife believed the generally poor showing of US submarines basing in Brisbane up to then was due to overcaution. When he had his feet firmly planted beneath his desk, he abandoned caution. Each of his skippers would give a good account of himself or else he would be summarily relieved.

Up to then, most of the skippers had been assigned an area to patrol and left pretty much on their own. Before the Japanese changed the codes in February, there was a steady flow of information from the codebreakers about Japanese maritime forces reinforcing the Solomons from Palau and Truk. Fife believed the submarine force could better capitalize on this information if the boats were more tightly controlled from Brisbane and shifted about frequently as targets became known. He believed he should take a direct – and firm – hand in the shifting – or as he told his staff, “playing checkers” with submarines.[2]

Fife sought to make use of the signals intelligence program referenced by its name “Ultra”. More specifically, the subdivision of Ultra that covered the Pacific and decoded Japanese communications, which was called “Magic”. Using information provided by Magic as a guide, Fife would frequently order his boats moved about the Solomons and the Bismarck Sea to potential targets he had identified.

While Captain Fife’s idea worked in theory, in practice it had a lot of problems, as the Germans had discovered in the Atlantic when they tried it with their U-boats. It meant a lot of messages sent out to his submarines. And a lot of acknowledgements and other messages back, which the boat could send and receive only when it surfaced, which was usually at night so its batteries could recharge, though as the war went on it became possible to transmit and receive at periscope depth. The big problem with such a system, as the Germans had learned, was that the constant steam of messages made the submarines easier for the enemy to identify and locate. As Clay Blair put it, Fife’s policy “led to many disasters and near disasters.”[3]

Communications with submarines were (and are) cumbersome and even dangerous. While surface ships and aircraft could usually radio a message when they encountered the enemy before they entered combat, a submarine usually could not. Headquarters regularly learned about submarine combat only after the fact, sometimes long after the fact, and other times not even then. If the submarine was sunk while submerged, unable to call for help, unable to abandon ship, the only ones outside the submarine who would have any idea it was sunk would be enemy.

Headquarters could be receiving a submarine’s radio reports regularly, if at uneven intervals. And then … silence. One day someone in fleet submarine operations realizes the submarine hasn’t checked in and radios a request for acknowledgment, only to be met with … silence. An order to end the patrol and return home is met with … silence. As far as headquarters knows, the boat just vanished. All attempts to contact it are met with silence. Frustrating, worrying silence.

Under water, no one can hear you scream.

The US submarine Argonaut was a curious beast. Laid down in 1925 with the designation V-4, she was part of a program of big “cruiser submarines” designed for long endurance more than high speed. And the Argonaut, a name she was formally given in 1931 (four years after her completion), fit the bill, but in so fitting, she was a big ‘un – 381 feet long, displacing 4,080 tons submerged. (By comparison, the Tambor-class submarines, of which the Grampus, Triton, and Grayback mentioned later were members, were 307 feet long and displaced 2,400 tons submerged.) The Argonaut was a monster – a very slow, ponderous monster with a maximum surface speed of 15 knots, slow to dive, and cumbersome, especially when submerged. She was the largest US submarine ever until nuclear submarines came on the scene.

The Argonaut was also the only US submarine ever designed as a minelayer. At her stern she originally had two 40-inch mine-laying chutes, handling equipment, and room for 60 mines. Forward she had four 21-inch torpedo tubes for more traditional submarine operations. Less traditionally, the Argonaut also had two 6-inch deck guns, the largest deck guns ever mounted on a US submarine, one forward of the conning tower, the other aft. In theory, the Argonaut had the weapons to wreak havoc on the enemy. In theory.

In practice, by the time the Pacific War came around, the Argonaut's great size, low speed, lack of maneuverability, and paltry torpedo armament made her next to useless. Rear Admiral Richard H. O’Kane, who had served aboard the Argonaut for four years before becoming famous as skipper of the Tang, said of her fighting capability, “If a fleet boat were stripped of one battery, two engines, six torpedo tubes, and could use no more than 15 degrees of rudder, she would still have greater torpedo attack and evasion ability than Argonaut[4] Clay Blair would describe the Argonaut as "ancient and clumsy". [5] Naturally, ancient and clumsy Argonaut would be called upon to make the first attack approach of the Pacific War.

Theoretically patrolling south of Midway atoll on December 7, 1941, the Argonaut under skipper Lieutenant Commander Stephen G. Barchet was entertaining herself fighting small internal electrical fires. Her lack of air conditioning combined with all-day dives sent the humidity inside the boat skyrocketing, leading to condensation, which dripped onto the electrical wiring, causing short circuits and knocking nearly half her major machinery offline.[6] After sunset, she surfaced to air out the stuffy boat. It was then that she received the radio announcement of the Pearl Harbor attack and the commencement of hostilities with Japan.

Moseying along the surface at 9:45 pm, the Argonaut saw what looked like gunfire west of Midway.[7] Nine minutes later, the Midway base reported it was being shelled.[8] Was it a Japanese invasion force? The big submarine moved in slowly to investigate.

What she found were two Japanese ships, “big destroyers or small cruisers,” later identified as the destroyers Ushio and Sazanami, bombarding Midway with their main batteries.[9] The Argonaut maneuvered to make a submerged torpedo attack, but one of the destroyers apparently saw her dive and headed in her direction. The Argonaut went deep, her attempt at an attack spoiled, while the Japanese crisscrossed and literally ran rings around her position, trying to locate her.[10] At dawn, the big submarine surfaced to recharge her batteries within the friendly confines of air cover from Midway – or not-so-friendly confines, since that air cover proceeded to bomb her unsuccessfully.[11] O’Kane was reminded of the adage “A submarine has no friends.”[12]

Convinced the Argonaut was useless in combat, Lieutenant Commander Barchet chose not to attack a force of three or four destroyers that she detected a week later, sparking an angry exchange with the executive officer and a trip back to Pearl, where the US Navy agreed that the Argonaut was useless in submarine combat.

But the US Navy was very resourceful. When it wanted to be. We have a big, slow submarine. What can we do with it? The obvious answer came in January 1942, when, the Argonaut was sent to Mare Island to have her minelaying gear ripped out and two external torpedo tubes bolted to her stern, and she was designated as … a troop transport submarine.

It was an ingenious concept because the Argonaut could transport about 120 troops. And she soon did so. In August 1942, the Argonaut and her sort-of sister boat Nautilus, another former V-boat, transported two companies of Marine Raiders to Makin in the Gilbert Islands for a raid intended to be a diversion from the imminent invasion of Guadalcanal. Despite the trip being uncomfortable and many of the Marines getting seasick, the raid was successful in damaging the Japanese, if not in diverting their attention from Guadalcanal.

The Argonaut made her way back to Pearl Harbor, where her official designation was changed from SM-1 (for “submarine minelayer”) to APS-1 (for “transport submarine”). It was appropriate for this one-of-a-kind ship … er, boat, who had gone from being the first minelaying submarine to being the first transport submarine. Somewhere along the line, however, she received the unofficial designation of SS-166 (for “regular submarine”), which, though unofficial, was painted on her hull. It may have had unforeseen and unfortunate consequences.

On November 24, the Argonaut, now skippered by Lieutenant Commander John R. Pierce, sailed from Pearl Harbor for the South Pacific. She was being transferred to Captain Fife’s Task Force 42 in Brisbane. Her job would be to land reconnaissance and raiding parties for General MacArthur in the Solomons and New Guinea area. It made sense.

As if to practice for her new assignment, after dusk on December 5 the Argonaut surfaced near the generically-named Ocean Island west of the Gilberts (now Banaba Island, Kiribati) and bombarded Japanese positions with her 6-inch guns.[13] But on December 6, she was ordered to break off her trek to Brisbane and make for Espiritu Santo, where she arrived on December 9 and was topped off and replenished.[14]

It was December 27 or so when Lieutenant Commander Pierce received orders from Captain Fife to make not for Brisbane but for the area between Bougainville and New Britain.[15] It was an area that was always dangerous, being so close to the Japanese bases at Rabaul and the Shortlands, but at this time it was even more so.

Because while it was still bad for the Japanese on Guadalcanal, at least the Imperial Army was holding its own. On New Guinea, not so much. On January 2, 1943, after a campaign that was much longer and more costly than it should have been, Buna on the north coast of Papua finally fell to General MacArthur’s forces. The Japanese 51st Division was on its way – too late to save Buna, but it could reinforce the Japanese fortress at Lae northwest of Buna. So on January 5, five transports – Brazil Maru, Nichiryu Maru, Clyde Maru, Chifuku Maru, and Myoko Maru – carrying the division left Simpson Harbor for Lae. They were escorted by the destroyer Maikaze and the 17th Destroyer Division with Isokaze, Urakaze, Tanikaze, and Hamakaze.

Magic was able to identify the convoy on January 9. The next day, Captain Fife vectored his checkers … er, submarines . . . Grampus and Argonaut to intercept. In the words of Kent Budge’s Pacific War Online Encyclopedia, the Argonaut “really had no business making a regular war patrol,” but to Fife, checkers was checkers.[16] Even so, Argonaut was making a game effort, attacking a Japanese vessel called either Ebon Maru (American sources) or Eban Maru (Japanese sources) in the Bismarck Sea on January 2, though it seems the target was not sunk.[17] A message was received from the Argonaut in the wee hours of January 10, presumably Lieutenant Commander Pierce acknowledging the intercept order.[18]

Nothing more was received from the Argonaut.

Under water, no one can hear you scream.

Skippered by Lieutenant Commander John A. Bole, Jr., the submarine Amberjack had left Brisbane on January 26 for her third patrol of the Solomon Islands area. She was a very active submarine, thanks to Captain Fife. In a brilliantly-written account, Clay Blair probably best described the Amberjack's trek:

Fife controlled these boats strictly, repeatedly moving them around on his checkerboard. Amberjack, for example, was first moved from west of New Guinea to west of Shortlands, then to west of Buka. She was next ordered west of Vella Lavella. En route, she was ordered west of Ganonnga Island. A few days later she was ordered north to cover traffic to the Shortlands. Then she was ordered farther north, and subsequently to the area between New Ireland and Bougainville. She was then shifted north of New Ireland, then ordered to a position west of New Hanover. Her last position was west of Cape Lambert.[19]

On February 3, Lieutenant Commander Bole made his first radio report, claiming to have made contact with a Japanese submarine 14 miles southeast of Treasury Island on February 1 and to have sunk a two-masted schooner by gunfire 20 miles from Buka the afternoon of February 3. She also claimed to have sunk an ammunition ship one night, but the ship had an unexpected bite, sweeping the Amberjack's deck with machine gun fire, killing Chief Pharmacist’s Mate A.C. Beeman, and wounding an officer.[20] On Valentine’s Day 1943, Bole reported the Amberjack had been forced down the night before by two destroyers after having fished an enemy aviator out of the water. The Amberjack was never heard from again.

Under water, no one can hear you scream.

The Grampus had been very active in her own right, if not particularly successful. In October she had landed Australian coastwatchers and supplies on Vella Lavella and Choiseul. More recently, Captain Fife had moved her to intercept the same convoy the Argonaut was supposed to be targeting, but she never made contact. She returned to Brisbane.

There, under skipper Lieutenant Commander John R. Craig, she refueled and replenished, then left on February 11 for a short exercise with her surface escort just outside of Brisbane.[21] Once done with that, she headed out to sea on February 12 for her sixth war patrol and a series of movements dictated by Captain Fife on his checkerboard.

On Valentine’s Day 1943, the Grampus was ordered to patrol the area of west of Shortland and south of latitude 06 degrees 30 minutes South, then the entire Rabaul, Buka, and Shortlands area, splitting the area with the submarine Triton, who was to cover the southern part of the zone. Six days later, Captain Fife sent out orders for the Grampus to patrol north of latitude 04 degrees 30 minutes South (roughly the latitude of Nissan Island, northwest of Buka) until dawn on February 21, and then to patrol east of Buka and Bougainville. On March 2 she was told to round Cape Henpan, proceed down the west coast of Bougainville, south of Treasury Island, north of Vella Lavella, and arrive in the Vella Gulf on the afternoon of March 5.

At that time, the Grampus was to join a rather meticulously planned operation. Admiral Aaron Stanton “Tip” Merrill would leave Espiritu Santo with three light cruisers, Montpelier, Cleveland, and Denver; and seven destroyers, Waller, Conway, Cony, Fletcher, Nicholas, Radford, and O’Bannon. Near the Russell Islands, destroyers Fletcher, Nicholas, Radford, and O’Bannon under Captain Robert Briscoe would split off to bombard Munda just after midnight on March 6. The remaining ships would go through The Slot to Kula Gulf and bombard Vila at the same time.

And for any Japanese who tried to escape through the back door of the Blackett Strait south of Kolombangara between it and Arundel Island, the Americans prepared a surprise. The submarines Grayback and Grampus. On March 2 both submarines were ordered to patrol separate areas of the Vella Gulf at the other end of the Blackett Strait from the Kula Gulf. Any Japanese trying to escape through the Blackett Strait could be ambushed by the submarines.

Assuming the submarines themselves were not ambushed. Late in the evening of March 5, Guadalcanal broadcast a warning that two Japanese light cruisers or destroyers had left Faisi, in the Shortlands, at 7:10 pm and were heading southeast at high speed.[22] At 11:30 pm on March 5, a PBY Black Cat radioed news that it had passed over two destroyers, undoubtedly the same ships, that were headed toward Wilson Strait, between Vella Lavella and Cannongga.[23] If the destroyers were on their way to Vila-Stanmore, they could skirt the southern edge of the Vella Gulf. They were a threat. And an opportunity. Both submarines were to be on the lookout.

The mission orders had gone out to both submarines and made each aware of the other’s assignment. The Grayback acknowledged the order but the Grampus did not, just as she, curiously, had not acknowledged any of the other orders she had received – had apparently not communicated at all with anyone – since she left on February 12. Orders radioed on March 7 and again on March 8 for the Grampus to check in were also met with silence.

After she left Brisbane, the Grampus was never heard from again.

Under water, no one can hear you scream.

Like the Amberjack and Grampus, the Triton had been a very active submarine. She had been the first US Navy submarine to fire torpedoes in the Pacific, and had continued that aggressiveness to become one of the navy’s most successful submarines.[24] The Triton left Brisbane on February 16 for her sixth war patrol, her first under skipper Lieutenant Commander George Kenneth MacKenzie, Jr. She would pass through the Rabaul-Buka-Shortlands area, the Grampus’ patrol area, on the way northward to her own patrol zone. At around noon on March 6, in the adjacent patrol zone of the submarine Trigger, the Triton engaged a Japanese convoy of five ships – the destroyer Yuzuki escorting the freighters Kiriha, Mito, Nagano, and Ryuzan Marus. MacKenzie managed to plunk a torpedo into the stern of the 3,057-ton Kiriha Maru, killing four of the ship’s crew as well as the ship itself. Another torpedo apparently thunked into the side of the Mito Maru, not exploding but leaving an unsightly dent. A third torpedo of the Triton’s made that special maneuver that seemed unique to American torpedoes – the Circular Torpedo Run, in which the torpedo is launched, and, instead of heading for the target on the assigned gyro angle, it circles around threatening to hit the submarine that launched it. The Triton had to dive deep to avoid her own munitions. Dangerous and, even worse, embarrassing. It also gave the Yuzuki the chance to drop depth charges on her, albeit without damage.[25]

Lieutenant Commander MacKenzie reported this action in messages of March 7 and 8. The Triton later tried two night attacks, one dawn attack, and one afternoon attack, all without success. She started heading back to her area, but not before one more night attack on a convoy. MacKenzie claimed five hits out of eight torpedoes fired, sinking two freighters, although he could not confirm results because an escorting destroyer charged in and forced the Triton to go deep.[26] Meanwhile, Admiral Halsey was impressed with the moxie of MacKenzie, who was on his first war patrol as skipper. The admiral radioed, “MacKenzie – You are doing extremely well for a beginner.”[27]

But any beginner’s luck the Triton was enjoying would not hold. On March 11, Captain Fife’s headquarters received a message from the Triton: “Two groups of smokes (funnel smoke from a ship’s stack), 5 or more ships each, plus escorts ... Am chasing.” In response, Fife ordered her to stay south of the Equator and reminded her that the submarine Trigger was in an adjacent patrol area.[28] Two days later, a warning was sent out that three Japanese destroyers had been sighted heading northward in the general direction of her patrol area, probably on a submarine hunt or a convoy cover who had missed contact with their charges. The Triton’s patrol zone was moved slightly to the east on March 16. Finally, on March 25 the Triton was ordered back to Brisbane. But after the submarine’s March 11 update, all messages sent to the Triton were met with silence.

The Triton was never heard from again.

Under water, no one can hear you scream.

On March 22, the US Navy listed the Amberjack and Grampus as overdue and presumed lost, and began the process of notifying the crew’s families. On April 10, the same determination was made for the Triton. Public announcements were made for the loss of the Argonaut in March 21, the Amberjack and Grampus on June 12; and the Triton on July 22.

Four boats lost in barely two months. As Clay Blair put it, “In peacetime, the loss of four submarines in so short a time would ordinarily prompt an extensive investigation to determine blame: heads would roll.”[29] Of course, in peacetime, no enemy would be trying to sink the submarines, so the loss of four submarines in so short a time would probably be much more cause for concern.

The Navy investigated the loss of each of the submarines, but it did not take that long to ascertain – mostly – what had happened to the Argonaut. Japanese radio was quick to brag about her sinking.[30] It was that convoy that Captain Fife had sent her to attack.

Because while Captain Fife was moving his checkers, General MacArthur’s Air Force had already found the convoy and mustered American Flying Fortresses, Liberators, and Marauders; and Australian and Kiwi Catalinas and Hudsons, supported by Lightnings and the venerable Warhawks, to hammer away at this convoy for a full five days. On January 7, a PBY hit the Nichiryu Maru and sent her to the bottom; Maikaze fished 739 of the 1,100 troops on board out of the water, but could do nothing to save their supplies. Then came the turn of the Myoko Maru. Wracked by American fighter bombers, she was run aground south of Arawe, but her beached carcass was bombed and destroyed the next day.[31]

Nevertheless, even after losing 40 percent of their transports, the Tokyo Express managed to land some 4,000 troops at Lae with most of their supplies. The 5th, Royal Australian, and Royal New Zealand Air Forces were left to bomb the supply dumps left behind and strike the empty retreating ships of the convoy on their way back to Rabaul, damaging the Brazil Maru,[32] One of the US Army Air Force bombers, a B-25 Mitchell with its bomb racks empty, was over the convoy on January 10 when it saw a flash and a column of water burst up next to one of the escorting destroyers.[33] A torpedo hit? Must be. Had to be a submarine attack.

But the Japanese were way ahead of the Army air crew. A destroyer was dropping several depth charges. The American aviators saw an unusually-large submarine bow break the surface of the water at a steep angle and hang there, like an accusing finger pointed at the heavens. It seems the submarine had tried to surface, probably by blowing main ballast. It was a desperate move, surfacing in the midst of enemy destroyers (unless you have, say, 6-inch deck guns), so her plight must have been desperate and her damage severe. This was a bad sign, the crew knew, but there wasn’t much they could do. Except watch in horror as two destroyers now circled the helpless boat like sharks, mercilessly blasting her with 5-inch gunfire. The submarine was evidently shattered, as the bow slid back into the sea from whence it had come.

The Mitchell crew headed back to report they had seen a crew of US Navy sailors die with their submarine. But it would be more than a month before the Navy would identify the submarine as the Argonaut, although when handed the bomber crew’s January 10 report the next day, O’Kane immediately figured out the identity of the submarine.[34] For reasons known only to Captain Fife, this ancient and clumsy submarine was sent to intercept a heavily-defended empty convoy as it headed back toward Rabaul.[35] Maybe he forgot the Argonaut's mission was to transport troops. Maybe he saw the “SS-166” on his chart and assumed the Argonaut was the typical fleet submarine instead of a plodding giant undersea beast. Checkers is checkers.

Despite her being such an old and limited submarine, the loss of the Argonaut had a crippling effect on Captain Fife’s submarine force. The Argonaut had been specifically refitted to carry out special missions such as making covert supply runs and landing reconnaissance and raiding troops. Because of her loss, those missions had to be handled by other submarines that were badly-needed on the front line, where the Argonaut had been useless. Immediate plans to land 30 Australian commandos and five tons of supplies on Bougainville had to be hurriedly reworked. The submarine Gato took over the mission, but had to make two trips where the Argonaut would have needed only one. Ultimately, six of Fife’s submarines would be engaged for varying periods in 1943 running these special missions that the Argonaut was supposed to run.[36] It underscored the abject stupidity of having the Argonaut intercept the heavily-defended, empty convoy.

But that was the end of the easiness for Navy investigators, because until they had access to Japanese records – and at this time the Japanese were being less than completely cooperative – they had very little to go on. Among the information they had to process was the experience of the Grayback.

The Grayback had been ordered to patrol the Vella Gulf with the Grampus the night of March 5-6. Having finally arrived in the Vella Gulf the night of March 5, her skipper Lieutenant Commander Edward C. Stephan was cautiously moving the submarine on the surface when, at 9:50 pm, her sonar picked up the sound of propellers. A look around the gulf revealed a dark silhouette to the northeast of the Grayback, seemingly creeping around all by itself.[37] It was in the part of the Vella Gulf assigned to the Grampus.[38] Must be the Grampus, Stephan figured. Right where she was supposed to be.[39] The Grayback did not attempt to exchange recognition signals with her, though, because her SJ radar, by which those signals would be transmitted, was malfunctioning, which also prevented her from tracking the Grampus. It was the Grampus, though. Had to be. Still, unable to establish his boat’s bona fides, Stephan decided to give the Grampus some space. Didn’t want a case of mistaken identity, and there was that chance she was something more sinister than the Grampus. It was about ten minutes later that the Grayback received the report of the two destroyers coming their way.[40]

Despite the Black Cat PBY passing only 700 feet overhead, those two destroyers, the Murasame and Minegumo, seem to have been unaware of its presence or the implications therefrom. Operating under the command of Destroyer Division 2 commodore Captain Tachibana Masao, the Murasame and Minegumo were on a quickie supply run to the airbase at Vila. While the Black Cat was sending out its warning, the two destroyers were stopped at Vila offloading supplies for the garrison onto waiting barges. It would lead to their destruction in the Kula Gulf at the hands of Admiral Merrill’s cruisers.

Meanwhile, the Grayback had been struggling with the islands in the Vella Gulf. Lieutenant Commander Stephan noted that the little islands made spotting ships difficult, especially against the dark backdrop of the land. “Our SJ radar would have helped a lot,” he commented acidly, “had it been working.”[41] The Grayback was able to witness the gun flashes and flares that marked the end of the Murasame and Minegumo. The flares actually illuminated the Vella Gulf at times. Even so, the night of March 5 into March 6 passed with the Grayback seeing and hearing nothing else of note, and afterwards she went back to her regularly scheduled patrol. Lieutenant Commander Stephan never did establish contact with the Grampus or confirm that she was the shadow he had seen.

The afternoon of March 6, there was something odd, however. As it is usually put in the histories, including the U.S. Navy’s U.S. Submarines Lost in World War II, “an oil slick was reported in the Blackett Strait.” Or “an oil slick was sighted in the Blackett Strait.” That is almost always how it is explained. Passive voice. No mention of who sighted it, when it was sighted, who sent the report to whom, how it was sent, when it was sent, or where in the Blackett Strait the oil slick was seen.

It seems that later efforts to locate the report or at least get more details on it fizzled out, leading to claims that the report “has no pedigree.”[42] Where the original report is now, if it ever existed, is indeed a mystery. But “an oil slick was reported in the Blackett Strait” would seem to describe a sighting documented in the Task Force 63 War Diary as a report from “Cactus” (Guadalcanal) of a “PM” (afternoon) sighing of an “Oil slick” “SE (of) Kolombangara”.[43] That description is not quite “in the Blackett Strait” and opens up the possibility that the slick was not in the strait but instead in the Kula Gulf, possibly a remnant of the sinking of the Minegumo and Murasame the previous night. Most of the reconnaissance missions from Guadalcanal during this time period were handled by the Lockheed Hudsons of No. 3 Squadron of the Royal New Zealand Air Force, which could explain the difficulty researchers have had in locating the original sighting report.[44]

The Navy also had a rather curious report from Lieutenant Commander Roy S. Benson’s submarine Trigger, whose patrol area was adjacent to that of Triton. The Trigger’s entries from March 15, 1943, quoted in relevant part (times are in 24-hour military format):

1035

(8th contact) (3rd and 4th attacks)

Lat. 0-00 N. Long. 145-00 E.

Convoy of five freighters with 2 escorts. One 7000 ton freighter probably sunk. One 7000 ton freighter damaged.

Picked up smoke on the horizon bearing 103. Commenced approach. Turned out to be a convoy of two columns, 2 freighters in the right hand column and three in the left. There was an escort, small destroyer or other type, on each outboard bow. The convoy was zigzagging with escorts patrolling station. We worked into position ahead in order to get between the columns. Maneuvers were successful.

1215 (K)

Fired three stern tubes at the leading ship in the right hand column at 1600 yards 90 port track. Two hits. While these torpedoes were on the way got set up on the leading ship of the left hand column. The last zig had placed the columns in echelon so that the angle on the bow of our new target was 10 degrees starboard, relative bearing 300 and at a range of 2000 yards. The first firing, due to smooth sea, disclosed our position. Our new target headed for us. There was not time for us to turn.

1217 (K)

Fired three bow tubes at zero angle on the bow range 700 yards, gyro angles about 45 degrees. Used normal dispersion as the spread. Two hits. Went deep on firing to avoid the target. The escorts were after us with depth charges instantly, alternating listening and dropping. The sea was smooth and sound conditions excellent so could not at the time come up to take a look. During the quiet periods the sounds associated with the breaking up and sinking of a ship were heard in the direction of our first target.

1420 (K)

Nothing in sound. Came to periscope depth. Nothing in sight.

1437 (K)

Depth charges at a distance. Nothing in sight. Assumed they were from planes. These distant explosions continued for the next hour.

1515 (K)

Sighted smoke on the horizon. Started approach. Its bearing remains constant.

1705 (K)

Could now see a ship’s masts. Continued the approach submerged. The new target was a ship which looked like our earlier second target, the one which we have fired and hit twice at 700 yards zero angle on the bow. Alongside of her on the far side and sticking out astern was the smaller ship resembling the freighter which had been third ship in the left hand column. This smaller ship was furnishing buoyancy and propulsion; speed about 2 knots, course 190. The pair was being protected by two small destroyers or corvettes. Continued the approach.

2142 (K)

Fired three bow tubes at 700 yards 90 port track to run at 15 foot. Apparently the torpedoes did not get up to the targets’ keels in time. Sound tracked them straight on the targets’ bearing. Immediately escorts started running around but apparently did not know where to look. It was bright moonlight. We reversed course and took position for stern shots. Possibly we had been too close but we turned and reversed course at one-third speed without crossing the target's track so it is believed that the range at firing was sufficient. The torpedoes, in any event, did not get up to the target’s keel depth.

2201 (K)

Fired three stern shots at 1100 yards 100 port track to run at 15 feet. The only one sound heard ran circles; ran right over our engine room. The commanding officer’s confidence in himself and his weapons was so shaken that he considered further action by this vessel against these targets futile. After the escorts gave up, surfaced and sent contact report. We were not able to get any station to respond to our calls and therefore sent it blind on two frequencies.[45]

The Trigger’s report is tantalizing on a number of levels. First, there is her frustration with the ineffectiveness of her torpedoes. She believed she sank one freighter and damaged a second, but a second attack on the second target was doomed because the torpedoes apparently ran under the target, and a third was doomed when a torpedo was heard to make a circular run and sprint over her own engine room. It was a good thing the Trigger was already submerged or else she would have been yet another combat craft sunk by her own torpedo making a circular run. There were at least two such cases documented in the Pacific War, sinking the submarines Tullibee and Tang; the aforementioned Richard O’Kane would be one of the few survivors of the Tang tragedy. Already in the Pacific War, the destroyer Porter had been sunk when a torpedo from a downed TBF Avenger whose pilot she was rescuing was released and made a circular run, striking the destroyer.

There is also the matter of the counterattacks by the convoy’s escorts. The Trigger endured a depth charging of her own, though she does not seem to have been seriously threatened. However, there was her entry for 2:37 pm: “Depth charges at a distance. Nothing in sight. Assumed they were from planes. These distant explosions continued for the next hour.”

The Japanese were not attacking the Trigger with these depth charges, so whom were they attacking?

After the war, it was hoped the four submarines’ full stories could be told. And in the case of the unfortunate Argonaut, it largely was.

According to Japanese records, an aircraft attached to the 582 (the former 2nd) Air Group had spotted the submerged boat and dropped bombs to call the attention of the escorting destroyers to its position.[46] One of them, Maikaze, raced out to the spot and dropped several depth charges. Shortly thereafter, the submarine’s large bow broke the surface at a steep angle. The Maikaze was joined by the Isokaze in circling the gravely-injured boat, pumping the exposed bow full of 5-inch shells. The bow slid back into the sea, leaving some gruesome debris. According to the Japanese, the “destroyed top of the sub floated.”[47]

To add insult to expiry, the Argonaut's torpedo had apparently exploded prematurely and hit nothing. Her war time record would be: one 1,200-ton vessel damaged.

Already having a good idea of the Argonaut's fate, naval investigators scoured Japanese records looking for anything that could explain the loss of the Amberjack, Grampus, and Triton. The Amberjack and Grampus were especially problematic because they disappeared in roughly the same area at around the same time with little or no communications from either. To further complicate matters, there was considerable submarine and antisubmarine activity in the area, any number of which could have involved the Amberjack and the Grampus. Investigators found a number of clues to examine.

According to Japanese records, at 10:00 am on February 16, a convoy consisting of the torpedo boat Hiyodori and the submarine chaser Ch-18 escorting the transport Noshiro Maru, with air cover provided by a seaplane, probably an Aichi E13A Type 0 reconnaissance seaplane, from the 958 Air Group, left Rabaul for Kolombangara. They were off Cape St. George when at 3:28 that afternoon, they spotted four torpedo tracks off the starboard beam, 4,000 meters away. The torpedoes were successfully avoided.

About six minutes later, the seaplane located the attacking submarine and dropped several depth charges. At 3:40, the Hiyodori arrived on site and dropped nine depth charges on the submarine’s suspected location. Some five minutes later, the Ch-18 got a fix on the boat and dropped six depth charges. This last attack brought oil welling to the surface. As the oil slick expanded, in the middle of it a conning tower appeared, its upper section bursting through the surface. That’s generally a bad sign for the submarine. Before the Japanese could train their guns on it, however, the tower disappeared back into the oil slick.

Smelling blood or at least oil in the water, Ch-18 dropped three more depth charges, and would have dropped still more except “parts of the hull” floated to the surface. That was it for the combat. The Japanese milled around looking for more debris. At 4:42 pm, the Hiyodori recovered several items, including a life raft printed with the words “Philadelphia Navy Yard”. The Japanese recorded it as a sinking.

The next day, February 17, the Japanese made two submarine sightings, both times apparently catching the same submarine. Which would soon make a nuisance of itself. On February 18, the armed auxiliary aircraft transport Keiyo Maru took a submarine torpedo and was “lightly damaged.” This was apparently not the result for which the boat’s skipper had been hoping, so she came back for Round 2 on February 19. This time the results were “unknown,” but the Keiyo Maru remained afloat and operational. Another proud performance for the Mark 14 torpedo.

Two attacks in two days by the same submarine on the same ship, resulting in, at best, light damage. That was embarrassing. For both the Americans and the Japanese. The Japanese were having none of it. The 958 Air Group sent out several seaplanes, probably Aichi E13A Type 0 reconnaissance seaplanes again, to bother the bothersome boat.[48] They found a submarine southeast of New Britain, evidently at periscope depth, and proceeded to bomb it. They reported one direct hit on the conning tower and a large amount of oil on the surface after the attack.[49]

But that was supposedly not the end of the subversive submersibles. The Japanese reported another submarine sighting on February 24. And on February 27, there was the somewhat mysterious case of the Kirikawa Maru.

The Kirikawa Maru, a 3,829-ton cargo ship, was her own one-ship convoy this February 27. And she was a busy convoy. At 4:00 am, she had left Shortland for Buin, on the opposite side of the Shortlands anchorage. She arrived two hours later. Then, at 9:00 am, she left Buin, packed with elements of the 7th Yokosuka Special Naval Landing Force, two 140-mm guns, four 80-mm guns, and 600 tons of supplies. Escorted by the subchaser Ch-26 and minesweeper W-22, the Kirikawa Maru started on a reinforcement run for Kolombangara.

She almost made it. Northeast of Vella Lavella, she was ambushed by US Navy SBD Dauntless dive-bombers and F4U Corsair fighter-bombers, escorted by US Army Air Force P-38 Lightnings and P-40 Warhawks of the joint Solomons air command with the disinfectant-sounding name of “AirSols”. The Kirikawa Maru was heavily damaged and her ammunition stores were set afire. The cargo ship was scuttled with three shots from one of the escorts, but the record is not clear as to whether it was the Ch-26 or the W-22.[50] The Imperial Japanese Navy Page suggests Kirikawa Maru was sunk “probably with all hands.”[51] Which would be curious by itself. It would mean either the Kirikawa Maru was scuttled by friendly ships with men still on board, or all the men aboard the Kirikawa Maru were killed by the air attack but the ship was left afloat.

In any event, later that day the W-22 found herself off Kolombangara, apparently by herself. Putting together the pieces of this incomplete puzzle, it sounds as though, after the air attack, the Ch-26 headed back to Shortland while the larger W-22 was left to rescue survivors of the Kirikawa Maru and probably to scuttle the ship as well. It was during this time that the minesweeper was the recipient of a submarine attack. Not a very good submarine attack, as the minesweeper was not sunk and is usually listed as “possibly” damaged.[52] Since she went back to Shortland and was immediately sent out again, whatever damage was inflicted seems to have been minor, though she was down for 10 days at the end of March for repairs in Rabaul.[53] It sounds like an approaching torpedo prematurely exploded just before hitting her.

When all of these reports are piled on top of the Grayback’s sighting of what she believed to be the Grampus on the night of March 5-6, the presence of the Japanese destroyers Murasame and Minegumo, and the reported oil slick in the Blackett Strait southeast of Kolombangara the afternoon of March 6, that’s a lot of puzzle pieces – too many sinkings and too many sightings for too few submarines.

Eyewitness accounts are usually the least reliable form of evidence, but unfortunately war often leaves interested parties with nothing but eyewitness accounts. The only physical evidence here of a sinking is the “parts of the hull” and the life raft printed with the words “Philadelphia Navy Yard” recovered after the February 16 attacks. And that evidence is not conclusive as to the submarine’s identity. The February 18-19 attacks on the Keiyo Maru and the Japanese counterattacks don’t even have that much, although, to be sure, the submarine force’s Admiral Charles Lockwood thought that the February 19 antisubmarine attack “presumably” sank the Grampus.[54]

Other evidence might be less ambiguous, if only a bit less. The February 24 Japanese sighting of an enemy submarine was in the Grampus's operating area. If the sighting was accurate, in theory it could only have been the Grampus.[55] On top of that is the February 27 attack on the already heavily harangued minesweeper W-22 off Kolombangara. The Grampus’ patrol area had been moved to east of Bougainville, so this attack is consistent with that area.

Which makes three sightings of a submarine where the Grampus was supposed to be, the third being the Grayback’s sighting on the night of March 5-6 in Vella Gulf.[56] Three sightings of a submarine where the Grampus was supposed to be suggests that the submarine was indeed the Grampus and the February 19 Japanese antisubmarine attack was not as successful as the Japanese claimed.

Which is hardly unknown. As submarine historian Theodore Roscoe noted, “Japanese credulity, rooted perhaps in egotism, accepted the flimsiest clue as an indication of success. As a result, the Japanese credited themselves with sinking about ten times the number of submarines actually sunk.”[57] Reporting a direct hit on a conning tower and an oil slick without debris doesn’t necessarily prove a sinking. Early in the Pacific War, three US Navy aircraft from the carrier Enterprise reported a direct bomb hit on a submarine and an oil slick that suggested a sinking. That submarine turned out to be the USS Pompano, who was not hit directly but one of her oil tanks was ruptured, leaving her rather annoyed – and giving a second potential explanation for the oil slick seen following the February 19 Japanese antisubmarine attack.[58] Again, eyewitness accounts are usually the least reliable form of evidence.

Moreover, we have the Court of Inquiry into the loss of the submarine Dorado, about which much more will be discussed below. The Court called veteran submarine Commander Jack H. Lewis, who had commanded the Trigger, interestingly enough, and the Swordfish to testify as an expert witness concerning certain questions about submarines. At the time of his testimony he was the Prospective Commanding Officer of the submarine USS Dragonet. On November 9, 1943, the Court asked Lewis a series of questions:

Court: In the event that a submarine has been attacked by aircraft and the submarine has been materially damaged, do you think that any evidence of that damage would subsequently be shown on the surface even though the submarine itself disappeared after the attack?

Lewis: I cannot conceive of a submarine being damaged to the extent that she is lost without very much evidence being left on the surface.

Court: What form would that evidence take?

Lewis: Undoubtedly oil; and lots of it, would be on the surface. Also, wooden sections of the deck grating. If the pressure hull has been penetrated, and it, undoubtedly would be, there would be segments of cork floating to the surface. I, also, believe that for many hours afterwards there would be a lot of bubbles breaking the surface of the water.[59]

That the Japanese reported only oil after the February 19 attack again calls the success of that attack into doubt.

Because significant debris was recovered afterwards, it is reasonable to conclude the February 16 attack did sink a submarine. After that attack, there do not appear to have been sightings of a submarine where the Amberjack was supposed to be, which by that time was west of Rabaul. It is therefore reasonable to conclude that the February 16 attack sank the Amberjack.

The Grampus is far more complicated. After her alleged sinking on February 19, it is possible to plot the submarine sightings moving east with the Grampus’ shifting patrol zone, which suggests the sightings are of the Grampus and that she was shifting her patrol area consistent with her orders. The problem is that whether the Grampus even received these orders cannot be confirmed because Lieutenant Commander Craig never acknowledged the orders. This brings to the forefront one very curious aspect of the Grampus case that is often overlooked – the Grampus’ radio silence. It is true that the submarine had not communicated with headquarters or anyone since the Japanese attack of February 19 that many believe sank the Grampus. But that fact both understates and misleads, because the Grampus had not communicated with headquarters or anyone else before the Japanese attack of February 19, either. Not a peep since she left the Brisbane area on February 12.

It was not unusual for a submarine operating in a hot area to delay surfacing and transmitting and receiving messages until the area cleared. But the Grampus was silent for at least one week – between her leaving Brisbane on February 12 and the February 19 attack that allegedly sank her. Her silence did not seem to concern Captain Fife or Task Force 42 at that time. It would be a concern by March 7, when Task Force 42 ordered the Grampus to report in and acknowledge.

The question must then be asked whether the Grampus was having a radio issue, starting early in her patrol, of which Lieutenant Commander Craig and her crew may not have been aware, at least initially. The submarine evidently was receiving communications, pointing again to the plots of the submarine sightings with Grampus’ shifting patrol area. Was she sending? And if she was not sending, did Craig and the crew know she was not sending? It was not unknown for a balky radio suite to be able to receive but not send.

The Grayback’s was the last reported possible sighting of the Grampus. Even though the two destroyers Murasame and Minegumo were headed her way, the Grayback neither saw nor heard the destroyers or anything else the night of March 5-6.

The next afternoon, there was that oil slick reported “SE (of) Kolombangara”.[60]

But is the reported oil slick in the Blackett Strait significant? One major problem in linking this report to the loss of the Grampus is that “southeast of Kolombangara” is not the same thing as “in the Blackett Strait”. That description and opens up the possibility that the slick was not in the strait but instead at the southern end of the Kula Gulf. Most of the reconnaissance missions from Guadalcanal during this time period were handled by the Lockheed Hudsons of No. 3 Squadron of the Royal New Zealand Air Force, which could explain the difficulty researchers have had in locating the original sighting report.[61]

The oil slick would seem to suggest something sank there during the night. But what? If the oil slick was in the Kula Gulf, it could have been a remnant of the sinking of the Minegumo and Murasame the previous night. If it was indeed in the Blackett Strait, then the Grampus starts to come into play. But why would a submarine sink there? The obvious inference is that either or both the Murasame and Minegumo sank the Grampus before they themselves were sunk in the Kula Gulf off Vila, taking their logs and records with them. And, indeed, that is the inference that has been made over the years.

Yet that inference is not completely solid. Though the Grayback saw the gun flashes and flares of the ambush of the Murasame and Minegumo on the other side of Kolombangara, and though she had heard propellers and sighted a dark object in the Vella Gulf that she believed to be the Grampus, she heard nothing else the night of March 5-6, including the Murasame and Minegumo passing through. Being only 15 miles away, she likely would have heard any depth charging.

Could the Grampus have been caught on the surface and sunk by gunfire? Yes. And, to be sure, for some reason Captain Tachibana did not want to go back through the Blackett Strait on his way out after dropping off the supplies at Vila. Maybe he saw something in the strait he didn’t like? Like an enemy submarine?

But there are multiple problems with this theory. Though the Grayback later saw the gunflashes and flares of the ambush of the destroyers off Vila, she did not see any other gunflashes beforehand that would indicate a separate surface action. The Grayback did not detect the Japanese destroyers at all, and there is the possibility that the Japanese destroyers had passed into the Blackett Strait even before the Grayback allegedly saw the Grampus. The biggest issue here, however, is that none of the Japanese survivors mentioned any antisubmarine action that could have sunk the Grampus before arriving in Vila. Both Captain Tachibana and skipper Lieutenant Commander Tanegashima Youji survived the sinking of the Murasame; neither reported an antisubmarine action either during or after the war. The Minegumo's skipper Lieutenant Commander Uesugi Yoshitake was killed off Vila, but presumably any surface action against a submarine would have involved the gunnery officer, yet the Minegumo's gunnery officer Lieutenant Tokuno Hiroshi also made no mention of a surface action or antisubmarine combat before Vila-Stanmore in his postwar interrogations. Nor is there any indication that the destroyers were delayed in arriving at Vila or completing their supply mission.

This leaves two intellectually unappealing possibilities of incomplete solutions. The first is that the Grampus was sunk by the Japanese air attacks of February 19. If that is true, then why was a submarine sighted on February 24 and March 5 and a submarine attack commenced on February 27 in areas where the Grampus was the only submarine operating? The second is, if the Grampus survived the February 19 Japanese air attacks, she must have sunk no earlier than the night of March 5-6, but why did she sink? If the Minegumo and the Murasame did not sink the Grampus, what did?

There is no clear answer here; the best we can do is mention the possibilities. The Grampus could have succumbed to whatever damage she suffered (and to be clear, she would have suffered damage, at the very least a ruptured oil tank) from that February 19 attack. She could have been the victim of a friendly antisubmarine attack; more on that later. An operational accident is a possibility, including that almost uniquely American phenomenon: the Circular Torpedo Run.

In this time period both the Trigger and the Triton had reported circular runs by their torpedoes. This does not suggest that circular runs by torpedoes were more likely during this timeframe, but there does seem to have been an unusually high number of them during this period. Circular runs in American torpedoes were often caused by improper installation of the gyros. A torpedo’s gyro was inserted just before the torpedo was slid into the tube. If the gyro was not installed at all or if it was installed backwards – the ability to put the gyro in backwards was yet another design flaw for American torpedoes – the torpedo could run wild. While she did not report a Circular Torpedo Run as had the Trigger and Triton, since the Grampus had reported nothing, the attack on the W-22, which was, at best, not completely successful, suggests the Grampus also was having issues with her torpedoes. But if she was indeed the victim of the Circular Torpedo Run, what was she attacking? And where? And when?

The fate of the Argonaut is definite and tragic. The Amberjack was almost – almost – definitely sunk by the February 16 attack by the Hiyodori and Ch-18. The Grampus is much more problematic because so much of her evidence is uncertain, but either she was sunk on February 19 by an air attack, or, more likely, she was sunk the night of March 5-6 or later by some means unknown.

Which leaves the fate of the Triton. The US Navy’s official position:

Information available now that the war is over shows that Triton was, without a doubt, sunk by the enemy destroyers of which she was given information on 13 March. Enemy reports show that these ships made an attack on 15 March at 0°-09’N, 144°-55’E. This position was slightly north and west of Triton’s area, but she undoubtedly left her area to attack the destroyers or the convoy they were escorting. The report of the attack the destroyers leaves little doubt as to whether a kill was made, since they saw “a great quantity of oil, pieces of wood, corks and manufactured goods bearing the mark ‘Made in U.S.A’.” In addition, Trigger, in whose area this attack occurred, reported that on 15 March she made two attacks on a convoy of five freighters with two escorts at 0°N, 145° E. At this time she was depth charged, but not seriously, and she heard distant depth charging for an hour after the escorts had stopped attacking her. Since she was only about ten miles from the reported Japanese attack cited above, it presumed that she heard the attack which sank Triton. Apparently by this time the destroyers had joined their convoy.

The US Navy’s official position tries to wrap everything up with a nice bow, but the wrapping paper isn’t quite taped down. The Navy does mention Trigger hearing depth charging in the distance, but is coy as to who exactly sank the Triton, mentioning only “the (three) enemy destroyers of which she was given information on 13 March.” Whose report said they saw “a great quantity of oil, pieces of wood, corks and manufactured goods bearing the mark ‘Made in U.S.A’”? The Navy never identifies the source of this report. Presumably, the Navy could have identified the source since its report is quoted, but chose not to do so. The puzzling question is why.

While such respected submarine historians as Theodore Roscoe and W.J. Holmes have seemingly endorsed the US Navy’s official position, it is so bereft of definite information aside from the reference to the Trigger’s report that it has drawn the skepticism of other such respected submarine historians as Clay Blair, SubSoWesPac historian Ed Howard, and Imperial Japanese Navy Page historian Allyn Nevitt.

Clay Blair wrote, “After the war, U. S. naval authorities made an intensive hunt in Japanese records to determine the causes for each loss [i.e., Triton, Grampus, and Amberjack]. There were clues, giving rise to various speculations, but nothing positive was ever learned about any of them.”[62] Howard, on his SubSoWesPac site, was more blunt:

The only thing that can be said with certitude is that Triton was never heard from again after March 11, 1943. At that time, she reported being in pursuit of two groups of enemy vessels, each having five or more ships with escorts. We know that the Navy was under a lot of pressure after the war to find explanations for the losses of the Amberjack, the Grampus, and the Triton. The final report they published on the Triton’s loss in 1946 does not bring final closure to the issue. Research published since its publication casts doubt on most of the explanations presented in the Navy’s report and renders it speculative, at best. The Triton’s loss remains an unsolved mystery.[63]

These may represent two extremes of the same spectrum, arguing either the US Navy report is completely accurate or completely inaccurate. Those are not the only choices, however. What if the Navy version is partially accurate? Or accurate except for the reference to “three destroyers” sinking the Triton?

Imperial Japanese Navy Page historian Allyn Nevitt performed a thoroughly thorough analysis of the movements of 35 Japanese destroyers in the South Pacific in the time period around March 15, 1943, the date of the Triton’s reported sinking. Nevitt could not identify the “three destroyers” responsible, but leaves open the possibility that it could have been a combination of small combatants, perhaps including destroyers.[64] Nevitt also suggests the source of the report of the mysterious “three destroyers.”[65]

Japanese records speak of two rather curious convoys that bear some importance to the mystery of the Triton. One was called “Hansa No. 1,” with transports No. 1 Shinsei, Momoyama, Yasushima, Oyo, Aso, Teiryu, and Sydney Marus, carrying elements of the 20th Division to, oddly enough, Hansa Bay, eastern New Guinea. It left Palau on March 6 with the escort of destroyers Kazagumo, carrying the commander of Destroyer Division 10, Captain Yoshimura Matake; Yugumo, Satsuki, Akigumo, and Samidare. It was while escorting this convoy that on March 11 the Satsuki dropped six depth charges on a suspected submarine, with “results unknown.”[66]

The convoy arrived at Hansa Bay on March 12. Offloading was completed in a day. Captain Yoshimura was ordered to take the Kazagumo, Yugumo, and Satsuki to Rabaul, where they were to make supply runs in the Solomons, so they headed out together. The Akigumo and Samidare got the privilege of escorting the transports back to Palau. But on their way out, at about 6:30 pm on March 13, they were caught by five B-17s from the 5th Air Force, who managed to hit the Momoyama Maru, killing nine and leaving the transport disabled. The Akigumo took survivors on board and scuttled the derelict with one torpedo.[67]

This convoy was the likely genesis for the warning of “three destroyers heading north” toward the Triton’s patrol area. Those three destroyers were the Kazagumo, Yugumo, and Satsuki heading to Rabaul. Either the B-17s or the submarine Greenling, who had witnessed the attack on the Momoyama Maru, probably saw Captain Yoshimura’s three destroyers in the distance and called it in.[68]

With one mystery potentially explained, it is the second convoy that is much more interesting and rather confusing. This convoy left Rabaul on March 12, 1943, bound for Palau. The convoy consisted of the Tonei, Asaka, Nishiyama (Seizan), Momoha, Toho, Tasmania , and Florida Maru s, with an escort of the old destroyer Mochizuki and the repair vessel Nagaura.[69] Not part of the convoy were the subchasers Ch-23 and Ch-24, but they left Rabaul the same day to meet an incoming convoy for escort.[70]

The trip was uneventful until March 15, when at 12:15 pm (11:15 am Tokyo Time), the Momoha Maru was hit by a torpedo in the engine room and lost power. She sank five minutes later, with no casualties.[71] The Nagaura was detailed to pick up survivors.

The available information starts to get much more complicated and much less certain after the sinking of the Momoha Maru. The escorts were transferring the survivors to the Florida Maru when, at 2:55 pm (1:55 pm Tokyo Time), the Florida Maru took a torpedo and was disabled. Subchaser Ch-23, though not part of the convoy, was directed to pick up her survivors.[72]

What stands out from the Japanese records so far is that they do not match the timeline from the Trigger. To quote the submarine’s report in relevant part:

1215 (K)

Fired three stern tubes at the leading ship in the right hand column at 1600 yards 90 port track. Two hits. While these torpedoes were on the way got set up on the leading ship of the left hand column. The last zig had placed the columns in echelon so that the angle on the bow of our new target was 10 degrees starboard, relative bearing 300 and at a range of 2000 yards. The first firing, due to smooth sea, disclosed our position. Our new target headed for us. There was not time for us to turn.

1217 (K)

Fired three bow tubes at zero angle on the bow range 700 yards, gyro angles about 45 degrees. Used normal dispersion as the spread. Two hits. Went deep on firing to avoid the target. The escorts were after us with depth charges instantly, alternating listening and dropping. The sea was smooth and sound conditions excellent so could not at the time come up to take a look. During the quiet periods the sounds associated with the breaking up and sinking of a ship were heard in the direction of our first target.

According to Lieutenant Commander Benson, the Trigger fired three of her stern tubes, recording two hits on one target. Two minutes later, he fired three bow tubes at a second target at an angle almost right on the bow, a “down-the-throat shot.” Benson recorded two hits, which is a good trick on a down-the-throat shot, but he was only 700 yards away. He had the submarine dive as soon as the torpedoes were launched because the target was only 700 yards away, so he evidently did not see the hits, but he heard what he thought were hits. While submerged, the Trigger heard the sounds of a ship breaking up from the direction of her first target, which would have been the Momoha Maru.

But what about the Florida Maru, later torpedoed as she took on survivors of the Momoha Maru ? Lieutenant Commander Benson indicates that the time period between the torpedoing of the Momoha Maru and the torpedoing of the Florida Maru was about two minutes. The Japanese records indicate that time interval was closer to two hours. Even if the Japanese records are off in terms of time – and they often are, while submarine records tend to be much more precise due to the deliberate nature of their attacks – there was at least enough time between the attacks for the Japanese to start transferring survivors of the Momoha Maru to the Florida Maru, which suggests the Japanese timeline of about two hours is much closer to the truth.

Moreover, according to Lieutenant Commander Benson, after her two torpedo attacks two minutes apart, the Trigger waited until well after nightfall to attempt another attack. Two attacks, to be precise, neither of which was successful.

When comparing these timelines, specifically Lieutenant Commander Benson’s two minutes versus the Japanese two hours, it is hard to escape the conclusion that the submarine Trigger did not torpedo and damage the Florida Maru. Whatever Benson thought were two torpedo hits on his second target, which may not have even been the Florida Maru, were not hits. Maybe premature detonations, which was common to the Mark 14 torpedo – just ask the Argonaut – but not hits.

Yet even if they were not hits, it is not disputed that the Florida Maru was torpedoed. If the Trigger did not torpedo her, someone torpedoed her. Who was it?

It may help to look at the Trigger’s timeline again for the period after Lieutenant Commander Benson’s attack on the second target:

1420 (K)

Nothing in sound. Came to periscope depth. Nothing in sight.

1437 (K)

Depth charges at a distance. Nothing in sight. Assumed they were from planes. These distant explosions continued for the next hour.

Benson heard these “distant explosions” starting about two hours after his first attack. That is more consistent with the Japanese timeline of an almost two-hour interval between the hits on the Momoha and Florida Maru s. These distant explosions lasted about an hour. He may not have heard the torpedoing of the Florida Maru, but he definitely heard the Japanese antisubmarine counterattack in response.

To go back to the Trigger’s timeline one more time:

1705 (K)

Could now see a ship’s masts. Continued the approach submerged. The new target was a ship which looked like our earlier second target, the one which we have fired and hit twice at 700 yards zero angle on the bow. Alongside of her on the far side and sticking out astern was the smaller ship resembling the freighter which had been third ship in the left hand column. This smaller ship was furnishing buoyancy and propulsion; speed about 2 knots, course 190. The pair was being protected by two small destroyers or corvettes. Continued the approach.

What Benson saw now was the damaged Florida Maru, moving very slowly, possibly with the current. Not unreasonably, Benson thought the Florida Maru was wallowing in the water as a result of his earlier attack. More interesting here are the three ships around the staggering transport. Two were “small destroyers or corvettes,” which is a description consistent with a subchaser. The Ch-23 had been assigned to help the Florida Maru. Was the Ch-24 with her?

Also of interest is what Lieutenant Commander Benson describes as “the smaller ship resembling the freighter” that was alongside the Florida Maru. That ship appears to have been the Tonei Maru, who at around this time tried to take the Florida Maru in tow. Evidently, that attempt failed or was abandoned for the time being. The Florida Maru was abandoned as well, and the convoy moved on.

The Japanese records, incomplete and scattered though they may be, contain a deceptively cryptic note from March 15, 1943, in the Tabular Record of Movement for the subchaser Ch-24. Northwest of the Admiralty Islands, it says, subchasers Ch-24 and Ch-22 and the destroyer Satsuki attacked and sank a submarine. The Imperial Japanese Navy Page’s Bob Hackett, Sander Kingsepp, and Peter Cundall, who assembled the TROM, note that the submarine was probably the Triton, but also point out that the Ch-22 and Ch-24 made multiple attacks, any one of which could have been against the Triton, and their target remains unidentified. Finally, there is this very interesting comment: “At about that time, Florida Maru was torpedoed and taken in tow by Tonei Maru. If Florida Maru's attacker was, indeed, Triton, the submarine probably was Ch-24 and Ch-22’s victim.”[73]

To emphasize just how incomplete the Japanese records are, it should be noted that this entry was only in the TROM of the Ch-24. It was not in the TROMs of either the Ch-22 or the Satsuki. This is the first appearance for the Ch-22 in the mystery of the Triton. The Ch-22 had left Wewak, New Guinea, earlier on March 15 headed for Palau.[74] She was in the area and thus seems to have been vectored in to deal with the submarine attacks on the convoy.

The Ch-24 would be making her first appearance in the Triton mystery. She had left Rabaul with the Ch-23 on March 12, the same day as the convoy, and so was close enough so that when the Florida Maru was torpedoed, the Ch-23 was vectored in to assist. She was likely one of the two “small destroyers or corvettes,” which Lieutenant Commander Benson saw with the damaged Florida Maru. Since the Ch-24 had left Rabaul with the Ch-23 and was in the area, likely she was the other one.

The Satsuki makes her second appearance in the Triton mystery. It will be recalled that, while escorting convoy Hansa No. 1 on March 11, she dropped a half dozen depth charges on a suspected submarine with “results unknown.” This was the same day that the Triton reported sighting “Two groups of smokes, five or more ships each, plus escorts.....Am chasing.”[75] Was the Triton tracking the convoy Hansa No. 1? Did the Satsuki depth charge the Triton? Did the Satsuki sink the Triton?

After dropping the convoy off in Hansa on March 12, the Satsuki left on March 13, in the company of the Kazagumo and Yugumo. They were headed to Rabaul to allegedly run supplies in the Solomons. Allegedly. Because the Satsuki next appears on March 29 in Kavieng, from where she was to join the destroyers Fumizuki, Minazuki, and Nagatsuki in making a troop transport run to Finschhafen. The run was cancelled when General MacArthur’s 5th Bomber Command made a large raid on the port and town of Finschhafen on March 30.

What the Satsuki was doing between March 14 and March 29 is anyone’s guess. Those incomplete Japanese records again. Was she running supplies in the Solomons as planned? Was she conducting antisubmarine operations around Rabaul? Was she heading to Kavieng? Was she making antisubmarine patrols between Rabaul and Kavieng? Was she delivering pizzas to the Admiralties? If she was vectored in to help with the submarine attacks on the convoy, the Satsuki could not have been in Rabaul, as that was too far away. She had to be in the area. The reference to the Satsuki in the sinking of the Triton is not entirely clear, as it’s possible her role could have been limited to the original March 11 attack and the damage it caused. To that end, Vernon Miller, in his respected analysis of the sinkings of US submarines in the Pacific War, says the Japanese are specific that there were no destroyer antisubmarine attacks on March 15 and thus credit the sinking of the Triton to the Ch-24.[76]

Also of interest here, though usually overlooked, is the old destroyer Mochizuki, the main antisubmarine platform in the convoy. Her records were apparently lost after the war, so her TROM had to be reconstructed from the records of other ships involved with her, as possible.[77] What antisubmarine attacks she made during this period, if any, are unknown. She could, conceivably, have made an attack on the Triton, the records of which have been lost. It is known that on March 18 she and the freighter Tonei Maru returned to retrieve the abandoned Florida Maru, which was still afloat. Another proud moment for the Mark 14 torpedo. With the Mochizuki as escort, the Tonei Maru towed the derelict to Möwe anchorage, New Hanover, where they arrived March 24.[78]

What should be clear from this complex and long-winded analysis is just how difficult it can be to solve mysteries such as the loss of the Triton (and, for that matter, the Grampus) in which records are incomplete, missing, or contradictory. You pull one thread out of the mystery and it leads you to another mystery, which must be solved first in order to solve the main mystery. Be that as it may, with the arduous analysis complete, a likely scenario can now be reconstructed, with certain parts definite and certain parts little more than speculation, but necessary to fill in gaps.

On March 11, the Triton reported that she had sighted “Two groups of smokes, five or more ships each, plus escorts.....Am chasing.” That was the last definite time anyone heard from the Triton. The convoy Hansa No. 1 was passing through the area in which the Triton was operating. It may have been the convoy the Triton was chasing. The escorting destroyer Satsuki dropped six depth charges on a suspected submarine with “results unknown.” If the Triton was indeed chasing Hansa No. 1, she was likely the target of this attack. The attack definitely did not sink the Triton, but it may have damaged her communications, making her unable to send any more messages. Like, maybe, the Grampus.

Hansa No. 1 reached Hansa on March 12, and three of the escorting destroyers – Kazagumo, Yugumo, and Satsuki – left on March 13 to go to Rabaul. These three destroyers were sighted either by B-17s attacking the convoy ships or by the submarine Greenling, and their location and course was reported. This became the basis for the warning sent out to the Triton that “three destroyers” were headed northward toward her area. In fact, the destroyers turned eastward toward Rabaul, where they arrived on March 14.

Also on March 12, a second convoy – one that included the Momoha, Florida, and Tonei Maru s and the destroyer Mochizuki, left Rabaul. Also leaving Rabaul on March 12 were subchasers Ch-23 and Ch-24 to meet a third, incoming convoy. They probably offered their antisubmarine services to the Mochizuki's convoy until they had to break off for their own convoy escort.

On March 15, the convoy was attacked by the submarine Trigger. Three torpedoes from the submarine’s stern tubes sank the Momoha Maru. Three torpedoes fired from the Trigger’s bow tubes two minutes later were believed to have resulted in two hits on a second ship, probably the Florida Maru, but were probably premature explosions and definitely did not hit the Florida Maru or anyone else. The Japanese rounded up survivors from the Momoha Maru.

About two hours later, the survivors were being transferred to the Florida Maru when the Florida Maru herself took at least one torpedo. The Ch-23 was ordered in to assist the stricken freighter. Meanwhile, the Ch-22, heading for Palau, was in the area and was apparently vectored in to assist. The Mochizuki was obviously also in the area and could have been involved in locating the sub, but seems to have continued escorting the remaining ships. The Ch-24, who had been with Ch-23, went out to help her sister subchaser Ch-22 get a fix on the submarine. Since, according to her timeline, the Trigger could not have launched the attack that damaged the Florida Maru, it had to be a different submarine. There was only one other American submarine in the area at that time – in Trigger’s patrol zone, in fact. That submarine was the Triton. The Florida Maru's assailant had to have been the Triton.

The Florida Maru would be the Triton’s last victim. The Ch-22 and Ch-24 were able to get a fix on the submarine’s position, as the smooth, glassy water that had revealed the Trigger would have also revealed the Triton. The subchasers engaged in a depth charge attack, which the Trigger heard.[79] It would seem that after these attacks, the Japanese saw “a great quantity of oil, pieces of wood, cork and manufactured goods bearing the mark ‘Made in U.S.A’.” If this report does or did exist and was not invented by the US Navy, it appears to have come from either Ch-22 or, more likely, Ch-24. Navy investigators perhaps seized on that report and called it a day. Perhaps.

There remains one other curious and overlapping possibility. As detailed by the Australian historian Peter Dunn on the website Australia & War, there have been rumors that the Triton was actually sunk on or about March 29, 1943, off Moreton Island, which serves as a breakwater for Moreton Bay, on which Brisbane is located. What caused the Triton’s loss is unconfirmed but there is a decent probability that it was friendly fire. It was this probability that she was lost to friendly fire that, according to the rumor, caused the Australian military to cover up the incident.

The genesis of the rumor is difficult to track; the rumor itself is more than a little messy, starting with its timeline. According to the rumor, on March 25 the Triton was ordered to leave her patrol area and return to Brisbane, to arrive on April 2. How the submarine was supposed to make it from the area northwest of Rabaul to Brisbane in eight days is not explained. To be sure, she was ordered on March 25 to return.[80] However, it should be noted that the Dictionary of American Fighting Ships entry for the Triton, rather curiously, does not mention the date of the last transmission to the Triton, as it does with other submarines. It may be that March 25 was a rebroadcast of an earlier order to return.

The rumor continues. The Triton was expected to be at the wharf in Brisbane on March 29. Why she would be there on March 29 and not on April 2 as the rumor says is also not explained. That timeline again. Supposedly, the Triton had made contact with Brisbane military authorities at about midnight on March 29, stating that she had spotted the Moreton Light and had surfaced and was entering Moreton Bay at 12 knots, estimating her arrival at the base to be at 8:00 am. Based on this message, a pilot boat was dispatched to meet the Triton and a welcoming committee was formed to greet her at the dock.

However, the pilot boat never found the Triton. The welcoming committee, including crewmen from US Navy submarines Albacore, Grampus, Grouper, Peto, Grayback, and Growler did, it seems, show up at the dock, complete with band, mail, fresh fruit, and ice cream. This was and is common for returning ships and, especially, submarines. The people at the dock were aware that nothing had been heard from the Triton, but attributed this to radio issues, which suggests just how common radio issues were in submarines. But the Triton never showed. One man at that dock waiting to board the Triton said he was told she had been sunk by accident when trying to enter Brisbane. He was reassigned that same day to the Albacore.

Not surprisingly, Australian military officials do not much appreciate the rumor. The Australian Defence Department refers inquiries concerning the rumor to the Australian War Memorial. The Memorial says it was highly unlikely Australian fire had sunk the submarine, and if there had been a cover-up during the war, the truth would have come out in the intervening years. In other words, if there was a cover up, someone would have said so. That line of defense (or defence, in Australia’s case) does not work when someone actually is saying so now.

Much better to attack the rumor based on known facts, or, in the legal world, stipulated facts. That a welcoming committee showed up at the dock the morning of March 29 does not seem to be disputed, which is curious by itself. Nevertheless, this is where the rumor really breaks down, as only four of the submarines mentioned were there. Of the remaining two, the Grayback was still on her patrol; the other, Grampus, supposedly last seen by the Grayback the night of March 5-6, was presumably sunk. Or was she? More on that anon.

In support of the rumor, or something, is a passage from the official Australian history, Units of the Royal Australian Air Force: A Concise History, Volume IV: Maritime and Transport Units has been cited.[81] The passage concerns the No. 71 Squadron of the Royal Australian Air Force, which was based near Brisbane. No. 71 Squadron was the unit responsible for performing antisubmarine patrols off eastern Australia. The passage reads as follows:

The Squadron, and its detached flights, flew anti-submarine patrols over shipping on the east coast. These duties were not without incident. On 17 March, a 250-lb bomb was dropped on a suspected submarine by Sergeant R.N. Walesby and, on the 28th, Walesby and his crew of Sergeant P.K. Yates and Flight Sergeant H.A. Yates reported being fired on by a similar vessel.[82]

How exactly this passage fits the rumor is also never explained. Perhaps the idea that a submarine was off the coast of Australia near Brisbane on March 28 is to suggest that the submarine off the coast was indeed the Triton. Again, though, how exactly the Triton was to get from northwest of Rabaul to Brisbane in the three days between March 25 and March 28 is unclear.

In making a possible case for the veracity of the rumor, Dunn points out, “The co-ordination in the recognition and understanding of signals was clearly an issue during the war. It has also been reported that many new pilots sometimes mistook proper signal flares for gun fire and that this might have been the case with USS Triton.”

Moreover, Dunn continues, “There was a mine field installed between Bribie Island and Moreton Island by the British Navy ship ‘Job Fifty-one’. The mine field was connected to a firing button in the detonation room at Caloundra. There was a direct underwater line from the Bribie Forts to the Caloundra Telephone exchange.” A minefield whose detonation was controlled from shore.

To suggest what could have happened, Dunn also points out the experience of Captain William J. Ruhe as detailed in Ruhe’s book, War in the Boats: My WW2 Submarine Battles. Ruhe details two separate instances when his submarine, the S-37, had difficulty getting into Brisbane because of recognition and communications issues with the fort controlling the mines. On both of these occasions, the shore station had complicated things by using an “incorrect challenging procedure.”[83] On one of these occasions, the S-37 [84]

Dunn closes the explanation of the rumor by hypothesizing, “If the USS Triton had been sunk by ‘friendly fire’, it is probably feasible to suggest that such an incident would have been covered up.”

In response, historian Ed Howard has gone so far as to say, “It is inconceivable the U. S. Navy had the ability or reason to conceal such a disaster. They would most likely have conducted a court of inquiry to gather the facts and recommend corrective or punitive measures, as had been done for other friendly fire incidents or accidents.”[85] Howard overstates it a bit, but only a bit. Inconceivable? No. Feasible? Yes. Probable, no, although perhaps not as improbable as we’d like to believe. Not for any nefarious reasons, however.

The danger of friendly fire has always been present in war, even in close quarters combat. Witness the Battle of Salamis in 480 BC in which Queen Artemisia of Halicarnassus, fighting under the Achaemenid Persian banner, in her zeal to escape pursuing Greeks, rammed and sank an allied ship also fighting for the Persians. In the Pacific War, at the November 13, 1942, Naval Battle of Guadalcanal, the disabled cruiser USS Atlanta drifted into the line of fire of the cruiser San Francisco, below the line of sight of her gun director, and a salvo from the San Francisco intended for the Japanese battleship Hiei blasted the Atlanta's bridge instead. More recently, a famous case of friendly fire caused the death of US Army Ranger Pat Tillman, a former American football player for the Arizona Cardinals. Tillman was killed in Afghanistan during a firefight that was a smaller version of Karánsebes in that it was actually the two halves of his own platoon shooting at each other.

Friendly fire cannot be eliminated, at least not without fatally compromising fighting effectiveness. At best, it can be minimized, with investigations of incidents leading to further minimization. With the stealthy nature of their design and their mission, submarines would seem to be especially vulnerable to friendly fire. Indeed, several friendly fire incidents with submarines have already been mentioned above, but truth be told the cases of submarines actually sinking as a result of friendly fire are relatively few.

The first such case involving a US submarine is generally believed to be that of the Dorado in October 1943. The Dorado left New London, Connecticut, on her way to the Pacific. Under strict radio silence, she was due to make a stop in Panama on October 14. Bases along the east coast and in the Caribbean were informed of the Dorado's passage, and a safety zone that restricted bombing was imposed at her projected waypoints. She never arrived in Panama and was never heard from again.

On the dark and stormy night of October 12, a Martin PBM Mariner flying boat out of Guantanamo Bay, Cuba, flying cover for a convoy spotted what she thought was a German mine-laying submarine. She attacked the boat, with at least two depth charges detonating. The PBM later made a second pass and dropped flares on the site, seeing only bubbles and foam that she attributed to the submarine diving. Two hours later, she spotted a surfaced submarine again. She attempted to exchange recognition signals but was fired upon instead and lost contact.

A Navy investigation found that the target of the PBM’s bombing attack was actually the Dorado. The PBM had been aware of the safety zone around the Dorado, but had been given bad information as to its location. That was about all that could be ascertained. An extensive Navy search after the attack found no debris, oil slicks, or survivors. Complicating matters was the presence in the area of a German minelaying submarine, U-214 – exactly the type of submarine the PBM had thought she was attacking. Postwar investigation of the U-214's logs found she was not the target of the attack, but had seen the flares dropped by the PBM and moved in to investigate. It was then that she was spotted by the PBM, at which she fired her antiaircraft guns.

While there was no indication the U-214 had directly attacked an enemy submarine, she had been a busy boat, laying mines off Colon, Panama, and in the Caribbean Sea west of the Antilles. It was in connection with this second group, two cumbersome EMS-type drifting mines that had to be assembled on deck and laid by hand, that the U-214 heard a series of explosions the afternoon of October 15. No ships are known to have been sunk by these mines; the cause of the explosions remains a mystery, but they were in the general area of the Dorado. Some have speculated that it was the Dorado that hit one of the mines and sank as a result.

The first confirmed friendly fire submarine incident and probably the most infamous of the war is the tragic case of the US submarine Seawolf.

The Seawolf left Brisbane on September 21, 1944, for her 15th war patrol. Eight days later she stopped at Manus Island to embark a US Army reconnaissance party and supplies to drop them off at Samar to scout for the upcoming invasion of the Philippines. During this time, Allied forces were invading Palau and Morotai, the latter of which was along Seawolf's route to the Philippines. The Allies established submarine “safety lanes” through which antisubmarine attacks were prohibited unless the submarine was ascertained to be enemy. Allied submarines scheduled to pass through the area and their relevant itineraries were revealed to local commanders. Unknown is whether it occurred to anyone that having a submarine safety lane so close to surface units was asking for trouble.

And trouble obliged. For the Japanese had sent five RO-class submarines to attack the Morotai invasion forces. One, later identified as the RO-41, picked her way into the submarine safety lane intended to protect US submarines and launched torpedoes at the aircraft carrier Midway.[86] The torpedoes missed the Midway, but hit the destroyer escort Shelton. The destroyer escort Richard M. Rowell dropped some depth charges on the submarine and proceeded to take off the Shelton’s crew. Later, while under tow, the Shelton would sink.

The Midway, finding the submarine attack on her more offensive than anything else, launched a vicious counterattack. Two of her TBF Avenger torpedo bombers found a submerged submarine. It must have been the Japanese sub that attacked. It was in the safety lane, but safety lane, schmafety lane. One Avenger crashed while making a bombing run on the submarine, the other managed to drop its bomb but missed – fortunately; the submarine was the USS Stingray. Another friendly fire incident. As Richard O’Kane said, submarines have no friends.

Meanwhile, the Seawolf had arrived in the area and exchanged recognition signals with the submarine Narwhal by radar. Rough seas had slowed her transit, so now the Seawolf was about a day behind schedule. She had reported her delay, but the report apparently never made it to the surface units.

At around 11:00 am, one of the Midway's aircraft spotted a submarine. It was in the safety lane, but as far as anyone knew, there were no US submarines within 70 miles. The TBF Avenger dropped two bombs and a dye marker. The bombs missed and the submarine dove.

The destroyer escort Richard M. Rowell sailed over and detected the submarine on its sonar at 1:10 pm. The Rowell proceeded with a depth charge attack. The sonar operator reported the submarine was behaving rather curiously, taking no evasive action. Moreover, his equipment was receiving beeping signals from the submarine that sounded like Morse code – long dots and dashes. The Rowell's skipper Commander Harry A. Barnard, Jr., later reported, “the stuttering transmission bore no resemblance to the proper recognition signal[,]” instead he believed the signals were an attempt to jam the Rowell's sonar. That was just … insulting. The Rowell went on to fire its “hedgehogs”. Not actual hedgehogs, but Mark 10 “hedgehog” mortars that blasted the water with 24 projectiles. Barnard later reported, “Three explosions heard. Two large boils (bubbles) observed off port beam. Debris observed in the boils.” Clearly, he had sunk the target of his attack.

But his target was not the Japanese submarine RO-41, which was not attacked and later returned to Japan. Urgent calls were made to the four submarines believed to be in the safety lane at the time of the attack to report their position. Three responded. The fourth, the Seawolf, was never heard from again.

The US Navy quickly held a board of inquiry at Manus. Despite Commander Barnard’s protestations to the contrary, it was determined that he had attacked and sunk the Seawolf. Barnard was censured for making insufficient efforts to identify his target, for dismissing the sound signals, and for attacking the Seawolf.

The US has been fairly open and honest about friendly fire incidents among US units. Friendly fire among units of different nationalities are often handled differently, perhaps not as openly. During the August 9, 1942, Battle of Savo Island, the Australian cruiser Canberra was sunk, along with three US Navy cruisers, in a Japanese surprise nighttime attack. Inquiries by both the US Navy and the Royal Australian Navy concluded the Canberra was disabled and sunk by Japanese gunfire from her port side that penetrated completely through the ship, causing her to list to starboard. The cases were quickly closed – too quickly for the survivors of the Canberra, who reported a mysterious underwater explosion on the cruiser’s starboard side, away from the Japanese. The survivors expressed a belief that a torpedo fired by the destroyer USS Bagley had actually hit the cruiser on her starboard side and knocked out all her systems. But did the two wartime allies, whose relationship was firm but not always smooth, want to discuss this serious potential friendly fire incident between them during the war, with political implications that could inflame the populace of one against the other? Coverups can have legitimate reasons.

The case against a coverup of the potential accidental sinking of the Triton by Australian forces is not helped by the conduct of US Navy investigators into the sinking of the Triton. As detailed earlier, the official explanation of the loss of the Triton is suspiciously explicit as to how she was sunk while suspiciously vague as to who was responsible for the sinking. The US Navy appears to quote directly from a Japanese report. Someone filed that report. Who was it? It should not take Allyn Nevitt and the crew at the Imperial Japanese Navy Page to sift through the records of every Combined Fleet ship in the South Pacific to determine who could have filed that report. It should be the US Navy, who quoted the report, to say who did file that report. Who filed the report of the sinking of the Triton?

Nor is that the only instance of curious conduct by the US Navy in the investigation of the loss of a US Navy submarine. Consider the Grampus. Without directly saying so, US Navy investigators have allowed the inference that the Grampus, if she had survived the air attack of February 19, was most likely sunk the night of March 5-6, 1943, by the destroyers Murasame and Minegumo immediately before the destroyers themselves were sunk in the Kula Gulf that same night. Since the destroyers were sunk, they left no logs, so the sinking cannot be confirmed. If one were trying to cover up a friendly fire incident, the lack of surviving logs would make it convenient to pin the blame on them. At the same time, the Navy had to be aware that the survivors of the destroyers made no mention of any antisubmarine action. Moreover, the submarine Grayback, who did see a submarine that she believed was the Grampus, witnessed no combat that night other than the attack by US surface ships that sank the Murasame and Minegumo. It is reasonable to deduce that the Murasame and Minegumo could not have sunk the Grampus. The missing submarine survived the night of March 5-6. The Grampus obviously sank somewhere. Where? Why?

It is true that the Grampus is one of the more difficult cases due to a relative lack of evidence because she made no transmissions after she left the Brisbane area on February 12. But that does not explain the relative lack of curiosity into her fate, or the seeming reluctance of anyone looking into what happened to the Grampus, unlike the case of the Triton, to consider that she may have been, and likely was, having problems with her radio suite.[87]

This is where that passage Peter Dunn quotes from the official Australian history, Units of the Royal Australian Air Force: A Concise History, Volume IV: Maritime and Transport Units, may actually offer a clue. To restate, No. 71 Squadron was responsible for antisubmarine patrols off Brisbane. According to the history:

On 17 March, a 250-lb bomb was dropped on a suspected submarine by Sergeant R.N. Walesby and, on the 28th, Walesby and his crew of Sergeant P.K. Yates and Flight Sergeant H.A. Yates reported being fired on by a similar vessel.[88]

While Dunn is focused on the Triton and the March 28 incident, the March 17 attack should also perk some interest. March 17 was 11 days after the last reported sighting of the Grampus in the Vella Gulf, 10 days after she had been directed to report in. She had sufficient time to return to the Brisbane area. If one accepts the hypothesis that the Grampus was having difficulties with her radio suite, that she could receive but not transmit, which was not an uncommon problem, it’s easy to develop a scenario in which she received the March 7 recall order and did the most logical thing: head back to Brisbane. In so doing, she could have been off Brisbane on March 17.

Which leads to the question: was the Grampus the target of the March 17 attack? If so, was she sunk by this attack?

There is no way to definitively answer the question. There are no details available as to the results of that attack. Presumably, if Sergeant Walesby had seen debris or an oil slick, any reason to claim credit for a sinking, it would have shown up in the aforementioned Volume IV of A Concise History. Then again, if he knew he had sunk a friendly submarine, he might have been more circumspect about reporting it. There is simply no way to know, not without further research into the March 17 attack that at this point in time is likely impossible.

Beyond a friendly fire attack, what else could have sunk the Triton and, for that matter, the Grampus? It should be pointed out that even the rumor does not go so far as to say what exactly sunk the Triton, only saying she was “accidentally” sunk. There are a few possibilities, but none stands out as realistic.

  • The submarine could have collided with a ship, always a possibility in a crowded wartime harbor; except there are two parties to any collision, and no second party has revealed itself.
  • The submarine could have been attacked by a Japanese submarine; while Japanese submarines did operate off of Brisbane during this period, there appears to be no record of such an attack.
  • The submarine could have struck a mine. This is the only feasible possibility, and just barely so. While there most definitely was a minefield protecting Brisbane and a Japanese mine could have been laid somewhere, no report of an explosion off Brisbane at this time has been revealed.

There is evidence of aerial antisubmarine activity during this period: March 17 and March 28. And it’s not clear if the March 28 encounter included an attack.

It all adds up to a possibility that should be investigated, but less so for the Triton than for the Grampus. It is already clear that the Triton was operational on March 15 and made an attack that damaged the Florida Maru. Afterwards, the Trigger heard a Japanese counterattack. The only recorded event that could have sunk the Triton is that March 15 attack overheard by the Trigger. Though the rumor says otherwise, there is no confirmed evidence that Triton communicated after March 15. Thus, even with the rumor, it is reasonable to conclude that the Triton sank as a result of that counterattack.

The Grampus is, well, a different story, though perhaps the Triton rumor can help focus it. One can assume here that the Grampus was having communications problems that allowed her to receive but not send messages. The only recorded event that could have sunk the Grampus is the February 19 Japanese air attack. By itself, that evidence is inconclusive at best due to the lack of debris. After that there were two sightings and one attack in areas where the Grampus was supposed to be, which indicates that these sightings were of the Grampus. While the theory of an attack by the Murasame and Minegumo the night of March 5-6 as being responsible for the Grampus’ loss is attractive, the lack of any witness to the attack from either destroyer or the Grayback means this potential cause must be ruled out. That leaves … nothing.

Until that rumor is considered. There is so much wrong with the rumor. But there is usually a kernel of truth to such rumors, even if it is garbled beyond all recognition. There was an antisubmarine attack by Australian aircraft on March 17. The results of that attack are ambiguous. What if the Triton rumor was actually a garbled recollection of a mistaken antisubmarine attack that sank the Grampus? A rumor in which the Triton was substituted for the Grampus because it came out on the day the Triton was to dock? Possible. Likely? No. But possible. And aside from some fatal operational accident after March 6, 1943, that March 17 attack is the only possible lead as to the Grampus’ fate. It warrants as much examination as is possible at this stage.

As for the Argonaut and Amberjack, there has never been much question that the Argonaut was sunk on January 10, 1943. It can be safely deduced that the Amberjack was the victim of the February 16 antisubmarine attack because of the behavior of the target and the presence of so much debris, some of which was actually recovered after the attack; the lack of sightings of submarines in areas where the Amberjack was supposed to be; and the elimination of the Grampus as a possible target.

The loss of four submarines in so short a time was a horrible beginning for Captain Fife. He was uniformly acknowledged as an odd duck, the descriptions of whom suggest a touch of Asperger’s. One division commander described him thus:

Surrounded by hard-drinking submariners fresh from combat, he was a strange, solitary, almost lonely figure. He didn’t drink or fraternize. He never seemed concerned about people. When a submarine came in from patrol, he wanted to know about the condition of the battery of the engines or the periscopes or electrical equipment. (Emphasis in original).[89]

Captain Fife was also universally acknowledged as a tough, perhaps stubborn officer. And his initial response did nothing to lessen that reputation. “Tough luck,” he wrote, “but they can’t get Japs without taking chances … don’t think the time has arrived to inject caution into the system because it is too difficult to overcome again.”[90]

It was a cold, but not revealing statement. Fife took the losses hard. He offered to resign, but his superior at Southwest Pacific Rear Admiral Arthur Carpender turned him down. Determined to get to the bottom of it, Fife went so far as to personally fly over enemy territory – six times. Attacked on two of those occasions. But his efforts to find out what had gone wrong were unsuccessful.[91]

His superiors noticed. His predecessor Rear Admiral Christie offered to head an independent investigation into Captain Fife, but Admiral Carpender instead appointed Captain Allan McCann to head an “informal” investigation. McCann did look into it “informally,” and gave his “informal” to Carpender, who locked it away in his office and did not share it with anyone.

Which did not sit well with Admiral Christie. Conducting an investigation of his own, Christie wrote to Fife:

The probability that the enemy is able to derive information of value from a large number of submarine operational dispatches must be reckoned with. A study of the patrols of the Grampus and Amberjack focuses attention on one method of operation which may conceivably have contributed to the loss of these vessels. Radio dispatches concerning Grampus and Amberjack totaled 106. Forty-six of these were reports of positions of submarines for higher command. Many of these dispatches gave specific names, locations, and times. … We have in numerous instances intercepted enemy transmissions revealing the positions of our submarines and ordering counter measures. Submarine transmissions have been repeatedly and expertly DF-ed. The obvious solution is rigid adherence to the fundamentals of radio security, particularly the elimination of all but essential traffic. The amount of traffic put on the air should be a minimum for reasons of external security. The number of addresses should be a minimum for reasons of internal security.[92]

Captain Fife defended himself in detail, stating that movement orders were usually short and locations were along traffic routes, not specific positions.[93] But he did seem to take the loss of the Argonaut, Amberjack, Grampus, and Triton to heart. After this experience, Fife loosened the reins quite a bit, the micromanagement ended, for the most part; and the number of transmissions went down.

Long after the war, Fife would say, “During the heat of this Solomons campaign, all the submarine captains I had out there had responded extremely well, and they were very anxious to inflict damage on the Japs, working down the Slot in the Solomons, and they may have gone over the other way, where they had abandoned caution.”[94]

Captain McCann’s report on the losses of the four submarines never did see the light of day. Years later, McCann would give a short summary of his report. “I exonerated Fife completely,” he said. “The losses were in no way his fault.”

To which he added this rather curious line: “The boats may have been sunk by our own aircraft. There was no way to tell.”[95]

* * *

Show Notes
* * *

© 2024 Jeffrey Cox

Published online: 07/26/2020.

About the author:
Jeffrey R. Cox is a litigation attorney in Indianapolis, IN, and an independent military historian specializing in World War II, ancient Greece and ancient Rome, which he has studied for decades. He holds a bachelor’s degree in National Security Policy Studies from The Ohio State University.

He is the author of:
  Blazing Star, Setting Sun: The Guadalcanal-Solomons Campaign November 1942–March 1943

  Morning Star, Midnight Sun: The Early Guadalcanal/Solomons Campaign August-October 1942 (first of the Guadalcanal/Solomons trilogy)

  Rising Sun, Falling Skies: The Disastrous Java Sea Campaign of World War II


* Views expressed by contributors are their own and do not necessarily represent those of MilitaryHistoryOnline.com.

An error has occurred. This application may no longer respond until reloaded. Reload 🗙