A single Japanese aerial torpedo, fitted the evening before, was hanging under a Nakajima attack aircraft in the sealed hangar deck of the carrier Akagi on the morning of June 4, 1942. It came off that morning, and nothing that happened afterward ever put it back. The entire Battle of Midway can be traced through that one weapon: why it was under that aircraft, why it was removed, what replaced it, and why the work could not be undone. The men who survived the morning gave a different account of all of it, and they gave it for 50 years.
It was wrong in the one place where being wrong mattered. Before the fleet sailed from the Inland Sea, the commander of the Combined Fleet, Admiral Isoroku Yamamoto, gave Vice Admiral Chuichi Nagumo an instruction about those torpedoes. Half of the torpedo aircraft in the carrier force were to remain armed against ships at all times, held back in case an American fleet appeared at Midway. Yamamoto said it aloud but never put it into the written operation order.
Two of his own staff officers said afterward that they wished he had. Nagumo sailed believing the Americans did not know he was coming. There was a reason for that. As late as June 1, the Naval General Staff in Tokyo was still telling the Combined Fleet that the American carriers were down in the South Pacific, and a Japanese search aircraft had reported seeing the Enterprise and Hornet east of the Solomons on May 16.
Admiral Chester Nimitz had arranged for them to be seen and then kept a stream of invented signal traffic running in that ocean after they had gone. Tokyo changed its mind on June 2. The Naval General Staff now thought the Americans had worked out the Midway operation and might be waiting with carriers, and it sent that judgment as an urgent signal addressed to both Yamamoto and Nagumo. Yamamoto received it.
The Akagi did not, because she carried a small island for her size and had no antennas long enough to take low-frequency long-distance traffic. A great deal of what the fleet flagship was hearing never arrived aboard her at all. Yamamoto assumed she was hearing it anyway, and his senior staff officer talked him out of passing it on on the grounds that radio silence mattered more. Nagumo’s chief of staff, Rear Admiral Ryunosuke Kusaka, had asked before sailing that the Yamato relay anything important across to them and had been turned down.
The most powerful carrier force in the world went out unable to hear the man who had sent it. So when Nagumo spoke to his staff a few hours before the first launch, he told them what the paperwork he had been given supported. The enemy did not know their plans. No serious force with carriers at its center was thought to be anywhere near.
The morning began the way that assumption said it should. At 4:00 in the dark, the four carriers put 108 machines into the air against Midway in equal thirds of dive bombers, torpedo aircraft, and fighters under Lieutenant Joichi Tomonaga. Seven aircraft went out on search lines across a wide arc to the east and south. It was a thin search by American standards.
Nagumo was not hunting a fleet; he was checking a box. The weather did the rest. One of the two search aircraft off the cruiser Chikuma flew almost over an American task force at around 6:00 and saw nothing at all because the cloud below it was thick and low and unbroken. The other Chikuma machine never finished its line.
It turned back at 6:35 with a storm sitting across its path. At 7:00, Tomonaga came up on the radio from over the island to report that Midway had been hit, that Midway was still working, and to recommend a second strike. At 7:15, Nagumo gave the order, and on four hangar decks, the mechanics began taking the anti-ship weapons off the reserve aircraft and fitting weapons meant for a runway and a fuel dump. The torpedo came down onto its trolley.
What has occupied naval historians ever since is a narrower question than why Nagumo gave that order, which was defensible. It is why the 20 seconds it takes to say the words could not be taken back over the next three hours. The first Americans arrived at 7:10, and they did not come from a carrier. Six Navy Avengers and four Army Marauders, torpedo aircraft belonging to two different services flying off one island airstrip, went in against the mobile force with no fighters and no plan between them.
Five of the six Avengers went down. The sixth came home with its control surfaces shredded, its gunner dead, and its pilot and radio operator both injured. Two of the four Marauders survived. Nothing was hit.
The Japanese fighters that killed them had to be brought back down to the decks afterward because a Zero carried 60 rounds for each of its two cannon. A Zero that is fought is a Zero that is out of ammunition. That is a detail about a fighter aircraft, and it is also the mechanism by which the battle was lost. At 7:28, the number four search aircraft off the cruiser Tone found ships, ten of them, steering southeast at better than 20 knots.
The signal said nothing about carriers. From where that crew were sitting, there were no carriers to see. The signal took its time getting home. The Akagi was not guarding the search frequency, so the cruiser that had launched the aircraft took it in instead, in code, on a set whose operator was not standing by for it, and then passed it across to the flagship by lamp.
Nagumo stated twice in his own report that it did not reach him until about 8:00. Almost every historian who came after him threw that out because the flagship’s message log has him acting on it at 7:45 to 8:00. The Akagi’s flag bridge log went down with the ship. What survives is a composite log that a staff officer named Chuichi Yoshioka stitched together from the signal logs of several ships.
Yoshioka wrote a warning into it that the entries were fragmentary and inconsistent and would not bear much weight. Whichever time is right, the work in the hangars ran on for a while. It could not have been undone in a hurry. In any case, nobody in the Imperial Navy had much practice at this.
Changing a squadron of torpedo aircraft over from torpedoes to 800-kilogram land bombs had been tried once before the operation as an experiment aboard the Hiryu, and it took an hour and a half to do and two hours to undo. The two weapons had nothing in common. A Japanese aerial torpedo was long and slender and weighed nearly a ton, and the land bomb was short and thick. The ribbed rack on the belly of a machine that carried one would not carry the other.
So it had to come off and a different rack had to go on. That alone ran about 20 minutes an aircraft. There were not enough trolleys to work a whole squadron at once. Each ship had enough wheeled carts with jacks on them to serve one division at a time, six on the Akagi and perhaps nine on the Kaga.
So the squadrons went through in shifts, and the carts went round and round. A torpedo came down onto a cart, traveled to a rack, and stayed there. Then the rack under the belly came off. Then a bomb came up.
All of it happened below in the dark because these ships were built with enclosed hangars. American hangars had roll-up sides and could be open to the air. Japanese hangars were boxed in on every side by storerooms and workshops and crew spaces, ventilated by fans with no way to clear exhaust fumes. So no aircraft engine aboard was ever warmed up below decks.
Warming happened up in the open for 15 minutes at the least, and there was no way around it. A radial motor taken to full power cold will wreck itself. So a strike could not be built downstairs and then squirted into the air. Every machine had to be lifted one at a time, spotted, unfolded, chocked, warmed, briefed, checked, and then flown off.
The deck was occupied for the whole of that, and a deck doing that cannot land a fighter. That raises a second question, and it has nothing to do with torpedoes. It is about what would be standing on those four decks when somebody finally turned up overhead. The Americans, without meaning to and without knowing they were doing it, spent the next three hours making sure that never happened.
At 7:55, 16 Marine Dauntlesses out of Midway under Major Lofton Henderson found the Hiryu. More than half his pilots were too green to dive bomb, so they came in on a shallow glide instead, and the Zeros took Henderson on the second pass. Eight machines got home, and six of those never flew again for a near miss. At around 7:45, 15 Army B-17s bombed the Hiryu from 20,000 feet and put 68 tons of high explosive into the ocean around her without touching her.
On the way through, they took photographs, and those photographs are the sole surviving pictures of any American attack on those carriers that day. At 8:20, 11 Marine Vindicators, aircraft so old the pilots made jokes about them, gave up on the big ships and went after the battleship Haruna instead and missed her too. At the same time, the search crew off the Tone, still circling, still watching, sent the signal Nagumo had spent the morning being told he would never need. The enemy, they reported, was accompanied by what appears to be a carrier.
That crew has been criticized by three generations of writers for the 52 minutes it took them to see one. The likeliest explanation is that there was nothing of the kind under them until then. Task Force 16 had begun to string out during its own launch, and the ten ships they had reported were the screening ships, several miles clear of the flat decks and under a thinner sky. They had also cut their own search line short by 65 miles on their own initiative, which is the reason anyone found anything at all that morning.
Having found something, they held station over an American fleet for an hour and a half at real risk to themselves and sent back seven signals in that time. It was careful flying. Nobody in that aircraft lived through the war to explain why they cut the corner. Now Nagumo had a decision and no good version of it.
Rear Admiral Tamon Yamaguchi, commanding the Second Carrier Division, sent across a recommendation to attack at once with whatever would fly, and Nagumo turned it down. His torpedo aircraft lay in pieces on the hangar decks. His Zeros were either up over the formation or down on the flight decks taking on fuel and ammunition, which left him nothing to escort a strike with. His first wave was overhead, shot up and short of fuel.
The position he had been handed for the American fleet put it far enough off for him to believe their fighters could not yet reach him, which meant the Americans would either wait to come nearer or come without escort, and either of those suited him. The position was wrong. The search crew’s navigation had put the American ships about 55 miles farther north than they were. So he chose the thing his whole service was built around.
He would take his aircraft aboard, arm everything properly, and send one enormous combined strike, dive bombers and torpedo aircraft and fighters together off four decks at once. It was the method that had worked at Pearl Harbor and in the Indian Ocean and everywhere since. It had a single requirement, and the requirement was quiet time on the flight decks. The recovery began at 8:37 and ran until about 9:00 for the strike aircraft.
At 9:17, the force swung to the northeast to run down the Americans, and the last fighter came aboard a moment after that. Then the carrier squadrons came, and they came the way the island squadrons had come, which is to say in pieces. Fifteen Devastators from the Hornet at about 9:00 under Lieutenant Commander John Waldron, with nothing over them, were every one shot down. Fourteen more from the Enterprise a few minutes behind, also unescorted, were almost all shot down as well.
Twelve from the Yorktown at about 9:30, with a handful of fighters above them that could not save them, went the same way. Between the three of them, they made no hits at all. None of that was designed. Rear Admiral Raymond Spruance had wanted his squadrons to arrive together and attack together.
His fighter cover got lost or took the wrong heading, and his torpedo crews flew on alone and died for it. From the Japanese decks, though, the shape of the morning was simpler than that, and what it amounted to was a queue. Every one of those attacks put the mobile force into hard evasive turns, which stops ordnance work in a hangar dead. Every one of them burned cannon rounds out of the patrol, which brought Zeros down onto the decks to reload, which took the decks.
Marines, Army crews, a Navy detachment, and three carrier torpedo squadrons out of two task forces under two admirals. None of it coordinated with any other part of it, and all of it arriving. The mobile force had been built to fight one enemy fleet doing one thing at a time. What arrived instead was a morning of unconnected American commands, each taking a turn at it.
At 10:20, the Akagi had just finished taking three fighters aboard. That is when the dive bombers arrived. Two squadrons off the Enterprise under Wade McClusky and one off the Yorktown under Maxwell Leslie. None of them escorted and none of them opposed because the last of the Zeros had been dragged down to sea level by the Yorktown’s torpedo crews, and there was nothing left upstairs.
In something under five minutes, the Kaga, the Soryu, and the Akagi were all burning. What the men on those bridges spent the rest of their lives declining to say plainly is that the decks were nearly bare. Japan’s own official war history, published in 1971, states that at the moment of the attack, every one of those ships still had its strike aircraft down in the hangars. The machines up in the open were fighters, and on the Soryu, a few of them were going up to reinforce the patrol.
The photographs the B-17 crews took between 8:00 and 8:15 show no strike aircraft on the decks of the Soryu, the Hiryu, or the Akagi, just a handful of Zeros. Lieutenant Richard Best, who led the three Dauntlesses that hit the Akagi and who was looking straight down at her out of his dive, said the six or seven machines on her deck were fighters, and that one of them was in the act of taking off. The strike was still below, fueled, half armed, packed into a sealed steel box with no ventilation and one way out, an elevator that took one aircraft at a time. That is also why three ships died in five minutes rather than burning through the afternoon.
A bomb through a Japanese flight deck went into a closed room full of aviation fuel and loose ordnance. Nine years later in Tokyo, Commander Mitsuo Fuchida published an account of the battle. Fuchida had led the aircraft at Pearl Harbor, and he was aboard the Akagi that morning, laid up after an emergency operation for appendicitis, unable to fly. In 1955, his book came out in English through the United States Naval Institute, and one passage in it became the most repeated paragraph ever written about Midway.
At 10:20, he wrote, Nagumo ordered the launch, and on the Akagi’s deck, all planes were in position with engines warming up. The ship began her turn, and within five minutes, he wrote, every machine she had would have been in the air. That number went into Samuel Eliot Morison, into Walter Lord and Gordon Prange, into the 1976 film, into a novel and a television series, and out from there into the general memory of an entire country. It is a magnificent line.
What it does is set a whole battle on an interval too short for anyone to be blamed for it, a defeat handed down by fate rather than by judgment, an honorable end for a navy that had just been taken apart in an afternoon. The paperwork those ships kept does not support any part of it. Nagumo’s decks needed about 45 minutes of uninterrupted work to put a strike of roughly 20 aircraft into the air on every deck at the same time: 40 of those to lift and spot and warm and brief, another five or ten to fly them off. That figure did not come from an American reconstruction.
It comes out of Japan’s own war history, which says flatly that even with everything armed and ready, it would take no less than 40 minutes to get them out of the hangar and finish preparing them. There is a fair yardstick for it as well. Six months earlier at Pearl Harbor, six carriers with no enemy anywhere near them needed 55 minutes to spot and launch a second wave of 171 machines. The Akagi had not held a clear deck for that long since 7:00 in the morning.
Her flight operations for the day are a wall of small entries. Fighters off at 7:10, one down at 7:20, one at 7:26, three at 7:36, two at 7:50, three up at 8:00, four up at 8:30. The whole Midway strike aboard between 8:35 and 9:00. One down at 9:10, five up at 9:30, three up at 9:45, two down at 9:51, three down at 10:00.
Ten minutes after that last entry, she was on fire. Almost everything the world believes about that morning came from one officer flat on his back in a sick bay describing a flight deck he could not see. There is a real argument among historians about which clock ran out. Dallas Isom held that the decisive thing was the delay in getting that first sighting signal to Nagumo and that everything followed from the extra 20 minutes of torpedoes coming off aircraft that could not then be put back.
Jonathan Parshall, David Dickson, and Anthony Tully answered that the arming clock was never the binding one, and the flight deck was, and that Nagumo, far from standing minutes from a launch, had barely begun. Isom never conceded it and published a book of his own to argue the point. Both positions are still in print and still opposed. The material that would settle them, the flag bridge log of the Akagi and the men who worked those hangars, went into the sea or into the ground a long time ago.
What both sides agree on is that the decks were bare. The man who said otherwise had spent the morning on his back in the Akagi’s sick bay. There is one more piece of evidence, and it is the cleanest of all because it is not a document and nobody had to remember it. The Hiryu was not hit at 10:20.
She was ten miles clear, undamaged, with an aggressive admiral aboard who had wanted to attack for two hours and every reason on earth to get aircraft off her deck the moment the other three began to burn. Her first strike left her at 10:57.