In the second week of April 1945, on open ground behind the American lines in southern Okinawa, an engineer battalion rehearsed a single task for days: they assembled a Bailey bridge, disassembled it, and built it again. The Bailey was a new design of steel panels that soldiers could carry and pin together by hand, and none of the battalion’s men had ever constructed one. They had shipped out from the United States before the Army adopted the design, and their previous island campaigns had required no bridging at all. A lieutenant named Irving Golden, who had joined them recently after building several Bailey bridges in Tunisia, walked them through the process panel by panel in a rear area where no one could see.
What looked like a routine work detail was actually preparation for a specific operation, one driven by a captured Japanese intelligence document. The reason for the rehearsals traced back to a single sheet of paper written by a Japanese officer. The Japanese 62nd Division, entrenched in the hills to the south, had issued notes to its own troops analyzing American behavior on the front lines. One line addressed the hours after dark: the enemy generally fires during the night, it observed, but very seldom takes offensive action.
That sentence was a practical battlefield observation, written to stop Japanese companies from flinching at the constant noise of night fire. But a copy of the document was recovered by American forces, translated, stamped with a file number, and eventually landed on the desk of an American division staff while they were planning their next major attack for April 19. The observation was accurate at the time. American guns fired through the night along entire fronts even where nothing was moving and no attack was planned.
To the Japanese soldiers in the hills, it looked like an enormous expenditure of ammunition thrown at the dark for no visible result. Japanese officers who wrote about that firepower were not complaining. They were drafting rules, because rules were the only defense they had against fire they could not counter. The 32nd Army had been analyzing American firepower since the year before.
One directive put it bluntly: fighting the American land army amounted in practice to fighting at Shermans. Everything else followed from that. The Japanese units were instructed to hold their fire until tanks and infantry were close. This protected positions from early discovery and gave anti-tank guns a surer shot.
The army’s battle instructions spelled out the logic: the surest way for the enemy to map the entire Japanese firepower system was to get defenders to open fire prematurely on a force large enough to be worth shooting at. Firing a weapon carried a price. A gun crew that shot at a patrol revealed its position. A machine gun that answered a scout bought its own destruction with a few seconds of noise.
The 47 mm anti-tank gun, the one weapon capable of punching through a Sherman at ranges up to 800 yards, weighed about 1,600 pounds and rolled on rubber tires. In practice, most were set inside cave pillboxes and never moved again. When Japanese troops did fire, they chose their targets carefully. Along the northern ridges, their habit was to let a few Americans cross an open saddle and then fire on the group from behind, preferring officers.
A man holding a map, waving an arm to point out a direction, or carrying a pistol instead of a rifle had identified himself. Marines climbing those slopes quickly learned to stop doing all three. The men on the receiving end of American fire left their own record in diaries and letters. A private in an independent battalion wrote in January that seeing enemy planes for the first time gave him the feeling of being in a war zone.
Three weeks later, after carrier aircraft returned, he filled a page with fury, describing bombing from 6:00 in the morning until 6:00 at night. On the island of Ie, another soldier watched a landing come in through the smoke and wrote a line that summed up American firepower: their firepower is so great that we dared not show our heads. By the numbers, the American 10th Army fired more than one million shells from its 105 mm howitzers alone during the campaign. That figure counts only field artillery, leaving out mortars, tanks, naval guns, and aircraft.
A Japanese soldier noted the ratio in his diary in late April: at least 10 rounds are guaranteed to come back for every one of ours. As the American attack pressed forward, Japanese tactical papers multiplied. On April 13, the 44th Independent Mixed Brigade issued an urgent battle lesson on flamethrowing tanks and yellow phosphorus shells. Two days later, the 22nd Regiment circulated a study of how Americans laid out positions after dark and how infiltration parties might get through them.
Army headquarters pressed units to fix in advance the exact spots from which men would attack tanks at arm’s length. On April 14, the 62nd Division told its troops what was coming in a plain sentence: the enemy is now preparing to advance on all fronts. Strong points were to be spread so losing one would not break the line. Weapons were to be kept under cover until they could be used effectively.
Secret papers were to be burned when positions became untenable. By mid-April, Japanese officers were writing less about tactics and more about survival inside the hills. A set of instructions issued on April 15 covered life under bombardment: intensify spiritual training, avoid useless work, get as much sleep as possible, send men out at night to breathe, build latrines inside and outside the hill, keep them clean, and watch for dysentery and waterborne sickness. Soldiers were drinking untreated water because lighting a fire underground seemed more trouble than it was worth.
The instruction to sleep stands out as a peculiar element in a battle paper, but armies write rules about the things that are killing them. By that measure, the Japanese papers from Okinawa amount to an inventory of American fire. The most revealing calculation in the documents concerned how many men a company could leave on the surface during a bombardment. Initially, the answer had seemed obvious: pull everyone underground, hold a few openings, and the bombardment achieves nothing.
That approach worked once. Then American tank and infantry teams began rushing the mouths of caves the moment the guns lifted. A company that was entirely below ground found itself sealed in with its own weapons, unable to shoot, waiting for demolition charges. The caves offered almost no field of vision.
Men inside often did not know where the Americans were on the hillside. They had names for everything: the American method of clearing caves, called blowtorch and corkscrew by the Americans, the Japanese called the cavalry charge. Narrow strips of ground a cave’s weapons could not cover were the dead angles. When Americans gained a foothold above a position and attacked its openings from outside the arc its guns could cover, it was called a straddle attack.
According to American historians who studied the translations, it was what the troops feared most. So the rule changed, and the new rule was a number. Each company was to keep 10 or 12 men on the surface through the bombardment. Their job was to survive it, then hold the position in the gap before the tanks arrived, long enough for the rest of the garrison to come up from below and retake the positions.
Even that was a compromise. There was not enough covered ground between caves to hold everyone outside. When shelling stopped, only about a third of the garrison could exit, while the other two-thirds stayed below, taking their chances on being buried or sealed in. That arithmetic was the closest surviving Japanese answer to American shellfire, not a complaint but a staff calculating how many men could be left in the open because the alternative was losing the whole company.
On April 19, the Americans attacked along the entire line, and both halves of the story converged the same morning. In front of the ridge at Kakazu, a battalion pushed a reinforced company of tanks through a cut between two hills, hoping to slip around a position that had already stopped one division. The Japanese had anticipated the move. Mortars, machine guns, anti-tank guns, and anti-aircraft cannon covered the road, and overnight infiltration parties had set up more guns behind the American front.
As the tanks advanced, fire fell between them and their infantry, cutting the two apart. The tanks got behind the ridge and shot up what was left of the village, but with no riflemen to keep satchel-charge teams off them, they were forced to withdraw. Eight of 30 tanks returned. Mines, anti-tank guns, and men carrying explosives accounted for the other 22.
The 62nd Division wrote up that engagement the next morning as a priority battle lesson, numbered it 19, and distributed it to its units so the next unit facing tanks would know what had worked. A few miles west that same morning, another regiment was already on the escarpment. They had crossed a tidal inlet in darkness using four bridges put in overnight: a flip bridge for assault troops, two Bailey spans totaling 90 feet for supporting weapons, and a rubber pontoon bridge strong enough for loaded trucks. Getting the bridging materials down to the water had been the hard part.
The only route was a rudimentary jeep track in full view of Japanese observers on high ground, and improving it in daylight would have announced the plan. For several days, a bulldozer worked the track in plain sight doing nothing in particular: hauling bogged jeeps out of the mud, pushing dirt and rock into the worst ruts. The operator slept, tinkered with his engine, and made broad gestures of irritation whenever another vehicle sank. After dark, he worked hard, and by April 18, the road ran almost to the water.
The night firing had never been the attack itself. It was harassment and denial, shells spent to keep a hillside from being worked on and to stop parties moving, and the Americans could afford to spend them that way for 82 days. The Japanese had watched that pattern, understood it, and written it down so their companies would not waste a night waiting for an assault that was not coming. Then one night, the assault came in the dark, because that captured sentence about Japanese expectations had reached an American staff in translation while they were choosing between a daylight crossing of a thousand yards of open ground and an alternative.
The 62nd Division had written its notes for a few hundred of its own officers, and the notes reached enemy headquarters instead. In early May, the Japanese command decided to attack. For the first time, whole artillery batteries came out into daylight. Crews hauled guns from the caves and toward the front to support the infantry.
The attack failed within two days. The artillery never recovered. American fire caught many tubes in the open and destroyed them, and the offensive had burned shells at a rate the island could not sustain. Afterward, the ammunition ration fell from 50 shells per gun per day to 15.
A Japanese lieutenant wrote the consequence in his diary: we realized that we were doomed once the failure came through. What is absent from the record is also telling. No surviving Japanese staff paper from Okinawa calculates shells fired against one man. The line most often quoted about 10 rounds coming back for every one of ours comes from a translated diary, and where it appears in print, it carries no file number and no name.
The Japanese Burning Order had already gone out, and much of what Japanese officers wrote on that island survives only because an American translator got to it first, stamped a number on it, and filed the English version. In November 1949 in Tokyo, a group of former Japanese officers finished a day-by-day account of the Okinawa fighting and handed it to American occupation authorities, who gave it a number in a monograph series. The men who compiled it had served in a different army and had not fought on the island. Almost everyone who wrote the papers they were working from was dead.
They assembled the account out of what remained, which was not much, and it now sits in the files alongside the translations.