In August 1945, aboard a small American warship anchored in the Marshall Islands, a wardroom table held open packs of cigarettes between Japanese naval officers and the Americans facing them. The men in white caps had not seen a supply ship in twenty months. They had come aboard to surrender their atoll to one of the ships that had been blockading them. The ship was the Levy, a slim warship of just over twelve hundred tons, built in New Jersey and armed for hunting submarines.

The surrender signed in her wardroom was the first by any Japanese garrison after Tokyo accepted the Allied terms, eleven days before the ceremony in Tokyo Bay that most people remember as the end of the war. To understand why it happened at Mili Atoll, and why it happened there first, it helps to know what kind of place Mili was. It is a ring of coral islets around a wide lagoon at the southeastern end of the Marshalls. Between late 1942 and late 1943, the Japanese built an airfield there with three runways, the longest a little under five thousand feet, along with two hangars, a wooden pier, repair shops, and a spread of wooden barracks and sheds.

Korean and Marshallese laborers did much of the building. Because the base came so late, there was very little concrete on it. The guns faced the open ocean, a mixed collection of British and Japanese pieces: eight six-inch coastal guns, three more fourteen-centimeter guns, four five-inch dual-purpose guns, and over a hundred anti-aircraft weapons. One radar set gave about ten minutes of warning.

By the end of 1943, more than five thousand men were stationed on that sand. Before the war ended, roughly half of them were dead. And yet when it did end, the men still standing on Mili knew, and knew quickly. That needs explaining because by then the codes they held were too old to unlock their own navy’s orders.

The best account from inside the garrison comes from its executive officer, Lieutenant Commander Hiroshi Tokuno, who held that post from July 1943 until the final days. In October 1945, an American naval captain named Stedman Teller questioned him in Tokyo for the American study of the bombing campaign. Tokuno told him that signals between the islands of the Marshalls had worked right up to the end, even after every link by sea was cut. But the codes went stale.

From then on, the official traffic of the Japanese Army and Navy was unreadable on Mili. So for news of the war, the garrison depended on the Japanese news agency Domei and on whatever American and Australian broadcasts its operators could pick up. That is how a garrison cut off from its own chain of command heard, in August 1945, that the war was over. It heard the news over the open air, like anyone else with a receiver.

That leaves the larger question. Mili was one of several Japanese garrisons that the Americans had simply flown past. Wotje, Maloelap, and Jaluit sat in the same chain, in the same condition, and Truk and Wake were bigger prizes. So why was Mili first?

Tokuno’s account begins in the winter of 1943. On the first of November, the garrison stood at nearly four thousand men, most of them naval guard troops and construction workers, with airmen and an army detachment alongside. As the Americans took the Gilberts, just to the south, reinforcements came in. Fifteen hundred arrived from Kwajalein.

More were diverted from a draft meant for the Philippines, and by the twenty-second of December, the count stood near fifty-one hundred. The food did not keep pace. The last convoy reached the atoll around the twenty-first or twenty-second of December. Its largest ship, the Nankai Maru number two, was sunk before she could unload, and Tokuno said the garrison was left with provisions and fuel for about six months.

The spring of 1944 began on the other side of the reef. In January and February, the Americans took Kwajalein, Majuro, and Eniwetok, and Marine air squadrons moved onto the captured fields almost at once, a dive bomber group at Majuro and a fighter group at Roi. From there, the bypassed atolls were a short hop away. The strikes began on the fourth of March.

Wotje, Maloelap, Jaluit, and Mili were all hit that first day, and the dive bomber squadron flying its own first mission came home with anti-aircraft damage. On the eighteenth of March, two things happened at Mili on the same day. Eight Marine Corsair fighters went in carrying bombs against the anti-aircraft positions, the first time Corsairs were flown as fighter-bombers in the Marshalls. And a task group that included the battleship Iowa stood offshore and shelled the island.

A Japanese coastal battery hit the Iowa twice, doing minor damage. By then, the air war over Mili itself was finished. Tokuno said the airfield was first damaged on the twentieth of November 1943, and that small bombers, more than the big four-engine ones, finally made it unusable on the fourth of February. The last fuel on the island went to aircraft staging through on the twenty-fourth of February.

After that, Mili’s fighters stayed on the ground, and about thirty aircraft lay wrecked or damaged around the runways. Summer came and the garrison changed jobs. Tokuno said that from the first of January 1944, the main occupation on the island had become making gardens to grow food. Work on fortifications carried on for a while, but by June the men had decided that growing food was more worthwhile than building more defenses.

The digging turned from gun pits to vegetable beds. Overhead, the tonnage climbed. The Marine wing responsible for the bypassed islands dropped more than seven hundred tons of explosives in July 1944 and about twelve hundred tons in August, its peak month. Sometime in mid-August, Mili’s radar was knocked out for good, and after that the garrison had no warning at all.

The Marines paid for their share. By the end of 1944, the squadrons working these islands had lost twenty-nine pilots, two gunners, and fifty-seven aircraft to enemy action. Then, in winter, the planes mostly stopped coming. On the twenty-third of January 1945, the Pacific commander ordered the bombing of the bypassed islands largely ended, except where the results were worth the cost.

The war had moved on to Iwo Jima and Okinawa, and Mili was left to its gardens. The deaths went on through the spring and the summer. When Tokuno gave his figures in Tokyo, he put the garrison’s losses to enemy action at about twelve hundred killed. He put the losses from lack of food, poisonous fish, beriberi, and dysentery at about twelve hundred more, and he blamed the high toll on the shortage of medical supplies and proper food.

So hunger and the bombs had each killed roughly the same number of men. After January, the bombs mostly stopped, and the hunger did not. Nobody ever had to land on Mili. In the spring of 1945, the cracks showed first among the people the garrison had brought there to work.

Hundreds of Korean laborers had been brought to the atoll to build the airfield. Early that year, a group of them turned on their overseers and killed several. Troops came across from a neighboring island and crushed the revolt. The numbers are not settled.

A South Korean government commission reported in 2010 that about a hundred of the laborers were killed, along with seven of their guards. A researcher in Japan, publishing in 2024, counted eleven guards dead, at least thirty-two laborers shot, and twenty-three who died by their own hand. The survivors later told investigators what had driven them to it, and the commission accepted their account while admitting it could not prove it. Late in March, five Koreans, one Japanese, and two Marshallese got off Mili and gave themselves up to an American landing craft numbered 392.

Over that summer, as Marine pilots dropped leaflets along with their bombs, another forty-two men came over from the bypassed atolls. In August, the Levy arrived. She had spent the last two months of the war blockading and shelling the bypassed atolls, and along the way she had picked up a boatload of Marshallese who had escaped from Jaluit. She came to Mili on the twelfth of August.

Two days later, Japan accepted the Allied terms, and on the nineteenth, Japanese officers came out to the Levy under a white flag. So why Mili first? The papers that survived do not give the Japanese captain’s reasons, and nothing written by him has come to light in English. What they give instead is a set of plain facts.

An American warship was already on station off Mili before Japan had even accepted the terms. The American command at Majuro, nearby in the same chain, had a flying boat and a senior officer ready to send out the same day. And the garrison had been cut roughly in half by twenty months of hunger and bombing, with no ships, no fuel, no aircraft, and no readable orders from home. Every bypassed garrison in the Marshalls was hungry, but only Mili had an American ship already waiting off its lagoon when the news came.

The officer in the flying boat was Captain Harold Bartley Grow, the atoll commander at Majuro, a Naval Academy man from Michigan who had come back into uniform in 1942. In the 1920s, the Navy had lent him to the government of Peru, where he built up its naval aviation and rose to inspector general of its air arm. On the twenty-second of August 1930, an army revolt broke out in southern Peru. Grow flew against it and was captured.

The new strongman in Lima called him a hired mercenary and talked of a court martial, and the State Department cabled its embassy to work for his release. Fifteen years to the day after that revolt began, Grow sat down in the Levy’s wardroom on the other side of a surrender, this time as the man accepting it. The signing came on the twenty-second of August. The Japanese commander was Captain Masanori Shiga, head of the 66th Naval Guard Unit and the senior officer on Mili.

A Navy photographer took several pictures. In one, Shiga salutes as he comes aboard with Grow facing the camera. In another, Shiga bends over the surrender document with Grow beside him. Even the moment of signing is not fixed.

A handwritten note kept with the surrender pen gives the time as 12:32 in the afternoon, while a later history of the ship says 1:00. A Navy chronology of the Marshalls puts the whole surrender a day earlier on the twenty-first. The photographs in the Navy’s own collection both say the twenty-second, and the weight of the evidence sits there. Nobody fired a shot.

The Levy was named for Commodore Uriah Levy, the officer who led the long fight against flogging in the old Navy. Her captain that summer was Lieutenant Commander William Clarenbach, a reservist from Alameda, California. Her sailors went ashore to take the island’s weapons under instructions that listed all arms up to and including thirteen-millimeter machine guns, swords, and bayonets. They built a new dock.

Six days after the signing, Rear Admiral William Keen Harrill arrived, and the American flag went up over Mili with a twenty-one-gun salute. The commanding officer of another small warship, the Parks, stayed behind as the American representative at Mili. He had about twenty-four hundred Japanese troops to account for, with ammunition to dispose of and minefields to clear out of a garrison that had once held fifty-one hundred men. And most of the missing had never come within rifle range of an American.

The Levy did not stay. A week after Mili, she carried Grow to Jaluit for the next round of surrender talks. Rear Admiral Nisuke Masuda commanded about two thousand men there. Then the Levy took aboard a Marine brigadier general and a party of officers and correspondents and sailed for Wake, where on the fourth of September another Japanese admiral came aboard to sign.

One small warship had become the place where the bypassed Pacific came to give itself up. The garrison of Mili went home. The pen is still in Washington. It sits in the headquarters collection of the Naval History and Heritage Command under accession number 07598P.

It is an ordinary dip pen with a black enamel wooden body, a brass ferrule, and a steel nib made by the Spencerian Steel Pen Company of New York and stamped as made in the United States. It came to the collection with a Navy Department envelope marked official business. On that envelope, someone wrote by hand that this was the pen used to sign the first surrender of any Japanese territory and gave the ship’s hull number as 166 instead of 162, an error the Navy’s curators point out in their entry. The envelope calls the pen a war relic, and that is still how the Navy keeps it.

A single steel nib in a collection at the Washington Navy Yard.