A gentle breeze sighs through the trees and leaves flutter gently in response. Robed priests slowly walk through the shrine’s precincts, stopping in front of the main altar to clap loudly and bow. The smoke of incense fills the air, and wooden placards painted with the prayers of the devout line the walkway. I am inside Hakozaki, one of Japan’s three most sacred Shinto shrines. Established in 923, Hakozaki has existed for more than a thousand years. The grounds of the shrine are filled with monuments and buildings, and it is in front of one of them that I stand gazing at a stone weight for an ancient ship’s anchor. A small plaque, in English and Japanese, explains that it came from a lost ship, part of a fleet sent by China’s Mongol emperor, Kublai Khan, to invade Japan in 1274.

A stone tablet nearby has musical notes and writing in kanji, or traditional Japanese script. I am told that it is a traditional song about the Mongol invasion. To my surprise and delight, our host stops a tour group of Japanese schoolgirls and requests them to sing the song. I ask my host and translator what the words are, and, with less grace than the girls but with gusto, he sings the song for us in English. The last stanza is the most significant:

Heaven grew angry, and the ocean’s

Billows were in tempest tossed;

They who came to work us evil,

Thousands of the Mongol host,

Sank and perished in the seaweed.

Of that horde survived but three;

Swift the sky was clear, and moonbeams

Shone upon the Ghenkai sea.

There it is, the story of how the gods sent a divine wind to sink the Mongol invasion fleet and save Japan. The anchor stone is displayed at Hakozaki as proof of that long-ago event and as a reminder of how Japan’s shores were protected by that wind—a wind whose name in Japanese is kamikaze.

The story of the kamikaze was used to lethal effect in the Second World War. In the name of that “divine wind,” nearly two thousand young Japanese men strapped themselves into airplanes and dove out of the sky to suicidally crash onto the decks of American and Allied warships. The deadly toll they wrought did not turn the tides of war, however. In defeat, the Japanese were told that their emperor was not a god and that the ancient story of the kamikaze was a myth. But the story of the Mongol invasion and the kamikaze remains a powerful part of the national consciousness of modern Japanese.

I’ve journeyed to Japan with fellow members of The Sea Hunters to visit an archeological site where a lost ship of Kublai Khan’s fleet has surfaced from the gray-green waters near Takashima, a tiny island off Japan’s southwest coast. History, myth or a combination of both? The remains of the ancient ship will tell us much about what really happened off these shores more than seven centuries ago.


Under Chinggis Khan, a great horde of “barbarians” swept out of the Mongolian plains in 1206 to win a series of military conquests that made them not only the masters of much of Asia but also of an army poised on the doorstep of Europe and the Middle East. History would have been very different had the Mongols achieved Chinggis Khan’s dream of absolute conquest. As it was, the world will never forget the saga of the Mongols and of battles like their capture of the Turkmen city of Merv in 1221. In revenge for the death of his son-in-law, Chinggis ordered the death of every living thing in the city, and seven hundred thousand people were put to the sword.

Battles against the Muslims, the Russians and other eastern European kingdoms continued under Chinggis’s son Ögodäi; however, the death of Ögodäi’successor not only doomed the Muslim campaign but stalled the conquest of China. The next Mongol leader, Kublai Khan, soon controlled more territory than any sovereign in history. But he wanted more territory, more riches and, above all else, recognition of his supreme status as ruler of much of the world.

Even while he was engaged in a bitter struggle to conquer China, Kublai sent envoys to the Japanese court in 1268 to demand subservience. The Japanese military dictatorship, the bakufu, ignored the Mongol demands. In response to this defiance, Kublai Khan ordered his vassals in the subjugated Korean state of Koryo to build a fleet of nine hundred ships to invade Japan. The relatively narrow straits of Tsushima, spanning 284 miles between Korea and the coast of Japan’s Kyushu Island, had been a trade route for centuries. Now it would become a highway for war.

The invasion fleet departed from Koryo on October 3, 1274, after embarking twenty-three thousand soldiers and seven thousand sailors. Two days later, the fleet attacked the island of Tsushima in the middle of the strait, overwhelming the eighty Japanese troops stationed there. The island garrison of Iki, closer to the Japanese coast, fell next. On October 14, the Mongol fleet attacked the Kyushu port of Hirado, and then moved north to land at various points along Hakata Bay (near modern Fukuoka). Groups of samurai and their retainers rushed to meet the invaders at Hakata Bay—in all, historians estimate that some six thousand Japanese defenders stood ready to fight the far more substantial Mongol army.

Among the defenders was a samurai named Takezaki Suenaga. He left the only contemporary pictorial records of the Mongol invasion on two scrolls that he commissioned later in order to petition the government for a reward for his services. The scrolls, known as the Moko Shurai Ekotoba, are one of Japan’s great cultural treasures.

Dating to around 1294, the first scroll unrolls to reveal samurai in armor riding off to battle in 1274. The battle was unequal not only in numbers but in weapons and tactics. Mongol weapons were more advanced than those of the samurai: their bows had greater range, firing poisoned arrows, and they also had explosive shells hurled by catapults. In battle, the Mongols advanced en masse and fought as a unit, while the samurai, true to their code, ventured out to fight individual duels. In a week of fighting, the Japanese were slowly forced to give way. The scroll shows the Mongol forces firing arrows as horses and men fall, and Suenaga himself bleeding and falling from his horse as a bomb explodes in the air above him. The Japanese retreated, falling back to Daizafu, the fortified capital of Kyushu. The Mongols sacked and burned Hakata, but time was running out for them: Japanese reinforcements were pouring in from the surrounding countryside. The Mongol commander was wounded, and the sailors aboard the invading ships were wary of an incoming storm that threatened the fleet in its crowded anchorage.

On October 20, the wind shifted, and a number of Mongol ships dragged anchor, capsizing or driving ashore. In all, some three hundred ships and 13,500 men were lost. Battered and depleted, the surviving Mongols retreated to Koryo, leaving the Japanese to cheer their salvation thanks to the storm that had ended the invasion.

Knowing that the Mongols would be back, the bakufu ordered the construction of defenses at Hakata Bay. In a six-month period in 1276, laborers erected a 12½-mile, 5- to g-foot high defensive stone wall set back from the beach. The samurai also organized their vassals into a compulsory defense force and requisitioned small fishing and trading vessels for a coastal navy.

Kublai Khan renewed his demand for Japan’s surrender in June 1279, just as the last remnants of the Sung dynasty in China crumbled before the Mongol onslaught. The bakufu cut off the heads of the Mongol envoys as they landed. Furious, Kublai Khan ordered Koryo to build a new fleet of nine hundred ships to carry ten thousand troops and seventeen thousand sailors; and in China, he ordered a fleet of nearly thirty-five hundred ships and an invasion force of a hundred thousand Chinese warriors to prepare for battle.


An artist’s rendition of a thirteenth-century Mongol ship wrecked at Takashima Island. Courtesy of Kyushu Okinawa Society for Underwater Archeology

Kublai Khan directed the two fleets—the Koryo Eastern Route Division and the Chinese Chiang-nan Division—to rendezvous at Iki Island to co-ordinate their attack. The Eastern Route Division sailed first on May 3, 1281, retaking Iki on June 10. Without waiting for the arrival of the Chiang-nan Division, the impatient commanders of the Eastern Route Division sailed to Hakata Bay. Takezaki Suenaga’s second scroll depicts the second invasion, showing him riding off to war, passing in front of the newly built stone wall at Hakata Bay as other samurai sit atop the wall and wait for the enemy. The stone walls thwarted the Mongols, who pulled back to occupy an island in the middle of the bay. The Japanese used their small navy to cut into the Mongol fleet, with armed samurai springing onto the enemy ships and killing the crews and soldiers. The second scroll also shows Suenaga in a small boat, running up alongside larger Mongol ships and fighting his way forward to cut the throats of the crew in deadly hand-to-hand combat. The brushstrokes of the artist convey the ferocity of the fighting, with blood spurting as sharp blades and arrows cut down men. The paintings are a graphic testimony as to why the badly mauled Eastern Route Division retreated to Iki Island with the Japanese in pursuit.

The Chiang-nan Division finally sailed in June and met up with the Eastern Route Division at the small island of Takashima, 30 miles south of Hakata. The Japanese fought the combined Chinese and Mongol forces in a running two-week battle throughout the rugged countryside. The crews of the invading ships chained their vessels together and constructed a plank walkway, forming a massive floating fortress in preparation for the inevitable waterborne assault by the small defense craft of the Japanese.

The Japanese ships, some of them filled with straw and set on fire, attacked the Mongol fleet but were unable to do much harm. As the story was later told, the Japanese beseeched the Goddess of the Ise Shrine for another storm to help them, and their prayers were answered. The legend states that “A green dragon had raised its head from the waves” and “sulfurous flames filled the firmament.” Driving rain, high winds and storm-lashed waves smashed into the Mongol fleet. Thousands of ships sank, drowning nearly a hundred thousand men. Mongol troops stranded on the beach, demoralized and cut off from escape, were rounded up and executed. The shores were strewn with debris and bodies; according to a modern Japanese history, “a person could walk across from one point of land to another on a mass of wreckage” at the entrance to Imari Bay. Kublai Khan abandoned his dreams of a Japanese conquest in 1286 when he abruptly cancelled the preparations for a third invasion.

Interestingly, Suenaga’s scrolls and the handful of Japanese documents from the 1281 invasion do not depict or mention a storm. Critics of the scrolls deride them as the work of a “self-aggrandizer,” while others point to persistent myth-building by Japan’s military and political leaders that glorified the emperor as a god and celebrated Japan’s divine protection and status. (Eventually, this led to a series of wars of conquest that greatly expanded the Japanese empire from the 1870s through the early 1940s.) But the Venetian adventurer Marco Polo, who allegedly spent several years in Kublai Khan’s court, wrote an account of the Mongol invasion in which he mentioned the storm that destroyed it:

And it came to pass that there arose a north wind, which blew with great fury, and caused great damage along the coasts of that Island, for its harbors were few. It blew so hard that the Great Kaan’s fleet could not stand against it. And when the chiefs saw that, they came to the conclusion that if the ships remained where they were, the whole navy would perish. So they all got on board and made sail to leave the country. But when they had gone about four miles they came to a small Island, on which they were driven ashore in spite of all they could do, and a great part of the fleet was wrecked and a great multitude of the force perished.

Given the prominent place of the story of the kamikaze in Japanese history, who knows where the truth lies? For a handful of young archeologists, the truth lies in the remains of the events, which now lie beneath the waters of Japan’s coast.


The beautiful views of Hakata and Imari bays and their gentle waves belie the violence of the storms that are said to have twice destroyed the Mongol fleet, as well as the tremendous battles waged on their shores in 1274 and 1281. Apart from memorials and monuments, few physical traces of the invasion remain on the land other than a handful of reconstructed sections of the stone wall in the heart of modern Fukuoka. Some scholars do not believe that the stone anchor weight at Hakozaki Shrine comes from the Mongol invasion; they think that it is one of many similar anchors lost on the bay bottom during the centuries Hakata Bay was an active port, because no other evidence—such as weapons or broken hulls—has ever emerged. But the waters off Takashima Island in Imari Bay have yielded traces of the Mongol fleet and its destruction.

Fishermen are usually the first to discover shipwrecks, and for years, Japanese trawlers operating in the waters of Imari Bay had been dredging up pottery and other artifacts from the lost Mongol fleet of 1281. Then, in 1980, Torao Mozai, a professor of engineering at Tokyo University, used a sonoprobe—a sound-wave device that geologists use to discover rocks buried in ocean sediment—to survey the seabed off Takashima Island. He discovered that buried artifacts appeared as different colors on his screen.

A year later, Professor Mozai’s team pinpointed many objects that divers then recovered. The artifacts attest to the diversity of the invading force and its weapons, as well as its need for provisions. In addition to spearheads, war helmets, stone balls for catapults and a cavalry officer’s sword discovered sticking upright in the mud—exactly where it had been dropped seven hundred years earlier—the divers found stone handmills for grinding gunpowder, iron ingots, stone anchor stocks and mortars for pounding rice or corn. The discoveries made international headlines (and a National Geographic magazine article) in 1981, the seven hundredth anniversary of the second Mongol invasion, and sparked the creation of a new museum on Takashima Island. The opening of the museum inspired a number of local fishermen to donate their own discoveries, including a bronze statue of Buddha dating to the twelfth century and a bronze seal of authority that had belonged to a Mongol commander of a thousand-man group.

Since 1991, the Kyushu Okinawa Society for Underwater Archeology (KOSUWA), under the leadership of Dr. Kenzo Hayashida, has been conducting surveys and excavations off Takashima’s shores. In 1994, they discovered three wood and stone anchors from the Mongol invasion fleet, buried in the mud 400 feet offshore and in 40 feet of water. One of the anchors is 21 feet long and weighs one ton. Analysis of the wood used in the anchor showed it was red oak dating to within a few years of the Mongol invasion. Analysis of the stone used in the anchor showed that it was granite from China’s Fujian province, from which most of Kublai Khan’s invasion fleet sailed to the shores of Japan. Of even greater interest were the remains of the anchor cable, which lay stretched out straight from the anchor to the shore, indicating the possible presence of a wreck. Excavations recovered 135 scattered artifacts, but the wreck itself remained elusive.

In October 2001, KOSUWA’S hard work paid off with the exciting discovery of a ship from Kublai Khan’s fleet. The wreck lay in Kozaki harbor, a small indentation on Takashima’s southern coast on the shore of Imari Bay. In all the years of work at Takashima, never before had the remains of one of the ships been discovered. In fact, only two other Asian shipwrecks of this age ever have been found, one at Shinan in Korea and the other at Guangzhou in China. Finding another ship from the thirteenth century, a time when Chinese ships were the best examples of shipbuilding in the world, made the wreck at Takashima a very significant discovery in the world of maritime archeology. What the excavation of the site revealed in 2002, however, made it one of the greatest underwater archeological discoveries of the century.

The catch was that the archeologists had to work fast, as construction of a new fishing harbor at the site meant that they had to completely remove the wreck before October 2002. They met the deadline and recovered nearly eight hundred artifacts, ranging in size from a small tortoiseshell comb to what may be part of the ship’s large keel or backbone. Now their work continues in the laboratory.


When The Sea Hunters team arrives, only half of the wreck has been cleared. Each morning begins with a briefing for all divers, and then the first Japanese team gears up to get to work. They wear masks that cover their faces, and they are connected to shore by air hoses and an underwater communications system. They are continuously fed air and report on what they are doing to the dive supervisor and the supervising archeologist in the control room. We gear up to go in with them, donning wetsuits, heavy tanks and our survey equipment. Stepping up to the edge of the concrete dock, I check my air, make sure all my straps are tight, then step off the edge, falling feet first into the water.

As the froth and bubbles from my jump clear, I check to ensure all my equipment is in place. A single line leads down the slope to the wreck, which rests in 43 feet of water. I swim in the gray haze of the warm sea, visibility only 5 feet, until I hear a loud humming sound. To my left, I see the air hoses of the first team and a thick, flexible tube that vibrates when I place my hand on it. This is the outtake for a large underwater suction dredge that the crew is using to uncover the wreck. I follow the tube to a cloud of silt and the excavation.

The seabed is covered with a thick, viscous, almost gelatinous ooze that the archeologists have to dig through to reach the wreck. The task of moving all that mud is immense, as the area of the site covers about two city blocks. The archeologists carefully sweep the handheld underwater suction dredge over the bottom, lying down alongside the thick corrugated hose and gently fanning the mud into the dredge with their hands.

The divers work in shifts, slowly cutting through 5 feet of mud to uncover the wreck, which lies on what was the seabed in 1281. This historic level is hard-packed, coarse gray sand mixed with shell. When the dredge exposes an artifact, a diver carefully fans away the silt and mud to clean it off while reporting his find back to the surface over his communications system. The dive supervisor and archeologist in the control room make notes on what has been found and assign a number to the discovery; the diver then sticks a large numbered tag into the seabed next to it. A team of diving archeologists will carefully map, photograph and draw the object before another team removes it to shore for analysis.

Swimming over the site, I pass through a maze of metal pins with tags—nearly a hundred of them—marking artifacts. A grid of metal legs and twine covers the entire site, dividing it into square units. I swim up to one unit and see scattered broken pots and dishes, timbers and a round object. The round object is only 5½ inches in diameter, but it is one of the most significant discoveries made to date. It is a tetsuhau, or an exploding shell. Chinese alchemists invented gunpowder around A.D. 300, and by the year 1100, huge bombs, much like giant firecrackers, were being used in battle. The first reference to exploding projectiles thrown by catapults appears around 1221, when Chinese sources describe hollow shells packed with gunpowder. Some historians have doubted that such shells were made this early, and even recently suggested, in a new book on the Mongol invasion, that the scene in Suenaga’s scroll, in which the wounded samurai is falling from his horse as a bomb explodes above him, was painted long after the fact because bombs did not exist then.


Some tetsuhau, hollow ceramic shells packed with gunpowder and metal shrapnel. They are the world’s oldest exploding projectiles and were found in the wreck of a Mongol ship off Japan. Courtesy of Kyushu Okinawa Society for Underwater Archeology

The discovery of not one but six tetsuhau proves that the old samurai was right. While four of them are broken, two are intact. X-ray analysis of the two intact bombs shows that one is packed just with gunpowder, while the other is filled with gunpowder and more than a dozen half-inch thick pieces of iron — shrapnel — to cut down the enemy. They are the world’s oldest exploding projectiles. They date to a century before Europeans first used guns at sea, and centuries before Europeans replaced solid stone and iron cannonballs with shells that exploded. Just a week before our arrival, the discovery of these tetsuhau made national news in Japan, though almost no one in the West has heard about the discovery. And here I am, hovering over this unique, technologically advanced and deadly weapon from more than 720 years ago.

Nearby lies a bunch of what looks like rust-colored twigs stuck together. It is a bundle of iron arrows. Japanese accounts of the invasion mention showers of Mongol arrows falling from the skies, impaling men and horses. Mongol soldiers used powerful laminated bows and could fire them rapidly — and from horseback. They were the undisputed master archers of their age. In 1245, a papal envoy, Friar John of Piano Carpini, visited the Mongols and described their bows and arrows: “They are required to have these weapons: two long bows or one good one at least, three quivers full of arrows . . . the heads of their arrows are exceedingly sharp, cutting both ways like a two-edged sword, and they always carry a file in their quivers to sharpen their arrowheads.” Interestingly, the rusted mass I am floating over is the third bundle of arrows found at the site, and I wonder about the “three quivers” comment of the old priest. Could these be the arms of a single Mongol soldier?

Each of the seventy arrows in the bundle could easily penetrate the armor of a samurai. According to Father John, this was because of the Mongol technique of dipping forged iron arrows “red-hot into salt water, that they may be strong enough to pierce the enemy’s armor.” Some of the Mongol arrows were dipped in poison to weaken their opponents, and looking at the bundle of arrows, which rust has melded into a nearly shapeless mass, it is ironic to see how the salt water that once hardened them to make them more deadly has now taken the sting from them.

Another exciting find, resting upright on the seabed, is a Mongol war helmet. Close beside it are small fragments of red leather from a suit of Mongol armor, originally made of laminated strips of leather bound with brass. The mud has preserved these fragile traces by burying them out of the reach of the water. Along with the armor, the dredge gently uncovers a small tortoiseshell comb, a fragment of red leather still clinging to one side. I think about another discovery nearby—the bones of a drowned member of the ship’s complement, perhaps a Mongol warrior. The proximity of bones, helmet, armor and arrows raises the question of whether or not they all belong to one victim of the wreck. In the laboratory, just before the dive, I had looked at a broken skull that was found lying face down in the mud, and wondered what stories this victim of an ancient shipwreck could tell.

Some artifacts do tell tales. A small bowl, broken and found upside down, is painted with the name of its owner and his rank. One of my dive partners, Mitsu Ogawa, later tells me that the man’s name was Weng and that he commanded a hundred troops. I wonder if it is Weng’s armor, helmet, weapons and bones that lie together in the mud. Other artifacts tell us that the preparations for the invasion were hasty. Many of the ceramic jars are sloppily made, misshapen and badly fired, rushed into production for the war. The ship’s massive anchor may also be proof of haste. Unlike the one-piece stone weight for the anchor at Hakozaki Shrine, this anchor’s stones—and others found nearby at Takashima— are made of two crudely shaped pieces. The anchor for the ship now being excavated dragged in the mud and broke apart where the two stone weights were joined by wood and lashings—a fatal shortcut.

After our dive ends, we get into a discussion on the dock with Kenzo Hayashida. What sank this ship? The anchor is set tight, as if the ship dragged in heavy waves and then broke up. A storm might have sent the ship into the shallows and smashed it into the many pieces the Japanese are recovering. “The question is whether there was one storm,” says Hayashida, or “several centuries of storms.” I get his point. The periodic typhoons that lash this coast sweep into Imari Bay and churn up the seabed. The breakup and scatter of the huge wooden wreck being mapped by the archeologists may be the result of generations of storms, not a single catastrophic kamikaze sent by the gods. The timbers of the vessel also show evidence of burning. Did this ship go to the bottom as a result of a Japanese attack with a straw-filled “fire ship?” The fragmented remains may never reveal all their secrets, but they have already enabled archeologists to refute a few stories. Hayashida, who bases his opinion of just how many wrecks should be on the bottom of the bay from 1281 on years of surveys and the information they provide, firmly believes that the figure of four thousand is a gross exaggeration. “How many ships?” I ask him. “Maybe four hundred,” he answers with a smile.

Over the next week, we make more dives and watch as more artifacts slowly emerge from the mud. Broken timbers from the ship, including the sockets where a mast would have fit into the bottom of the hull; shattered planks; ceramic bowls and pots once filled with provisions; weapons and armor; and personal possessions, like a small delicately cast bronze mirror, are reminders of the individuals behind the myths and the big sweep of history. The personal items and bones are all that remain of the forgotten warriors who came here, on the orders of Kublai Khan, to expand an empire and an emperor’s prestige, and instead met their deaths far from home. I think of all the dead of 1281. And I think of the millions who died later in the 1930s and ’40s, victims of what was, if not a false legend about the kamikaze, then a distorted and exaggerated one that was used to justify the militaristic expansion of a “divine empire” and a brutal war.

If you find an error please notify us in the comments. Thank you!