The Castle That Was Already a Ruin: Hara-jō, the Siege of 1637–1638, and What the Excavators Found

The Fortress the Government Had Already Killed

What Was Hara Castle? The Abandoned Arima Fortress Occupied in 1638

In mid-January 1638, tens of thousands of people walked out onto a promontory at the southern tip of the Shimabara Peninsula and occupied a place that had officially not existed for more than twenty years.

Hara Castle, 原城, had been abandoned when the Arima clan was transferred out of the district, in 1614 by most accounts, though 1616 also appears. The abandonment was not neglect. It was policy. Under the ikkoku ichijō principle, one castle to one province, the Tokugawa regime required its lords to reduce their surplus fortifications, and when Shimabara Castle went up on the northern side of the peninsula, Hara came down. The gates were removed, the buildings taken away, the roofs gone. What the shogunate left behind was a set of crumbling stone revetments and earthen embankments on a headland, of no military value to anyone, which was precisely the intention.

The intention was wrong, and the reason it was wrong is geography.

The promontory rises steeply out of the water on three sides. There is one land approach, from the west, and it runs across flooded rice paddies. An attacker coming from that direction had to cross soft ground in the open, uphill, and could not go around, because there was no around. The whole complex of citadels retained the plumbing of the garrison town it had once been: a well inside the works, two large ponds, cultivable fields within the perimeter, and a small sheltered beach below the cliffs where boats could land out of sight.

The capacity of the place, as later estimates reckon it, was about forty thousand people.

Roughly thirty-seven thousand came. That number is the standard figure in the Japanese tradition, though the range runs from twenty thousand at the low end, which George Sansom preferred, up to figures above a hundred thousand that plainly count the whole movement across two districts rather than the people behind the walls. C. R. Boxer's breakdown remains the most-used: about fifteen thousand men of fighting age, some two hundred masterless samurai, and around twenty-two thousand women, children and the elderly.

The two hundred rōnin are the detail that explains everything that followed. They were, for the most part, former retainers of Konishi Yukinaga, the Christian daimyō dispossessed and executed after Sekigahara, men who had spent thirty-eight years as farmers in a district where nobody had forgotten what they used to be. They knew how a castle was supposed to be laid out, how a firing line was supposed to be dressed, how a sortie was supposed to be timed. Layered on top of them was an organization the prohibition had never dissolved: the confraria, the lay Christian brotherhoods, run by village headmen, the shōya, accustomed to collecting and distributing on behalf of a whole community.

Thirty-seven thousand people did not stumble into a ruin. They moved into it, in order, in January, with an inventory.

Fifty Boats, Cut Up for Timber

How the Rebels Refortified Hara Castle with Boats, Bamboo and Clay

The first thing the defenders did was destroy their own means of escape, and they did it for the wood.

Somewhere between fifty and a hundred fishing boats had ferried the Amakusa contingents across the bay, and once the last of them was over, the boats were broken up on the beach and carried up the cliff in pieces. Masts became ridge poles. Prows and keel timbers became beams. Planking became decking and revetment. In a district stripped bare by three years of failed harvests and a tax regime that had taken everything above subsistence and then a good deal below it, seasoned structural timber was not something anyone could buy.

For everything else they used bamboo, which the peninsula grows in quantity. Bamboo went up as watchtower frames and as the palisades along the parapets, and bamboo and rice straw went on as thatch. The command post was dug in as a turf bunker in the gap between two cliffs, which is a solution a professional soldier arrives at and a farmer does not. And along the parapets, where any other garrison in Japan would have hung the banners of a house, the defenders mounted crucifixes.

The result was a fortification with three entirely different structural personalities, and the distinction is not decorative. It decided the entire artillery question of the next three months. The uppermost fortress, the honmaru, sat behind a good high wall of heavy dressed stone, the surviving Arima-period work, on top of a cliff. The lower and outer defenses were topped with parapets of packed clay on earthen embankments. And the accommodation for thirty-seven thousand people, packed into every enclosure, was straw, matting and salvaged planking.

Smoothbore round shot, fired flat, deflects off good ashlar. Fired at a clay bank, it buries itself. Fired at a straw hut, it goes through and keeps going and hurts nobody unless it happens to hit a person on the way. There was almost nothing on that headland for solid iron shot to usefully break, which was a fact that would take the finest gunners available in East Asia about four days to discover and rather longer to admit.

The nominal head of this enterprise was a boy. Amakusa Shirō Tokisada, born Masuda Shirō, son of a former Konishi retainer, was fifteen or sixteen or seventeen depending on which account you follow. His function inside the castle was symbolic and sacramental rather than tactical, and the men who arranged the fields of fire had names of their own: Yamada Emonsaku, a painter and former Jesuit dōjuku, or lay catechist, who commanded eight hundred men.

They had food for perhaps ninety days and no prospect of resupply. Everyone inside the walls knew it. The political story of how they got there was already finished. What remained was engineering.

The Arithmetic of the Landward Side

Itakura Shigemasa's Failed Assaults on Hara Castle in 1638

The shogunate's first commissioner arrived on 15 January 1638 and spent a month proving the castle could not be taken by assault.

Itakura Shigemasa was fifty, and he had been sent with a force the sources place between thirty and fifty thousand men. His instructions were to end an embarrassment quickly. His problem was that the only ground he could attack across was wet paddy, and that a man wading uphill through mud with a matchlock is not a man who is going to shoot straight or move fast.

On 3 February he tried anyway. The assault failed and cost him more than six hundred dead.

Reinforcements kept arriving through the first half of February, and with them arrived the political problem that would define the rest of the siege. Contingents came in from twenty-five daimyō by the count in the entrenchment records, and the men at the head of them were not junior figures. Hosokawa Tadatoshi came from Kumamoto, five hundred and forty thousand koku. Nabeshima Katsushige came from Saga, three hundred and fifty-seven thousand koku, with his son Naosumi. Kuroda Tadayuki came from Fukuoka, Tachibana Muneshige of Yanagawa came at seventy years old with the authority of a Korea veteran, and Terazawa Katataka came from Karatsu with an interest of his own in the outcome, because Amakusa was his.

The total is one of the least stable numbers in the whole affair. James Murdoch's figure of 100,619 men exactly, said to come from Japanese military records, is quoted everywhere. The supportable range runs from around 120,000, the most common modern figure, to the 200,000 that Jonathan Clements gives and that the Portuguese captain Duarte Correa recorded, accurately, as the number people in Nagasaki were saying to each other. The reduction of one ruined castle on one headland became the largest domestic military mobilization of the entire two hundred and fifty years of Tokugawa peace.

Itakura, watching that army assemble around him and understanding exactly what it meant for his own reputation if somebody else finished the job, attacked a second time on 14 February.

The date was lunar New Year's Day, Kan'ei 15/1/1. The attack was unauthorized. It was also a total failure, and it killed him: Itakura was shot dead in front of the walls, and the shogunal casualty figure for the day runs to about five thousand.

The peasants had now destroyed two frontal assaults and one commissioner.

Drying to Death

Matsudaira Nobutsuna's Blockade: How the Shogunate Starved Hara Castle

The man Edo sent to replace him did not attack the castle once.

Matsudaira Nobutsuna, Izu-no-Kami, was in his early forties and had the kind of administrative mind the Tokugawa system rewarded, which is to say that he looked at Hara Castle and saw a logistics problem rather than an honor problem. Tachibana Muneshige told him the same thing: do not storm it, sit on it. The Japanese sources call the resulting strategy something close to drying to death.

The mechanism had three parts and they were all about closing an aperture.

The first was entrenchment. The besieging army dug in around the entire landward perimeter, threw up its own earthworks, and settled into covered positions from which it kept up continuous harassing fire at the parapets. This did not kill many defenders. It was not meant to. It meant that no one inside could stand on a wall, work the fields, or move between enclosures in daylight without being shot at.

The second was the sea. The shogunate assembled a blockading fleet of fifty junks from Higo and Chikuzen plus five large junks out of Nagasaki, fifty-five vessels, and closed the water on all three sides. That fleet had two jobs, and only one of them was the garrison. The other was Manila and Macao. The regime had convinced itself, without evidence but not without logic, that a Catholic rebellion of this size implied Catholic sponsorship, and the junks were there to intercept a relief expedition that was never coming. They also shot at the castle from close range with everything they had, and their guns were too light to make any impression on the stone, which is the fact that produced everything that happened next.

The third part was time, and time was the only weapon on the field that actually worked. Hara Castle had a well and two ponds, so thirst was never going to end it. It had fields inside the perimeter, which is why the garrison lasted as long as it did, and which is also why the covered fire mattered so much. Take away the ability to work the ground and a self-sufficient fortress becomes a very large sealed room with a fixed quantity of rice in it and thirty-seven thousand people breathing.

Every day the siege continued, the shogunate's most powerful western lords sat in the mud in front of a peasant garrison, and every failed assault cost something more expensive than lives.

So he sent to Hirado for heavier guns, which the Japanese did not have and the Europeans did.

Six Barrels of Powder and a Great Deal of Wine

Why the Dutch VOC Aided the Shogunate at Shimabara in 1638

The first approach to the Dutch had been made a month earlier, by Itakura, on 13 January 1638, and it was a request for gunpowder and weapons rather than for ships.

The Dutch East India Company, the VOC, had been trading out of Hirado since 1609 on terms that made them, in Japanese eyes, something between foreign merchants and the retainers of a Japanese lord. Their factory operated under the territorial authority of the daimyō of Hirado, and so the formal order reached them through him. The lord in question was Matsuura Shigenobu, who had succeeded that July; the older literature names his predecessor Matsuura Takanobu, who had been dead since 16 July 1637 and was in no position to order anybody to shell anything. Behind him stood Matsudaira Nobutsuna, with the Nagasaki magistrates Sakakibara Hida-no-kami and Baba Saburōzaemon applying the pressure, and the hereditary Nagasaki daikan, or deputy governor, Suetsugu Heizō, applying rather more than pressure.

Suetsugu is on record with the sentence that summarizes the entire transaction. One ought, he told the Dutch, to do some service for the profits he enjoys.

The opperhoofd at Hirado, Nicolaes Couckebacker, spent the next six weeks conducting a masterclass in the bureaucratic arts. He was, by his own factory diary, highly reluctant to fire on fellow Christians, and rather than refusing, which would have been suicidal, he complied slowly and incompletely in every particular. He sent florid greetings. He sent gifts of wine and candy, which bought days. He shipped six barrels of gunpowder and stated that this was the entirety of the Company's stock, which it was not. On 19 February he ordered the second Company ship at Hirado, the Petten, to sail for Taiwan, removing her from any list of available vessels. And when he finally did sail for Shimabara, he did it in the manner of a man being taken to a dentist, moving only in daylight and anchoring at offshore islands each night.

He complied because the alternative described to him was confiscation of every Dutch ship, cargo and asset in Japan, and because the regime already suspected the Dutch of sympathy with the Catholic rebels.

What this looked like from the other side was entirely different. Adam Clulow's reading of the shogunal record finds the service filed under hōkō, loyal service, the same category used for the military obligations of Japanese lords, with commanders noting approvingly that the Dutch had presented their service as was done by the other lords. On that reading the Dutch were not coerced hostages but participants performing vassalage, and the loyalty test is a story assembled afterward.

Both readings rest on real documents, and they are looking at different documents. The shogunal archive records what the act was made to mean. The factory diary records a merchant hiding a ship and lying about his powder stocks. The gap between them is roughly the gap between a press release and an internal memo.

On 24 February the ship arrived off the promontory, fully manned.

Three Hundred and Sixty-One Rounds

How Many Rounds the Dutch Ship de Rijp Fired at Hara Castle

The vessel was the Rijp, sometimes written De Rijp or Reip, carrying twenty guns.

Her most important contribution was made before she ever anchored. On 10 February her five largest guns had been dismounted and carried overland to Arima to be emplaced in a shore battery. A Dutch warship standing off a headland and pounding a castle is a good picture. What actually happened was that a Dutch gun crew built an artillery position on land, alongside the samurai, and served it.

Couckebacker's factory diary records the daily expenditure of shot, one of the very few sets of figures in this siege written down by a participant in the same week. Firing opened on 26 February with fourteen rounds from the ship, or possibly on the 25th. Twenty-seven on the 27th, thirty-five on the 28th, eleven on 1 March, thirty-two on the 2nd, eight on the 3rd, eighteen on the 4th. On 5 March the diary reports a steady barrage while the battery was being reinstalled and gives no figure; on the 6th the ship was silent and the battery fired thirteen; the entry for 7 March is missing. Eighteen on the 8th. Then the tempo rose: seventy-three on 9 March, thirteen from the ship and sixty from the shore, forty-two on the 10th, forty-four on the 11th, twenty-six on the 12th.

The sum is 361, split eighty-nine from the ship and two hundred and forty from the shore, with two undocumented days unaccounted for. The shore battery fired something close to three rounds for every one that came from the sea.

The figure that appears in almost every account of the siege is 426, usually bundled with the phrase "in fifteen days." The contemporaneous Dutch log supports about 360 plus two blank days.

What all of it achieved was assessed, in writing, by the man directing it:

"…we saw clearly that we could do nothing important with our guns, as the houses are merely made of straw and matting, the parapets of the lower works of defence being made of clay and the uppermost fortress being surrounded by a good high wall, built with heavy stones."

The castle sat on a cliff, which meant that shipboard fire had to be lobbed upward at a target whose range changed with every swell, and gunnery from a rolling deck against an elevated target is a matter of hope rather than mathematics.

The psychological effect was another matter, and it was real. Continuous plunging fire, however inaccurate, drove the defenders to dig cellar-like shelters and crowd into them, which meant that thirty-seven thousand starving people spent the first half of March underground. On 1 March a Dutch ball landed in the shogunate's own camp. Later a round struck the main keep, injured several retainers, and tore the left sleeve of Amakusa Shirō's robe, which the defenders read as divine protection and the besiegers read as evidence that the guns were at least pointed the right way.

The only confirmed Dutch fatality of the operation was caused by a Dutch gun. On 11 March an iron culverin in the shore battery burst near the muzzle, and a fragment struck a bricklayer and stonemason named Gillis in the abdomen. Portuguese and French Catholic accounts, which had an obvious interest in recording Dutch blood spilled in the service of a pagan against Christians, variously report two Dutchmen shot down by fire from the castle. The Dutch and Japanese records know of one man, and he was killed by his own artillery.

The Letter on the Arrow

The Arrow Letter and Why the Dutch Were Dismissed on 12 March

On 1 March, while the batteries were firing, someone inside Hara Castle tied a letter to an arrow and shot it into the Tokugawa lines.

The factory diary's translation is restrained: the besieged asked why the Netherlanders had been called to give assistance, there being so many courageous and faithful soldiers in Japan. The version in the samurai diaries is not restrained at all. Although there are many honorable soldiers in Japan, why do you seek the aid of the Dutch?

Other pasquils, lampoons, followed it over the parapet. One of them observed that the shogun's finest were masters of the abacus rather than of arms, which is a remarkable thing to write when you have not eaten rice in a month and are being shelled by two nations at once, and which was aimed with complete precision at the one place the besieging army could be hurt. Kyūshū's great houses had spent twenty years being converted from warriors into administrators by a regime that no longer had wars for them, and had now spent six weeks failing to take a ruined castle from farmers with the help of hired heretics.

It worked.

On 12 March the Dutch were formally thanked and dismissed, and the reasons in the two record traditions converge on five. Samurai pride had been publicly wounded, and Hosokawa and Nabeshima made their resentment plain; Matsudaira himself agreed that the arrangement looked strange and made Japan appear weak. The siege trenches had been pushed so close to the walls that blind high-angle fire from the sea was endangering the besiegers, a risk the round that landed in the camp on 1 March had already demonstrated. The gunnery was not working, and everyone including Couckebacker had said so. Gillis's death was awkward, because the shogunate had guaranteed Dutch safety. And the point of the exercise had already been made: the Dutch had fired on Christians at Japanese request, in front of an audience of assembled daimyō, and there was nothing further to prove.

The Rijp weighed and went back to Hirado. The shogunate turned to mining, and to a conversation with a traitor.

What They Found in the Stomachs

The 4 April Sortie and the Stomachs That Showed Hara Was Starving

Yamada Emonsaku had been in correspondence with the besiegers for some time. His final message left the castle by arrow on 1 April. Shortly afterward the rebels caught him at it and locked him in a cell to await a sentence that events overtook.

By then the garrison was three months into a food supply meant for three.

On Easter Sunday, 4 April, and into the following day, roughly three thousand of them came out. It was not a breakout; there was nowhere to break out to. It was a raid for provisions against the Nabeshima, Kuroda and Terazawa camps, mounted by starving people at night against entrenched troops, and it was extraordinarily well executed. They inflicted about five hundred casualties and lost about four hundred.

What the shogunal officers did next is the single most informative act of the entire siege.

They cut open the stomachs of the dead.

This was not desecration, or not only desecration. It was an intelligence operation, and it was the correct one. The one thing Matsudaira could not observe from outside the walls was how much food remained inside them, and the only reliable sample available to him was the interior of the men who had just been eating it. The examination found seaweed, tree leaves, and unripe barley cut green from the fields inside the perimeter.

There was not one grain of rice in any stomach.

That settled it. A garrison eating leaves has days, not weeks, and a starving enemy cannot counterattack, cannot sustain a long perimeter, and cannot man the walls he has. The blockade had cost the samurai essentially nothing, and it had done what six hundred dead on 3 February and five thousand on 14 February had not.

Neither the European nor the Japanese record for Hara preserves clear evidence of tunneling, though mining was under discussion in the second week of March, and mine warfare was standard Japanese siegecraft, used at Osaka in 1614. Nobody has yet found the shafts.

They did not need them. The castle had already been beaten by arithmetic: thirty-seven thousand mouths, one harvest's worth of stored grain, a perimeter that could not be worked under fire, and a sea sealed by fifty-five junks.

Two Days of Hand-to-Hand

The Final Assault on Hara Castle and the Killing of the Garrison, 1638

The general assault was supposed to begin on 11 April.

It began on the evening of the 10th, because Nabeshima Naosumi moved roughly eighteen hours early to claim the honor of first entry, and the resulting fighting was confused, unsupported, and marked by units shooting into each other in the dark. The formal attack followed on the 11th and ran into the 12th. The outer rings were carried by storm, enclosure by enclosure, over the clay parapets and the bamboo palisades and through the straw shelters, and the defenders, who had been eating leaves for a fortnight, fought for every one of them.

The honmaru was breached on 12 April 1638, Kan'ei 15/2/28. That equation is worth stating plainly because a widely repeated alternative gives the fall as 15 April, equated with Kan'ei 15/3/1, and the equation is arithmetically false: 15/3/1 fell on 14 April. The lunisolar calendar has scattered wrong dates through this whole episode, and the Dutch bombardment dates are the most secure in the account for the unglamorous reason that a Dutchman was keeping them in Gregorian time.

What followed inside the walls is recorded in numbers rather than description, which is how it should be recorded.

Clements gives thirteen thousand killed where they stood and twenty-three thousand, chiefly women and children, burned alive, beheaded, or thrown from the cliffs into the sea. The garrison was not offered terms and did not ask for them. One hundred and five people were taken prisoner. Yamada Emonsaku was found alive in his cell, which is why he is habitually described as the sole survivor, a claim that holds only if it is narrowed to the sole survivor of the defending garrison proper, since a number of his own men were also found alive in the castle prisons.

Amakusa Shirō was discovered badly wounded in the ruins of the main keep, wearing dirty white robes, and beheaded on the spot. His mother Martha, in shogunal custody throughout, was brought to identify the head. It was then displayed at Nagasaki, in the city where a generation of Christians had already been executed on the hill above the harbor.

The shogunate's own losses are reported with the usual instability. The official return submitted to Edo gave thirteen thousand killed and wounded. Neil Fujita's figures are 1,992 dead and 10,656 wounded. Clements puts total casualties above twenty thousand. The surviving per-clan returns give Hosokawa 274 dead and 1,824 wounded, Kuroda 213 and 1,657, Tachibana 181 and 779.

Roughly one hundred thousand of the best soldiers in Japan had taken four months and something like twelve to twenty thousand casualties to reduce a fortress the government had already demolished, held by farmers.

The Flattening

After Hara: Matsukura Katsuie's Execution, the 1639 Expulsion and Dejima

When it was over, the shogunate did to Hara Castle what it had done in 1614, only properly.

The stone walls were pulled down, not slighted at the gates in the usual formal manner but thrown down along their length. The enclosures were leveled. Dressed blocks that had stood since the Arima period were tipped off the revetments and left where they landed. The instruction was not to decommission a fortress. It was to make certain that the ground could never again be held by anyone, which is a fair measure of how badly the four months had frightened the state that gave it.

The accounting extended to the living. Matsukura Katsuie, whose family's fiscal regime had produced the rebellion by reassessing a domain of sixty thousand koku at a hundred thousand and then collecting on the fiction, was beheaded on 28 August 1638. That sentence is the point. A daimyō of the Edo period who fell out of favor was permitted seppuku; Katsuie is standardly described as the only one denied it. Terazawa Katataka lost Amakusa outright.

For everyone else, the consequences ran in one direction and did not stop for two centuries. In 1640 the shogunate created the Office of Inquisition, the shūmon-aratame yaku, under Inoue Chikugo-no-Kami, and made every subject register at a Buddhist temple under the terauke system, converting religious affiliation into a permanent administrative record.

The Portuguese were finished too, though not for the reason usually given. The 1639 edict banned their shipping outright, and when Macao sent an embassy in 1640 to plead for the trade, sixty-one of the seventy-four men were beheaded on the hill at Nagasaki. But the Dutch were not handed the Japan trade as payment for Hara Castle. The Portuguese went on trading under surveillance after the rebellion was suppressed, and the shogunal council refused to sever the Macao link until it was satisfied that Dutch and Chinese shippers could deliver the same volumes of raw Chinese silk that the domestic clothing industry ran on. Once the Company's Taiwan pipeline was stable, the ban followed, on 4 August 1639. It was a supply-chain decision.

The Dutch reward, such as it was, arrived in the autumn of 1640, when shogunal officials inspected the Company's new stone warehouses at Hirado and ordered them demolished because the gables carried the Christian-era date A.D. 1640. In May 1641 the Dutch were moved to Deshima. Engelbert Kaempfer, the German physician who served there in the 1690s, wrote that it was the auri sacra fames, the accursed hunger for gold, that had made his employers behave in a manner that shocked even the nobler-minded Japanese. Jonathan Swift put them in Gulliver's Travels.

They had fired three hundred and sixty rounds at a straw roof and been laughed at for two hundred years.

What the Spades Found

What Archaeologists Found at Hara Castle: Bones, Lead Crosses and a Gold Reliquary

Systematic excavation of the site began in the 1990s, and it has confirmed the documents to a degree that is unusual and, in places, unwelcome.

The first thing the archaeologists found in the honmaru was bone, in a density the excavators described as being like a carpet: skulls, loose teeth, fractured limbs, men and women and children mixed together in the same deposits, which is itself a finding, because it demonstrates a population that shared the defense and the logistics rather than a garrison with dependents behind it. The bone was never a secret. In 1766 the fields around the ruin were still turning it up, and the local response is attested in the erection that year of a memorial stone figure for the dead of both sides, the Honekami Jizō.

The most abundant small finds are crucifixes, and they are the reason this site matters to the history of the Nanban century rather than only to the history of Japanese siege warfare.

They are small, crude, and cast from lead. The defenders made them inside the castle during the siege, and the metal was the shogunate's own musket balls. Excavators have recovered them lying beside skulls and inside jawbones, which is generally read as evidence that the dying held them in their mouths.

Then somebody ran the isotopes.

Lead from different ore bodies carries different ratios of its stable isotopes, and the ratios survive smelting and casting, so a lead object is stamped with the geology of the mine it came from whether or not anyone intended it to be. Analysis of shot and cast crosses from Hara returned signatures matching neither Japanese nor Chinese nor Korean ore. The lead had come from Southeast Asia, which is to say that it had reached Japan through the European maritime trade of the previous ninety years, sold into the archipelago as ammunition by the same commercial system that brought the priests. It was fired at the castle by the shogun's troops, picked up off the ground by the people it had been fired at, melted, and cast into devotional objects.

The rest of the assemblage is quieter. Bronze medals. Rosary beads in blue and green and white glass, many of them plainly old when they went into the ground, family pieces carried through a generation of prohibition into a castle that everyone inside knew was going to fall. And, in February 1951, a farmer digging in the honmaru turned up a reliquary cross of twenty-two-karat gold, identified in local tradition as a gift from Pope Gregory XIII to the boys of the Tenshō embassy, carried home in 1590 and presented to the Christian daimyō Arima Harunobu. Its provenance chain rests on that tradition rather than on documentation, which is worth saying out loud before anyone repeats it.

One category of object is conspicuously missing. No excavated cannonball at Hara has been securely attributed to the Dutch guns. There are cannonballs, and there are no cannons, and the most famous bombardment in the history of the Japanese Christian century has left no physical trace on the ground it was aimed at.

Which is exactly what Couckebacker said would happen, in his diary, at the time.

The Field of Broken Stone

Hara Castle Today: A UNESCO World Heritage Site Since 2018

On 30 June 2018 the UNESCO World Heritage Committee inscribed a serial property under criterion iii with the English title Hidden Christian Sites in the Nagasaki Region. One of its components is listed as the Remains of Hara Castle.

The word remains is doing considerable work. There is no keep, no gate, no restored turret, none of the white plaster and black tile that constitute a Japanese castle in the popular imagination. There is a headland with the sea on three sides, a view across the water to Amakusa, grass, the outlines of enclosures in the ground, and the demolition field: heavy dressed blocks lying where the shogunate's engineers pushed them off the walls in 1638. What the visitor walks on is not a castle. It is the negation of one, performed twice, in 1614 by administrative order and in 1638 by an army that had learned the hard way what the first demolition failed to accomplish.

The site was chosen by the rebels for exactly the qualities it still has. Cliffs on three sides, a single approach across wet ground, water inside the perimeter, arable land inside the walls, a beach out of sight below. Thirty-seven thousand people took a ruin the government had personally decommissioned and, with fifty fishing boats, local bamboo, packed clay and rice straw, made it hold for four months against the largest army the Tokugawa state ever assembled on its own soil. It did not fall to European artillery, which achieved nothing and knew it. It fell because a competent administrator sealed the sea with fifty-five junks, pinned the defenders inside their own walls with entrenched fire, and waited until the corpses of the men who came out to raid could be opened and read.

The lead they melted for their crosses had come to Japan in Portuguese and Chinese ships, and it went into the ground with them, in the jawbones, on the promontory, under the grass.

Sources & Further Reading

Boxer, C. R. The Christian Century in Japan, 1549-1650. University of California Press, 1951. The standing English account of the period and the source of the garrison breakdown used here, fifteen thousand fighting men and two hundred Konishi rōnin among roughly thirty-seven thousand people, together with the translations of Duarte Correa that carry the Portuguese view of the siege.

Clements, Jonathan. Christ's Samurai: The True Story of the Shimabara Rebellion. Robinson, 2016. The fullest modern English narrative of the campaign, and the source of the castle's forty-thousand capacity, the massacre breakdown of thirteen thousand and twenty-three thousand, and the per-clan casualty returns.

Clulow, Adam. The Company and the Shogun: The Dutch Encounter with Tokugawa Japan. Columbia University Press, 2014. The argument that the bombardment was performed vassalage filed under hōkō rather than coercion, which is one half of the interpretive disagreement this article preserves rather than resolves.

Correa, Duarte. Relação do alevantamento de Ximabara. Lisbon, 1643. Written in Ōmura prison by a Portuguese captain who had been in custody since 4 November 1637 and saw none of the fighting, it preserves the fiscal explanation of the revolt as it circulated among non-Christian Japanese informants.

Dougill, John. In Search of Japan's Hidden Christians: A Story of Suppression, Secrecy and Survival. Tuttle, 2012. Connects the site to the later kakure kirishitan communities of the Nagasaki region and to the landscape the World Heritage inscription eventually recognized.

Elison, George. Deus Destroyed: The Image of Christianity in Early Modern Japan. Harvard University Press, 1973. Supplies the lunisolar reconciliation that makes Kan'ei 14/10/25 into 11 December 1637 and exposes the transposed dates that circulate elsewhere.

Fujita, Neil S. Japan's Encounter with Christianity: The Catholic Mission in Pre-Modern Japan. Paulist Press, 1991. Source of the precise shogunal casualty return of 1,992 dead and 10,656 wounded that competes with the round official figure of thirteen thousand.

Kaempfer, Engelbert. The History of Japan, Together with a Description of the Kingdom of Siam. Translated by J. G. Scheuchzer. Thomas Woodward, 1727. The Deshima physician's condemnation of his own Company's conduct as auri sacra fames, and the channel through which the episode reached the European reading public.

Murdoch, James. A History of Japan, Volume II: During the Century of Early Foreign Intercourse, 1542-1651. Kegan Paul, Trench, Trübner, 1925. The origin of the frequently quoted count of 100,619 shogunal troops and of the standard English narrative of Itakura's two failed assaults.

Oka, Mihoko. The Namban Trade: Merchants and Missionaries in 16th and 17th Century Japan. Brill, 2021. Establishes the commercial framework behind the silk-supply decision of 1639, which is the actual reason the Portuguese were expelled and the reason the monopoly was not a reward.

Sansom, George. A History of Japan, 1615-1867. Stanford University Press, 1963. The low-end estimate of the garrison at twenty thousand, and the older institutional account of the shogunate's response to the rebellion.