Toba fell ill, and his condition progressively worsened, until a bold fortune-teller said the words no other courtier would utter: that his mistress, with her odd mastery of scripture and her propensity to glow in the dark, was not a Buddhist saint at all, but a malicious demon who intended to kill Toba and supplant him. Tamamo-no-mae supposedly disappeared at this point, leading to a savage cull of foxes in the surrounding countryside until Toba regained his health.

I repeat the story here not for its historical accuracy, which is nonexistent, but for the glimpse it offers of the whispers and petty jealousies of Heian life, with bedroom companions influencing political decisions, and courtiers hiding behind coincidence and innuendo in their fox-shaming campaign against some poor concubine. Tamamo-no-Mae was never seen again, although her angry spirit was said to influence many of the scandals that followed. Even in the afterlife, it seems, there were intrigues and scandals, dead emperors and wronged courtiers who might be persuaded to avenge forgotten insults. It was, some said, the curse of Tamamo-no-mae that brought down Toba’s young proxy, Konoe; the young boy was always sickly, and reigned for barely more than a decade, dying at the age of seventeen, before he had the chance to sire an heir of his own.

The year was 1155. Retired Emperor Sutoku hoped to regain the throne, but Retired Emperor Toba still had seniority, and managed to recommend that his own fourteenth son, Sutoku’s brother, should be crowned as Japan’s seventy-seventh emperor, Go-Shirakawa (1127–92). Sutoku had hence been passed over in the succession three times—forced to abdicate against his will, and then replaced by two of his siblings when he regarded himself as a prime candidate for restoration. There was also a scurrilous rumor, never quite discounted, that Toba hated Sutoku because he wasn’t really his son at all, but the secret love-child of Toba’s father, sired on Toba’s wife in some tawdry incident.

If all that looks confusing, it’s only half the story, since these feuding emperors were themselves merely the outward manifestation of another conflict underway over who got to be the emperor’s chief minister. In fact, it hardly mattered who the emperor was; the real issue was who his mother was, with the various fallings in and out of imperial favor masking internal conflicts within the Fujiwara family, which had supplied most of the brides and concubines, and hence most of the regents.

Nobody dared challenge the decision directly, and the new emperor Go-Shirakawa, a man who had never expected to be emperor and rather seemed taken by surprise by the whole thing, endured a tense first year on the throne, ending in the summer of 1156 with the death of his father Toba. Toba had taken two months to die, on a sickbed attended by hushed whispers and intense conferences, in a mansion guarded by stern samurai.

It was Toba who had held everything together, and whose factions had crushed any resistance. With him gone, Sutoku was the new senior retired emperor, and he was ready to pounce.

Emperor Go-Shirakawa knew trouble was brewing. Three days after his father’s death, his officials were ordering samurai to steer clear of the capital. Two days after that, known associates of Retired Emperor Sutoku were directly ordered not to recruit troops. Forty-eight hours later, samurai loyal to the incumbent emperor and samurai loyal to the retired emperor clashed in open combat on the streets of Heian.

It was a landmark moment. The intrigues of the court had erupted into open violence, and had done so not at the border, but within the very capital itself. That, at least, was how things felt to the court at large—the attentive reader will recall that some of the courtiers’ own ancestors were not above stabbing their enemies to death in the emperor’s presence in ages past—but it seems that many of the contemporary courtiers had come to believe their own hype, and were ill-prepared for violence returning to their doorstep.

The samurai in play amounted to several hundred on each side, but the only prize was Go-Shirakawa himself, who might be persuaded to abdicate if he fell into the hands of his brother’s rebels.

There were Fujiwara courtiers and Minamoto samurai on both sides of the conflict. Unfortunately for the pro-Sutoku faction, their nominal leader, Fujiwara Yorinaga, was very much an armchair general whose ideas about warfare were based solely on the idealized, rather ceremonial events described in old stories and songs. His Minamoto advisers, veterans of many an asymmetrical skirmish in the northern wars, suggested that the best thing to do was to start a fire at the emperor’s residence, which was sure to lead their target to flee in his palanquin with a small group of bodyguards. They could then overwhelm the guards, seize the palanquin, and thereby obtain control of the only figure who could order the enemy to stand down. The conflict would be over before it started, with minimal loss of life.

Yorinaga was not interested. The whole thing sounded sneaky and underhanded to him, and he very much preferred to imagine things the way they were in the old songs, with a few hundred samurai marching out to a nice area of flat ground, stating their names and lineages, and then taking each other on in single combat until the victor was revealed.

It does not seem to have occurred to Yorinaga that if his own samurai had come up with the idea for such a ruthless, surgical strike, then the enemy, whose samurai hailed from a different branch of the same family, was liable to have a very similar idea. In fact, his enemies had already apprehended one of his men, who had spilled all their plans, leading the incumbent emperor to authorize the seizure and search of Yorinaga’s house.

At dawn on the eleventh day of the seventh lunar month, 1156, the emperor led his court in prayers while his loyalists converged on Yorinaga from three directions with several hundred mounted men. Within an hour, there were flames and smoke in the east of the city. The battle was bloody but brief, although its aftermath would stretch on for two generations.

Several of the rebel leaders were killed in the skirmish. The pretender Sutoku was packed away into monastic exile on a remote island, where he lived for another eight years, muttering curses against his enemies, and, it was said, forming a malicious faction in the afterlife with the fiery fox spirit Tamamo-no-mae. In subsequent years, his angry ghost would get the blame for many famines, earthquakes, and misfortunes, becoming one of the great bogeymen of Japanese history.

For centuries, the Kyōto aristocracy had boasted of the civilized nature of their capital. It was a mark of the drastic changes in attitudes and expectations that the uprising ended with a round of beheadings. Courtiers had prided themselves on the peaceful capital for the last three and a half centuries—nobody had been executed in Kyōto since the failed coup of Retired Emperor Heizei in 810. Now, Sutoku’s surviving supporters were executed, sometimes in cruel situations in which their own relatives were ordered to carry out the task.

In the most infamous case, the Minamoto loyalist Yoshitomo was ordered to behead his own father. He was unable to carry out such a terrible command, but one of his lieutenants, seeing that a Minamoto would die at the hands of a Taira unless he took action, did the deed himself. Shortly after he had spared his lord from committing patricide, the loyal lieutenant killed himself in contrition.

It was by no means the first reference to suicide in the tales of the samurai, nor even in the events of the Hōgen Insurrection. But it is during this failed rebellion that the chronicles of the samurai first start referring not only to suicide, but to a particular kind of suicide. The cult of the samurai had already begun to take on certain new elements. One was the desire to wear flashy armor, decorated with striking icons or tied with distinctive color strings, in order to make it clear who was winning fame on the field of battle. Samurai helmets, in particular, became notorious for their ostentatious adornments; these have included, among many other things, a giant snail shell, insect’s wings, antlers, devil horns, sunbursts, and rabbit ears. The samurai had started to develop a sense of themselves that placed them on a hierarchy of bravery and battle prowess, and that meant it was necessary for their victories to be obvious to all. A side effect of this ease of identification was that it would also be clear who was running away. The distinctive nature of samurai battlefield adornments promoted a gung-ho sense of always charging, never retreating.

There were times when victory was impossible. Samurai might be surrounded with no possible retreat. They might be disarmed. They might find themselves just about to fall into enemy hands, where they might suffer the further shame of being used as hostages or bargaining chips, or tortured for information. Or, like Yoshitomo’s lieutenant, they might find themselves in an impossible situation, where they had done the right thing by their lord but could not possibly be expected to go on living after having done so.

Instead, they chose to kill themselves, but not with the throat-slitting or defenestration favored by women in search of a quick death. Instead, they killed themselves in the most painful way imaginable, by slicing open their own abdomen as a mark of their bravery and inner strength—the belly was thought to be the seat of the soul, and hence also a mark of sincerity. Cutting the belly, seppuku (more vulgarly, hara kiri) was a one-way trip to agony. There was no cure; only a slow, lingering death. The decision to slice open one’s abdomen was also a get-out clause for one’s underlings—they would not dare lift a finger against their master, but would be justified, once he had voluntarily wounded himself in such a fashion, in ending his suffering by beheading him.

Over the years, seppuku would take on new rituals. Samurai would wear a white kimono, symbolising death and purity. They would write a death poem, ensuring that parting words, criticisms, or curses were encapsulated in repeatable form. The nature of the wound would become deliberately cruel, with “tradition” demanding four cuts through the abdominal muscles—shi, meaning four, being a homonym for death, but also demanding incredible determination and strength of purpose in the self-harming samurai. Seppuku started as a battlefield compromise—a last resort by besieged men in burning castles, determined not to surrender to enemies who would torture and humiliate them. But once it became enshrined in tradition, it became the default means of repentance, and even criticism. It faded out after the era of the samurai, but still occasionally returns to haunt the country.

If this seems shocking to the modern reader, we should bear in mind that religious belief played an important part. Buddhism had taken hold, but with a certain nihilistic angle. The concept that “all life is suffering” had been embraced by the Japanese with a melancholy sense of poetry, as well as a certain sense that the end of the world was nigh. Certain Buddhist scriptures predicted the rise, peak, and subsequent fall of the Buddha’s teachings: five hundred years of struggle for success, a thousand years of worship and achievement, and then five centuries of worsening conditions as things fell apart. It was, hence, widely believed among the medieval Japanese that they were living in the “Latter Days of the Law” (mappō). Any natural disasters, reversals of fortune, or atrocities could be written off as further evidence that the teachings of Buddha were under attack, and that any ends available would justify the means of sustaining them.

One particular Buddhist sect, the Essence of the Pure Land (Jōdo Shinshū) gained ground in medieval Japan. Pure Land Buddhism regarded the country’s troubles as yet another example of the Latter Days of the Law, in which it was almost impossible for anybody to engage in correct Buddhist devotion. In a sense, Pure Land Buddhists all but gave up trying, instead paying a new form of devotion to Buddha that recognized that things were terrible—people were trapped in cycles of toxic karma, eating meat, drinking booze, fornicating, and otherwise coping with the onrushing end of the world—but that it was still possible to at least make it obvious to Buddha that you bore him in mind. You would do this by chanting “Namu Amida Butsu” (I take refuge in Amida Buddha) as often as possible, as a little spell to hold back the worst of the world. Most importantly, Pure Land Buddhism was a sect that offered the chance of rebirth in a Buddhist paradise to absolutely everybody. It was not restricted to monks or the rich who could afford costly demonstrations of devotion; literally anyone could find refuge in the Pure Land—even warriors.

Buddhism was actually abundantly clear about killing people being a sin. “A disciple of the Buddha,” said the fifth-century Sutra of Brahma’s Net, “should not possess swords, spears, bows, arrows, pikes, axes, or any other fighting devices. Even if one’s father or mother were slain, one should not retaliate.”

It was, however, the Zen flavor of Buddhism, originating in the Shaolin Temple in China, which achieved prominence among the samurai. Yes, killing people would bring about bad karma, but what about standing up for what was right, if that involved breaking a few heads? What about killing an assassin hell-bent on killing one’s lord? In such cases, presumably we would not be talking so much about bad karma, but about the least-worst.

Zen found plenty of adherents in Japan’s warrior class, in part because of some of its teachers’ habit of cutting through knotty issues of philosophy with seemingly dismissive put-downs. In fact, there was substantially more to it than that, but the nature of certain Zen parables and questions for meditation did lend itself well to a breed of anti-intellectualism. The Chinese Zen master Linji, for example, once famously said, “If you see the Buddha on the road, kill him.” He meant that the earnest Zen scholar should question all presumptions, and never lean on credentials or blind faith. But in the hands of the samurai, this became a recipe for a nihilistic battlefield philosophy.

It is often necessary to read between the lines in comments from the history books about “Buddhist monks” in medieval Japan. We already know, for example, that certain retired emperors were shaving their heads and ruling “from the cloisters,” even though their lives (and loves) continued in much the same way as they did in lay life. We also know that wily landowners were evading their tax responsibilities by “donating” their lands to Buddhist monasteries. With such deceptions at all levels of Japanese religious life, it should come as no surprise that there was an entire class of Buddhist “monks” who were little more than shaven-headed militia employed as military muscle to deal with their institution’s widening secular responsibilities. Even legitimate temples got in on the act, employing mercenaries to protect them from their newly proactive rivals.

Despite proscriptions against violence in other areas of Buddhism, and indeed within Zen itself, the interpreters of Zen among the samurai came to regard it as a warrior’s creed. Meanwhile, monasteries of doubtful provenance—some established as tax refuges—were prepared to offer prayers for the soul of a samurai who killed in the name of justice. Although not quite like the selling of indulgences in a European sense, it did give rise to a warrior class whose members felt that their religion entitled them to fight.

It was during the time of the wars of the Taira and Minamoto that Zen Buddhism first began to take hold in Japan, brought back to Japan, like so many other things, by monks who had studied in China. Zen was an offshoot of Buddhism that emphasized self-reliance. As brought to China by the monk Bodhidharma, Zen was a teaching “outside the scriptures”; sometimes this was interpreted as an extremely brawny, no-nonsense dismissal of much scripture and philosophy in favor of sparks of insight and moments of direct action.

Zen Buddhism hence threw away many of the accretions of Buddhist religions in favor of the cultivation of enlightenment (satori)—a perpetual moment of clarity. The version brought to Japan by the monk Eisai (1145–1215) was keen on short, punchy aphorisms designed to function as tools for thinking. Known in Japanese as kōan, these parables have come to characterize much Zen thought, as acolytes meditate on such questions as “What is the sound of one hand clapping?”; “What is the face you had before you were born?”; and that old favorite from Tang-dynasty China, “If you see the Buddha on the road, kill him.”

Later sects postdating the Taira-Minamoto war would introduce other ideas, such as zazen, “sitting meditation,” in which the aspirant emptied his mind of all thought except for a single mantra or goal. This was particularly appealing to the samurai, who loved the idea that there was no difference between life and death—there was only the singleminded pursuit of one’s mission.

Buddhism, particularly Zen Buddhism, soon turned in the hands of the samurai into an elaborate game of death in which killers accepted the risk of bad karma balanced against the accrual of merits for loyal service and just actions. As Buddhism splintered and evolved in Japan, there were plenty of sects that could offer warriors the chance to buy off bad deeds with donations and penances, and priests who spoke of the wheel of reincarnation. The samurai believed that the relationship between a lord and vassal was, if not immortal, then sure to last for at least three lifetimes. Die well in this life, and you were assured of respawning at a higher social station, under better conditions, perhaps even having been dealt a better hand. Die badly or with dishonor and you might not return as a samurai at all, but as a peasant, or a woman or an animal. In the multiple reversals of fortune and wars over nothing that would come to characterize medieval Japan, a “good death” became one of the primary aims of a samurai life.

And the result? As implied by the opening lines of The Tale of the Heike, you might say that it was all for nothing. Go-Shirakawa, the reigning emperor in whose name so many fought and died, sat on the throne for barely two years before deciding that he, too, would abdicate in favor of his own teenage son, the seventy-eighth emperor, Nijō (1143–65).

Go-Shirakawa would remain the main power broker for the next thirty years, through the troubled reigns of five successors. He gained such a reputation among historians for cunning plans and dastardly schemes that he is still referred to as the “Grand Crow-Demon” (Dai Tengu) or even the “Shadow Lord” (Anshu). Meanwhile, there were mixed feelings among his supporters in the skirmish. Taira no Kiyomori (1118–81), the scheming, moustachioed courtier who brokered the power behind the scenes, gained an impressive promotion and a nearby coastal fief to rule over. Minamoto no Yoshitomo, however, who had done the actual fighting in a conflict that had cost him the deaths of his own relatives—sometimes at his own hands—received much less. As far as the court was concerned, he was a loyal servant being granted some great concessions of noble rank and title. Yoshitomo felt that Kiyomori was getting the glory for his own hard work, and that once the fighting was done, the courtiers had suddenly remembered again how much they despised the samurai.

The Fujiwara, meanwhile, were up to their usual tricks, making sure that the new emperor had a Fujiwara bride. The one they found had previously been the child-bride of her new husband’s uncle, the sickly teenage emperor Konoe. Kiyomori made sure one of his own daughters was married to the new emperor’s chief minister, and, it seems, dismissed Yoshitomo’s complaints that he was not getting what he deserved.

Yoshitomo took action in January 1159, waiting until Kiyomori and his cronies were on a pilgrimage. His men snatched both Emperor Nijō and his father Go-Shirakawa, who were then obliged to sack many of their ministers and replace them with appointees favorable to the Minamoto clan.

This was by no means the first time such a power grab had occurred, but the outcome was different. It used to be that whoever had lost the upper hand would run for the provinces, to lean on their power base there. But Kiyomori had observed the fate of such former figures: absent from the capital, they had been branded by the captive administration as “rebels,” which led all loyal samurai to take arms against them. Kiyomori had seen several such examples in recent memory, and was determined not to be another one. Accordingly, instead of running for the coast of the Inland Sea, he rode straight back to Kyōto, daring his enemies to make their move.

Kiyomori and his Taira samurai were unable to act for as long as commands were issued in the name of the emperor—the confidence of the samurai had yet to achieve that arrogant tipping point whereby they acted out of regard for what the true emperor’s orders might be. Instead, the capital endured a tense ten-day standoff of messengers and conferences, with a substantial number of samurai at battle readiness. Four years earlier, the troops fielded had numbered in the hundreds; tellingly, there were now thousands ready to strike.

The impasse was broken through subterfuge. Two aristocrats switched sides and dolled up the teenage emperor Nijō in makeup and women’s clothes, sneaking him out of his palace in disguise and whisking him away to Kiyomori’s compound in the middle of the chaos caused by a convenient fire at the palace. Go-Shirakawa was even bolder, sneaking from the palace by simply dressing in commoner’s clothes and riding out the gate.



Now the power shifted once more. Emperor Nijō issued a new proclamation naming Kiyomori’s residence as the “new palace”—effectively declaring a state of emergency that implied anyone in the original palace was an imposter or rebel. A force of some 3,000 Taira cavalry, with at least as many foot soldiers in support, was already marching on the former imperial palace, leading to a running battle in the streets of Kyōto. Depending on whom one believed, either the Taira were chasing the Minamoto across town, or the Minamoto successfully pushed the Taira back to Kiyomori’s mansion. Either way, Yoshitomo’s only chance of survival was to snatch back his imperial bargaining chips before they could officially declare him to be an enemy of the state.

But it was too late. Yoshitomo had been outmaneuvered, and now it was his turn to flee, splitting up and leading a dwindling band of faithful samurai in a fighting retreat amid a driving midwinter snowstorm. Few of them lasted more than a few days, with even their allies turning against them. Yoshitomo himself was murdered while bathing at a house that he believed to have been run by friends.

Yoshitomo’s newest mistress, the twenty-year-old Tokiwa, took a different route with her three young sons, leading two by the hand with the third, a new-born baby, nestled against her chest beneath her robe. She was soon apprehended and brought before Kiyomori, who informed her that the menfolk of the Minamoto were being purged from the Earth. He did, however, have an offer for her that she could not refuse. Her three sons would be spared if she sent them away to a monastery…and agreed to become Kiyomori’s concubine.

The Taira were appalled that Kiyomori could even consider such an offer. His own stepmother warned him that Yoshitomo’s children were sure to grow up with a desire to avenge the fall of their clan. But Kiyomori was arrogant in victory, utterly convinced that he had stripped the Minamoto of all their power. Raping their leader’s woman would be the final insult.

Tokiwa would live another three decades, although Kiyomori soon tired of her; she ended her days married to a Fujiwara courtier. Kiyomori, meanwhile, achieved all his desires, and was the first samurai to be made chief minister in 1160. Not long afterward, his sister-in-law attracted the eye of Go-Shirakawa, fell pregnant, and persuaded the retired emperor that the child of their union should be the next infant sovereign requiring a regent. The boy was crowned as Emperor Takakura in 1168, and would eventually marry Kiyomori’s daughter. Kiyomori then “retired” from his official posts, enjoying the glory but rejecting the responsibilities that might actually be required to carry out those roles. What could possibly go wrong?

In fact, things had already started to go wrong, on a dark, snowy night when Emperor Takakura’s regent, Fujiwara Motofusa, found his retinue’s path blocked by a bunch of teenage samurai. The regent’s men demanded that they move, but the samurai, celebrating after a day’s hawking and hunting, told them to shove it. The regent’s men dragged them from their horses, and the lead teenager—another of Kiyomori’s grandsons—went home and whined to his dad about it.

His father, wise to court etiquette, immediately apologized to the regent, but Kiyomori had other ideas. He rounded up sixty country samurai with allegiance directly to him, and ordered them to avenge the “insult” to his grandson. They lay in wait for the regent’s entourage, ambushed their target on the road, wrecked the carriage, and cut off the hair of the captured guardsmen. The humiliated regent arrived at the palace in a cart dragged by one of his retainers, the oxen having been cut loose.

In the aftermath, the grandson was packed off to the provinces, and the perpetrators of his vengeful drive-by were dismissed. But the incident had made Kiyomori ample enemies among the Fujiwara. The fateful nuptials of Emperor Takakura and Kiyomori’s fifteen-year-old daughter were only a few days later; although they would eventually produce the Emperor Antoku in 1178, they occurred amid an atmosphere of resentment.

The child-emperor Antoku was the culmination of all Kiyomori’s scheming, and also the seed of his downfall. From being stripped of imperial status, the Taira were just about to supply Japan’s next ruler. In bundling Antoku onto the throne, Kiyomori made a permanent enemy out of Retired Emperor Go-Shirakawa, who was far too wily a politician to say anything in public. Instead, Go-Shirakawa got his own son, Prince Mochihito, to proclaim that Antoku was a usurper and that he, Mochihito, was the rightful heir, and to call on any samurai with a sense of justice to come to his aid.

In the beginning, he had few supporters. In fact, he spent the remainder of his short life on the run, protected by a small group of loyal guardsmen and “warrior monks” (some of whom were former Minamoto who had taken holy orders to avoid persecution in pro-Taira times), pursued out of Kyōto by samurai loyal to the Taira. He made it as far as the bridge at Uji, falling off his horse six times. His men pulled up the planks on the bridge to delay their pursuers, and commandeered the nearby Phoenix Hall temple to give the pretender a rest.

It was, however, a fatal delay. The pursuing enemy samurai plunged straight into the rushing waters of the river—200 men and horses were swept away by the current, but plenty made it to the opposite bank, their fellow samurai providing cover with a hail of arrows from the far side. One warrior, the fighting monk Jōmyō, did not bother with the river, but made an acrobatic barefoot assault across the scaffolding of the bridge. The Tale of the Heike reports him reaching the other side ready for action, firing off all twenty-four arrows from his personal supply (killing twelve men and wounding eleven—even a breathless story allowing for one miss). He then grabbed his spear, killing another five men before it broke. Drawing his sword, he dispatched another eight opponents before snapping his blade on the helmet of another, and dropping it into the river. Then he drew his dagger—at which point The Tale of the Heike appears to lose count. It does, however, return to Jōmyō when all the fighting is done, counting sixty-three arrow dents in his armor, five of which have pierced the leather, although none of them seriously.

The fighting spread to the Phoenix Hall, with many of the Minamoto loyalists choosing to make a last stand, dooming themselves in order to allow Prince Mochihito to escape. The Tale of the Heike offers a catalogue of last stands and acts of seppuku, although at least one samurai lived to fight another day. The warrior monk Tayū Genkaku somehow fought his way back to the bridge, jumped into the river, and hit bottom in his heavy armor before clambering out on the Kyōto side, hurling insults at his enemies, and commencing the long, damp trudge back to the capital.

But all the heroism amounted to nothing. The prince’s foster brother, trembling amid the duckweed in a roadside ditch, saw a troop of Taira samurai heading home, bearing the headless body of Prince Mochihito on a window shutter. The prince’s head, along with the heads of some 500 of his allies, was taken to Kiyomori’s mansion by the evening, where victory celebrations soon soured, as nobody could be found to make a positive identification.

Since he had been sequestered for years in a remote palace, living largely in the company of an entourage who were now dead, nobody knew what Prince Mochihito actually looked like. Tense hours passed while the Taira scoured the capital for someone who could identify him, eventually dragging in the mother of one of his children, whose distraught reaction was all that Kiyomori really needed to see.

And there things really ought to have ended, with the pretender dead—except the momentum of Mochihito’s rebellion kept on without him. Despite his protestations that Mochihito was dead, Kiyomori still had to contend with the news of Minamoto armies assembling to the east. Those three surviving Minamoto boys were now all grown up, married into Kantō plainsman aristocracy, and ready for revenge. Their cousin, too, a man called Yoshinaka, had been adopted into the Kiso clan, and hence had not shown up as a Minamoto clansmen when the purges were all the rage. Now he, too, rediscovered his Minamoto roots and came after Kiyomori.

Kiyomori did not live to see the endgame he had set in motion. Bedridden and in his sixties, he died in 1181 as Minamoto forces advanced on the capital; his grandson, the Emperor Antoku, was moved for safety’s sake to the Taira heartland on the coast of the Inland Sea.

The Minamoto flooded into the capital, where they were welcomed by the scheming retired emperor Go-Shirakawa. Although Antoku was still on the run with the sacred imperial regalia—the mirror, the sword and the jewel—the Minamoto wasted no time in proclaiming that he had abdicated, and that his half-brother, Go-Toba (1180–1239), the son of a Fujiwara mother, was the new emperor. In the battles that followed, the Minamoto would hound the Taira across the Inland Sea until their final showdown at sea at Dannoura in 1185.

Realizing that all was lost, the last of the Taira began jumping into the sea, their armor dragging them straight to the bottom. Kiyomori’s widow, Tokiko, turned to her grandson, the six-year-old Emperor Antoku, and told him to say prayers to the east, toward the Shintō shrine at Ise, and west toward the homeland of Buddha.

“Beneath the waves lies our capital,” she said. Then, hugging Antoku close to her along with the ancient sword Kusanagi, she hurled herself into the sea.

The conflict between the Taira and the Minamoto was finally resolved, with the Taira almost entirely wiped out and excluded from the capital. Scattered survivors, including Antoku’s mother, who was pulled by her hair from the water by sailors using a rake, would live on as impoverished local fishermen or religious devotees. The sacred mirror and jewel were, at least officially, retrieved by divers, although the sword Kusanagi was never found—Japanese authorities are deliberately vague about it; although a sword still forms part of the imperial regalia of Japan, the one carried most recently during the coronation of Emperor Heisei in 1989 is believed to be a replica.

The Minamoto were victorious…but—as with every other event in The Tale of the Heike, as foreshadowed by its opening lines—it was all for nothing. In the aftermath, the Minamoto turned on each other, as the eldest surviving son of Yoshitomo, Yoritomo, unleashed his simmering resentment against his half-brother Yoshitsune, who had been instrumental in many of the Minamoto victories of the war with the Taira. Yoritomo largely stroked his chin and looked at maps in his distant headquarters at Kamakura, a fort chosen for strategic reasons—it was approached by seven roads, every one of them traversing steep, defensible mountain passes. But it was Yoshitsune who was on the front line—often against the wishes of his fellow Minamoto generals, but winning forever the support of his men.

Yoshitsune is another of the iconic figures of Japanese history whose life story has lent itself well to legend. From his first appearance in Japanese stories (and indeed, in this book), tucked into his mother’s robe as she flees in a snowstorm, to his legendary tutelage at the feet of crow demons in the hills outside Kyōto, he has gained an enduring presence in Japanese plays, books, and movies. It is Yoshitsune, so the legend goes, who bested the warrior-monk Benkei on Kyōto’s Gojō Bridge; who seduced a nobleman’s daughter so he could read her father’s copy of an ancient Chinese military manual; who led a foolhardy cavalry charge down a steep cliff, surprising the enemy by hitting them from behind their camp at Ichinotani. It was Yoshitsune who lit fires on the landward side of the Taira base, frightening his enemies into taking to their ships and thereby setting up the ultimate showdown at Dannoura.

Yoritomo hated that his half-brother was getting all the credit. He seemed to find fault in every one of Yoshitsune’s victories, criticizing him for minor details like escaped prisoners, rather than praising him for his incredibly effective strategies. Yoshitsune even managed to charm Go-Shirakawa, the retired emperor, although Yoritomo regarded that as yet another example of scheming. On the apparent belief that his half-brother was planning to betray him, Yoritomo ordered his arrest, ending the war with a tragic coda in which the greatest Minamoto general became a fugitive in the north, fleeing his own family.

Yoshitsune the loyal lieutenant was eventually hunted down and killed, his henchmen and son murdered, all so Yoritomo could feel secure. “Sympathy for the lieutenant” (hōgan biiki) remains a popular term in Japanese for championing the underdog. Yoritomo was left with a large holding of his own lands along with the lands of Minamoto vassals and over 500 estates taken from the defeated Taira. It made him a substantial rival for the imperial court itself, which for its own part now lacked any military allies on which it could call for assistance.

With the Minamoto now dominating the court, and with the death of the manipulative retired emperor Go-Shirakawa in 1192, his grandson the child-emperor Go-Toba was persuaded to recognize the possibility of another war breaking out against unknown enemies of the state, and appointed Yoritomo as Shōgun. Despite the continued use of the archaic title for “suppressing barbarians,” Yoritomo was more of a government-appointed autocrat, running Japan in the emperor’s name under a state of martial law. The term he used, which would be used by his successors for the next seven centuries, was intended to imply that this situation was merely a temporary fix until the trouble had died down: the authorities came to be known as the bakufu, or “tent government,” taking their name from the baku windbreaker behind which samurai generals would hide from enemy archers while plotting their next move.

You would be forgiven for thinking that it was a happy ending for the Minamoto, but they had taken heavy losses in the war, not helped at all by Yoritomo’s paranoid postwar purges. Ruling Japan from Kamakura, Yoritomo became the first leader of the Kamakura shōgunate, which would technically run Japan in the emperor’s name for the next 200 years—except much of his military success had been funded by his father-in-law, Tokimasa, leader of Kamakura’s Hōjō clan. After Yoritomo’s death in 1199, his sons were swiftly elbowed aside in favor of “regents” (shikken) from the Hōjō clan. It was these regents who held the true power of the Kamakura shōgunate thereafter, while the Minamoto disappeared in a bout of stabbings and assassinations—Yoritomo’s son, the shōgun Sanetomo, was assassinated by his own nephew, who was then executed for murder, bringing the line to an end while the ghosts of the Taira laughed on the bottom of the sea.

Exactly what kind of unrest was the Kamakura shōgunate expecting? The biggest problems they might expect to encounter often seemed to come from the imperial family itself, whose members did not take kindly to being the puppets of their leading general. Crowned during the conflict as a three-year-old child by the Minamoto in 1183, Japan’s eighty-second emperor, Go-Toba, was forced to abdicate in 1198, but remained inconveniently alive for the next forty-one years, watching from the sidelines as his sons were pushed onto the throne and then off again in the service of the shōgunate’s power games.

In 1221, Go-Toba made his move. Without waiting for the shōgunate to recommend its own candidate, he put his two-year-old grandson on the throne. He then invited all the important samurai in the vicinity of Kyōto to a celebration.

It was a brilliant strategic move. Those who accepted his invitation were clearly willing to support him in any further resistance to the shōgun. One prominent lord did not show up, and soon died under suspicious circumstances—by implying even for a moment that he disapproved of Go-Toba’s actions, he had signed his own death warrant. The others were ready to hear Go-Toba’s new proclamation in the style of his ill-fated ancestors: that anyone who was truly loyal to the emperor and the court should rise up against the Kamakura usurpers. The Hōjō clansmen were officially declared outlaws, and disaffected samurai in the Kyōto region began to flock to Go-Toba’s banner.

Well, maybe not “flock.” Go-Toba attracted a few followers, but the bulk of Japan’s samurai were persuaded to support the so-called outlaws. Hōjō Masako, Yoritomo’s widow, rallied the troops by reminding them of the improvements they had enjoyed under the bakufu. She proclaimed that this was a crucial turning point in history, where the samurai could choose either to remain masters of their own destiny, or to return to the days when they were mere patsies for the court. She must have struck a strange figure addressing the samurai—her head was shaved, as was the custom for widows, leading to her nickname among the samurai: ama-shōgun, the Nun Shōgun.

A Kamakura army marched on Kyōto, scoring a string of successes against the lesser numbers of Go-Toba’s followers. Go-Toba went to the fighting monks on nearby Mount Hiei, pleading for them to come to his aid, but they refused, unwilling to take on the shōgunal forces. The imperial forces made their last stand at the bridge over the river at Uji before giving up and fleeing. Kamakura forces occupied Kyōto, and Go-Toba and his “retired” sons were exiled to remote islands. The grandson became known as the “Dethroned Emperor,” having ruled for barely two months, the shortest reign in Japanese history; he was not even recognized as an emperor at all until the nineteenth century.

The defeat of Go-Toba’s attempted restoration played into the shōgunate’s hands, allowing for the confiscation of some 3,000 estates that could be used to buy favor with the samurai faithful. It secured the shōgunate two generations of relatively stable rule until the 1270s, when the conquest of China by the Mongols led to the threat of an invasion by Khubilai Khan.

Japan Approves Plans to Convert Izumo-Class Into F-35-Carrying Aircraft Carriers

JS Kaga (DDH 184)

JS Izumo (DDH-183)

Japan’s two largest warships will be converted into aircraft carriers over the next five years.

By Franz-Stefan Gady

The cabinet of Japanese Prime Minister Shinzo Abe officially approved on December 18 plans to convert the Japan Maritime Self-Defense Force’s (JMSDF) two largest warships — the helicopter destroyers JS Izumo, lead ship of the Izumo-class, and its sister ship the JS Kaga — into aircraft carriers capable of launching the F-35B, the U.S. Marine Corps variant of F-35 Joint Strike Fighter capable of vertical or short takeoffs and vertical landings (STOVL) without requiring a catapult launcher.

Japan’s new National Defense Program Guidelines (NDPG), which lay out the Japan Self Defense Force (JSDF) capability targets and needs for the next decade, call for the first deployment of Japanese aircraft carriers since the end of the Second World War. According to the fiscal 2019-2023 midterm defense buildup program, the conversion of the two 27,000-ton  helicopter destroyers is expected to take place over the next five years.

During the same time period, Japan is also slated to receive its first batch of 18 F-35B fighter jets capable of short takeoffs and vertical landings. Japan is set to procure 63 extra F-35As — the aircraft’s conventional takeoff and landing variant –and 42 F-35Bs, the U.S. Marine Corps’ short take-off and vertical landing (STOVL) variant of the aircraft, in addition to 42 F-35As already on order over the next decade. Japan currently has a total of 10 F-35As operational deployed.

Notably, the NDPG re-designates the Izumo-class as multi-purpose escort destroyers to comply with Japan’s pacifist constitution that limits JSDF capabilities to self-defense. “The planned modification to the Izumo-class carriers is to increase their applications,” Japan’s Minister of Defense, Takeshi Iwaya, said last week, downplaying the carrier’s offensive capabilities once retrofitted.

As I reported in May, a study released by the Izumo-class builder, Marine United Corporation (MUC), earlier this year concluded that the warships can be converted into full-fledged aircraft carriers with some modifications. Indeed, as I pointed out elsewhere, while large sections of the study remain classified, the Izumo carrier class has reportedly been designed to operate STOVL fighter jets right from the inception of the program and requires only minor modifications to accommodate the F-35B.

    [A] consensus was privately reached at the inception of the Izumo project that the warships should be designed allowing for a future conversion into a F-35B-carrying naval platform. (…)

    Modifications on the Izumo-class, among other things, would require the installation of a ski-jump. Contrary to earlier reports, the JMSDF executive claims that the flight deck has already been coated with paint that can withstand the exhaust heat generated during F-35B landings and takeoffs. Additionally, the aircraft elevators, connecting the flight deck with the hangar, were reportedly specifically designed to accommodate the F-35B.

Following modifications, the carriers would each only be capable of accommodating around a dozen F-35Bs.

Both the JS Izumo and JS Kaga have been principally designed for anti-submarine warfare since entering service over the past three years. “Its usefulness as an [anti-submarine warfare] ASW platform would be substantially reduced following the ship’s conversion as it would no longer be capable of carrying its full contingent of Mitsubishi-built SH-60k ASW helicopters and other ASW-related equipment,” I wrote last week.

Nakajima Ki 84 Hayate ‘Frank’ Part I


The Nakajima Ki-84 Hayate (Gale) was numerically the most important fighter serving with the Japanese Army Air Force (JAAF) during the last year of the Pacific War, and was probably the best Japanese fighter aircraft to see large-scale operation during this period of the war. The Hayate was fully the equal of even the most advanced Allied fighters which opposed it, and was often their superior in many important respects. It was well armed and armoured, was fast, and was very manoeuvrable. Although it was generally outnumbered by Allied fighters which opposed it, it nevertheless gave a good account of itself in battles over the Philippines, over Okinawa, and over the Japanese home islands. So desperate was the need for Ki-84s in the last months of the war, Japan was building underground factories with a planned rate of 200 aircraft per month.

The history of the Ki-84 can be traced back to just after the beginning of the Pacific War between Japan and the United States. Just three weeks after Pearl Harbour, the Koku Hombu instructed the Nakajima Hikoki K.K. (Nakajima Aeroplane Co Ltd) to begin the design of a replacement for the Ki-43 Hayabusa, which itself had just entered service with the JAAF. The JAAF wanted a general-purpose long-range fighter that would be superior to those that were then under development in the USA and Britain. The specification called for an aircraft with the manoeuvrability of the Ki-43 Hayabusa coupled with the speed and climb of the Ki-44 Shoki. In addition, the aircraft was to be provided with armour protection for the pilot and was to be fitted with self-sealing fuel tanks.

The aircraft was to have a maximum speed of 398-423 mph, and was to be capable of operating at combat rating for 1.5 hours at distances as far as 250 miles from its base. The wing loading was not to exceed 35 pounds per square foot. The manoeuvrability requirements were relaxed somewhat as compared to those of the Ki-43, but were to exceed those of the Ki-44 which had been designed strictly as a bomber destroyer. The engine was to be the Nakajima Ha-45 eighteen-cylinder double-row air-cooled radial. The armament was to be two 12.7 mm Type 1 (Ho-103) machine guns and two 20-mm Ho-5 cannon.

Koyama was named as the project engineer, and work on the Ki-84 began in early 1942 at Nakajima’s Ota plant in Gumma Prefecture. The first prototype was ready in March of 1943. The aircraft was a fairly conventional low-wing monoplane that bore an obvious family resemblance to the Ki-43 and Ki-44 fighters that preceded it. The 1800-hp Nakajima Ha-45 engine that powered the Ki-84 was a JAAF version of the Navy’s NK9A Homare. Experimental models of the Homare engine had been test run as early as May of 1942, but the development of the Homare was long and difficult, and few Homares were available until August of 1943, and experimental production did not begin until late 1943 at Najajima’s Musashi engine factory.

A big exhaust collector pipe was mounted on each side of the engine behind the cowling gills. The all-metal airframe followed the common Japanese practice of building the wing integral with the central fuselage in order to save the weight of heavy attachment points. The fuselage was of oval section, with flush-riveted stressed skin. The two-spar wing carried metal-framed, fabric-covered ailerons and was provided with hydraulically-operated Fowler flaps. A total of 220 US gallons of fuel was carried in tanks aft of the cockpit and in the wings. The engine mounting and cowling incorporated the oil cooler and intakes for the carburettor and supercharger. The three-part canopy had an aft-sliding central section. All three undercarriage members were hydraulically retractable. The main gear retracted inward and horizontally into the wings and was fully covered with flush-fitting doors. The non-steerable tailwheel retracted into the fuselage and was covered by a flush-fitting door. The rudder was of metal construction but was covered with fabric.

The tailplane was set well ahead of the vertical surfaces. Two 12.7 mm Ho-103 machine guns with 250 rpg were mounted in the upper cowling, and a 20-mm Ho-5 cannon with 150 rounds was mounted in each wing outboard of the main undercarriage leg. The pilot was protected by a 70-mm armoured windshield and by 13-mm armour plate in the rear and floor of the cockpit. Provision was made under the fuselage centreline for a single drop tank.

The Ki-84 prototype flew for the first time from Ojima Airfield in April of 1943. The second prototype flew in June. The first prototypes were assigned to the JAAF for trials at the Tachikawa Air Arsenal under the direction of combat-experienced pilots, and the modifications recommended were incorporated into the fourth prototype. The fourth prototype had a maximum speed of 394 mph at 21,800 feet, and could achieve a speed of 496 mph in a dive.

The test program went well, and a service trials batch of 83 machines were ordered in August of 1943. These were built between August of 1943 and March of 1944. The pre-production machines differed from each other in minor details, but fuselage changes were incorporated to ease production, and the area of the fin and rudder was increased to improve control on takeoff.

A few service trials machines were handed over to the Tachikawa Army Air Arsenal. JAAF pilots commented favourably on the machine, although its maximum speed was below the requirement. The aircraft had a maximum speed was 388 mph, could climb to 16,405 feet in 6 minutes 26 seconds, and had a service ceiling of 40,680. This made the Ki-84 the best-performing Japanese fighter aircraft then available for immediate production.

A few service-test Ki-84s were fitted experimentally with a ski undercarriage. The legs retracted into the normal wheel wells, with the skis lying flat underneath the wing roots. These aircraft were tested in Hokkaido during the winter of 1943-44. The ski installation increased the maximum weight, and thus had an adverse effect on maneuverability and reduced the maximum speed by 8 mph. Consequently, skis were not incorporated on production machines.

The Ha-45 engine entered full-scale production in April of 1944 as the Type 4. Production of the Type 4 engine was hampered by many setbacks, most of which were due to inadequate preparation, with shortages of jigs, tools, and skilled personnel being significant problems.

Service tests of the Ki-84 began in Japan under operational conditions in October 1943. The type was accepted for production as the Army Type 4 Fighter Model 1A Hayate (Gale) or Ki-84-Ia.

A second pre-production batch of 42 Ki-84s was started in April of 1944. These were built between March and June of 1944. These were built in parallel with the first production aircraft, which began to roll off the production lines in April of 1944. Both types were fitted with individual exhaust stacks, which provided some thrust augmentation, and could increase the maximum speed by some 9-10 mph.

Each of the two wing racks could carry a 44 Imp gall drop tank or a 551-pound bomb. Some of the aircraft of the second service test batch were tested with wings of increased span and area to serve as development aircraft for the projected Ki-84N and Ki-84P projects.

Early production machines had the 11 and 12 models of the Ha-45 engine, with takeoff ratings of 1800 hp and 1825 hp respectively. Later models had the model 21 version of this engine, delivering 1990 hp for takeoff. These engines were rather unreliable and were subject to numerous quirks.

Sudden loss of fuel pressure was a constant source of difficulty, and this was addressed by the adoption of the Army Type 4 radial Model 23 ([Ha-45]23) for even later production machines. This Model 23 engine was a modification of the Model 21 engine fitted with a low-pressure fuel injection system.

The Ki-84-Ia was followed on the production line by the Ki-84-Ib Army Type 4 Fighter Model Ib. In the Ki-84-Ib, the fuselage-mounted machine guns were replaced by a pair of 20-mm Ho-5 cannon, giving the aircraft a total armament of four 20-mm cannon.

The Ki-84-Ic was a specialized bomber destroyer variant armed with two 20-mm Ho-5 cannon in the fuselage and two wing-mounted 30-mm Ho-105 cannon. Only a small number of this version were built.

In March of 1944, the experimental squadron that was conducting the service test trials of the Ki-84 was disbanded, and its personnel transferred to the 22nd Sentai. This unit was re-equipped with production Hayates and transferred to China where it entered combat against the USAAF’s Fourteenth Air Force in August of 1944. The Ki-84-Ia quickly established itself as a formidable foe that compared favourably with the best Allied fighters then available. The Hayate had an excellent performance and climb rate, and had none of the shortcomings of the earlier generation of Japanese fighters, being well armed and possessing adequate armour protection for the pilot. In addition to the penetration and interception roles, the aircraft was used as a fighter-bomber and dive bomber. The 22nd Sentai was later moved to the Philippines, where it was joined by the 1st, 11th, 21st, 51st, 52nd, 55th, 200th, and 246th Sentais.

Following encounters with the Ki-84-Ia, the Allied Technical Air Intelligence Unit (ATAIU), commanded by Colonel Frank McCoy, assigned the code name FRANK to this fighter. This code name had previously been assigned to a fictitious aircraft known as the “Mitsubishi T.K.4”, which was erroneously believed to be under development in Japan. When the T.K.4 failed to materialize, Colonel McCoy decided to name the new Ki-84-Ia after himself.

The FRANK later appeared in the battle for Okinawa, serving with the 101st, 102nd, and 103rd Hiko Sentais. Two new Sentais, the 111th and the 200th were activated with Hayates. The Hayates were used for long-range penetration missions, fighter sweeps, strafing, interception and dive-bombing missions with considerable success. The Ki-84 proved faster than the P-51D Mustang and the P-47D Thunderbolt at all but the highest altitudes. At medium altitudes, the FRANK was so fast that it was essentially immune from interception. The climb rate was exceptionally good, 16,400 feet being attained in 5 minutes 54 seconds, which was superior to that of any opposing Allied fighters.

The Ki-84 had a close resemblance to the Ki-43 Hayabusa, which caused many Allied fighter pilots to confuse it with the earlier Nakajima fighter during the stress of combat. Many an American pilot, having sighted a Japanese fighter he believed to be a Ki-43 and salivating at the prospect of a quick and easy kill, suddenly found he had latched onto a different bird entirely. The Ki-84 even did well at the fighter-bomber role. On April 15, 1945, a flight of eleven Hayates from the 100th Sentai made a surprise air attack on American airfields on Okinawa, damaging or destroying a substantial number of aircraft on the ground. However, eight of the Hayates were destroyed in the attack, and one made a forced landing on a small islet near Kyushu.

Although the Ki-84 was intended for the offensive, penetration role, Hayates were assigned to the defensive role over the Japanese home islands during the last few weeks of the war, operating with the 10th Division responsible for the defence of Tokyo. The units assigned to home defence included the 47th, the 73rd, the 111th, and the 112th and the 246th Sentais. Since the Hayate was regarded as being essential for the interception role, relatively few were expended in Kamikaze attacks.

The Hayate was simple to fly, and pilots with only minimal training could fly the type with relatively little difficulty. However, the aircraft did have have certain poor control characteristics to which a veteran pilot could easily become become accustomed but which could be deadly in the hands of an inexperienced pilot. Taxiing and ground handling were generally rather poor. On takeoff, once the tail came up, continual pressure had to be maintained on the starboard rudder pedal to counteract a tendency to swing to port caused by the high engine torque. In flight, the controls were sluggish in comparison with those of the Hayabusa, and the elevators tended to be heavy at all speeds. The ailerons were excellent up to about 300 mph, after which they became rather heavy. The rudder was mushy at low speeds for angles near neutral.

However, most of the defects with the Ki-84 can be laid to poor quality control during manufacture, especially during the last few months of the Pacific war. When the Ki-84 was being designed, emphasis had been placed on ease of production, and the manufacture of the Ki-84 required less than half the tooling needed by the Ki-43 and Ki-44 which preceded it. However, many experienced workers had been drafted into the military, and this loss, acting in concert with the accelerated rates of production ordered by the Japanese Ministry of Munitions, resulted in a steady drop in quality standards of both the engine and the airframe of the Hayate as the war progressed.

The performance and reliability of production Hayates was seldom as good as that of the service test machines. As the quality of the workmanship steadily deteriorated, the performance of the Hayate steadily declined as production progressed, with later machines having successively poor and poorer performance and mechanical reliability. The hydraulic and fuel pressure systems were both poorly designed and were subject to frequent failures. The wheel brakes were notoriously unreliable, and the metal of the landing gear struts was often inadequately hardened during manufacture, which made them likely to snap at any time. This caused many Hayates to be written off in landing accidents, without ever having been damaged in combat.

In Richard M. Bueschel’s series on Japanese WWII fighters (i.e, “Nakajima Ki-84 Hayate in Japanese Army Air Force Service) and Rene Francillon’s “Japanese Aircraft of the Pacific War”), the authors repeatedly mention that the latter-day Japanese fighters, while technically equal to their Allied counterparts, were plagued by variable and poor workmanship; Bueschel writes that the Ki-84 could only rarely achieve a level speed of 400 km per hour. For example, the level speed of the Nakajima Ki-84 fighter was rated at 620 km per hour, yet reportedly most late-production specimens could not achieve even 400 km per hour due to poor workmanship of the airframe and engine.

I recall reading accounts of trials of captured Japanese aircraft, in particular the Ki-84, which stated that the fitting of US spark plugs made a vast difference to the aircraft’s performance. The fitting of US brake pads too, where possible was found to be a must too! Material shortages and quality problems in seemingly minor areas can have a dramatic effect on overall performance.

Engine shortages and delays were a constant problem for the Hayate. Although the Ha-45 engine had been plagued with production difficulties all throughout its life, most of the delays in deliveries were caused by frequent visits of 20th Air Force B-29s to the Musashi engine plant during the last year of the war. This plant was hit by B-29 raids on no less than twelve occasions between November 24, 1944 and August 8, 1945. Production was able to continue at the Musashi plant until April 20, 1945, when it was finally put out of business for good and all production came to a standstill.

Operations were transferred to an underground plant at Asakawa. and to a new plant at Hamamatsu, and a trickle of engines still continued to flow, but the supply of engines never reached the previous peak. Because of the production delays and components shortages, the quality of the Ha-45 engines delivered steadily deteriorated as the months passed, and later engines were considerably less powerful and less reliable than those initially delivered. By June of 1945, the lowering of manufacturing standards had cut the climb rate of the fighter so severely that the aircraft was virtually useless at altitudes over 30,000 feet.

A total of 1670 Hayates were built during 1944, making the aircraft numerically the most important Japanese fighter in production at that time. However, this was still far below JAAF requirements. Orders for 1944 alone totalled 2525 machines, almost a thousand more than were actually delivered. This shortfall was partly a result of the failure of subcontractors to deliver components on schedule, but became increasingly caused by Allied air attacks on Japanese industry as 1944 neared its end. On February 19. 1945, Nakajima’s Ota plant was attacked by 84 B-29s, which seriously damaged the plant and destroyed some 74 Hayates on the assembly line. Further attacks on the plant by US carrier-based aircraft further damaged the plant to such an extent that an extensive dispersal program had to be carried out, with an accompanied sharp drop in production.

In May of 1944, Nakajima opened up a second Hayate manufacturing line at its Utsonomiya plant. This facility had built 727 fighters by July of 1945, less than half the number scheduled during this period. Construction of the Hayate was also assigned to the Mansyu Hikoki Seizo K.K. (Manchurian Aircraft Manufacturing Company), which started production in the spring of 1945 at its Harbin plant in Manchuria. However, only a hundred or so Hayates were built at Harbin before the end of the war brought production to an abrupt end. Total production of the Hayate by all factories was 3514, including prototypes and service trials aircraft.

In 1946, a captured late-production Hayate was restored and tested at the Middletown Air Depot in Pennsylvania. At a weight of 7490 pounds, the aircraft achieved a maximum speed of 427 mph at 20,000 feet, using war emergency power. This speed exceeded that of the P-51D Mustand and the P-47D at that altitude by 2 mph and 22 mph respectively. These figures were achieved with a superbly maintained and restored aircraft and with highly-refined aviation gasoline, and were not typical of Japanese-operated aircraft during the later stages of the war.

Nakajima Ki 84 Hayate ‘Frank’ Part II

Hayate and Fuel Grade

After testing by US air force, the Hayate was owned by a collector. After that, the owner presented it to FUJI HEAVY INDUSTRY which used to be ‘NAKAJIMA AIR PLANE”. At that time, the Hayate was full flying condition. However, one day FUJI H.I. gave up on maintaining it and sold the Hayate to a museum. It was the start of this sad story. The new owner set it outside and the rain came. The Hayate could not be flown under this condition. After that, the museum sold the Hayate to another muesum. Then, the main undercarriage was cut for moving it to a new museum. The sad Hayate has been exhibiting in Chiran museum in Kagoshima Japan now. http://www.d1.dion.ne.jp/~r_dom/haya_index.htm

Allow me to quote from a book published by Dai Nippon Press, detailing the development of the Nakajima Ki-84 Hayate, code named the ‘Frank’ by the Allies:

“The following spring, another captured Hayate (believed ultimately scrapped) went through more rigorous testing at the Middletown Air Depot in Pennsylvania. Taking place in April and May of 1946, to their great surprise, the Americans found that in many performance categories, the Ki-84 was the equal of, or superior to, the P-47N and P-51K to which it was compared. Roughly the same in maximum speed at altitude, the Hayate proved superior in early climbing performance, maneuverability, and overall handling. On the down side, the testers noted the aircraft’s relatively short range, and poor protection, as well as shoddy materials and workmanship in some components, especially the exhaust stacks. It must be noted that during these tests, the plane had an advantage it never had during the actual war: high grade gasoline. 87-octane was the best the Japanese forces could provide their planes during the conflict, and although the ‘official’ data on the Ki-84 indicated a top speed of 624 kph, in practice the aircraft could only reach 580-590 kph. However, with 140-octane gas in its tank, the normally troublesome Ha-45 [the engine] was able to pull the Hayate through the skies at a top speed of 689 kph in level flight at 20,000 feet, a stunning 60 kph more than the Japanese were ever able to produce, and just slightly faster than the P-51D-25 and P-47D-35 at the same altitude.”

I’d guess the Japanese were even worse off with regard to high octane fuels than the Germans were. Although they had access to a considerable supply of high grade crude (at least until US submarines cut the shipping lines) I don’t believe they had sufficient processing capability to take full advantage of it.

In a confidential report listing titled COMPARATIVE PERFORMANCE AND CHARACTERISTICS REPRESENTITIVE ENEMY AND ALLIED AIRCRAFT the FRANK 1, Nakajima is listed as follows:

Engine: Nakajima Ha.45: 1970hp/S.L. 1695hp/21,000ft. Armament: 4x20mm. Range: 1795mls/156mph/359gallons of fuel.

Performance at the test weight of 7940lbs.: Climb: 3780fpm/S.L. 3290fpm/21,000ft. 10,000ft/2.7min. 20,000ft/5.8min. Service Ceiling: 39,000ft. Maximum Speed: 348mph/S.L. 422mph/21,000ft.

Due to all the problems it encountered in production, but taking into account the ingenuity of some chief mechanics in the field, IMHO it was likely the Ki.84 the Allied pilot would meet in combat would be capable of 325-350mph/S.L. and 384-425mph/20,000ft. Just an opinion though. I have read that the performance of the Ki.84 varied greatly in the field during 1945.

The Ki.84 is the most controversial Japanese fighter to evolve from WW2. It definitely had the potential to be a world beating contender if it hadn’t been for its constant failure of exhaust stacks due to poor materials, inefficient welding and problems with the hydraulics.

According to Wright Field Report No. F-1IM-1119B-ND released in January 1947, a Ki.84-1 with the Ha.45 Model 21 engine, Serial No. 302, performance was found to be as follows: Speed: 350mph/S.L. 389/10,000ft. 412/20,000ft. 426/23,000ft. 400/30,000ft. 370/35,000ft.

This was listed under Normal Fighter, Military Power. Climb Normal Fighter: 3790fpm/S.L. and 3195fpm/20,900ft under Military Power. 3615fpm/17,900ft under W.E.P. Service Ceiling: 38,800ft. No test weight or engine power settings given.

On another official PERFORMANCE AND CHARACTORISTIC sheet 156A-2 the following performance is listed: Engine at T.O. and WEP: 1970hp/S.L. 2040hp/3000ft. 1850hp/17,900ft. The following performance is at 7,940lbs.: Speed: 369mph/S.L. 427mph/20,000ft. Climb: 4275fpm/S.L. 3615fpm/17,900ft. Maximum Range: 1,815mls/173mph/1500ft./359gallons of fuel.

These test figures are under ideal situations and using USA high grade fuel. However, quoting Richard Dunn: “I have a copy of a translation of a captured document of unknown reliability which indicates two maximum speeds for the Ki.84-1 (light) and Ki.84-1 (improved versions). The two speeds apparently relate to the Ha 45 rating of 2000hp at 4,020ft and 1800hp. at 19,680ft: Ki.84-1 (light): 412mph and 430mph respectively. Ki.84-1 (improved): 409mph and 427mph respectively. The light version is listed at 7885lbs. and the improved is listed at 8507lbs.

Ever since the Frank and Mustang first met in China the dogfights that ensued were reputed to be among the most ferocious of WW2!

The Ki 84 potentially had everything a balanced air superiority fighter needed except reliability. It was fast, far ranging, light on its feet, fast climbing, armored and hard hitting, and in strong numbers. Only lacking high altitude performance to the degree of the USAAF or RAF. However it had the agility to turn inside all of them. It’s like a Bf 109K or Yak-3 but with long range. Terminal dive was good at 495 mph. Initial climb was in the 4,000ft/min class and level speed was better than any mass produced fighter from Japan at 427 mph (with high octane fuel). But as the war was ending, factory quality was fading – even muzzle velocity of its fast (850 rpm unsynchronized!) world beating 20mm cannons decreased.

Specification of Nakajima Ki-84-1a:

Engine: One Army Type 4 eighteen-cylinder air-cooled radial (Nakajima Ha-45). The following engine models were used: [Ha-45]11 rated at 1800 hp for takeoff and 1650 hp at 6560 feet. [Ha-45]12 rated at 1825 hp for takeoff and 1670 hp at 7875 feet. [Ha-45]21 rated at 1990 hp for takeoff and 1850 hp at 5740 feet. [Ha-45]23 rated at 1900 hp for takeoff and 1670 hp at 4725 feet.

Performance (early production): Maximum speed 392 mph at 20,080 feet, cruising speed 277 mph. An altitude of 16,405 feet could be reached in 5 minutes 54 seconds. An altitude of 26,240 feet could be attained in 11 minutes 40 seconds. Service ceiling 34,450 feet. Normal range 1053 miles, maximum range 1347 miles.

Weights: 5864 pounds empty, 7955 pounds loaded, 8576 pounds maximum. Dimensions: Wingspan 36 feet 10 7/16 inches, length 32 feet 6 9/16 inches, height 11 feet 1 1/4 inches, wing area 226.04 square feet. Armament: Two fuselage mounted 12.7-mm Type 1 (Ho-103) machine guns and two wing-mounted 20-mm Ho-5 cannon (Ki-84-Ia). Two fuselage-mounted 20-mm Ho-5 cannon and two wing-mounted 20-mm Ho-5 cannon (Ki-84-Ib). Two fuselage-mounted 20-mm Ho-5 cannon and two wing-mounted 30-mm Ho-105 cannon (Ki-84-Ic). External stores included two 551-pound bombs or two 44-Imp gall drop tanks.

Ki-84-II Hayate Kai

The Ki-84-II or Hayate Kai was an attempt to conserve valuable supplies of aluminum by employing large numbers of wooden components in the manufacture of the Hayate. The rear fuselage, certain fittings, and modified wingtips were made of wood, with all the wood work being carried out at a shadow factory at Tanuma. The engine was the Nakajama [Ha-45] 21, 25 or 23 with low-pressure fuel injection. Armament consisted of four 20-mm or two 20-mm and two 30-mm cannon. The designation Ki-84-II was actually a Nakajima designation, the aircraft in JAAF service retaining the Ki-84-Ib or -Ic designation, depending on armament.

Specification of Ki-84-II:

Engine: One Army Type 4 eighteen-cylinder air-cooled radial (Nakajima Ha-45). The following engine models were used: [Ha-45]21 rated at 1990 hp for takeoff and 1850 hp at 5740 feet. [Ha-45]23 rated at 1900 hp for takeoff and 1670 hp at 4725 feet. [Ha-45]25 rated at 2000 hp for takeoff and 1700 hp at 19,685 feet.

Performance: Maximum speed 416 mph

Weights: 8495 pounds loaded.

Dimensions: Wingspan 36 feet 10 7/16 inches, length 32 feet 6 9/16 inches, height 11 feet 1 1/4 inches, wing area 226.04 square feet. Armament: Two fuselage-mounted 20-mm Ho-5 cannon and two wing-mounted 20-mm Ho-5 cannon. Alternatively, the two wing-mounted 20-mm Ho-5 cannon could be replaced by two 30-mm Ho-105 cannon. External stores included two 551-pound bombs or two 44-Imp gall drop tanks.


The Ki-84-III was a high-altitude version of the Hayate powered by a Ha-45 Ru engine with a turbosupercharger in the fuselage belly. This version was still on the drawing board when the war ended.


The Ki-106 was an all-wood version of the Ki-84 Hayate designed by Tachikawa Hikoki K.K., with the goal of achieving further savings of aluminium. Three airframes were built for Tachikawa by Ohji Koku K.K. (Prince Aircraft Co, Ltd) at Ebetsu, in Ishikari Prefecture on Hokkaido. The use of wood made it possible to employ lots of unskilled labour in the manufacture of the airframe. The Ki-106 was powered by a 1990 hp Nakajima [Ha-45] 21. The Ki-106 retained the external configuration of the Hayate, but the vertical surfaces had increased area and the skin was of plywood with a thick lacquer coating. The armament was four 20-mm cannon on the first Ki-106, but was reduced to only two cannon on the second and third prototypes to save weight.

Flight tests started in July of 1945. The use of wood rather than metal had raised the normal loaded weight to 8958 pounds (an increase of some 600 pounds), and this had an adverse effect on climb rate and manoeuvrability. An altitude of 26,240 feet could be attained in 13 minutes 5 seconds, this being nearly a minute and a half greater than that for the standard Hayate. However, because of the aircraft’s exceptionally fine finish, the maximum speed of 384 mph at 24,000 feet compared closely to that of the standard metal Hayate.

During trials with the first prototype, the plywood skinning failed during a test flight and began to rip away. The aircraft managed to land safely, and steps were taken in order to anchor the skin more firmly to the airframe. Although the flight tests were quite satisfactory, the end of the war brought the Ki-106 project to an abrupt halt.

Specification of Nakajima Ki-106:

Engine: One Army Type 4 eighteen-cylinder air-cooled radial (Nakajima Ha-45/21) rated at 1990 hp for takeoff and 1850 hp at 5740 feet. Performance: Maximum speed 385 mph at 21,080 feet. Cruising speed 310 mph at 20,100 feet. An altitude of 16,405 feet could be reached in 5 minutes. Service ceiling 36,090 feet. Normal range 497 miles plus 1.5 hours of combat.

Weights: 6499 pounds empty, 8598 pounds loaded.

Dimensions: Wingspan 36 feet 10 7/16 inches, length 32 feet 7 3/4 inches, height 11 feet 9 5/16inches, wing area 226.04 square feet. Armament: Two fuselage-mounted 20-mm Ho-5 cannon and two wing-mounted 20-mm Ho-5 cannon. External stores included two 551-pound bombs or two 44-Imp gall drop tanks.


The Ki-113 was a version of the Ki-84-Ib partially built of steel. It was an attempt to conserve light alloys by using steel in place of aluminium in as many sub-assemblies as possible. It employed steel sheet skinning and the cockpit section, ribs, and bulkheads were made of carbon steel. The aircraft retained the Ha-45 Model 21 engine and had an armament of four 20-mm cannon.

The Ki-113 was designed in the autumn of 1944, and a single example was completed in early 1945. However, it never flew since it was decidedly overweight.

Specification of Nakajima Ki-113:

Engine: One Army Type 4 eighteen-cylinder air-cooled radial (Nakajima Ha-45/21) rated at 1990 hp for takeoff and 1850 hp at 5740 feet. Performance (estimated): Maximum speed 385 mph at 21,325 feet. An altitude of 16,405 feet could be reached in 6 minutes 54 seconds. 33,800 feet service ceiling. Normal range 621 miles plus 1.5 hours combat.

Weights: 6349 pounds empty, 8708 pounds loaded.

Dimensions: 36 feet 10 7/16 inches, length 32 feet 6 9/16 inches, height 11 feet 1 1/4 inches, wing area 226.04 square feet. Armament: Two fuselage-mounted 20-mm Ho-5 cannon and two wing-mounted 20-mm Ho-5 cannon (Ki-84-Ib). External stores included two 551-pound bombs or two 44-Imp gall drop tanks.


No effort on the part of Najajima seemed to succeed in turning the Ha-45 engine into a truly reliable powerplant, so the JAAF began to look for other sources of engines for the Hayate. In view of the successful adaptation of the Kawasaki Ki-61-II Hien airframe to take the Mitsubishi Ha-112-II air-cooled radial engine, the JAAF thought that the Ha-45 engine problems might be solved by replacing this engine with the Mitsubishi Ha-112 in the Ki-84.

The designation Ki-116 was applied to the fourth Mansyu-built Ki-84-I adapted to take a 1500-hp Mitsubishi [Ha-33] 62 (Ha-112-II) driving a three-bladed propeller. This engine was borrowed from a Ki-46-III twin-engined reconnaissance aircraft. This engine was substantially lighter than the HA-45 that it replaced, and required that the engine mounts be lengthened in order to maintain the centre of gravity. In order to compensate for the additional length, the tail surfaces had to be enlarged. The Ki-116 weighted only 4850 pounds empty, a full 1000 pounds lighter than the standard Ki-84-Ia. The Ki-116 showed considerable promise and had a performance approximating that of the Ki-100. Test pilots were extremely enthusiastic about its capabilities, but the Japanese surrender brought an end to further development.

Specification of Ki-116:

Engine: One Army Type 4 fourteen-cylinder air-cooled radial (Mitsubishi Ha-33) rated at 1500 for take off, 1350 hp at 6560 feet and 1250 hp at 19,030 feet.

Performance: Maximum speed 385 mph.

Weights: 4938 pounds empty, 7039 pounds loaded.

Dimensions: wingspan 36 feet 10 7/16 inches, height 11 feet 3 13/16 inches, wing area 226.04 square feet.

Armament: Two fuselage mounted 12.7-mm Type 1 (Ho-103) machine guns and two wing-mounted 20-mm Ho-5 cannon.


The Ki-84N was a projected high-altitude interceptor version of the Hayate powered by a 2500-hp eighteen-cylinder twin-row Nakajima [Ha-44] 13 (Ha-219) air cooled radial engine. The wing area was increased from 226 square feet to 249.19 square feet. The production version of the Ki-84N was assigned the Kitai number of Ki-117, and the aircraft was in the initial design stage when the war in the Pacific ended.


The Ki-84P was another high-altitude interceptor version of the Hayate powered by the 2500-hp eighteen-cylinder twin-row Nakajima [Ha-44] 13 (Ha-219). The Ki-84P differed from the Ki-84N in having the wing area further increased to 263.4 square feet. The Ki-84P was abandoned in favour of the less-ambitious Ki-84R.


The Ki-84R was a projected high-altitude version of the Ki-84-I Hayate powered by a 2000 hp Nakajima [Ha-45] 44 with a mechanically-driven two-stage three-speed supercharger. At the time of the Japanese surrender, the first prototype was eighty percent complete.

Staving Off Aerial Immolation

Japan’s ability to repel an American bombing campaign began with very few prospects in 1942 and sharply declined thereafter. Yet an enduring question is why Tokyo squandered more than two years after the Doolittle Raid, and why so little interservice coordination was attempted once B-29s appeared in homeland skies. The answer lies in the Japanese psyche more than in its military institutions.

In defending its airspace, Japan’s army and naval forces were tasked with a nearly impossible mission. Nonetheless, they failed massively in even approaching their nation’s potential to ameliorate the effects of the Allied onslaught.

Japan’s only prospect for staving off aerial immolation was to inflict unacceptable losses upon B-29s. Because of the Superfortress’s exceptional cost—some $600,000 each—a downed B-29 represented the financial equivalent of nearly three B-17s or B-24s, plus an invaluable crew. Development of ramming units demonstrates that some Japanese understood the value of a one-for-one or even two-for-one tradeoff, but the tactic largely failed for technical and organizational reasons. Therefore, defense of the home islands reverted to conventional means: flak guns and ordinary interceptors.

The resulting failure was systemic, crossing all boundaries of government and military-naval leadership. Probably the major cause was Japan’s national psychology: a collectivist culture possessing a rigid hierarchy with unusually strict protocols that inhibited breakout thinking and instilled extreme reluctance to express contrary opinions. Japan poses an intriguing puzzle for sociologists and political scientists: how an extremely well-ordered society permitted itself to make a series of disastrous decisions, each threatening its national existence. Ironically, the situation was partly explained by the atmosphere of gekokujo (“pressuring from below”) in which strident subordinates often influenced their superiors.

If interservice rivalry constituted a “second front” in Washington, D.C., it was a full contact sport in Tokyo. The postwar United States Strategic Bombing Survey concluded, “There was no efficient pooling of the resources of the Army and Navy. Responsibility between the two services was divided in a completely impractical fashion with the Navy covering all ocean areas and naval targets . . . and the Army everything else.”

In June 1944, the month of the first B-29 attack, Imperial General Headquarters combined army and navy assets in an air defense command but the navy objected to army control. A compromise was achieved with naval air groups at Atsugi, Omura, and Iwakuni assigned to the respective army district. Phone links from JAAF command centers were provided to each of the three naval units, but operational integration was seldom attempted. In fact, throughout Japan, the two air arms operated jointly in only three areas: Tsuiki on Kyushu plus Kobe and Nagoya.

A major part of the problem was astonishingly sparse allocation of fighters to air defense. As late as March 1945, Japan allotted less than one-fifth of its fighters to home defense, and the actual figure only reached 500 in July. By then very few were flying, as Tokyo hoarded its strength for the expected invasion.

In the crucial realm of radar, Japan got a jump on the world—and almost immediately lost its lead. The efficient Yagi-Uda antenna had been invented in 1926, the product of two researchers at Tohoku Imperial University. Professor Hidetsugu Yagi published the first English reference two years later, citing his nation’s work in shortwave research. But such was military secrecy and interservice rivalry that even late in the war few Japanese knew the origin of the device that appeared on downed Allied aircraft.

The Allies rated Japanese radar as “very poor,” and fighter direction remained rudimentary. While land-based radar could detect inbound formations perhaps 200 miles out, the data included neither altitude nor composition. Consequently, picket boats were kept 300 miles at sea to radio visual sightings—of marginal use in cloudy weather. However, what radar systems did exist were easily jammed by American radio countermeasures—aircraft dropping aluminum foil that clogged enemy screens.

Furthermore, the Japanese army and navy established separate warning systems, and seldom exchanged information. Even when unit-level pooling was attempted, navy officers generally refused orders from army officers.

Civilian observers were spread throughout Japan to report enemy aircraft, but predictably there was no unity. The army and navy established their own observer corps, and neither worked with the other.

Japanese navy doctrine contained an internal contradiction for air defense. A 1944 manual asserted, “In order to overcome the disadvantages imposed on fighter plane units when the enemy raids a friendly base—that is, getting fighter planes airborne on equal terms with the enemy airplanes—full use must be made of radar and other lookout methods. . . . These must be employed in the most effective manner.” But as noted, use of radar remained rudimentary.

Some pilots dismissed the state of their nation’s electronics. “Why do we need radar? Men’s eyes see perfectly well.”

Excluding mobile radar sets, at least sixty-four early-warning sites were built in the homeland and offshore islands: thirty-seven navy and twenty-seven army. But the rare assets often were squandered by duplicating effort: at four sites on Kyushu and seven on Honshu, army and navy radars were located almost side by side. The southern approaches to Kyushu and Shikoku were covered by some twenty installations but only two permanent radars are known on all of Shikoku.

Though the huge majority of Japanese radars provided early warning, some sets directed AA guns and searchlights. But apparently there was little integration of the two: some B-29 crews returned with harrowing tales of ten to fifteen minutes in a searchlight’s probing beam with minimal or no flak damage.

Apart from inadequate radar, some of Japan’s technical focus was badly misdirected. From 1940 onward, the military devoted over five years to a “death ray” intended to cause paralysis or death by very short-wave radio waves focused in a high-power beam. The nonportable unit was envisioned for antiaircraft use, but the only model tested had a range much less than firearms.

Tactically, the lack of army-navy cooperation hampered the already limited potential of Japan’s interceptors. With unit commanders conducting their own localized battles, there was little opportunity to concentrate large numbers of fighters against a bomber formation as the Luftwaffe repeatedly achieved.

B-29s from Saipan

The pilots who flew the first B-29s from Saipan took with them a valuable fund of knowledge about what their bombers could do and could not do in the skies of Japan, and that knowledge had been amassed-sometimes very painfully-by the men who had flown the big bombers out of Chengtu and Kharagpur. First of all, the bombers could be operated both by day and by night without serious loss; rarely did the loss rate exceed 5 percent, and for all B-29 operations during the war, was under 2 percent. At thirty thousand feet the Superfortress had little to fear from flak. Enemy fighters could operate at that altitude but could rarely manage more than a single pass through a formation, because of the big bomber’s speed. Sometimes, when the weather conditions were right, the B-29 could place its bombs with remarkable accuracy. But the weather proved to be the great limiting factor in the precision bombing for which the plane had been built, since, as in the case of the European theater, targets were all too often obscured by cloud cover. And whereas in Europe it was fairly easy to determine from England what the weather would be like over Mannheim, since the weather generally moved from west to east, this same phenomenon made it extremely difficult to know what kind of weather might move from Siberia or central Asia over the Japanese home islands.


The problem of Japanese weather tended to grow even worse in the fall and winter, as the men of Brig. Gen. Haywood S. Hansell, Jr.’s Twenty-first Bomber Command soon discovered. Hansell believed strongly in the precision-bombing doctrine, which he had helped formulate, so he set his men and planes to work on the Japanese aeroengine industry, most plants of which were well known. The very first raid from Saipan was directed at the Musashi engine works in northwest Tokyo, which produced 27 percent of all Japanese aircraft engines. The Musashi plant, “target no. 357,” was destined to become famous, or infamous, to the men who flew B-29s. During the raid of November 24, there were strong winds at thirty thousand feet, and the target below was almost completely obscured. Three days later, the Superfortresses returned to Tokyo to find the Musashi works completely blanketed by cloud. On December 3 the plant was visible, but bombing was scattered because of high winds.

In all, there were eleven major raids on the Musashi works between November 1944 and May 1945; they cost the attackers fifty-nine Superfortresses. Air crews drilled relentlessly to hit the works. (Some still in the United States made practice bombing runs on the Continental Can Company’s plant in Houston, which was about the same size.) Only the last two raids were effective; all the others were balked by adverse weather. At thirty thousand feet, wind was often more of a problem than cloud, for it could reach in excess of 150 knots. On one downwind bombing run, a B-29 went rocketing over the Musashi plant at a ground speed of more than five hundred mph. The story was not much more encouraging at the other eight high-priority targets. In three months of effort, not a single one had been destroyed. No more than 10 percent of the bombs dropped seemed to be landing anywhere near the objective. Even the Japanese noticed the erratic pattern of the bombing. So many bombs exploded in Tokyo Bay that a joke started to make the rounds of the Japanese capital: The Americans were going to starve the Japanese into submission by killing all the fish.


In the meantime, an alternative approach to strategic bombing was emerging in Washington. General Arnold’s Committee of Operations Analysts had pursued its investigations into incendiary raids to the point of building models of Japanese structures and testing their flammability. The committee proposed several Japanese cities for incendiary attacks, and General Arnold sent out instructions in November to conduct a test raid. General HanselPs heart was not in this sort of bombing. He made a small and inconsequential fire raid on Tokyo on the night of November 29-30, but when he received word to mount a full-scale incendiary effort on Nagoya, using a hundred B- 29s, he protested. Nevertheless, Hansell was a good soldier, so he sent his bombers to Nagoya on the night of January 3-4. The damage caused was slight; bad weather kept reconnaissance planes from getting the photographic evidence for some twenty-seven days. By that time, General Hansell was no longer leading Twenty-first Bomber Command; on January 20 his command had passed to Maj. Gen. Curtis E. LeMay.

The official history of the Army Air Forces indicates strongly that Hansell’s preference for precision bombing cost him his job, and this may indeed be the case. The man who succeeded him did not have the same commitment to doctrine. He had the reputation of a “driving operator” who had already taken over Twentieth Bomber Command and breathed energy into its operations. But, for a month and a half, LeMay made no radical departures in operations from the Marianas. At first, he rode two horses at once: he continued the high-altitude daylight precision raids against the aircraft plants that were now becoming so familiar to his crews; at the same time, he pushed experimentation with incendiary attacks, with which he already had some experience-his XX Bomber Command had succeeded in burning much of Hankow in December 1944. On February 3 he sent the B- 29s to Kobe, where they dropped 159 tons of incendiaries and burned out a thousand buildings, a fairly encouraging result. On February 25 a maximum-effort fire raid on Tokyo produced an impressive level of destruction: a square mile of the city was burned out and over twenty-seven thousand buildings were destroyed. It was early in March that LeMay made the basic changes in B- 29 operations, and on those changes he no doubt staked his career. The fact was that up to that point his bombing force had not “delivered the goods”; that is to say, it had not justified its existence by striking telling blows at the enemy. After three months of operations, the big bombers had delivered about 7.000 tons of bombs, a very modest figure: half of the sorties had ended with the bomber unable to attack the primary target. The clear solution was to drop more bombs and drop them where they would count.

LeMay felt that massive incendiary raids carried out by night against the cities of Japan offered several advantages. First of all, very often the precision targets were located within an urban matrix, so that if the city were burned, the factory or arsenal would go up in flames as well. That the cities were particularly vulnerable to fire was already well established; in many of them 95 percent of the structures were flammable. The attack on a city was an area attack, so it could be conducted in adverse weather and. if necessary, by radar. An attack of this sort had several advantages if delivered by night. It would help neutralize Japanese defenses, which at night were nowhere near as formidable as those LeMay had known in Germany, for the Japanese night fighter was still in its infancy and lacked airborne radar. Japanese flak was sometimes intense but not a grave danger at night. The night attack paid another dividend in that it could be executed at fairly low altitude, as low as five thousand feet. At this height there was less strain on the engines than at thirty thousand feet, and fuel consumption was appreciably lower, so that the bombload could be increased accordingly. And LeMay took a further gamble by ordering his bombers to fly stripped of guns and ammunition; normally the B- 29 carried 1.5 tons of armament. This weight too would now be carried in bombs.

The key to the successful raid was saturation and just the right concentration, as Air Marshal Harris had proved over Hamburg, so when LeMay sent his bombers against Tokyo on the night of March 9-10 he sent an extremely large force-a total of 334 bombers carrying 2,000 tons of bombs, the vast majority of them incendiaries. The first pathfinder planes passed over the city shortly after midnight to mark the target area: a rectangle about three miles by four, containing a hundred thousand inhabitants per square mile, or roughly 1.25 million people. There was no tightly organized bomber stream that night, and the last bombers did not pass over Tokyo until about three hours after the attack had begun. By then, Tokyo was a sea of flames. Tail gunners in the returning B-29s could see the glow of the city 150 miles away; it was a man-made dawn on the horizon, and the first of many that would light the skies over Japan.

The raid on Tokyo on the night of March 9-10, 1945, was the most destructive air raid ever carried out, not excluding the nuclear attacks on Hiroshima and Nagasaki. The loss of life that night has officially been fixed at 83,793, but other estimates have placed it at over 100,000. The vast fires burned out some sixteen square miles of the immense city and destroyed a quarter of a million structures. Several factors contributed to making the attack particularly destructive. Both the air defense and the Tokyo fire brigades were caught off guard by the new tactics, over a hundred firemen lost their lives in the conflagration, and nearly that number of fire trucks were consumed by the flames. Worst of all, that night the Akakaze, or “Red Wind,” was blowing across Tokyo, and it took the flames with it. There was no true fire storm over Tokyo that night. “Because of the wind, the potential fire storm was transformed into an even deadlier force-the sweep conflagration. A tidal wave of fire moved across the city, the flames preceded by superheated vapors that felled anyone who breathed them.

Forty-eight hours after their attack on Tokyo, the B-29s struck Nagoya and then moved on to Osaka and Kobe. Within a ten-day period beginning March 9 the bombers dropped 9,373 tons of bombs and burned out 31 square miles of city. LeMay pushed the firebombing with such energy that by the end of March his depots began to run low on incendiary bombs, and the shortage was not overcome until June. City burning was becoming something of a science, as LeMay’s men tried various weapons and techniques. The M50 thermite incendiary used in Europe had “excessive” penetration. It would often pass entirely through a Japanese structure and ignite in the earth beneath it. occasionally perforating water mains. The best weapon was the M69, a small incendiary bomb, many of which were dropped in a single casing: “Each of these clusters, arranged to explode at 2500 feet altitude, was constructed to release thirty-eight incendiary bombs, made to fall in a random pattern, this arrangement furnishing the basis for the big bombing success to come. The orderly design or distribution from one bomber with an intervalometer setting, or spaced fall, of one bomb every fifty feet, could burn about sixteen acres, as each Superfort had a full bomb load of 16,000 pounds.” The basic procedure, concludes this passage, “was like throwing many matches on a floor covered with sawdust.”

As these descriptions indicate, the destruction was most effective if carried out systematically. With “impressionistic” bombing-that is, with each bombardier trying to place his bombs where they would extend the damage-the ultimate yield was less than if there was a general pattern. In some cases radar bombing was more effective than visual aiming. Two hundred and fifty tons of bombs per square mile, adequately distributed, virtually guaranteed total destruction of the area. Everything combustible would be consumed, and the fierce temperatures generated would ensure that by radiant heat alone the conflagration would cross streets and canals. In some cases the heat would soften the asphalt in the streets, so that fire equipment mired down and was lost to the flames. Water sprayed on the fire would simply vaporize; glass panes would soften and drip from metal window frames. Here and there, incredibly, concrete melted. No living thing could survive in such an atmosphere.

Hapless Defence

There was very little that the Japanese government could do, short of capitulation, to prevent the incineration of its great cities one after another. The menace from the Marianas was growing every day. By June, General LeMay was mounting raids with five hundred Superfortresses, and by September he would have a thousand at his disposal. In March, American P-51 fighters began to move to bases on Iwo Jima, and by April they were appearing over Japan. From February on, the attacks from LeMay’s B-29s were supplemented by those from carrier-based planes, which periodically appeared to harass the home islands.

Japan’s early-warning network had begun to disintegrate, like that of Germany. The increasingly mighty American navy had destroyed Japanese picket ships or driven them toward the shelter of the home islands. The type-B radar, with its range limited to 150 miles or so, was an inadequate substitute. The Japanese fighter force probably made its heaviest impact on the raids in January 1945, when B-29 losses rose to 5.7 percent; thereafter, the Japanese fighters had less success, although the pilots were plucky and aggressive to the end. The Tenth Air Division held the Kanto Sector, covering the highest-priority targets, Tokyo and Yokohama. On the night of the great March raid in Tokyo, they put eight fighters in the air; there were at that time only three hundred fighters for the defense of all Japan plus two hundred machines available in the training schools. Some pilots tried to make up the deficiencies by extraordinary measures, such as ramming the B-29s. This tactic was first used against the B-29 in August 1944 and from time to time afterward; late in 1944 the Japanese high command ordered the formation of “special duty” units whose pilots were to ram the American bombers. In statistical terms, the policy seemed justified. The Japanese pilot took with him eleven American crewmen and a bomber twelve times the size of his fighter plane. But many Japanese commanders violently opposed the policy of ramming. Japan was already running short of experienced pilots, and this practice would take the lives of those who were left.

Some Japanese fighter pilots pinned their hopes on the Shusui jet-powered fighter, which could climb to thirty thousand feet in a scant four minutes, but the fabled weapon came too late. In July, air force authorities were working on a daring plan called the Ken operation. Transport planes would fly special demolition teams to the Marianas, where they would storm the airfields and destroy the Superfortresses on the ground. The scheme collapsed when the transport planes were destroyed in an air raid. For lack of radical solutions, the air defense authorities continued with traditional methods. They decided not to challenge every air attack, but to husband their strength for the big bomber incursions. Japanese intelligence tried to “read” American radio traffic and predict when and where attacks might take place. The flak forces, woefully insufficient, were moved about according to the- readings; at one point, nearly a third of Japan’s flak units were being shifted about between potential targets.

The Japanese authorities did what they could in the way of passive defense. Beginning in June 1944, they began evacuating young children from urban areas and ultimately other groups as well. Although Japan was losing much of its industrial capacity with the burning of its cities, the authorities did not order dispersal and relocation of critical industries until the spring of 1945. They probably delayed because they knew that war production, already slumping in late 1944, would dip further as the firms shifted their operations to new localities. Within each Japanese city, the local authorities tried to prepare for fire attacks, filling water reservoirs and cutting firebreaks, often by demolishing whole blocks; municipal authorities made agreements to lend fire-fighting apparatus back and forth between threatened cities.

Overall, Japanese fighters were spectacularly ineffective against B-29s. From more than 31,300 Superfortress sorties over the homeland, only seventy-four were known lost wholly to interceptors and perhaps twenty more in concert with flak guns. Japanese pilots logged their best performances in January and April 1945, each with thirteen bombers downed. But during fifteen months of combat, losses to interceptors amounted to merely 0.24 percent of effective B-29 sorties.

The Strategic Bombing Survey concluded, “The Japanese fighter defense system was no more than fair on paper and distinctly poor in practice. One fundamental matter stands out as the principal reason for its shortcomings—the Japanese planners failed to see the danger of allied air attacks and to give the defense system the requisite priorities.”

Lieutenant General Saburo Endo of Army Air Force Headquarters stated, “Those responsible for control at the beginning of the war did not recognize the true value of aviation . . . therefore one defeat led to another. Although they realized there was a need for merging the army and the navy, nothing was done about it. There were no leaders to unify the political and the war strategies, and the plans executed by the government were very inadequate. National resources were not concentrated to the best advantage.”

In short, in Japan’s military, parochialism trumped efficiency at every turn.


Nakajima B6N Tenzan

In 1939 the Imperial Japanese Navy drew up its specification for a carrier-based torpedo-bomber to supersede the Nakajima B5N. The specifications issued by the navy called for very modern characteristics. A maximum speed of 288 mph (463 km/h), a cruising speed of 230 mph (370 km/h) and a range of 1,000 nautical miles (3335 km) without a bombload. To meet the requirement, Nakajima decided to use an airframe very similar to that of the earlier aircraft, differing primarily in its vertical tail surfaces. The navy had specified use of the Mitsubishi Kasei radial engine, but Nakajima decided to use instead its own 1,870 hp (1395 kW) Nakajima NK7A Mamoru 11 radial engine of similar output driving a four bladed Hamilton type propeller. The first of two prototypes was flown in spring 1941, but initial flight testing revealed a number of problems, including engine vibration and overheating, but the most serious was that of directional stability, requiring revised vertical tail surfaces. Final flight testing carried out aboard the aircraft carriers Ryuho and Zuikaku in the end of 1942, revealed further problems with the tuning of the engine and the need to reinforce the arrester hook and landing gear. It was not until February 1943 that the type entered production as the Navy Carrier Attack Bomber Tenzan Model 11, company designation Nakajima B6N1, incorporating a number of refinements as a result of extended flight testing. However, after only 135 production Tenzan (heavenly mountain) aircraft had been delivered a new crisis arose when Nakajima was ordered to terminate manufacture of the Mamoru engine, and use the more reliable 1,850 hp (1380 kW) Mitsubishi MK4T Kasei 25 engine, a step also taken to allow greater emphasis to be placed on production of the widely-used Nakajima Homare and Sakae engines.

The company was now compelled to use the engine which the navy had specified originally, the Mitsubishi Kasei, but fortunately the adaptation of the B6N airframe to accept this powerplant presented no major difficulties. The resulting aircraft, which was also the major production version, had the designation B6N2 and differed only from the B6N1 by the installation of the Mitsubishi Kasei 25 engine. The B6N2a variant had the rear-firing 7.7 mm (0.303 in) machine-gun replaced by one of 13 mm (0.51 in) calibre. When production ended, Nakajima had built a total of 1,268 B6Ns of all versions, this number including two modified B6N2 airframes which had served as prototypes for a proposed land-based B6N3 Model 13. The powerplant had been the improved 1,850 hp (1380 kW) Mitsubishi MK4T-C 25C version of the Kasei engine and the strengthened landing gear had larger wheels for operation from unprepared runways, but production did not start before the war ended. Allocated the Allied codename ‘Jill’, the B6Ns saw intensive use during the last two years of the war for conventional carrier operations and, in the latter stages, in kamikaze roles.


Nakajima B6N2 – Nakajima was ordered to cease using the Mamoru engine and use instead the Mitsubishi Kasei 25 engine, thus resulting in the redesignated B6N2. Although the Kasei 25 was slightly less powerful, this was offset by introducing a less drag version of the exhaust ports which also gave a slight jet-thrust like boost effect.

Nakajima B6N2a – This type differed from the B6N2 only by having a rear firing machine gun of 13 mm (0.51 in) calibre, instead of the 7.7 mm (0.303 in) type used on the B6N2.

Nakajima B6N3 – Two conversions of the B6N2a resulted in the B6N3 prototypes equipped with 1,850 hp (1380 kW) Mitsubishi MK4T-C Kasei 25C engines for evaluation as land based bombers.

(Navy Carrier Attack Bomber Tenzan “Heavenly Mountain” Model 11 – Nakajima B6N2)

Allied Codename: Jill

Type: Three Seat Carried based Torpedo Bomber

Design: Nakajima Hikoki KK with Kenichi Matsamura as led Technical Director

Manufacturer: Nakajima Hikoki KK

Powerplant: (B6N1) One 1,870 hp (1395 kW) Nakajima NK7A Mamoru 11 14-cylinder radial engine. (B6N2) One 1,850 hp (1380 kW) Mitsubishi MK4T Kasei 25 14-cylinder radial engine. (B6N3) One 1,850 hp (1380 kW) Mitsubishi MK4T-C Kasei 25C 14-cylinder radial engine.

Performance: Maximum speed 298 mph (480 km/h); service ceiling 29,660 ft (9040 m); initial climb rate 1,885 ft (575 m) per minute.

Range: Normal 1,084 miles (1745 km); Maximum (overload) 1,892 miles (3045 km) on internal fuel.

Weight: Empty 6,635 lbs (3010 kg) with a maximum take-off weight of 12,456 lbs (5650 kg).

Dimensions: Span 48 ft 10 1/2 in (14.90 m); length 35 ft 8 in (10.87 m); height 12 ft 5 1/2 in (3.80 m); wing area 400.42 sq ft (37.20 sq m).

Armament: One 7.7 mm (0.303 in) Type 98 machine gun manually aimed from rear cockpit and one manually aimed by middle crew member from rear ventral position and one fixed 7.7 mm (0.303 in) Type 98 machine gun in left wing (often absent from the B6N1). A 1,764 lbs (800 kg) 18 inch torpedo carried offset to the right of centreline, or six 220 lbs (100 kg) bombs carried under the fuselage.

Variants: B6N1 (Mamoru engined), B6N2 (Kasei engined), B6N2a, B6N3 (prototypes for land based version).

Avionics: Some later models were equipped with ASV radar for night operations.

History: First flight March 1941; service delivery (B6N1) early 1943; service delivery (B6N2) December 1943.

Operators: Japan (Imperial Japanese Navy).