The Weapons: Converted Aircraft, the Ohka, the Kaiten and the Shinyo

A Yokosuka MXY7 Ohka Model 11 piloted rocket bomb captured by U.S. forces at Yontan airfield, Okinawa, April 1945. American intelligence called it “Baka” (“fool”), a contemptuous wartime name; the Japanese name means “cherry blossom”. Photo: U.S. Navy, U.S. Navy photograph, Public domain via Wikimedia Commons.

Kamikaze · Chapter 4 of 9

The Weapons

Most special attacks were flown in ordinary aircraft carrying one bomb; around them Japan built a family of purpose-made weapons, the Ohka, the Kaiten and the Shinyo boats among them, that cost far more Japanese lives than they took.

October 1944 – August 1945Kanoya · Ulithi · Okinawa · Kure · Yokosuka~25 min read

The word “kamikaze” calls up a single picture: a fighter plane diving on a ship. That picture is accurate for most of the attacks. From the first sortie of the Shikishima unit on 25 October 1944 to the last crashes off Okinawa in August 1945, the great majority of special attacks were made by standard combat aircraft, Zero fighters above all, carrying one bomb and flown into a ship by their pilots. But the special attack (tokubetsu kōgeki, tokkō) was also an industrial programme. In 1944 and 1945 the Imperial Japanese Navy and Army designed, built and issued weapons that had no purpose except to be steered into a target by a man who would die with them: a rocket-powered glider bomb, a manned torpedo, plywood motorboats with an explosive charge in the bow, divers with mines on poles, and a simple single-seat aircraft meant to be built by unskilled workers.

This chapter describes those weapons, what they were made from, how they were used and what they achieved. It is a chapter about machines, but every machine in it carried a man, usually a young one (see Chapter 3), and many of the American sailors they reached are named in Chapter 5 and Chapter 6.

A note on evidence. The specifications here come from museum records (the Smithsonian National Air and Space Museum, the Australian War Memorial, the Pima Air and Space Museum), from the Naval History and Heritage Command (NHHC) and from specialist studies, above all Steven Zaloga’s Kamikaze: Japanese Special Attack Weapons 1944–45 and Bill Gordon’s research on kamikazeimages.net. Contemporary evidence includes U.S. Navy photographs of captured weapons, gun-camera film and action reports; the Japanese side survives mainly in postwar testimony, such as the Ohka veterans interviewed for Naitō Hatsuho’s Thunder Gods, and in unit records compiled after the war. Where figures differ, as they often do, the differences are given.

A Weapon Made from What Was Left

The first special attack units were formed in the Philippines in October 1944 out of the aircraft and pilots that were available, and that principle never changed. When Vice Adm. Ōnishi Takijirō proposed crash attacks to the officers of the 201st Air Group at Mabalacat on 19 October 1944, the weapon he described was a Mitsubishi A6M Zero fighter carrying a 250 kg bomb (see Chapter 1). The Zero was what the 201st Air Group had. Its pilots, many of them young and short of training, could not hope to survive a conventional dive-bombing attack against American fighters and radar-directed guns, but a Zero diving into a carrier’s flight deck with its bomb attached needed no release point and no aiming skill beyond holding the aircraft on the target.

The arithmetic was put plainly by the U.S. Navy after the war. The Commander in Chief, U.S. Fleet’s Anti-Suicide Action Summary of 31 August 1945, a contemporary analysis of hundreds of attacks, calculated that the Japanese needed about 3.6 suicide planes for each ship hit, against 37 aircraft per hit in conventional attacks. That figure, more than any other, explains why the special attack spread from a few volunteer units in the Philippines to the main form of Japanese air attack in 1945. A weapon that was ten times more likely to hit was, for a commander who had already lost his trained aircrew, a rational choice, whatever its human meaning.

The pilots of the converted aircraft were usually accompanied by escort fighters, whose tasks were to fend off American fighters on the way in and to report the results. Those reports, made from a distance, in smoke and under fire, were the main basis of the Japanese claims of ships sunk, and they were consistently and greatly exaggerated. The U.S. Strategic Bombing Survey later concluded that the commanders who directed the campaign had been misled by them; the effect was that Japanese planners believed the weapon was doing far more than it was.

The bomb was the limiting factor. A fighter could carry a 250 kg bomb and later, in the A6M5 and A6M7 versions adapted for the role, a 500 kg bomb. Bombers carried more: the twin-engined Yokosuka P1Y Ginga (Allied name “Frances”) of the Azusa unit that flew from Kanoya to Ulithi on 11 March 1945 each carried an 800 kg bomb. But none of these charges was designed to penetrate the decks and armour of a large warship, and the impact of an aircraft at a few hundred kilometres an hour did little to drive them deeper. The result, as later chapters show, was a weapon that devastated destroyers and small ships, burned carriers, and sank no ship larger than an escort carrier.

The Zero and the Navy Bombers

The Navy flew most of the special attacks and lost most of the airmen: by the count of the Tokkōtai memorial association, about 2,500 Navy airmen against about 1,300 to 1,400 from the Army (see Chapter 8). Its aircraft were the types that remained in production or in reserve in 1944–45. The Zero, called “Zeke” in Allied reporting, was the commonest. Its light construction, the reason it had been so agile in 1941 and 1942, meant that it broke up on impact, but its fuel, its engine and its bomb were enough to wreck a destroyer’s superstructure or set a carrier’s parked aircraft on fire. The two Zeros that struck USS Bunker Hill on 11 May 1945 each carried a bomb of about 250 kg (550 lb, as NHHC gives it); the one that knocked USS Enterprise out of the war three days later carried a bomb of about 500 kg (1,100 lb).

The Navy’s dive bombers and torpedo bombers were used the same way. The Aichi D3A (“Val”), the fixed-undercarriage dive bomber of Pearl Harbor, was obsolete by 1944 and was relegated to special attacks in large numbers; Vals and the newer Yokosuka D4Y Suisei (“Judy”) make up most of the attackers in the action report of the destroyer Laffey on 16 April 1945. The Nakajima B6N Tenzan (“Jill”) torpedo bomber and the twin-engined Ginga carried heavier bombs and a crew of two or three. Floatplanes and trainers followed. On 4 May 1945 the destroyer Morrison was sunk off Okinawa in an attack by two groups of Kawanishi E7K (“Alf”) biplane floatplanes; NHHC records that the five-inch proximity-fuzed shells “didn’t work against the wooden Alfs.”

The slowest aircraft of all were trainers. The Yokosuka K5Y (“Willow”), a fabric-covered biplane trainer, was sent on special attacks from Formosa and Kyushu in the summer of 1945; on the night of 28–29 July one of them sank the destroyer Callaghan on the picket line off Okinawa, the last American destroyer lost in the war. In June 1945 the Navy formed special attack units from the Kyushu K11W Shiragiku, a crew trainer of wooden and metal construction never meant for combat. Their low speed and wooden structure gave American radar operators and proximity fuzes some difficulty, but their bombs were small and their losses heavy.

Army Fighters, Bombers and Trainers

The Imperial Japanese Army formed its own special attack units in the Philippines in November 1944, among them the Banda and Fugaku units of the 4th Air Army, and from early 1945 the Shinbu units of the 6th Air Army flew from Chiran, Bansei, Miyakonojō and other Kyushu fields to Okinawa (Chapter 3). The Army’s commonest type was the Nakajima Ki-43 Hayabusa (“Oscar”), a light fighter that by 1945 was outclassed in air combat and was flown south with a single 250 kg bomb. The Mitsubishi Ki-51 (“Sonia”), a light attack aircraft, and the Kawasaki Ki-61 Hien (“Tony”) were also used, and so, in smaller numbers, was the Nakajima Ki-84 Hayate (“Frank”), the best army fighter of the late war. A Ki-84 was the aircraft that struck the battleship New Mexico, Adm. Raymond Spruance’s flagship, on 12 May 1945, killing 54 of her crew (NHHC). A Ki-61 crashed into the attack transport Callaway in Lingayen Gulf on 8 January 1945.

The Army also converted bombers. The Fugaku unit, formed in the Philippines in October 1944, flew Mitsubishi Ki-67 Hiryū (“Peggy”) twin-engined bombers fitted with heavy explosive charges for special attacks. The aim was a heavier blow than a fighter could deliver; the practical result was a larger, slower target. Like the Navy, the Army in 1945 also sent trainers and obsolete types on special attacks as front-line aircraft ran short.

Why a Bomb Was Not Enough

The weakness of the converted aircraft was understood in Japan during the war. The United States Strategic Bombing Survey, interviewing Japanese officers in late 1945, reported that the Japanese commanders were misled by their own inflated claims of ships sunk and ignored their technicians’ advice that a heavier explosive head was needed to sink large ships. The Survey’s summary judgement is one sentence: “However, no ship larger than an escort carrier was sunk.” The three escort carriers lost, St. Lo, Ommaney Bay and Bismarck Sea, were thin-skinned ships converted from merchant hull designs, and in each case the crash set off fires and explosions among the ship’s own aircraft, bombs and torpedoes.

The purpose-built weapons were in part an answer to this problem. The Ohka carried a warhead of about 1,200 kg, more than four times the bomb of a Zero, and was designed to reach its target at a speed no Japanese aircraft could match. The Kaiten carried about 1,550 kg of explosive under the waterline, where a torpedo does its damage. The chart compares the main explosive loads.

Weight was not the whole story. Where the explosion happened mattered as much. A bomb that went off on or above a flight deck could start catastrophic fires among fuelled and armed aircraft, as on Bunker Hill, but it rarely threatened a large ship’s buoyancy. A warhead that exploded against the hull below the waterline, as the Ohka that broke the back of the destroyer Mannert L. Abele did, could sink a ship in minutes. The Navy’s planners knew this; the weapons in the rest of this chapter were their attempt to act on it.

The Ohka: Design

The Ohka (桜花, “cherry blossom”; Navy designation MXY7) began with an idea put forward in the summer of 1944 by a Navy aviator, Ensign Ōta Shōichi, whose name some American sources, including the Smithsonian, give as Mitsuo Ohta. He proposed a small piloted glider bomb, carried to the target area under a bomber and released for a final rocket-powered dive. The First Naval Air Technical Arsenal (Kūgishō) at Yokosuka produced a design within weeks, according to the National Air and Space Museum, and the weapon was in production before the end of 1944.

The Model 11, the only version used in combat, was a small mid-wing aircraft of wood and light alloy with a pointed nose holding the warhead, a cockpit behind it, and three Type 4 solid-fuel rocket motors in the tail. According to the figures gathered by aerospaceweb.org from Japanese and American sources, it was about 6.07 m long with a span of 5 m, weighed about 440 kg empty and 2,140 kg loaded, and carried a warhead of about 1,200 kg. The rockets burned for only a few seconds each, so the Ohka glided most of the way to its target and fired them for the final approach, reaching in a dive a speed often given as about 1,000 km/h, far beyond the reach of anti-aircraft gunners tracking a normal aircraft. About 755 Model 11s were built.

The Ohka had no undercarriage and no means of returning. Its pilots trained on K-1 training gliders at the Navy air base of Kōnoike in Ibaraki from November 1944, by the 721st Air Group, which took the name Jinrai Butai, “Thunder Gods Corps”. Its commander is generally named as Capt. Okamura Motoharu, one of the officers who had argued for special attacks in the summer of 1944. Later versions were designed but never used: the Model 22, with a smaller 600 kg warhead and a Tsu-11 motorjet to give a range of about 130 km from a faster carrier aircraft, the Ginga; and the Model 43, intended to be launched by catapult from caves on the Japanese coast against an invasion fleet. The Kūgishō completed 50 Model 22 airframes, but only three of their engines were built; the Smithsonian notes that “Unlike the Ohka 11, the Ohka 22 never became operational.”

The Americans were able to examine complete Ohkas after U.S. forces captured examples at Yontan airfield on Okinawa in April 1945. U.S. intelligence gave the weapon the name “Baka”, Japanese for “fool” or “idiot”. The name was a wartime expression of contempt, used in reports, the press and sailors’ talk; it appears here only as the historical term it was. The Japanese name came from the cherry blossom, the image of the short life and early fall that wartime propaganda attached to the special attack pilots (see Chapter 9 on how the symbol was used and remembered).

The Betty Problem and 21 March 1945

The Ohka’s flaw was its range. The Smithsonian’s object record states that because the Model 11 could fly only a short distance under its own power, the Mitsubishi G4M (“Betty”) bombers that carried it “had to fly within about 23 miles (37 km)” of a target, well inside the range of the American combat air patrols. Other figures for the range, 60 km and 90 km, appear in the literature, depending on the release height assumed; all of them left the carrier aircraft exposed. A G4M2e loaded with a two-tonne Ohka under its belly was slow and heavy, and the Betty had never been well protected. To reach the American fleet the Jinrai crews needed fighter escort strong enough to hold off the Hellcats, and by 1945 the Navy did not have it.

The first Ohka mission tested that in the worst way. On 21 March 1945, while Task Force 58 was operating off Kyushu, Lt. Cmdr. Nonaka Gorō led the first Jinrai attack from Kanoya. The Smithsonian record counts 16 Bettys carrying Ohkas; kamikazeimages.net, from Japanese records, gives the unit’s loss as 18 Betty bombers and 160 men. Before they came within range of the carriers, they were intercepted by F6F Hellcats, which shot down the whole group. The Smithsonian records that most of the Ohkas fell into the sea. Nonaka died with his crews.

The lesson, kamikazeimages.net notes, was that formation attacks in daylight were reckless while the Americans controlled the air. After 21 March the Jinrai Butai sent its Bettys in small numbers or singly by separate routes, and many of its fighter pilots were used instead as conventional special attack pilots, flying Zeros with 500 kg bombs.

The cost of the mother planes is the part of the Ohka’s story most often left out. Every Ohka mission put a bomber crew of about seven at risk for one Ohka pilot. Those crews were not special attack volunteers in the formal sense, but their survival odds were little better; the 160 men lost on 21 March were overwhelmingly Betty crewmen. Counts of special attack dead that include only the men “who crashed” leave them out, which is one of the reasons the figures in Chapter 8 differ.

Mannert L. Abele and the Ohka’s Record

The Ohka’s one sinking came three weeks later. On 12 April 1945, during the second Kikusui mass attack (Chapter 5), nine Bettys of the 3rd Ohka Special Attack Squadron took off from Kanoya, each with an Ohka, and flew to Okinawa by separate routes. North of Okinawa, the destroyer USS Mannert L. Abele (DD-733), on radar picket duty, was first struck by a Zero; about a minute later an Ohka hit her near the waterline and broke her in two. She sank within minutes. NHHC records 84 of her crew killed and 30 wounded, and calls her “the first ship to be sunk by an Ohka rocket-assisted manned suicide bomb launched from a bomber.” Japanese records credit the hit to Lt. (jg) Dohi Saburō, whose Ohka was reported released that day. The story of the picket line where she died is told in Chapter 6.

Mannert L. Abele was also the only ship the Ohka sank. Ohkas damaged several others. The best documented is the destroyer-minelayer Shea (DM-30): on 4 May 1945, at 08:59, an Ohka crashed through her superstructure, passed clean through the ship and exploded beyond it; NHHC records one officer and 31 enlisted men killed and 91 wounded. Specialist studies list a handful of other ships hit or near-missed by Ohkas; the lists differ, and the total number of Ohka sorties is not established in the sources used for this section. In 2022 the wreck of Mannert L. Abele was found off northern Okinawa by Tim Taylor’s Lost 52 Project, and NHHC confirmed the identification; it is the grave of her 84 dead.

The Kaiten

The Navy’s second purpose-built weapon came from submariners. In 1943 two young officers, Lt. Kuroki Hiroshi and Lt. (jg) Nishina Sekio, proposed turning the Navy’s most successful weapon, the Type 93 24-inch oxygen-fuelled torpedo, known in the West as the “Long Lance”, into a piloted one. The Navy approved development in February 1944; the first prototype was ready at the end of July 1944. The name Kaiten (回天) is usually rendered “turn of heaven”, in the sense of a reversal of fate.

The Type 1 Kaiten, the only one of six models used operationally, was a Type 93 torpedo lengthened to take a small control compartment with a periscope, steering controls and a seat for one man. NHHC gives its warhead as 3,420 lb (about 1,550 kg), its speed as 12 knots cruising and 30 knots maximum, its range as 42 nautical miles and its maximum operating depth as 250 ft. It was about 15 m long and 1 m in diameter. The Kaiten was carried on the deck of a submarine, four to a boat on the first operation. Once released, he navigated by periscope and stopwatch toward his target and made the final run submerged. About 330 Type 1s were built. The training was dangerous: Kuroki himself died in a training accident in September 1944; NHHC counts 15 Kaiten pilots killed in training.

As the NHHC H-Gram on the weapon puts it: “Once the Kaiten launched, however, there could be no return.” Unlike the Navy’s earlier midget submarines, which had at least in principle been expected to recover their crews, the Kaiten was designed as a one-way weapon.

Ulithi, 20 November 1944

The first Kaiten operation, by a group the Navy named the Kikusui unit, was aimed at the two great American anchorages of the western Pacific. Three submarines, I-36, I-37 and I-47, left the Kaiten base at Ōtsushima in the Inland Sea on 8 November 1944, each with four Kaiten. I-47 and I-36 were to attack the fleet anchorage at Ulithi Atoll in the Carolines; I-37 was to attack Kossol Roads in the Palaus. I-37 never launched: on 19 November she was sighted by the net tender Winterberry, and the destroyer escorts Conklin and McCoy Reynolds sank her that afternoon with all hands and her four Kaiten.

At about 03:00 on 20 November I-47 released all four of her Kaiten outside Ulithi; I-36 managed to launch only one. At 05:47 the fleet oiler USS Mississinewa (AO-59), at anchor in the lagoon and loaded with aviation gasoline and fuel oil, was torn open by an explosion, followed seconds later by a second. Her captain, Capt. Philip G. Beck, ordered the ship abandoned about fifteen minutes later; she rolled over and sank at about 09:00. Of her 20 officers and 278 enlisted men, 63 were lost (NHHC). Nishina Sekio is believed to have been the pilot who hit her; he carried Kuroki’s ashes with him. Of the other Kaiten, NHHC records that the destroyer Case rammed one, one ran aground on the reef and one was probably destroyed by depth charges in the lagoon; I-36’s single Kaiten was never heard from again.

Japanese claims for the attack, based on the submarines’ reports of explosions and smoke, were far larger than the single oiler actually lost, and the programme went ahead on the strength of them. Later attacks on anchorages achieved little; one sank the infantry landing craft LCI-600 at Ulithi in January 1945, killing three men. American anti-submarine forces, alerted by Ulithi, sank eight of the Kaiten-carrying submarines before the war ended.

Underhill and the Kaiten’s Balance Sheet

In the spring of 1945 the Sixth Fleet changed tactics and sent its Kaiten submarines against ships at sea along the American supply routes, where the targets were moving but less heavily guarded than the anchorages. The only success of this phase was also the second-largest loss of life the weapon caused. On 24 July 1945 the destroyer escort USS Underhill (DE-682), escorting a convoy from Okinawa to Leyte north-east of Luzon, attacked a submarine contact and was hit by a Kaiten. The explosion tore her in two. DANFS records 113 men lost, among them her commanding officer, Lt. Cmdr. Robert M. Newcomb, and 125 survivors of a crew of 238. Other sources identify the launching submarine as I-53; the DANFS entry does not name it.

NHHC’s accounting of the programme is blunt. Against Mississinewa, Underhill and LCI-600, three ships and about 179 American lives, the Japanese Navy lost eight Kaiten-carrying submarines with 846 of their crews, 106 Kaiten pilots (15 of them killed in training and two who took their own lives at the end of the war) and 156 maintenance and support personnel. About 100 of the 330 Type 1s built were used. In the judgement of the NHHC director’s H-Gram, “the Kaiten program was a miserable failure.”

Shinyo and Maru-re: The Boats

The most numerous special attack weapon was also the simplest. The Navy’s Shinyo (震洋, “sea quake”) was a plywood motorboat about 5 m long, driven by a Toyota or Nissan truck engine of 70–80 horsepower and designed for a speed of 20 knots or a little more, with an explosive charge in the bow, usually given as 250–300 kg. The boat was meant to be driven at night into the side of a transport or landing craft at anchor; the pilot died with it. The Australian War Memorial, which holds a surviving example powered by a six-cylinder Chevrolet engine, and Japanese production accounts gathered by kamikazeimages.net put the number built at about 6,000 to 6,200. The Army built its own version, the maru-re (連絡艇, nominally a “liaison boat”), armed with depth charges, and organized it in sea raiding battalions.

The boats were deployed in hundreds to the Philippines, Formosa, the Ryukyus and the coasts of Japan, hidden in caves and under camouflage on islands where an invasion was expected. Their record is thin. Off Luzon they sank the submarine chaser PC-1129 on 31 January 1945. Off Okinawa, a suicide boat sank the gunboat LCI(G)-82 in late March, and on 27 April a single boat disabled the destroyer Hutchins so badly that she was not repaired (NHHC). Many of the Okinawa boats were destroyed before they could be used, because American troops landed in the Kerama Retto on 26 March 1945, a week before the main invasion, and found many of them in their hiding places.

Gordon’s compilation counts 1,082 Navy Shinyo dead and 266 Army boat dead, together more than the Navy’s and Army’s losses in any other special weapon except aircraft. The number includes men killed in accidents and in explosions at Shinyo bases, as well as those lost in attacks. One Shinyo unit commander, the future novelist Shimao Toshio, waited with his boats on Kakeromajima in the Amami islands until the surrender; his unit never sortied.

Fukuryū: The Divers Who Were Never Sent

The last and most desperate of the Navy’s special weapons was a man. In 1945 the Navy trained units of divers, the Fukuryū (伏龍, “crouching dragons”), to wait on the seabed off the invasion beaches in diving suits and to thrust a contact mine on a pole, with a charge of about 15 kg, against the bottom of a passing landing craft. Training was carried out at Yokosuka in 1945. The divers were never used in combat, because the war ended before the invasion of Japan. The equipment was primitive and the training itself killed men; the number of training deaths is uncertain and is not given here. No wartime photograph of the Fukuryū in this collection’s archive could be verified, and none is shown. A memorial to them stands at the Yūshūkan (Chapter 9).

The Fukuryū belong with the Kaiten bases, Shinyo caves and catapult-launched Ohkas prepared on the Japanese coast in 1945. All were part of the planned defence of the home islands, Ketsu-gō, which is described with the American invasion plans in Chapter 8.

The Ki-115 Tsurugi and the Weapons for the Homeland

The Army’s answer to the shortage of aircraft was a machine designed from the start for one flight. The Nakajima Ki-115 Tsurugi (“sabre”) was, in the words of the Pima Air and Space Museum’s record of its surviving example, intended to supplement normal combat aircraft for organized suicide attacks on Allied ships. It was meant to be built by relatively unskilled workers from non-strategic materials: much of the structure was wood, and the fuselage was covered in thin steel. Its fixed landing gear was dropped after take-off to save weight and complexity. It could take whatever radial engine was available, though only the Nakajima Ha-35 of about 1,130 hp was actually fitted. Test flights showed it was difficult to fly, particularly for the barely trained pilots it was meant for. Just over 100 were built between May and August 1945. None was used operationally.

The Tsurugi and the Ohka Model 43 show where the programme was heading in the summer of 1945: aircraft made to be expended, launched from dispersed fields and caves, against an invasion fleet off Kyushu. At the surrender, the U.S. Strategic Bombing Survey reported, more than 9,000 aircraft were available in the home islands for special attack, and more than 5,000 had already been specially fitted for it. Most were ordinary types, trainers among them; the purpose-built weapons were a small fraction.

What the Weapons Cost

The men who died in special attacks are counted differently by different compilers, and the differences follow from what is counted: aircraft only or all weapons; the Ohka pilots only or the Betty crews as well; training deaths or only combat sorties. The most detailed breakdown by weapon is Bill Gordon’s, in his 2020 thesis on the last writings of special attack corps members, drawn from the memorial association’s 1990 compilation. It counts 5,843 dead in all special attacks. Aircraft account for about two-thirds; the boats for nearly a quarter.

Gordon’s categories show how broad the Japanese definition of “special attack” became. They include the crews of the Navy’s midget submarines, beginning with the nine men of the five Type A boats lost at Pearl Harbor on 7 December 1941, who were honoured in wartime Japan as the “nine war gods” and whom the postwar memorial association counted among the special attack dead (see the Pearl Harbor collection). They include the Army’s Giretsu airborne commandos, who were flown onto Yontan airfield on Okinawa on the night of 24 May 1945 to attack American aircraft on the ground; and the Army fighter pilots of the Shinten units who rammed B-29 bombers over Japan. None of these men flew a converted Zero into a ship, but all were sent on missions from which return was not expected.

The comparison with what the weapons achieved belongs to Chapter 8, but the weapons in this chapter point to its conclusion. The converted aircraft, for all their crudeness, did nearly all the damage: they sank most of the 34 to 57 ships counted as lost to special attack aircraft, damaged several hundred more, and killed thousands of American and Allied sailors. The purpose-built weapons, on which the most engineering effort and the most hope had been spent, sank four ships between them, unless the boats’ few victims are added. The Ohka was defeated by the fighter screen before it could be launched, the Kaiten by the difficulty of finding and hitting a target through a periscope from a moving torpedo, the boats by the speed of the American advance. Each was, in the end, a way of turning a young man into a guided weapon at a time when Japan had lost the industrial and technical means to build any other kind.