We're currently closed.Todays Hours: 9am – 5pm HST

Pearl Harbor: Thunderfish in the Sky

December 28, 2015

Posted on December 28, 2015
By Ray Panko | [email protected] | Pearl Harbor Aviation Museum

Japan’s Type 91 Modification 2 Torpedo Fins at Pearl Harbor


Key Points

  • The Japanese had to modify their Type 91 Modification 2 aerial torpedoes for the Pearl Harbor attack.
  • They needed to limit the initial plunge so that the torpedoes would not strike the bottom mud.
  • The big stabilizing fins at the rear of the tail cone were an older modification unrelated to this problem.
  • The enabling innovation was small wooden fins near the front of the tail cone
  • These were gyroscopically driven ailerons. They ensured that the torpedo dropped without roll.
  • This allowed the torpedo’s horizontal rudders to be used to pitch the nose up immediately on impact with the water. This would not be possible if the torpedo were rolled when it hit the water.
  • This reduced the initial plunge.
  • It allowed the aircraft to drop their torpedoes from an altitude of 66 feet (20 m) and at a speed of 185 mph (160 kt, 300 km/h).
  • The British used an entirely different system for their shallow-water attack at Taranto Harbor a year earlier.
  • They attached a wire to the torpedo’s nose. Playing out from a drum attached to the airplane, the wire exerted upward resistance on the nose, keeping it from rotating downward too steeply.

Fin Beliefs and Myths

When an aerial torpedo weighing almost two tons slams into the water at 200 mph, it barely slows down. In the open sea, Japan’s aerial torpedoes plunged 150 feet before climbing back to attack depth.[1] Pearl Harbor was only about 40 feet deep,[2] so Japan needed to modify its tactics and torpedoes to attack successfully. This was not impossible. A year earlier, the British had attacked Italian ships in Taranto Harbor, which was 60 to 75 feet deep where the British torpedoes launched.[3] On June 13, 1941, Adm. Royal E. Ingersoll sent a report to Adm. Husband E. Kimmel warning that, contrary to previously sent information, successful attacks could take place in harbors less than 75 feet deep.[4] However, what this meant for Pearl Harbor, which was much shallower, was not clear.

Figure 1 : Water strike by a British aerial torpedo (aircraft is a Fairley Swordfish)
A 1,600-lb. torpedo could make a sizable splash. (Nigel Clarke / Alamy)


On Dec. 7, 1941, Japan proved that it had mastered shallow-water torpedo strikes. After the attack, a shocked U.S. Navy examined the remnants of Japanese torpedoes. They saw that the dropped torpedoes had big wooden fins at the back of their tail cones. These big wooden fins, which were fitted over smaller metal fins, broke off when the torpedo hit the water. The Navy also saw that there were two smaller fins near the front of the tail cones. These were again made of metal covered with wooden gloves that snapped off at water entry.

With no evidence or aerodynamic analysis, the Navy guessed that the big tail fins had been the key to the Japanese torpedo attack. Again with no evidence or aerodynamic analysis, they guessed that the big tail fins had pulled the torpedo’s nose up so that it would not dive so deeply. In Congressional hearings on Pearl Harbor, Adm. Kimmel called the big fins “a device which all the brains in our own Navy Department, who had been seeking such a solution, had been unable to arrive at.”[5] Neither these guesses nor Kimmel was correct.

Later, many argued that the Japanese, being mere imitators, must have copied the technology that the British used in their shallow-water attack at Taranto Harbor a year earlier. In 1995, however, a pilot on the Taranto mission revealed that the British had used an entirely different technology.[6] They attached a wire cable to the nose of the torpedo. The rest of the wire was wound around a drum attached to the Swordfish torpedo bomber. As the torpedo dropped, the wire played out. As it did so, however, it kept upward resistance on the nose of the torpedo, preventing the nose from rotating down too much.

Two-engine torpedo bombers also used a drum spooling system, but it worked differently and had a different purpose.[7] The problem with torpedo drops from twin-engine bombers at sea is the need to tip the nose down. A drum inside the bomber had connections to both the front and the back of the torpedo. The front connection broke slightly earlier than the rear connection, tipping the torpedo down.

The Type 91 Modification 2 Koku Gyorai

Japan’s war plans for dealing with the U.S. battleship fleet had long called for waiting for the fleet to steam into Japanese waters, and then destroy it.[8] There, the battleships of the two sides would wage a climactic battle that would decide the war. After the naval treaties of 1922 and 1930, Japan knew that if it had to fight the United States, its battleships would be outnumbered. To address this imbalance, the Japanese planned to reduce the U.S. battle fleet as it crossed the Pacific.[9] Japan focused on the traditional equalizer in sea battles, the torpedo.[10] As the U.S. battleships crossed the ocean, they would undergo frequent torpedo attacks from submarines, surface ships, and aircraft. Japan’s big ace in its planning was the Type 93 torpedo, which Morrison[11] later called the Long Lance. This enormous torpedo could be carried only by surface ships. It had a 1,080-lb. warhead, and by using oxygen as the oxidizer instead of air, it had incredible speed and range.[12] It could travel an amazing 20,000 meters at 48 kt or 40,000 at 36 kt.[13] Nobody else had anything like it.

During the great depression, most countries were hoarding their small caches of torpedoes. In contrast, the Imperial Japanese Navy fired its torpedoes freely, often in live-fire exercises.[14] One reason for this liberal use of torpedoes was that the Japanese economy pulled out of the Great Depression in 1935, many years earlier than the United States and other countries.[15] Another was that by 1936, the Imperial Japanese Army effectively ruled the country, assassinating those who opposed them. By the mid-1930s, they were spending 43 percent of the government budget on their army and navy.[16] Free firing allowed the Japanese to encounter problems, overcome them, and constantly improve the torpedoes that were so crucial to their battle plan. In contrast, when the United States entered the war, U.S. submarine, ship, and aerial torpedoes usually failed during attacks.

Thanks to extensive experience that led to constant development, the Japanese Type 91 torpedo was by far the best aerial torpedo in the world in December 1941. It was reliable at a time when most aerial torpedoes were fragile and quick to stop, sink, or turn away from their targets. Under water, its depth-keeping limits were plus or minus 18 inches. Its yaw limits were a mere 1 percent of the range.[17] Its warhead was big enough to sink even battleships. Perhaps best of all, from the bomber crew’s point of view, it could be dropped at considerable speed and from considerable height. This made torpedo attacks more survivable for the air crew. Although the attack on Pearl Harbor at point-blank range did not require these performance attributes, the Japanese used them three days later, when G3M and G4M two-engine, land-based bombers used Type 91 torpedoes to sink the HMS Prince of Wales and the HMS Repulse off the coast of Malaya. Three days into the Pacific War, the allies had no working battleships.

The Japanese called their torpedoes gyorai, which means “thunderfish.” Their aerial torpedo was the koku gyorai,[18] literally “the thunderfish in the sky.” Specifically, it was the kyuujuuichi shiki koku gyorai (Type 91 aerial torpedo). The shiki/type designation indicated the year in which the weapon was first accepted or used. The 91 meant that the torpedo was created in the Japanese Imperial year 2591, which was Western year 1931. In 1944, the Japanese tentatively developed another aerial torpedo. This became the Type 4 aerial torpedo.[19]

The initial design was improved frequently as its flaws became apparent during intensive training. Torpedo designations were given kai (modification) numbers when the weight of the warhead changed.[20] The Type 91 Kai 1 (Modification 1) design in 1936[21] or 1937[22] had a warhead of 471 lbs., of which 330[23] lbs. was the explosive charge. It could be dropped at 180 kt or faster. The Type 91 Mod 2 design in 1940[24] raised the warhead to 595 lbs., of which 450 lbs. was the explosive charge.[25] Each warhead change required substantial modification to the torpedo fuselage. The Japanese also made improvements between formal modifications. These changes did not change the kai number. This can lead to confusion.

The Large Wooden Stabilizing Tail Fins at the Rear of the Tail Cone

The idea that the big fins were the key advance for the Japanese has been stated in many books (see Appendix A, Fin Quotes). However, most of these statements had little detail or seemed to make up detail, and few big-fin sources had references. Often, they evidenced a lack of knowledge, for example referring to the Type 91 Modification 2 torpedo as the Type 91 Model 2.

The U.S. Navy Technical Mission to Japan interviewed senior Japanese-torpedo developers after the war. After these interviews, the mission was not impressed by the rear big fins, which the Japanese called the hokyoban.[26] The mission’s report noted that, “Wooden tail frames, similar to those used on U.S. aircraft torpedoes, were used for air stabilization.”[27] By this time, such fins were known to be simple stabilizing surfaces, like tail feathers on an arrow or the fins on a bomb. They merely prevented the torpedo from wobbling in flight.

The Japanese encountered the wobble problem in the mid-1930s. To control it, they added the wooden stabilizing tail fins.[28] The hokyoban were given the shiki (type) number 97, indicating that they were implemented around 1937.[29] These rear fins were created by Lt. Cdr. Haruo Haruta and Lt. Makoto Kodaira.[30]

Limiting wobble was important because torpedoes had to hit the water within a small range of vertical angles. Entering the water too steeply could cause the torpedo to dive very deeply. Even if the torpedo did not hit the bottom after this deep initial plunge, the torpedo might not be able to rise to attack depth in time to hit its target.

Striking the water at too shallow a vertical angle, in turn, would cause a belly flop that would make the torpedo porpoise out of the water or even break open. Torpedoes look solid, but they are filled with sensitive devices, including gyroscopes, internal combustion or turbine engines, high-pressure air flasks, an igniter system for the warhead, and a complex rudder-control system.
Their external fins, rudders, and propellers are also subject to damage at splashdown. In one test using American aerial torpedoes, deceleration forces of 18 Gs and 53 Gs were measured, with peak transients as high as 70 Gs.[31] To compound these issues, aerial torpedoes needed to be smaller and lighter than standard torpedoes.. The need for miniaturization makes everything more difficult, and lightness is not a facilitator for designing something that must endure a heavy shock. If you had a choice. A World War II aerial torpedo is the last thing you would drop from an airplane.

During an attack, the pilot or observer used a chart that gave the entry angle based on launch speed and altitude. Figure 2 shows a chart used by U.S. torpedo bombers. In this chart, if the aircraft was flying 260 kt at 800 feet, the drop would take around 7 seconds, and distance traveled through the air before water entry would be around 3,000 feet (1,000 yards).[32] The torpedo would enter the water at a desirable angle of about 28 degrees.[33] The torpedo dropper would add another 400 yards of travel after the drop so that the torpedo could arm and come to the correct depth.[34] This complex task needed to be done in seconds.

Figure 2: U.S. chart for determining the angle of water entry
Source: U.S. Navy, Aerial Torpedo Attack: High Speed, High Altitude (1945).

Stabilization permitted faster and higher drops while still providing a suitable vertical drop angle. For open-ocean releases, airplanes could drop mature versions of the Type 91 torpedo at 260 kt.[35] Preferred drop parameters, however, were 180 kt at an altitude of 330 feet.[36] This gave a 17- to 20-degree entrance angle, which balanced depth of dive with protection against porpoising. In comparison, the Royal Navy’s Mk XII had a maximum release speed of 150 kt.[37] For the dismal early models of the U.S. Mark 13 aerial torpedoes, operational squadrons limited torpedo drops to only 110 kt and 50 feet of altitude.[38] In fairness to the Mark 13’s developers, its design specifications in 1930 were 100 kt and 50 feet,[39] so poor performance was designed into the torpedo. In all cases, the vertical angle of water entry was critical.

If the torpedo wobbled in flight, this would also add uncertainty to the horizontal strike angle. Stabilizing fins also limited wobble in yaw. Consequently, the big fins also helped ensure that wobble would not cause the torpedo to yaw left or right significantly at water entry, which would cause it to hook away from its path.

Again, the Japanese encountered and overcame the air-drop wobbling problem during the 1930s. These problems would later plague the aerial torpedoes of other countries, especially the United States. In response, other countries also added wooden stabilizing fins for the air drop. The Italians had wooden stabilizing fins by 1940 when they torpedoed the HMS Kent with a U.S. observer on board.[40] America introduced them in 1944 after extensive development,[41] helping to finally turn the “outstandingly bad”[42] Mark 13 aerial torpedo into an effective weapon. The British used them, as well.[43],[44]

Figure 3: Fins on an Italian torpedo (aircraft is an SM-79)

Source: United States Navy.

Figure 4: Fins on an American Mark 13 torpedo (operational in 1944)
U.S. Navy, Aerial Torpedo Attack: High Speed, High Altitude, United State Navy Training Film, 1945, Walter Lantz Productions, Inc.

In any case, given that the Japanese had the large wooden fins in 1936 or 1937, hokyoban were part of the koku gyorai technology long before the Japanese struggled to develop shallow-water torpedo enhancements for the Pearl Harbor attack in late 1941. They were not the Japanese solution to the initial dive problem.

The Smaller Forward Aileron Rudder Fins

Although unimpressed by the big rear stabilizing fins, the Technical Mission focused on the smaller fins near the front to the tail cone.

There were no remarkable features about the aircraft torpedoes in operational use during the war, with the exception of the anti-roll stabilizers. The Japanese considered this innovation of great importance in improving torpedo performance. (Technical Mission, Page 1).

To understand how these smaller fins work, it is necessary to understand roll control on torpedoes. Torpedoes do not seem to have an up and down, but they emphatically do. Figure 5 shows that horizontal and vertical metal stabilizing fins terminate in horizontal and vertical rudders. The horizontal rudders act like elevators on an airplane. The vertical rudders act like the rudder on an airplane’s vertical stabilizing tail. Fins gave depth-keeping stability under water, while the rudders produced directional change.

Figure 5: Horizontal and vertical fins and rudders
To see why roll control is necessary for using the horizontal rudders, consider two examples. First, suppose that the torpedo is upright when it hits the water, as Figure 6 illustrates. Also, suppose that the horizontal rear metal rudders are pitched up upon water entry. This will cause the nose to pitch up, which is necessary to reduce depth.

Figure 6: Pitching up when a torpedo is upright.

However, what will happen if the torpedo is rolled when it hits the water? Figure 7 shows a torpedo that has entered the water with a 45-degree roll to the right (clockwise). Again, the horizontal rudders pitch up immediately upon water entry. Now, however, the horizontal rudders yaw the torpedo to the right, causing it to veer off course. In addition, only part of the raised horizontal rudder control authority goes into pitching the nose up to reduce the initial plunge. In the worst case, suppose that the torpedo has rolled through 180 degrees when it strikes the water. Now, pitching the horizontal rudder “up” will actually pitch the nose down, exacerbating the initial plunge. Quite simply, the horizontal rudders cannot be raised on water entry until the torpedo is known to be upright. With proper roll control, the rudder can be pitched fully up as soon as the torpedo hits the water.[45]

Figure 7: Pitching “up” when a torpedo is rolled 45 degrees clockwise

In normal torpedoes, roll was controlled by weighing the torpedo more heavily on the bottom than at the top. Bottom heaviness promoted roll stabilization, which eventually returned a rolled torpedo upright again. However, this method has disadvantages. Figure 8, which is a drawing of a U.S. torpedo warhead, shows that the warhead was tipped down 30 degrees when it was filled with explosive. When the explosive solidified, more weight would be on the bottom than at the top. This weighted the torpedo properly, but it also meant that the warhead could not be filled completely with explosive, reducing its damage potential.[46]

Figure 8: U.S. torpedo warhead poured to be bottom-heavy
Source: Johnson, 1937.

The Japanese took a different approach to roll control. They created active roll control by adding small metal side fins that acted like ailerons on an airplane’s wing. These fins (rudders) flipped in opposite directions to rotate the torpedo clockwise or counterclockwise. To control these rudders, the torpedo had a gyroscope dedicated to their use. The control mechanism was highly sensitive, sensing angular rotation speed. As the torpedo neared the upright position, the rudders flipped in opposite directions, stopping the torpedo precisely upright.[47]

Figure 9: Anti-roll rudder fins on a captured Japanese Type II aerial torpedo used in Pearl Harbor
Source: Ray Panko, photo taken at the Valor in the Pacific Memorial, Pearl Harbor, June 14, 2013.

The aileron had small metal rudders sufficient for rapid roll control underwater. Each metal aileron was only 8 cm2 in area.[48] This was fine underwater, but in the thin air, they did not have enough control authority. The solution to aileron control authority in the air was the same one the Japanese used for tail fins. They covered each metal aileron with a larger wooden glove that broke off when the torpedo hit the water. Each wooden glove was 240 cm2.[49] This gave a 30-fold increase in control authority.

The Torpedoes for Pearl Harbor

When was the roll-control mechanism introduced? It was not part of the Kai 2 design, which appeared in 1938.[50] Instead, it was introduced in late 1941, barely in time for the Pearl Harbor attack. It was an example of a change that did not take place with the introduction of a formal modification.[51]

By the summer of 1941, the Japanese knew that their torpedo attack plan was in trouble. At Kagoshima Bay, the Type 97 Carrier Attack Bombers, which Americans code-named “Kate,” practiced endless at radically low speeds and altitudes. Kagoshima Bay was much deeper than Pearl Harbor,[52] so torpedoes had to be caught in nets to determine how deeply they were plunging. Lt. Cmdr. Shigeharu Murata, who was in charge of the torpedo attack, devised many combinations of launch conditions. This included making drops at 100 kt (115 mph/185 km/h) and an altitude of 10 meters (33 ft).[53] To fly this slowly, the Type 97 had to have its wheels down to add drag and its flaps down to provide enough lift not to stall.[54] However, the torpedoes continued to dive to about 20 meters (66 feet),[55] which was twice the depth needed for success at Pearl Harbor. Speed and altitude control had helped, but they would not be enough.

This was not a new problem for the Japanese. In 1939, they had experimented with harbor attacks using aerial torpedoes. When mock attacks were made on warships in Saeki Bay in Kyushu as part of the Imperial Japanese Navy’s 1939 exercises, the Japanese discovered that the torpedoes pushed themselves into the bottom mud.[56] Lt. Cmdr. Fumio Aiko noted that Manila, Hong Kong, Singapore, Pearl Harbor, and Vladivostok all had an average depth of 50 to 80 feet (15 to 25 meters).[57] To succeed with aircraft in the kind of ships-in-harbor attack that had worked so brilliantly with torpedo boats in the Russo–Japanese War, a shallow-water attack technology and strategy would be needed.

Led by Lt. Cmdr. Haruo Hirota, engineers at the Yokosuka Naval Arsenal in Kanagawa Prefecture worked for two years to find ways to get around this problem.[58] At the same time, the engineers realized that a recent increase in launch speeds from 130 kt to 180 kt was causing the roll problem discussed earlier.[59] However, active roll control was not seen to be possible then,[60] and two years would pass before the problem was addressed. Finally, in the spring of 1941, two solutions appeared. They were examined, and the better solution was developed and put into testing in August 1941.[61]

In early September, the Yokosuka Air Group conducted test drops with torpedoes using the new mechanism.[62] This lasted about a month, and the pilots found that they could achieve initial plunges of about 12 meters.[63] Sometime between October 30 and November 4, Cmdr. Minoru Genda brought 15 to 20 of the new torpedoes to Kagoshima for evaluation by the Pearl Harbor attack force.[64] These first tests were “effective but irregular.”[65] Although the attack group’s selected pilots could keep the initial plunge to 10 meters in most cases, some thunderfish dove deeper.[66] Of almost equal importance, the pilots could now launch their weapons from an altitude of 185 mph (160 kt. 300 km/h) and an altitude of 66 feet (20 m).[67] There was no longer a need to lower the landing gear and fly precariously with lowered flaps during the launch approach.[68] The added speed also would reduce the crew’s exposure to anti-aircraft fire.

Just before the fleet left for Hitokappu Bay to rendezvous for the Pearl Harbor strike, the torpedoes began to arrive. Lt. Cmdr. Murata wanted each torpedo crew assigned to the Pearl Harbor mission to drop at least one of the new torpedoes to familiarize themselves with it.[69] This request was refused, but Murata arranged for an operational test in which three of his B5N2 crews were selected.[70] One crew was in the top third of all crews based on proficiency. The other two were in the middle and bottom tiers. Two of the three launches were successful. The third, executed by the lowest-tier crew, stuck in the bottom mud.[71] This was good enough for the attack, as the Japanese soon demonstrated.

Now the problem was producing enough of the modified Kai 2 torpedoes for the attack. Japan produced Type 91 torpedoes in the Nagasaki Weapons Factory operated by Mitsubishi Heavy Industries. Its production schedule would not make enough torpedoes available until the end of November.[72] That would be too late. Mitsubishi would not permit overtime work, and the factory’s manager, Yukio Fukuda, could not be told about the dire time need for the final torpedoes.[73] However, Fukuda was able to read between the lines. He authorized the overtime.[74] Even so, only 50 of the torpedoes were available when the fleet sailed for Hitokappu Bay, where they would sortie against Pearl Harbor. Akagi remained behind to receive the final batch. Akagi reached the fleet on November 24, only two days before the fleet steamed out of the bay.[75]

Historiographical Notes

A Revisionist Argument?

Given many statements that the big wooden fins at the end of the tail were the key to the shallow-water attack, it is appropriate to ask if this paper is revisionist. I would argue that it is not, in the sense that it is in line with the information the Japanese provided consistently about the role of the roll mechanism compared to the role of the wooden stabilizing tail fins. Furthermore, this information has been available to American scholars since the end of the war. In the Naval Technical Mission to Japan, interviews clearly showed that the Japanese viewed the roll-control mechanism to have been the key development. In fact, the hokyoban were dismissed in the report as nothing special. As noted earlier, other countries had been using large rear fins to stabilize the torpedo during its drop.

Furthermore, the Japanese have been consistent in their statements that the small roll-control mechanism fins, rather than the large tail fins, were the secret to the shallow-water attack. For example, Capt. Mitsuo Fuchida described the situation this way:
Around that time, we were provided with torpedoes fitted with stabilizing fins using gyroscopes to improve the launched torpedo’s entry angle into the sea surface. In actual use, they performed very well, almost reaching the targeted sinking degree of 10 meters. But performance was still uneven, and we wanted one more month of practice to gain more confidence. (Fuchida, Page 67, emphasis added.)

Of course, stabilizing fins could refer to the hokyoban or the aileron rudder fins. The term entry angle into the sea surface is also ambiguous. It could refer to the vertical entry angle or the roll entry angle. However, only the aileron rudder fins used gyroscopes. In addition, “around that time” indicates that the innovation is new, which the hokyoban were definitely not.
In another publication, Fuchida and Pineau (1981) say, “It was not until early November that the torpedo problem was probably solved by adding additional fins to the torpedo” (emphasis added). The hokyoban was not new, so “additional fins” would have to refer to ailerons farther forward.

Yoshino Haruno, who was the observer (teisatsu) on one of the torpedo-armed Kates during the attack, also emphasized the roll-control mechanism.[76] Lt. Yoshino said:
Aerial torpedoes were considered useless in shallow bays or harbors because the weapon needed a depth of 150 feet to dive before coming up to proper running depth. Without proper depth and distance to recover, the torpedoes would either stick into the seabed or pass under their intended targets. This problem was overcome by combining two developments: the training of crews to drop their torpedoes at lower speeds at lower altitudes, and an ingenious new design. It involved a gyro and a rudder-controlled break-away extension fin that was mounted to both sides of the rear of the aerial torpedoes. This kept the torpedo level during its brief drop to the sea. The wooden fins were attached to the torpedo’s existing metal side fins. The wooden fins acted like wings to adjust up or down on each side of the torpedo to keep it stable in flight. Once in the water, the wooden fin extensions broke away. (King, 2012, Page 141, emphasis added)

Again, the statement is not ambiguous only because of the terms in italics. The hokyoban was not merely attached to the sides of the torpedo. And, again, the hokyoban did not involve a gyroscope.

Yoshino was clearly talking about the wooden aileron fins, which were only on the two sides of the torpedo and that used a gyroscope to control the movable rudders. The hokyoban were not just on the two sides of the torpedo, were not controlled by a gyroscope, and had no movable rudders.

The idea that the wooden tail fins were the critical mechanisms for the Pearl Harbor attack appears to be primarily an American idea, as Appendix A indicates. Furthermore, sources for this claim have one thing in common: None cites a source, apart from those that cite earlier writings that do not cite a source. In addition, these sources tend to use the term “Model 2” rather than “Modification 2” when they refer to the Pearl Harbor torpedoes. This indicates a limited understanding of even basic Japanese torpedo terminology. However, as noted above, the Pearl Harbor torpedoes used two sets of wooden extension fins, and this may have caused subsequent confusion, especially in interviews that involved translation.

The Kai 2 Modification versus the External Aileron Fins

It is clear that the addition of roll control during air drops was the key to the Pearl Harbor attack’s success. It is also clear that the mechanism was not finalized until just before the Pearl Harbor attack. However, the Modification 2 design was not created in 1941. Wikipedia gives its date as 1938. The Naval Technical Mission to Japan lists it at 1941, but this would have meant that its total production life was only a few months.[77]

Then what definitive change occurred in 1941? Although this was the introduction of a roll control mechanism for the flight portion of the launch, there are two possibilities. The first is that countersteering roll control did not exist before the Pearl Harbor preparation. The second is that the active roll control mechanism existed since the beginning of the Modification 2 design and that the wooden fin covers for the controller’s metal fins were the addition for Pearl Harbor. Peattie (2007) appears to support the latter (see 44).

I have not found definitive evidence either way. Adding roll control for underwater roll control would certainly make sense in the original design. It would allow the warhead to be filled fully, giving rise to an increase in warhead weight. This would be enough of a benefit to add the mechanism with its original small underwater fins. The wooden gloves would then have been added just before Pearl Harbor. However, it took several months to develop whatever modification was made for Pearl Harbor. Would this have been necessary for adding gloves? In either case, adding roll control during the air drop was the key innovation needed for the shallow water attacks.

Although rear stabilizing fins were definitely a prior invention, was the rear-fin design changed for Pearl Harbor? Again, we have no data. It may have been, and it may have had some impact. However, Japanese sources make it clear that roll control was the definitive addition. One might be able to turn to the writings of the aircrews of B5N2 attack bombers that flew in the raid. I have come across two. Both cite the wooden fins as the definitive innovation, but they give totally different explanations for why the big wooden fins worked. Lt. Juzo Mori, one of the two pilots to hit the USS California, gave the convention explanation — that the fins caught the water at splashdown, tipping the nose up.[78] Lt. Yoshino Haruno, who was noted above, gave an opposite explanation — that the fixed fins pushed the torpedo’s nose down to prevent it from hitting too flatly.[79] (However, he gave primary credit to the roll-control mechanism.) This discrepancy regarding the stabilizing fins may exist because the crews were not told details about the torpedo. Mori’s book seems to reflect that.

Might there have been other contributors? The torpedo on exhibit in Pearl Harbor appears to show more rear metal fins than one would expect from other descriptions of the Kai 2 design. Yet other innovations may have been added. Overall, however, the main innovation still appears to have been active roll control during the air drop, allowing the rear horizontal rudders to be tipped up immediately upon water entry.

Were the Pearl Harbor Torpedoes Modification 2 or Modification 3 Torpedoes?

Were the torpedoes used at Pearl Harbor the Kai 2 version of the design? They probably were, but this has to be determined deductively rather than from direct proof. The Naval Technical Mission to Japan collected production data that can lead to a reasonable deduction. The Type 91 Mod 2 torpedo had been in use for some time before 1941. In the year of the Pearl Harbor attack, the Japanese produced 237 of these weapons before production ended in November. As noted earlier, tests of the torpedoes with the roll-control mechanism began in August 1941. In August, September, October, and November, 121 Modification 2 torpedoes were built. The fleet left Hitokappu Bay on November 26 to begin the attack, so the total available for Pearl Harbor might have been somewhat less. However, there would still be a large enough number of Mod 2s to account for the 100 sent to Pearl Harbor and for the dozen or more used in tests. If some July torpedoes had not been delivered, even more would be available for production, even if not all November torpedoes could be delivered. In Modification 3 torpedoes had been built. This would not have been enough for Pearl Harbor, much less testing. It appears reasonable to conclude that the torpedoes used at Pearl Harbor were Type 91 Modification 2 torpedoes to which the roll control mechanism had been added late in 1941.

Table 1: Japanese Aerial Torpedo Production in 1941
Source: U.S. Navy 1946 (54.) The production data begins in April because that was the beginning of the fiscal year.

Simulation Evidence

In a television documentary,[80] a 1/5 scale model of a Type 91 Kai 2 torpedo with mock wooden roll control fins was dropped at scale speed into a long water tank. This was done five times. High-speed cameras captured each water entry and subsequent motion of the torpedo model. In this demonstration, the wooden tail frames caught briefly when they hit the water, then broke off. The model quickly returned to the surface. This demonstration was led by Dr. Guy Meadows of the Marine Hydrodynamics Lab at the University of Michigan—Ann Arbor. In a private communication with the author, Dr. Meadows confirmed that the model was indeed 1/5 scale in size and was dropped at scale speed. However, he confirmed that it was not 1/5 scale in weight. If it had been, the model would have weighed more than 350 pounds. In the video, it was clearly light enough for the person who did the dropping to do it easily. Given that the model was largely hollow, it probably would have quickly popped to the surface regardless of the fins. In addition, the horizontal rudders were full up at water entry, which would not work if the torpedo had been in a rolled condition.

Another scale-model drop could be done, this time with 1/5 scale weight as well. However, the model would have to have the precise weight distribution of items inside the torpedo. In addition, the angle of entry would have to be precisely controlled. More generally, water entry is an extremely complex process—far more than air drops and underwater travel. Extensive British experiments with torpedo water entry found that it could not be modeled from aerodynamics, in part because the torpedo creates a partial vacuum around itself in the first moment of water entry.[81] It is not clear if a scale model could duplicate these conditions. In addition, to improve the U.S. Mark 13 aerial torpedoes, Caltech scientists had to launch many full-size torpedoes into a reservoir at different speeds and angles of attack.[82] Only then were they able to develop the shroud rings that greatly improved the Mark 13’s performance when entering the water. In addition, the idea that the tail fins caught the water just enough to pull the nose upward would depend heavily on the physics of the material join between the frangible fins and the body. There would be no way to model this. Most importantly, as noted earlier, extensive Japanese experiments with the big wooden tail fins available in the months before the torpedoes with roll control appeared all failed to stop the initial plunge sufficiently during the run-up to the Pearl Harbor attack.

Using Wikipedia

I would like to close this research issues section with a note on the use of the Wikipedia article on the Type 91 torpedo. I have used it because it is the only English source that provides considerable information from Ichikawa, Kodaira, and Kawada’s book, Kyu Ichi Kai – Koku Gyorai Note (91 Association – Aerial Torpedo Notebook). This book was created in 1985 by key members of the Type 91 development team. It provided detailed information not available from other sources. The Wikipedia article uses the information from the notebook, often without attribution. However, the Naval Technical Mission to Japan and other sources provide ample information on key elements of the argument.


Admiralty Research Laboratory, Teddington, Middlesex (1945). Note on the Underwater Behaviour for a Mk XII Torpedo When Fitted With a Collapsible Nose Cap or So Called “Drag Ring,” A.R.L/N.18/H.36. Read by the author at the National Archives II in 2014.
Aldridge, Arthur (2014). The Last Torpedo Flyers: The True Story of Arthur Aldridge—Hero of the Skies, London: Simon & Schuster.
Barker, Ralph. (2009). Ship-Busters: British Torpedo-Bombers in WWII, Mechanicsburg, PA: Stackpole. Original version published in 1957 by Chatto and Windus.
Boyne, Walter J. (1994). Clash of Wings: World War II in the Air, New York: Simon & Schuster.
Branfill-Cook, Roger (2014), Torpedo: The Complete History of the World’s Most Revolutionary Naval Weapon, Naval University Press, Annapolis, Maryland.
Campbell, John. (1985), Naval Weapons of World War Two, US Naval Institute Press: U.S. Naval Institute, Annapolis, MD.
Clough, Carl E. (1944, October 11). G.B.—Navy—Torpedoes—Decelerations in Drops from Aircraft. Read by the author at the National Archives II in July 2014.
Combinedfleet.com, Japanese Torpedoes, http://www.combinedfleet.com/torps.htm. Last viewed May 31, 2015.
Discovery Channel (2003, December 7). Myths of Pearl Harbor, Unsolved History, Season 2, Episode 7.
Edwards, Peter T. (2010). The Rise and Fall of the Japanese Imperial Naval Air Service, Pen & Sword: Bransley, U.K.
Fuchida, Mitsuo (2011). For That One Day: The Memoirs of Mitsuo Fuchida Commander of the Attack on Pearl Harbor, translated by Douglas T. Shinsato and Tadamori Urabe, Kamuela, Hawaii: Xperience.
Fuchida, Mitsuo, and Pineau, Roger (1981). “I Led the Air Attack on Pearl Harbor,” In Stillwell, Paul, ed. (1981) Air Raid: Pearl Harbor! Recollections of a Day of Infamy, Annapolis, MD: Naval Institute Press,
Gannon, Michael (2001, December 7). Pearl Harbor Betrayed, Washington Post online discussion.
Greaves, Percy L, Jr. and Greaves, Bettina B., ed. (2010), Pearl Harbor: The Seeds and Fruits of Infamy, Mises Institute: Auburn, Alabama.
Groom, Winston (2005). 1942: The Year that Tried Men’s Souls, Grove Press: New York.
Ichikawa, Hidehiko; Kodaira, Makoto; Kawada, Teruyuki (July 25, 1985). “Kyu Ichi Kai – Koku Gyorai Note” or 91 Association – Aerial Torpedo Notebook. Tokyo, Japan: Iyeno Hikari Private Publishing Service. ISBN. Used in the Wikipedia article.
Ito, Mansori with Pineau, Roger, Trans. Andrew Y. Kuroda and Roger Pineau (1956). The End of the Japanese Navy, New York: W. W. Norton.
Johnson, I. C., Captain (April 14, 1937). U.S. Navy Torpedo Mark XIII, Ord 629, Torpedoes.
Jolie, E.W. (1978, September 15). A Brief History of U.S. Navy Torpedo Development, NUSC Technical Document 5436, Weapons Systems Department, Newport Laboratory. http://archive.hnsa.org/doc/jolie/index.htm.
King, Dan (2012). The Last Zero Fighter: Firsthand Accounts from WWII Japanese Naval Pilots, Pacific Press: Irvine, California.
Kuwae, Michinoba; Okuda, Noboru; Miyasaka, Hitoshi; Omori, Koji; Takeoka, Hidetaka, and Sugimoto, Takashige. (2007). “Decadal- to Centennial-Scale Variability of Sedimentary Biogeochemical Parameters in Kagoshima Bay, Japan, Associated with Climate and Watershed Changes,” Estuarine, Costal, and Shelf Science, 73, 279-289.
LeFeber, Walter (1997). Clash: U.S.–Japanese Relations through History, New York: W. W. Norton.
Lord, Walter (2001). Day of Infamy, 65th Anniversary Edition, New York: Holt. (Originally published in 1957.)
Lowry, Thomas P. and Wellham, John W. G. (2000). The Attack on Taranto: Blueprint for Pearl Harbor, Mechanicsburg, Pennsylvania: Stackpole Books. (Originally published in 1995).
Matsuo, Kinoaki (1942). How Japan Plans to Win the War, Boston: Little, Brown and Company. Translation by Kilsoo J Hahn of the Japanese book, The Three-Power Alliance and the United States-Japanese War, published in 1940.
Mori, Juzo (2015). The Miraculous Torpedo Squadron — A Japanese Pilot’s Account of Pearl Harbor, translated by Nicholas Voge, Kojin Publishing. Originally published in Japan in 1973 as Kiseki no Raigekitai.
Morison, Samuel Eliot (1948). History of the United States Naval Operations in World War II. Volume III: The Rising Sun in the Pacific 1931–April 1942, Little, Brown, and Company: New York.
Morison, Samuel Eliot (1950). History of United States Naval Operations in World War II. Vol. VI: Breaking the Bismarcks Barrier: 22 July 1943–1 May 1948, Little, Brown, and Company: New York.
O’Connor, Christopher Patrick (2010). Taranto: The Raid, The Observer, and the Aftermath, Indianapolis, IN: Dog Ear Publishing.
Peattie, Mark R. (2007). Sunburst: The Rise of Japanese Naval Air Power: 1909-1941, Naval Institute Press, Annapolis, Maryland. ISBN 9781591146643.
Pool, Bob (March 27, 2003). “Torpedo Test Site Launched New Arms Era,” Los Angeles Times. http://articles.latimes.com/2003/mar/27/local/me-surround27.
Prange, Gordon W. with Goldstein, Donald M. and Dillon, Katherine V. (1981). At Dawn We Slept: The Unknown Story of Pearl Harbor, Penguin Books, New York.
Reynolds, Clark G. (1982). The Carrier War, Time-Life Books, New York.
Scheid, Ann (1981) Frederick C. Lindvall- How It Was (Part 2). Engineering and Science, 44 (3). pp. 21-25. ISSN 0013-7812, http://resolver.caltech.edu/CaltechES:44.3.Lindvall.
Scutts, Jerry (2000). War in the Pacific: From the Fall of Singapore to the Japanese Surrender, Thunder Bay Press: San Diego, California.
Stille, Mark E. (2011). Tora! Tora! Tora! Pearl Harbor 1941, Osprey: Oxford, UK.
Tagaya, Osamu (2012). Imperial Japanese Naval Aviator 1937-1945, Oxford, UK: Osprey.
Tillman, Barrett (2008). TBF/TBM Avenger Units of World War II, Oxford: Osprey.
Tillman, Barrett (April 03, 2013). “The Outstandingly Bad Mark 13 Torpedo,” Flight Journal Magazine. http://www.flightjournal.com/blog/2013/04/03/iconic-firepower-the-outstandingly-bad-mark-13-torpedo. Last viewed December 12, 2014.
United States Congress, Joint Committee on the Investigation of the Pearl Harbor Attack (1946). United States Congress, 79th Congress, 2nd Session, Investigation of the Pearl Harbor Attack, Appendix F: Geographical Considerations and Navy and Army Installations, Washington, D.C.: United States Government Printing Office, pp. 489-491.
United States Navy (1946). U.S. Naval Technical Mission to Japan: Japanese Torpedoes and Tubes, Article 2, Aircraft Torpedoes, Index Number O-01-2.
United States Navy (1945). Aerial Torpedo Attack: High Speed, High Altitude, United State Navy Training Film, 1945, Walter Lantz Productions, Inc.
Walling, Homer N. (1968). Pearl Harbor: Why, How, Fleet Salvage and Final Appraisal, Washington, D.C.: Naval History Division.
Wellham, John (1995). With Naval Wings, Kent, UK: Spellmount.
Wikipedia (May 21, 2013). Type 91 Torpedo, http://en.wikipedia.org/wiki/Type_91_torpedo. Last viewed December 12, 2014.
Wilburn, Bryan. (2001, February 25), IJNAF and IJAAF Aircraft Ordnance: Part I Aircraft Torpedoes. http://www.j-aircraft.com/research/bryan_wilburn/ijnaf_torpedos.htm.
Zimm, Alan D. (2011). Attack on Pearl Harbor: Strategy, Combat, Myths, Deceptions: Strategy, Combat, Myths, Deceptions, Casement: Havertown, PA.
Zocchi, Louis (2007). When it Counted: Japanese Bombing and Torpedo Accuracy at Pearl Harbor … and After, Osprey Publishing. https://archive.is/lg4D0.

Back to Top