WW2 Japanese AT Weapons


Imperial Japanese Army Japan (1939)
Anti-tank Weapon – ~1,200 Made

There is a saying, Proper Preparation and Planning Prevents Piss Poor Performance (also known as the 7 P’s). In 1939, the Imperial Japanese Army proved this to be true by winning a battle against an overwhelming enemy armored force, without a single tank of their own. At the heart of this preparation was a small bottle of soft drink.

Japanese soldiers posing with captured Soviet armor at Nomonhan.

The story starts along the China/Mongolia border, near the town of Nomonhan. In the first part of the 20th century, this wilderness was inaccurately mapped. There was a small patch of land that both the Japanese client of Manchuria and the Soviet client of Mongolia claimed. The competing claims would lead to five months of fighting between the Russians and the Japanese. The Japanese named this war after the town closest to the border, Nomonhan, while the Soviets named it after the river in the area, Khalkhin Gol (the Japanese called the river the Halha).

To recount the entire story of the battle would be a major undertaking, and there are many such works already in existence. However, suffice to say that, from the initial skirmishes starting on the 11th of May 1939, both sides began to escalate, drawing in more men, tanks, guns and aircraft as time wore on.


One of the units swept up in this escalation of forces was the veteran and fully motorized Japanese 26th Regiment, commanded by the competent Colonel Shinichiro Sumi. When his regiment arrived at the logistics base at Hailar on the 22nd of June, Col Sumi dispatched officers to visit the various units that had already been in combat, and find out more details about what facing the Russians would be like. It is almost certain that these officers would have encountered stories of the Soviet’s tanks, the BT-5 and BT-7. At the time, the Japanese infantry would have had what they termed ‘rapid-fire infantry guns’, but today we would recognise them as 37 mm anti-tank guns. These would, of course, wreck the lightly armored BT tanks. However, the 26th Regiment had none of these weapons. Indeed, it was extremely short of heavy weapons, having just six machine guns and an equal number of battalion guns. The other anti-tank weapon the Japanese infantry had was the Type 93 mine, dismissively nicknamed Anpan by the troops, as it resembled the small sweet bread rolls of the same name. This small round mine was fixed to bamboo poles and shoved under the tracks of any attacking tank. The problem was that, on the sandy soil of the area, a tank would push the mine into the ground and not trigger the fuse.

Japanese soldiers of the 72nd Regiment loading into trucks for movement to the front. Type 93 mines can be seen attached to the poles used to shove them under attacking tank’s tracks.
(Source: Unknown Japanese publication, quoted in Coox, 1985)

It is quite possible that, during these investigations, the officers would have interviewed Private, First Class Okano Katsuma from the 23rd Division. During the skirmishes in May he, along with two other men, were assigned as truck drivers to help bring supplies forward. During one such trip, they were chased by a Russian tank. In desperation, PFC Katsuma started to throw cans of petrol off the back of the truck in an attempt to impede the pursuing Soviet tank. Much to the soldier’s surprise, when the tank hit one of these cans, it burst into flame, allowing them to escape.

The idea of petrol as a weapon against tanks and AFVs was not entirely new to the Japanese. Major Nishiura Susumu had been an observer during the Spanish Civil War and had seen the combatants use wine bottles filled with petrol to attack armored vehicles. In July 1937, he had sent a report back to Japan. This was seen with incredulity by the Ordnance Bureau. However, Major Susumu’s insistence convinced them to conduct trials. These failed utterly. In the cold Japanese weather, the stationary tank stubbornly failed to burst into flame. Thus, the Ordnance Bureau concluded there was nothing to this idea.

Back at the supply base supporting the Japanese efforts, Colonel Sumi had no other ideas to help defend his soldiers from tanks, and he had been ordered to move forward to the front. When the Regiment marched out, he left behind 26-year-old 2nd Lieutenant Negami Hiroshi from the regiment’s Quartermaster detachment. He had orders to secure as many bottles as he could from the army supply chain and ship them to the regiment via truck. Lieutenant Hiroshi found the supply dump stocked with thousands of bottles of soft-drink, and he immediately attempted to requisition these. Like in nearly every army ever, the Quartermaster did not want to issue the bottles. ‘Stores are for storing, not for issuing’. Lieutenant Hiroshi’s task was made all the harder, as he could not divulge what he wanted such a large number of drinks bottles for, due to security concerns. It seems odd to consider security in this situation, however, a large portion of the logistics effort was entirely civilian. Indeed, the trucks that the 26th Regiment were mounted in were commandeered from civilian service, and many still were driven by their original owners in their civilian clothes.

Eventually, Lt Hiroshi managed to obtain crates of the soft-drink by being persistent and striking some form of deal with the Quartermasters. He obtained around 1,200 bottles and shipped them to the regiment. The supplies caught up with the soldiers at Chaingchunmiao. There, they were distributed and the men warned not to throw the bottles after they had emptied the contents. Trials were held to determine the best way of creating the weapon. It was determined the best design was to fill the bottle about ⅓ with sand to give it ballast and the ability to be accurately thrown, and the rest topped off with petrol. To complete the weapon, a small wad of cotton, taken from the soldier’s rifle cleaning kit, acted as a bottle stopper and fuse when lit. This weapon was named Kaenbin. There was still one unsolved flaw. The flat open countryside often had a strong wind blowing, which made lighting even something like a cigarette difficult, if not impossible, let alone having to light the wick in battle. With this problem unsolved, each man temporarily filled his bottle with water and tied it to his waist with a string. Lieutenant Hiroshi had acquired enough drink to provide one bottle to every man in the regiment, including Colonel Sumi. There were a few other bottles leftover and these were shared with neighbouring infantry units.

Illustrative picture only: These Japanese bottles were prepared for use as Molotov cocktails during the Sanrizuka Struggle (1966-present) against the Narita International Airport.

To Battle

Starting on the 1st of July, the Japanese launched their counter offensive. They were to cross the river at its narrowest point, forces would hold the bridgehead, and the 26th regiment in its trucks would push round behind the Soviet forces and encircle them, at the same time overrunning the large Russian artillery reserves that had caused so many casualties in the previous two months.

Like so many plans from the Japanese command structure, this plan was powered by no small amount of delusion, passing over some very critical problems that the command structure simply ignored or talked themselves into not believing the issues were important.

The foremost of these was the pontoon bridge to be used to cross the river. It was the only pontoon bridge that the Japanese had in all of China, and it dated from 1900. What is more, there was insufficient construction material. Thus, the bridge was only 2.5 m wide and the pontoons had to be spaced out further than was desirable. The infantry crossing the bridge had to take off their packs. Only one truck was allowed on the bridge at a time, and that had to be unloaded first. Even with these precautions, the bridge still took damage, and so crossing had to be halted every 30 minutes to repair the structure. To make matters worse, the current at the narrowest point of the river was also the strongest, which made the bridge curve.

Japanese soldiers marching across the inadequate bridge over the Halha.

It is no surprise that, by the morning of the 3rd of July, only one of the 26th Regiment’s three battalions was across the river, along with the 71st and 72nd Regiments to hold the bridgehead. The choice was simple, attack with one battalion, or wait for all three to cross. It will come as no surprise that the Japanese chose to attack. Colonel Sumi ordered his men to cross in boats as fast as possible to join the defence, as the lead battalion began its attack.

Faced with a Japanese bridgehead, the Russians reacted immediately. Elements of the 36th Motorized Rifle Division were based at Tamsag. These were the 11th Tank Brigade, 7th Motorized Armored Brigade, and the 24th Motorized Rifle Regiment. In total, they had 186 tanks and 266 armored cars. These were ordered forward to assault the Japanese position. This required a long fast road march in the baking sun and 40 degree Celsius heat. The Soviet armor surrounded the Japanese bridgehead and began probing attacks, while the main column, in no formation, ploughed straight into the lead battalion of the 26th Regiment, and shortly afterwards the remaining two battalions who were trying to advance on foot to catch up.

The terrain of the battlefield was utterly flat and desolate. There were no features, trees or bushes to hide behind, just endless flat soft sandy soil, with very short grass. In such a situation, the tanks should have obliterated the Japanese infantry caught out in the open.

The 71st and 72nd Regiments had access to rapid-fire infantry guns, as well as the 13th field Artillery Regiment, armed with modern Type 90 75 mm guns. Thus, they were able to hold off most of the attacking tanks. Where these guns or Kaenbin were not available, the infantry resorted to Nikuhaku Kogeki (Human Bullet) attacks. In these, the Infantry would hold their ground until the target tank was within about 40 m, then leap up and charge at the tank. The infantry would swarm the tank, attempting to wrench open hatches or cause damage with grenades. This was pure close combat, man against machine in the blistering heat. Soviet tanks would hose their colleagues down with machine gun fire, or, if the crew was quick enough, they could rotate their turret at full speed, throwing Japanese soldiers off. The scalding hot metal plates of the tank’s hull, further heated by running the engine for so long in the direct sun, also proved somewhat of an impediment.

Painting by Tsuguharu Fujita from 1941 showing Japanese Nikuhaku Kogeki teams assaulting Soviet tanks.

At the 26th Regiment, they had no rapid-fire infantry guns. Their only support was from twelve Type 38 75 mm regimental guns. These dated from 1905 and only had HE ammunition. As the tanks bowled towards the 26th Regiment, these guns opened fire at a range of 1,500 m, but were largely ineffective. At 800 m, the handful of Type 90 70 mm battalion guns the regiment owned opened fire, but these could only score a hit with about a third of their shots and were also largely ineffective. At 500 m, the few HMG’s the regiments owned opened fire. As there was no Russian infantry, these machine guns aimed for vision slits, and also had no effect.

A Japanese Type 92 machine gun set up on the plains of Nomonhan. You can clearly see the exposed nature of the terrain and total lack of cover. Two destroyed Soviet armored cars lie in the background.

Then the tanks reached 40 m, and the Nikuhaku Kogeki teams began to attempt to light their Kaenbin. The harsh wind kept preventing ignition. As a tank bore down on him, in desperation, one soldier hurled his unlit bottle. It smashed on the armor of the tank. To everyone’s surprise, the tank burst into flames. Eyewitness accounts describe how a tank struck by Kaenbin burned:

‘…the bottle would shatter, the gasoline contents would splatter quickly, and the sheet of fuel would ignite in the heat of the sun and vehicle. Flames would appear from the bottom of the tank, the way newspaper burns, giving the impression the ground was on fire. When the flames licked the top of the tank, the fire would subside with a puff, for the fuel tank had been entered. Now the inside of the tank would catch fire and burn furiously.’

The suggestion by surviving soldiers was the heat radiating off the armor plate was sufficient to ignite the fuel. However, the accounts miss a few important details. First, from the information we have on ammunition usage, it seems each of the tanks destroyed by Kaenbin were hit by multiple bottles, on average approximately three each, although an accurate figure is hard to determine. This would mean that the tank would be absolutely drenched in petrol, seeping into every opening, especially the engine bay. Here, there are several possible means to ignite the fuel, such as the exhaust, which would be running at several hundred degrees from the long hard drive. Equally, the hours of driving, in the extreme heat, would have meant the transmission in the tank was scaldingly hot.

In the swirling dust, heat haze and smoke-shrouded battlefield, confusion reigned supreme. However, it was a situation the Japanese were ideally suited to. Any officer or NCO would take charge of the men around him, indicate a target and it would be hit by a volley of Kaenbin. Even Colonel Sumi was directing and organizing his soldiers. The Russian tankers were largely ignoring the infantry, trying to concentrate their fire on the support weapons that the Russians assumed to be wreaking so much havoc on their armored force, when it was the infantry who were the main threat. As the battle progressed, some Russian tankers abandoned their vehicles before they were hit, attempting to flee on foot. Those crews that had bailed from burning tanks were also trying to retreat to friendly lines. They had to endure the attention of the Japanese heavy machine guns.

However, the Japanese were not having it all their own way. Casualties were mounting, and on a few occasions, poor coordination between the Battalion Guns and the Infantry meant that Nikuhaku Kogeki teams were killed by friendly fire. By 1500 that afternoon, mere hours after the attack was launched, the Russians withdrew. As they pulled back, they left a field of burning vehicles. These would burn for 3-4 hours after they were hit. Ammunition would suddenly cook-off in the flames, randomly sending turrets flying, or sprays of small arms fire out from their wrecks.

That evening, Colonel Sumi tallied events. The regiment had claimed 83 tanks knocked out, although Col Sumi reckoned this involved some overclaiming. He calculated that the total was around 70. The force, as a whole, had knocked out some 280-230 AFVs from the attacking Russians.

However, the Japanese force was spent. It had taken about 10% casualties and was all but out of ammunition. For example, the 26th Regiment could find just thirty-six Kaenbin. The lead battalion had no ammo left for its Battalion Guns, the other two battalions only had one serviceable gun each, with just one box of ammunition left.

With no hope of resisting the following day, and with Russian Artillery coming more into play, the Japanese began to withdraw. However, through a miscommunication, the lead battalion of the 26th Regiment did not get the message until too late, and took even heavier casualties.

Like many of the Japanese plans from this campaign, the attack was overly ambitious. This overconfidence and lack of ability from the Japanese chain of command would lead, in September, to the total destruction of the Japanese force, and utter victory for the Soviets. Throughout this long battle, the Kaenbin would serve where possible. Today, Nomnhan/Khalkhin-Gol is largely overshadowed by the Second World War, which started just as the battles were winding down.

In the Pacific

The Kaenbin or some other variant of the idea would see service in the latter part of the Second World War. Once again, the Japanese would face a superior armored force in the shape of the Allies. A standard part of the Japanese anti-tank tactics was the Kaenbin. Japanese anti-tank tactics called for an ambush, preferably where terrain limits the mobility of the tank and slows it down. In an ideal engagement, the tanks supporting infantry would be pinned, or forced to withdraw. Then the tank would be immobilized by mines, or whatever was on hand. Then the crew of the tank would be forced to dismount. One such tactic suggested for this was to attack the tank with Kaenbin, although other weapons, such as the Type TB gas grenade could be used.

Japanese soldiers attack a M3 Stuart tank with a Kaenbin, or similar device. From the quality of the picture, and the utter lack of covering fire against an exposed enemy, it suggests that this is a propaganda picture.

With the tank unmanned and immobilized, it could be destroyed, or booby-trapped at leisure by engineers. Of course, if it was the only weapon the Japanese infantryman had, he would go straight to the attack with the Kaenbin, although success was unlikely. Even in the last days of the fighting at Nomonhan, the Japanese reported that Russian tanks had tarpaulins draped over their rear decks to render the Kaenbin ineffective.


Drea, E. J. (1981), Leavenworth Papers: Nomonhan. Fort Leavenworth: Combat Studies Institute.URL: (accessed 1/1/2021)
Coox, A. D. (1985), Nomonhan: Japan Against Russia, 1939. Stanford: Stanford University Press.ISBN: 0804718350.
Japanese tank and Anti-tank warfare (1945) Washington: United States Government Printing Office. Series #34. URL:,_No._34_Japanese_Tank_and_Antitank_Warfare_1945.pdf (accessed 1/1/2021)
Taki’s Home Page (2004) Imperial Japanese Army Page Available at: (accessed 1/1/2021)

2 replies on “Kaenbin”

We were thought in the Belgian army to add washing powder to the fuel, to make the liquid more gooey. I think the sand in this Japanese molotov cockttail had the same function, increasing viscosity and not really adding weight to the bottle.
The rest of the recipe I will skip, as I do not want to be blocked for terrorist urgings.

Leave a Reply

Your email address will not be published.