February 8th, 1942. Makassar Strait, 22 miles south of the port of Makassar on the island of Celebes. It is a little after half past eight at night. Four Japanese destroyers are steaming in column, screening an invasion convoy, and one of them is Natsushio. She is 18 months old. She displaces 2,000 tons, makes 35 knots, and carries six 5-inch guns and eight tubes for the best torpedo in the world. Eight hundred yards off her beam, on the surface in the dark, is USS S-37. She was commissioned in July 1923. She is 219 feet long, she has four torpedo tubes, all in the bow, and 42 men aboard. The weapons in those tubes are older than she is. They are Mark 10 torpedoes, a design that went to sea during the First World War. The Mark 10 is slow. It is short-ranged. It has no magnetic exploder, no second speed setting, nothing that the Bureau of Ordnance considers modern. The Navy stopped developing it more than a decade ago. Lieutenant James Dempsey fires one torpedo at each destroyer in the column. Thirty seconds after the third one leaves the tube, it strikes Natsushio between her stacks. The destroyer buckles in the middle and folds into a V. She is the first Japanese destroyer sunk by an American submarine in the Second World War. In those same weeks, the newest submarines in the American fleet, armed with the newest torpedo the United States had ever built, were firing spread after spread at Japanese ships and watching nothing happen. This is how a leftover from 1918 kept sinking ships while its replacement failed, and why it took the Navy 21 months to find out what was wrong.

On December 8th, 1941, the United States Asiatic Fleet had 29 submarines based at Manila. Twenty-three of them were modern fleet boats. Six were S-boats, numbered S-36 through S-41, all commissioned in the early 1920s. War Plan Orange had always assumed the surface fleet in the Philippines could not stop a Japanese invasion. The submarines were the defense. Twenty-nine boats, positioned across the approaches to Luzon, were supposed to bleed the invasion convoys before they reached the beaches.

It did not happen. Two days into the war, Japanese bombers destroyed the Cavite Navy Yard, and 233 torpedoes went with it. On December 22nd, more than 70 Japanese transports anchored in Lingayen Gulf and put an army ashore. Across the whole of December, the Asiatic Fleet submarines made roughly 45 attacks, fired close to a hundred torpedoes, and sank three ships. The Philippines fell. The boats retreated to Java, and then to Australia.

The captains came home with the same story. Lieutenant Commander Tyrrell Jacobs of USS Sargo made attack after attack in December with good setups and close ranges. He fired 13 torpedoes on that patrol. Not one exploded against a ship. Jacobs was a torpedo specialist with postgraduate ordnance training. He checked his fire control, found nothing wrong, and broke radio silence to tell his command that the torpedoes were defective. His command told him the fault was his.

So by February 1942, the Navy's position was that its submarine captains could not shoot. And that made the results coming from the six oldest boats in the fleet very hard to explain.

To understand why nobody expected anything from the Mark 10, look at what the Navy had built to replace it.

The Mark 10 was designed in 1915 by the E.W. Bliss Company and the Naval Torpedo Station at Newport, Rhode Island. It was 21 inches in diameter, just over 16 feet long, and weighed 2,215 pounds. It carried 497 pounds of TNT. An alcohol and water steam turbine drove it at 36 knots to a maximum of 3,500 yards. It had one speed. It had one way of exploding, a Mark 3 contact exploder in the nose, a simple inertia mechanism that fired when the torpedo ran into something solid. To sink a ship with a Mark 10, a captain had to get inside two miles, and realistically inside one, and physically hit the hull.

The Mark 14 was everything the Mark 10 was not. Design work started at Newport in 1931. The new torpedo was 20 feet 6 inches long and weighed over 3,000 pounds. It had two speed settings: 46 knots to 4,500 yards, or 31 knots to 9,000. Its warhead began at 507 pounds of TNT and grew to 643 pounds of Torpex. Each one cost about 10,000 dollars, at a time when a new family house cost less than that.

And in its nose sat the most secret weapon in the American arsenal, the Mark 6 exploder. The Mark 6 had been in development since 1922. It carried a contact trigger, but its real purpose was magnetic. It was designed to sense the steel hull of a ship overhead and detonate the warhead underneath the keel, where there is no armor belt and no torpedo bulge. One hit under the keel would break a ship's back. On paper, the Mark 6 made every battleship's side protection irrelevant.

The Mark 6 was so secret that most submarine officers never saw one before the war. The manuals were locked in safes. Crews were trained on the older exploders and handed the Mark 6 when the shooting started.

One viewer of this channel asked a fair question under the S-44 episode: if the submarine torpedo went from Mark 10 to Mark 14, what happened to Marks 11, 12, and 13? The answer is that they were never submarine weapons. The Mark 11 and Mark 12 were destroyer and cruiser torpedoes of the late 1920s. The Mark 13 was the aircraft torpedo. For submarines, the line jumped straight from a 1915 design to a 1931 design, with nothing in between.

The S-boats could not carry the new weapon at all. Their tubes were too short for a torpedo 20 feet 6 inches long. They went to war with the Mark 10 because nothing else would fit. In the view of the Bureau of Ordnance, that made them the worst-armed submarines in the Navy.

The Bureau was looking at the wrong column of the ledger. Here is what it missed.

The Mark 10 had been in the fleet for more than 20 years. Thousands had been fired in exercises, year after year, by crew after crew. Its faults were known. It did have one: the Mark 10 ran about four feet deeper than the depth set on it. But the Bureau worked that out and told the submarine force in the first weeks of 1942, and four feet was an error that a careful captain could correct with a dial. Set it shallow, and it ran where you wanted.

The Mark 14 had almost no history at all. Through the 1930s the Navy built torpedoes slowly and treated each one as too valuable to destroy. In 1937, Newport was producing about one and a half torpedoes a day. A live test meant blowing up 10,000 dollars of government property and a target ship besides. So the Mark 6 exploder had exactly one live test before the war, in 1926, against the hulk of an old submarine. The first shot ran under the target and did nothing. The second one worked. That was judged sufficient. In the 1930s the cruiser Indianapolis fired about a hundred instrumented trial shots near the equator to measure magnetic fields, with no warheads. No production Mark 14 with a production Mark 6 was ever fired at a real hull and allowed to explode. Not once.

Depth tests were run with exercise heads that weighed less than the real warhead. A torpedo with a heavier nose runs deeper. Nobody checked.

That produced three separate defects, each one hiding the next.

First, the Mark 14 ran about 10 feet deeper than set. A torpedo set to pass just under a destroyer's keel went so far beneath it that the magnetic exploder never felt the hull. A torpedo set to hit a freighter's side passed clean under the ship.

Second, the magnetic exploder itself had been calibrated in the waters off Rhode Island. The Earth's magnetic field is a different shape in the western Pacific. There, the Mark 6 tended to fire early, 50 yards short of the target, throwing up a column of water that looked like a hit through a periscope. Or it did not fire at all.

Third, and buried under the other two, was the contact trigger. It was essentially the mechanism inherited from the Mark 10's generation, where it had worked for decades at 36 knots. At 46 knots, a square hit against a ship's side stopped the torpedo so violently that the firing pin's guides deformed and the pin jammed before it reached the primer. The better the shot, the more likely the dud.

And that is the hidden advantage of the old weapon. The Mark 10 had no magnetic feature to misfire. Its depth error was small and known. And it hit its target 10 knots slower, inside the speed its simple exploder had been built for. Every part of it had been fired in anger or in practice for a generation. It was the only proven submarine torpedo the United States Navy owned in 1942.

The boats that carried it had a hidden advantage of their own. A fleet boat captain with a 9,000-yard torpedo and a torpedo data computer could shoot from a distance, and doctrine encouraged him to. An S-boat captain had a 3,500-yard torpedo and fire control not much advanced from a slide rule. He had no choice except to close until he could not miss. Eight hundred yards. A thousand yards. At those ranges a slow torpedo arrives before the target can turn.

Now the record.

December 22nd, 1941, Lingayen Gulf. While the fleet boats outside reported misses, Lieutenant Wreford Chapple took S-38 into the gulf itself just before dawn, into shallow water full of anchored transports and patrolling destroyers. At 7:10 in the morning he fired all four bow tubes at a convoy from about a thousand yards. All four missed. Chapple judged that they had run under shallow-draft targets, reloaded, and set the next torpedoes to run shallower. At 7:58 he fired two at an anchored transport. Less than a minute later, the 5,445-ton Hayo Maru blew up. S-38 then spent most of the day on the muddy bottom at 80 feet while destroyers hunted her. She was one of the very few American submarines to sink anything at Lingayen. Notice what Chapple did. He missed, guessed the cause, turned a dial, and hit. With a Mark 10, that worked, because there was only one thing wrong with it.

February 8th, 1942. S-37 sighted destroyer masts at five miles just after six in the evening. Dempsey tracked them for more than two hours in the dark, first hoping for the transports behind them, then settling on the screen itself. Between 8:36 and 8:40 he fired four torpedoes, one at each destroyer, from 800 yards. The fourth destroyer saw the submarine and turned away. The third did not. Natsushio took the hit between her funnels, lost ten of her crew, and did not survive the night. Her sister Kuroshio took off the survivors. S-37 dived and sat under three destroyers for an hour and a half, counting depth charges at ten to fifteen minute intervals. She was leaking through her engine room hatch for the rest of the patrol. Dempsey received the Navy Cross.

March 4th, 1942. S-39 sank the 6,500-ton fleet oiler Erimo south of Borneo.

Then August. On the 8th, S-38, now under Lieutenant Commander Henry Munson, fired two torpedoes at the transport Meiyo Maru off Cape St. George. Both hit. She was carrying troops to reinforce Guadalcanal. Two days later, S-44 put three of four Mark 10s into the heavy cruiser Kako from 700 yards, as this channel has covered. Kako was the first major Japanese warship sunk by an American submarine in the war.

Line those up. A transport at Lingayen. The first destroyer. A fleet oiler. A troopship. A heavy cruiser. Every one of them fell to a boat that was 17 to 19 years old, firing a torpedo designed in 1915.

And in the same months, the modern force was proving the other half of the case, the hard way.

Rear Admiral Charles Lockwood took over the submarines at Fremantle in the spring of 1942 and did what the Bureau of Ordnance had not done in ten years. He bought a fishing net. On June 20th, 1942, in Frenchman Bay near Albany, Western Australia, USS Skipjack fired a Mark 14 set to run at 10 feet through the net from 850 yards. The hole was at 25 feet. Two more shots under Captain James Fife averaged 11 feet deep. The Bureau rejected the tests as unscientific. Lockwood ran them again. On August 1st, 1942, the Bureau conceded that the Mark 14 ran deep. It had taken eight months of war.

Fixing the depth only exposed the magnetic exploder. With torpedoes now running close under their targets, premature explosions multiplied. Captains reported warheads detonating halfway to the ship. On June 24th, 1943, Admiral Nimitz ordered the magnetic feature switched off in every Pacific Fleet submarine. In Australia, Rear Admiral Ralph Christie, who had helped develop the Mark 6 at Newport, kept his boats using it until Admiral Kinkaid ordered him to stop at the end of the year.

And switching off the magnet exposed the third fault. July 24th, 1943, west of Truk. Lieutenant Commander Dan Daspit in USS Tinosa found the Tonan Maru Number 3, a 19,000-ton whale factory ship converted to a tanker, one of the largest ships in the Japanese merchant fleet, steaming alone. He fired four torpedoes from 4,000 yards at an awkward angle. Two hit and exploded, and the tanker stopped. He fired two more. They hit, and she was dead in the water. Daspit then moved Tinosa to 875 yards, directly off her beam, the textbook position, and fired nine torpedoes one at a time, deliberately, watching each one through the periscope. Every one ran straight. Every one hit. None exploded. He took his last torpedo home to Pearl Harbor as evidence.

Lockwood's staff fired live Mark 14s into the cliffs of Kahoolawe on August 31st, 1943, recovered a dud, and found the crushed firing pin. Then they filled warheads with sand and dropped them from a crane 90 feet onto a steel plate, a height chosen to match a 46-knot impact. At a square 90-degree hit, seven out of ten exploders failed. At a glancing angle, the failure rate dropped by half. The glancing shots that Tinosa opened with had worked. The perfect ones had not.

The fix was a lighter, stronger firing pin. The first batch was machined at Pearl Harbor from the toughest light alloy available on the island, which turned out to be the propeller blades of Japanese aircraft shot down on December 7th. New exploders reached the boats in September 1943.

That was 21 months after Pearl Harbor. For 21 months, the United States fought a submarine war in which its best captains in its best boats were firing a weapon with three independent defects. Through that period the Bureau blamed the crews.

The measurable proof is in what happened on either side of that date. In 1942, American submarines sank about 180 Japanese ships. In 1944, with a working torpedo, they sank over 600. By the end of the war the submarine force, less than two percent of the Navy's personnel, had sunk roughly 1,300 ships and more than five million tons, over half of all the Japanese shipping destroyed by any means.

The Mark 10 did not do that. It is important to be exact about what the old torpedo was. It was slower, shorter-ranged, and lighter in the warhead. The S-boats that carried it sank a modest number of ships, and by the end of 1943 most of them had been pulled back to training duty. Several fleet boats carried Mark 10s in their stern tubes during the 1942 torpedo shortage, and were glad to have them, but nobody proposed going back. A fixed Mark 14 was the better weapon by every measure, and it proved it. It stayed in American service until 1980.

What the Mark 10 provided was the evidence. Through 1942 it was the control group in an experiment the Navy did not know it was running. Same ocean. Same enemy. Same training pipeline for the officers. Older boats, worse fire control, worse living conditions, and a torpedo from the First World War. And those boats fired torpedoes that exploded. When a captain with a Mark 10 missed, he could work out why. When a captain with a Mark 14 missed, he was told to shoot better. The S-boat record was one of the reasons men like Lockwood stopped believing that.

S-37 never sank another warship. She made five more patrols, sank the freighter Tenzan Maru in July 1942, and finished the war as a training boat. In February 1945, stripped and under tow to be used as a bombing target off San Diego, she broke her cable and sank on her own.

February 8th, 1942. Eight hundred yards from a column of four destroyers. A 19-year-old submarine with a leaking hatch fires four torpedoes designed before America entered the First World War, and the newest destroyer class in the Imperial Navy loses its first ship. The Bureau of Ordnance had a torpedo that was better in speed, range, warhead, and exploder, and it had fired that torpedo at a real target one time in 15 years. The old one had been fired thousands of times. Its flaws were small, and every one of them was written down. The Mark 10 was obsolete. It was also the only torpedo in the submarine force that anyone had ever bothered to prove. In 1942, that was the specification that mattered.