On August 28, 1940, a ship called the American Legion tied up in New York harbor carrying a strange piece of cargo: a crated twin 40-mm anti-aircraft gun built by the Swedish firm Bofors, along with spare parts, 3,000 rounds of ammunition, and a Swedish engineer named Janson. The gun had traveled an extraordinary route. With Norway under German occupation, it had been hauled across Finland to the Arctic port of Petsamo and loaded there. The United States Navy had bought the weapon to find out whether it deserved its international reputation.
Within weeks of firing it for the first time at Dahlgren, Virginia, nearly everyone who watched it agreed that it did. The drawings that accompanied the gun were another matter entirely. The engineers who studied them found notes scrawled across the prints: “File to fit at assembly” and “Drill to fit at assembly. ” Many of the dimensions on the paper did not match measurements taken from the gun itself because the design deliberately left gaps for a skilled workman with a file to take up by hand.
The prints described a weapon that was finished at a workbench, one gun at a time. Each piece was, in effect, handmade. For a mass-production line, this was the whole problem. Before the Chrysler Corporation could build the gun in Detroit, its engineers had to create a new set of drawings from which any part could be cut to a specified number and fitted to any other part without a file.
The Navy’s own history later placed a figure on the first stage of this work alone: 8,500 man-hours at Chrysler just to transpose the drawings from British to American standards. The first senior American officer to come home sold on the gun was a Navy gunnery specialist, Captain William Blandy. In August 1940, he flew to Trinidad to watch a Dutch warship, the Van Kinsbergen, fire its Bofors guns under Dutch fire control. The island had no target planes, so the cruiser Tuscaloosa traveled 1,500 miles from Guantanamo Bay to launch aircraft that could tow them.
The Navy’s history later called it about as international an affair as it could be: American planes towing targets for a Dutch ship firing Swedish guns off a British island. Blandy returned unimpressed by the Dutch fire control system but full of enthusiasm for the gun itself. The British wanted him to choose their own weapon. The Royal Navy’s 2-pounder pom-pom was already performing well at sea, and London offered help with production.
But the pom-pom fired a slower shell and ran on cordite, a British propellant the American Navy could not replace with its own powder. The choice narrowed quickly to the Bofors. Blandy asked the Dutch naval attaché in Washington for the gun’s drawings. Holland had been overrun that spring, and no prints had reached London, but a set had survived in the Dutch East Indies at Surabaya on Java.
That September, the drawings were microfilmed and flown to Washington. The Navy gave the job of converting them to the York Safe and Lock Company of York, Pennsylvania, a firm best known until then for bank vault doors. The Army approached the problem from a different direction. In November 1940, the British supplied it with a single Bofors, complete with sights and a British crew.
The crew fired it at Aberdeen Proving Ground on November 25 and again at Fort Monroe on December 9. Then the gun went to Detroit, where the same men took it apart on the factory floor. The Army asked Chrysler to prepare detailed drawings from the British prints and from the physical pieces in front of it, and then to build two pilot guns from those drawings. The initial price for that job was $66,000.
At the same time, the Army hired Firestone, the tire company, to do the same work for the gun carriage, converting every metric measurement into inches and every thread into American standard. Now there were two translations of one Swedish gun. York, working from the Dutch prints for the Navy, converted the metric dimensions into decimals of an inch. Chrysler, working from the British prints for the Army, converted them into fractions.
The two versions did not agree. When Navy historians later counted, about 200 parts on the two versions were not guaranteed to interchange. There remained the small matter of the right to build the gun at all. The Swedish government first demanded licenses for American aircraft engines and airframes in return, and the talks stalled until late April 1941.
Two discoveries seem to have ended the standoff. The Swedes found that they held an American patent on almost nothing in the gun. And they discovered that the American Army and Navy intended to build it with or without a deal. The two sides signed an agreement in Stockholm on June 24, 1941, for a total of $600,000, with $100,000 of that dependent on Bofors sending two production engineers to America for a year.
The engineers never came. The United States paid the remaining $500,000 and its own engineers solved the gun’s problems without them. The man tasked with settling the dispute between the two sets of drawings was Blandy, who in February 1941 had become chief of the Bureau of Ordnance. On November 14, 1941, three weeks before Pearl Harbor, he wrote to the Army that every Chrysler part common to both Army and Navy guns would be made to the Army’s prints and inspected with the Army’s gauges.
He accepted that about one part in ten on Chrysler’s naval guns would not fit York’s naval guns, and he judged the chance of trouble from that to be remote. Some Navy men had argued for York’s set because the Navy had already checked it. Blandy chose the drawings from Detroit. Chrysler finished its first production gun mechanism on February 5, 1942.
The men who built it had spent their careers on automobiles, and they approached the Swedish design the way they would have approached a car. The milling on the breech was split into seven setups instead of one. Dozens of small curved corners were standardized so a machinist did not have to change tools for every curve. The flash hider, the bell-shaped piece at the muzzle, had originally been machined out of a solid 40-pound forging to end up weighing only ten.
Chrysler’s engineers replaced it with a length of tube narrowed at one end by a die ring and flared at the other by a punch, and the change saved the company nearly 5,000 man-hours a month on that single part. The Navy placed its first quad mount aboard the old training battleship Wyoming on June 22, 1942, and its first twin mount aboard the destroyer Coghlan nine days later. Somewhere in those months, an American production engineer was said to have remarked that the Bofors must have been designed to cure the unemployment problem. The remark is repeated in gun histories with no name attached, and no source says who made it or where.
The company’s own account of the work, sent to the War Department, boasted that its gun fired 120 rounds a minute, a considerable improvement on the original. The Army’s historian added a footnote pointing out that the original Bofors had fired 120 rounds a minute too. The Navy’s drawings fared worse. York Safe & Lock kept the master set for the naval version, and by 1943 the Bureau of Ordnance feared its drawings threatened the entire program.
Dimensions were wrong, revisions had not been entered, and some sheets carried no tolerances at all. By one estimate, no more than one sheet in five could be trusted completely. Subcontractors complained endlessly. In August 1943, the Navy took the entire set away from York and gave it to the naval gun factory in Washington, which reviewed more than 2,200 sheets.
In October, York failed to deliver a single complete mount because it was short about 25 parts. In January 1944, under an executive order, the Navy took over the plant. The Detroit set, by contrast, had been drawn under a supplement to the very first order on tracing linen. The difference appeared where the guns were assembled.
A British officer from Sheffield toured the Lynch Road plant in Detroit and reported that British factories took between 260 and 340 hours to assemble one gun. At Lynch Road, the same work took fourteen. The reason, he said, was that American precision made hand fitting unnecessary. The file had been designed out of the gun.
This was the capability Japanese kamikaze pilots faced in the Pacific. At first, the Navy took about 15 percent of the gun mechanisms built in Detroit, water-cooled versions for twin mounts on ships. In January 1944, the Army cut back its own orders for the gun, freeing those production lines to build for the Navy. The Bureau of Ordnance later tried to credit kills to particular guns, and of the enemy planes it could assign to a specific caliber between October 1, 1944, and February 5, 1945, about half fell to the 40-mm.
On the morning of May 11, 1945, the destroyer Evans was on radar picket station 15, northwest of Okinawa, when the Japanese attacked in waves. Her action report reads like a railroad timetable. At 8:39, a Zeke came down in a steep dive from high on the starboard quarter, was taken under fire by the 5-inch battery at 7,000 yards, by the 40-mm at 3,000, and was shot down in flames at 2,000. Two minutes later, another Zeke came in the same way and was knocked down by the 40s and 20s, hitting the water 800 yards off the beam.
Before the morning was over, the Evans had destroyed fourteen planes and held down three more, and then four kamikazes hit her in quick succession. She lost thirty-two men. Nearby, the destroyer Hugh W. Hadley took a 250-pound bomb squarely on her port quad 40-mm mount, and all nine men at that gun died.
By the end of that morning, the Hadley’s gunners had shot down more planes than any American surface ship managed in a single engagement, nineteen or twenty by the Navy’s count, before a bomb and three kamikazes left her dead in the water. The Evans was towed into Kerama Retto, an anchorage off Okinawa whose capture in late March had been part of an operation commanded by Admiral Blandy. He had left the Bureau of Ordnance in December 1943 to command an amphibious force at sea, ran the preliminary bombardment at Iwo Jima and the assault on Kerama Retto, and after the war commanded the atomic tests at Bikini Atoll. He retired as a full admiral in 1950 and died in January 1954.
Blandy was also the man who demolished the favorite version of how the drawings reached America, the story in which they were flown out of Java just ahead of the Japanese. He called it a lovely story but untrue. The microfilm had left Surabaya in the autumn of 1940, more than a year before Japan attacked anyone in the Dutch East Indies. Even the Bofors was not enough by the spring of 1945.
A 40-mm shell could break up a plane only if it hit squarely, and a kamikaze that had already been hit would often keep coming. The proximity fuse that made the 5-inch guns so effective would not fit inside a shell that small. The Navy began a crash program for a rapid-fire 3-inch gun to replace the Bofors, but it did not reach the fleet before the war ended. The Army could have had the weapon years earlier.
In 1937, its Ordnance Department asked to buy one for testing and was told that Bofors did not sell guns for demonstration, on principle. In May 1938, the company changed its mind and offered to send a pilot gun for trials along with a company officer, an engineer, and a mechanic, free of charge. The cable quoting prices for ammunition had almost no punctuation, and it placed the word dollars only near the end. The Ordnance Department misread a price of $10 followed by a quantity of 1,000 as 10,000 shells, believed it was being asked to buy more than 20,000 rounds at $12 apiece, and turned the offer down as impractical for a peacetime test.
The Army’s own historian, writing in May 1944, concluded that the misreading delayed for about two more years the adoption of the greatly superior Bofors. The explanation was filed in a footnote.