The de Havilland Comet, the Square Windows, and the Metal Fatigue Investigation That Changed How Every Pressurized Airplane Is Built

The de Havilland Comet became the world’s first commercial jet airliner on May 2, 1952, and within two years it had killed 56 people and grounded itself permanently in its original form. The investigation that followed - centered on a water tank at Farnborough, England - discovered that the Comet’s roughly square windows were concentrating stress at their corners with every pressurization cycle, growing invisible fatigue cracks until the fuselage catastrophically failed. That finding gave every pressurized aircraft flying today its oval windows.

Britain’s Lead in the Jet Age

Geoffrey de Havilland had been building aircraft since before the First World War. His company produced the Tiger Moth trainer and the Mosquito, the fast plywood twin that outran most Luftwaffe aircraft. During the final years of World War II, the Air Ministry asked British manufacturers to think ahead - a jet transport, pressurized, long range.

The prototype, designated the DH 106 Comet, first flew on July 27, 1949 - four years before the Boeing 707 made its first flight and three years before Douglas had a comparable design on paper. Britain was first. Genuinely, unambiguously first.

What the Comet Was

The production Comet carried approximately 40 passengers, powered by four de Havilland Ghost turbojet engines buried cleanly into the wing roots - no propellers, none of the vibration prop-liner passengers endured for hours. Cruising altitude was 35,000 feet, above the weather systems that punished piston-engined routes. Speed approached 500 miles per hour, cutting route times nearly in half.

The cabin was pressurized to the equivalent of 8,000 feet altitude - extraordinary engineering for the early 1950s. The windows were large and nearly square, with rounded corners, big enough for passengers to see the curvature of the earth on a clear day. Airlines could not order fast enough. Air France, UAT, and Japan Air Lines all committed to the type.

BOAC launched the first commercial jet service on May 2, 1952: London to Johannesburg. For two years, the Comet carried passengers in a style and at a speed the world had never seen. Geoffrey de Havilland’s bet was paying off.

The Two Fatal Accidents

Earlier incidents - two Comets lost in takeoff accidents during initial service - were treated as the expected learning curve of introducing a genuinely new type. De Havilland refined training procedures. The Comet kept flying.

On January 10, 1954, BOAC Flight 781 departed Ciampino Airport in Rome bound for London. The aircraft, registered G-ALYP, carried 29 passengers and a crew of 6 - 35 souls aboard. Twenty minutes after takeoff, climbing through the upper twenties in clear sky over the Tyrrhenian Sea, the aircraft came apart without warning. There was no distress call, no mayday, no crew report of anything wrong. The wreckage fell near the island of Elba. All 35 were killed.

Investigators pursued every theory: engine failure, turbulence-induced overload, military collision, explosive decompression, sabotage. The Comets were grounded for inspection, modifications were made, and in late March 1954 the type returned to service.

On April 8, 1954, a second Comet departed Rome on a southbound service to Johannesburg, operated that day by South African Airways, with 14 passengers and 7 crew - 21 souls. Near Naples, the aircraft disappeared. Again, no distress call. The Comet was grounded worldwide that afternoon and never flew passengers again in its original configuration.

The Farnborough Investigation

Arnold Hall, director of the Royal Aircraft Establishment (RAE) at Farnborough, ran two simultaneous investigations.

The first was underwater recovery. The Royal Navy sent divers and salvage ships to the floor of the Mediterranean and recovered whatever remained of G-ALYP. That wreckage was brought to Farnborough and reassembled on a skeleton framework, matching torn metal edges and logging each fragment’s position - working backward through the breakup sequence to find where it had started.

The second track was the water tank. Hall’s team built a tank large enough to hold a complete Comet fuselage, submerged it, and pressurized it from inside repeatedly - simulating the pressurization cycle of a normal flight. Thousands of simulated flights, run over weeks.

Why Water Instead of Air

The choice of water over compressed air was methodologically critical. Air stores enormous energy under compression; if a pressurized metal structure fails with air inside, the result is an explosion. The failure propagates so fast that fracture surfaces are obliterated - impossible to read, impossible to trace back to an origin point.

Water is incompressible. It stores no significant energy in compression. When a water-pressurized fuselage fails, it tears and deforms rather than exploding. Fracture surfaces remain intact. Investigators can see exactly where the crack started and how it grew.

After enough simulated cycles, the test fuselage failed. Under magnification, the fracture surfaces revealed fatigue cracks - microscopic fractures in the aluminum skin that had grown a tiny amount with each pressurization cycle, invisible to any external inspection, growing silently until the structure could no longer hold.

The cracks had started at the corners of the windows.

Why Corners Fail

The fuselage skin in flight carries significant tensile load. At cruise altitude, cabin pressure inside is substantially higher than ambient pressure outside; the skin wants to expand outward. That load must route around every opening in the structure - every window, every cutout. At a circular or oval opening, stress flows smoothly around the curve and distributes evenly, the way water routes around a rounded stone.

At a corner, there is no curve. The stress concentrates at the right-angle point, stacking many times higher than the stress in the surrounding skin. With every pressurization cycle - every single flight - the metal at that corner works far harder than anywhere else on the fuselage. Metal that works hard repeatedly fatigues.

On the actual accident aircraft, investigators traced the critical failure to the corner of a rectangular antenna cutout in the upper fuselage skin, forward of the main passenger cabin. The crack had grown through enough cycles until rapid fracture became inevitable. Over the Tyrrhenian Sea on January 10, 1954, it let go. Once it started, it finished in less time than a human thought.

The Fix, and Its Legacy

The solution, once the cause was understood, was geometric: remove the corners. Make window openings oval or round. Any shape that allows stress to flow evenly around the perimeter without concentrating at a point.

Every jet airliner flying today has oval windows. That shape is not aesthetic. It is not coincidence. It is the direct consequence of what Hall’s team found in that water tank at Farnborough in 1954.

Boeing absorbed every lesson the RAE published. When the Boeing 707 entered service in October 1958, it incorporated oval windows, fatigue test requirements built into certification standards, and a design philosophy around pressure cycling that had not existed before the Comet disasters. Douglas built the same thinking into the DC-8. Every pressurized transport aircraft certified since has been required to demonstrate fatigue life and damage tolerance standards that trace directly to what Farnborough found.

The Comet 4 - Too Late

De Havilland rebuilt the Comet. The Comet 4 entered service with BOAC in autumn 1958: oval windows, thicker skin, structure tested to standards that hadn’t existed when the original was designed. BOAC actually beat the 707 into transatlantic jet service by a matter of weeks.

But the American manufacturers had watched the Comet’s disasters and absorbed every published lesson without paying the price in lives or lost production time. They arrived with aircraft that were larger, longer ranged, and already on order at every major airline. The lead Britain had spent years building was gone by the time the Comet 4 was airborne.

Why This Matters for Pilots Today

Every time a pressurized aircraft climbs through the flight levels, the number of pressurization cycles accumulates in the maintenance logbook alongside flight hours and calendar time. Mandatory structural inspections come due not just on the calendar but on the number of pressurization events the airframe has seen. Damage tolerance requirements, fatigue testing standards, the oval geometry of every passenger window - all of it traces to knowledge bought at the cost of 56 lives lost over the Mediterranean in 1954.

The passengers and crew aboard G-ALYP were flying on the most advanced airliner in the world. Neither they nor the engineers who built it knew what square window corners would do over thousands of pressurization cycles. No one did, because no one had ever built anything like it and operated it long enough to find out.

That knowledge now lives inside every pressurized aircraft built since.

Key Takeaways

• The de Havilland Comet inaugurated commercial jet travel on May 2, 1952 - years ahead of American competitors - but two fatal crashes in early 1954 grounded it permanently in its original form, killing 56 people.
• Investigators at Farnborough discovered that square window corners concentrate fuselage skin stress at a right angle, generating fatigue cracks that grow invisibly with every pressurization cycle.
• The water tank test - pressurizing a submerged fuselage thousands of times - was the methodological breakthrough that preserved fracture surfaces and made the failure mode readable.
• The fix was geometric: oval windows allow stress to flow smoothly around the opening without concentration at any point. Every commercial jet flying today uses this geometry for exactly this reason.
• The investigation established fatigue life and damage tolerance certification requirements now fundamental to every pressurized transport aircraft worldwide - a standard that did not exist before the Comet disasters.