The de Havilland Comet: the first jet airliner and the 1954 crashes

Published

2026-11-06

Updated

2026-11-06

Facts checked

2026-10-26

On 2 May 1952 a BOAC Comet left London for Johannesburg, the first jet airliner in scheduled passenger service. Less than two years later two Comets broke up over the Mediterranean, off Elba and near Naples, and every Comet was grounded. Investigators at Farnborough found the cause: metal fatigue at the corners of cut-outs in the pressurised cabin. The redesigned Comet 4 opened jet service across the Atlantic in 1958, only 22 days before Pan Am's Boeing 707.

A jet airliner before anyone else

After the war de Havilland chose not to build another propeller airliner. With experience of jet fighters and an engine company of its own, it set out to build a civil aircraft that would be, in the words of the 1955 court of inquiry, "a step ahead of the current type". Design work began by the end of September 1946. BOAC and the Ministry of Supply did not wait for a prototype: they signed a contract for Comets on 21 January 1947 (Court of Inquiry, 1955).

The first prototype flew on 27 July 1949. The chief test pilot, John Cunningham, took it up from Hatfield aerodrome in Hertfordshire for 31 minutes and reached 8,000 feet. It was Sir Geoffrey de Havilland's 67th birthday and Cunningham's 33rd (This Day in Aviation).

The jet engine changed the rules of airliner design. To carry a paying load over a useful distance at jet speed, the court explained, the Comet had to cruise above 35,000 feet, double the height of the airliners of the day, and its structure had to weigh as little as possible. At 40,000 feet the cabin was kept at the pressure of 8,000 feet. The difference between the air inside and outside reached 8.25 pounds per square inch, about 50% more than was then usual.

British and ICAO rules called for a proof test at 1.33 times the working pressure difference and a design strength of twice the working pressure. De Havilland designed the fuselage for 2.5 times and tested it to twice the working pressure (FAA). The company believed, and the Air Registration Board and other experts agreed, that a cabin which survived double its working pressure would not fail from fatigue in service.

2 May 1952: the first jet passengers

The Comet received its full certificate of airworthiness early in 1952. On 2 May G-ALYP entered scheduled passenger service, the first jet-propelled passenger aircraft in the world to do so. It left London at 15.12 GMT with 36 passengers on a journey of about 7,000 miles to Johannesburg, with stops at Rome, Beirut, Khartoum, Entebbe and Livingstone. The timetable allowed 23 hours 40 minutes, and the Comet landed in Johannesburg at 14.38 GMT on 3 May, fourteen minutes early (This Day in Aviation).

Captain Michael Majendie was at the controls as the Comet left London with 30 bags of mail aboard, Smithsonian magazine records. The Comet cruised at 35,000 feet and 460 miles an hour, more than 100 miles an hour faster than the fastest propeller airliner (Smithsonian, 2002). Its 36 seats were split between two cabins, and in first class passengers sat around tables, as in a railway carriage. "It was the Concorde of its day," Alistair Hodgson, curator of the de Havilland Aircraft Museum, told BBC Future: it flew higher, faster and smoother than any other airliner of its time (BBC Future, 2017).

The business case looked good too. Four months in, TIME reported, the Comet flew to Johannesburg three times a week, the waiting list was a month long, and it had made an operating profit of £16,000 (TIME, 1952). American airlines took notice. Juan Trippe of Pan Am had a ride, and after a 90-minute flight Eddie Rickenbacker of Eastern Air Lines said he would put up $100 million for 35 Comet IIIs if de Havilland could deliver them by 1956. In October 1952 Pan Am ordered three Comet IIIs for an estimated $6.3 million; according to Smithsonian magazine, it was the only American airline to order the Comet (TIME, October 1952). What Pan Am did next is told in October 1958: Pan Am, the 707 and America's first transatlantic jet service.

Warning signs, 1952-1953

The first accidents happened on take-off. On 26 October 1952 the Comet G-ALYZ ran off the end of the runway at Rome; there was no fire, and none of the 35 passengers or crew were hurt. Peter Duffey, a BOAC pilot who watched it, said the captain had raised the nose too far on the take-off run. In March 1953 a Canadian Pacific Comet crashed on take-off at Karachi and burned, and 11 people died. The investigation found a flaw in the wing, which was cured by new instructions to pilots and a modified leading edge.

The next loss was harder to read. On 2 May 1953, a year after the first service, the Comet G-ALYV broke up near Calcutta in a tropical storm of exceptional severity. It had been climbing for six minutes, and all 43 people aboard were killed (FAA). An Indian court found that the structure had been overstressed, either by violent gusts or by the pilot's handling in the storm. Nobody suspected fatigue of the cabin, and the British court of 1955 saw no reason to doubt the Indian verdict.

De Havilland's own tests seemed to settle the question of the cabin. By mid-1953 no Comet had flown more than 2,500 hours, about 800 pressurised flights. From July the company pressurised a test section of the forward cabin again and again; by September it had withstood 18,000 cycles of working pressure, on top of about 30 earlier loadings at up to double the working pressure. The test ended with a fatigue crack at the corner of a window, which started at a small defect in the skin. But the number of cycles was so large that it was taken as proof that the cabin was safe with an ample margin (Court of Inquiry, 1955).

Elba and Naples, 1954

On 10 January 1954 G-ALYP, known by its call sign Yoke Peter, left Ciampino airport in Rome for London at 09.31 GMT. At about 09.51 the captain was talking by radio to another BOAC aircraft, and his last message broke off: "George How Jig from George Yoke Peter did you get my". The Comet was probably approaching 27,000 feet. Witnesses on the island of Elba saw wreckage, part of it in flames, fall into the sea at about 10.00. All 29 passengers and the six crew under Captain Alan Gibson were killed, among them the Australian war historian Chester Wilmot. An Italian fisherman told TIME: "I heard a roar, very high. Then there was a series of blasts. The next thing I saw was a column of smoke plunging straight down into the sea" (TIME, January 1954).

BOAC took the Comets out of service at once. A BOAC committee under C. Abell went through every possible cause, from control flutter and wing fatigue to engine fire, and thought fire the most likely. The modifications, in the words of the Air Registration Board's chairman, Lord Brabazon, were meant to cover "every possibility that imagination has suggested as a likely cause of the disaster". Meanwhile Royal Navy ships searched the sea bed off Elba, using underwater television for the first time in such a search, and by the end of August 1954 about 70% of the aircraft had been recovered.

The first Comet went back into passenger service on 23 March 1954. Peter Duffey later told BBC Future that the decision had been put to a vote at meetings at Heathrow and carried by a majority of one; he had voted against it. "They gambled and lost; I was one of the chips," he said (BBC Future, 2017). The court of inquiry blamed no one for the decision: at the time it was not the practice, in Britain or anywhere else, to ground a whole type after one unexplained accident.

On 8 April 1954 the Comet G-ALYY (Yoke Yoke), on charter to South African Airways, left Ciampino for Cairo at 18.32 GMT. At about 18.57 the crew reported that they were abeam Naples and climbing to 35,000 feet. At about 19.10, near the top of the climb, something catastrophic happened, and the aircraft fell into the sea south-east of Naples. Fourteen passengers and seven crew were killed. Barely 16 days had passed since the ban was lifted, TIME noted, and searchers found an oil slick some 30 miles north of Stromboli. Before take-off an American passenger had told his companion: "This is progress. Sure, they've had accidents, but everything is O.K. now" (TIME, April 1954).

This time BOAC stopped all Comet services, and on 12 April the House of Commons heard that the Minister had withdrawn the British certificate of airworthiness from every Comet. De Havilland halted all work on the Comet II and III until the cause was found. Thirty-three Comet IIs and eleven Comet IIIs were on order, three of the IIIs for Pan Am. "We know we're in a crisis," a company executive told TIME. "Even if the cause of the crashes were found tomorrow, we would have lost between four and six months" (TIME, May 1954). As Smithsonian magazine quotes him, the Prime Minister, Winston Churchill, declared that "the cost of solving the Comet mystery must be reckoned in neither money nor manpower".

The tank at Farnborough

The Minister of Supply handed the investigation to the Royal Aircraft Establishment at Farnborough and its director, Sir Arnold Hall. On 18 April Hall decided to test the whole cabin under repeated loading.

The cabin was tested with water rather than air. If a pressurised cabin full of air burst, the court explained, it would go off like a 500-pound bomb and destroy the evidence of where the failure began, so the whole fuselage was immersed in a tank (Court of Inquiry, 1955). TIME described the steel tank, 112 feet long, 60 feet high and 20 feet wide, built around the Comet G-ALYU, known as Yoke Uncle. In five minutes the tank could put the aircraft through the fatigue of a typical three-hour flight (TIME, November 1954).

Yoke Uncle had made 1,230 pressurised flights before the test. After the equivalent of 1,830 more in the tank, 3,060 in all, its cabin split, and the failure started at the corner of a cabin window. The FAA identifies it as the squarish window of the forward escape hatch (FAA). Strain gauges then showed that near the corner of a window the stress in the skin probably exceeded 40,000 pounds per square inch at the working pressure difference of 8¼ pounds per square inch, significantly higher than the designers had believed, and the court accepted that this explained a fatigue failure after so few cycles. Yoke Peter had made 1,290 pressurised flights when it was lost off Elba, and Yoke Yoke 900. Dr Walker of the RAE estimated that 3,060 flights in the tank were worth about 2,500 in real service, because the tank left out vibration and other sources of fatigue.

The decisive piece came from the sea. An Italian fishing boat trawled up a section of skin from the top of Yoke Peter's cabin, above the front spar of the wing, with two windows for the aerials of the automatic direction finder, the ADF radio compass. From it the RAE established that the cabin had first fractured near the rear ADF window and that the crack had run forwards and backwards from there. The court judged it probable that the break began at the starboard rear corner of that window, at the edge of a countersunk bolt hole, where experts found that fatigue had been present (Court of Inquiry, 1955).

The verdict was firm for Elba. The RAE concluded that the accident was caused by structural failure of the pressure cabin, brought about by fatigue; for Naples there was no wreckage, but the same explanation appeared to apply. "I unhesitatingly come to the conclusion that R.A.E. were right," wrote Lord Cohen, who chaired the inquiry, and placed the failure "in the region of the A.D.F. window". His report is dated 1 February 1955. When the hearings opened in October 1954, TIME reported, Sir Lionel Heald for the Crown began by citing Nevil Shute's novel No Highway, whose hero predicts that an airliner will crash from metal fatigue. The court did not blame de Havilland. The company had followed what was then good engineering practice, and in the state of knowledge of the time it could not be blamed for making little use of strain gauges.

Square windows or a radio-compass window?

The version most people know is that the Comet was killed by its square windows. Alistair Hodgson of the de Havilland museum names two fatal flaws: a skin made as thin as possible to save weight, and rectangular windows. "If you have a square opening in a sheet of metal, like the skin of an airliner, and you stress it, the place where the crack is going to start is at the corner," he told BBC Future. And when de Havilland resumed work on the Comet II and III in 1955, TIME reported, they were to get thicker skins and oval windows instead of rectangular ones (TIME, 1955).

The record is more exact. The tank fuselage failed at the corner of the escape hatch window. The break-up off Elba began at the rear ADF window in the roof, an opening for radio aerials rather than a window for passengers. That roof section, with fatigue cracks around the ADF hatches, is now object 1993-2155 in the collection of the Science Museum Group (Science Museum Group). The FAA describes the ADF windows as "also squarish" and draws the lesson in general terms: the squarish windows and window frames created high stress concentrations that testing did not reveal. Both accounts point to the same weakness, the corners of openings in a thin pressurised skin. But the crack that destroyed Yoke Peter started at an aerial window.

Why did the 18,000 cycles of 1953 mislead everyone? According to the FAA, de Havilland ran the fatigue test on the same section it had earlier loaded to double the working pressure. The overload cold-worked the metal at the corners and improved its resistance to fatigue, which masked the real weakness. Production Comets had never been overloaded, and by the FAA's estimate they were reaching only about 1,000 cycles. In 1956 the British requirements added a rule that the fatigue test must use a complete cabin not previously used for strength testing (FAA).

The Comet 4 and the lost lead

The first of the redesigned Comets, TIME wrote in 1955, would probably not be ready before 1957. The Americans did not wait. On 15 July 1954 the Boeing 367-80 made its first flight. It became the Boeing 707, with more than three times the passenger capacity of the Comet 1, and entered service in 1958, at the same time as the much smaller Comet 4 (Smithsonian, 2002). In the FAA's reckoning it took four years before the Comet carried airline passengers again, this time as the Comet 4. By then, the Science Museum notes, the Douglas DC-8 and the Boeing 707 dominated the new market for jet airliners.

The Comet 4 still had one first to claim. On 4 October 1958 BOAC opened the world's first transatlantic jet service, between London and New York (Guinness World Records). Two Comet 4s crossed that day in opposite directions. G-APDB, now kept at Duxford, flew from New York to London in a then record 6 hours 12 minutes, half the time of the fastest piston airliners (Duxford Aviation Society). The westbound G-APDC left Heathrow at 8.45 a.m. and, with a fuel stop of an hour and ten minutes at Gander, took 10 hours 20 minutes. Each of the two was fitted for 48 passengers and had four Rolls-Royce Avon engines (This Day in Aviation).

The lead did not last. Just 22 days later, the Duxford museum writes, Pan Am put its bigger and faster Boeing 707s on the New York - Paris route, and Britain's lead was lost for good: 74 Comet 4s were built, against more than a thousand Boeings. Meanwhile Aeroflot had been flying the Soviet Tu-104 between Moscow, Omsk and Irkutsk since 15 September 1956; Guinness World Records counts it as the first sustained jet airline service and the Tu-104 as the second jet airliner to enter service after the Comet (Guinness World Records).

After the Comet

BOAC flew the Comet across the Atlantic until mid-1960 and on other routes until 1965. Charter airlines bought the second-hand aircraft, and Dan-Air, which acquired every flyable Comet 4, became the largest operator with 49 of them. The last flight of a Comet took place on 14 March 1997, and its military descendant, the Hawker Siddeley Nimrod maritime patrol aircraft, served the Royal Air Force until 2011, more than 60 years after the Comet first flew (Duxford Aviation Society).

The bigger legacy was what the industry learned. The Comet, the FAA explains, was designed to a "safe-life" of 10,000 cycles, while Boeing and other manufacturers of the time built aircraft such as the 707 to be "fail-safe", with redundant structure and features to stop cracks. Today both approaches have merged into damage-tolerance standards. TIME wrote in 1954 that Britain's aircraft industry could be proud of the inquiry's utter frankness. "Today's jet aircraft owe a huge amount to the Comet," Hodgson said in 2017. "It taught the lessons that the other manufacturers learned and paved the way for modern safety and design". More aircraft of the period are on the aircraft shelf.

Fly an airline of 1955

• Era: Jet Age, 1955
• Base: London, LHR
• In the catalog: de Havilland Comet 4, from 1958

Start in 1955

No signup, in the browser

This text was generated by AI from a dossier of facts, each with a source. A separate AI check compared every number, date, name and quotation with the sources, and the English and Russian versions with each other. No human editor read the text before publication. The person responsible for publication is named in the byline. Sources and the check date are provided in the article. Game parameters are distinguished from historical facts.