It took 395 days to establish with cast-iron certainty what had never really been in doubt: that 39-year-old Janet O’Hara of Irvington, New York, was among the 230 people who died on board TWA Flight 800; that on a sultry July evening in 1996, on her way to Paris, she perished in a fireball above the Atlantic Ocean off the shore of Long Island, together with her TV-producer husband, Jack, and their 14-year-old daughter, Caitlin.

The positive identification of Janet’s remains by the Suffolk County Medical Examiner’s office - much maligned at the time of the crash for its supposed insensitivity and lack of dedication - was remarkable because it marked the first time that the fate of every victim of an airline disaster has been definitively confirmed by DNA testing - by matching the shattered human debris salvaged from the sea with samples taken from close relatives or from personal effects of the dead: clothes, hairbrushes, toothbrushes - in one case, nothing more than stubble from an electric razor.

But what comfort that certainty has brought to the relatives left behind - who include the O’Hara’s 13-year-old twin sons - is marred by the gnawing uncertainty of what killed them: of what caused the Boeing 747 to explode in a cloudless sky at 13,700ft, eleven and a half minutes after taking off from New York’s Kennedy airport.

In some ways the investigation of the accident has been as exemplary and even more enduring than the forensic detective work undertaken in the Suffolk County morgue. The search for debris from the plane, scattered over six miles of ocean bed, continued for ten months, until all but the smallest pieces had been found. The task of reconstructing the fuselage, foot by foot, inch by inch, continues still.

Meanwhile, though the FBI’s early certainty that the plane was brought down by a bomb or a missile has faded away, starved by the lack of any physical evidence, neither theory has been completely abandoned. In June, eleven months after the crash, in a "let’s-leave-no-stone-unturned" mode of investigation, missile warheads were detonated alongside an old 747 fuselage, parked in the desert in the southwest United States. The aim was to establish if the warheads, exploded at various distances from the fuselage - from a few feet to inches – produced gaping holes similar in pattern to those found in the hull of the TWA plane. They did not.

In July, FBI agents, assisted by British explosives experts who investigated the 1988 sabotage of Pan Am Flight 103 over Lockerbie, explored what might be called the "vanishing bomb" theory: that although no suspicious trace of explosives was found on the wreckage of Flight 800, a small but powerful "shaped charge" might have been used to detonate an explosion in one of the plane’s fuel tanks. In a second junked 747 fuselage, this one parked on a military base north of London, five "shaped charges" were detonated to see if they produced tell-tale similarities. They did not.

So, on July 14, three days short of the anniversary of the crash of Flight 800, the US National Transportation Safety Board, or NTSB, the primary investigators of the disaster, flew the first of nine carbon-copy flights from Kennedy airport, using a 747 of the same early vintage as the TWA plane, taking off from the same runway, after the same delay in similar weather, carrying the same amount of fuel, following the same flight path at the same speed to the same altitude, reproducing every action the doomed crew is known to have taken. There were no passengers or cabin crew on the copy-cat flights, but installed in the plane - and, crucially, its fuel tanks - were more than 150 sensors, to precisely monitor temperature, pressure and vibration.

No investigation in the 30-year history of the NTSB has been more thorough or cost more - and if the costs of the FBI’s futile criminal investigation are included they now top more than $50 million. Worth it, one would think, for NTSB investigators firmly believe they have established beyond reasonable doubt the proximate cause of the disaster, if not its precise catalyst.

Yet for all the impact the investigation has had, or is likely to have, on aviation safety - on preventing a reoccurrence of the set of malign circumstances that turned Flight 800 into a flying bomb - it might as well have been abandoned months ago. For the NTSB’s conclusions are hotly disputed, and their recommendations have been rejected out of hand or ignored. Fifteen months on, the NTSB believes it could happen again on any Boeing 747 - and thousands of other commercial airliners with similar design characteristics.

*****

Airline accident investigation is rarely a precise science. Even if most commercial jets were equipped with the latest flight data recorders, which, for reasons of cost, most of them are not, the "black boxes" (actually coloured bright orange) that preserve the final statistics of a fatal flight are often incapable of explaining what went wrong. In the case of Flight 800, every parameter the black box measured was entirely normal, until the recording abruptly stopped. The cockpit voice recorder also registered no sign of alarm.

But through the patient charting and identification of almost every scrap of debris collected from the sea, NTSB investigators were able to establish with near-certainty the sequence in which the plane broke up: from the first explosion at 13,700ft that punched holes in both sides of the fuselage, immediately behind the 747’s distinguishing dolphin-like hump, to a second explosion some 5,000ft lower that tore off the front section as far back as the leading edges of the wings.

For the NTSB’s investigators, there has never been much doubt that the first, destructive explosion was of an overheated, highly-volatile mixture of air and jet-fuel vapour in the 747’s centre fuel tank, which sits in the belly of the plane, between the wings, some 18ins beneath the passenger cabin floor. On Flight 800, the centre tank was virtually empty, containing about 50 gallons of kerosene, rather than the 12,890 gallons it can hold, allowing fuel vapour to build up in the void. And because the plane sat on the ground at Kennedy airport for three hours in July temperatures - delayed by minor mechanical problems and a late-arriving passenger - that vapour had inevitably heated to or beyond the flashpoint of commercial jet fuel, of around 100 degrees Fahrenheit.

What the NTSB has not established with any certainty is what ignited the vapour, causing it to explode. Nor can it explain why such a catastrophe has not happened before. Commercial jets regularly take off with tanks that are less than full, and containing vapour, for airlines are loath to carry more fuel than is necessary. Extra fuel means extra weight, which means more fuel is burned, which wastes money. And in the quarter-century or more that jumbo jets have made millions of flights, Flight 800 was scarcely the first 747 to be delayed on the tarmac in baking heat.

The reason why jumbo jets have not routinely turned into flying bombs in the past is that extraordinary precautions were taken in the design of the fuel system to prevent sparks or electrical discharges occurring inside the tanks. For example, no electrical wires run through the centre tank of a 747, and the electric motors of the fuel pumps are mounted on the outside. And while on earlier models of the 747 - like the 25-year-old TWA plane that crashed - bundles of wires do run through the three fuel tanks on each wing, they are shielded with aluminium and then wrapped in two layers of protective Teflon.

So what was the fatal difference on Flight 800? Among the crash investigators there was no shortage of theories, but also no physical evidence to support them. More than 90 per cent of the wreckage was salvaged from the Atlantic without providing a single clue as to what triggered the explosion. By last December, when the salvage operation was scaled down and left to shrimping boats to trawl the sea bed with their nets, the NTSB knew that it was highly likely that the definitive evidence, if it existed, would never be found.

But if the trigger of the explosion could not be identified and eliminated, the source of the explosion most certainly could - or so the NTSB came firmly to believe. In December, James Hill, the chairman of the NTSB’s governing board, accepted his investigator’s urgent recommendation that draconian measures were needed to guard against fuel-vapour explosions, however they might be triggered, by preventing the build-up of volatile vapour in the first place. It was, as one investigator said, as obvious an expedient as disarming a petrol bomb by removing the petrol, rather than the fuse.

Easier said than done, but not impossible. To begin with, the NTSB proposed, no passenger jet should be allowed to take off with empty or half-empty fuel tanks. If there are no large voids in the tanks until long after a plane has reached the chilling temperatures of its cruising altitude, vapour will not - cannot - build up or become heated to beyond its flash point. To make doubly sure, the NTSB urged that refuelling should no longer be done from mobile tankers, which bake in the heat as readily as planes, but from underground tanks where the kerosene could be kept cool. And, third - or perhaps as an alternative - the NTSB argued that commercial airlines should be forced to do what the US military does to guard against fuel explosions on most of its transport aircraft: equip them with cylinders of a heavy inert gas, such as nitrogen, which is pumped into the tanks to replace the fuel as it is burned off.

Had all or even any of those precautions been accepted, the NTSB believes, the risk of another Flight 800 catastrophe would have been virtually eliminated overnight. But the NTSB had no power to impose its recommendations. That authority belongs to the Federal Aviation Administration, which approves every aspect of aircraft design, down to the smallest rivet, and which holds sway over every airline and commercial aircraft flying to and from the US, as well as within it. But while safety is the FAA’s priority, it also has a regulatory duty, imposed by Congress, to foster the aviation industry, and it must draw the balance between, on the one hand, eliminating risk and, on the other, imposing burdens the industry could never meet (following the maxim that a wholly safe plane would never take off).

The industry’s first response to the NTSB’s proposals was to describe them, in the words of Boeing, as "far reaching" with "broad implications for the entire industry" - which is to say expensive. Estimates of the cost of retrofitting inert gas cylinders to the more than 1,000 747s in service range up to $2 million per plane - and it is not just 747s that are at risk, but thousands of other aircraft with similar fuel systems, including models made by the European consortium, Airbus.

To nobody’s surprise, the FAA’s response to the NTSB recommendations has been to stall. In June, six months after the investigators pressed for "urgency", Barry Valentine, then acting administrator of the FAA, wrote to NTSB chairman Hall to say: "More research is necessary to understand fully what happened within the center fuel tank to cause it to explode." Only then, Valentine said, would there be "a more complete understanding of the next steps that must be taken on this critical issue". Hall didn’t see it that way at all. He responded by repeating his demand for immediate action, and by setting down a marker: "As there still appear to be significant differences between us on this point, I believe the position of the board needs to be reiterated now." Publicly, the exchanges between the NTSB and the FAA have been frostily polite, but Hall has made no secret of the fact that he regards the FAA’s position as "unacceptable". Bernard Loeb, the NTSB’s director of aviation safety who heads the investigation into the crash of Flight 800, has gone further, publicly accusing the FAA of "recalcitrance" that, he claims, was and is putting passengers lives at constant risk.

But to no avail - and that situation is unlikely to change any time soon. On July 8, some 100 representatives from 40 airlines and manufacturers from the US, Europe and Asia, gathered at Boeing’s headquarters in Seattle, Washington to debate the NTSB’s recommendations. In what one participant described as an unprecedented display of unity, they agreed to flatly oppose them - in particular the expensive proposal to use inert gas as a smothering blanket. The objections were based on safety as well as cost, for, the industry claims, the plumbing required to pump in inert gas could make the fuel tanks more susceptible to leaks. The representatives came up with detailed counter-recommendations of their own, including a call for the development of a safer commercial jet fuel. In the view of the NTSB, those recommendations will do little or nothing to prevent airliners taking off with one or more of their fuel tanks filled with overheated explosive vapour.

*****

The industry has not been entirely idle. At the urging of Boeing and the FAA, the fuel tanks and systems of all 747s in service have now been minutely examined. While no defects were found that might cause a spark in centre fuel tanks, the inspection did turn up an unexpected flaw in some older models of the jumbo: what the FAA describes as "numerous" cases of chafing of the supposedly impregnable protective casing of the wires running through the fuel tanks in the wings.

Though no bare wires were found, the discovery has led the FAA to develop a complex if not convoluted theory as to what might have happened to Flight 800. By this theory, a "flame front" started in a wing tank by faulty wiring travelled to the wing tip and then "blew back" into the centre tank - ironically along a venting tube through which vapours from the tank are supposed to escape. (The venting system was designed to deal with a gradual build-up of fumes, not a huge tank filled with them even before take off.)

If this theory becomes accepted as the "probable cause" of the TWA crash, the industry will breathe a collective sigh of relief, because it means the problem is confined to a few hundred aging aircraft, and can easily be dealt with by regular inspections of the wiring and, if necessary, replacement of the protective coating. Perhaps that is why the NTSB is highly skeptical of the FAA’s theory, dismissing it as no more than "one of many possibilities".

The NTSB’s favourite theory is much more disturbing - for both the industry and passengers - because it implicates every jet in which, in the absence of inert gas, fuel can slosh around in a near-empty centre tank, allowing a build-up of static electricity - and, eventually, a spark. In theory, the fuel tanks are designed to prevent the build-up of static, but Loeb, the NTSB’s lead investigator, believes that a small crack in the TWA tank may have allowed fuel particles to gather into a lethal cloud, before discharging in thousands of volts of electricity - more than enough to ignite the vapours.

In an attempt to force the FAA to take action, Loeb plans to conduct a series of experiments, igniting vapour inside fuel tanks. He says the carbon-copy flights conducted over Long Island "gave us tremendous insight into the volatility of vapours" and "helped us far more than we ever imagined possible". Building on that knowledge, his experiments will culminate in the blowing up of an entire 747 on the ground, somewhere in the Nevada desert, early next year.

Even so, his theory will never be more than that: a possible explanation for what triggered the explosion on Flight 800, unproven by physical evidence, and therefore unprovable. Faced with a slew of lawsuits filed in California, New York and Pennsylvania by relatives of many of the victims, Boeing and TWA are certain to continue resisting Loeb’s explanation of the "probable cause" – as firmly as the industry as a whole will continue to resist the NTSB’s proposed measures to prevent it from happening again.

It will require another crash, another explosion of fuel vapour under the passengers’ feet - and one that occurs preferably over land, where the wreckage is easier to retrieve - before there is any slim hope of identifying the fatal flaw. Even then, the chances of finding, say, the guilty bit of wiring, or minute evidence of a static discharge, must be hugely remote.

The NTSB believes that as things stand, sooner or later, another mid-air explosion of fuel vapour is probable. It is also probable that in its aftermath, if those required to deal with it have the persistence and skills of the pathologists of Suffolk County, it is the remains of the dead that will be positively identified - not the precise cause of their dying.©