Federal officials were yesterday speeding up efforts to reassemble wreckage of the plane in the main hangar of a Long Island airfield not far from where it fell to the ocean in a fireball on 17 July, killing all 230 on board. The work, which involves rebuilding the aircraft around a skeleton of chicken wire, is focusing on the section where the wings joined the fuselage, where the worst damage has been found. The pattern of debris on the ocean floor also indicates that parts of the plane's underbelly beneath the wing junction fell to the sea first.
Evidence has also been found suggesting one of the explosions was in the fuel tank between the wings, which was almost empty at the flight's outset. Sources have reported that that blast may have occurred up to 20 seconds after an earlier explosion elsewhere in the aircraft.
That finding, if proved, militates against an earlier theory that an electrical spark may have ignited the fuel tank, causing the initial blast that downed the airliner. Thus investigators are again being drawn to the theory that a primary explosion, which may have triggered the tank blast, was caused by a bomb. In the absence of forensic evidence of an incendiary explosion, the two other primary theories remain under consideration: mechanical failure or a missile attack.
As to where a bomb may have been planted, wreckage retrieved this week clouded two earlier theories: that a detonation occurred in the front cargo hold, where passenger bags were stowed, or in the cockpit. Both scenarios had been fed by evidence that the front of the aircraft was severed from the rest of the fuselage. Items from the hold and the cockpit have meant both areas have been more or less ruled out.
Discovery of fire damage in an area near where the leading edge of the right wing joined the plane has propelled new theories to the fore: that a bomb may have been placed in that area of the cabin, above the fuel tank, in a food trolley or inside a carry-on bag. The inside-right engine also appears to have suffered much greater damage than the other three.
The "fingerprint" that forensic evidence might supply is so far eluding investigators. Preliminary testing by equipment at the hangar this week showed residue of PETN, a prime ingredient of an explosive known as detasheet or detcord and which can also be used in Semtex. But subsequent testing in Washington DC failed to confirm the presence of PETN.
With it being possible that all the hoped-for evidence may have been washed away - about 50 per cent of the plane has been retrieved - investigators are also looking for clues from the directions in which the metal was twisted as well as the pattern of metal fragments in the bodies of passengers.
Shards have been found in passengers' feet and legs, pointing to a blast from below, presumably from the fuel tank. But the extent of damage in that area suggests an explosion there was not sufficient to cripple the plane so decisively.
There may be divisions in the investigation team as it approaches the one-month anniversary of the crash with no clear evidence of what happened that can be offered to the public. A source told the Associated Press news agency: "Investigators are split. Some believe there was an explosion in the forward part of the aircraft. As much as 20 seconds later the fuel tank blew up ... Others speculate that the tank may have been part of the initial blast."
The possibility of a missile attack is thought unlikely. It would almost certainly have been heat-seeking and therefore made initial contact with an engine as it roared to raise the plane to cruising altitude. All four engines have been located and none shows evidence of having been hit. The discovery of the engines has so far provided no evidence of mechanical mishap, including the possibility that one became unhinged.
Three theories on how airliner was downed
The near-perfect safety record of the Boeing 747 has discouraged investigators from seeing mechanical failure as a likely cause.
Scenarios for such a calamity have also, one by one, been ruled out by what they have found.
The cockpit voice recorders or black boxes, which showed a routine ascent and ended abruptly with a brief, unexplained noise, weigh against pilot error.
The possibility that an engine became unhinged and spun into the rest of the plane is seen as unlikely, since all four engines have been found in reasonable condition.
Suggestions that an electrical spark might have ignited vapour in the central fuel tank have also been discounted. While parts of the tank show fire and smoke damage, other parts appear in pristine condition.
It is now thought the tank may have ignited 20 seconds after an initial blast elsewhere on the plane.
Officials have spoken little about a further possibility: that the plane may have been torn apart by a catastrophic deployment after take-off of the engines' reverse thrusters.
A favourite theory of anyone familiar with the 1990 novel Stinger. Written by Doug Hornig, the book opens with the downing of an American airliner taking off from Boston by a man on a boat with a shoulder-launched Stinger missile.
This scenario was lent early credence by eyewitness sightings of an "object" streaking towards TWA800 just before it exploded as well as by an unexplained second blip on traffic-control radar stills.
A missile attack has been all but ruled out by investigators since the discovery of the fourth engine this week, however.
A missile would almost certainly have been guided by a heat-seeking system and initial contact would have been on an engine. All four have shown damage consistent only with impact on the ocean surface in the crash. Nor, so far, have forensic scientists found any evidence of an incendiary blast on the engines.
And a missile strike would not easily explain the instant loss of electrical power shown by the data and voice tapes. Forensic tests are still continuing.
Bomb placed on board
Investigators want to believe a bomb doomed the aircraft after decapitating its front section. But nothing has surfaced yet to provide them with conclusive evidence.
An early theory that a bomb may have been detonated inside passenger luggage stowed in steel cargo containers in the front cargo hold was dashed after all the containers were retrieved and found to be in "unremarkable" condition.
Suspicions about a cooler containing corneas placed in the cockpit just before departure were also negated by the discovery of some undamaged glass covers from cockpit instrumentation. The focus now is on the section where the forward edge of the wings joined the fuselage.
The shape of retrieved plane parts suggests two blasts, one that sent its force downwards from the cabin and another that came from below.
Forensic tests at the hangar where the plane is being reassembled showed initial signs of the presence of PETN, the chief ingredient of a plastic explosive known as detasheet or detcord.
But subsequent and much more reliable tests of the same wreckage parts conducted in Washington DC have failed to confirm these first results.