Investigators would draw no conclusion from the new information, supplied from military radar, but one theory was that it suggested an effort by the crew to keep control of the jet.
John Clark, a US National Transportation Safety Board investigator, said radar signals showed that Flight 990 plummeted from 33,000ft to 16,000ft, then climbed to 24,000ft before diving to 10,000ft, where it seems to have broken up. "The Air Force tells us they see multiple primaries [radar tracks] at 10,000 feet," Mr Clark said.
The plane had taken off from New York's John F Kennedy airport half an hour before it crashed and was heading for Cairo on the fourth and last leg of its journey. There was no distress call, and all communications with the plane had been normal until its precipitate fall.
Stormy seas prevented the recovery of wreckage and the crucial "black boxes" for a second day yesterday, and no improvement is expected before the weekend. Signals believed to be from the boxes were picked up on Tuesday. Investigators hope the flight data recorder will show whether a thrust reverser on an engine had been activated in mid-air.
Attention has focused on the thrust reverser - the engine- braking mechanism that is supposed to deploy only when the plane is on the ground - because this was tentatively singled out as the cause of the only other mid-air crash of a 767, a Lauda Air aircraft, over Thailand in 1991. EgyptAir has confirmed that Flight 990 had a problem with a thrust reverser, which had been deactivated before it left Cairo. However, experts say this is a standard procedure that should not have affected the flight.
Lobstermen fishing off Nantucket offered what could be the nearest thing to an eyewitness account of the crash. Skipper Christopher Lutyens told the New York Times he had heard "an immense crash" that "sounded like aluminium hitting the water, and tearing apart" around the time the plane came down. It was bit like a rumble of thunder, he said.
Mr Lutyens and his crew were interviewed by FBI agents. Three of the five crew reported hearing a crash; the other two, on the other side of the boat, heard nothing.
But the fishermen did not see anything. "No flares, no fireballs, no explosion - nothing," Mr Lutyens said. The sound, lasting two to three seconds and from the horizon, not the sky, was nothing like a sonic boom or an explosion. "I don't think that's what happened. I think it hit the water," he said.
t The cockpit voice recorder from Payne Stewart's Learjet, which crashed on 25 October, has yielded various noises but no voices from the final moments of the flight. However, the noises may help establish the cause of the accident that killed the golfer and five others.