By robert davey

The elusive ignition source for the extremely rare center tank explosion that the National Transportation Safety Board believes destroyed TWA Flight 800 on July 17, 1996, remains a troubling and controversial mystery. Although the investigation is continuing, the NTSB admits that it may never be able to explain what led up to the horrifying last moments of the Paris-bound, 25-year-old 747 and her 230 passengers and crew. 

But, in truth, the crash left a host of anomalies in its wake. The Voice has recently uncovered many unexplained elements in the investigation, among them: a piece of wing debris bearing clues that excited one investigator, a veteran pilot; a loud sound on the cockpit voice recorder and erratic readings on the flight data recorder, both suggesting a high explosive blast; hundreds of eyewitness accounts, finally being examined by the NTSB, of which more than a hundred suggest that a missile brought down the plane; and, lastly, a recommended test that investigators failed to carry out. 

The wing debris seemingly shatters the NTSB's theory. "You ever shot a .22 through a tin can? You know how the holes look where it punctures the metal and it rolls the metal back and tears it as it stretches?" the veteran pilot asked. "Well that's what these holes looked like, except they were oval-shaped." He was recalling three holes— each at least six inches long by around three inches high, he said— which had been punched through the thin aluminum panelling of a structural piece from inside the right wing of the 747. The holes were punched out "from the airplane toward the wing tip," he added. The piece, called a rib, came from within the wing's leading edge about five feet out from the fuselage, he said, where the landing lights would be. 

The pilot, an accredited accident investigator, found the piece four days after the crash while touring the Calverton hangar where the recovered debris was taken. He spotted Dr. Merritt Birky, a top NTSB scientist who led the agency's effort to document damage caused by fire or explosion, and carried the five-foot-by-six-inch rib over to him. "I said, 'Look, I think that these holes were caused by a high explosion,' " the pilot recalls. 

According to the pilot, Birky said it had already been determined that the holes in the rib were made by impact with the water. Dissatisfied, the pilot took the piece over to the FBI field lab at Calverton, where technicians gave him a demonstration of their explosive-sniffing machine. The piece promptly tested positive for nitrates, a possible sign of explosive residue. Before the rib was taken off next day for further tests at the FBI's Washington lab, another crash investigator had a chance to examine the holes. "They were not caused by water," he told the Voice. (The NTSB did not respond to Voice queries about the rib. Birky didn't respond to repeated Voice requests for comment.) 

"I do remember a piece in that general vicinity [of the right wing] was of great interest but on further examination by the [FBI] metallurgists it proved to be nothing," retired assistant director of the FBI, James Kallstrom, recently told the Voice from his office at the Delaware bank where he's now employed. The piece tested negative for nitrates in Washington, Kallstrom said. 

Kallstrom said he couldn't recall details of the investigation. For instance, a December 1997 report recommended that investigators should fill the inboard wing fuel tanks of a 747 with water and fire shoulder-launched missiles at them. Only then could a missile be ruled out as the cause of the crash. (A spokesman for the NTSB told the Voice the agency does not consider the test necessary.) 

Noting that the "severe shattering of the left wing upper skin" had puzzled investigators, military expert Richard Bott speculated in the report, obtained by the Voice, that a missile striking the inboard left wing fuel tank would create "a significant hydrodynamic ram event" that would account for the wing's peculiar fragmentation. Some wing pieces were recovered near JFK, suggesting that they fell from the aircraft in the first moments after the plane exploded. "You know," Kallstrom said, "there are some things you can't explain." 

Yet while dismissing this evidence Kallstrom seemed at the same time less emphatic in his rejection of the missile theory than he had during previous Voice interviews. "Clearly there's a mountain of evidence that says it wasn't [a missile], and maybe there's a little pile over here that says it was," he said. Previously he had insisted that there is "not a scintilla" of evidence a missile was involved. 

Whether they amount to "not a scintilla" or part of that "little pile," the eyewitness accounts remain for many the most vexing element of the TWA 800 story. Kallstrom over a year ago unveiled the CIA video that explained that the more than 100 witnesses who told of a streak of light ending in an explosion, fireball, or flash had seen not a missile but the burning plane, which the CIA thinks climbed steeply after it exploded. But now when Kallstrom was asked about the eyewitnesses, he said, "Let me say this— to this day I still believe that the eyewitnesses were reporting what they saw." Confoundingly, this appears to endorse accounts challenged by both the CIA and the FBI. 

The FBI would not allow NTSB investigators near the eyewitnesses in the early days of the investigation, but much later transferred a mountain of witness statements to the NTSB for examination. One member of the NTSB's Witness Group told the Voice the group began examining the statements last fall and hopes to finish sometime in March or April, after which, he said, the NTSB will put the accounts into its TWA 800 public docket and post them on its Web site. This senior accident investigator said he hopes the group will take a crack at explaining what the witnesses saw, and does not expect to be constrained by the CIA's analysis. "If once it's explained it doesn't make sense, I have no problem going on the record and saying so," the investigator said. 

If eyewitnesses really did just imagine they saw a missile, that leaves the NTSB searching for something that could have ignited the vapor above the 50 or so gallons of fuel in the huge center tank. Yet the safety board stated last year that it may never find the ignition source. 

That is not surprising, considering the odds it faces. After all, its $30 million investigation was unable to find a single flaw in any of the recovered aircraft components that could have allowed a spark into the tank. That, in turn, is no surprise to the man whose familiarity with 747s is second to none. "When we designed the airplane we did every damn thing we could to make a fuel tank explosion not happen," said Joe Sutter, the retired Boeing chief engineer who more than 30 years ago led the 747 design team. Sutter told the Voice, "There've been thousands of 747s, sitting on the ramp, cooking in violent heat for hours due to delays, and they've taken off with hot empty tanks hundreds of thousands of times, so if it happened this way, that was a real freak accident." (Boeing has built 1200 747s.) 

The Flight 800 disaster was truly a freak occurence. As noted in Insight Magazine (February 8), it is one of two unexplained center tank explosions. The other was a new Philippine Air Lines 737 that exploded before takeoff in Manila in 1990— and although a cause for that explosion has never been found, it is known that the airplane had been modified by the owners since it left the factory. 

To show it has indeed found the cause of the crash, the NTSB has identified a breakup sequence that it says led in two or three seconds from the center tank explosion to disintegration of the aircraft, specifically when the nose and first-class section fell off. But although its investigators determined which parts of the tank were damaged in the initial explosion, not all those features were included in the scenario developed by the Sequencing Group. In addition, the group made no progress toward identifying where in the tank the explosion may have begun, admitted senior metallurgist James Wildey in his summary to the Sequencing Report. The breakup sequence, then, remains merely a best guess for the order in which things happened. 

But one expert who examined the report suggested that the burst of energy that tore the center tank partitions from their rivets, snapped the keel beam beneath the tank, and ripped apart the fuselage skin may not have come from inside the tank at all. "If the tank had 50 gallons of water in it, and you put enough energy into it, it will blow up," said Professor Richard Schile of the University of Bridgeport. Schile, who has degrees in mechanical and aeronautical engineering and has worked on failure analysis at Pratt & Whitney Aircraft, said that in examining combustion in the tank the NTSB may have missed the main event. "This sounds to me like the pressure rise was a hell of a lot higher than we think it was and it occurred so rapidly that it just blew the structure apart," he said. "The fire may have come later and been incidental." 

A sudden increase in pressure would be expected to leave its mark on the airplane's black boxes. There is a sonic signature at the very end of TWA 800's cockpit voice recording that investigators have described as a very loud, very abrupt sound. So far, however, the NTSB has barely acknowledged the sound. To electronics engineer James Cash, who as chairman of the Cockpit Voice Recorder Group leads the NTSB's study of the sound, that's as it should be. 

"Analysis is never released . . . it's for our internal use," he told the Voice. But one investigator said that the NTSB has turned down offers from outside labs to interpret the CVR sound. (The NTSB did not reply to a query about this.) One member of the CVR group is skeptical that the NTSB has made any real effort to analyze the sound at all. "It certainly should have been investigated thoroughly," he said. 

Recently, one investigator said, TWA and the Air Line Pilots Association, two parties to the investigations, have urged the NTSB to say what could have caused the CVR sound, and to release a report on some tests conducted on a retired 747 in Bruntingthorpe, England, in the summer of 1997. In one test the safety board filled the airplane's center tank with propane, exploded it, and recorded the sound on a CVR. "We need to use the Bruntingthorpe data in our analysis. I want the data to be published," said one investigator. The CVR group has not met since the Bruntingthorpe tests, which were finished in August 1997, he said. The NTSB did not respond to a question about the report on its Bruntingthorpe tests. 

Some independent investigators remain convinced that the wildly erratic readings included on the last line of data from the flight data recorder, which the NTSB drew a line through on its FDR Report, also indicate that a high-pressure wave rocked the airplane (Voice, July 21, 1998). Retired TWA captain and flight engineer Howard Mann examined an addendum to the FDR Report, obtained by the Voice, and concluded, "They are not accounting for the erratic data at all." 

These anomalies— the witness accounts, the CVR sound, the damage to the wing rib, the FDR's last line— all nourish a band of TWA 800 conspiracy theorists, some of whom still believe that an accidentally fired U.S. Navy missile hit the plane. A founding member of this group recently got the chance to scrutinize the wreckage inside the Calverton hangar. Author James Sanders, who wrote The Downing of TWA Flight 800 and who is awaiting trial on charges of conspiring to steal evidence from the hangar (Voice, April 21, 1998), was permitted to photograph the fuselage reconstruction as part of pretrial discovery. He and his attorney first had to sign an agreement that they would not share any of the photos with the media, he told the Voice. 

Sanders said that after examining the reconstruction he concluded that the NTSB has given a misleading impression of some of the damage. He mentioned, for example, a center tank partition that, according to the NTSB, sustained "accordion" damage, meaning "folding directly inboard" from the direction of the right wing. But Sanders said that the damage is far more extreme than the description suggests. "Spanwise beam two is crushed inward about eight feet from an external force— it's extraordinary when you see it in person," he said. (The NTSB suggests that the accordion damage was caused by water impact.) 

Retired navy commander William S. Donaldson also is convinced a missile hit the plane (Voice, July 21), but unwilling to believe the Navy capable of such a ghastly error, he blames terrorists. He told the Voice recently that an 800 number his group had placed in a Long Island newspaper has drawn an overwhelming response from eyewitnesses who are not satisfied with the government's explanation of what they saw.