An object that detached from the space shuttle Columbia on its second day in orbit resembled a carrier panel from near the leading edge of a wing, investigators said yesterday.

The evidence from Air Force radar analysis fits with a growing body of data indicating that the orbiter's left wing was damaged before it began its Feb. 1 reentry into the atmosphere. Investigators believe the damage created an opening in the orbiter's heat shield that allowed superheated air to burn its way into the structure, with disastrous results.

The carrier panels serve as a bridge between the ceramic tiles that cover the orbiter's underbody and the larger carbon fiber panels -- shaped something like horseshoes -- that are bolted to the wings' leading edge to protect against the hotter temperatures that build up there.

Following extensive analysis of 29 shuttle materials, "we've concluded that, right now, only the carrier panel remains a viable candidate for the Day Two object," said Maj. Gen. John L. Barry, a member of the Columbia Accident Investigation Board, at a Houston briefing yesterday.

However, the picture shifted again soon after the briefing ended, when engineers analyzing shuttle wreckage delivered to Kennedy Space Center identified a carrier panel from the bull's-eye area of the wing where investigators believe the initial breach may have occurred. It was carrier panel number 6 and appeared to be in good shape, a member of the board said later.

"That means that not a whole lot burned around it," he said. "It probably means the carrier panels on either side didn't burn."

The development cast fresh doubt on the radar object's identity, he said, and indicated that "this is a detective story. Now we have one more part of the puzzle. It's a long, long process."

The team previously identified 10 other carrier panels and is still looking for panels number 5, 8, 9 and 12.

The mystery object that drifted away from the shuttle in space was detected in extensive post-accident studies of about 3,100 radar images that Defense Department trackers took of orbiting objects during the mission, prompting fleeting speculation that a piece of space junk or a micrometeoroid had struck Columbia.

The object was first spotted floating near the orbiter. It began to tumble and eventually fell out of orbit.

The hardware could have been weakened or loosened by the impact of foam debris that flew off the shuttle's external tank during launch. In one scenario, the same hardware a day later could have gotten jarred free as the orbiter fired its jets to complete a series of maneuvers.

To identify the object, the investigating board sent samples of 29 shuttle materials to Wright-Patterson Air Force Base in Ohio, where analysts have been studying the "signatures" of their radar reflections to try to match one with the object seen in orbit.

If the team had sent more samples, more items might have fit the radar profile, said Barry, who is director of plans and programs for the Headquarters Materiel Command, at the base.

Additional tests are planned to see if, for example, a partial section of a leading-edge carbon fiber panel would fill the bill. Other tests will be done on carrier panels with varying numbers of ceramic tiles and other hardware attached, he said.

The carrier panels are bolted onto the wing, and the bolts in turn secure heat-shield tiles to the panel's exterior. The bolt holes are covered with ceramic plugs, which are relatively weak. Investigators plan tests to determine what damage might result from an impact on those points.

Investigators have also been studying NASA's handling of pinhole defects that develop in the carbon fiber panels as they age, another possible source of weakness in the heat shielding. Inspections have found that each panel typically has 20 to 40 such pinholes, and any measuring more than 0.04 inches across are repaired or more extensively refurbished, Barry said.

If Columbia began its reentry with a missing carrier panel, said board member Roger E. Tetrault, this would have allowed the flow of enough superheated air to trigger the chain of events observed in telemetry data and measurements from a flight data recorder recovered from Columbia's wreckage two weeks ago.

However, he too cautioned that several other scenarios have to be ruled out before the board reaches a firm conclusion. "You can't rule out all these [possibilities] and make this leap of faith" without risking another shuttle accident, he said.

Retired Adm. Harold W. Gehman Jr., chairman of the board, said the evidence of heat damage has to match up with several other independent lines of evidence -- a process he compared to lining up the holes in pieces of Swiss cheese.

As a direct result of the new information from the data recorder recovered by searchers in Texas on March 19, the board said it is making accelerating progress in filling out and expanding the known sequence of events during reentry.

The tape recovered from the flight data recorder contained measurements of temperatures, vibrations, pressure and other variables taken during both launch and reentry. Tetrault said an initial reading of the launch data showed no clear evidence of the debris impact, which occurred about 80 seconds after liftoff.

Investigators said the National Imagery and Mapping Agency has further enhanced a video that shows the left wing impact during the ascent. Tetrault said the team now believes a two-pound piece of foam approximately 24 inches by 15 inches by 5 inches, traveling about 430 mph, struck the left wing somewhere in a two-foot-wide "footprint" centered on the lower side of carbon fiber panel number 6. The footprint includes portions of two carrier panels.

The board has delayed tests set to begin this week at an independent research institute in order to incorporate the new video data. Researchers will fire foam debris into samples of leading-edge carbon fiber panels, carrier panels and other shuttle materials in order to assess accident scenarios.