The chunk of insulating foam that broke free from shuttle Columbia's external fuel tank Jan. 16 was the biggest piece of debris ever to hit a shuttle during launch, and only two previous launches have encountered even moderately large impacts from the tough insulating material, a NASA official said yesterday.

That suggests that NASA engineers had little practical experience to rely upon as they weighed the potential importance of the event -- and as they made their decision, days later, to write off the collision as probably no threat to the shuttle's safety.

The new detail about foam debris impact described by Michael Kostelnik, a NASA deputy associate administrator, was one of several pieces of new information to shed light yesterday on the frustrating level of uncertainty that engineers had to work with as they decided whether to ignore the incident or perhaps design a radical plan to bring the crew home safely.

Among the uncertainties were the potential effects of the debris not only on the shuttle's ceramic tiles -- the major focus of NASA's attention -- but also on the reinforced carbon leading edges of the shuttle's left wing, which are stronger than the tiles in some respects but have their own vulnerabilities.

Also yesterday, NASA officials sent teams to California and Arizona to check reports that pieces of Columbia had been found in those states, which would be the farthest west that debris from the high-altitude crash had been found. Officials are very interested in looking at the most westward pieces from the breakup, on the assumption that those pieces would reveal which part of the spacecraft was the first to fail.

Last night, local officials reported finding a portion of what they believed to be a shuttle wing in a pond in Texas. And in a tantalizing development that could offer additional clues, the military yesterday confirmed that one of its Apache helicopters flying over Texas early Saturday had caught images of the shuttle on its approach. Details of the footage were not available yesterday.

Investigators have said the apparent collision between a dislodged piece of fuel tank insulation and the shuttle's left wing -- noted on video about 80 seconds after launch -- is their leading candidate as the cause of Saturday morning's disintegration of the shuttle.

At the same time, NASA officials have repeatedly expressed doubt that such a collision would be capable of damaging the shuttle's protective surfaces enough to jeopardize the spacecraft through the heat of reentry as it hurtled back to Earth.

Their confidence, they have said, was based on both empirical evidence -- the many previous times that tank foam has pelted the craft upon liftoff with no serious effects -- and on computer modeling they conducted both before and after Columbia's voyage, which suggested the impacts did not pose a "safety of flight" issue.

But Kostelnik's comment at a Washington news briefing yesterday that the slab of hardened foam -- estimated to weigh 22/3 pounds -- "probably is the largest piece" ever to hit the shuttle on takeoff undermines the agency's supposition that the lack of trouble in the past was predictive of a safe return on this flight, experts said.

If one in only three large-scale foam impacts -- or "instances of significance," as Kostelnik called them -- can result in the loss of a shuttle and its crew, experts said yesterday, the overall safety of the shuttle may be less than NASA presumed, and it becomes important to know the probability that similar sized or even larger chunks of foam might break off on future flights.

"We need to find out how big a piece could come off and how often might this happen," said Robert K. Weatherwax, president of Sierra Energy and Risk Assessment, a Roseville, Calif., company that has analyzed shuttle safety. "This is reminiscent of the bad old NASA," he said, referring to what he said was an old habit at NASA of being overly optimistic in its assessments of dangers.

Other experts noted that computer models for predicting how impacts might affect the structural integrity of high-tech materials are notoriously susceptible to error.

"Fracture from simple stress is already a complex phenomenon," said John Bassani, a professor of mechanical engineering and applied mechanics at the University of Pennsylvania and an expert on high-temperature mechanics and material failures.

"A lot of work has been done to better model it, but it's complicated," Bassani said. "And when you have an impact, that adds complexity because one never knows precisely what the impact loads are."

NASA investigators have acknowledged that they had to make several assumptions about the size and weight of the foam, its velocity and the angle at which it hit the craft. They have yet to describe exactly what models they used and what, precisely, the results were -- other than to say they concluded that the flight was not jeopardized.

One potentially important factor is whether the debris that hit the shuttle was foam, ice or a combination. The shuttle sat on the launch pad in wet weather for more than a month, and water in or on the foam when the tank was filled with fuel likely would turn to ice. Inch for inch, ice weighs about three times as much as the insulating polyurethane foam.

Kostelnik yesterday said he had "confidence that they did that as well as anyone can do."

But beyond the question of whether ice as factored in, is the complexity of predicting how materials respond to unusual stresses, experts said.

Eugene E. Covert, professor emeritus of aeronautics and astronautics at the Massachusetts Institute of Technology, who helped investigate the Challenger disaster, agreed that computer models leave a lot to be desired.

Referring to NASA's previous use of predictive computer models, he said: "They've said that every time they use this computer model and then they look at the shuttle back on the ground they find that it predicted more widespread damage and more severe damage than they actually have seen," suggesting that their models, if anything, overestimate the risk. "Still that's not necessarily that good, because it may be for one reason or another that they haven't imagined all the possibilities of failure and there may be certain things you just don't know."

Part of the problem, Bassani said, is that materials with different responses to heat can behave in unpredictable ways where they are joined. That is exactly the situation with the ceramic tiles -- which can withstand extremely high temperatures -- and the aluminum shuttle body to which they are cemented, which deforms at far lower temperatures.

Further complicating predictions about how fractures will evolve, Bassani said, is the fact that even small variations in the atomic structure of materials under high stress can send microscopic cracks growing in unexpected directions in a series of "cascading effects."

Mathematicians are trying to use sophisticated "chaos theory" computer programs to predict how these tiny alterations can have major effects on outcomes, but the science is still too young to be very practical, he said.

Ultimately, of course, in situations like Columbia's, a decision must be made.

"You have to go with what you've got," Covert said. "It's really a very difficult position to be in. You don't have a great deal of flexibility. You can't send a tile up by Federal Express."

Some experts have said that if the extent of damage had been known, NASA could have tried something radical, such as flying the shuttle alongside the international space station for a risky spacewalk transfer of personnel -- a theoretical possibility but one that Columbia lacked the fuel to accomplish.

But in all likelihood, several experts said, Columbia's fate was sealed after launch, and the point of the investigation should be to make future flights safer, not to second-guess what should have been done before.

Former NASA space flight official Joseph Rothenberg, now president of Universal Space Network, underscored the importance of good detective work in locating the root cause of the accident.

"They can't just fly again and hope it doesn't happen. They're going to have to find the root cause, or at least have a high probability they've found it."

Staff writer Kathy Sawyer contributed to this report.

Smithsonian Air and Space Museum visitors watch President Bush's remarks during the memorial service. "Their mission was almost complete, and we lost them so close to home," he said. The museum crowd listens to President Bush's nine-minute remarks in Houston hailing the seven astronauts as having lived "seven lives of great purpose and achievement."