When the Twin Towers Fell

When New York City's giant World Trade Center towers plunged to earth following successive suicide terrorist attacks on September 11th, the world was confronted with one of most shocking—and sickening—sights of modern times. The mechanisms by which these huge and seemingly solid edifices suddenly collapsed, snuffing out the lives of thousands, was the subject of a preliminary postmortem conducted last week in Cambridge, Mass. A panel of Boston area-based civil and structural engineers convened to discuss the fate of the superskyscrapers, struck by hijacked passenger planes, in front of an overflow audience on the campus of the Massachusetts Institute of Technology. Their starkly sobering analyses highlighted the vulnerabilities of ultra-tall buildings to fire and pointed out steps that could be taken to lessen them.

After first describing the highly redundant structural system that kept the 110-story twin towers standing for decades despite hurricane-force winds and a terrorist truck bomb, the engineers then delineated how that system was breached and finally overcome on that fateful day when America was attacked. The main culprits in bringing the famously lofty buildings down, they concluded, were the two intensely hot infernos that erupted when tens of thousands of gallons of aviation fuel spilled from the doomed airliners. Once high temperatures weakened the towers' supporting steel structures, it was only a matter of time until the mass of the stories above initiated a rapid-sequence "pancaking" phenomena in which floor after floor was instantly crushed and then sent into near free fall to the ground below. Significantly, the panel stated that any mitigating reinforcements and redundancies added to these buildings could have only delayed the inevitable failure, though they would have bought more time for the evacuation of the occupants. No existing or foreseeable economically viable skyscraper structure, they agreed, could have withstood this kind of cruel onslaught. Clearly, prevention is the best defense against this kind of assault.

"Though the twin towers were not much taller than their famous uptown predecessor, the Empire State Building, the World Trade Center rose during the late 1960s, a new era of construction characterized by rapidly erected, lightweight steel structures rather than heavy masonry walls," explained Robert Fowler, senior engineer at the structural engineering firm of McNamara and Salvia. Fowler was then a junior member of the WTC's engineering firm of record, Worthington, Skilling, Helle & Jackson, later renamed Skilling Helle Christiansen Robertson. "As the Trade Center was so much lighter in comparison to earlier designs, it was a watershed building in the history of skyscrapers," he added. Leslie E. Robertson, then the project manager, was the engineer most responsible for the superskyscraper's design, Fowler noted. He is currently principal partner at Leslie E. Robertson Associates, the current structural consultants to the WTC. The late Seattle-based architect Minoru Yamasaki designed the World Trade Center.

Read the rest here

Prof. Eduardo Kausel

Inferno at the World Trade Center towers

by Prof. Eduardo Kausel

As I anxiously watched the TV coverage of the terrorist attack on the World Trade Center towers, my training in Structural Engineering instantly elicited in me visions of doom, and a feeling that the towers were in imminent danger of collapse. Still, knowing that in 1993 the towers had resisted massive damage in a terrorist attack, and being unaware of similar cases of skyscraper collapse, I hoped against reason that they might survive yet again. To my horror, I then witnessed the unthinkable unfolding in front of my eyes. In retrospect, I should have been 100% sure that they would fail, but the idea was so disgusting that I allowed my wishful thinking to prevail instead. Soon after the tragedy occurred, cooler thoughts and the engineer in me returned, and I began to ponder about the mechanics that led to the catastrophe.

Why did they collapse?

There were three causes for the massive structural damage that led to ultimate failure: the impact of the aircraft, the subsequent explosion, and most importantly, the raging fire caused by the vast amounts of jet fuel. Burning fuel must have also cascaded down floor openings to the levels below.

The towers were reportedly designed for the impact of a Boeing 707 aircraft, the largest of its day. The takeoff weight of a fully loaded Boeing 707 320 is 336,000 lbs., including 23,000 gallons of jet fuel, while the maximum takeoff weight of a Boeing 767-200 is some 395,000 lbs., with 24,000 gallons of fuel. (The fuel accounts for roughly half the weight of a fully loaded aircraft). Thus the 767 is not vastly larger than the 707, and it carries approximately the same fuel load. In addition, both ill-fated planes were only lightly loaded with passengers, so they did not carry their full takeoff weight. The implication is that the buildings may indeed have been designed for the impact load caused by a commercial airliner, but the designers never considered the ensuing inferno from the fuel. Suggesting that the buildings were designed for the crash of an aircraft is ultimately self-delusion-and perhaps public relations-on the part of the design team, because other aspects of a crash, i.e. the explosion and fire, were not taken into account. Perhaps the probability of such an occurrence was deemed insignificant.

From information available on the web, it appears that the weight of each building was mainly carried by an inner core of columns surrounding elevator shafts and stairways, while a dense lattice of external columns spaced 39 inches on center formed an outer tube intended principally to prevent the building from overturning when subjected to strong lateral forces, such as those elicited by hurricane winds. The floors were supported by a grid of truss beams that carried the weight of the floors to the inner core, while the floors in turn provided lateral support that prevented buckling of the columns.

The North Tower was hit at 8:46 AM above the 96th floor, and remained erect until 10:28 AM, nearly two hours after initial impact. By contrast, the South Tower was hit at 9:03 AM above the 80th floor and collapsed less than an hour later at 9:59. The damage to the latter was more severe, perhaps because the second plane traversed the building at an angle and blew off external columns on two adjacent faces. This asymmetry, combined with the greater weight of the 31 stories above the crash elevation led to some tilting of the upper portion down the damaged corner, causing large overturning forces in the remaining members of the floor.

Memorial posters in the Bldg. 10 lobby filled up with tributes to those who died on Sept. 11, including seven MIT alumni. The Alumni Office has set up a web page on so that MIT affiliates (or their families) can write in and let their friends know they are safe. Photo: Donna Coveney/MIT

The initial impact of the aircraft caused massive structural damage to the external columns, to the floors in the proximity of the impact, and perhaps also to parts of the inner core. The ensuing explosion must have significantly exacerbated this damage, possibly collapsing several floors, and setting the buildings ablaze in a virtually uncontrollable, fierce fire. Still, both buildings did not give way for a remarkably long period of time after the crash. This extraordinary capability allowed many lives to be saved, and is a major credit to the designers. Ultimately, however, the intense fire heated the structural steel elements well beyond the thermal limit of some 800° F, which caused the steel to lose resistance or even melt. Supporting members gave way, initiating the final failure of the building.

Various mechanisms may have been at play in this failure. Witnesses who escaped the buildings reported seeing large cracks develop on the walls of the staircases. This would suggest a steady redistribution of vertical forces and propagation of structural failure down the building. However, the immediate failure mechanism was almost certainly initiated locally at the elevation of the crash. Truss beams heated by the fire were probably more vulnerable than columns, and may have been the first to go. As parts of the floors then collapsed and rained down onto the floors below, the weight of the accumulating debris steadily increased beyond the support capacity of those floors, and they collapsed in turn.

At the same time, local collapse of the floors caused the heat-weakened columns to lose their lateral support, and to buckle and collapse under the intense weight of the floors above the level of the fire. At that point, the upper floors began to fall wholesale onto the structure below, and as they gained momentum, their crushing descent became unstoppable. Indeed, with two fairly simple dynamic models, I determined that the fall of the upper building portion down the height of a single floor must have caused dynamic forces exceeding the design loads by at least an order of magnitude. There was no way in the world that the columns below could have taken this large overload, and these failed in turn and collapsed, creating a domino-effect down the building. The towers then collapsed in practically a free fall.

Why did they not fall like a tree?

Some observers have wondered why the buildings telescoped down, instead of overturning and rolling to their side like a tree. Unlike trees which are solid, rigid structures, buildings such as the WTC towers are mostly open space (offices, staircases, elevator shafts, etc.). Indeed, a typical building is 90% air, and only 10% solid material. Thus, it is not surprising that a 110- story structure should collapse into 11 stories of rubble (actually less, because the rubble spreads out laterally, and parts are compressed into the foundation).

In addition, the towers did not fail from the bottom up, but from the top down. For a portion of the tower to roll to either side, it must first acquire angular momentum, which can only occur if the structure can pivot long enough about a stable plane (e.g. the stump in a tree). However, the forces concentrated near the pivoting area would have been so large that the columns and beams in the vicinity of that area would simply have crushed and offered no serious support permitting rolling. Also, both building sections above the crash site were not tall enough to significantly activate an inverted pendulum effect. Thus, the upper part could do nothing but simply fall down onto the lower part, crushing it. While photographic evidence shows the upper part of the South Tower to be inclined just as it began to collapse, it may not necessarily have rolled to the side, but instead fallen down onto the lower floors in a tilted position. (A careful review of collapse videos and additional photos should help clarify this contention.) Indirect evidence points to minimal vertical resistance to telescoping or pancaking of either tower: the duration of the collapses was nearly the same as that of an object in free fall, while any serious resistance would have slowed down the collapse. In essence then, the towers did not collapse like trees because the structures, despite their strength, were too fragile to sustain such motions.

During the dedication of the Memorial Wall on Sept. 14, members of the community set lighted candles afloat in the moat surrounding the MIT chapel. Photo: Donna Coveney/MIT

Corollary to the WTC collapse

An important lesson from the WTC collapse is that buildings are like chains in that they are only as strong as their weakest link. If the structural integrity of any floor in a building should be seriously endangered by a blast or a massive fire (perhaps excepting the very top floor or those immediately below it), that building is highly likely to collapse and pancake to the ground. However, inasmuch as catastrophic damage to all load bearing members is very rare and the vast majority of modern high rise buildings are well-engineered and designed to resist office fires (but not jet fuel fires), these buildings are and will continue to be very safe indeed.

Can we design buildings to resist collapse?

The answer to this question depends on what is meant by design. If we make buildings as solid as the containment structures in nuclear power plants, it might be possible to design not only for impact and blast forces, but also for the massive fires caused by the jet fuel. But nobody would wish to live or work in such fortresses. In addition, they would be unbearably ugly. From a practical viewpoint, the chance that any individual building out of hundreds of thousands (millions?) in the nation might suffer an attack is so small that it would not make economic sense to make them jet-crash proof. (But do not confuse this chance with the probability that some building in the US may be hit this way.) As for retrofitting existing buildings, my view is that making them jet-crash proof would make no sense whatsoever. However, it would make eminent sense to retrofit at least some buildings, perhaps as part of an overall escape system overhaul, to ensure that load bearing elements have sufficient thermal protection and the buildings can survive a fierce fire for several hours. By providing adequate redundancies in the form of both alternative escape routes and sufficient escape time, we can prevent deadly consequences to people even when we should not able to avoid ultimate structural collapse. These improvements may be needed if for no other reason other than to allay the concerns of people whose fear of a similar tragedy will persist for years to come. I, for one, would not wish to live or work in a mouse trap with insufficient escape routes.

Dr. Shyam Sunder

08 April 2005 -- Ineffective fireproofing and a shortage of staircases were the main reasons for the collapse of the World Trade Center, according to a report released Tuesday and led by Dr. Shyam Sunder, Acting Deputy Director of the Building and Fire Research Laboratory (BFRL) at the National Institute of Standards and Technology (NIST).

Dr Sunder's study is likely to have a significant effect on the future design of skyscrapers and on building regulations in the US.

The NIST report concludes that a combination of factors caused both buildings to collapse shortly after terrorists flew hijacked commercial airliners into them on 11 September 2001. Computer simulations have been used to help piece together the chain of events that unfolded between impact and the collapse of each structure.

The report says the initial collisions severely damaged several of the columns at the core of each building. Critically, they are also thought to have dislodged fireproofing on both the columns and the floors - the floors linked the inner columns to the supports on the outer structure.

"While the buildings were able to withstand the initial impact of the aircraft, the resulting fires that spread through the towers weakened support columns and floors that had fireproofing dislodged by the impacts," says Sunder, who led the NIST investigation. "This eventually led to collapse as the perimeter columns were pulled inward by the sagging floors and [became] buckled."

"The reason the towers collapsed is because the fireproofing was dislodged," according to Sunder. If the fireproofing had remained in place, Sunder said, the fires would have burned out and moved on without weakening key elements to the point of structural collapse.

At a news conference Sunder drew an analogy with the 2003 Columbia space shuttle disaster when the absence of a small piece of insulation foam - knocked off during launch - allowed fire to seep into the shuttle's entire wing span during re-entry with catastrophic results.New alternatives to traditional fireproofing should be explored, Sunder said, citing a paint-like substance which, if applied in sufficient layers, would stick "even if a plane hit it".

Nearly 3 000 people were killed in the attack on the World Trade Centre. Roughly 17 000 people were in the skyscrapers at the time of the attack, and Nist estimated that the death toll would have been closer to 14 000 if the two towers had been filled to their 50 000-person capacity.

Photographs show that the walls of the north tower to have deformed by as much as 140 centimetres just a few minutes before collapse and the walls of the south tower to have arched by 50 cm.

Sunder told a press conference that newly developed fireproofing could perhaps have sustained the structure for longer. "Even with the aeroplane impact and jet-fuel-ignited multi-floor fires - which are not normal building fires - the buildings would likely not have collapsed had it not been for the fireproofing that had been dislodged," he says.

The report further concludes that more lives might have been saved if both structures had been built with more than just three staircases. And the stairs were also surrounded by lightweight drywall that was immediately destroyed upon impact. Reinforced surrounding walls "might have provided greater opportunities for escape", Sunder says.

The complete report is comprised of more than 10,000 pages, the preliminary 3400 pages of which were released on Tuesday. The remainder of the study is planned for release in July 2005, when the institute will also make recommendations concerning building design and construction.

BFRL’s mission is to meet the measurements and standards needs of the building and fire safety communities by serving as the source of critical tools - metrics, models, and knowledge - used to increase productivity, facilitate trade and enhance public safety through technical innovations and improved codes, standards, and practices.

In his current position, Dr. Sunder also: • serves as the lead investigator for the federal building and fire safety investigation into the World Trade Center disaster; • leads NIST activities related to the National Earthquake Hazards Reduction Program (NEHRP); • oversees NIST activities related to the National Construction Safety Team Act; • guides effective implementation of the NIST strategic plan within BFRL and the four BFRL goals: Homeland Security, Fire Loss Reduction, Enhanced Building Performance, and High-Performance Construction Materials and Systems; • chairs, as designated by the NIST Director, the Interagency Committee on Seismic Safety in Construction (ICSSC) - a group that recommends policies and practices to its 32 member-agencies on improving the seismic safety of federal buildings nationwide; and • serves as the U.S.-side chair of the Wind and Seismic Effects Panel established under the U.S.-Japan Cooperative Program on Natural Resources (UJNR).

Prior to joining NIST, Dr. Sunder held a succession of positions at the Massachusetts Institute of Technology (MIT) beginning in 1980: instructor, assistant professor, associate professor, principal research scientist, and senior research scientist.

Dr. Sunder’s awards include the Gilbert W. Winslow Career Development Chair (1985-87) and the Doherty Professorship in Ocean Utilization (1987-89) from MIT, the Walter L. Huber Civil Engineering Research Prize (1991) from the American Society of Civil Engineers, and the Equal Employment Opportunity Award (1997) from NIST.

Dr. Sunder holds a Bachelor of Technology (Honors) degree in civil engineering from the Indian Institute of Technology, Delhi (1977), a Master of Science degree in civil engineering from MIT (1979), and a Doctor of Science degree in structural engineering from MIT (1981).

Twin tower collapse theory challenged

Twin tower collapse theory challenged

• 19:00 05 February 2003
• Exclusive from New Scientist Print Edition.
• Duncan Graham-Rowe

Enlarge image

The south tower, with thinner insulation was hit second but fell first (Image: SUSAN MEISELAS/MAGNUM)

The US National Institute of Standards and Technology will soon be testing a controversial theory about the collapse of the World Trade Center towers.

According to an analysis by a leading fire-safety expert, had the fire-proofing insulation on the towers' steel structures been thicker, the towers would have survived longer and might even have remained standing after they were hit by the hijacked planes. The work is being seized on by lawyers representing victims' families and insurance companies.

If confirmed, it could also lead to changes in building codes. NIST is responsible for drawing up the final report on the towers' collapses and recommending if any changes are needed.

It is widely accepted that the collapses were caused by the failure of the buildings' steel structure as it was weakened by the heat of the fires. But Jim Quintiere of the University of Maryland, College Park, thinks the thickness of the surviving fire insulation, rather than the destruction of insulation during the impacts, explains why the towers collapsed when they did.

The south tower was the first to fall even though it was hit after the north tower. The insulation on its burning floors was only half as thick. According to Quintiere's calculations, if the insulation had matched that in the north tower, the south tower would have stayed standing longer.

50 millimetres

No one doubts that the planes killed many people on impact and started the fires that led to the buildings' collapse, says Quintiere. But if both towers had had insulation over 50 millimetres thick, he says, they might not have collapsed at all. His analysis calls into question the safety of other buildings constructed to the same standards as the twin towers. However, the Port Authority of New York, the owner of the twin towers, rejects his theory.

Quintiere, whose previous work includes investigating the 1993 fire at the Branch Davidian compound in Waco, Texas, was struck by a statement in last year's preliminary report of the Federal Emergency Management Agency (FEMA). It noted that there was a discrepancy in the thickness of the fireproofing in the towers. On the floors of the south tower where the plane hit it was just 19 millimetres thick, half that on the floors struck in the north tower.

The diagonal rods in the trusses supporting the floors were particularly vulnerable, he says, since they were the thinnest structures and would heat up fastest. "The implications of these insulation differences are astounding," Quintiere says.

Together with Marino di Marzo, also at College Park, and Rachel Becker at the Technion, the Israel Institute of Technology in Haifa, Quintiere calculated how long it would take for the trusses to fail at the temperatures they were subjected to in the fires.

The results, presented in June 2002 at a meeting in New York organised by NIST, have now been published in Fire Safety Journal (vol 37, p 707). The team calculated that the south and north towers would collapse after 75 and 115 minutes respectively. In fact, they fell after 56 and 103 minutes. "It's the only calculation I've seen that has any correlation with events," Quintiere says.

Dislodged fireproofing

Frank Lombardi, the Port Authority's chief engineer, insists that the thickness of the insulation is irrelevant. He says the impacts dislodged much of the fireproofing on the trusses. Without this protection, he says, it was inevitable that the heat would make them buckle.

This view is supported by Gene Corley, who led the FEMA's investigation last year. "I do not believe the insulation was substandard," he says. However, he concedes that if extra insulation had been applied and had remained in place after the planes hit, the buildings would have remained standing longer.

Quintiere is not convinced by Lombardi's account. He thinks that unprotected steel trusses would have given way after just 10 to 15 minutes. The fact the buildings stood as long as they did suggests the insulation remained intact on many structures, he says.

The FEMA report acknowledges that it is surprising the buildings stood for so long. "The fact that the structures were able to sustain this level of damage and remain standing for an extended period of time is remarkable," it states.

FEMA investigators should have looked more closely at this issue, Quintiere says. "We have a distorted, flawed and incomplete presentation to the public of why the WTC twin towers collapsed due to fire," he wrote in an email sent to fire-safety experts.

Gaining ground

Quintiere acknowledges that further work is needed to prove or disprove his theory. But despite initial scepticism, it appears to be gaining ground. Shyam Sunder, the lead investigator at NIST, told New Scientist that his team will assess it. "We plan to conduct tests at NIST with different insulation thicknesses beginning in February and likely to end in March," he says. The results are unlikely to be revealed until the end of the investigation, which is due around September 2004.

If the NIST tests back Quintiere's theory, attention will turn to why the insulation was thinner in the south tower than the north tower. The New York City building code stipulates that the insulation on steel structures should be at least 38 millimetres thick. However, the Port Authority's special legal status means it does not have to comply with the code.

When the twin towers were built in the early 1970s, fire insulation just 19 millimetres thick was sprayed onto the trusses. But in 1996, Lombardi recommended the thickness be doubled. "I made the decision, since there was a question from a general contractor as to how much thickness is needed to provide a two-hour fire rating of the floor joists and floor assembly that would be in conformance with New York City building code," he says.

"Why would you be taking the dramatic measure of doubling the thickness if it wasn't for safety concerns?" says Brian Alexander, an attorney with Kreindler & Kreindler who is representing some of the families of victims. "They should have had double the building code requirements, given the size of the building and its design," he says. "The whole point of fireproofing in this building was to provide a certain amount of time for folks to get out." He will argue in court that they were not given this time.

Empty floors

Despite the recommendations by Lombardi, thicker insulation had been applied to fewer than a third of the trusses in the twin towers by 11 September. This, Lombardi says, was because it could only be done as floors became empty.

Six months before the attacks, the Port Authority received a copy of a report it commissioned from British consulting engineers Buro Happold to see if there was a more cost-effective alternative to applying thicker insulation. The authority declined to provide New Scientist with a copy of this report. But on being told of the report's existence, Alexander said he would be seeking a copy as part of legal proceedings.

Quintiere has also received the support of some of the families of those killed when the towers fell. Sally Regenhard, the founder of the Skyscraper Safety Campaign and the mother of one of the firefighters lost on 11 September, has said she would fund his research if NIST did not address the issue. "Right now we do not have the truth. We have people who have a vested interest in not knowing the truth," she says.

Fireproofing key to Twin Towers' collapse

Fireproofing key to Twin Towers' collapse

• 17:28 06 April 2005
• NewScientist.com news service
• Will Knight

Ineffective fireproofing and a shortage of staircases are highlighted in a preliminary federal safety report into the attacks on the World Trade Center, issued by the US National Institute of Standards and Technology (NIST) on Tuesday. The study is likely to have a significant effect on the future design of skyscrapers and on building regulations in the US.

The NIST report concludes that a combination of factors caused both buildings to collapse shortly after terrorists flew hijacked commercial airliners into them on 11 September 2001. Computer simulations have been used to help piece together the chain of events that unfolded between impact and the collapse of each structure. Several videos of the simulations can be seen here and here (both require Realplayer).

The report says the initial collisions severely damaged several of the columns at the core of each building. Critically, they are also thought to have dislodged fireproofing on both the columns and the floors - the floors linked the inner columns to the supports on the outer structure.

"While the buildings were able to withstand the initial impact of the aircraft, the resulting fires that spread through the towers weakened support columns and floors that had fireproofing dislodged by the impacts," says Shyam Sunder, who led the NIST investigation. "This eventually led to collapse as the perimeter columns were pulled inward by the sagging floors and [became] buckled."

Photographs show that the walls of the north tower to have deformed by as much as 140 centimetres just a few minutes before collapse and the walls of the south tower to have arched by 50 cm.

Multi-floor fires

Sunder told a press conference that newly developed fireproofing could perhaps have sustained the structure for longer. "Even with the aeroplane impact and jet-fuel-ignited multi-floor fires - which are not normal building fires - the buildings would likely not have collapsed had it not been for the fireproofing that had been dislodged," he says.

The report further concludes that more lives might have been saved if both structures had been built with more than just three staircases. And the stairs were also surrounded by lightweight drywall that was immediately destroyed upon impact. Reinforced surrounding walls "might have provided greater opportunities for escape", Sunder says.

Full report

But some experts remain unconvinced by the study's conclusions. James Quintiere, of the University of Maryland, US, says he does not understand how fireproof insulation could have been dislodged from the buildings' floors and columns.

"Everything I see points to the fact that there may not have been enough insulation," he told New Scientist, adding that the fuel loads used in the report's calculations may have been too low.

And Barbara Lane, leader of the Structural Fire Group at UK engineering company Arup, adds: "[We] don't believe that [the dislodging of fireproof material] has been substantiated in any of the published data to date.” She adds that it is difficult to extrapolate heat assessments of a material to what might happen when it is actually in place in a building.

Lane also questions recommendations concerning the use of thermally-resistant window assemblies to slow the spread of fire. "This is of considerable concern as even this form of glass can fail under direct flame impingement," she says.

The complete report is comprised of more than 10,000 pages, the preliminary 3400 pages of which were released on Tuesday. The remainder of the study is planned for release in July 2005, when the institute will also make recommendations concerning building design and construction.

New Evidence Is Reported That Floors Failed on 9/11

New Evidence Is Reported That Floors Failed on 9/11

By JAMES GLANZ

Published: December 3, 2003

AITHERSBURG, Md., Dec. 2 — Federal investigators said here Tuesday that new evidence supported earlier suggestions that the floor supports in the World Trade Center began failing in the minutes before the towers fell and might have played a major role in their collapse.

The investigators, who are carrying out a two-year, $16 million analysis of the collapses, made it clear that they had not yet settled on a final explanation. They said, though, that their findings gave new weight to a theory that the failure of the floors weakened the towers' internal structure to the point that the entire buildings came down.

S. Shyam Sunder, who is leading the investigation for the National Institute of Standards and Technology in the Commerce Department, said, "We are seeing evidence of floors appearing to be sagging — or that had been damaged — prior to collapse." Still, Dr. Sunder said, "The relative role of the floors and the columns still remain to be determined in the collapse."

According to an alternative theory of the collapse, the planes that smashed into the towers damaged the towers' vertical structural columns so severely that the buildings were virtually certain to fall. In that view, none of the buildings' many structural novelties — the towers were daring engineering innovations in their day — would have played a significant role in the collapses.

Last spring, the standards institute found the first photographic evidence on the east face of the south tower that a single floor — with its lightweight support system, called a truss — had sagged in the minutes before it started collapsing. Now, detailed analysis of photos and videos has revealed at least three more sagging floors on that face, said William Pitts, a researcher at the institute's Building and Fire Research Laboratory.

In addition, Dr. Pitts said, sudden expansions of the fires across whole floors in each tower shortly before they fell suggested internal collapses — burning floors above suddenly giving way and spreading the blaze below.

Finally, an unexplained cascade of molten metal from the northeast corner of the south tower just before it collapsed might have started when a floor carrying pieces of one of the jetliners began to sag and fail. The metal was probably molten aluminum from the plane and could have come through the top of an 80th floor window as the floor above gave way, Dr. Pitts said.

"That's probably why it poured out — simply because it was dumped there," Dr. Pitts said. "The structural people really need to look at this carefully."

The investigators also said that newly disclosed Port Authority documents suggested that the towers were designed to withstand the kind of airplane strike that they suffered on Sept. 11.

Earlier statements by Port Authority officials and outside engineers involved in designing the buildings suggested that the designers considered an accidental crash only by slower aircraft, moving at less than 200 miles per hour. The newly disclosed documents, from the 1960's, show that the Port Authority considered aircraft moving at 600 m.p.h., slightly faster and therefore more destructive than the ones that did hit the towers, Dr. Sunder said.

The towers did withstand the plane strikes at first, allowing thousands of people to escape, but then the fires, stoked by burning jet fuel, softened the steel of the towers. Potentially challenging other statements by Port Authority engineers, Dr. Sunder said it was now uncertain whether the authority fully considered the fuel and its effects when it studied the towers' safety during the design phase.

"Whether the fuel was taken into account or not is an open question," Dr. Sunder said. It is also unclear, he said, "whether the extent of the loss of human life as a result of that" was taken into account.

The studies of the floor trusses and the design of the towers are just two elements of the investigation, which is carrying out computer calculations of the collapses, rebuilding pieces of the towers in order to test them in real fires, and piecing together a highly detailed chronology of the response to the attack.

In one set of laboratory tests concerning the floor trusses, researchers used earthquake simulators to violently shake assemblages much like the ceilings in the twin towers. The shaking was meant to simulate the impact of the aircraft.

The findings, said Richard Gann, a senior research scientist at the Building and Fire Research Laboratory, showed that many of the fire-protecting ceiling tiles near the impact probably crumbled, exposing the undersides of the trusses directly to the fires.