For a decade now, some of the architects and engineers who helped create a row of shimmering glass skyscrapers in the heart of a hurricane zone have been haunted by the private fear that, in the words of glass subcontractor Mel Lebo, "we may have built a time bomb."

No matter how much reassuring data their wind-tunnel and stress tests spat out, no hurricane had ever roared through a skyscraper canyon. So nobody knew from experience what would happen when the wind eddies, surges and suction forces within a hurricane bounced from one glass-skinned building to another.

Eleven days ago, Lebo and his colleagues got a non-laboratory test, courtesy of Hurricane Alicia.

The buildings themselves appear to have come through with flying colors: the taller ones swayed an average of a foot or two apiece, according to design and without ill effects.

But the glass skins of a half dozen proved perilously fragile. An estimated 2,000 window panels were knocked out, a toll that eventually may force engineers to rethink many of their assumptions and codes governing the action of wind on glass.

There's already a hot whodunit debate under way. Houston architects and engineers blame the breakage on wind-borne debris, especially gravel from rooftops. Others suspect that the culprit was the classic hurricane-type suction force created when fast-moving wind swirls past large objects. Several investigations are being conducted, including one funded by the National Science Foundation.

Most of the afflicted buildings were new structures with glass that had been fortified to more than twice the stress-load requirments of the Houston building code. Although the code is considerably less stringent than Miami's, for example, it is in line with nationally accepted standards for an inland hurricane zone.

They also were tested in wind-tunnel models designed to pick up the "channeling" effects that occur when wind bounces off nearby structures.

Still, Alicia's winds were relatively modest by hurricane standards, leaving the damaged buildings well within their theoretical margins of safety. What happened, and why?

"An event like this reminds us how much we have to learn about the wind," said Professor Kishor Mehta, a wind expert at Texas Tech.

Lebo and most of the Houston architectural and engineering fraternity are less inclined to play agnostic. They say they think high winds hurled tiny pellets of gravel at the windows like bullets.

"We found gravel inside many of the offices that had broken windows, and you'll notice that the pattern of breakage corresponds to the tops of neighboring buildings which had loose gravel," said Dr. Joseph Colaco, engineer for the 55-story Interfirst Plaza Building, which lost 631 glass panels, or 16 percent of its total.

The building's northern face lost two-thirds of its panels, mostly from the 31st floor down, tending to implicate, in many peoples' minds, the rooftop gravel of the 31-floor Tenneco Building northeast of Interfirst.

"It seems absolutely clear it was the gravel," added Richard Keating of Skidmore, Owings and Merrill, the national firm that designed both Interfirst and the Allied Bank next door, which lost about 350 panels.

For his part, Lebo, who served as a glass consultant on most of Houston's skyscrapers, said he was "thrilled" that the lab theories held, but disappointed that the city continues to allow loose-gravel rooftops.

He has called for a new code, similar to one in Miami, that would require rooftop material to be secured or protected by walls. (Gravel is cheap, efficient ballast for roofing insulation, but technological innovations with single-membrane roofs may be making it obsolete.)

Not everyone is buying the flying-missile theory, however.

"The gravel could have blown into the offices after the windows were already out," said Horace Cude, deputy building inspector for the city. "And as for the breakage tending to stop above the height of a neighboring building, well, that could be where the canyon effect on the wind stopped, too."

Paul Kennon, president of CRS, Inc., the biggest local architectural firm, suspects a combination of debris and suction.

"You could get a missile hitting a window, creating a small crack that weakens it and makes it susceptible to being sucked out." In hurricanes and tornadoes, the suction force can exert three or four times the pressure on a surface as the head-on wind force.

Whatever triggered the glass storm, most engineers agree that once some breakage occurred, a kind of avalanche effect took over in a highly localized area between two or three of the hardest-hit buildings.

"It was an almost hypnotic thing to watch," said city Public Works Department spokesman Dan Jones, who observed from a police squad car. "A panel would break out, but it wouldn't fall directly to the ground. It would get whipped around in the wind, hit another panel maybe in the building across the street, and then there would be more broken glass flying around. It seemed to keep feeding on itself."

Because the storm struck Houston at dawn, with plenty of advance warning, downtown was deserted. There were no injuries from the falling glass.

Another theory implicates the whole curtain-wall construction style that has been standard operating procedure for skyscrapers for a generation.

Unlike the Empire State Building-vintage high rise, in which the masonry walls are part of the structure, modern walls are merely a cladding against the elements, tacked onto a steel girder frame with lightweight metal. It is economical, and even though there is a post-modern rebellion against the esthetic starkness of the glass box, economic forces are not likely to yield easily.

Howard Barnstone, a prominent local architect who does not design skyscrapers, finds another lesson altogether from the curious pattern of damage, which found many architecturally similar buildings faring completely differently within a seven- or eight-square-block area. He assigns great weight to the fact that two downtown jewels of his former mentor, renowned architect Philip Johnson, came away virtually unscathed, while two blocks south along Louisiana Street, two Skidmore, Owings and Merrill creations took heavy losses.

"My hunch is that the problem may have been the inadequate holding of the glass by the rubberized gasket," which sits, much like a caulking, between a window and the frame, Barnstone theorized. "The difference is that in Johnson, you have the power of an eminent architect to demand safety factors which a group of young architects working against the power of the financial hatchet men cannot match."

Keating said the Skidmore buildings were wind-tested to roughly double the building code--glass, gaskets, frame, everything. "We have consciences in the business. We build in huge margins of safety," said Lebo.

The flying-debris theory, however appealing it may be to architects, does not entirely let them off the hook. Debris is always picked up in a hurricane, and if one piece of gravel is capable of starting a glass avalanche, does it make sense to put up buildings clad entirely in glass in hurricane zones?

"Absolutely not," said Dr. Neil Frank at the National Hurricane Center in Miami.

"We can't build refrigerators," responded Kennon at CRS.

"Maybe there is something to be said for reducing exposure by reducing the amount of window on the wall," said Charles Thomsen, president of 3D/International, another major local firm. "On the other hand, I'm sitting here in a corner office, because like everyone else, I enjoy what glass gives you--a view."

"There is no reason to panic over glass buildings," said Cude at City Hall. "You can get killed just as fast walking across the street."