Tim MacFarlane's glass canopy for a subway entrance on the plaza of Raphael Vinoly's Tokyo Forum

Tim Macfarlane holds a thick book in his hand. "This has everything you need to build a building," he says. "It tells you how to use steel, brick, timber, concrete. It has everything except for glass." The book of standards he's holding, found on every engineer's shelf, explains how much weight a support material can carry. Without it, an engineer is blind, with no idea how to arrange things so a bridge, a building, or a house stays standing. With it, an engineer can build almost anything. Macfarlane dreams of the day when this book will contain one of the strongest, most versatile, most exciting materials known to man. He hopes that in the future every good engineer will be able to build with glass.

The 49-year-old Scottish structural engineer says we are just beginning to understand what glass can do. Macfarlane's recent breakthroughs prove that architects and engineers can now create structures entirely of glass. There is no longer a need to have any other material holding a building up; he imagines suspension bridges held by glass chains, or a gigantic geodesic dome of pure glass surrounding a college campus or an entire city. "That's just waiting to be done," he says. "You could build all of Buckminster Fuller's structures [out of glass]."

Ours is a pivotal time for building with glass, Macfarlane says. The last decade has been revolutionary: A new development about every other year has opened up dramatic possibilities for using glass. New or impending technologies allow glass to be switched, at the push of a button, from transparent to solid black, green, or yellow--to the image of an Impressionist masterpiece or the Jerry Springer show. And glass is no longer a passive solar-heat transmitter. It can be coated so that it reflects heat outwards, keeping a house cool but well lit all summer long. In the winter, a new type of glass can generate heat, offering a dramatic improvement over radiators or forced-air. With glass, Macfarlane says, "you could go a long way to a completely self-sufficient environment."

At a moment when many architects want to highlight structure, glass is a particularly attractive option. Often invisible, glass actually calls attention to itself when used structurally, says Guy Nordenson, an engineer who has built several structures with the material. "You can create something that looks impossible and invites you to figure it out," he says.

Consider some of the buildings in the world that have most delighted their viewers: the Crystal Palace, the Farnsworth House, I.M. Pei's Louvre Pyramid. All are structures in which glass dominates and structural steel is kept to a minimum. Glass allows a structure, or at least parts of it, to seem to disappear or float in midair. Glass amazes. For centuries--even with thick-stone construction--builders have sought to leave as much room as possible for glass.

It's tempting to say that the home of the future is only a few years away and that it will be built completely out of glass. It's an exciting prospect, but it won't happen. We are entering a period when it is becoming clear that glass has tremendous limitations. Surprisingly, they are not structural ones. Macfarlane has discovered that glass can be even stronger than steel. But if it is imagination and excitement that push glass forward, it is fear and money that keep it back. And fear and money, it appears, are stronger.

Fear exists at every point on the chain of construction decision-making. Clients don't believe glass can hold up a building. Architects can't imagine designing with only glass. Engineers don't know how to use glass. Contractors and glass manufacturers are terrified of being responsible for an all-glass structure that fails. And then there is money. Each of these players can imagine the lawsuits if they approved some building that killed people.

Even the celebrated German engineer Jörg Schlaich, famous for building beautiful walls, roofs, and domes made mostly with glass, holds his creations in place with a steel mesh. "We have not taken glass as load-bearing," he says. "I'm very reluctant to do that. Glass is very good in compression [pushing] but very poor in tension [pulling or bending]."

"He's convinced about something he's not thought through!" Macfarlane says sharply. "He shouldn't pronounce on something with such conviction without holding it out as a debate. Every piece of glass that graces your window undergoes extreme cycles of tension. His prejudice is based on lack of familiarity with the process of using glass. He's scared of it. And instead of saying he's scared, he says it's better to use glass in compression."

Macfarlane overcame his own prejudice in 1995. That's when he was asked to make a small Tokyo subway entrance out of nothing but glass. "I thought, This is completely mad. There is no way in the world you would build a 10-meter cantilever in earthquake country. I couldn't imagine it." He put the faxed proposal in his pocket and promptly forgot about it. What happened over the next few months showed how much more we can do with glass--and how unlikely we are to do it.

By the time of the Tokyo proposal, Macfarlane, who operated out of an office in London, had already made some dramatic structural breakthroughs. He'd created the first large-pane glass walkway--a sidewalk made of a single 12-foot-by-4-foot sheet of glass for a restaurant in London. Previously, glass walkways were made of small glass bricks or sheets no bigger than two-and-a-half square feet. He had also created the first structure--a small garden patio--with no steel, concrete, or brick supports at all: the beams and columns were sheets of glass.

These innovations dramatically expanded how architects could use glass, but the material was still far from viable as an alternative to steel or reinforced concrete. The Tokyo subway entrance was much more complex than what Macfarlane had done previously. This all-glass canopy was to be part of Rafael Viñoly's Tokyo International Forum, a grand conference center. The Forum included a curved-glass building several stories high in front of a plaza. Unfortunately, says Charles Lumberg, the design architect under Viñoly, the Tokyo city government insisted on placing a subway entrance in a key part of the plaza. "We didn't want anything blocking the view to the building," Lumberg says. "We had done 18 different schemes of a steel structure with glass on it. The consulting engineer said, eThis is ugly,' which embarrassed the hell out of me." The engineer said that the entranceway should be made entirely of glass. "He thought we should have a technical as well as an aesthetic breakthrough on the project," Lumberg says.

The project's engineer, Kunio Watanabe, had spent three months trying to create a stable all-glass canopy. He couldn't do it. According to Macfarlane, engineers have a hard time creating glass structures because there isn't enough information available on them. "Every other material has engineering specifications," he says. "Within seconds, you have them at your fingertips. But it never happened for glass as a structural material. How does glass break; what level of weight makes it break?"

The information isn't available because glass manufacturers--the ones who have the data--don't want engineers using it. "Glass does not let you make a mistake," says John Colvin, a scientist who is developing new glass standards for the European Union. He says manufacturers are terrified of supplying data to an engineer who would use them improperly. "The mistake can be trivial," he adds. "Glass doesn't have any yield. If you put a concentrated stress on steel, it flattens around, bends a little. Do that with glass, and you end up with broken glass."

Macfarlane ignored the fax from Tokyo for a week. Then one night he began absentmindedly sketching on a napkin while eating dinner at a Chinese restaurant. The first problem was he knew he couldn't use just one piece of glass. "I realized I had to use pieces that were joined together," he says. "I knew that one shouldn't attempt to glue them together, because all the glues I had used had ripped the surface off the glass and caused complete havoc every time. Then I realized the only other method was bolting"--something no one had done in an all-glass structure. But normal bolts would concentrate the load in one place on the sheet, and that's what makes glass shatter.

In Macfarlane's sketch, three perpendicular glass beams would support the glass roof of the canopy. The beams and roof panes would be connected at their midpoints and edges by a specially made bolt that would distribute the load evenly. Since each sheet of glass--roof pane or beam--would be bolted at four separate places, the load would not be concentrated in any single area.

Macfarlane realized two things. The first was that this idea just might work. The second was that no one had ever done anything like it. No one had cantilevered 33 feet of glass. No one had built a glass structure in which a heavy load was carried by a bolt through a small hole. He flew to Tokyo a week later.

Macfarlane's solution was a great shock to the entire Tokyo team. They had spent months failing to solve a problem that he had solved over dinner. Several team members expressed doubt that such an untested process could work. As it happened, the man directing the project for the Tokyo city government was an engineer who had a fondness for glass and a belief in its strength. He insisted that the project go forward.

The contractor, meanwhile, hoped to kill it right away. "Ultimately, he would have to take responsibility for the canopy in its entirety," Macfarlane explains. "He had no experience of imaginative leaps and no desire to have them, either. He liked to go home at night and sleep, thank you very much. With a lot of pressure from the architect and the engineer, we eventually dragged the contractor along with us. And edrag' is the word because you could see he was almost hell-bent on stopping the process at numerous points along the way."

The contractor insisted on presenting the plan to engineers at Asahi Glass, Japan's leading manufacturer. "Asahi didn't believe in the techniques," Macfarlane says. "In the first meeting, the Asahi engineer brought a cartoon along because he didn't speak English very well. He produced this roll of paper on which he'd drawn a canopy with a Japanese man lying on his back with a glass blade through his stomach. He said, eI'm very worried.' They had no framework in which they could think about the problem. They could only think of catastrophic results."

To assuage the contractor's fears--and his own--Macfarlane was given a budget of 50,000 pounds ($80,000) to fully test the canopy. After calculating the likely loads placed on the canopy during a typhoon or earthquake, Macfarlane realized that the whole project came down to one question: Could a bolt through a hole in a sheet of glass support a weight of one ton? He hired his architect friends Jon Corpe and Sheila Bull to spend months in a lab at London's City College finding out. They would test the glass by cutting holes in it, applying greater and greater loads, and seeing how much weight made the whole thing break. They were able to make special stress images, multicolor scans that show exactly where in a sheet of glass the stresses are strongest. Corpe soon realized that the holes for the bolts needed to be cut with extreme precision. The slightest imperfection would concentrate the load at one part of the hole and cause the glass to shatter. But nobody in the industry cut to such standards.

Macfarlane had learned from his earlier projects that most glass manufacturers and glaziers would have nothing to do with an idea like this; they had said no to far more modest proposals. He turned to the bravest and best glass man he knew, John Hodgson of the British company Firman Glass. "The crazier the project, the better we like it," says Hodgson, who has worked with glass since he apprenticed under his father as a boy.

Corpe pushed Hodgson, who at first felt the standards were impossible to meet, and together they discovered ways to reconfigure Hodgson's glass-cutting machine until it reached tolerances higher than had been achieved by anyone they knew. Once the holes were cut properly and without error, the tests began to amaze Macfarlane and Corpe. They attached half a ton. Three-quarters of a ton. A ton. Then, Corpe recalls, "We heard a loud bang. We all turned around expecting to see shattered glass." Instead, they saw the glass beams dangling limply but completely undamaged. It wasn't the glass that had broken; it was the steel pin holding the glass in place.

The tests showed that a sheet of glass supported by a bolt on only one side could support a load of 19 tons before it would break--that's 19 times what the project required. This figure was only reached under ideal conditions, but in every test, the glass supported from six to 10 tons. "It's astonishing that you could drill a hole in glass, put in a bolt, and hang four or five Rolls Royces from it," Macfarlane says.

The testing took months and a great deal of money. Clearly, few projects can support so extensive a process. Worse, explains Corpe, these tests apply only to this specific project. If a materials-science professor or engineer spent a year testing Macfarlane's structure through the full range of possible loads, the idea would be easily available to whoever wanted to use it. But Corpe says that almost all glass manufacturers have stripped down their research and development departments. "And what developer does any R&D?" he asks.

Macfarlane now had to convince the Japanese contractor that the project was safe enough to build. The contractor insisted on sending Asahi Glass engineers to London to make sure the canopy was being fabricated correctly. "I didn't want Asahi to find out who John [Hodgson] was because he's an East End barrow boy," Macfarlane says. "His factory looks like a scrap heap. There's broken glass everywhere. It looks terrible as compared to the Japanese-hospital conditions that their factories are run under." Eventually, the Asahi engineers demanded to see the yard. "It's way out in the East End of London, and they're getting further and further from civilization and their faces are getting longer and longer," Macfarlane recalls. "You drive into this yard and you can see that they're absolutely appalled."

But that changed when they saw the product. "There's the glass beam lying on the table, and it just looks spectacular," Macfarlane says. "Every inch of skill is visible. The Japanese immediately pounced on it and started measuring it, top to bottom. They kept on gasping. They couldn't measure an error. That was shocking to them."

Now confident, Macfarlane and Corpe sent the glass-roof sheets and beams to Tokyo, and Asahi fabricators assembled the canopy. In four years it has withstood two major typhoons and an earthquake measuring six on the Richter scale. But Macfarlane says the Japanese team never accepted the project. "Even on the day they finally took away the supports, there wasn't that kind of eruption of confidence and happiness that accompanies the fully committed process," he says. The contractor simply asked Macfarlane, "When will it break?"

Macfarlane won several prestigious engineering awards, and his canopy was praised in architectural and engineering journals around the world. Many people recognized that his solution applied not only to subway entrances but to any sort of structure someone might like to create.

"You suddenly give yourself a whole new way of joining glass to glass, and it releases a whole new set of structural possibilities," Macfarlane explains. He pauses, imagining more and more absurd glass structures. "If you want to build a dome that has a 120-foot radius out of individual pieces of glass, you could certainly do that."

This is the point in the story where you're supposed to read that the new bolting method--so elegant and brilliant--is being used on countless buildings all over the world. But as far as Macfarlane knows, it has been used exactly two other times. Last year, he used the system, again for Viñoly, to create a 60-square-yard wall of glass on the Samsung Cultural and Education Center in Seoul, Korea. And a friend of Macfarlane's has reported seeing the identical system on a canopy at the Ventura Building, designed by Davis Brody Bond Architects, on East 86th Street in New York City. "They completely plagiarized without calling me," Macfarlane says. "I would have helped them." On the other hand, such audacity is exactly the attitude he respects. "If ASI [the company that manufactured the New York canopy] have the balls to knock off my canopy, I might work with them."

In fact, Macfarlane is so eager to have companies use his idea that he didn't patent it. "People who patent things drive me nuts," he says. "People are born to share things, not to stick them in fucking safety vaults."

But the idea hasn't spread as much as it could, and the Tokyo canopy project, for all its success, can serve as a study of the reasons why. First of all, few people are aware of glass's structural potential. Clients--be they home-owners or developers--are unlikely to be knowledgeable enough about glass to suggest that their buildings be made out of it, and few architects have had the chance to work with glass structurally. The rare architect who does explore the structural possibilities of the material during the design phase is unlikely to find an engineer willing to carry out his ideas.

Even with an architect as adventurous as Viñoly and an engineer as creative as Macfarlane, the Tokyo canopy project shows how reluctant contractors will be when confronted with new and potentially hazardous ideas. But contractors might follow the lead of the one group that could create a groundswell for structural glass: the glass manufacturers themselves. There are only a handful of companies in the world that start with sand and end up with flat glass. The manufacturers, more than anyone, know how strong glass is, but they're far from eager to open up a new market.

"There's no particular incentive to change," says Adrian Ashfield, a former executive at Pilkington, England's leading glassmaker, who now works as a consultant to the glass industry. "Rather than competing with the manufacturers of different materials [such as steel], glass companies view each other as the competition," he says.

With so few glass companies all churning out the same product in the same standard sizes for the same clients, the only thing they can compete with each other on is price. "Pile it high and sell it cheap," is the industry's credo, Corpe explains. To create a market for structural glass on a big scale--and the glass industry only does things on a big scale--would cost a lot of money. First manufacturers would have to conduct the comprehensive sorts of tests Corpe hopes for. Then they would have to spend a great deal to educate contractors and others about ways glass could be used structurally. All this when their primary clients--contractors--are not at all interested in such an innovation. Meanwhile, by promoting structural applications for glass, the manufacturers would open themselves to liability lawsuits whenever structural glass failed.

All of these costs would increase the price of their bread-and-butter product: mass-produced nonstructural glass. Whichever company raised its prices first would suffer a long time before seeing any benefit from structural glass. And why do it when the industry already makes a fortune on windows, bottles, computer screens, tables, and drinking glasses?

Some believe that the way to make structural glass a reality is to encourage manufacturers to make stronger glass. Or, rather, to unlock the strength already present in glass.

"The theoretical strength of glass is between 2 million and 5 million pounds per square inch," says Arun Varshneya, professor of glass science and engineering at the New York State College of Ceramics at Alfred University. "The best of steel is around 90,000 pounds per square inch. That's counterintuitive, isn't it? Glass is 20 to 50 times stronger than steel."

The 5 million figure represents "the force [it would take] to pull one atom away from the other," he says. But glass rarely exhibits such enviable strength in practice, Varshneya adds, because "when you have a flaw in glass, it acts as a stress concentrator." He explains that any sheet of glass has countless "flaws," places where there is a space between two chains of glass atoms. These flaws are often so small they can't be seen on the strongest electron microscope, and they exist even in the best manufacturing processes. As a result, most commercially viable sheets of glass can withstand loads of only around 2,000 pounds per square inch.

Glass companies can overcome the problem created by these flaws by toughening glass through a lengthy, expensive process that effectively creates an invisible shield on the surface of the glass. Current technology, used mostly on airplanes, can produce glass capable of withstanding forces of 100,000 pounds per square inch, or more than 10 percent stronger than steel. Varshneya believes that manufacturers could be able to make a cheap version of this super-strong glass within 10 years. But it looks as though glass companies will ignore Varshneya's ideas and continue to charge exorbitant rates for stronger types of glass, prohibiting all but the wealthiest developers from using them.

Cutting costs to remain competitive, glass manufacturers have dramatically slashed research and development budgets in recent years. What research is done focuses on other areas. The 1997 "Glass Technology Roadmap," created by a consortium of glass manufacturers and the U.S. Department of Energy, spelled out the industry's research priorities. The greatest focus is on improving the process of making glass, which now requires tremendous heat, drains fossil fuels, and creates pollution. The industry is also excited about continuing to develop smart windows that can change colors or reflect or convert heat. And there are many applications for fiber optics. Some imagine that the guts of computers will soon be composed mostly of glass. There is a section of the report called "innovative uses," a sort of dreamy wish list of possible new markets for glass. In it, there is not a single mention of glass used structurally in the way Macfarlane imagines. Instead, researchers are examining how to reinforce concrete with glass threads or improve the strength of fiberglass so that it could become a structural solution. Apparently, opaque fiberglass is--both functionally and to the human imagination--far more like concrete than transparent glass is.

As Macfarlane explains, the ability to use glass structurally relies not only on major technological breakthroughs but also on subtle personal ones. "It's a common prejudice that glass is a weak, brittle material," he says. "It's a natural response. I had it. To actually believe that I could handle this as a structural material was the hardest of ideas to overcome psychologically." Now he says, "I would like glass technology to be a familiar tool for an architect to use. I would like the technology to be as cheap and familiar as it can be so that architects can use it well."

Yet for all its strength, Macfarlane says, glass is tricky. He imagines that if it were commonplace, some people might use it incorrectly, and glass structures would fail. But he's more afraid of something else. After fascinating humankind for thousands of years, what if glass structures sprouted up on every block? Widely accessible, structural glass might lose some of its wondrous appeal. "It could stop being interesting," Macfarlane says, his normally confident and fluid voice becoming uncertain. He explains that, precisely because it is so difficult and expensive, the few all-glass structures he has seen are important, beautiful works. "Glass structures can truly become significant in the hands of someone who wants to say something," he says. "A glass structure is almost sculptural, almost a piece of high art." If structural glass were to become as widespread as plywood, "it would probably be highly abused and would end up losing its potency," he admits. "It could come to a Wal-Mart near you."