Antigravity in Pisa

Engineers have been tinkering with this lovable leaning bell tower for hundreds of years. Now it is so close to actually falling over that they had to try something radical

By Robert Kunzig|Tuesday, August 01, 2000

The control room of the Leaning Tower of Pisa is not very impressive, as control rooms go— just a handful of technicians and computers in a construction— site trailer. But if the tower ever decides to stop leaning and start falling, those technicians will be the first to know. Every five minutes the computers receive data from 120 sensors inside the tower that monitor its inclinations. The tower has its harmless daily moods. In the late morning it leans away from the sun, like some giant antimatter sunflower, tilting imperceptibly northwest as its southeastern side warms up and expands. At night the tower settles back to its current southward tilt of around 5.3 degrees.

It is that persistent angle that is alarming. It is bigger than it sounds or than it looks on postcards. When you walk the streets of Pisa, and the tower pops into view for the first time, it is shocking— the visual equivalent of a prolonged screech of brakes. For a split second you wait for the crash. People have been waiting for centuries, of course, and so you might reassure yourself that the crash can't really happen. After all, it is hard to imagine 177 feet and 32 million pounds of marble simply falling, in an instant, after 800 years. But some people have no trouble imagining it. "It is pretty terrifying," says John Burland, a specialist in soil mechanics at Imperial College in London. "The tower is literally on the point of falling over. It is very, very close."

Not quite as close as it was last year, though: Lately the tower has been moving ever so slightly in the right direction. From his London office Burland is supervising a delicate operation in which dirt is being extracted through thin drill pipes— the geotechnical equivalent of laboratory pipettes— from under the north, upstream side of the tower foundations, allowing it to settle toward the upright direction. The rate of soil extraction amounts to just a few dozen shovelfuls a day; anything faster might jolt the tower over the brink. Its condition is considered so precarious that it has been closed to visitors for a decade: The top leans a full 15 feet out of plumb. Burland and his colleagues on an expert committee appointed by the Italian government are hoping to bring it back 20 inches by next summer.

There are 13 members of the committee, but Burland, for this crucial operation, is the "responsible officer." Every day he gets faxes from the control room in Pisa telling him how the tower is doing; every day he sends back instructions on where to remove dirt next. He takes care to sign his messages. "That's absolutely essential," he says. "Someone's got to take responsibility. Unless you do that, you get another Black September." Burland is referring to September 1995, when it seemed for a while as if the committee, which was charged with saving the tower, might manage to knock it down instead.

In 1902 the campanile collapsed in St. Mark's Square in Venice, and the Italian government appointed an expert committee, the third, to consider what to do about the Leaning Tower of Pisa. In 1989 another medieval bell tower collapsed in Pavia, south of Milan, killing four people, and the Italian government appointed its 16th (or 17th, depending on who's counting) expert committee to consider what to do about the leaning bell tower of Pisa. Burland had never been to Pisa and little knew how his life was about to change when he took a phone call early in 1990 from his friend Michele Jamiolkowski, a geotechnical engineer at the Polytechnic in Turin. Burland remembers the conversation this way:

Burland: Michele! How are you?

Jamiolkowski: I was fine until this morning. Then I opened my newspaper and read that Prime Minister Andreotti has set up a commission to stabilize Pisa, and I'm chairman.

Burland: Oh, Michele, I'm sorry. What a terrible job!

Jamiolkowski: Keep your sympathy. Your name is there as well.

There followed a telex— it all seems so long ago, Burland says; he and Jamiolkowski both are gray-haired now— a summons to a meeting in Rome. Thus began a decade during which Burland devoted much of his energy to Pisa. He was known in his profession for another delicate excavation, in which he built a below-ground parking garage alongside the Houses of Parliament without toppling Big Ben; he is still working for the London Underground on the extension of the Jubilee Line. But he has spent more time in recent years analyzing various models of the Italian tower. One morning last spring, in his office at Imperial, he demonstrated the simplest one. Taking a cardboard box from his bookshelf, he extracted some cylindrical plastic blocks and a two-inch-thick piece of foam rubber. "The problem of Pisa," Burland said, laying the foam on his worktable and stacking the blocks on it, "is that it's not built on rock. It's built on soft clay."

Under the Tower of Pisa, under all of Pisa, 1,000 feet of sediments cover the bedrock. The sediments come both from the Arno River, which flows through the town on its way to the Mediterranean, about six miles to the west, and from the sea itself, because as recently as the Roman period the area around Pisa was still a coastal lagoon. The tower sits on 30 feet of fairly dense river silts, below which lies a 100-foot-thick layer of marine clay. Called the Pancone Clay, it is made of flat, jumbled, loosely packed particles, and it is thus especially compressible. The tower, bearing down on a foundation just 65 feet wide and 10 feet deep, has compressed it.

The first three stories— the tall ground story and the first two loggias, or open galleries— were built between 1173 and 1178. The next four loggias were added between 1272 and 1278; the belfry was finished in 1370. In other words, there were two construction delays of nearly a century— and that was lucky, because otherwise the clay would have failed right then under the growing load. "In both cases the masons stopped just in the nick of time," says Burland. "Because they left it, the weight of the tower squeezed a lot of the water out of the clay, and the clay became stronger."

It's possible they stopped because they were worried about the lean; it's certain, anyway, that the tower was leaning, right from the beginning. When the new generations of masons resumed work, at the fourth story and then again at the belfry, they tried to correct the lean by building substantial northward kinks into the tower, thus giving it a banana shape. They were trying to curve the center line of the tower back over the center of the foundations, Burland thinks, just as a child would when faced with a leaning stack of blocks. Any child who has stacked blocks on a soft carpet knows, though, that sooner or later you add one block too many. At Pisa, the belfry was one block too many.

The tower had already sunk 10 feet into the soil, according to Burland's calculations, but the belfry caused it to sink another few inches, which quickly caused a big jump in its tilt, to about four degrees. The tower tilts south because one of the shallow silt layers happens to be more compressible on that side— it has some soft clay mixed into it. Today that shallow layer has become the seat of the tower's problem, Burland believes. Analyzing data collected by previous committees, he has found that the tower as a whole, even as its tilt continued to increase, had stopped sinking in the 20th century, apparently because the Pancone Clay strengthened again. Instead, the tower is rotating: As the south side of the skimpy foundation digs deeper into that soft shallow layer, the north side is moving up toward the surface, ready to pop out like the roots of a storm-felled tree.

Every little nudge moves the tower closer to that fate. According to Burland, ever since the addition of the belfry, it has been "metastable," like a ball on a flat table. Give it a nudge and it does not come back, as it would if it were truly stable, like a ball at the bottom of a bowl. It just rolls along toward the edge of the table— toward what a geotechnical engineer calls "leaning instability."

In 1838 the tower received a big nudge: An architect named Gherardesca decided that people should be able to see the base of the tower— which had disappeared into the dirt— and so he excavated a walkway around it. The tower jumped half a degree south. In 1934 an engineer named Girometti decided to stabilize the foundations by drilling 361 holes into them and injecting 80 tons of grout; the tower jumped another 31 arc seconds. (There are 3,600 arc seconds in a degree.) More recently, the gradual increase in the tilt has been caused, Burland thinks, by groundwater rising up under the base of the tower during the annual rainy season. For some reason it pushes up on the north side of the tower more than on the south. "It starts in September, and it ends in February," Burland says. "The tower ratchets in one direction, and it never comes back. It's just moving inexorably toward falling over, and accelerating as it gets closer."

Burland and his colleagues have developed a computer model that reproduces the tower's tilt history from the 12th century on. The one thing it can't quite reproduce is the tilt of 5.5 degrees, the angle it had reached before soil extraction. At any angle above 5.44 degrees, the computer tower refuses to remain standing— which suggests how close to the edge the real one has been. On the worktable in his office, Burland slowly adds blocks to his plastic tower. It teeters as it presses into the foam foundation. At block number seven it topples.

When Jamiolkowski's committee convened for the first time in 1990, the tower was increasing its tilt by around six arc seconds a year. An equally pressing danger, though, was that its masonry wall would fail first, causing the tower to collapse on itself, as the Pavia tower had. The wall is not solid; it consists of external and internal facings of marble encasing an infill of rubble and lime mortar. The stress exerted by the weight of the building is concentrated in these foot— thick facingsÑand the tilt has concentrated it dangerously at one point in particular: on the south side, at the bottom of the first loggia. That also happens to be where the wall suddenly shrinks from 13 feet to nine feet in thickness, and where it is hollowed by the internal stairway, which spirals around the tower inside the wall and arrives at the first loggia on the south side. In 1990 the external facing there was already badly cracked.

The tower was threatened with a hernia— and the first solution, says Jamiolkowski, was "like a belt for your belly." In 1992 the committee ordered the installation of 18 plastic-sheathed steel tendons around the first loggia and the ground story, pulled tight to hold it together. Early this year workers finally finished the rest of the committee's wall-strengthening program, which included injecting grout into the wall to fill air pockets in the infill and inserting stainless steel bars between the inner and outer facings to tie them together.

The committee also decided that they had to take some simple, temporary measures to stabilize the lean, to give themselves time to develop a long-range solution. If the north side of the foundation was rising, as Burland had found, there was an obvious option: Add a counterweight to stop it. In 1993, 600 tons of lead ingots were stacked on the north quarter of the tower, atop a concrete ring cast around the base. "For the first time in the history of the tower the tilt was stopped," says site engineer Paolo Heiniger. By the summer of 1994 the tower had moved some 50 arc seconds north, around two thirds of an inch.

The counterweight worked, but it was also very ugly. Six years later the ground floor of the tower remains obscured on the north side by that 15-foot pile of lead and concrete. The committee, which includes art restoration experts alongside its engineers, started to worry about this ugliness soon after creating it. In an effort to remove the pile, they came close to bringing the tower down.

By 1995 Burland had done much of the research to develop a permanent solution: soil extraction. It was not a new idea, having been suggested as early as 1962 by an engineer named Fernando Terracina. At Imperial, Helen Edmunds, a student of Burland's, had built a simple scale model of the tower on a bed of sand and sucked sand from under the model with a syringe. She found that, as long as she kept the point of the needle north of a certain line, there was no danger of the tower being inadvertently tipped into oblivion.

But a large-scale field test still needed to be done, and then a test on the tower itself, and it was all taking a long time. The committee had endured funding troubles and ministerial turf squabbles and periodic lapses in its mandate; the Italian parliament had never gotten around to ratifying the decree that had created the committee in the first place. Some members began to fear that the committee would go out of business, with the lead blight still in place as their one legacy to Pisa.

An idea for a new, temporary solution popped up: Why not replace the lead weights with 10 anchors buried 180 feet underground, in the firmer sand below the Pancone Clay? The anchors would hang from cables attached to yet another reinforced-concrete ring, this one hugging the foundations underneath Gherardesca's sunken walkway. To install it would require digging under the walkway and under the shallow water table. The committee knew that digging the walkway had caused the tower to lurch back in 1838, but they figured it would be safe to excavate their own trench in short sections. To avoid a groundwater escape that would flood the trench and possibly cause the tower to lurch again, they decided to freeze the ground first by injecting it with liquid nitrogen. The procedure worked on the north side of the tower. In September 1995, at the beginning of the rainy season, when the tower is at its most mobile, freezing started on the south side.

"The operation," says Heiniger, "had unexpected effects. The tower showed a tendency to move south, a tendency that developed quite suddenly." South was the wrong direction for the tower to be going.

"It was hair-raising, really," says Burland, who rushed out of a conference in Paris to fly to Pisa. "As soon as they switched the freezing off, the tower began to move southwards at a rate of four to five arc seconds a day, which is the normal rate for a year. For three weeks we were watching the tower day and night." Burland suspects that by freezing the groundwater under the walkway on the south side, he and his colleagues had compressed the soil underneath— water expands when it freezes— creating a gap for the tower to settle into once the freezing stopped. Ultimately, though, another 300 tons of hastily added lead halted the southward excursion, and the tower shifted only seven arc seconds.

The committee now faced loud criticism. Piero Pierotti, an architectural historian at the University of Pisa, told The Guardian, a leading British newspaper, that Burland had done "incalculable damage" to the tower. "I just hope for the sake of the good people of Britain," he added, "that he doesn't do to your Big Ben what he has managed to do to the Leaning Tower." James Beck, a professor of art history at Columbia University, compared the Pisa committee to the Keystone Kops— and also to Mussolini, for the committee's supposedly authoritarian disregard of outside criticism.

Meanwhile, Jamiolkowski was finding he had plenty of internal dissension to deal with. The government disbanded the committee for most of 1996, and when it was finally reconstituted with many new members, there was heated debate on how best to proceed. "To keep together a large group of university professors is quite a difficult task, especially when these university professors must make important decisions," says Jamiolkowski. "I believe after this experience I will come to New York and open a psychoanalytic practice."

For the moment, the argument seems to be over; what the committee is doing now is working. In 1998 they added one more ugly prophylactic to the tower, intended to catch it should anything go drastically wrong while soil is being extracted. Two steel cables looped round the second loggia were attached to giant anchors partially concealed behind a neighboring building. The final underexcavation program began in February. "There are no more polemics at the moment," Pierotti says. "People have accepted this solution."Forty-one drill pipes are now arrayed around the north quadrant of the tower.

They enter the soil at different points along an arc about 40 feet from the tower and at an angle of 30 degrees; their tips lie about 12 feet under the north edge of the foundation. Inside each eight-inch-diameter pipe is an augur, a corkscrewlike bit that traps soil between its blades and channels it to the surface. The tower then settles into the resulting yard-long cavities. Burland steers the tower, and tries to keep it on an even northward course, by deciding how much soil to remove through each pipe on any given day. As of late May, Heiniger's crew had removed more than 10 tons of soil. The tower had rotated 513 arc seconds north, and the crew was one third of the way to its goal. There had been no ominous lurches.

Every day now the workers wind the tower's tilt clock back by months or even as much as a year. By next summer the committee hopes to restore the tower to five degrees, an angle it last saw early in the 19th century. That should buy the tower roughly two centuries of stability. Visitors are not likely to notice a half-degree decrease in tilt. The mayor of Pisa hopes to reopen the tower next year on June 17, the feast of San Ranieri, the city's patron saint.

But they are not there yet, the tower-savers. Jamiolkowski is looking forward to closing the work site, disbanding his fractious committee, enjoying life— but he balks superstitiously when you mention how well things seem to be going. Heiniger points out that the greatest threats to the tower have always come from people trying to give it a friendly nudge. "I hope it won't happen this time," he says. Burland, in the driver's seat, has perhaps the most reason to carry a rabbit's foot. Everywhere he goes these days the faxes from the control room follow him— two a day telling him how the tower has responded to the latest gentle suctioning underneath it. Every night Burland sends back the next day's instructions, signed.

"It's kind of taxing," he said recently, scanning the first fax of the day as he rode the elevator up to his office at Imperial. "It's like trying to ride a bicycle by fax. It's such a dangerous structure, and so many people have come unstuck on it. But yesterday was very good. We got the biggest north movement yet: four arc seconds in a single day."

The Leaning Tower of Pisa official Web site ( has all you could possibly want about the tower.

See the Unofficial Leaning Tower of Pisa site for tower humor and plenty of serious info:

Comment on this article

Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar

Log in to your account

Email address:
Remember me
Forgot your password?
No problem. Click here to have it emailed to you.

Not registered yet?

Register now for FREE. It takes only a few seconds to complete. Register now »