By 1990 the Leaning Tower of Pisa was tipping a little further every year, and the engineers studying it worried it could fall. Italy closed it, and an international team spent the next eleven years nudging it back. They did not try to stand it up straight. They pulled out small amounts of soil from under the high side, a few buckets at a time, until the tower settled back by about half a degree. Here is why it leaned in the first place, the fixes that made things worse, and how the one that worked was done.
A tower built on soft mud
The tower is the bell tower of Pisa Cathedral. It is a hollow marble cylinder about 56 meters (183 feet) tall on its low side, and it weighs about 14,500 tonnes. Its stone foundation is only about 20 meters across and reaches no more than 5.5 meters below the ground.
That is not much of a footing for so heavy a building, and the ground under it is poor. The top 10 meters or so are soft silt and sand left by old rivers and lagoons. Below that is a layer of very soft marine clay that runs down to about 40 meters. Engineers who bored into the ground found the top of that clay layer is shaped like a shallow dish, pressed down by the tower’s weight. From that shape, they worked out that the whole tower has sunk about 3 meters into the ground since it was built.
The soil is also a bit softer on the south side than the north. So as the weight went up, the south side sank more, and the tower began to tip that way.
Two long pauses that probably saved it
Work began in August 1173. By about 1178 the builders had reached partway up the fourth level, and then they stopped. Nobody knows exactly why, though Pisa was often at war with its neighbors in those years. Work did not start again for nearly a hundred years.
That delay turned out to matter a great deal. Under the weight of the half-built tower, water slowly squeezed out of the clay, and the clay grew stronger. Engineers who later modeled the tower concluded that if the builders had kept going in the 1170s, the soil could not have carried the load and the tower would have fallen over.
Building resumed around 1272 and reached the seventh level by about 1278, when it stopped again. By then the tower was clearly leaning south. The builders tried to correct for it by making each new floor slightly taller on the low side, so the tower actually bends back toward the north as it rises, a little like a banana. The bell chamber on top was added around 1370, almost two hundred years after the first stones were laid. Even there, the builders added six steps on the south side and only four on the north to level it out.
Modern analysis found the tower was built right up to the edge of what the soil could hold. Engineers call this leaning instability. John Burland, a British soil engineer who worked on the rescue, compared it to stacking toy bricks on a soft carpet. You can go up to a certain height, but no higher, however careful you are.
The fixes that made it worse
In 1817 two British architects hung a plumb line from the tower and measured a lean of about 5 degrees. Over the next century and a half, almost every attempt to help made things worse.
In 1838 an architect named Alessandro della Gherardesca dug a walkway, called the catino, around the base so visitors could see the column bases and foundation steps that had sunk out of sight. On the south side the dig went below the water table, and water rushed in. The lean grew by more than a quarter of a degree, and the tower came close to falling.
In 1934, workers drilled into the foundation and pumped in cement grout to strengthen it. The tower lurched about 10 millimeters south. In the 1970s, pumping of groundwater from the deep sands beneath the Pisa plain made the ground sink and added about 12 millimeters more.
Careful measurements started in 1911. They showed the lean creeping up every single year, and the rate had doubled since the 1930s. By 1990 the top was moving about 1.5 millimeters a year, and the seventh floor hung about 4.5 meters out past the base.
Closed, wrapped, and weighed down
In March 1989 the medieval Civic Tower in Pavia, another Italian city, suddenly collapsed and killed four people. Italy took the warning. The Leaning Tower of Pisa was closed to visitors on January 7, 1990, and a new international committee, chaired by the Polish-born engineer Michele Jamiolkowski, was given the job of saving it.
There were rules on what they could do. Conservation guidelines for great monuments say you keep their character and history, flaws included, so straightening the tower or propping it up with visible supports was off the table. Propping it could also have cracked the fragile masonry and set off the very collapse they feared. Anything temporary had to come off again without a trace.
The first worry was the stone itself. The walls are marble on the inside and outside, but the middle is rubble and mortar with gaps in it, and the most stressed part was low on the south side. In 1992 the team wrapped a few lightly tightened steel bands around the tower near the first and second levels to hold it together.
Then they went after the lean. In the second half of 1993 they cast a concrete ring around the base and stacked lead ingots on its north side, about 600 tonnes in all. The weight pulled the tower back only a sliver, about one-sixtieth of a degree. More important, it cut the force trying to tip it over by about 10 percent.
Black September
Nobody liked the look of a wall of lead at the foot of a world landmark. In 1995 the team tried to swap it for steel cables anchored deep in the sand below the clay. To dig for the anchors below the water table, they froze the ground with liquid nitrogen.
It went badly. When the freezing stopped on the south side one night in September, the tower began leaning south at about 4 arc seconds a day. An arc second is one 3,600th of a degree, so that sounds tiny, but for a structure this close to the edge it was alarming. The work was stopped, and the team piled on more lead, bringing the total to about 900 tonnes. Committee members came to call that stretch Black September.
Taking soil out from under the high side
The idea that finally worked is called soil extraction, or underexcavation. Instead of pushing the tower north, you let the north side sink a little by removing soil from beneath it. It had been used before, notably to even out settling under the Metropolitan Cathedral in Mexico City. But no one had tried it on a tower that might fall over.
The team tested it slowly, first with small physical models, then with computer models, then on a large trial foundation built on the square north of the Baptistery. The models found a critical line about halfway in from the northern edge of the foundation. Taking soil from north of that line always helped. Taking it from south of the line made things unstable. The trial proved the point the hard way. Overeager drilling took soil from the wrong side, the trial foundation began to tip the wrong way, and it took six weeks to bring it back under control.
The trials also produced the drill. It was a screw-shaped auger inside a rotating steel casing about 18 centimeters wide, angled down under the foundation. When the auger was pulled back, it left a small hole in the soft soil, and the soil above gently closed it. Each closing let the north side settle a tiny bit.
Before anyone touched the real tower, in December 1998, the team tied safety cables around its third level. They ran about 100 meters north over two big steel A-frames, held taut by lead weights. If the tower started moving the wrong way, more weight could be hung on the cables to hold it still. They were never meant to pull it north.
A few buckets a day
The first real extraction took place on February 9, 1999. The plan was cautious, about 20 liters of soil a day from 12 holes along a narrow 6-meter stretch of the north side. The site sent readings to the engineers twice a day, and every next step came back as a signed instruction. Green, amber, and red trigger levels set out what would happen if the tower moved the wrong way.
For a few days nothing happened. Then the tower began to rotate north. On February 23 it suddenly moved 2 arc seconds south in a day, and the team held its breath. It turned out a cold northerly gale had dropped the temperature, which the records showed usually nudges the tower south for a while. When the weather warmed, it started north again.
By June 1999 the tower had come back about 80 arc seconds, and the trial phase stopped. That was enough to go ahead. Between December 1999 and January 2000 the team installed 41 extraction holes across the full width of the north side, half a meter apart. Full extraction began on February 21, 2000.
From then on the tower moved north about 6 arc seconds a day, from roughly 120 liters of soil a day. It kept wanting to drift east, so the crews took about 20 percent more soil from the western holes to keep it on a straight path. In May 2000 they began lifting off the lead ingots, a couple a week at first, and the last one came off on January 16, 2001. The final extraction was on June 6, 2001.
Half a degree
All told, the lean came back by about 1,830 arc seconds, a touch over half a degree. That is too little for a visitor to notice, but it was enough to take the foundation back from the edge and ease the strain on the stonework. The seventh floor moved about 44 centimeters (17 inches) north. By the team’s reckoning the tower was back to the lean it had in 1838, just before Gherardesca’s walkway made it worse.
Along the way the team quietly strengthened the stonework too. They filled gaps in the rubble core, added stainless steel ties where the marble face could bulge outward, and swapped the steel bands for slimmer wires set in resin. Gherardesca’s old concrete ring under the walkway was tied into the foundation, which made the footing effectively wider.
The tower reopened to visitors on December 15, 2001. Guinness World Records later measured its lean at 3.97 degrees. In 2008 engineers said it had stopped moving for the first time in its history, and in 2018 the group that now watches over it reported that it had come back another 4 centimeters on its own over the past two decades. Nunziante Squeglia, a soil engineering professor at the University of Pisa, said its overall stability was better than expected.
The soft soil has a bright side
The same mud that tipped the tower may also have protected it. The area has had at least four strong earthquakes since 1280, and the tower was never badly damaged. In 2018 a team of 16 engineers led by Camillo Nuti at Roma Tre University explained why. The soft ground under such a tall, stiff tower changes how it sways, and it slows its natural rocking so much that it does not fall into step with the shaking of a typical earthquake. As George Mylonakis of the University of Bristol, one of the team, put it, the very soil that brought the tower to the verge of collapse helped it survive those quakes.
One more story is attached to the tower. Galileo, who lived in Pisa, is said to have dropped two balls of different weights from the top to show they fall at the same speed. The account comes from a biography by his student Vincenzo Viviani, written in 1654, and historians are not sure it happened. Either way, the tower is still standing for anyone who wants to try.
Further reading
Structures: Or Why Things Don’t Fall Down — J. E. Gordon’s witty classic on how buildings, bridges, and everyday objects carry loads, and what happens when they can’t.
Why Buildings Stand Up: The Strength of Architecture — Mario Salvadori explains in plain language how structures from the pyramids to skyscrapers stay up.