The Skyscraper That Almost Fell: Inside Citicorp Center's Secret Fix

A Skyscraper Built Around a Church That Refused to Move

The site at 53rd Street and Lexington Avenue in Midtown Manhattan came with a condition attached. St. Peter's Lutheran Church owned the northwest corner of the block, and it would sell Citicorp the rest of the site only if the bank tore down the old church and built it a brand-new one — on that exact corner, untouched, with no columns from the tower passing through or above it. Citicorp agreed. Then it handed the problem to a structural engineer named William LeMessurier: design a 59-story office tower that leaves one entire corner of its footprint empty.

LeMessurier's solution, reportedly sketched on a napkin at a Cambridge restaurant, was to abandon the corner columns every skyscraper is normally built on. Instead, he placed four massive columns at the center of each face rather than at the corners, lifted the tower nine stories into the air on top of them, and let the northwest corner cantilever 72 feet out into open space — directly over the new church's roof. To carry the building's full weight down through those unconventional, off-center supports, he used a skeleton of giant inverted-V "chevron" braces running up all four faces, funneling every floor's load diagonally inward to the four stilts below. It was elegant, it was unlike anything built before it, and it made Citicorp Center — later Citigroup Center, now known by its address, 601 Lexington Avenue — the seventh-tallest building in the world when it opened in 1977.

The building was so lightweight and flexible for its height — a deliberate choice, since a lighter frame is a cheaper frame — that it would sway noticeably in strong wind. LeMessurier's fix for that was itself a first: a 400-ton block of concrete installed on the 59th floor, floating on a film of oil and nudged by hydraulic pistons to slide in the opposite direction of the building's sway. It was the first tuned mass damper ever installed in a skyscraper in the United States, and it worked well enough that Citicorp Center became a case study in structural engineering courses almost immediately. Nobody, including LeMessurier, expected the next chapter of that case study to be about something the building's own engineer had missed.

The Phone Call That Undid a Year of Confidence

In June 1978, a phone call reached LeMessurier's Cambridge office from an engineering student — later reported to be Princeton undergraduate Diane Hartley — working on a thesis about the tower. The question was simple: why were the building's columns placed at the center of each face rather than at the corners, and had that choice been checked against wind hitting the building diagonally, at 45 degrees, instead of straight-on? These diagonal gusts are known in structural engineering as quartering winds.

LeMessurier gave a confident answer on the spot — the column placement was, if anything, better suited to resist quartering winds, he told the student. New York City's building code at the time didn't even require diagonal wind cases to be checked; only perpendicular, straight-on wind loads were mandatory. But the question stuck with him. Preparing what he thought would be an interesting lecture for his own engineering students, LeMessurier sat down to actually run the quartering-wind numbers on his own building for the first time since it opened.

A 40% increase in force on a properly welded joint, designed with a normal margin of safety, usually isn't fatal to a structure — steel has some give. But LeMessurier's braces weren't all welded anymore. Sometime during construction, structural steel supplier Bethlehem Steel had proposed replacing the chevron braces' full-penetration welded joints with bolted connections instead — cheaper, faster to install, and, under ordinary perpendicular wind loads, perfectly adequate. LeMessurier's own office had approved the substitution as a routine, minor change. LeMessurier himself, by his own later account, never personally re-checked it against the diagonal wind case, because that case had never been part of the design brief in the first place.

The Math That Should Never Have Been Allowed to Add Up

Once LeMessurier combined the two facts — bolted joints instead of welded ones, and roughly 40% more force on some of those joints under quartering wind — the increase in stress on the bolts themselves worked out closer to 160%, because the bolted connections had far less reserve capacity than the welds they replaced. He brought his numbers to Alan Davenport, a leading wind engineer, who confirmed the worst version of the finding: gusts of around 70 miles per hour, striking the building diagonally, were enough to shear the bolts and start a progressive failure of the bracing system. New York experiences wind of that strength, on average, about once every sixteen years — squarely within hurricane season, which was roughly six weeks away.

LeMessurier later said he considered three options: stay silent and hope, resolve the matter through his liability insurance and effectively end his career quietly, or disclose the flaw and fix it. He described weighing whether he could live with the outcome of each. He chose the third option.

Project SERENE: Fixing a Skyscraper at Night, in Secret, While It Stayed Open

LeMessurier first notified his professional liability insurer, then the architect Hugh Stubbins, then Citicorp's own executives, including John Reed and Walter Wriston. Rather than evacuate or shut the tower down — which risked triggering exactly the public panic everyone wanted to avoid — the decision was made to repair all 200-plus bolted joints in place, welding two-inch-thick steel plates over each one, while the building continued operating normally during business hours.

The work, internally referred to as Project SERENE, ran through nights from August into October 1978. Welders worked after the offices emptied out; carpenters and cleanup crews worked the following mornings to erase any visible trace before employees returned. An emergency backup generator was installed to guarantee the tuned mass damper never lost power, since a wind-driven collapse became far more likely if the damper went offline during exactly the kind of storm the fix was racing to prepare for. Engineers monitored strain gauges around the clock, and the National Weather Service supplied forecasts multiple times a day.

Roughly halfway through the repair, the timeline nearly ran out. Hurricane Ella formed on August 30, 1978, strengthened into a Category 4 storm, and appeared to be tracking toward the U.S. East Coast during Labor Day weekend. City officials quietly prepared an evacuation plan for the blocks around the tower, ready to be activated with almost no public notice. By the time forecasters confirmed Ella had turned north and out to sea on September 13, engineers estimated the partially completed repairs already let the tower withstand roughly a 200-year storm — better than the original design, even unfinished. A newspaper strike affecting several major New York papers around this period meant the drama played out with far less press scrutiny than it otherwise would have.

The Crisis, Month by Month

What the Original Design Assumed vs. What the Building Actually Needed

The uncomfortable engineering lesson isn't that LeMessurier made an error — wind engineering in the 1970s genuinely didn't treat quartering winds as a standard check, and the code didn't require it. The lesson is what happened after the design was finished: a single mid-construction substitution, reasonable on its own terms and approved without incident, quietly invalidated an assumption nobody thought to re-examine until a student asked an inconvenient question.

Why This Matters Even If You'll Never Design a Skyscraper

Almost nobody using Estima's estimator is designing chevron-braced towers on center-column stilts. But the failure mode underneath the Citicorp Center crisis shows up constantly at every scale of construction: a change made for good, ordinary reasons — a cheaper material, a faster method, a substitution nobody flagged as significant — quietly invalidates an assumption made earlier in the design, and nobody re-checks it because on paper, nothing looks different. A swapped beam size, a downgraded fastener spec, a foundation detail value-engineered out during bidding: these rarely fail loudly at the moment they're approved. They fail later, under a load case nobody thought to ask about.

The deeper lesson from Citicorp Center isn't really about wind loads at all — it's that the cost of re-checking an assumption is always smaller than the cost of discovering, after the fact, that it was wrong. That instinct is worth carrying into a project of any size, at any budget: get a complete, honest estimate of what a project actually requires before a single substitution gets waved through as "basically the same."