The Citicorp Tower Fix

A 25-year-old engineering student called William LeMessurier in June 1978 to ask why his year-old skyscraper rested on four columns placed at the midpoints of its sides rather than the corners. LeMessurier explained the design, hung up, and started recalculating. Within weeks he knew his tower would topple in a quartering wind that statistics said arrived every 16 years. A hurricane was forming in the Atlantic. The repair happened at night, behind plywood, while tenants slept above. The story stayed buried for 17 years.

What was wrong

Citicorp Center, completed 1977, is 59 stories tall and stands on four nine-story stilts notched into the middle of each face. The cantilever was forced by St. Peter's Lutheran Church, which refused to move from the northwest corner. To control sway, LeMessurier installed a 410-ton tuned mass damper on the roof — the first of its kind in a tall building.

His original calculations treated wind hitting the building face-on (perpendicular loading), which the New York code required. The geometry made quartering winds (hitting a corner at 45°) the real threat: they loaded two faces at once and routed force through the diagonal bracing. After the student's call, LeMessurier ran the quartering case. The bracing was over-stressed by 40% in some members.

A second discovery made it worse. During construction, the contractor had substituted bolted joints for the welded ones in the bracing — a change LeMessurier's office approved as standard practice for a fully-welded design. Bolted joints carry less load. With the quartering-wind case and the bolted joints together, the building's failure wind was a 70 mph storm with a 16-year return period in New York. With the tuned mass damper disabled by a power cut, the failure threshold dropped to a 55-year storm.

The repair

LeMessurier told Citicorp's chairman in late July 1978. Within weeks, a plan was approved: weld two-inch-thick steel plates over each of the 200 bolted joints, working at night so office tenants would not notice. The Red Cross drafted an evacuation plan for a ten-block radius. Weather Service forecasts were monitored hourly. Hurricane Ella tracked toward New York in early September and turned out to sea at the last moment.

A press strike that began on August 10, 1978 ran 88 days and covered three of New York's largest newspapers. No reporter noticed the welders. The repair finished in October. The building's failure threshold is now once every 700 years or more for the same wind.

Why the story matters

LeMessurier did not hide the discovery from his client or his insurer. He told Citicorp, told the city's building department, told his own carrier (Northbrook Insurance settled for $2 million on a potential liability his lawyers estimated at $4 billion). He did keep it from the public and the press, with the city's cooperation. The ethical defense was that public knowledge would cause panic without preventing harm. The ethical objection is that tenants and the surrounding neighborhood were assuming a risk they had not been told about.

The case became a fixture in engineering ethics curricula after Joe Morgenstern broke the story in The New Yorker in May 1995. LeMessurier himself, by then, taught the case at Harvard, treating his own conduct as the worked example. He kept his license. His firm kept its reputation. Most retellings treat this as a vindication of professional candor. A minority view holds that the secrecy is the part that still needs defending.

What's contested

The student's identity remained disputed for decades. Diane Hartley, then at Princeton, said in 2002 she had asked her thesis advisor the question, who relayed it to LeMessurier's office; she was not told what her question had triggered until she saw a BBC documentary 20 years later. LeMessurier's published recollection credited a male student. The two accounts have never been fully reconciled.

The deeper contested question is whether a similar concealment could happen today. Building codes have tightened: ASCE 7 explicitly requires quartering-wind cases. But the systemic conditions — a contractor substitution approved as routine, a code that did not require the load case that mattered, an engineer's calculations not independently rechecked after construction changes — are not specific to 1977.

Why this has to do with other realms

The Citicorp fix sits next to concept normal accidents as a case where a system avoided catastrophe through a single person's choice to recompute. Charles Perrow's argument is that tightly coupled complex systems will eventually fail in ways no individual can prevent. Citicorp is the inverse case: one phone call, one engineer willing to take the call seriously, one client willing to spend the money. It is the story Perrow's framework does not predict. Whether that makes it a counter-example or a survivorship-biased outlier is the harder question.

An open question

How many buildings standing today have an undetected quartering-wind error or a contractor substitution that no one rechecked? The honest answer is that no one knows, because the only way the Citicorp problem was found was a student asking a question the original engineer had not thought to ask himself.

Key sources

"The Fifty-Nine-Story Crisis" by Joe Morgenstern, The New Yorker, May 29, 1995 — the article that broke the story publicly.
Why Buildings Fall Down by Matthys Levy and Mario Salvadori (1992) — chapter on the Citicorp case, written with LeMessurier's cooperation.
BBC Under Pressure documentary segment, 1996 — first broadcast attribution naming the student as a woman.
to verify: LeMessurier's own Harvard GSD case lecture transcripts, available via the school's archives.

Abhishek's take

I see a smaller version of this when a vendor changes a fabric mill after sign-off and the spreadsheet still carries the old assumption. The dangerous part is not the change itself, it is the quiet approval path that makes every later calculation look clean. On the floor, the repair is a forced reread of the buy sheet before the drop, while there is still time to reduce units or move dates.

Abhishek