The connection held for a year, and then a thousand people came to dance under it.

October 14, 1979.
39.0836° N, 94.5825° W

Crown Center was Hallmark's idea of what a company could do with the ground around its own headquarters. The card company had bought up a large tract on the hill south of downtown Kansas City and spent the 1960s and 1970s turning it into offices, shops, apartments, and open space, developed through a Hallmark subsidiary. A convention hotel was part of the plan, and the hotel was going to be the piece people looked at.

The design put a wide atrium at the center of it, a room several stories tall with a glass wall, a floor of polished stone, and daylight coming down into it. Guests crossed that room on suspended walkways rather than on enclosed corridors. Three of them ran across the open space at the second, third, and fourth floor levels. The second and fourth floor walkways were stacked, one directly above the other, hung from the same line of steel rods anchored in the atrium roof frame. The third floor walkway was offset to one side on rods of its own.

The architects were a Kansas City firm. The structural engineering was done by Jack D. Gillum and Associates. Eldridge Construction was the general contractor. The steel for the walkways was fabricated by Havens Steel Company, also of Kansas City.

The original detail for the stacked walkways was a single continuous rod running from the roof, passing through the box beam under the fourth floor walkway, and continuing down through the box beam under the second floor walkway. Each walkway would hang from a nut and washer bearing up against the underside of its own beam. Every rod would carry the weight of both walkways, but each beam would carry only itself.

The beams were not solid steel. Each one was made from two C-shaped channels turned toward each other and welded along their edges to form a hollow box. The rod passed through a hole drilled near the welded seam. What held the walkway up was the washer bearing against the bottom of that box, and the welds holding the box together.

Somewhere in the fabrication process, the detail changed. The continuous rod became two shorter rods. The upper rod ran from the roof to the fourth floor beam and stopped there, with its nut under the beam. The lower rod started at that same beam a few inches away, hung from its own nut, and ran down to the second floor walkway. Building it this way was easier. A continuous rod would have had to be threaded along most of its length so the upper nut could be run down it, and the fourth floor walkway would have had to be lifted over the whole thread. Two rods eliminated that.

It also meant the fourth floor box beam was now carrying the second floor walkway as well as its own. The load at that connection doubled. The change appeared on the shop drawings and the shop drawings came back stamped.

By later analysis, the connection as originally drawn would have carried roughly sixty percent of what the Kansas City building code required. As built, it carried about thirty.

On October 14, 1979, with the hotel still going up, a large section of the atrium roof came down. Nobody was killed. The owner brought in an outside engineering firm to look at the structure, and questions went to the engineers of record about whether the rest of the building was sound. Assurances were given. Construction went on, and the hotel opened in the summer of 1980.

It did well. The atrium became a room people came to on purpose, and one of the things the hotel put in it was a Friday evening tea dance, a big band playing on the lobby floor with dancing in front of the bandstand. The dances drew crowds from all over the metropolitan area. Older couples came to dance. Younger people came to watch. It was free, and it was popular enough that the hotel counted it as a fixture.

On Friday, July 17, 1981, the lobby held somewhere around fifteen hundred people. The dance had been going for a while. The floor in front of the band was full. Along the sides, people stood at the rail with drinks. And on the second and fourth floor walkways, dozens of people stood watching the dancing below, which is what the walkways over an atrium invite a person to do.

At about 7:05 in the evening, the fourth floor connections let go.

The nuts and washers pulled up through the bottoms of the box beams, splitting the welds, and the beams tore open along their seams. The fourth floor walkway came off its rods and dropped, and it landed on the second floor walkway directly beneath it, and both of them, steel and concrete deck together, went down into the crowd on the lobby floor. It took a few seconds. People at the far end of the room saw the walkways move before they heard anything.

The mass on the floor was a stack of two concrete slabs with steel underneath, lying flat, pinning everyone who had been standing where they came down. Water lines and sprinkler piping ripped out with them, and water began filling the lobby, which meant that some of the people trapped under the edge of the slabs were in rising water while rescuers worked.

The fire department, police, and ambulance services came in and could not lift what they found. Jackhammers were brought in to break the concrete. To get heavy equipment inside they took out the glass wall of the atrium and drove a crane through it. Physicians came in from the hospitals and set up triage on the lobby floor, sorting the living where they lay. One man who could not be freed any other way had a leg amputated in the wreckage. The work ran through the night and into the next day.

One hundred and fourteen people died. More than two hundred were injured. It was the deadliest structural collapse in American history until the World Trade Center.

The National Bureau of Standards investigated the connections and reported the following year. The finding was that the walkways came down because of the hanger rod connection at the fourth floor, and that the change from one rod to two had doubled the load there, and that even the original detail would not have satisfied the code.

Who approved the change was disputed. The fabricator's position was that it had been discussed and agreed to by telephone. The engineering firm's position was that no such approval had been given and that no record of it existed. What was not in dispute was that the shop drawings carrying the two-rod detail had gone to the engineer of record and had come back with the engineer's seal on them.

A Jackson County grand jury looked at the collapse and returned no indictments. No criminal charges were filed against any person or company.

The civil claims were consolidated and settled. The total paid out ran well over one hundred million dollars.

The Missouri Board of Architects, Professional Engineers and Land Surveyors brought an administrative case. It took years. Hearings ran through the mid 1980s, and the decision found Jack D. Gillum and Daniel M. Duncan, the project engineer, guilty of gross negligence, misconduct, and unprofessional conduct in the practice of engineering. Their Missouri licenses were revoked. The firm lost its certificate of authority to practice in the state.

The hotel reopened that autumn, a few months after the collapse, after the atrium was rebuilt. The stacked walkways were not put back. What replaced them was a single crossing at the second floor level, carried on columns that stand on the lobby floor. The building changed hands and changed names in the decades that followed, and it is still an operating hotel.

The names of the dead were cut into a memorial in a park about a mile from the lobby. It was dedicated in 2015, thirty-four years after.

The atrium is still there, and the glass wall, and the stone floor. Nothing hangs in it now.