On the night of May 12, 1976, the southern half of the longest steel arch in the world shifted. It was past midnight over the New River Gorge in West Virginia, and both halves of the arch had been reaching toward one another for nearly a year, jutting out from the cliff tops, held against gravity not by cranes but by a web of steel cables pulling them backward. Then a pressure pipe inside the southern support system burst. The sound of the break carried three miles across the valley.

The southern half of the arch moved. The gap between the two halves closed by a few inches. Come morning, the governor would be arriving by helicopter. The bridge was supposed to be finished, the arch closed, the ribbon cut.

Instead, engineers faced a structure that had settled into a position no one had calculated, with dawn and rising temperatures threatening to close the gap until the steel tore itself apart. The closing had to happen in the cold of early morning, before the sun heated the metal, or not at all. What follows is the story of how that arch came to be hanging there in the first place—and how the men on the bridge got it closed in time. In 1974, when construction began, there was not a crane on Earth that could reach high enough or lift heavy enough to place steel at that height above the valley floor.

The New River cuts through the Appalachian plateau to a depth of nearly 800 feet. The walls are almost vertical, and the valley floor is too narrow for the staging area any conventional bridge would need. Below the cliff tops, abandoned coal mines had hollowed out the ground, making tall supports nearly impossible without first filling the voids with cement. A cable-stayed design would have required towers so tall and so costly they would have blown the budget.

A simple truss on intermediate supports meant lifting steel from railroad cars at the bottom of the valley up to positions 900 feet above, piece by piece, at a site with no road access. The designers, Clarence Knudson and Frank Kempf of Michael Baker, spent years on calculations equal to one person working full-time for fifteen years before committing to a steel arch spanning the full 1,700 feet between the cliff tops, touching nothing in between. The material they chose added another constraint. Cor-ten weathering steel forms a protective rust-colored layer when exposed to the atmosphere, which meant the bridge would never need painting—saving hundreds of thousands of dollars up front and about a million dollars each time a conventional bridge would have required a repaint.

But it also meant every stress on every piece during construction had to be calculated precisely, because flaws in unpainted, weather-exposed steel accumulate on the surface and never get hidden by a coat of paint. The design called for steel plates four inches thick at the arch joints, with the largest joint containing 2,300 bolts. These were not decorative numbers. Every bolt existed because the engineers had calculated exactly what the steel would do.

Before any part of the arch could be installed, the American Bridge division of U. S. Steel erected two 330-foot towers on opposite sides of the gorge. A helicopter carried the first test cable across the valley—half an inch in diameter and 5,000 feet long, strung at treetop height.

Workers used it to pull progressively heavier cables across until load-bearing cables three inches in diameter hung between the towers. The final system was a pair of aerial cables, each 3,500 feet long, capable of lifting 50 tons. Used together, they could move arch pieces weighing 90 tons. Everything that went into the bridge—every beam, every bolt, every part of the truss structure—crossed the air on those cables.

The ironworkers, members of Local 301, rode to their work positions in a low-walled steel box suspended from the cable, lowered onto the structure at heights where the bridge deck narrowed from about 40 feet to 15 feet at the arch tip. There was no scaffolding beneath them. No safety net. The valley floor was 850 feet below.

These were structural ironworkers, trained on the high steel of skyscrapers, used to narrow beams and open air. That experience was not accidental. It was the reason they were there. The backstay system that kept each half of the arch standing on its own is the mechanism most people never see.

Imagine the southern half of the arch jutting outward from the cliff top. The farther it extended over the gorge, the stronger the pull at its base. Without support from below, gravity would eventually rotate the entire half-arch downward. The backstay cables prevented this by connecting the advancing tip of the arch to a fixed point behind it.

The completed approach span above the cliff top was used for this, with heavy wire ropes running forward over a rocker device mounted on the leading edge of the span, then back to jacking rods that could tension the cable precisely. Each new section of arch added weight and changed the load distribution. Engineers recalculated, and workers retensioned the jacks. The arch did not simply grow outward.

It was constantly rebalanced, section by section, against a web of cables pulling backward toward solid ground. The geometry was also deliberately wrong. Each half was built slightly higher than its final design height. This was not a mistake.

When the two halves finally met and the backstays were released, the arch would settle under its own weight into the correct position—a deflection calculated in advance on an old central computer. The process required modeling every temporary loading condition during construction, not just the finished bridge. This was the most complex part of the entire calculation effort. By the spring of 1976, nearly a year after the arch erection began, the gap between the two halves had narrowed to the distance waiting for the final closing pieces.

On the night of May 12, the project figures, including the construction planning manager for American Bridge, were having dinner at the nearby Hawks Nest park. Around seven in the evening, the American Bridge officials quietly disappeared. The other guests did not seem to notice. A pressure pipe inside the backstay anchorage on the south side had broken.

The sound carried three miles. The southern half of the arch moved. The gap closed by a few inches into a position the engineers had never planned for. Throughout the night, calculations continued without stopping.

The structure was stable, but steel expands when it heats. If the sun rose in the morning, the lengths of the arch members would increase. The gap would narrow further, and the stress state would spiral out of control. The closing had to happen in the cold of early morning before the temperature rose, or not at all.

The governor’s helicopter was already on its way for the opening ceremony. The ironworkers were told everything. They agreed to proceed. When the crane cabin lowered them to their position midspan, they stopped.

The sound of the structure was different. It moved in ways they had not known before. The pipe had failed. The arch was settling into a new, uncalculated position, and they could feel it.

They needed someone standing beside them. Superintendent Pete Spaulding and the construction planning manager climbed into the cabin and rode themselves to midspan. They stood with the crew in the open air above the river. The crew went back to work.

Before the governor’s helicopter landed on the south landing pad, the final steel pieces were already moving through the air on the crane cable. The bolts were installed. The backstays were released. The arch settled almost quietly into its design geometry, exactly where the computer had predicted it would be.

The arch was closed, but it was not yet a bridge. Thirteen vertical steel towers were then erected symmetrically from the arch center outward, ranging in length from 26 to 305 feet, each carrying the bridge trusses on top. The final roadway steel was placed in November 1976, and in the spring of 1977, crews poured more than 1,000 tons of concrete for the roadway, which also served as a dead load pressing down on the arch. The bridge opened to traffic on October 22, 1977.

Sixteen thousand cars cross it every day now. A trip that once took 40 minutes on winding roads down into the gorge and up the other side now takes 60 seconds. No one sitting in traffic on Route 19 thinks about what is holding the arch up. Early maintenance work revealed a problem no one had fully expected.

The road salt spread on the deck to melt ice was corroding the Cor-ten steel, which depends on cycles of moisture to form its protective layer properly. West Virginia eventually switched to calcium magnesium acetate, then to a liquid potassium acetate solution that can be sprayed while vehicles are moving without blowing off the surface. The material that made painting unnecessary created its own maintenance puzzle. Later arch bridges, including the Lupu Bridge in Shanghai, which surpassed the New River Gorge span in 2003, used the same principle: cantilever construction from both sides simultaneously, each half held by backstay cables, then released when the arch was complete.

What changed was the measuring tools, the rope crane technology, and the speed of calculations. What remained constant was the basic physical logic that Knudson’s team had worked out in those 4,759 sheets of calculations. The original Cor-ten steel arch members are still in service. A deck overlay has been added.

The closing splice—the bolted joint where the two halves finally met in the cold air before sunrise on the morning of May 13, 1976—is still there, 876 feet above the river, carrying every car that crosses.