The Silver Bridge Didn't Fail From Corrosion — It Failed From a Crack No One Could Ever Have Seen

Chapters

1. 0:00 5:00 p.m., December 15th, 1967
2. 0:38 A flaw no inspection could have found
3. 1:26 A bridge that hung from a chain
4. 2:04 Joint C13N
5. 2:45 The instant it let go
6. 3:11 Proving the sequence, not assuming it
7. 3:59 Ruling out the last alternative
8. 4:33 One eighth of an inch
9. 4:58 Forty years in a water pocket
10. 5:51 The edge of what the steel could take
11. 6:12 Five ordinary choices
12. 7:16 The sixth factor
13. 7:42 The bridge closed before the report
14. 8:06 398,000 bridges, no uniform standard
15. 8:38 The first National Bridge Inspection Standards
16. 9:17 Eight episodes, one pattern
17. 9:58 Next: a seal that stiffened overnight

Transcript

5:00 p.m., December 15th, 1967 0:00

December fifteenth, nineteen sixty-seven, about five in the afternoon. The U.S. thirty-five highway bridge, connecting Point Pleasant, West Virginia, to Kanauga, Ohio, collapses — the entire structure, in about a minute. The seven-hundred-foot center span goes first, then both three-hundred-eighty-foot side spans, then the towers themselves. Forty-six people die. Nine are injured. Thirty-one of the thirty-seven vehicles on the bridge that afternoon fall with it — twenty-four into the Ohio River, seven onto the Ohio shore.

A flaw no inspection could have found 0:38

Every bridge in this series failed the same basic way: a control that existed on paper, and was never run. A calculation nobody made. A checklist that was never updated to cover the part that actually broke. This one is different. Investigators checked whether the bridge had simply been overloaded that afternoon, one too many cars for a span built decades earlier — and found nothing unusual. Ordinary holiday-season traffic, well under half of what the structure was rated to carry. The final report goes further: it says the flaw that actually brought this bridge down could not have been found — by any inspection method known at the time — without taking the joint apart. So what was there left to investigate?

A bridge that hung from a chain 1:26

This bridge didn't hang from woven cable, the way most suspension bridges do. It hung from a chain — pairs of forged steel eyebars, each one a flat bar with a ring forged into either end, linked pin to pin, tower to tower, all the way from anchorage to anchorage. The steel itself was newer and stronger than older chain bridges used, which let engineers build the whole span with far fewer, thicker bars than a design from decades earlier would have needed. It opened to traffic in nineteen twenty-eight, and for almost forty years it carried U.S. thirty-five across the Ohio River without incident.

Joint C13N 2:04

One joint held an unusual amount of that design together: C13N, the first link in the north chain west of the Ohio tower. Two eyebars ran side by side there, sharing a single pin — eyebar three thirty to the north, eyebar thirty-three to the south, both cut from the same higher-strength steel as every other bar in the chain. Both of the bridge's towers rested on rocker seats, free to rotate under load rather than fixed rigidly in place. The moment that one joint let go, total collapse of the entire structure was a certainty, by the design itself — there was nothing in the geometry of the bridge that could have stopped it.

The instant it let go 2:45

At five o'clock, it let go. A brittle fracture split the lower half of eyebar three thirty's eye. A ductile tear followed on the upper half, and the bar came free of the pin entirely. Its partner, eyebar thirty-three, slid off that same pin an instant later. The north chain was cut in two, and the collapse moved eastward across the bridge in under a minute.

Proving the sequence, not assuming it 3:11

Investigators had to prove that sequence, not assume it. Wind loads, secondary stresses, fatigue from wind-driven motion — each was tested directly on this bridge's own near-twin upstream, and each was shown incapable of producing stress anywhere near what a fracture like this required. Other fractures on the wreckage itself looked like candidates for where the collapse began — one in a stiffening truss member, two more at hanger connections. Every one of them showed the signature of a high-strain-rate break: metal torn fast, under violent, sudden load. That signature only appears after a structure is already tearing itself apart. Eyebar three thirty alone showed none of it — the mark of a fracture that started a collapse, not one that followed it.

Ruling out the last alternative 3:59

One more possibility remained: could eyebar thirty-three simply have rotated off the end of its pin, no fracture required at all? Investigators examined the worn edge of its hole under a microscope, looking for the terraced wear pattern that kind of motion leaves behind. It wasn't there. A scale model, built to reproduce the actual fracture in eyebar three thirty, matched it closely. The same model, pushed deliberately to walk a bar off its pin — even with the pin tilted — could not reproduce that at all.

One eighth of an inch 4:33

With everything else ruled out, one crack was left to explain the entire disaster. It had a radius of about one eighth of an inch. A second, smaller crack sat right beside it. Both were cut into the inside surface of the pinhole, angled across the bar's own line of stress — invisible from outside the joint, and invisible without pulling the pin apart.

Forty years in a water pocket 4:58

That crack didn't appear overnight. It began at a pinpoint corrosion pit, in the one place on the joint most exposed to it. A spot where the pin and the hole rubbed against each other under load. And where the shape of the hole itself formed a narrow, tapered pocket, built by geometry alone to hold rainwater against bare steel. Engineers had measured how much ordinary stress concentrated at that exact edge of the hole — as much as three times the plain stress running through the rest of the bar, even before any crack existed at all. Two laboratories that examined the fracture called the growth mechanism stress corrosion. A third found evidence pointing to corrosion fatigue instead. The final report credits both, working together, across forty years in service — since this bridge opened to traffic in nineteen twenty-eight.

The edge of what the steel could take 5:51

The calculated failure stress for a crack that size came out to seventy-five thousand pounds per square inch — within a percent of this steel's own specified elastic limit. The flaw had grown, slowly, until it was sitting almost exactly at the edge of what this particular material could carry before it stopped behaving elastically at all.

Five ordinary choices 6:12

None of that explains why one crack took down an entire bridge. The investigation names five things that had to converge, and not one of them was unusual on its own. High-strength, heat-treated steel, letting designers use fewer and thicker bars than an older chain bridge would have needed. A high allowable stress — set at two-thirds of this steel's elastic limit — reasonable because most of the load was the bridge's own dead weight: known, steady, predictable, not the kind of load that swings wildly from one day to the next. A design practice of skipping certain secondary stress calculations wherever that load range was small, standard practice for eyebars at the time, tested only under static load in the shop. A slow crack-growth mechanism barely documented in bridge steel anywhere in nineteen twenty-seven. And a water pocket, sitting right at the point of highest stress, built into the geometry of every eyebar joint like this one, on every eyebar bridge like this one.

The sixth factor 7:16

A sixth factor made the first five fatal. This chain used only two eyebars per link. That made total failure of the link inevitable the instant one bar cracked through. With three bars, or four, the report says plainly: the failure of one might not have led to disaster at all. There was no backup path for the load to take. None was ever designed in.

The bridge closed before the report 7:42

Only one other bridge in the entire country combined every one of these factors at once: St. Marys, West Virginia, a short distance upstream, built to a closely related design by the same era of engineering practice. It was closed to traffic before this investigation was even finished — nobody was willing to wait for the final report to find out whether it had the same hidden flaw.

398,000 bridges, no uniform standard 8:06

The report also asked a harder question: how many other bridges share even part of this? At the time, the country had roughly five hundred sixty-three thousand highway bridges. Investigators found that close to four hundred thousand of them sat entirely outside the federal-aid system — county roads, rural roads, city streets, owned and inspected however the local authority saw fit, under no uniform standard at all. Ninety-four percent of those were built before nineteen thirty-five.

The first National Bridge Inspection Standards 8:38

This is the collapse that changed that. Within four months, Congress passed the Federal-Aid Highway Act of nineteen sixty-eight, ordering the first national standards for bridge safety inspection. A guide went out to every state, every county, and twenty-four hundred selected cities, with a deadline to inspect the oldest and most critical bridges first. By April of nineteen seventy-one, the Federal Highway Administration had its first National Bridge Inspection Standards on the books. Mandatory inspection, for the first time in this country, of every bridge in the federal-aid system, on a fixed schedule.

Eight episodes, one pattern 9:17

Eight episodes into this channel, one pattern keeps repeating. A theory correct only within its tested range. A checklist complete before a new shape existed. A switch never once exercised. An alarm whose silence looked exactly like good news. A connection check that existed only on paper. A calculation that nobody, anywhere in the country, was ever even asked to make. And now: a flaw with no check that could have caught it at all — answered, in the end, not by finding flaws like it sooner, but by never again building a bridge that depended on just one part holding.

Next: a seal that stiffened overnight 9:58

January twenty-eighth, nineteen eighty-six. A rubber seal in a rocket booster joint, stiffened overnight by unusual cold, on a space shuttle called Challenger. The night before launch, the engineers who built that seal recommended against flying. Hours later, their own company reversed them.

Description and sources

On 15 December 1967, at about five in the afternoon, the U.S. 35 highway bridge connecting Point Pleasant, West Virginia, to Kanauga, Ohio, collapsed in roughly a minute. Forty-six people died — the highest toll in this channel's cases so far. Thirty-one of thirty-seven vehicles on the bridge fell with it.

The bridge hung from a chain of forged steel eyebars rather than woven cable. At joint C13N, the first link west of the Ohio tower, a brittle fracture split the lower half of eyebar 330's eye; a ductile tear followed on the upper half, and its partner, eyebar 33, slid off the shared pin an instant later. Because the towers rested on rocker seats, total collapse became a certainty the moment that one joint let go.

Investigators had to prove that sequence, not assume it — ruling out other fractures on the wreckage by their high-strain-rate signature, and ruling out the possibility that eyebar 33 simply rotated off its pin. What was left was a single crack, about one eighth of an inch in radius, cut into the inside surface of the pinhole: invisible from outside the joint, and invisible without disassembling it. It began at a corrosion pit at the point of highest stress on the bar, in a water pocket built into the geometry of every eyebar joint of this type, and grew across forty years through stress corrosion and corrosion fatigue together. The calculated failure stress at that flaw came out within a percent of the steel's own elastic limit.

The report names five ordinary, defensible engineering choices that had to converge, and a sixth — only two eyebars per link, no redundant load path — that turned one crack into total collapse. Only one other American bridge combined all six factors; it was closed before the investigation even finished. Nationally, over 70% of America's 563,500 highway bridges sat outside federal oversight. This collapse produced the Federal-Aid Highway Act of 1968 and, by 1971, the first National Bridge Inspection Standards in U.S. history.

PRINT-READY, FROM THIS CHANNEL

The Failure Atlas, Vol. 01 — Tacoma Narrows · Citicorp Center · Millennium Bridge · Apollo 13 · the 2003 blackout · Hyatt Regency

https://therepository.gumroad.com/l/failure-atlas

PRIMARY SOURCES

• National Transportation Safety Board, Highway Accident Report: Collapse of U.S. 35 Highway Bridge, Point Pleasant, West Virginia, December 15, 1967, NTSB-HAR-71-1, adopted 16 December 1970 — read from the scanned original (Foreword, Synopsis, Conclusions, Cause, laboratory sections). https://www.ntsb.gov/investigations/AccidentReports/Reports/HAR7101.pdf
• Wikipedia, "Silver Bridge" — cross-checked for context only. https://en.wikipedia.org/wiki/Silver_Bridge
• StructureMag, "Silver Bridge Failure 1967" — cross-checked for context only. https://www.structuremag.org/article/silver-bridge-failure-1967-aka-point-pleasant-bridge/

The 23 technical plates in this video are illustrations generated for the channel by a diffusion image model, styled to match its cyanotype identity. They are diagrams of the system, not photographs of the hardware, and no person is depicted in any of them.

Root Cause investigates why engineered systems fail, using the official investigation reports and the primary technical literature. Sources for this episode are linked above.

The technical drawings in this video are cyanotype-style illustrations produced for the channel. They are diagrams, not photographs of the real hardware or the actual bridge. The charts and dimensioned comparisons are drawn from the figures and text in the sources listed above.

#engineering #failureanalysis #rootcause #structuralengineering #silverbridge