Why Every Bridge in America Is Inspected Every Two Years

This is a broken piece of steel, photographed on a laboratory bench at the National Bureau of Standards.

Three-thirty was this bar's number on the bridge. The curve at the bottom is a hole, and a pin went through it.

This is the other half of the same hole.

Forty-six people died when it came apart.

This is the cover of the report that says so. It carries two dates.

December fifteenth, nineteen sixty-seven. And the day the government finished explaining it.

Three years and one day apart.

The bridge carried U.S. 35 across the Ohio River, from Point Pleasant, West Virginia, to Kanauga, Ohio. It had stood since nineteen twenty-eight.

At about five in the evening, two weeks before Christmas, it went into the river.

Forty-six people died. Nine were hurt.

Thirty-one of the thirty-seven vehicles on the bridge went down with it. Twenty-four into the water. Seven onto the Ohio bank.

The whole structure — a seven-hundred-foot center span, two three-hundred-and-eighty-foot side spans, and both towers — was gone in about a minute.

This is the story of what the government did next, and it is almost entirely a paper story.

An interim report in October nineteen sixty-eight. A public hearing in Charleston that May. And then two more years of laboratory work.

They pulled the bridge out of the river and put it back together on the bank, member by member, and drew every piece of it.

Early on, the investigators drew a chart.

They titled it "Possible Causes of Failure." Every branch is a thing the bridge might have died of.

Read the top row. Substructure. Overstress. Wind excitation. External event.

And under external event, in a government flow chart, printed for the record: sabotage, vehicle collision, sonic boom.

Point Pleasant had heard a noise. People said it sounded like a sonic boom.

So the Safety Board checked with nearby military installations. No aircraft likely to produce such effects were operating in the vicinity.

Then they did the arithmetic. The bridge chains were built to carry fourteen hundred pounds per foot at deck level — a heavier loading than most roofs are designed for.

A pressure wave big enough to drop that bridge would have taken the windows out of the town. Nobody in Point Pleasant reported so much as minor damage.

The report says the search for evidence of that cause was, therefore, abandoned.

That is what the boxes on the chart are for. One by one, they get crossed off.

The towers went first. Then pier movement. Then wind. Then the trusses, the hangers, the gusset plates.

The last row is a list of mechanisms, and every one of them works out of sight: hydrogen embrittlement, stress corrosion, corrosion fatigue, fretting.

What is left at the bottom of the page is a single chain of boxes, and it narrows.

Eyebar chain. Ohio side span. Joint C-thirteen north.

A suspension bridge usually hangs from spun wire cable. This one didn't.

It hung from a chain of flat steel bars, each with a hole punched at either end, pinned together like a bicycle chain scaled up to carry a highway.

The report's own caption for this drawing ends: "This plate shows makeup of joint C-thirteen north."

They drew the joint that failed before they knew it had.

And there is one more design detail, which the Board lists among the contributing causes.

There were only two eyebars per link. Two. Which made the total failure of the chain inevitable once one of them broke.

Had there been three or more, the report says, there would have been the possibility that the failure of one bar would not have led to disaster.

The bar that broke was numbered three-thirty. You saw it at the start.

Forty years in the weather over the Ohio River, holding up a highway.

The work was done for the Bureau of Public Roads by a metallurgist at the Bureau of Standards named John Bennett.

He photographed the fracture before he cleaned it.

Then after.

Then the surface of the hole itself, where the steel had been in contact with the pin for forty years.

And then he went inside the metal.

This is a corrosion pit, magnified two hundred times, with a crack running out of it.

The caption calls it the crack origin.

These are secondary cracks beside the fracture.

At a thousand times, the black is not empty. It is oxide, laid down in layers.

More cracks, just starting, four tenths of an inch under the face of the bar.

Bennett's first conclusion states the size of the thing that killed forty-six people.

The fracture propagated from pre-existing cracks having a depth of less than one eighth of an inch.

An eighth of an inch. Less than the thickness of a pencil.

And his third conclusion says how long they had been there. Almost all the cracks were filled with oxide — meaning they had been held open by tension for a long time.

Not an accident that happened in December. A condition that had been developing since the Coolidge administration.

Here is the Safety Board's finding, on page one hundred and twenty-six.

The cause was the fracture in eyebar three-thirty, from a critical size flaw developed over the forty-year life of the structure, by the joint action of stress corrosion and corrosion fatigue.

Then three contributing causes. The first is a kind of absolution. In nineteen twenty-seven, when the bridge was designed, nobody knew these mechanisms happened to this class of steel in ordinary rural weather.

The second: the location of the flaw was inaccessible to visual inspection.

And the third is the sentence this whole episode turns on.

"The flaw could not have been detected by any inspection method known in the state of the art today without disassembly of the eyebar joint."

Read that again. Not "nobody looked." Not "the inspectors missed it."

In nineteen seventy, three years after the collapse, with the pieces on a laboratory bench, the government's finding was that there was no way to have found it while the bridge was standing.

Which raises an obvious question about the rule this disaster produced.

Because the law had already moved, and it had moved fast.

August twenty-third, nineteen sixty-eight. Eight months after the bridge went down, and more than two years before the Safety Board said why.

Buried in the Federal-Aid Highway Act of nineteen sixty-eight, between a section on apportionment and a section on emergency relief, is a heading.

Bridge inspection. Section twenty-six.

The Secretary shall, as soon as possible, establish national bridge inspection standards. Specifying the method. The maximum time lapse between inspections. And the qualifications of the people doing them.

Before this, there was no federal answer to the question of how often an American bridge should be looked at. After it, there had to be one.

It took the Federal Highway Administration a little over two years to write it.

September twenty-fourth, nineteen seventy. A proposal, published for comment.

A hundred and forty-five comments came back. More than half the state highway departments wrote in.

Some of them objected to the inventory — to being made to count their own bridges.

It would require collecting data that were not readily available, they argued, and an expenditure of time and effort out of proportion to their value.

Others asked that the qualifications for inspectors be lowered.

The Administrator declined to do so — and instead made two changes raising them.

The final rule ran on April twenty-seventh, nineteen seventy-one.

Part twenty-five. National Bridge Inspection Standards.

A bridge, for this purpose, is anything carrying traffic over an opening of more than twenty feet.

The person running the program has to be a registered professional engineer, or qualify as one, or have ten years in bridge inspection plus a formal course.

Every state has to keep an inventory. Every structure has to be rated for the load it can actually carry, and posted if it can't carry the legal one.

And then the line that outlived everything else on the page.

Each bridge is to be inspected at regular intervals not to exceed two years.

That is where the two-year cycle comes from. Every bridge you have ever driven over in this country has a file, because of a piece of steel photographed on a bench at the Bureau of Standards.

The rule was signed a little over four months after the Safety Board adopted its report. Which brings us back to what that report actually said.

The Board called the new standards a positive step. Then it did some counting.

There were about five hundred and sixty-three thousand five hundred highway bridges in the United States. Over seventy percent of them were not in the federal-aid system at all.

Of the three hundred and seventy-three thousand on county roads, rural roads and city streets, ninety-four percent had been built before nineteen thirty-five.

The oldest bridges. Owned by the governments with the least money to fix them. Covered by nothing.

The standards would become known to those authorities, the Board wrote. But there is no requirement that such authorities implement the intent of these documents.

And then, plainly: the 1968 Act had created a situation where federal-aid bridges must meet rigorous standards while the majority of the bridges in the country are not subject to those standards, except as voluntarily adopted.

So the Board recommended that the Secretary explore how to make it apply to all of them. And that he consider federal money to repair the ones left out.

There is one more thing in the conclusions, and it is the coldest sentence in the file.

There was a second bridge in the country combining every one of these factors. It was at St. Marys, West Virginia. It had been closed since early in the investigation.

Bennett had seen it coming from the metal. There is another structure virtually identical to the Point Pleasant bridge, he wrote, so it is of great importance to determine whether anything about this eyebar was unique.

If not, one must assume that cracks may be present in other bars with a similar history.

He found nothing unique. Nothing defective about the eye when it was installed, nothing to suggest it had been damaged.

It was an ordinary bar that had done an ordinary job for forty years.

And the Board would not say it was over. Many other structures possess one or more of these factors, it wrote, and it is not precisely known what the critical combinations may be.

Which is the honest shape of it. The disaster that gave America bridge inspection was caused by something inspection could not have found.

The rule that followed is not a promise that it won't happen again. It is a decision to keep looking anyway, on a schedule, and to write down what is seen.

Five signatures. Adopted, December sixteenth, nineteen seventy.

Three years and one day after five o'clock at Point Pleasant.