They Came to Watch It Gallop

▶️ Prefer to watch? Every story's a short video → youtube.com/@GetSafetyIntel · 📣 Got a near-miss? Share it anonymously — link at the bottom.

I'll be straight on the date: our exact day has no strong, well-documented workplace event, so we're using a landmark from this same week — and it is one of the most useful failures an engineer or a safety leader will ever study.

The Tacoma Narrows Bridge opened to traffic on July 1, 1940. It collapsed four months and seven days later, a little after 10 o'clock on the morning of November 7.

Here is the part that matters. It did not fail without warning. It had been moving from the day the deck went up. Construction workers named it "Galloping Gertie" because a mild wind would make alternate halves of the center span visibly rise and fall several feet over four or five seconds. Drivers would watch the car ahead of them disappear into a trough and come back up. People drove out to ride it on purpose. It became an attraction.

The bridge was a known-moving structure that the public had been invited to enjoy.

Things were tried. Tie-down cables were run from the plate girders to fifty-ton concrete blocks on shore — they snapped shortly after installation. Inclined cable stays were added from the main cables to the deck at mid-span; they stayed in place until the collapse and did nothing. Hydraulic buffers were fitted to damp the deck's motion, and then the seals were destroyed when the bridge was sandblasted before painting, so the buffers were dead long before the wind came.

And the engineering was underway. The Toll Bridge Authority had hired Professor Frederick Burt Farquharson at the University of Washington, who built a 1:200 model of the whole bridge and a 1:20 section of the deck. His first studies concluded on November 2, 1940 — five days before the collapse. He proposed two fixes: drill holes through the girders to let air through and kill the lift, or add fairings and deflector vanes to the deck. The first was not favored. The second was not carried out.

On the morning of the 7th, in a wind of roughly forty miles an hour, the motion changed character. The gentle vertical galloping everyone had grown used to became a violent torsional twist — the two halves of the deck rolling in opposite directions, the roadway tilting past 45 degrees. WSDOT's own account names the mechanism torsional flutter: a self-feeding instability where the structure's motion pulls more energy out of the wind, which drives more motion. Not a gust that was too strong. A structure that fed itself.

Two vehicles were on the span. Leonard Coatsworth was thrown to the pavement and crawled several hundred yards to the towers. A delivery truck tipped over; its two occupants got out and walked off. Nobody was killed. The only death was Tubby, a dog trapped in Coatsworth's car, who bit the hand of a man who came out to save him. The deck went into the water.

Strip the suspension cables off this and you have the most common shape of industrial disaster there is.

The hazard was visible for four months. It had a nickname. The nickname made it familiar, and familiar is the most dangerous thing a hazard can become. Nobody walked past that bridge thinking "unresolved structural instability." They thought "that's just Gertie." You have one of these. Every site does — the pump that always runs hot, the door that has to be shouldered, the alarm everyone knows is nuisance. The moment a deviation earns an affectionate name, it has stopped being a finding and become a feature of the landscape.

Every barrier failed quietly and nobody re-checked. The cables snapped. The stays were useless. The buffers were killed by the paint crew — a crew doing their own job correctly, who had no way to know they were destroying a safety control, because to them it was a seal on a piece of equipment. That is the purest systems failure in the whole story: a control defeated as a side effect of unrelated, competent work, with no process to notice.

And the answer arrived in time and sat there. Farquharson's study finished five days out. The fix was known, specific, and not done. Procurement, cost, review, the fact that the bridge had stood for four months and would presumably stand for four more — whatever the reason, the gap between "we know what to do" and "we did it" is where the deck ended up. In your world that gap has a name: it's the corrective action with a due date and no owner.

Where it lands today. None of this was OSHA's — the OSH Act wasn't signed until December 29, 1970. But the discipline it created is yours now: 29 CFR 1910.147 for controls that must stay effective, 1910.303 and the general duty clause for equipment maintained in safe condition, and above all a real management-of-change process. OSHA's only hard MOC requirement sits inside Process Safety Management at 29 CFR 1910.119(l), which means that if your facility is outside PSM scope, nothing in federal OSHA obliges anyone to tell you when a control gets modified or destroyed. The sandblasting crew would be fully compliant today. You have to require that yourself.

New section, same five questions every day. Barriers, not people.

(E) A steel and concrete deck in a tidal strait, plus whatever was in the paint in 1940 — lead, almost certainly. No one was tracking it. The same collapse today triggers a sediment survey, a debris-recovery plan and a notification clock.

(H) Zero fatalities and almost no injuries, which is exactly why it is safe to study without flinching. But don't miss the responder angle: men walked out onto a deck twisting past 45 degrees to try to reach a trapped animal. Spontaneous rescue into an actively failing structure is a predictable human response and a recognized killer. Your emergency plan should name who is authorized to re-enter, because somebody will go.

(S) The whole lesson is above, and it reduces to one line: a structure can be strong enough to carry traffic and still be feeding energy into its own destruction.

Miners sue over a silica standard that still doesn't exist (reported Nov. 6, 2026). The United Mine Workers of America and a coalition of current and former miners filed an emergency petition on Oct. 28 with the Supreme Court of Appeals of West Virginia, asking the court to force the state's Board of Coal Mine Health and Safety to act. The claim: the board has never established a respirable silica standard despite previous requirements to do so. The petition asks the court to compel a meeting, a vote on a limit and an enforcement program within 180 days — and if the board still won't, asks the court itself to set a PEL of 25 micrograms per cubic meter over an 8-hour TWA. Backdrop: MSHA's April 2024 rule cut its silica PEL to 50 µg/m³, matching OSHA's 2016 limit, and enforcement has been delayed amid litigation in the 8th Circuit, with the metal and nonmetal compliance deadline postponed indefinitely.

So what for safety leaders: don't wait for the number to settle. Silica dose is cumulative and irreversible, and the argument in front of that court is about whose job it is to write a limit — not about whether the dust hurts people. If your exposure assessment is older than the 2024 rule, it's stale. Go sample.

The HazCom deadline is twelve days out (reported Nov. 2, 2026). Employers using covered products have until Nov. 20 to update hazard communication programs, workplace labels and training for substances under OSHA's revised 1910.1200, which OSHA aligned with the seventh revision of GHS in 2024. Manufacturers, importers and distributors had to update substance labels by May 19, 2026, so the new labels are already arriving at your dock. Mixtures run on a separate clock: manufacturer labels by Nov. 19, 2027, and the employer deadline May 19, 2028.

So what: this is the compliance item most likely to bite you this month, and it is unglamorous. If the SDS and the label changed and your written program and training still describe the old pictogram set, you are out of date on the 21st. One afternoon fixes it. Also worth saying plainly: handing someone SDS access is not training.

A composite, built from the kind of call that comes in more than you'd think — no names, no company.

A plant has an emergency eyewash in the acid transfer bay. It gets tested weekly, logged, initialed, fine. During a shutdown, a contractor crew re-runs conduit through that bay and needs the wall. They move the eyewash eighteen inches left to clear a tray, re-plumb it correctly, pressure-test it, and leave. Good work, properly done.

Eighteen inches left puts it behind the open swing of the bay door.

Nobody notices for five weeks, because the weekly test is performed by a man who comes in through the other door, props it, and tests the unit from the far side. It works perfectly. It is logged perfectly. Then an operator takes a splash, turns for the eyewash, and finds a door between him and it. He gets there. He's fine. It took eleven seconds longer than it should have, and eleven seconds is the whole ballgame with a corrosive in your eye.

Nobody was careless. The contractor did the plumbing right. The tester did the test right every single week. The test was the problem — it proved the unit functioned and never once asked whether a person in trouble could reach it. Same shape as the hydraulic buffers on a bridge in 1940: a control defeated by competent unrelated work, with nothing in the system positioned to notice.

The fix wasn't discipline. It was two lines: the weekly check now starts from the operator's normal position and walks to the unit, and any contractor work order that touches a wall in that bay routes past the person who owns the emergency equipment.

Got a fail or a near-miss? Hit reply, or share it anonymously at scantoconfess.com — no names, no company, no blame. Just the lesson, so the next crew doesn't learn it the hard way.

Pick your Galloping Gertie. Walk out and find the one thing on your site that everybody knows about, that has a nickname or a knowing laugh attached, and that has been that way long enough to feel normal. The valve you have to hit with a wrench. The interlock that's been bypassed since before you started. The alarm everyone silences without looking.

Then ask the only two questions that matter about it. One: is there a finished recommendation sitting somewhere about this — an inspection finding, a vendor report, an engineering study — that nobody executed? Two: if someone did unrelated work near it tomorrow and quietly defeated it, is there anything in our system that would tell us?

If the answers are "yes, there's a report" and "no, nothing would tell us," you are standing in front of the same two failures that put a bridge deck in Puget Sound. The engineering got solved in 1941. Those two didn't.

———————————————————————————

🎬 Every Safety Intel story as a short video → youtube.com/@GetSafetyIntel · 📸 Share a near-miss, 100% anonymous → scantoconfess.com · 📤 Forward this to someone who walks past your Gertie every day.

Please stay Safe & Hydrated!!!