Why the Tacoma Narrows
Bridge Collapsed
On the morning of November 7, 1940, a bridge began to dance. The Tacoma Narrows Bridge — opened just four months earlier, the third-longest suspension bridge in the world — started twisting in a 42-mph wind. Not violently at first. Just a gentle torsional oscillation, the deck rolling from side to side in slow, nauseating waves. A reporter drove his car onto the bridge, felt it move, and turned around. A local professor walked out to the center to take measurements. The waves grew. At 11:02 AM, a center cable band slipped. At 11:08 AM, a 600-foot section of the main span broke free and fell 190 feet into the Puget Sound below.
The only casualty was a dog named Tubby, trapped in an abandoned car on the bridge. But the engineering consequences rippled through the entire field. Tacoma Narrows didn't just fail — it failed in a way that no one had predicted, that exposed a fundamental gap in how structures were being designed, and that permanently changed what it means to design a bridge.
"The most dramatic failure in the history of bridge engineering. Not because the bridge fell — but because of what engineers didn't know when they built it."
The bridge that was supposed to be safe
The Tacoma Narrows Bridge was not an amateur production. It was designed by Leon Moisseiff, one of the most respected bridge engineers of his era — the man who had contributed to the design of the Manhattan Bridge, the Golden Gate Bridge, and the Delaware River Bridge. His reputation was impeccable. His mathematics were meticulous. His design had been reviewed and approved. And it met every structural specification required of it.
The design philosophy was rooted in what Moisseiff called the "deflection theory" — a more elegant, more refined approach to bridge analysis than the conservative methods of earlier generations. Where older engineers used heavy, deep trusses that acted as stiff beams between the towers, deflection theory allowed for more slender, flexible designs. The Tacoma Narrows deck was only 8 feet deep for a 2,800-foot main span — a ratio far more slender than any previous major suspension bridge. This made it graceful, economical, and modern. It also made it thin as a ribbon in the wind.
The bridge had actually shown a tendency to oscillate vertically — to heave up and down in the wind — almost from the day it opened. Workers called it "Galloping Gertie." People drove across it for the sensation. Local authorities were concerned enough to contact Moisseiff and other experts. Hydraulic buffers were installed. They didn't help. But no one was particularly alarmed, because the oscillations were within the design parameters for vertical movement, and bridges had been known to move without failing. The movement was novel; it wasn't understood as dangerous.
Tacoma Narrows is the most extensively documented structural failure in history. Barney Elliott, owner of a local camera shop, filmed the bridge's death from beginning to collapse. The footage — shaky 16mm film of a massive structure twisting itself apart — was used in physics classrooms for decades. The university professor who walked to the bridge's center to observe the motion, Frederick Burt Farquharson, also filmed portions of the event and subsequently spent years researching aeroelastic effects in bridge design. The film didn't just document a disaster. It created a visual record that forced engineers to confront what their equations had missed.
Static loads vs. dynamic behavior
The core of the Tacoma Narrows failure is the gap between static analysis and dynamic behavior. Static analysis — which is what Moisseiff's calculations were — asks: given a wind blowing at X miles per hour, what horizontal force does it exert on the structure? You calculate the wind pressure, you multiply by the exposed area, you check whether the structure can resist that force. If the answer is yes, you're done. This is the same analysis used for virtually every bridge built before 1940.
What static analysis completely ignores is the question: what happens when the structure starts moving? A rigid structure sitting still in wind is a very different engineering problem from a flexible structure that has begun to oscillate. Once the bridge started twisting, it was no longer just resisting a steady wind force. It was interacting with the wind in a dynamic feedback loop — and that feedback loop had the potential to become self-reinforcing.
The specific mechanism was aeroelastic flutter. As the deck twisted in the wind, its angle of attack to the wind changed. The changed angle altered the aerodynamic forces on the deck. Those altered forces caused more twist. More twist changed the angle further. The forces grew. The oscillation grew. Under the right conditions — a wind speed near the critical flutter speed, a structure with the right ratio of torsional stiffness to aerodynamic properties — this feedback loop doesn't converge to equilibrium. It grows without bound until the structure fails.
Run a wet finger around the rim of a wine glass and it sings — a clear, sustained tone. You're adding energy to the glass at exactly the frequency at which it naturally vibrates. The glass oscillates with growing amplitude until it reaches a limit set by the damping in the system. Tacoma Narrows was the same mechanism at a different scale. The wind was the "wet finger" — continuously adding energy at (or near) the bridge's natural torsional frequency. The difference from a wine glass: the bridge had very little damping, and the forces available were orders of magnitude larger than the restoring forces in the structure. The amplitude grew until something broke.
🤔 Why didn't the bridge's weight damp the oscillations — a suspension bridge is enormously heavy?
▼Weight provides resistance to vertical deflection but doesn't necessarily damp oscillation. Damping requires energy dissipation — the conversion of kinetic energy to heat through internal friction, material hysteresis, or aerodynamic drag on the moving structure. The Tacoma Narrows deck was a solid plate girder (not an open truss), which actually reduced aerodynamic damping of the vertical motion. And crucially, torsional oscillation of a suspension bridge deck depends primarily on the torsional stiffness of the deck relative to its torsional inertia — not its weight directly. The bridge was light for its span precisely because of the slender plate girder design, which reduced both the mass resisting torsion and the torsional stiffness simultaneously. The combination was near-optimal for producing a dangerous flutter condition.
What the engineers didn't know
It would be easy to blame Moisseiff for negligence or ignorance. That would be unfair and inaccurate. Aeroelastic flutter in bridges was not a well-understood phenomenon in 1940. The theoretical framework for analyzing it didn't fully exist. The wind tunnel testing methods needed to predict it hadn't been developed for bridge applications. The connection between aircraft flutter (which aeronautical engineers were actively studying) and bridge oscillations hadn't been made. Engineers were working at the edge of their collective knowledge — and the edge gave way.
What's more revealing is that Moisseiff's deflection theory, which enabled the slender Tacoma Narrows design, was genuinely better in many respects than the conservative truss approaches it replaced. It produced designs that were lighter, more economical, and structurally efficient in every way that existing analysis could measure. The problem was that existing analysis couldn't measure everything that mattered. The theory was correct as far as it went. It just didn't go far enough.
This is a recurring pattern in engineering failures of the highest consequence: the engineers weren't stupid, weren't negligent, weren't cutting corners. They were applying the best available knowledge correctly. The knowledge was incomplete. The gap between what they knew and what they needed to know was invisible to them — because the tools to see it didn't exist yet.
Donald Rumsfeld famously distinguished known unknowns (things you know you don't know) from unknown unknowns (things you don't know you don't know). Tacoma Narrows was an unknown unknown at the time of design: aeroelastic flutter in bridges was not in the design engineer's vocabulary. This is the most dangerous category of risk in engineering — not the risk you've identified and chosen to accept, but the risk you haven't imagined. It's why engineering codes evolve after disasters rather than before them, and why the history of catastrophic structural failures is largely a history of engineering learning what questions to ask.
What changed after Tacoma Narrows
The aftermath was swift and comprehensive. Every major suspension bridge design in progress was immediately reviewed for aeroelastic stability. Wind tunnel testing of scaled bridge models — evaluating their behavior under dynamic wind conditions, not just static wind pressure — became standard practice. The Bronx-Whitestone Bridge, which had a similarly slender plate girder deck, was retrofitted with open trusses within months. New design standards required open deck structures (lattice trusses rather than solid plate girders) to allow wind to pass through rather than act on the full deck area.
The deeper change was methodological. Bridge design, which had been largely a static structural problem, became also a dynamic aerodynamics problem. The Strouhal number, the flutter speed ratio, the mass ratio, torsional frequency ratios — concepts borrowed from aeronautical engineering — entered the bridge engineer's vocabulary. Wind tunnel testing became as standard as structural analysis. The Severn Bridge (1966), the Humber Bridge (1981), and every major long-span bridge since has been designed with explicit aeroelastic analysis and wind tunnel validation.
The Tacoma Narrows Bridge was replaced in 1950 by a wider, stiffer structure with an open truss deck. It stood until 2007, when a parallel bridge was built to handle increased traffic. The original deck still lies on the bottom of Puget Sound, now an artificial reef. And the footage of it tearing itself apart still plays in structural engineering classrooms everywhere — a reminder that the most dangerous failures aren't the ones you can calculate. They're the ones you haven't thought to calculate yet.
🤔 Are modern long-span bridges immune to flutter now that we understand it?
▼Not immune — but far better protected. Modern bridges are designed with explicit flutter speed requirements: the critical flutter speed must exceed the design wind speed (typically a 10,000-year return period wind) by a comfortable margin, usually at least 40%. Wind tunnel tests at 1:100 or smaller scale subject bridge deck models to turbulent wind at all angles of attack and all wind speeds up to well beyond design conditions. Computational fluid dynamics supplements physical testing. Aeroelastic tailoring — shaping the deck cross-section to alter the aerodynamic derivatives — allows designers to increase flutter speed without adding structural weight. The Akashi Kaikyo Bridge (1991m main span) was specifically designed with slotted center boxes and a truss stiffening girder to maximize flutter speed. Multiple checking disciplines and multiple testing methods provide defense in depth that 1940s practice lacked entirely.
🤔 Why did the deck twist rather than just heave up and down — what determines whether a bridge flutters torsionally?
▼Flutter mode depends on the relationship between the vertical bending frequency and the torsional (twisting) frequency of the deck. When these two frequencies are similar in magnitude, their coupling through aerodynamic forces creates a particularly unstable flutter condition called "coupled flutter." If the torsional frequency is much higher than the bending frequency — as in modern truss-stiffened decks — the two modes are well separated and don't couple easily. Tacoma Narrows had a torsional-to-bending frequency ratio of about 1.6 — close enough for coupling. Modern bridge design typically targets a ratio above 2.5–3.0. The plate girder deck also had poor torsional stiffness relative to its aerodynamic moment arm, making it particularly susceptible to the torsional mode. Open trusses, box girders, and aerodynamic deck fairings all increase effective torsional stiffness relative to aerodynamic excitation — specifically to avoid the condition that killed Tacoma Narrows.
Order the Events
Drag to arrange these events in chronological order — from design to collapse.
- Hydraulic buffers installed — oscillations continue
- Leon Moisseiff applies deflection theory to the design
- Main span collapses into Puget Sound
- Galloping Gertie opens in July 1940
- Aeroelastic flutter begins on November 7