In 1879 the Tay Bridge was the longest bridge in the world, spanning two miles across the Tay estuary in southeastern Scotland. On the evening of December 28, 1879, the central part of the span—the so-called high girders—suddenly collapsed, leaving a gap of well over a half-mile. Most disturbing was that the two-year-old bridge collapsed while an express passenger train from Edinburgh was making its way across. The resulting accident claimed the lives of 75 victims, making it the most catastrophic structural failure in Britain's history.
More than 125 years later the cause of the disaster remains in doubt. In my recent book, “Beautiful Railway Bridge of the Silvery Tay” (Tempus Publishing), I attempt to put an end to the uncertainty. By reexamining the wealth of surviving evidence—in particular the photographic archive and the records from the formal accident investigation of 1880—I reassess the various theories of how and why the bridge came down.
So what happened on that fateful night? A strong gale was blowing, the sky partly cloudy, and a full moon illuminated the landscape. A local train traversed the bridge at a quarter past six and the journey across had been noticeably difficult. Sparks flew from the wheels as the wind tilted the carriages against the guardrail—rails that were designed to prevent toppling in just such circumstances. Passengers later emphasized the violent shaking of the carriages.
At about 7:13 pm, an express train drawn by a much larger and heavier locomotive was seen by witnesses in Dundee passing over the southern part of the bridge, again with some difficulty and with sparks flying from the wheels. An especially severe gust of wind was felt on land just as the train was passing through the high girders at about 7:20 pm, and several observers saw what appeared to be flashes of light coming from the metalwork. The towers in the high girders collapsed progressively and the train plunged into the water below.
Rescuers arrived on the scene by boat at first light the next morning, but found no survivors or bodies. What they did find was remarkable. The high girders were resting on the estuary bed, partly exposed at low tide and remarkably intact. Divers found the train resting between the fourth and fifth piers, also having suffered little damage. In fact, the locomotive would later be restored to a long and working life.
In the aftermath, designer Sir Thomas Bouch alleged that the wind blew the train from the track into the bridge, and that the shock caused the lugs on one of the towers to break, leading to the collapse. However, Bouch's analysis failed to explain why all twelve towers collapsed and not just the one nearest the point on the high girders that the train allegedly hit.
The final report of investigators—delivered by June 30, 1880, a remarkably short turnaround time, especially compared with present practice—disagreed with Bouch's assessment. The Court of Inquiry faulted the structure for its design and material defects, and held Bouch—who died a few months later—personally responsible for its collapse. Chief investigator, Henry Rothery, condemned the construction of the bridge in no uncertain terms, describing it as “badly designed, badly constructed, and badly maintained.” However, the Court did not specify exactly how or why the structure failed.
My reappraisal confirms the general conclusions of the original inquiry, but it also extends them by suggesting that lateral oscillations were induced in the high girders section of the bridge by trains passing over a slight misalignment in the track. The amplitude of these oscillations grew with time, because joints holding the bridge together were defective, and this in turn resulted in fatigue cracks being induced in the cast iron lugs. Although wind loads contributed to the disaster, the bridge was already severely defective owing to failure of its most important stabilizing elements.
Critical evidence of the state of the bridge a few months before the accident came from the crew that painted the structure. They experienced severe vibrations on the piers whenever a train passed over (regardless of whether the wind was blowing). Passengers also reported disconcerting vibrations, especially those who had traveled over the bridge from south to north.
And movement of a different kind had been observed much earlier. After completion of the bridge in the spring of 1878 an inspector was appointed to maintain the structure. While near one of the pier platforms in October 1878 the inspector heard a rattling noise when a train passed overhead, and upon further investigation discovered that some of the joints were defective. Yet, he neglected to report the problem, expecting that he could remedy it himself. He purchased lengths of wrought iron bar and cut them down to make shims—thin pieces of material used to fill gaps. Then he hammered the shims into the loose joints (as many as 150) to stop the vibrating and rattling. But by doing so he jammed the joints into a fixed state bearing little or no strain, effectively destabilizing the towers.
Taken together the evidence of the painters and the inspector point toward serious deterioration of the towers of the high girders section—after the bridge had been tested by the Board of Trade in February of 1878, a year-and-a-half before the accident. The tests involved running six heavy locomotives (total weight of well over four hundred tons) at high speed (forty miles per hour) over the bridge and observing the effect on the pier towers. The Board of Trade inspector measured little effect on the structure. However, by October of that year the joints were coming loose, probably as a result of high frequency vibrations from passing trains. Hammering shims into the gaps may have kept the joints from rattling, but it also meant they were no longer effective. The steady cumulative loosening of the structure on all of the towers allowed the lateral movement felt by the workmen on the bridge in the summer of 1879.
So where does this analysis of the collapse take us? The evidence for steady deterioration of the pier structure is convincing. It was produced by two mechanisms: poorly designed joints in the bracing bars, which allowed play to develop (chattering joints); and large stress concentrations at the bolt holes of the lugs, which allowed fatigue cracks to grow. The history of the bridge from its opening to the final collapse is thus important to an understanding of why it collapsed so dramatically. The bridge had been open to traffic since September 1877, heavily loaded with trains carrying stone and coal, and traffic grew as passenger trains were added to the route. Such conditions led to the loosening of the joints and the swaying of the towers felt in 1879.
On the day of the disaster, extra loads were added to the high girders by the westerly gale, especially during the passage of the six o'clock train. The rear carriages were swaying severely enough to cause sheets of sparks from the wheels as they met the guardrail. But how much of the sway was caused by wind acting against the carriage sides, and how much by the bridge itself swaying on its joints? If joint looseness and fatigue cracking had progressed far, then the sway of the bridge itself must have been considerable. Many more tie bars must have broken and swung free during the passage of the six o'clock train, leaving the bridge in a parlous state for the following express train. None of the damage would have been visible because night had already fallen.
When the express train entered the high girders, the greater weight of the train (well over a hundred tons) would have produced critical movement to aid toppling of the towers over which the train passed. Each tower behaved as though composed of two separate towers linked by struts and tie bars alone. The train nearly reached the fifth tower before collapse overtook it, probably starting at the southern end and working progressively forward until the entire high girders section had been swept away.
Like many other collapsed bridges the Tay was rebuilt—parallel to the line of the original structure, using surviving girders from the low section of the old bridge. The new support piers were much wider, giving a much higher safety factor against toppling, and were placed upstream of the old piers, which now served as breakwaters. In fact, the collapsed piers remain in the water to this day, a haunting reminder of the tragedy.
Notably, the Tay Bridge disaster inquiry pioneered systematic investigation and recording of the evidence visible at an accident site. Other accidents of the railroad age were systematically investigated prior to the late 1880s, but the Tay Bridge was probably the first time a systematic photographic survey was made for an accident investigation. The photos have proved to be an invaluable archive, which has enabled reexamination of the disaster with the benefit of modern knowledge of likely failure modes.
Dr. Peter R. Lewis is senior lecturer in materials engineering at the Open University in Milton Keynes, United Kingdom. In addition to Beatutiful Railway Bridge of the Silvery Tay he has co-authored two books and published numerous review and papers in such journals as RAPRA Review Reports and Engineering Failure Analysis.