Skip to content
Science, Maths & Technology

Tay Bridge Theories: Bill Dow

Updated Wednesday, 9th May 2007

Bill Dow gives his expert view on how The Tay Bridge came to collapse

This page was published over five years ago. Please be aware that due to the passage of time, the information provided on this page may be out of date or otherwise inaccurate, and any views or opinions expressed may no longer be relevant. Some technical elements such as audio-visual and interactive media may no longer work. For more detail, see our Archive and Deletion Policy

Bill Dow gives his opinion on why the Tay Bridge collapsed.

How did you come to investigate the Tay Bridge disaster?

"I was brought up on the story of the Tay Bridge, with a particular family connection. My great grandfather had his own foundry and he, together with the owners of another big foundry in Dundee, seriously thought about tendering for this bridge.

My mother always declared that he had said that the wrong bit of the bridge fell down. Now, I don’t know what he meant, but I have a suspicion that what he probably meant was that the most difficult part of the bridge to construct was the bit between pier 15 and pier 22 which was founded on piles driven into mud flats.

Of course, it was also within the living memory of my grandparents, who used to tell me all sorts of lurid stories about what happened on that night - half of which weren’t true!

When I was a student at St Andrews University I used to travel on the present Tay Bridge at least once a week. When you look down from the present bridge you can see the remains of the old bridge, and this really prompted me to take an active interest. There were always umpteen theories as to why the bridge fell down and I thought to myself, well why not investigate this. Since then it’s really been a life long interest!

Back to top

Theory Summary

There was always a question, was the train involved in the disaster? Gradually, as I teased out a lot of the material which had been put together by the people who salvaged the remnants of the bridge, it became pretty obvious to me that at least one vehicle had been off the rails for quite some distance, at least the length of one girder.

Although there were probably many things involved, I believe that this derailment played a key part in the fall of the bridge. In short, I think that the rear carriages of the train derailed and ran into one of the coverplates. The force of this impact would shatter the cast iron lugs leaving the bridge in a high wind without its proper structural support.

I also think that the girder closest to Wormit, in the high girder section of the bridge, had bent during the lifetime of the bridge. This girder had been dropped during construction and was bent. It was straightened out and reused, but I believe over time it started to return to the bent shape. As a result the rail tracks which ran over it developed a kink. I think that this kink, combined with the high wind, may have been responsible for the derailment. The presence of the 'kink' was well documented in accounts of the permanent way, but they never came to light during the inquiry.

Back to top

What do you think happened on the night of the disaster?

I went through the papers in the old North British Railway Company's archive and I came across two absolutely crucial letters that showed that two girders had actually fallen into the river in February of 1877. And this fall is quite well documented, it concerned the two girders spanning piers 28, 29 and 30.

The inquiry was told that two girders had actually fallen into the river in February of 1877, during the bridge’s construction. It was fully published in the press, it wasn’t hushed up at all. But what isn’t so well known is that the construction company realised that if they were going to get the bridge finished by September of 1877 they just didn’t have enough time to make two extra girders over and above what they already had to make anyway.

Each girder took about four weeks and required a spring tide for raising. So they left one at the bottom and made a completely new one to replace it, but they repaired the other girder that had fallen in. It was literally, according to the manager, fished out, straightened and put back up. And, in my view, that was the girder, between piers 28 and 29, which had a major part to play in the disaster.

The girder needed to be straightened out because when it fell into the river bed it bent due to the fact that the river bed was not flat, but curved as a result of scour. Now when you straighten anything out it’s never quite as strong as it ever was before. And when that girder was put back up they’d obviously managed to straighten it sufficiently well that during all Major Hutchinson’s tests early in 1878 everything appeared to be OK.

Now that girder was riveted to the next 4 girders, so the end of the girder must have been held straight by means of the other girders behind it. But because of the hammering from trains going over it that girder would have reverted to the shape, or as close as it could, that it had on the bottom of the river.

As the girder bent, it would have forced the rails to bend with it, and there were certainly reports that as the trains came along the low girders from Wormit and entered the high girders the engines nodded into the girder. In other words it went slightly downwards and then went slightly eastwards. After that the rails gradually crept back into the normal line of the rest of the girders.

Back to top

On the night of the disaster the wind was coming from the west (so it’s described as a westerly wind), which was virtually at right angles to the bridge and it was definitely about force ten or force eleven. The training ship Mars was in a good position to give this information. Its true speed in miles per hour as far as I know was never measured, nor was the speed of the gusts.

As the train entered the high girder section it would have gone slightly, but sharply, towards the right, or the east. This would have put an impact on the girder which would have tended to drive the girder initially to the west and then to the east as it travelled among the bent girder. That is exactly what happened; the engineer declared that this girder, had moved 20 inches to the east before it was pulled off its rollers. Apparently a tracing of that exist to this day, but I’ve never seen it. But it certainly was referred to in the inquiry.

Whether anything came off the line there or not I just don’t know. But the bridge was such that if a carriage, but not an engine or a tender, came off the rails it could be towed for quite some distance. It might have torn up some of the floorboards, but it wouldn’t have done any major structural damage to the bridge.

In fact, the wooden floorboards came loose as the girders landed in the river, most of which were simply swept out into the North Sea. However, some of the floorboards were washed up by the outgoing tide at, what was then the village of, Broughty Ferry. Some of the floorboards which were recovered had marks from the treads of the carriage ‘tyres’.

As the train was somewhere about the third or the fourth high girder out of Wormit there was an incredibly hard gust of wind. As a result, either the second class carriage was blown off then, or it was already off the line. Now, the mathematics which was done for the inquiry about the force required to upset the second class carriage was completely wrong. It doesn’t allow for the Bernoulli effect, where you get a lifting force when there is a high speed wind across the top of the carriage and a low speed wind below it. They actually reckoned that it would take a force of about 33 pounds per square foot to displace that carriage. British Rail during my time would not trust their mark two steel carriages on the present bridge if the wind could exert a force of more than 30 pounds per square foot. There’s obviously something wrong here if the inquiry gave greater stability to an old lightweight wooden carriage than to a modern steel carriage.

That severe gust pushed it sufficiently far east that it actually hit one of what we call the cover plates at the end of the fourth girder. At that point the 135 tonnes of train is going at about 25 miles per hour, which is quite enough to demolish essential parts of the bridge - particularly the brittle cast iron lugs. The train was effectively a 135 tonne battering ram travelling at 25 miles per hour..

Back to top

When the second class carriage hit this plate the big, heavy guard’s van ran into it from behind and totally demolished the second class carriage, and in the process demolished itself as well. What was found were the undercarriages of the second class carriage and the guard’s van telescoped so hard together that in fact they initially thought they were recovering one vehicle when in fact there were two vehicles there. And it was quite difficult to disentangle the two.

It required very serious forces to drive two vehicles together. That just does not happen by chance. That does not happen by the two vehicles landing in the water. This happens when in fact the vehicles are travelling at speed and one is being stopped and the other one is running into it. So that was where I began to reckon that the train played more than just a cursory part in it.

So, effectively, you’ve got a hammer blow from the train’s impact with the bridge which would break all the cast iron lugs. And incidentally Bouch did not design the lugs, that was one of the things that Bouch left to Gilkes (the contractor who built the bridge). He said to Gilkes, look you can put the ties on either by wrought iron collars or wrought iron bands, or you cast them on. And Gilkes realised it was a lot easier to cast them than to use wrought iron bands, so he cast them. Bouch shouldn’t have passed that responsibility on to the contractor.

As a result of the impact the bridge had lost all the cross bracing of the wind ties which were there to resist wind pressure - they’d all broken off at the lugs. With a wind blowing there was not enough stability in the vertical columns by themselves, so the bridge just went down to the east.

Now the other question you need to answer is why did all the high girders fall and not any of the low girders. All the girders were riveted in sets, the high girders were riveted into three sets. At the Wormit end there were five 245 foot girders riveted together. The middle four girders were riveted together as another set and at the Dundee end there were another set of four girders riveted together.

It is essential to realise that the rails were laid from one end of the bridge to the other in 24 feet lengths with half inch expansion joints - there were no special expansion joints at the point where one girder met the next. Now, before any bit of the high girders started to fall you not only had to pull the girders off their rollers (which sat between the piers and the girders) and off various joints, but you also had to sever four steel running rails. These were of 75 pounds per yard and made of rolled steel, so they were tough stuff.

Back to top

So the rails had to be broken somehow; from what we can see in the photographs the rails broke at the fishplates (the joints between two lengths of rail), which would be the easiest place to break. But the photographs of the surviving piers seem to show one of the rails actually severed, and that would really take some serious force.

So what I think happened is: the set of high girders nearest Wormit went first due to the impact of the carriages with the coverplate. In falling they pulled the next set of girders over because although the sets of girders were separated by expansion joints, the rail tracks did not have an expansion joint at the same place - so there were some very strong ties between the sets of girders. As the second set of girders fell it would have pulled the third set of girders over, in the same way.

But this didn’t bring down the low girders at either the Wormit end or the Dundee end, although the photographs show that the surviving piers had all their wind ties broken and the rails are bent towards the east showing that as the high girders went over they had a jolly good pull on these surviving piers towards the east as well as downwards.

Not only did those piers survive the pull, they also survived the break and they survived the rest of the gale. And because these two surviving piers were constructed in the same way as the piers that fell, this is proof to me that something more than just the wind was needed to bring the bridge down. Because despite having lost all the wind ties they still survived the rest of the gale. The sole reason that they did was that the girders each of them was supporting were riveted into sets of four.

So, you’ve got four low girders, riveted together and supported by five piers. But at one end the pier is effectively out of action because it's lost all its wind ties, and at the other end the pier is effectively only half a pier as it is supporting the end of another set of girders. But the three in the middle are still in business and what it means is that four girders are managing to survive in a wind when there’s only three, or three and a half, piers able to withstand that wind. Now that shows a pretty high factor of safety.

Back to top

How does this theory match the actual evidence on the ground?

The high girders were riveted together in three sets; one set of four closest to Dundee, one set of five nearest Wormit, and one set of four in the middle. As the set nearest to Dundee and the middle set fell into the river on the east side of the bridge, each set of girders remained riveted together. The reason we know this is that they were found in the river still riveted together, and they were in lovely little arcs.

But the five girders that made up the set closest to Wormit were found with only four of them still riveted together, the girder nearest Wormit had broken away and was found at quite a different angle from the rest. It was obviously suffering from forces which none of the other girders were suffering from, in order to sever the rivets which were holding it to the other girders and land in such a jaunty angle.

That’s why I think that this kink in the rails had a lot to do with the actual disaster. You’ve got to explain why this one particular girder broke away and why it lies at a jaunty angle. I think that over time the girder started to return to the shape that it had bent into when it had fallen into the river during the bridge’s construction.

You can also see in the photographs that all of the cast iron lugs have been broken, but there’s no failure of wrought iron. There have been those who said that the wrought iron wind ties should have been much thicker - but they’d burst their lugs. It was the cast iron lugs that failed in all cases, which ties in with my theory of the train hitting a cover plate.

Back to top

Does the speed of the wind play a factor in your theory?

You can’t get away with the train and the train alone bringing the bridge down, that would be pretty speculative. You had to get a series of things; not only did you have to get the vehicle derailed, you had to get it pushed far enough towards the east that it would have hit the cover plate - which would have meant a deviation of at least two feet to the east.

But I also think bits of the bridge were damaged and the ‘repairs’ carried out by Noble were really putting the bridge out of alignment. He was stuffing wedges down in-between the sling plates and this would have destroyed the line of some of the wind ties and they really distorted the positions of the lugs as well.

The wind speed, whatever it was, was critical to derailing the train, and probably combined with some existing damage to the bridge, it helped to bring down the high girders. But it didn’t bring down the rest and that is the big mystery, especially thinking about the piers at either end of where the high girders had been. The gale went on for at least an hour after the high girders fell and these piers, despite the quite serious damage they endured as the high girders parted company from them, still survived the rest of the gale.

So, despite the damage there was enough factor of safety left in them to survive the winds. And this is why I’m quite convinced that the wind, and the wind alone, just would not do it. This is where the critical speed comes in. The wind has got to be strong enough to bring down the high girders, but not strong enough to bring down the low girders. So I think if you’ve got a wind and a wind alone theory you’ve got to explain why anything survived at all really.

Back to top

Do you have an ethos that guided your investigation?

During my time with the Air Force I was trained that if ever a disaster occurred, ‘For God’s sake get the information.’ What was rammed into us was, ‘Get where everything was. Find out the position of everything , it’s very vital to know where things are relative to the others.’ Now that kind of information was just never taken for the Tay Bridge disaster.

This training has definitely shaped how I approached this investigation. I look for the damage. What was actually done? That was the Air Force rule, ‘We don’t want just what you can explain, we want to know all the damage whether you can explain it or not. Don’t start making up theories before you know all the evidence, particularly without knowing where everything has landed.’

So, essentially, my ethos is to make sure that I shape the theory to answer all the evidence; where was the train found, where were the bodies found, where were the carriage roof lights found, where were the wheels?

Back to top

How do you rate the quality of the BOT inquiry, both against today’s standards and also those of the 1880s?

I think it was a top class inquiry of its day, but it had its faults.

After the disaster, the Board of Trade appointed three assessors. It was a wreck so they got the Wreck Commissioner, Rothery. It was a railway wreck so they got the Chief Inspector of Railways, Yolland. It was a civil engineering disaster so they got the President of the Institution of Civil Engineers. Who better than these three? It shows that the Board of Trade got, in their opinion, the top three men.

But they acted like church mice while the person who conducted the inquiry was the Sheriff of Forfarshire, a man called Trayner, who later became one of the top judges in Scotland. Now what happened there was that Trayner conducted the inquiry but the assessors wrote the report.

Recently, when Lord Cullen conducted the "Piper Alpha" inquiry, he asked the questions. He had assessors sitting alongside him, giving advice on what to ask and giving an opinion as to the quality of the answers. In other words, the assessors could "steer" the questions towards areas which they thought important and suggest further questions depending on the answers. But in the final analysis, it was Lord Cullen who came to his own conclusions and wrote the report.

In the case of the Tay Bridge Inquiry, the two engineers could not agree with Rothery, the inquiry chariman, so two separate reports were written, one by the engineers and one by Rothery. The engineers did not apportion blame but Rothery blamed Bouch and Bouch alone. There were other essential differences between the two reports. The Board of Trade should have insisted on the three assesors presenting an agreed report, but it never did."

Back to top

Anatomy of a disaster





Related content (tags)

Copyright information

For further information, take a look at our frequently asked questions which may give you the support you need.

Have a question?