Within days of the flood, two government inspectors, Robert Rawlinson, and Nathaniel Beardmore, travelled to Sheffield to investigate the cause of the dam's failure; and in their final report, which was 'presented to parliament by Her Majesty's command', gave their opinion that the structure's break-up was a result of various aspects of bad design and workmanship. Rawlinson particularly criticised the mode of laying the outlet pipes - set in trenches beneath the embankment. He made much play on the possibility of a fractured pipe - leading to leakage and consequential erosion of the embankment - being the reason for the disaster, and put forward a convincing case; but, at best this was only professional guesswork. The pipes remained buried under tons of embankment: they could have been damaged in the collapse, in which event it would be difficult to prove the point one way or the other.
The chief engineer of the Sheffield Waterworks Company, John Gunson; and the designer and consulting engineer, John Towlerton Leather, strongly refuted the allegations made by the Government engineers: after much deliberation, they concluded that the catastrophe was most probably caused by a landslip; basing their theory on the fact that the downstream toe of the dam was built on an old landslide area, and the walls of some nearby cottages had cracked on the night of the disaster, suggesting that the old landslip had moved again. However, it seemed natural that the Waterworks engineers would look for some explanation to absolve themselves of any responsibility; and given the authoritative evidence of the two government engineers, which was presented at the inquest, the jury had little option but to find that 'there has not been that engineering skill and that attention to the construction of the works, which their magnitude and importance demanded . . .'. They did decide, however, under the influence of Rawlinson, and much to the relief of Gunson and Leather, that no one should be prosecuted. Rawlinson had expressed the opinion that nothing would be gained by attempting to incriminate the engineers or the Water Company and that the aim of the inquest should be 'to obtain facts so as to arrive at reliable conclusions for future use'.27
In the months that followed, the Sheffield Corporation, who had been wanting for some time to take over the Sheffield Waterworks Company, employed no fewer that nine leading civil engineers to try to discover, with more certainty, the cause of the dam's failure. The Waterworks Company themselves employed five of the country's civil engineering giants for the same purpose. These included Thomas Hawksley and John Frederick Bateman, who were probably the two most successful dam-builders of the Victorian era, and Sheffield's own John Fowler - later famed for his part in the design and construction of the Forth Bridge. All made their various inspections of the dam, and of all the available evidence. The engineers employed by the Sheffield Corporation agreed, with the exception of a few minor details, with the findings of the two government engineers; however, the five employed by the Waterworks Company strongly disagreed, stating that in their opinion both design and workmanship were perfectly adequate for a work of this magnitude. These engineers had also excavated the outlet pipes and in their report stated that the pipes had been 'carefully examined. They have been tested under a hydraulic pressure far exceeding that which they would have had to sustain in use. They have been inspected internally by a man passing through them and their lines and levels have been observed from without by means of candles of equal lengths placed centrally within them. In these several manners, the pipes have been proved to be accurate in their position, not having even bent under the pressure of the embankment, and perfectly sound and watertight. This is a state of perfection we never anticipated and speaks well not only for the mode of construction adopted by the engineers, but also for the excellent character of the workmanship.' 28 They supported Gunson and Leather's theory that the collapse had been caused by movement of the old landslip area on which the toe of the dam was built; again citing the nearby cracked cottage walls as their evidence; and that the tragedy had been 'an unavoidable accident'. 29
It seems no one was really certain why the dam had so tragically failed: and in all likelihood, Gunson and Leather went to their graves, in 1878 and 1893 respectively, without really knowing what had caused the dam to collapse. The 'did it fall, or was it pushed?' debate went on, and continued well into the twentieth century, with, over the years, numerous engineers and 'experts' giving a multitude of explanations as to what might have caused the embankment to break up.
In 1978, G. M. Binnie, who was at that time Vice President of the Institute of Civil Engineers ('He is also a senior partner of the well known firm of consulting engineers founded by his grandfather which is responsible for many waterworks at home and abroad.'), and who, for many years, has nurtured a keen, personal, interest in what causes dams to fail, did an extensive study of all the recorded evidence concerning the design, construction and destruction of the Dale Dyke Dam - including an intense study of the area surrounding the site of the old embankment. In consequence we now have, what seems almost certain to be the true answer to what happened to the Bradfield dam on that fateful night in March 1864. His 'report' came in two stages: on 10th. January 1978, he presented a written report at a meeting of the Engineering Group of the Geological Society, in which he concluded that the dam's puddle clay corewall (watertight membrane in the centre of the dam embankment - explained later) had ruptured resulting in leakage and consequential erosion of the inner central part of the embankment: however, there was, at that time, still an element of uncertainty as to what had caused the puddle wall to crack in the first place. The following year, however, he discovered a longitudinal diagram of the original Dale Dyke Dam which had belonged to Watson-Hawksley.30 A certain aspect of this diagram led him to declare that there is, now, 'no need to look any further for the basic cause of the accident.' 31 A report on the cause of the rupturing of the puddle wall was then presented in his book of that year - 'Early Victorian Water Engineers'.
The following section is my explanation of the dam's gradual break up and eventual collapse, based on the combined two Binnie reports.
Firstly, it is necessary to understand a few basic principles of earthen dam design. Diagram 1, opposite, shows a cross section of the original Dale Dyke Dam as designed by John Towlerton Leather. It consists of two sloping embankments (in this instance the slopes are designated to be of a 2½:1 gradient), supporting a central 'puddled clay' corewall. The purpose of this 'puddle wall' is to provide a watertight membrane - essential to make the whole structure leak-free, and thus avoid any gradual erosion of the embankment. To ensure water-tightness, the foundations of this 'puddle wall' need to be built on bed-rock ('watertight strata'); and in the case of the Dale Dyke Dam the engineers had to dig down to the unusually great depth of 60 ft. to reach solid rock (as indicated in the diagram). The puddle clay corewall of the Dale Dyke Dam stood 95ft above ground level, was 16ft wide at 'ground', tapered to 4ft. wide at the top, and, as already mentioned, was sunk to a maximum depth of 60 ft. below ground. The embankment stood 95 ft. above ground level - the same height as the puddle wall - the crest being 12ft. wide and 1254ft long.
The dam had two 18" diameter outlet pipes, both sunk in trenches beneath the embankment, and terminating at a 'Valve House' at the bottom of the downstream slope of the embankment. This aspect of dam design (conveying water from the inside to the outside of the dam) becomes important to the understanding of the break up of the Bradfield embankment: in later years, and for a time, it became the practice to run the 'outlet' pipes through solid concrete tunnels which passed directly through the embankment itself, and this feature soon highlighted a problem with puddled clay corewalls. Puddle clay, due to its high water content, is a highly compressible substance, and over a period of time following construction, any wall built of this material would sink a by certain amount, as the upper layers compressed those beneath them.
2 shows the longitudinal view of a typical puddle wall, with the
discharge pipe passing through such a concrete 'tunnel'. This particular
diagram does not relate directly to the Dale Dyke Dam, but was used at
a much later date to demonstrate the effect of 'differential settlement'
of the puddle wall. What happens is, due to the far greater depth of the
puddle wall at 'A' compared to that at 'B' (see diagram), the upper area
of 'A' sinks to a much greater degree than that at 'B', resulting in a
vertical crack forming at the point between these two zones - shown at
'C' in the diagram. In 1887, A.R. Binnie wrote: '
. . . clay puddle is a very plastic and compressible substance, and apt
to crack under unequal strain, consequently it has been found that, owing
to the unequal depth of puddle which occurs at the steps, the superincumbent
weight has caused that on the deeper side to settle more than on the higher,
and so produce a vertical crack or fault in the puddle which has led to
This problem of 'differential settlement' would not have been understood at the time that the Dale Dyke Dam was constructed: dams of this size and magnitude were in their early days; and in any case, this scenario could never have been suspected as having occurred to the Bradfield embankment, as the outlet pipes were not passed through such a concrete tunnel, but were laid in trenches beneath the embankment (as previously mentioned).
Figure 3. shows the longitudinal diagram of the original Dale Dyke Dam, which was discovered by G. M. Binnie in 1979. It can be clearly seen that, close to the centre of the bottom of the 'puddle trench' (the bed-rock on which the puddle wall was built) is a very unusual near-vertical step in the rock formation of 35 feet depth (marked 'X' in the diagram). It can also be seen that this lies directly beneath the central part of the embankment which was washed away in the flood. It was this that led Binnie to declare: 'we need look no further for the basic cause of the accident': he went on to explain that this abrupt change in the puddle trench 'must have caused a rupture in the puddle clay wall, most probably before the corewall reached its full height.' As shown in the previous description and figure 2, the differential settlement between the greater depth of puddle wall at 'B' compared to that at 'A' would have resulted in the clay wall cracking just above this vertical step (marked 'X' in the diagram). No details of this irregularity in the rock foundation were revealed at the inquest nor in subsequent discussions. Interestingly, in April 1860 John Towlerton Leather reported:
'The works of the Dale Dyke reservoir have been retarded in some degree, in consequence of the difficulties experienced in arriving at a satisfactory foundation in the middle of the valley, for the puddle wall, which difficulties have been enhanced by the severity of the past winter. . . . '
The puddle trench was completed in 1861 and the placing of fill began in the autumn of that year. On 20 April, 1863, Leather wrote:
'I am happy to report that notwithstanding the more than ordinary difficulties which have been encountered in the construction of the Dale Dyke reservoir, the embankment is now so far completed as to be ready for the reception of the Water, and the Waste Weir, which is not quite finished, may be soon so in the course of a few weeks from the present time, and the Works throughout are well and substantially executed.' 33
Sometime after the flood, Joseph Ibbotson - one of the Bradfield villagers who had been employed on the construction of the dam - sent a letter to Samuel Harrison, editor-proprietor of the Sheffield Times, which read: ' . . . Another important fact, well known here, but which I have not so far seen any inquiry into is, that a large spring of water issued from the foot of the embankment where the breach has occurred, and was conveyed away by a drain'.34 It seems that several springs encountered by the engineers during the construction of the dam were believed to be no more than 'natural' ones emerging from the hillside. Binnie remarks 'If, as later became established practice, Gunson had constructed a weir for measuring potential leakage at the toe of the embankment and if, before going home [on the day of the flood], he had inspected it, he would probably have found some leakage, and also that the water was carrying sediment. Erosion had probably been going on a long time without being noticed.' 35
As already mentioned, following the official inquest into the bursting of the Dale Dyke Dam, a number of leading engineers were engaged by both the Sheffield Waterworks Company, and the Sheffield Corporation, to investigate, and report on, the cause of the disaster (more details are given in the section 'Aftermath': from the 'main page', follow link 'Photo Gallery . .'). One of the Corporation's engineers was Mathew B. Jackson. Jackson recorded statements made by several of the workmen who had been employed on the dam's construction. One of them reported that:
'Several weeks, or a month, before the bursting, I observed the pitching inside the bank had settled forming a hollow, as near as I can tell, about the place where the hole was first blown through, just above the surface of the water at that time, I suppose this was bout 10 or 12 feet below the level of the waste weir when I observed this sinking of the pitching. I was standing on the bank about opposite, inside the mason work, inside where the water enters the pipes, when I was watching the water boil through the bank at a distance of about 30 yards from it.' 36
It is not clear whether Gunson was ever aware of this situation: he certainly 'did not report any depression in the pitching nor any whirlpool, but when he was giving his evidence the Coroner, who was very dictatorial, hardly allowed him to do more than answer directly the questions put to him, and no question about either was raised.'
Another Waterworks employee who was interviewed by Mathew B. Jackson, revealed more details that were never presented at the inquest:
[On the night of the flood] 'Mr Gunson and the contractors sent for drills and powder to blow out a stone from the top course of the waste weir. Whilst the men were engaged drilling, a man was trying to cut out a stone with wedges. It was observed that the water was lowering faster than the pipes could draw it off. This was named to the contractors. They thought it was not so; but they noticed themselves afterwards, and they were satisfied that it was lowering. Then they sent a man along the bank with a light to see whether the water was escaping. He came running back to say there was a hole blown through. Mr Gunson and George Swindon and others immediately went over the top of the embankment. I went as far as I thought safe, and saw the water boiling through. I stood about ten minutes and then the top fell in, which appeared to stop the water a minute or two, until the water ran over the top in sheets of foam. An immense gap was speedily opened . . .' 37
Curiously, in all the evidence given by John Gunson, there is no mention of this hole which had 'blown through'; and which, according to the Waterworks' employee, occurred about 'ten minutes' before the final collapse.
The following paragraphs present the 'evidence' that was given at the inquest by John Gunson relating to the hours leading up to the dam's collapse:
After being summoned to attend the embankment following the discovery of the horizontal crack in the outer slope by one of the Waterworks' employees, Gunson eventually arrived at around 10 p.m.
'In his evidence Gunson stated that when he arrived he could just get his fingers into the crack edgeways.' He stated that his 'first impressions on the cause of the crack were as follows:
'I thought it was the action of the wind and the waves that had been beating against it all the afternoon, and that it might have loosened the material that the inner slope of the top of the embankment is composed of, above the water mark: and if that had been the case, I thought it would be like taking the inner slope of the puddle wall from it, and cause a slight crack outside.'
Gunson decided to relieve the water pressure against the crest of the embankment by blowing up the top stones of the weir. He described what happened next:
'After putting the shot in, I said to the workmen, 'We will go carefully back and examine it again', and I thought of measuring from the top to the crack to see whether it was above or below the surface water of the reservoir. We walked over the crack and all seemed perfectly right just as it did when we had passed over before, all the men and the rest of us, perhaps half an hour before. I had a lantern in my hand, and when I got to the end and saw what was there, I said to Swindon [the contractor] 'George, good God, the water is over the embankment'. It came right under my feet and dropped down the crack.' 38
Gunson went down the slope to the valve house to see if he could get some idea of the quantity of water passing over which initially was 'no great current'. He continued his evidence:
'After I had been at the bottom a short time, I turned me round to go out, and cast my eye up the embankment, and could see an opening about 30 yards wide, perhaps, just as though we had blasted a way in the middle of the embankment, and in another moment one tremendous rush came and shook the ground under my feet.'
A huge gap was quickly cut in the embankment down to the level of the top of the bottom outlet chamber on the upstream side and down to river level on the down stream side. The time was approximately 11.30 p.m. . . . ' 39
The dotted 'Parliamentary line' shown in Figure 4, indicates the site originally chosen for the 'centre line' of the dam, however, it was soon discovered that this line lay in the middle of an ancient landslip area. The dam centre line was consequently moved about 150 yards upstream at the north end; which still left the downstream 'toe' of the dam encroaching on the old landslide area, and a few hundred yards to the north of the embankment (on the northern edge of the ancient landslide area) were several cottages (which no longer exist)(see figure 4). 'When the dam collapsed some movement of the ancient landslide, sufficient to cause cracks on the walls of these cottages, occurred. Hawksley and his supporters [the engineers engaged by the Waterworks Company] used this evidence of cracking of the cottage walls to support their theory that the dam had been destroyed by the ancient landslide moving again.'
Mathew B. Jackson later remarked on this, pointing out that the dam failed immediately the reservoir became practically full for the first time, and asked 'Was this the precise moment for a natural and unavoidable landslip to occur?'.40 On the other hand, one might ask: could not the increased stresses applied to the foundations of the downstream toe of the dam, due to the weight of the impounded water, have caused the old landslip to move again?
Binnie writes: 'However, as a consequence of reservoir impounding, the piezometric levels in the river bed foundations beneath the dam must have been raised considerably on the upstream side of the cut-off trench and, when the dam collapsed, the speed with which the load was removed from the foundations in this particular area almost certainly exceeded the speed with which the piezometric levels were able to adjust themselves. As a consequence of this rapid drawdown, some movement of the dam foundations, which also affected the adjacent landslip area, probably occurred. In other words this movement is [now] believed to have been a consequence of, and not the cause of, the collapse.' 41
All the evidence points to leakage having taken place due to the puddle wall being ruptured near its base, and this having occurred long before the dam's completion. Having clearly determined this, the postulated sequence of events that would have subsequently occurred - leading to the eventual collapse of the dam, can now be explained (Refer to Fig. 5):
1: A cavity forms in the upstream side of the embankment due to erosion
caused by leakage through the cracked puddle wall.
Stage 2: The roof of this cavity progressively caves in resulting in the cavity migrating upwards. As a consequence of the reduced support against the clay corewall (possibly combined with the undermining of the 'wall' by the crack itself), the upper part of the crest leans in the upstream direction (towards the water) - thus 'stretching' the outer slope of the embankment, and resulting in a horizontal crack forming a little way down the slope. This was the crack that was discovered earlier on the day that the dam collapsed.42
Stage 3: 'When the migrating
cavity reached the surface a 'swallow hole' was formed very close to the
crest with the consequence that the adjacent material slid into it. It
is conjectured that this is the stage seen by Gunson when the waves suddenly
started coming over the crest and down the crack at his feet.'
Flood at Sheffield - 1864 (main page)