United Marine Aggregates Ltd v G.M.Welding & Engineering Ltd & Anor

This is a fire case. The Claimant ("UMA") owns a large aggregate processing plant on the Thames, near Greenwich, at which there was a very serious fire on the morning of Sunday, 3 February 2008. UMA contends that the fire was caused by hot work being carried out by the Defendants setting fire to the rubber lining in a steel hopper or chute.

The Defendants deny that they caused the fire but, if they are liable, they claim an indemnity from their public liability insurers, Novae Syndicates Ltd ("the insurers"). The insurers deny liability on the ground that the Defendants were in breach of a Burning and Welding Warranty in the policy, which is expressed to be a condition precedent to any liability.

The Defendants also assert positively that the fire was caused by hot work being done in an adjacent building at the same time under the supervision of an employee of UMA.

Thus the principal issues in the case are as follows:

The second of these issues also raises the question of whether UMA knew the exact method of working that was to be adopted by the Defendants and approved it, even though it may not have complied with UMA's own written procedures for hot work.

The primary case of the insurers is that the warranty required that all combustible materials in the immediate vicinity of the work that could not be moved had to be covered and protected by some non-combustible material. They contend that the rubber lining of the underpan to the A side primary screen to which hot work was being carried out by the Defendants was not so covered and protected. Alternatively, they contend that the Defendants did not carry out a fire safety check at regular intervals during the work and between 30 and 60 minutes after completion of the work.

UMA’s aggregate processing plant at Murphy's Wharf, Greenwich, is the largest, or one of the largest, such plants in Europe.

Ballast, in the form of material dredged from the river or sea bed, is brought to the site by barge. It is discharged at the site onto a spoil heap from which it is transferred into the sorting and processing system. The material is transferred by conveyor belt to a building known as the scalping (or scalpen) house where the largest stones and lumps are removed. This is done by passing the material over a screen which permits material with a diameter of less than 75 mm to pass through it into a type of hopper or chute, known as an underpan, below.

This underpan has an opening at its base measuring about 2 m by 500 mm, through which the material falls onto a conveyor belt below (CV6). CV6 then transports the material out of the scalping house and up a 35 m inclined section to another building, known as the screening house. Within that building it continues for about another 20 m or so and finally it deposits the material into a container known as the top box.

The top box is a rectangular container, at right angles to the line of CV6, which contains a mechanism for feeding the material into two screens either side of CV6. These are known as the A side and B side primary screens. The screens are large steel vibrating containers which permit material below a particular size to pass through into an underpan below. From the underpan the material is discharged to feed further screens below.

Each of the primary screens is a large steel rectangular hollow box - shaped like a shoe box with no top and bottom. It is braced by a series of hollow horizontal tubes running from side to side that are bolted at the ends to each side. Square steel flange plates are welded to the ends of each tube in which four holes are drilled to take the bolts which fix the flange plates to the side of the screen. There are two layers of tubes, and the tubes on each layer are about 350 mm apart.

To the upper surfaces of the tubes, and parallel to the sides of the screen, are welded horizontal steel bars that support the screening mats. The screening mats are made of polyurethane and resemble the wooden gratings found in 18th century warships, only much finer. To judge by the photographs, each mat is about 1 m long and about 300 mm wide. They are kept in place by steel bars that are bolted to the sides and the central bar of the screen. The square apertures in the mats permit the smaller particles in the material to pass through into an underpan below the screen. In order to achieve this, the screen vibrates. However, this produces substantial stresses in service to the flange plates at the ends of the horizontal tubes. The most common form of failure takes the form of cracks that propagate outwards from the bolt holes in the flanges. From time to time these cracks have to be “vee-ed out” and then repaired by welding. In order to do this the bolts have to be removed and, using a small angle grinder, a groove is cut along the line of the crack which is then filled by welding. Cutting a groove in this way in preparation for welding is known as “vee-ing out”.

Below each primary screen is an underpan, which is a little like an elongated builder’s skip. Its dimensions at its widest part, being the top, are the same as those of the screen above. It is some 7-8 m long and about 2.5 m wide. The bottom of the underpan is roughly horizontal, but the shorter sides are at a fairly shallow angle to the horizontal so that the dimensions of the horizontal base are much smaller than the dimensions of the top. The underpan is divided into two sections by a central partition running along its length, which has the shape of an inverted V. This partition does not extend to the full height of the underpan and so its ridge is significantly below the sides of the underpan. In the middle of the base of each half of the underpan is a chute which permits the material to be discharged from the bottom for further processing.

In the centre of the underpan, the vertical distance from the base to the underside of the lower layer of the tubes of the primary screen is about 3 m, so that a man standing on the base of the underpan could probably not reach the tubes above him and certainly could not carry out any work on them without some form of platform. However, if he was to walk along the length of the underpan and up the sloping ends there would come a point at which he could reach the tubes quite comfortably, but the gap between the tubes and the sloping side continues to decrease so that at each end of the underpan the tubes are only about 300 mm or so above the sloping side (or base).

Between the lower edge of the long sides of the screen and the top of the side of the underpan there is a gap. It has been reduced in width by the addition of a long strip of metal which has been welded to the top of each side of the underpan. The strip has been fixed at a slight angle to the vertical so that it projects beyond the line of the side of the screen. In other words, if one held a small marble against the side of the bottom of the screen and let it go, on falling it would probably be caught by the strip of metal and deflected into the underpan. The reduced gap between the top of the strip of metal and the underside of the screen is sufficient for a man to put his arm through into the underpan.

From time to time it is necessary for someone to enter the underpan for maintenance or inspection purposes. This can be done in one of two ways. Either the person can wriggle through the tubes of the screen above and descend at a point along the sloping end of the underpan where there is suitable clearance, or access can be gained from below by way of the chutes in the base of the underpan.

In order to get into the underpan through the screen it is of course necessary that the mats have been removed at the point of access. If hot work is being carried out on one of the screens then the screening mats above it will usually be removed because they are combustible and thus present a fire risk.

Each screen is supported by means of four large steel brackets, known as rosta brackets, that are bolted onto the side of the underpan and the side of the screen.

The sides and base of the underpan are lined with a red rubber material known as Semperit. The purpose of this is to prevent attrition of the metalwork of the underpan from falling stones. The tubes are also encased in the same or a similar material, presumably for the same reason. From time to time the lining has to be replaced. According to the evidence of Mr Smith, a charge hand who worked for the Defendants, which was not challenged on this point, the current lining material was introduced in 2006 to replace a similar material that had been used before. This former lining material was black.

Mr Smith said also, again unchallenged, that the Defendants were concerned that the new red material might be Linotex, which is also red but is readily combustible. For this reason carrying out hot work in proximity to Linotex was regarded as being out of the question. Accordingly, the Defendants asked UMA about the properties of this new lining material, particularly in terms of its combustibility, and, according to Mr Smith, they were told that it was less combustible than Linotex and, indeed, less combustible than the material that it replaced. However, there was some inconsistency in the evidence as to precisely what the Defendants were told about the combustibility of the new lining material but, whatever may have been actually said, it appears that the change did not lead to any changes in the established hot work procedures from which I consider that the only reasonable inference is that no-one thought that any such changes were necessary. Accordingly, I find that the effect of what the Defendants were told by UMA was that the new Semperit linings need not be treated any differently from the material that it replaced.

However, it was nevertheless accepted by the Defendants that the lining material in the underpan was potentially combustible so that precautions had to be taken to protect it when carrying out hot work, such as oxy-propane cutting, welding or grinding.

I have already mentioned that in the scalping house there is also a screen with an underpan below it. This underpan is fitted with what are known as “dead boxes”. These are strips of metal (I assume in a hollow E cross section, or similar) that are welded horizontally to the sides of the underpan. Their purpose is to absorb the impact of the larger stones that are screened out in the scalping house. The inside of the underpan in the scalping house is also lined with a lining material, which at the relevant time I assume was Semperit. One of the routine tasks often undertaken by the Defendants was to replace the dead boxes from time to time when they became damaged or broken.

On the morning of the fire, Sunday, 3 February 2008, four men were working on the primary and secondary A side screens in the screening house - they were all employed or engaged by the Defendants - and two men were engaged in fitting dead boxes to the underpan in the scalping house. The latter were a Mr Clancy, an employee of UMA, and a Mr Davies, an employee of the Defendants. In the Permit to Work issued to Mr Clancy dated 2 February 2008, he was described as the person in charge of the work. Mr Davies was named against the heading "Others Covered by Permit". This permit was issued by Mr Clancy to himself, a procedure that Mr Farla said was irregular.

In the screening house there were three tasks that required hot work. The first involved welding 1 or 2 lugs (the evidence was not clear as to which number) as lifting brackets for the secondary screen. This work was carried out by a Mr Shaw and a Mr Garnham.

The second involved cutting the nuts off 20 bolts on the sides of the A side primary screen. 16 of these nuts were on bolts which fixed the flanges of the tubes to the side of the screen. 4 went through the rosta bracket, and the evidence suggested that these four bolts may also have secured one of the tube flanges. This cutting was done by Mr Smith using an oxy-propane cutting torch. Mr Smith said that it was not possible to cut off these nuts using an angle grinder without damaging the metal side of the screen, although his evidence on this was challenged. In case anything turns on it, I should say that I accept Mr Smith's evidence on this: as he explained, it is difficult to use an angle grinder so that it is flush with the surface being worked on because of the guard and the presence of the projecting bolt that secures the disc. At any rate, no other witness gave evidence to the contrary.

The third task that required hot work to the A side primary screen was the vee-ing out of the cracks in the tube flanges. As I have explained, this was done using a small angle grinder once the bolts had been removed. According to Mr Smith, whose evidence on this was not disputed, all the cracks in the flanges were on the top - that is to say, in the area of one or both of the upper two bolt holes in the flange. When the cracks had been vee-ed out the flanges were cleaned using the second angle grinder with the wire brush attachment. For the reason given later in this judgment, in my view the use of this wire brush attachment was not “hot work”.

Turning to the work in the scalping house, Mr Clancy and Mr Davies had to make new dead boxes in the workshop and then fit them to the inside of the underpan. Initially, this involved tack welding the new dead boxes in position. Thereafter three or four seam welds would be laid down at points along their length to hold them firmly in place. By the time of the fire on 3 February 2008, Mr Clancy and Mr Davies had tack welded about five dead boxes, but had carried out no seam welding. That work was carried out between about 8 am and 9:30 am that morning.

Later in this judgment I will consider in more detail the work that was being carried out at these two locations. For present purposes it is sufficient to say that between about 9:30 and 9:40 am nearly all the Defendants’ men left site to go and have breakfast nearby. Mr Clancy remained on site because he wished to carry out repairs to the exhaust of his car.

The facts as I have summarised them above are, in my judgment, all established by the evidence of the witnesses or by those parts of the contents of the agreed documents that were not in issue.

I find that at about 9:55 am, or shortly afterwards, on Sunday, 3 February 2008, the Wharf Manager, Mr Stephen Farla, arrived at the site by car. He had come in to catch up on his paperwork. Whilst parking his car he noticed a wisp of black smoke coming from the top of the screening house. Having got out of his car, he noticed a lot more smoke coming from the screening house as he was walking towards his office. He then realised that there was a fire and he ran to his office to telephone the fire brigade. The call was recorded as having been made at 10:01 am.

He then left his office in order to find the men whom he expected to be working on the site. The only person he could find was Mr Garnham, who was in the Defendants’ canteen. He told him about the fire and then went straight back to his office to put on his protective clothing. Having done so he ran towards the plant and, as he approached the switch room he heard an explosion from the direction of the screening house.

Mr Farla said that he could see that the conveyor CV6 was on fire at its top end. He could see no evidence of fire on the lower part of the conveyor or any smoke coming from the scalping house. He then telephoned Mr Paul Burgess, the site Operations Manager, to tell him of the fire. He said that by that time the conveyor was burning fiercely and that it was not long before it "burnt through and dropped down to the scalpen house".

This, at any rate, was the evidence of Mr Farla, although his eyewitness account of his first sighting of the fire and its spread along the conveyor belt was seriously challenged on behalf of the Defendants.

In this section of this judgment I will give a brief description of the evidence given by each of the main witnesses and the impression that I formed of that witness. I will take the witnesses in the order in which it is most convenient to consider the evidence, rather than the order in which they were called at the trial. In this context it is unfortunate that proceedings were not started until three years after the fire, with the result that four years have now elapsed between the events giving rise to the fire and the witnesses giving evidence at the trial.

Mr Steven Farla was UMA’s Wharf Manager from September 2007 to December 2008. He was called on behalf of UMA. He made a fairly short witness statement which dealt only with his observations of the fire. However, in cross examination his evidence ranged much wider. I found him to be a straightforward and truthful witness.

He said that when he first noticed the fire smoke was coming out of the gap between the south wall and the roof of the screening house (as he had marked on the photograph at A/103). He then ran into his office, the back window of which looked on to the scalping house, in order to telephone the Fire Brigade. He said that his office, which was on the first floor, was at much the same level as the point where the conveyor CV6 came out of the scalping house. He said that he did not think it was possible that the conveyor could have been on fire in the area of the scalping house otherwise he would have seen it and told the Fire Brigade. However, he accepted that from his office he would not have been able to see the top of the conveyor where it entered the screening house. But he said that if there was any visible smoke coming from the lower part of the conveyor he would have seen it from the office.

In his witness statement Mr Farla said that after he had telephoned the Fire Brigade he put on his personal protective equipment and ran towards the plant. As he approached the switch room he said that he heard what sounded like an explosion in the screening house. He could then see that the conveyor was on fire between the screening house and the scalping house. In his witness statement he said: "The fire that I saw was at the top end of the conveyor only".

In a hand written witness statement that he made two days after the fire, on 5 February 2008, he said this:

The impression given by this passage is that he did not see the conveyor belt break or fall to the ground. However, in his witness statement, he said that it was not long before the conveyor burnt through and dropped down to the scalping house. He said that he had marked the area on a photograph where he believed the conveyor snapped. In evidence he explained that he was describing a point about three quarters of the way along the open length of the conveyor (from the scalping house end). I will return to this evidence later in this judgment.

Mr Farla said that he had a close working relationship with Mr Marshall of the Defendants and that he had no issues with their safety record. He had a favourable opinion of the Defendants. He said that UMA had Hot Works Procedures already in place when he joined the company and that overall responsibility rested with him, together with his Assistant Manager and Mr Curtis, the site supervisor.

Mr Farla was taken to paragraphs 2.3 and 2.4 of the note of the Hot Work Procedure briefing that was held on 16 January 2006. He agreed that where there were combustible materials that could not be moved when hot work was taking place they should be covered where possible. He accepted that there might be such materials that could not be covered and he said: "I would accept that the underpan is one area that couldn't be covered with blankets or the like”.

He was asked about the risk assessment that was carried out on 28 January 2006, in relation to which he said this:

“The normal procedure was to use running water to protect the underpan, but this was not cut and dried. There were places that could not be covered and so we would spray them with water. I would accept that the A side underpan was one such area.

UMA was aware of the practice of spraying the lining of the underpan with water instead of covering it and approved it.”

When asked about the method statement he accepted there was no reference in it to covering combustible materials with metal sheeting or welding blankets and that it only provided for live water to be at hand at all times during the repairs. He said that overall he would have approved the arrangements as described in the risk assessment and the Permit to Work.

Mr Farla said in evidence that he had not told Miss Russell that hot work had not commenced in the scalping house on the Sunday morning, although he recalled that hot works were due to be carried out in that building over that weekend. He said that the first he heard of the allegation that the fire started in the scalping house was only a few weeks ago.

In relation to the new rubber lining in the underpan, he said that "we were not sure quite how combustible it was until after the fire. It turned out to be very combustible".

Mr Alan Davies was a fitter formerly employed by the Defendants, who called him as a witness. On the weekend of 2/3 February 2008 he was working in the scalping house with Mr Clancy, who was employed by UMA.

Mr Davies remained on site after the fire along with the other employees of UMA and the Defendants. The Defendants’ employees, including Mr Davies, were marshalled in a rest room belonging to a nearby company where they spent the rest of the day. Mr Davies was never asked to make a statement after the fire, although a short statement was taken from Mr Clancy. In his witness statement he said that a few days after the fire he heard from Mr Smith that Mr Clancy had told those investigating the fire that he had not been carrying out any hot work in the scalping house immediately before the fire. This made Mr Davies very angry.

As a result he was asked to make a statement by Mr Marshall, which he did on 28 February 2008 in the presence of Mr Marshall and Mr Smith. This statement read as follows:

“On Sat 2/2/08 and Sun 3/2/08 myself & Terry Clancy carried out repairs to Scalpens underpan. No hot work done on Sat. Started hot work on Sun morning 8 am - 8:30 am. Tack welding dead boxes into underpan. Both Alan and Terry doing this. Finished this approx 9:30 am. Measured up next pieces to be fitted. Left Scalpens House approx between 9;35-9:40 am.

Both in underpan.

No man outside. The burning gear in underpan hatch (did not use).

Cut rubber in underpan to fit dead boxes.

Standing on CV6 belt.

Fire hose in situ.

Did not wet down after completion.

Mats were not removed from Scalpens screen.

Went to breakfast approx 9:40 am.”

In evidence, Mr Davies said that there was one error in this account, in that it was wrong that no hot work had been carried out on the Saturday. He said that they did carry out some hot work on the Saturday, because they had to burn out the old dead boxes in the underpan with an oxy-propane cutting torch before fitting new ones.

Mr Davies left the Defendants about 6-8 weeks after the fire. He said that he was not sacked - “it was after the banks went down”. The next occasion on which he was asked to make a statement was in December 2011, when Mr Marshall spoke to him. He said that the next day "I received this", meaning his witness statement.

In that statement he said that both he and Mr Clancy had been carrying out tack welding on the morning of the fire. In evidence he said that Mr Clancy was in the workshop when he arrived. They had to make the dead boxes before they could install them. As I have already mentioned, the dead boxes consisted of lengths of steel bar which were used to absorb the impact of ballast as it fell into the underpan and there was also a rubber lining. They wore out fairly regularly and had to be replaced.

Mr Davies explained that after they had made the dead boxes they had to fit them in the underpan. They fitted 5 dead boxes in the underpan on the Sunday morning, each of them doing some of the welding. He said that Mr Clancy put the last one in at about 9:30 am and that he, Mr Davies, was standing beside him. At that point, he said, Mr Smith came down and said that he was going to breakfast. Mr Clancy then came down the ladder and went off to repair his car whilst the others went off to breakfast.

Mr Farla told the court earlier that the conveyor was protected by a layer of ballast or mixed material. I accept this. Mr Davies said that neither he nor Mr Clancy hosed down the underpan or the conveyor below it on the Sunday morning. He said that he could not remember if there was any rubber left on the inside of the underpan or on the surfaces of the old dead boxes, but if there had been any they would have burned it off when cutting out the old dead boxes.

I found Mr Davies to be a reliable witness and I accept his evidence, which was not challenged to any significant extent.

Mr Clancy was called by UMA. He used to be employed by the Defendants but later moved to UMA as a fitter.

Mr Clancy's first description of what he did in the scalping house on the morning of Sunday, 3 February 2008, was brief. The hand written statement that he made on 4 February 2008 read as follows:

“Arrived on site approx 5 to 7. Went into canteen got changed into work gear, went around to workshop/Scalpen House about 7.25. Measuring up and fabricating dead boxes in workshop. Went to tea about 9.45-9.50. . . "

Although he said in his witness statement that he had no recollection of doing any welding on the morning of the fire, in cross examination he accepted that he may have done some tack welding that morning: he said "I'm not saying no welding took place".

I did not find Mr Clancy to be a reliable witness or an impressive person and, to the extent that his evidence conflicted with that of Mr Davies, I unhesitatingly prefer the latter. I am quite satisfied that Mr Clancy was involved in tack welding in the underpan of the scalping house on the morning of the fire. I have a strong suspicion also that he told Mr Macdonald that no hot work had been carried out in the scalping house on that morning. Nevertheless, it is surprising that Mr Macdonald, UMA’s Safety Health and Environmental Manager who directed the investigation into the cause of the fire on the company’s behalf, did not ask the other person named on the Permit to Work for the scalping house, namely Mr Davies, for his account of what happened on the morning of the fire.

Mr Smith was a charge hand engaged by the Defendants. He is an engineer/welder and fitter by trade. By the time of the fire he had been working at the site for about 20 years. He clearly knew it very well. In most respects I found him to be a reliable witness whose evidence was not shaken in any significant respect during a probing and sustained cross-examination. I have no doubt that he is a very skilled and conscientious craftsman.

He is, perhaps unsurprisingly, adamant that his work was not responsible for the fire. Since he had carried out much the same work in the same way many times without incident, this is understandable.

Mr Smith made a short handwritten witness statement on the afternoon of the day of the fire under the supervision of Mr Macdonald. In relation to the events of Sunday, 3 February 2008, it read as follows:

“Approx 7:30 am myself & Don & Ray left workshop to start work upstairs. I myself turned on fresh water pump,

Don went inside screen and underpan to clean pipework ready for weld repairs

Ray had one lifting lug to finish welding on A-side secondary.

I myself had to burn 20 bolt heads off of tube flanges. 12* drive side, 8 walkway side, I do this by placing a shovel under each bolt so as not to drop hot bolts inside underpan. I had done 16 this way and changed bolts, the last remaining 4 were on rosta bracket at walkway front, don't have use a shovel because of bracket shelf. 100% sure no hot bolt fell into the underpan, as I discarded every hot bolt onto mezzanine floor. I never left area for at least a approx 35-40 m after last bolt was removed, checked area turned off water, went to breakfast at 9:30 am.”

(* Initially written as 8 and then changed to 12)

In cross examination a draft witness statement was put to Mr Smith that had been taken by a Mr Wookey, a loss adjuster appointed by the insurers. Mr Smith did not sign this witness statement, and the copy in the trial bundle had various annotations that were made by Mr Marshall. Neither Mr Walker nor Mr Shepherd put the statement to Mr Smith during his initial cross examination, but Mr Walker then applied to put it to him after Mr Shepherd's cross-examination had concluded. This I allowed.

After referring to an inspection of the A side primary screen that was carried out on Friday, 1 February 2008, by Mr Farla and representatives of the Defendants, probably Mr Marshall and Mr Smith, the draft statement described the events on the morning of Sunday, 3 February 2008, in the following terms:

“9.

I would estimate that we were at the primary screen ready to start work by about 7:30 am. Ray Shaw was working on the secondary screen which is adjacent to the primary screen. Ray was undertaking the final works to attach lifting eyes to the secondary screen.

10.

I proceeded to remove the guards from the day side/drive side of the primary screen and instructed Don to clean up the central rail within the primary screen using an angle grinder and a wire brush, ready to be worked on later in the day.

11.

I placed the flexible hose in the underpan and turned the same on, leaving a continuous flow of water down approximately half of one side of the underpan. At the bottom of the underpan there was a bed of wet sand. I then proceeded to cut the nuts off eight bolts that hold the metal tube within the screen in situ using an oxy-propane torch. I placed a shovel under each bolt to catch the hot slugs before dropping them onto the metal floor out of harms way. I would estimate that it took me no longer than four or five minutes to cut all eight bolts on the drive side. WATER RUNNING IN UNDERPAN AT ALL TIMES. MOVED AROUND TO EACH AREA OF HOT WORK.

12.

I then proceeded to undertake exactly the same task on the aisle/walkway side of the primary screen.

13.

The only difference is that on the last four bolts it was not necessary to catch the slugs with the shovel as they were above a metal shelf known as a rosta hanger and so there was no risk of them falling into the underpan. I would estimate that all of the cutting works were completed by no later than 8 am. The oxy-propane torch was switched off immediately after use. I am 100% certain that none of the hot slugs fell into the underpan.

14.

After checking the work that Ray Shaw was undertaking on the secondary screen I then proceeded to remove the remains of the bolt (sic) from the primary screen. Essentially I placed my hand through the gap between the primary screen and the underpan and attempted to catch each shank so as to prevent the same falling into the underpan.

15.

I then proceeded to replace all eight bolts on the day side/drive side and tightened them hand tight. This took no more than five minutes per flange to complete. I then returned to the walkway/aisle side and replaced four bolts. Unfortunately, I had not taken sufficient bolts from the workshop to complete the task and so left the cut bolts in situ above the rosta hanger.

16.

As it was approaching 9 am Don and I decided to leave the area to go for breakfast at a local cafe. Ray Shaw had left the area 10-15 minutes earlier for the same purpose.

17.

Don left the area first, leaving me to undertake a final fire check of the primary screen. I undertook a routine inspection of both sides of the screen, including looking inside the underpan. I was also smelling the air as, if there had been any ignition of the rubber lining of the underpan, it produces an obvious odour. I then removed the hose from the underpan, opened the valve fully and sprayed the underpan and Don Percival's area of work to ensure there were no hot spots in the same. I would estimate that it took me approximately two minutes to undertake this task. I then switched the water off and left the area at approximately 9 am."

Since this was nothing more than a draft statement, it is only necessary to make two observations on it. First, the underlining of the words in paragraph 11 and the words added in upper case at the end of it are by Mr Marshall, not Mr Smith. In relation to the words underlined, Mr Marshall had noted on the side "this description is not true". It was Mr Marshall’s evidence that this witness statement was not agreed by Mr Smith, which is why it was never signed and returned. Second, when the statement was put to Mr Smith at the conclusion of his cross-examination, only certain passages were put to him. As a general observation, he said that the statement accorded with the events as they occurred but that he could not comment now on the precise timings.

In his witness statement, made in December 2011 for the purposes of the trial, Mr Smith said, at paragraphs 11 and 13 to 22:

“11.

In accordance with standard practice, the entire area was wetted thoroughly using pumped water from a 50 mm hose located in the primary screen tower.

. . .

13.

. . . Don Percival went inside the primary screen and used a 4½ inch angle grinder and wire brush to prepare the central rail within the primary screen for welding work later that day.

14.

I removed 20 bolt heads from the tube flanges using an oxy-propane torch: 12 bolt heads on the drive side and 8 on the walkway side.

15.

Before commencing at this task, I placed the flexible hose in the underpan and directed towards the area we were working on. The hose was wedged between the top of the underpan and the screen. I turned the water on, giving a continuous flow of water down the underpan, hitting a central divide and then spraying the complete underpan.

16.

A bed of wet sand was laid at the bottom of the underpan to extinguish any hot sparks or platter (sic) that dropped to the bottom.

17.

I then proceeded to cut the necks off the 12 bolts on the drive side of the screen using an oxy-propane torch. I placed a shovel under each bolt and each one was caught. None dropped into the underpan. The task was repeated for 8 bolts on the walkway side, save that 4 of these bolts were above a steel shelf and so no shovel was required. I accounted for all 20 bolts and none fell into the underpan.

18.

Throughout this process Don Percival and I exercised extreme care and vigilance to ensure none of the sparks from the welding equipment ignited anything. We also directed the hose in the underpan awards the side on which the burning* work was being undertaken. (* this was changed by correction from “welding”)

19.

After the hot work had finished, the torch was switched off. The area was again hosed down thoroughly and check for any fire, smoke or burning smells and none was found. The hose was relocated in the underpan as before.

20.

The hot work finished at approximately 8 am. For approximately the next hour or more, I was in primary screen area. At no time was it left unattended. There was no sign at all of fire, smoke or burning smells. Absolutely nothing was found to indicate that any fire had started. I am very experienced in my job. I believe that I would have smelled or detected any sign of fire or smouldering over this hour period. There was none.

21.

At approximately 9 am Don Percival left the primary screen tower. I made a final fire check and then turned the hose to full and thoroughly doused the entire area in cold water for at least 2 full minutes.

22.

At approximately 9:30 am Ray Shaw, Don Percival and I left site to go to breakfast. Mark Garnham remained in the our (sic) office on site. This was 1½ hours after the hot works had finished. There was absolutely no indication of any fire, smoke or burning smells on site."

In evidence, Mr Smith said that he made his first, hand written, statement under "duress". By this he explained that he meant that it had been a traumatic day and he was still feeling very shaken by the occurrence of the fire earlier that day. He wanted to make a statement the following day. He says that Mr Macdonald refused his request and insisted that he make the statement then and there. This, he said, must have been why he referred to Mr Percival going "inside screen & underpan" when what he was referring to was the screen alone.

His evidence at the trial was that he left the primary screen area at approximately 9:30 am (which accords with the handwritten statement of Mr Percival), not 9 am as the draft statement suggests. Mr Smith told Mr Heath that the hot work finished at some time between 8 - 8:30 am and that he remained in the area for 1-1½ hours after the last hot work had been carried out. According to Ms Russell’s notes, he said that he stopped burning by 08:15 am.

I find as a fact that Mr Smith had completed his cutting of the bolts by or shortly before 8:15 am and that thereafter he remained in the area until shortly before 9:30 am.

As I have mentioned, Mr Smith said in evidence that the lining material that was used in the underpan was changed in 2006. The original material had been black. The new material was red and Mr Smith and others were concerned that it might be Linotex, which was a highly combustible rubber material that was also red. He said that carrying out hot work in the immediate vicinity of the Linotex was a "no-no". The new material was called Semperit and Mr Smith said that the Defendants were not prepared to carry out any hot work close to it until they knew how combustible it was. He said that UMA's Wharf Manager at the time, Mr Langton, obtained a specification and told them that it was less combustible than Linotex and, indeed, less combustible than the material that it replaced (the evidence of Mr Marshall was that the Defendants were told that it was not as combustible as Linotex).

A technical specification for Semperit was included in the trial bundle, but it is silent about its combustibility. It has annotated on it in someone's handwriting "Spec for lining rubber. From Gary Marshall". Mr Smith, who was shown the document, could not identify the handwriting. However, notes taken by Mr Heath of the conversation that he had with Mr Gary Marshall on 6 February 2008 record that the data sheet for the rubber lining "was obtained by UMA Gary Cauthen (now works elsewhere)". The evidence did not disclose who Mr Cauthen was, or what position he held within UMA. On the basis of this note, I find that the specification was obtained by UMA, not by Mr Marshall (although he may well have received a copy of it).

Mr Smith said that after the Defendants had been given this assurance about the new lining material, the system of working near the primary A screen underpan did not change. I accept this evidence.

Mr Smith explained in evidence why the flange bolts had to be removed before vee-ing out of the cracks could take place, which was because otherwise the heads of the bolts would be in the way. If the cracks were propagating from the bolt holes in the flanges, as Mr Smith said they were, then this is self-evident: if the bolts were left in position the angle grinder would cut into the head of the bolt as it was vee-ing out the crack. Accordingly, said Mr Smith, Mr Percival could not start this part of his work until Mr Smith had finished his. This accords with common sense and I accept it.

Mr Smith said that after he had cut the nuts off all the bolts and they had cooled he removed the remains of the bolts by tapping them out with a hammer and letting them fall into his hand, which he put through the gap between the screen and the underpan. He said that he did not catch all the bolts - he dropped about four - but that presented no fire risk because by then they were cold. When this was done Mr Percival started vee-ing out the cracks, and then cleaning the flanges using the wire brush attachment. He said in cross examination that one or other of them kept a fire watch whilst the other was doing hot work.

Mr Smith said that when he was removing the nuts from the bolts he used the oxy-propane torch in a “washing” mode, not a “lancing” mode. The nozzle of the oxy-propane torch was fed by three pipes: one contained propane and was controlled by a red knob, the second contained oxygen and was controlled by a blue knob and the third was the oxygen injection feed that was controlled by a lever. The oxy-propane flame is used to heat the steel to melting point and the operator then applies a jet of oxygen to heat the steel still further and blow away the resulting molten metal.

Mr Smith explained that if a high jet of oxygen is used the process is known as lancing, and when that happens a lot of splatter is produced that can fly in all directions. Washing is a process that involves a reduced flow of oxygen, which can result in the level of splatter being reduced by about two thirds. He agreed that it was possible that some particles of molten metal could pass through the gap between the screen and the underpan, but that he regarded the risk of this as minimal. He said that if any particle went through the gap, the water from the hose would cool the molten steel as soon as it hit it.

Mr Smith said that Mr Percival was sitting on the screen keeping a fire watch whilst he was cutting the bolts. He said that Mr Percival did not enter the underpan at any stage, and that Miss Russell had misunderstood the position (in thinking that he did so). He said that the underpan was being sprayed throughout the hot work and so Mr Percival could not have been working in there.

Mr Smith said that he had done this job many, many times in the same way and there had never been a problem.

At a meeting that was held at a hotel on 19 March 2008 to investigate the fire, Mr Smith was recorded as saying, apparently in relation to Mr Percival, that he "did not need boards due to slope”. This note of the meeting, which was taken by Mr Burgess (who was present in court but was not called as a witness), was not put to Mr Smith initially, and so he was recalled so that he could deal with it. He said that he did not remember making the remark, but it probably referred to the fact that when keeping a fire watch “you wriggle between the tubes” at the end of the screen - where the sloping floor of the underpan was fairly close to the lower layer of tubes in the screen - so that it is possible to stand or sit with one's feet on the sloping floor. He said that this was a perfectly good position from which to keep a fire watch because, if the screening mats have been removed, it is possible to see quite clearly between the tubes into the bottom of the underpan. He thought that there may have been a scaffold board on the screen on which Mr Percival had been sitting. He said that there would not have been a scaffold board in the underpan.

Mr Smith was asked about the reference in paragraph 10 of the draft statement prepared by the loss adjuster to Mr Percival being instructed to "clean up the centre rail within the primary screen". He said that this must have been a mistake, because it was no part of Mr Percival's job on that day to clean up the central rail. He was then referred to paragraph 13 of his witness statement prepared for the trial in which he had said exactly the same thing. He said again that he thought that the reference to the central rail was a mistake and that it should have referred to the flanges, because there was no work that had to be done on the central rail.

I discuss the question about whether Mr Percival was told to clean the central rail later in this judgment, but it is obvious that, if Mr Percival was cleaning the central rail whilst Mr Smith was cutting off the bolts with the oxy-propane torch, he could not at the same time have been carrying out a proper fire watch. Mr Smith was adamant that Mr Percival was carrying out a fire watch whilst he was cutting the bolts.

In relation to the use of fire blankets, which were identified in the Risk Assessment as being an action required to start or continue the job, Mr Smith said that they had fire blankets available but they were not "on the actual job". He said they were to be used if it was deemed to be practicable. He said that it was his practice to identify in the risk assessment all the measures that they might need to use. I will return to the risk assessment later in this judgment.

He was working with Mr Shaw who was welding the lifting lugs. He said that Mr Shaw was doing the welding and he was the dedicated fire watch man. He said that that was the normal system; it was not because there were any particular combustible materials in the immediate vicinity. He said that he had a hose and was spraying the back of the screen on which Mr Shaw was working. They were doing this work on both Saturday and Sunday.

He said that by Sunday there were only 2 lugs to weld and the welding was finished before 8 am. Until then no hot work was being carried out in the primary screen because there was only one hose available. He said that after the welding was finished Mr Smith came and took the hose back and began cutting off the nuts shortly after 8 am. He identified the hose as being the one shown in the photographs. At about 8:30 am he went back to the workshop.

As soon as Mr Garnham was told about fire he went to the scalping house, but he did not see Mr Clancy there. He said that the belt of CV6 was straight in front of him: it had come down between two rollers, and was hanging in the shape of a U, both sides alight. He seized a fire extinguisher from the workshop in case Mr Clancy was still in the underpan and he called out to him. He said that he was in something of a panic and could not get the pins out of the fire extinguishers.

He said that the left-hand side of the belt caught fire more vigorously and the fire then went up into the scalping house; he could not really tell whether or not there was a fire already in the scalping house.

He believed that the lining of the underpan had been assessed as low risk, from a combustibility point of view, and that when the new lining material was introduced Mr Smith had asked Mr Langton, UMA’s Wharf Manager at the time, about its flashpoint. He said that he recalled a discussion about it prior to the carrying out of some work on a previous occasion.

Mr Garnham said that if it was not possible to cover combustible materials with a suitable blanket, then they would hose down the material thoroughly and keep running water over it during the hot work. After the work was completed they would stay back and hose down the area. He said he was aware of the need to keep a fire watch after the hot work was finished. He said that he remembered Mr Percival being on the screen before the hot work started on the A side primary screen.

I found Mr Garnham to be a truthful and reliable witness and I accept his evidence, save that I do not necessarily accept his recollection of times where it is at variance with that given in other accounts.

Mr Marshall made two witness statements. The second witness statement was made to address certain points that had arisen out of witness statements previously served and some further disclosure by UMA, including numerous photographs. Much of the content of these statements was directed to UMA's investigation following the fire, which was carried out under the direction of Mr Ian Macdonald. I discuss this in more detail in a later section of this judgment.

Mr Marshall was on occasions a rather testy witness, and a little inclined to be argumentative. However, I suspect that he is probably deeply concerned that the survival of his business may depend on the outcome of this case and so his attitude is not hard to understand. However, where he was giving evidence on matters of fact - his evidence sometimes strayed into matters of opinion - I consider that, with one or two exceptions, he was trying to the best of his ability to give truthful evidence.

He was asked about a document (which is discussed in more detail later in this judgment) entitled "Safe System of Work for rebuilding support frames including . . . general fractured or worn parts on Primary, Secondary, Dewatering and scalping screens". He agreed that the work that was being carried out on Sunday, 3 February 2008, fell within that description. Whether or not this was a correct concession is a point that I will have to decide. He agreed that the Defendants had all the equipment available that was identified in the document including, in particular, fire blankets.

Mr Marshall was then referred to a note of a talk on fire risk and prevention on 8 October 2004 which stated that work which involves burning and/or welding was to include a 30 minute fire watch, and he said that it was mandatory for someone to remain on the site after hot work had been completed. He was also referred to paragraphs 2.2 and 2.3 of a briefing on UMA's Hot Work Procedure that was given on 16 January 2006, which referred to covering or protecting combustible items, but he said that he thought constant running water was preferable. He agreed that any job needed a risk assessment and a method statement before a Permit to Work could be issued, and he accepted that just because a precaution was not mentioned in the Permit to Work that did not mean that it did not have to be taken.

He said that if the Defendants saw something wrong, they would mention it to UMA but as a general rule it was for UMA to decide what was to be done.

He said that he was involved in discussions on Friday, 2 February 2008, about the work that was to be carried out over the weekend. He said that he actually got in to the underpan on the A side primary screen to have a look at things and that Mr Farla was with him at the time. The screen had not been working properly, so they knew that something was wrong with it. However, there was no work to be done in or from the underpan.

He agreed that he saw the draft statement from Mr Smith that was prepared by Mr Wookey, and that he made the annotations on it. He said that he was not happy about it (amongst other things, he did not think that the timings given at paragraphs 16 and 17 were correct) and he did not think that Mr Smith should sign it. He discussed it with Mr Smith who was not happy with it either and so it was never agreed.

In his witness statement he said that the Defendants had installed a pumped water system throughout the site with some 20 hose outlets. He explained that these outlets, including the one feeding the hosepipe used by Mr Smith, carried pumped water that was at three times mains pressure. If the water was directed at the side of the screen it would hit the central divide. The work being carried out by Mr Smith had been done "hundreds of times" and he knew how it was done. He was not aware that any work had to be done to the central rail of the screen.

Mr Marshall was asked why Mr Percival had not given evidence. He said: "I'm not sure - I would need to check that". And again, when asked if he knew that Mr Percival had been sent a draft witness statement by his solicitors, he said that he would need to check that. He then said: "I've got no objection to Mr Percival's comments", to which Mr Walker retorted by putting to him that he didn't know what those comments were. Mr Marshall responded by saying "obviously I don't".

I am bound to say that I find this evidence very difficult to accept. Mr Marshall was, on his own evidence, a manager who kept a tight grip on his organisation and he was present throughout the trial, effectively giving instructions to Mr Stuart Hornett, who represented the Defendants. I have no doubt whatever that Mr Marshall knew why Mr Percival was not being called to give evidence but was reluctant to say so. I appreciate that the exercise of investigating why a witness has not been called to give evidence can involve trespassing on questions of legal professional privilege, but it was open to Mr Marshall to say that he knew the reason but did not see why he should have to disclose it, or something similar. However, when he said that he did not know the reason why Mr Percival was not being called I do not believe that he was telling the court the truth.

By reference to a letter that he had written to his insurance brokers on 28 February 2008 it was put to him that this was the first time the possibility had been raised that the fire started in the scalping house. He said that this was not so and that it had been discussed with Mr Macdonald, and possibly Mr Burgess also on 4 February 2008. However, I am doubtful about this.

Mr Marshall said that in his view this type of work on the screens could be done safely with sufficient water in the right place and a proper fire watch. He considered that the precautions taken by the Defendants on the day of the fire exceeded what was required. He described his men as trusted employees who had an excellent safety record. He said that in the 21 years during which the Defendants had been working at the plant they had had an unblemished safety record. He said in his witness statement that he had not received a single complaint from UMA in relation to the defendant's working methods. These last two points were not challenged.

When cross-examined by Mr Shepherd, he agreed that he would not have shown his men the Burning and Welding Warranty in the insurance policy, but he said that his men, including Mr Smith and Mr Garnham, knew how they had to work. It was put to him that it would have been simplicity itself to fill the gap with a fire blanket in the manner suggested by the insurers’ expert, Mr Heath, but he disagreed. He said he liked to be able to see through the gap so that he would know where the sparks were. He did not agree that fire blankets were mentioned in the Risk Assessment because they were necessary to protect the lining.

He was asked whose decision it was to play a hose on the side of the underpan as a method of protecting the lining and he said that it was agreed with UMA's management. He had discussed it with Mr Farla and, although it was a joint approach, UMA made the ultimate decision. He said that they went through the method statement and Risk Assessment and would have discussed the procedures. He said that fire blankets were included in the Risk Assessment because when preparing for work all equipment that might be possibly needed is listed.

He said that when the new lining material was introduced the Defendants were assured by Mr Langton that it was not as combustible as Linotex. They understood that it was a material into which they could carry out hot work.

Mr Marshall accepted that when the cutting was being carried out by Mr Smith some particles of molten metal might pass the shovel and enter the underpan through the gap. He said that the protection of the lining was provided by the running water which, because of its high pressure, would spray all over the area in the vicinity of the work providing complete coverage to the lining. He said that he would do the same again if the same situation arose, because the methods used were the most comprehensive.

In re-examination he said that he did not think it would be easy to clip welding blankets in position as suggested by Mr Heath because once they were saturated with water they would become very heavy and they would also be a potential source of obstruction. He said that UMA's management, including the Operations Manager, Mr Burgess, were fully aware of the methods used by the Defendants. The former Wharf Manager, Mr Langton, had been at the plant for 20 years. Mr Curtis, who issued the Permit to Work, had been at the plant for 10-15 years and had issued similar permits on numerous occasions. He said that no-one within UMA had ever suggested that the Defendants should use fire blankets or screens in the manner now being suggested. He said that the site was well covered with CCTV cameras and Mr Farla had a TV monitor on his desk and could watch the work being carried out by the Defendants if they wished to do so.

He accepted that references to a "dedicated fire watch" referred to a person who was fire watching and nothing else. He said that the Defendants were paid on an hourly basis per man, not by the job. That was one of the reasons why he walked the job with Mr Farla beforehand: the number of men required was one of the things discussed. Mr Farla knew in detail what they were proposing to do that weekend. I accept this evidence.

It will be apparent from this summary that parts of Mr Marshall's evidence were based on what he would have expected to happen, rather than on what he knew from his own experience. I consider that Mr Marshall was, save where I have specifically found otherwise and at least in relation to working practices, a truthful witness, if sometimes a little inclined to reconstruct events, particularly when he was being asked about matters such as his men's knowledge of the terms of the insurance policy. However, I accept his evidence, which is entirely consistent with that of Mr Farla, that UMA knew precisely what precautions were going to be taken by Mr Smith when he was cutting the nuts off the bolts on the A side primary screen. I also accept his evidence that the Defendants had a very good track record for safety and that there had never been any safety related problems of any consequence.

Mr Ray Shaw worked for the Defendants between October 2007 and February 2008. Accordingly, he could only give limited evidence about the Defendants working procedures in general, apart from saying that he used to work to set procedures. On the morning of the fire he had been welding lugs to the secondary screen.

He described his witness statement as drafted by someone else, he could not say who, and that it was just a “rough statement". Although I am quite satisfied that he was perfectly truthful, it was equally clear that he did not have a very good recollection of events. However, as it turned out he was able to give little evidence that was of any real relevance to the issues in the case.

Mr Stephen Lynch is a Senior Claims Adjuster with Novae Syndicates Ltd. He was an unimpressive witness who was seldom able to give a useful answer to any question. It was not clear to me why he was called to give evidence at all. The only thing that emerged from his evidence was that it disclosed a failure by the insurers to make proper disclosure of their own documents, amongst which was a proposal form completed by the Defendants in 2001 in which it was stated that they had had no claims within the previous 5 years.

Louise Russell was the fire expert instructed on behalf of UMA (or, more accurately, its insurers). She is a partner of Dr J H Burgoyne and Partners, the well known consulting scientists and engineers. She has a degree in Chemical Engineering from the University of Nottingham and graduated in 1993. From 1993 to 1997 she worked as a laboratory scientist in the Fire and Explosion Division of a company specialising in process safety. She then moved to Pirelli Cables where she worked as a chemical process engineer before joining Burgoynes two years later in 1999. Since then she has specialised principally in the investigation of fires and explosions.

I found her to be a candid and competent expert, although I consider that, for reasons that were very largely not of her own making, her investigation of this fire was not as satisfactory or thorough as it might have been. She was not supplied with many of the relevant documents until after she had completed her site visits, in spite of the fact that she had asked for them, with the result that she did not know their contents when talking on site to those involved in the relevant events. In addition, she was given an introductory briefing by Mr Macdonald parts of which appear to have been either wrong or misleading. As I will explain in more detail in the next section of this judgment, I consider that this led her to misunderstand what work had been carried out and how it had been carried out.

Miss Russell arrived on site at about 3 pm on Monday, 4 February 2008, where she met her assistant, a Mr Griffiths, who had arrived a little earlier. On that day she spoke to Mr Macdonald and Mr Farla, of UMA, and Mr Beasley, of the London Fire Brigade. Her notes suggest that she had quite a long talk with Mr Macdonald, who explained in some detail what he understood had been done by way of work in the screening house. He told her that hot work in the scalping house had not started. She said that she asked for documents such as any witness statements, permits to work, risk assessment and so on.

Two days later, on 6 February 2008, she returned to the site and met up with Mr Heath, the expert instructed by the insurers. Whilst they were in the primary screen area Mr Smith came up and spoke to them both. She said it was a fairly informal conversation and that it was the only time that she spoke to Mr Smith. She wrote her first report on 26 February 2008, and was not instructed again until 29 December 2011 which is when she first found out that the Defendants were alleging that hot work in the scalping house was the probable cause of the fire. When she wrote her second report she had not seen any report by Mr Heath.

She accepted, with hindsight, that she should have taken a sample of the conveyor CV6, and she said that she would have liked to have spoken to all the other people involved. I will deal with her evidence about the whereabouts of various objects in the primary screen area in a later section of this judgment.

In relation to the cause of the fire, her view was that the pattern of the damage indicated that the fire had started in the screening house. However, she accepted that in her first report dated 26 February 2008 she did not discuss the possibility that the fire might have started in the scalping house, although she said that she had considered it. She agreed that the area of greatest fire damage was not always where the fire started, and she accepted that the tendency was for fire to travel upwards as a result of convection effects.

She accepted that flames would be expected to lie reasonably flat if a fire was travelling up the conveyor. However, whilst she agreed that the shape of the conveyor would cause a slight "trench" effect which would increase the rate of spread of a fire along it, this would be mitigated to some extent by an eddy effect caused by the wind: that, she thought, would remove heat from the upper surface of the conveyor. In response to a question from me, she said that if the conveyor caught fire at the bottom end, her best guess was that it would probably take 5-10 minutes to spread up the conveyor to the screening house: but she explained later that the belt would have to be sufficiently well alight for this to happen, and this in turn would require there to have been a significant fire at the lower end of the belt in the scalping house.

Mr Hornett suggested that a fire in the screening house would vent much more easily because part of the roof was missing, and would therefore be more visible than a fire in the scalping house. Miss Russell disagreed: she said that there were always gaps between profiled metal sheets such as those that formed the cladding to the scalping house and that these would give rise to a chimney effect. She thought that a fire would be very obvious if it was spreading up the conveyor.

Miss Russell did not agree that there would be no dynamic to drive a fire down a conveyor belt. She said that although fire would travel upwards preferentially, once it was raging it could spread quite quickly down the belt so that it could have been seen in the position described by Mr Farla just before the belt broke (if that is what he saw).

She agreed that one can never eliminate the risk of fire, and that even if fire blankets had been used a fire might still have resulted. She did not think that individual fire blankets could be draped over the inside of the underpan to protect the lining (as suggested by Mr Heath): there would have to be something specially made.

In response to a question from me, she said that a risk assessment would usually precede the method statement and that the two documents would be required in order to obtain a permit to work. However, in her report (at paragraph 4.3) she referred to the method statement as if it was part of the Risk Assessment. I will have to refer to these documents in more detail later in this judgment.

For reasons that are set out later in this judgment, I accept Miss Russell's opinion that the fire started in the screening house. But insofar as she thought that Mr Percival had been working in the underpan itself, I consider that she was mistaken.

Mr Richard Heath was the fire expert instructed by the insurers. He has a degree in Electronic Engineering from the University of York and became a Chartered Engineer in 2003. In April 2005 he attended the University of Denver Fire Science and Fire Investigation course. The first eight years of his professional life were concerned with networks and communications systems. He joined Hawkins & Associates in 2004, of which he became a Senior Associate in 2008. It seems that it was not until he joined Hawkins & Associates that he became involved in the investigation of fires.

He attended the site on 6 February 2008, having agreed to meet Miss Russell there. He was met by Mr Marshall, who gave him an initial briefing, and later he met Mr Smith. He spoke also to Mr Farla, whom he met on the site. He was shown (but not given) copies of the handwritten statements of Mr Marshall, Mr Smith, Mr Shaw, Mr Garnham and Mr Percival, from which he made notes, although he did not speak to any of the last three. In addition, on site he met a Mr Fasham, who was an electrician engaged by UMA.

During his visit on 6 February 2008 Mr Heath was not told by Mr Marshall, or any other of the men who worked for the Defendants, that hot work had been carried out in the scalping house immediately before the fire. He learned of this much later. As a result, he did not consider the scalping house as a potential source of the fire during his initial investigation.

Mr Heath was told that Mr Smith had used a shovel to catch the hot nuts that had been cut off the bolts. In his report he wrote that "The shovel was wedged into a gap in the vibrating screen and the underpan", and this was illustrated in one of the photographs attached to his report. Other considerations aside, it is perfectly clear to me that a shovel in the position shown in Mr Heath's photograph, positioned as it was at an angle of 15° or 20° to the horizontal, would not have been a safe or sensible way to catch a hot bolt that was going to drop onto it from a height of about 6-8 inches: the chances were that the head of the bolt would either bounce or roll off it. Although the shovel appears to have been a typical builder’s shovel, with slightly scalloped edges, I am very doubtful that these would have been sufficient to prevent a falling bolt from rolling off the shovel. If a hot bolt end dropped onto the blade of the shovel in that position it might well bounce off it and strike the operator. It would, in my view, have been a very dangerous way of working. Mr Heath said that that is how he found the shovel when he first went to the area of the A side primary screen and so he assumed that that was how it had been used. He said that he could not be sure whether or not Mr Smith had told him that himself.

I have to say that I regard Mr Heath's assumption that the shovel was used in the manner shown in his photograph as surprising, for the reason that I have given above.

Mr Heath was not a welder and so he was not able to comment on Mr Smith's evidence about the difference between "lancing" and “washing", although he thought that with a lower oxygen pressure the splatter particles would be larger but would not go so far.

Mr Heath agreed that metal halide lights had a potential to explode, and he had retained one of the lights that was in the area of most concern. That light was fitted with a safety shield to contain the glass if the light should explode. Mr Heath was unable to check the second light, because he was told subsequently that it had been disposed of, but he thought it unlikely that an exploding light was the cause of the fire. I do not consider that the fact that he had not seen the other light was a material weakness in his investigation.

Mr Heath agreed with Miss Russell that the most likely cause of the fire was the hot work that was being carried out to the A side primary screen. He made a number of suggestions as to how it could have been prevented.

In his report he said that the simplest method would have been to stand sheets of non-combustible rigid material (for example, thin sheets of steel, or possibly plasterboard) on the walkway beside the underpan so as to cover the gap between the screen and the underpan. In the course of his cross-examination he withdrew this suggestion, partly because it had been pointed out to him that there were various obstructions on the walkway that would have made it difficult to lean a rigid sheet against the side of the underpan and partly because, whilst the walkway consisted of horizontal sections, the underpan was not horizontal.

His next suggestion was that welding blankets could be clipped to the underside of the screen and then fed through the gap between the screen and the underpan so that they hung down the sides of the screen (on the outside) and therefore closed the gap; alternatively he suggested that welding blankets could have been placed inside the underpan so as to cover the lining.

Miss Russell did not favour the first of these two suggestions, although initially she agreed with the second. I reject the suggestion that it would have been practicable to cover the lining inside the underpan with fire blankets. There are three reasons for this. First, whilst I agree with Miss Russell that if this was to be done, it could be done only by some specially made cover, I am not persuaded, in the light of the configuration of the screen and underpan, that it would have been practicable to insert and fit such a cover given the very limited access. Second, such a cover would have to be tailor-made for every different underpan and there is no evidence that UMA would have been prepared to adopt such a solution if it had been suggested: the covers would probably have been very expensive to make and they would have been time consuming to install. Finally, if this was a practicable and cost-effective solution one might have expected UMA or the Defendants to have thought of it and adopted it at some time during the 20 years before the fire. Even now, there is no evidence that any consideration has been given to adopting this solution since the occurrence of the fire. So far as I am aware, the Defendants continue to carry out this operation in exactly the same way today as they say they did on the day of the fire.

As to Mr Heath's first proposal, that welding blankets could be clipped to the underside of the screen and draped through the gap, on the face of it this appeared to be a reasonable proposal - although not one that Miss Russell was prepared to adopt. However, it has not been tested - and indeed it has not even been demonstrated that such blankets could be clipped to the underside of the screen in the manner suggested by Mr Heath in such a way as to ensure that they did not fall down unexpectedly during the carrying out of hot work, particularly if they were to be used in conjunction with a high-pressure hose (which I understand to be the suggestion). In the light of what I regard as Mr Heath's surprising approach to the use of the shovel and the abandonment of his suggestion of rigid screens, I am reluctant to accept his evidence on a practical proposal such as this when no attempt has been made to demonstrate that it is viable in practice and when it is not supported by his fellow expert. It may be that it would be a viable option, but the onus is on UMA or the insurers to demonstrate this and in my judgment they have not done so. In the four years that have elapsed since the fire there would have been every opportunity to do so.

One undamaged piece of lining that was recovered from the B side primary screen underpan had several “pockmarks” on it. Mr Heath's opinion, with which Miss Russell agreed, was that these marks were probably the result of hot pieces of metal, or spatter from cutting, landing on the surface and partly burning it. He said in evidence that he thought it unlikely that such marks would be made by a spark or hot article falling vertically, he thought that it would have to be a particle travelling on a path that was nearly normal to the plane of the surface of the lining. For this reason amongst others, Mr Heath excluded the possibility that the pockmarks were the result of hot material falling onto the lining during the fire.

I accept view of Mr Heath and Miss Russell that these pockmarks were probably caused in the manner that they suggest but, contrary to their view, it seems to me that this demonstrates that the method of using running water to protect the lining does work in practice. The existence of the pockmarks, but no more serious damage, suggests that the offending particles were extinguished by the water on or very soon after impact. Whilst the water may not have been sufficient to have prevented the lining from being damaged, it must have prevented that damage from becoming a fire which is, after all, its purpose.

It was put to Mr Heath on behalf of the Defendants that the difference in severity of the fire damage to the A and B primary screen underpans was consistent with fire spreading from the conveyor above because the B side underpan was to some extent protected by the screening mats which had not been removed on that side. Mr Heath said that he did not consider the conveyor CV6 as source from which the fire spread to the underpans, because there was no evidence of any source of ignition in that area. However, this does not really meet the point because the premise underlining the question was that the fire had spread up the conveyor from the scalping house to the top of the screening house.

In relation to the possibility that the fire started in the scalping house, Mr Heath thought it unlikely that the rubber of the conveyor could have been ignited by hot material from tack welding because when tests have been carried out involving fires on conveyor belts considerable heat is required to ignite the belt in the first place (such as the use of a blow torch).

Mr Heath said that if the fire had spread up the conveyor one would have expected to have seen black smoke and more damage at the point where CV6 entered the screening house. He said that lack of damage could be explained if a section of the belt was on fire inside the screening house and the belt then broke so that the flaming end was pulled back down the conveyor under gravity thereby passing quickly through the entrance.

He said that he disagreed with Mr Townsend's conclusion about the trench effect. He said that you could not compare the situation of a belt that is in the open air with one that is in a tunnel, even making allowance for the fact that CV6 was on an incline.

Mr David Townsend retired from the London Fire Service in October 2006 having completed over 30 years continuous service. During that time he acted as an operational crew commander and watch officer and was responsible for the investigation of all fires that he attended. He estimates that he attended in excess of 4,000 such fires. During the final 11 years of his service he specialised in fire investigation.

During the last five years Mr Townsend has acted as an independent Fire Consultant and expert witness in relation to fires. He is an Associate Consultant to both the Building Research Establishment and the Fire Protection Association. He is a visiting lecturer at the Fire Service College.

Mr Townsend was not instructed immediately after the fire and did not become involved with this claim until, as I understand the position, some time in 2011. He first visited the site in October 2011.

Several pages of Mr Townsend's report were devoted to a detailed critique of the initial reports that had been prepared by Miss Russell and Mr Heath in February 2008. One of his criticisms of Miss Russell's report was that it was based on witness statements taken by site personnel, instead of statements that she had taken herself from those involved. Whilst there is force in this point, the fault does not lie with Miss Russell. She was not given a free hand when she visited the site shortly after the fire.

Carrying out what he described as "a cold case desktop study", Mr Townsend reached the conclusion that the most likely place of origin of the fire was the scalping house. He said that hot work in the underpan in the scalping house immediately above the conveyor CV6 set fire to the top surface of the conveyor belt and the fire, assisted by a following wind, spread up the conveyor to the screening house, where a major fire resulted. He considered that the spread of the fire along the conveyor would have been assisted further by what is known as the "trench effect", a phenomenon that was identified following the major fire at King's Cross in 1979.

It was his view that the fire would have been confined initially at the top surface of the conveyor belt so that it would not have burnt through the belt until it became involved in the major fire in the screening house. The ruptured burning belt, falling back down the conveyor rollers, then caused a substantial fire in the scalping house.

One difficulty that faced Mr Townsend is that he was mistaken about the wind direction. The wind was not blowing from the direction of the scalping house towards the screening house, but in the opposite direction. It would not, therefore, have assisted the spread of the fire up the conveyor as Mr Townsend postulated. His response, when he was challenged with this, was to say that an adverse wind would have produced an eddy effect which would have turned the wind in to the trough of the conveyor so as to boost the spread of the fire. However, he said that this would not make either the flame or the smoke more visible. He thought that the flame would lie low on the conveyor and that the smoke would be light in colour, and therefore not readily visible. His conclusions in this respect were to an extent based on some tests that he had carried out on material similar to that of which the conveyor was thought to be made.

Mr Townsend mentioned the wind direction several times in his report and obviously placed some reliance on it. His response about the eddy effect was a little too glib to be convincing. However, Mr Townsend's overall point that neither Ms Russell nor Mr Heath gave any consideration to the scalping house as a source of the fire when carrying out their initial investigations is a valid one. If they had not been put off course by being told that no hot work had been carried out in the scalping house on the day of the fire, I suspect that they would have looked at matters differently. The responsibility for this state of affairs rests squarely with UMA.

On the whole, I thought that Mr Townsend was a reasonably reliable witness, if a little partisan, who made some good points in spite of coming to the scene so late in the day. The extent to which I accept any of his conclusions is discussed in the next section of this judgment.

Numerous papers were examined in the course of the evidence, the first was produced by Miss Russell at the experts’ meeting and the remainder by the Defendants at the trial, about tests involving the spread of fire on horizontal conveyor belts in tunnels with a view to establishing the likely rate at which a fire might have spread up CV6. I derived almost no assistance from these and the result of the cross examination of the experts really did not advance matters. I accept Mr Heath’s evidence about the difficulty of drawing any conclusions relevant to this case on the basis of experiments conducted on horizontal conveyors in tunnels.

As I have mentioned, when Mr Farla arrived at the car park he noticed a wisp of smoke from the top of the screening house. I find that this was at about 9:55 am. As he walked towards his office, having parked his car, he saw a lot more smoke coming from the screening house.

The photographs suggest that from the point where Mr Farla parked his car the scalping house may have been completely obscured by the much higher screening house in the foreground, so it is possible that if the scalping house was on fire he might not have seen it when he was parking his car. However, if the scalping house or CV6 was on fire by the time Mr Farla was driving towards the car park - which it would have to have been if the defendant's case is correct - it is surprising that he did not see smoke rising from the area of the scalping house as he approached the site.

As Mr Farla walked, or possibly ran, towards his office the scalping house, or at least the airspace immediately above it, would have been in his view and so I would have expected Mr Farla to see smoke in that airspace if there was a fire at the scalping house or on CV6 at that stage. In fact, he said in evidence that as he approached his office he could see the scalping house through the framework that supported the screening house.

As I have already noted, Mr Farla said in evidence that he did not think it was possible that at this stage the conveyor could have been on fire in the area of the scalping house, because if there was any visible smoke coming from the lower part of the conveyor he would have seen it when he was in his office and he would have told the Fire Brigade. Mr Farla said also that there was a rubber skirting in the scalping house around the base of the underpan, which suggests that if the conveyor caught fire below the underpan the rubber skirting would catch fire as well. In my view it follows from all this evidence that if there had been a fire of any significance on CV6 at the scalping house, Mr Farla would probably have seen it.

The other important eyewitness was Mr Garnham. He was one of those who made a handwritten statement shortly after the fire, in his case this was on the day after the fire, 4 February 2008. In that statement he described his actions just after Mr Farla had told him that there was a fire:

“I ran round to the work shop and call (sic) Terry Clancy who was working on the under pan of the scalpen screen. After getting no reply I noticed No 6 belt had come down under the scalpen screen house. I grabbed two extinguishers and attempted to extinguish the fire on the belt but the pins would not come out of the extinguishers. By now the fire had taken hold of the belt and had spread up into the scalpen screen. I then retreated to the car park and waited for the Fire Brigade.”

In cross-examination he said that whilst he saw that the belt hanging down from the scalping house was on fire, he didn't really see whether or not the scalping house itself was on fire. However, what he said on the day after the fire suggests strongly that the fire was not in the scalping house when he first arrived on the scene.

There was no very reliable evidence before the court as to the speed at which the fire might have spread up the conveyor, if that is what it did, but no one was disposed to challenge Miss Russell's estimate of 5-10 minutes. I accept it as the best estimate. Assuming that it would have taken about 2-3 minutes, or so, for the fire to develop in the scalping house to an extent sufficient to set the rubber belt of the conveyor well alight, then 5-10 minutes to spread up the conveyor and a further few minutes for the fire to develop in the screening house once it had reached the top of the belt, the part of CV6 in the scalping house (and surrounding area) would have to have been on fire for some 10 minutes, or possibly more, before any smoke would have been visible coming from the roof of the screening house. This means that there must have been a fire in the scalping house by 9:45 am. In his handwritten statement Mr Garnham said that Mr Farla told him about the fire "just before 10 o'clock", immediately before going to ring the Fire Brigade. This estimate coincides well with the time of the call to the Fire Brigade at 10:01 am. Accordingly, by the time Mr Garnham arrived at the scalping house at least part of it must, on the defendant's case, have been well alight for some 15 minutes (being 10 minutes before Mr Farla first saw the wisp of smoke at 9:55 am). Yet Mr Garnham did not recollect seeing any fire in the scalping house itself.

There is also the evidence of Mr Davies. In his handwritten statement made on 28 February 2008 he said that he left the site to go to breakfast at about 9:40 am. By that time the conveyor immediately below the underpan, or something combustible on it, must have caught fire. It is remarkable, to say the least, that he left the area without noticing anything. As an experienced fitter who had been involved in welding work, one might have expected him to make a quick check of the area before leaving it.

In my judgment, the evidence of these three eyewitnesses alone is sufficient to dispose of the defendant's case that the fire started in the scalping house. Another pointer to the same end is that no-one saw fire spreading up the conveyor CV6. Whilst I accept that it might have been difficult for someone standing on the ground below to see the flame of any fire spreading up the conveyor, I accept the evidence of Miss Russell and Mr Heath that there would have been visible smoke. One thing that is quite clear is that the conveyor belt broke, having been burned through by the fire, somewhere along the upper section in or near the screening house, and so, if the fire spread up the conveyor to at least that point, no one saw it do so.

The only difficulty in the way of a conclusion that the fire started in the screening house is the evidence of Mr Farla that he saw the conveyor on fire at a point which he said was about 10 m below the point where it entered the screening house. Mr Hornett submits that it is far more likely that the fire would have spread up the conveyor to this point, rather than down the conveyor from the screening house. He relies on two matters. First, fire has a tendency to spread upwards rather than downwards, although I accept the evidence of Miss Russell that fire can spread downwards. Second, that the absence of any significant fire damage to the wall of the screening house at the point where the conveyor passed through it is more consistent with a small fire spreading up the conveyor than a raging fire in the screening house which started to spread down the conveyor.

There is some force in these submissions, but I consider that the answer to them may lie in the fact that Mr Farla's recollection, some four years after the event, as to precisely where he saw the conveyor belt on fire could be mistaken. In his handwritten statement, made two days after the fire, Mr Farla said that he telephoned Paul Burgess at approximately 10:05 am to tell him of the fire, by which time the conveyor belt was on fire. He said the Fire Brigade arrived within 20 minutes of this, by which time CV6 had fallen to the ground burning fiercely and the fire had spread to the scalping house. This does not suggest that he was watching the belt at the time that it actually broke.

What Mr Farla said in his witness statement dated 15 December 2011 was this:

“5.

I could see that the conveyor CV 06 between the main screen house and the scalpen house was on fire. The fire that I saw was at the top end of the conveyor only. I saw no evidence of fire or smoke at the bottom end of the conveyor or any fire or smoke near or coming from the scalpen house. I phoned Paul Burgess at approximately 10:05 am and informed him of the fire. By that time CV 06 was burning fiercely.

6.

It was not long before CV 06 burnt through and dropped down to the scalpen house. I have marked the area where I believe the conveyor snapped on exhibit SF 2. The conveyor belt rolled down and collected that (sic) the bottom in the scalpen house."

(My emphasis)

The point marked on the photograph by Mr Farla was about 10 m below the point where the belt entered the screening house. I do not consider that I can place any great reliance on Mr Farla's estimate of the point at which the conveyor broke. It seems to me far more likely that the fire started in the screening house in the area of the primary screens and spread upwards to the conveyor and burnt through it in that area. I find that what Mr Farla probably saw was the conveyor belt on fire in the screening house, but since it is not clear from the evidence precisely where Mr Farla was when he telephoned Mr Burgess and saw the conveyor I cannot resolve this question more positively.

But on any view the upper section of the belt, once broken, would then have fallen by gravity down to the scalping house. The lower (return) would probably have done the same (although whether it would then have been restrained from falling further by the drive mechanism halfway along the conveyor, I cannot say). However, it is possible that the lower (return) part of the conveyor would have burnt through first, which would probably have had the result that the topmost section of the lower part would have been pulled by the weight of the upper part of the belt round the roller at the top of the screening house so that it all fell back down to the scalping house. It does not matter much which of these two alternatives is correct; either way virtually all of the upper section of the belt would have fallen back down the rollers under gravity to the scalping house. The top part of the belt would have been on fire, which would then have taken the fire to the scalping house.

The experts agreed (at paragraph 4.6 of their Memorandum of Agreement) that:

“We agree that the fire did not significantly develop in the Scalping House until after the conveyor belt (CV6) had broken."

The absence of any significant damage in the area where the conveyor passed into the screening house may be explained by the fact that the conveyor was on fire in the area of the primary screens, which must be some 10-15 m away from the point of entry, so that, when the conveyor burnt through, the part of the conveyor that was on fire passed fairly quickly through the entrance in the wall of the screening house causing relatively little fire damage as it did so.

Taking the evidence as a whole, but more specifically for the reasons that I have given, I am quite satisfied that the fire started in the screening house and spread to the scalping house. In the light of this conclusion, I do not propose to lengthen this judgment by discussing in any detail the work that was being carried out in the scalping house by Mr Clancy and Mr Davies. I consider that the tack welding that they were doing would not readily have set fire to the rubber of the conveyor belt on which they were standing if it was covered, as I accept it was, by ballast and other non-combustible debris that had been hosed down from the screen above. As I have already mentioned, the experts are agreed that tack welding is a far less potent source of fire then oxy-propane cutting.

The remaining parts of this judgment are based on the premise that the fire started in the screening house as a result of the hot cutting work carried out by the Defendants.

UMA relied on two documents. The first was the record of a tool box talk on Hot Work Procedure that was given on 16 January 2006. The note of the briefing was signed by those who attended, including Mr Marshall, Mr Smith and Mr Garnham and Mr Percival.

The note stated that all persons carrying out Hot Work were to follow the procedures set out. Those that have relevance to this case were as follows:

“2.2

All sources of fuel within a (10 m radius) shall be removed where possible. Any, which cannot be removed, shall be adequately protected from heat and sparks.

2.3

Where it is not possible to remove sources of fuel, e.g. conveyor belt, wooden walkways, rubber decks or chute linings, etc, these should be protected by spreading non-flammable dust, fire blankets, covering with steel plates, etc, where possible the area should be damped down using water.

2.4

Where it is not possible to protect such items, extreme care and attention are required whilst carrying out hot work to prevent a fire occurring.

2.5

Pay particular attention to areas below the hot work where sparks, hot metal, etc can fall and create a fire later.”

The Defendants contended in their pleaded case and in oral closing submissions, in my view correctly, that this document did not have contractual force at the time of the fire. They submitted that the relevant document was one entitled "Safety, Health and Environmental Rules for Contractors", Version 02 - May 07. This formed part of UMA's H&S Policies and Procedures Manual, and carried the reference: UMA H&SP.19 APP1. Section 22, entitled “Hot Work", stated that “all persons carrying out Hot Work must follow the procedures detailed in Policy 39 App 1”. Paragraph 22.1 defines Hot Work as including "the use of blow torches, oxy/acetylene, oxy/gas or electric welding/cutting equipment". I regard an angle grinder fitted with a grinding disc suitable for grinding or cutting metal as being electric cutting equipment falling within this definition.

Policy 39 App 1 contained the following provisions, in Section 2:

“2.2

All sources of fuel within a (10 m radius) shall be removed where possible. Any which cannot be removed shall be adequately protected from heat and sparks.

2.3

Where it is not possible to remove sources of fuel, eg conveyor belt, wooden walkways, rubber decks or chute linings, etc, these should be protected by spreading non-flammable dust, fire blankets, covering with steel plates, etc, where possible the area should be damped down using water.

2.4

Where it is not possible to protect such items, extreme care and attention are required whilst carrying out hot work to prevent a fire occurring.

2.5

Pay particular attention to areas below the hot work where sparks, hot metal, etc can fall and create a fire later.”

Punctuation apart, these precautions are word for word the same as those set out in the note of the tool box talk on 16 January 2006. Section 5 of the document, headed "Carrying out the Work", contained the following additional provisions:

“5.3

A constant check of the area to be made for any signs of fire.

5.4

The working area shall be monitored after the hot work has been completed for signs of heating or fire. The period of time will vary according to circumstances and the degree of risk involved (30 minutes to 4 hours or longer).

5.5

In high-risk areas, such as around fuel tanks, rubber lined screens/shoots, buildings of a combustible nature, areas where they fuel is present, the area will need to be monitored until the area has cool and there are no signs of heat, i.e. services cool, no smell of combustion, etc. Pay particular attention to areas below the hot work where sparks, hot metal, etc can fall and create a fire later.

5.6

. . .

If in doubt, ask the Local Management for further advice”

These also reproduced, verbatim, similar provisions in the note of the tool box talk.

Although Mr Ronald Walker QC, who appeared for UMA, relied on the note of the talk of 16 January 2006, that was mainly because the relevant operatives of the Defendants had all signed it. However, it may not matter for the purposes of this case which document has contractual effect, since the relevant provisions are, for all practical purposes, identical in both documents.

However the parties were not agreed on the relevance of a further document which was headed:

This document, which came into existence in about February 2005, and was signed by Mr Marshall, Mr Smith and Mr Garnham, contained the following provisions:

The Particulars of Claim referred, at paragraph 6(b), to an earlier tool box talk that was said to have taken place on 6 October 2004. In fact it took place on 8 October 2004. The paragraph then set out what were said to be extracts from the note of that talk. However, this was an error. The extracts were in fact extracts from the Work Procedure section of the "Safe System of Work" document that I have set out in the previous paragraph, not from the note of the tool box talk of 8 October 2004.

In paragraph 17.3 of the Defence, paragraph 6(b) of the Particulars of Claim was not admitted. It was admitted that there was a fire risk and prevention briefing dated 8 October 2004 (not 6 October), but it was then pleaded that "none of the words set out in paragraph 6(b) appear on the same". There was a request for further information of the Particulars of Claim which asked UMA to provide a copy of the briefing notes dated 6 October 2004. As far as I can tell, the response to this request did not provide a copy of the briefing note requested, but merely said "The briefing referred to is that of 8 October 2004". There, it seems, the matter rested.

It was not until I raised a query, on the first day of the trial, about the origin of the extracts that were pleaded at paragraph 6(b) of the Particulars of Claim, that it became clear that those acting for UMA had assumed, wrongly as it turned out, that the "Safe System of Work" document to which I have already referred formed part of the note of the tool box talk that took place on 8 October 2004. Mr Walker then applied to amend the Particulars of Claim to plead the "Safe System of Work" document, an application that I granted. However, Mr Hornett quite properly raised a marker to the effect that he had not had the opportunity to take proper instructions in relation to the amendment and he reserved his position.

Having had an opportunity to consider his client’s position, Mr Hornett then submitted that the "Safe System of Work" document did not apply to the work that was being carried out on the A side primary screen over the weekend 2/3 February 2008. He submitted that it was clear from the title that this document was intended to be a self contained procedure applicable to major work of rebuilding parts of the plant, rather than to routine work of repair. He submitted, in my judgment correctly, that the work carried out over the weekend 2/3 February 2008 fell into the latter category. In support of his submissions he referred to the various parts of the Work Procedure that referred to cutting up steelwork into manageable pieces, using lifting gear and the installation of new steelwork, and so on, and submitted that these were clearly directed at major rebuilding work rather than routine repair.

Mr Walker submitted that the document, with its reference in the title to "general fractured or worn parts on Primary . . . screens", was applicable to the work being carried out by the Defendants on 2/3 February 2008 because that work consisted of repairs to fractured parts on a primary screen.

In my judgment, Mr Hornett's submissions on this point are to be preferred for a number of reasons. First, as a matter of language, "rebuilding" is different from "repair". Second, I note that the document, which is undated but was probably created in February 2005, is not in the same form as other parts of the May 2007 version of UMA's H&S Policies and Procedures Manual. This suggests that it was brought into existence to cover special situations, as opposed to routine operations. Third, there is no evidence that it was still in force in February 2008. Fourth, if it was intended at the time of its issue to cover routine hot work, as Mr Walker submits, then it must have been superseded by the provisions relating to hot work in the May 2007 version of UMA's H&S Policies and Procedures Manual. In the absence of any evidence from UMA to the contrary, I consider that this is the only inference that I can draw.

For these reasons, therefore, I consider that UMA's reliance on the "Safe System of Work" document is misplaced and that it has no application to the work that was being carried out by the Defendants over the weekend of 2/3 February 2008.

I should add that Mr Walker put the "Safe System of Work" document to some of the Defendants’ witnesses in cross-examination, including Mr Marshall, on the basis that it was a document with which they were bound to comply and they did not challenge this premise. However, given the circumstances in which the document came to be relied on at the trial, I do not consider that it would be either right or fair to treat the witnesses’ apparent acceptance of the applicability of the document as binding on the Defendants.

Having disposed of this point, I now turn back to the Hot Work Procedure document, 39 App 1. In his oral closing submissions, Mr Walker accepted, correctly in my view, that adequate protection of the lining from heat and sparks could be achieved with water. He therefore accepted that "etc" in paragraph 2.3 could include the protection of rubber linings by continuous spraying of the relevant area with running water during the hot work, as an alternative to the other precautions listed in paragraph (namely, "spreading non-flammable dust, fire blankets, covering with steel plates").

Accordingly, it followed also, as Mr Walker accepted in his oral closing submissions, that paragraph 2.4, with its reference to "extreme care and attention" did not apply. As he put it, this was because "an adequate amount of water would do" to satisfy the requirement imposed by paragraph 2.3. I consider that, in the light of the evidence, Mr Walker was right to make these concessions.

Mr Walker’s submissions eventually amounted to this: if there was no proper fire watch or if the water was not used so as to provide adequate protection to the lining, then the claim must succeed. As to the latter, he submitted that it should have been possible to use the water in such a way as to provide adequate protection and so, in effect, the occurrence of the fire shows that the water was not properly applied.

In relation to the fire watch, Mr Walker submits that it has not been shown by the Defendants either that Mr Percival kept a proper fire watch whilst Mr Smith was cutting the bolts or that Mr Smith kept watch whilst Mr Percival was doing the vee-ing out with the angle grinder. This submission raises questions of causation, which I discuss below. Further, I did not understand Mr Walker to criticise the fire checks that were carried out once the hot work was completed, but in any event I need to return to this aspect below.

The Risk Assessment describes, in the relevant box, the task or work operation as "Secondary Screen A and B side repair". To the right of the box have been added the words "A Side Primary", with what appears to be Mr Smith's initials beside them. This suggests that these words were added later, a conclusion reinforced by the fact that the date of assessment is stated to be 28 January 2006. Against the heading “Risk Assessment” is written "GM 11", which indicates that it was probably a template document used whenever repairs were to be carried out to the A and B side secondary screens. I did not understand there to be any real dispute about this.

The document was initialled at the bottom, again it seems by Mr Smith, and dated 1 February 2008. From this I conclude that Mr Smith took the standard GM 11 Risk Assessment and modified it slightly by including in its scope work to the A side primary screen. He did this on 1 February 2008. There is no evidence that the original document GM 11 was modified in any other way at this time.

At this stage it is worth noting that if, as Mr Smith said, the Semperit lining material was introduced in 2006, then it is likely that this risk assessment was carried out at a time when the previous lining material was in use. Accordingly, it seems that the risk assessment may not have been revisited after the new lining material was introduced. Whilst one explanation for this could be oversight, it seems to me to be much more likely that the result of the inquiries that the Defendants say were made when the new material was introduced was that no re-assessment was necessary because the new material was thought to be no more combustible (and possibly less so) than the material it was replacing.

I accept the evidence of both Mr Smith and Mr Marshall that this was what the Defendants were told by UMA. In my judgment it does not lie in UMA’s mouth to complain if that information was incorrect: the Defendants were entitled to rely on what they were told unless there was some reason to doubt it. There is no evidence of any such reason here.

Reverting to the sequence of events, on the same day, 1 February 2008, Mr Smith completed the safe working practice document (which has been referred to in evidence as the method statement). In the box beside the side heading “Task or work operation” was written "Repair tubes on A side primary screen". Under the heading "significant hazards" was written "See Risk Assessment. GM11. Vol 1 ". The method statement did not mention fire blankets; the only precautions identified being described as follows:

The following morning Mr Curtis issued the Permit to Work for the A side primary and secondary screens. The relevant work was described as “A side primary screen repairs to tubes etc”. The relevant control measures set out in that document were in the following terms:

“fire watch and hoses to hand 2 people at all times (see risk ass)”

Apart from the reference back to the risk assessment, there was no reference to welding blankets.

In my judgment these documents must be considered in their context. The risk assessment, as its name suggests, is intended to identify the risks and hazards associated with particular pieces of work. In this case the risk assessment was a generic document which was intended to encompass routine repairs to the A and B side secondary screens. It was not intended to cover one specific piece of work. It was then extended by Mr Smith to cover the work to the A side primary screen, being work of a similar nature, that was to be carried out over the weekend of 2/3 February 2008. However, so far as one can tell, Mr Smith did not make any other changes to the document. By contrast, the method statement that he prepared is a document that was task specific. In my judgment, the method statement should reflect the contents of the risk assessment: in other words measures to deal with the risks and hazards identified in the risk assessment, so far as applicable to this particular job, should be set out in the method statement. If they are not, the person who is to issue the Permit to Work should either refuse to issue a permit or make good any deficiencies in the method statement, if practicable to do so, by inserting any necessary additional precautions in the box on the permit entitled Control Measures.

In these circumstances, it seems to me that the primary sources of the precautions that are required for any particular task are the method statement and the Permit to Work. That is particularly so where the risk assessment was a generic document and not confined to a particular operation. In this case Mr Curtis appears to have resolved any ambiguity in the method statement about whether or not the "constant fire watch" had to be carried out by a second person by stipulating in the Permit to Work that two people were to be present at all times during the work.

When one considers these documents, together with UMA’s Hot Work Procedure and what actually happened on site prior to the weekend of 2/3 February 2008, the evidence as to how the work was to be carried out is, in my judgment, all one way. It was clear from his oral evidence that Mr Farla, like his predecessor no doubt, knew exactly how the Defendants were intending to carry out the work to the A side primary screen. In particular, he knew that the Defendants would not be using welding blankets to protect the lining of the underpan, because that would be done by protecting it with a constant supply of water. So, apart from the skill of the operator of the cutting torch, the precautions to prevent fire that were known to or were expected by him consisted of:

To the extent that these precautions were more limited than those required by UMA's Hot Work Procedure, which in the ordinary course of events would be contractually binding on the Defendants, I consider that strict compliance with the Hot Work Procedure was waived by UMA. I do not reach that conclusion from the issue of the Permit to Work alone, because I do not regard Mr Curtis (the person who issued it) as a person who can be regarded for these purposes as a directing mind of UMA. However, the position is different in the case of Mr Farla, who was the Wharf Manager. His position in the organisation was such that I consider that if he gave his informed consent to a particular set of precautions that did not comply strictly with UMA's Hot Work Procedure then his doing so constituted a waiver by UMA of strict compliance with that procedure.

This conclusion is consistent with the concession made by Mr Walker, in his written closing submissions, that the agreement between Mr Marshall and Mr Farla that it was not necessary to de-mat the B side screen amounted to a variation of the contract. Whether it is properly characterised as a variation of the contract or a waiver of the requirement to comply with a particular contractual obligation probably does not matter in the context of this case.

As it happens, with one possible exception, I cannot detect any material difference between the precautions that the Defendants were proposing to adopt and the precautions that were required by UMA's Hot Work Procedure as applied to this situation. The possible exception is that paragraph 5.3, which requires a constant check of the area to be made for any signs of fire, does not say that this has to be done by a second dedicated fire watcher. It may be that it could be complied with if the person carrying out the work paused every minute or so and carried out a visual check of the area. However, that question does not arise in this case because, as I have already mentioned, the Permit to Work required a second person to be present.

For the sake of completeness I should mention that the experts agreed that there was "no physical, forensic evidence that the UMA hot work procedures were inadequate".

In addition, I consider that whilst carrying out the hot work and implementing the relevant precautions the Defendants owed UMA a duty to carry out the work with reasonable skill and care, that is to say a degree of skill and care commensurate with the risks involved in carrying out hot cutting work. In the circumstances of this case, I consider that that would also involve hosing down the working area with water after the conclusion of the hot work (which was not a requirement that, in my view, was imposed by the Hot Work Procedure). I find that this precaution was taken (see paragraph 257 below).

This provided as follows:

“It is a condition precedent to liability that in respect of the use away from the Assured’s own premises of an electric, oxy acetylene or similar welding or cutting equipment like blow lamps or blow torches, hot air guns or even hot air strippers and asphalt bitumen tar or pitch heaters, the following precautions will be complied at every occasion:

1

Adequate and suitable portable fire extinguishers to British Standards 5423:1987 or its predecessors or successors in full working order will be kept at each area of work and used immediately if smoke or smoulder or flames are detected.

2

The area in the immediate vicinity of the work shall be cleared of all movable combustible material. The combustible materials which cannot be moved and (sic) must be covered and protected by over-lapping sheets or screens or non-combustible material.

3

A fire safety check of the working area to check for smoke, smouldering or flames (including spaces behind the walls and screens or partitions and above false ceilings) shall be made at regular intervals during work and between 30 and 60 minutes after completion of each period of work and immediate steps taken to extinguish any smouldering or flames when discovered.

. . .”

It is not disputed that this warranty is a condition precedent to liability, as it states, or that it must be exactly complied with. Further, it is a warranty that precautions will actually be taken, not a warranty that arrangements will be made for precautions to be taken. It is also accepted on all sides that it is not necessary for the insurers to prove that a breach of warranty caused the loss: it is settled law that a breach by the insured of a warranty such as this discharges the insurer from any further liability under the policy from the time of the breach.

On the day that the fire took place UMA set up an internal investigation under the direction of Mr Macdonald. Between about 5 and 7 pm that evening two of the defendant’s employees, Mr Phil Smith and Mr Don Percival, were asked to attend his office and to make statements about what had happened. From internal notes made at the time it appears that Mr Marshall asked if he could be in attendance when the statements were taken and that this request was refused. I have already mentioned Mr Smith's evidence that he did not want to make a statement that evening because he was still feeling very shaken by the fire. He said that Mr Macdonald refused his request and insisted that he make the statement then and there.

When reading the papers after the case had been opened, which by then included a supplementary witness statement prepared by Mr Marshall that was given to me on the first day of the trial, I became concerned about the course which this internal investigation had taken. Partly prompted by some submissions made by Mr Hornett, and partly because of my concerns about this investigation, I directed disclosure of the notes prepared by UMA’s and the insurers’ fire experts that had been made during their visits to the site. These notes proved to be very informative.

Attached to Mr Marshall's supplementary witness statement were some photographs taken after the fire showing the cutting torch that had been used by Mr Smith. In one photograph it is shown in the position where Mr Smith says that he left it shortly before he left the area of the fire, and in another it is shown partly hooked over the side of the primary A screen in the screening house - which is where UMA say the fire started. Mr Marshall said that on the day after the fire Mr Macdonald told him that he wanted to take some photographs of the relevant area in the screening house.

Mr Marshall says that Mr Macdonald took a series of photographs of the area around the primary A screen and then started to move things around, including the cutting torch, which Mr Marshall says was taken and hooked over the primary screen.

Miss Russell said that she first saw the cutting torch hanging over the screen as shown in one of the photographs appended to Mr Marshall's supplementary witness statement, which she herself had taken. However, the cutting torch had become separated from its hoses, which had burnt in the fire, and Miss Russell said that the remains of the hoses for the cutting torch were coiled up on the walkway immediately below the place where Mr Smith said he had left the cutting torch. That is the place shown in the second photograph attached to Mr Marshall's supplementary witness statement, a photograph also taken by Miss Russell. Miss Russell said that she took the torch from the place where she first found it and put it in the place where Mr Smith said that he had left it and took a photograph of it. She then put it back where she first found it. She said that her reason for doing this was that Mr Smith had told her that various items had been moved and that it was clear to her that the position of the cutting torch immediately before the fire was not where she first found it, but where Mr Smith said he left it.

The importance of this possibly innocuous sequence of events becomes apparent when one reads the report of the investigation that was prepared by Mr Macdonald the following month. Paragraph 4.0.10 of the report said "Propane/oxygen cutting torch hung over inboard screen side oxygen valve closed and propane ¼ turn open". Paragraph 4.0.13 of the report reads as follows:

“Fuel

The ¼ turn open propane valve on the cutting torch is a potential source of fuel. The cutting was completed at 0900 hrs and the torch hung over the screen side. The bottle was not isolated until 09:25 hours (Verbal confirmation with P Smith) This would allow 25 minutes of gas to leak. Propane is a heavy gas that will gravitate to the bottom of the sump. The chutes will allow air flow which could supply the necessary mixture of oxygen to propane that will support combustion. The cutting torch might have been turned off on completion of works but the valve disturbed when the torch was hung over the screen side. This was discussed with the fire Station Officer reference noise (Explosion) I was advised that if the propane ignited in the open sump it would not cause an explosion as in a strictly vented container."

Under the heading "Immediate causation" two options were presented as the possible cause of the fire. The first was that procedures were not followed and a hot bolt/spark ignited the rubber lining. The second was in the following terms:

“Contractor completed the cutting of the final four nuts. The gas torch propane/oxygen was turned off and hung over the screen down side knocking the propane open. After checking the secondary screen works and washing the area down with fresh water he then turned the propane off and left the screen house. The mixture of propane and oxygen ignited due to the 110 volt splitter and created an immediate fire.”

I accept the evidence of Mr Marshall about what Mr Macdonald did in relation to moving some of the equipment; it is also supported by the evidence of Miss Russell. It was not challenged - unsurprisingly, perhaps, since I was told that Mr Macdonald was no longer employed by UMA. However, in the absence of any explanation from Mr Macdonald the only inference that can be drawn from these facts is that Mr Macdonald moved the cutting torch as I have described and then relied on its position to found the basis for this alternative theory as to how the fire might have started. In other words, he manipulated the evidence in order to create a possible explanation for the start of the fire that did not in any way implicate UMA.

Unfortunately, it appears that this may not have been the only example of manipulation of evidence by Mr Macdonald. I have already referred to the photograph of a shovel waged in the gap between the screen and the upper edge of the underpan below it and Mr Heath's evidence that he had understood that this was the position in which the shovel had been when Mr Smith was cutting the bolts above it.

In the bundle of photographs there is a photograph (at C/26) taken by Mr Macdonald on the day after the fire, 4 February 2008, which shows the gap between the screen and the underpan in which the shovel was said to have been placed. The shovel can be seen standing on the floor leaning up against the side of the underpan (which is consistent with Mr Smith’s evidence that he left it on the floor). The shovel is shown in the same position in photographs (at C/55-6) taken by Miss Russell on the same day. However, a photograph (at C/81) taken by Mr Macdonald of the same area on the following day, 5 February 2008, shows the shovel wedged in the gap at precisely the same place as is shown in the photograph (at A/296) appended to Mr Heath's report (the precise place can be identified with reasonable confidence by, amongst other things, a distinctive pattern of marks on the side of the underpan at this point). As I have already mentioned, Mr Heath saw the shovel in this position when he inspected the site on 6 February 2008, and knowing that Mr Smith had used a shovel to catch the heads of the cut bolts, assumed that this was the position of the shovel when Mr Smith was carrying out the cutting. Again, in the absence of any other explanation, this suggests that the shovel was deliberately left in this position, either by Mr Macdonald himself or to his knowledge, in order to lead others to assume that that is how the shovel had been used during the work. So far as Mr Heath is concerned, that is exactly what happened.

This is deeply unsatisfactory. In these circumstances it seems to me that the only proper approach that the court can adopt in relation to the evidence as a whole is to view with particular scepticism any evidence that may have been influenced wholly or in part by anything said or done by Mr Macdonald. The possibility also has to be borne in mind that there may be evidence relevant to the cause of this fire, or more importantly the responsibility for it, which has never emerged because it was suppressed or glossed over in the course of the post-fire investigation. I must make it clear that I am not making any finding to this effect against UMA, and certainly not against any of the independent experts, rather I am simply noting that where the evidence on a particular topic is scanty and that evidence might well have been under the control of UMA, the court needs to proceed with caution given that some manipulation of evidence has clearly taken place.

A striking example of why such scepticism or caution is required is the evidence given by Miss Russell and Mr Heath as to their understanding of how the work was carried out. Miss Russell went to the site on the afternoon following the fire and then again on 6 February 2008, when she met Mr Heath who was visiting the site for the first time. Before Mr Heath went round the site with Miss Russell he had a conversation with Mr Marshall and Mr Smith as I have already mentioned. Mr Heath did not speak to Mr Macdonald.

By contrast, when Miss Russell visited the site on 4 February 2008 she was met by Mr Macdonald who told her what had happened in some detail, as her notes demonstrate. He told her that during the works to the primary screen and underpan one man had been inside the underpan itself. In fairness to Mr Macdonald, he had some basis for saying this because of the reference in Mr Smith's hand written statement of 3 February 2008 to Mr Percival going “inside screen & underpanto clean pipe-work ready for weld repairs".

I have already quoted from Mr Smith's witness statement that was prepared for these proceedings, in which he said that Mr Percival "went inside the primary screen and used a 4½ inch angle grinder and wire brush to prepare the central rail within the primary screen for welding work later that day".

In her report prepared in January 2012, at paragraph 3.13, Miss Russell said this:

“Mr Percival entered the underpan of primary screen A at approximately 07:30 hours carrying two hand-held grinders, one fitted with a wire brush and the other with a grinding disc. He started to clean areas of the tubes in preparation for welding [PS2 paragraph 13]."

The reference quoted, PS2 paragraph 13, is to paragraph 13 of Mr Smith's witness statement in these proceedings which contains the passage that I have quoted above. It refers to the screen, and not to the underpan.

At paragraph 3.18 of the same report Miss Russell said:

“Mr Smith sprayed some water into primary screen A underpan, being careful not to wet Mr Percival, who was still inside the underpan, and then he sprayed water into secondary screen B, where Mr Shaw had last been undertaking hot work."

In her notes of her conversation with Mr Smith on 6 February 2008 Miss Russell recorded the following:

“One person inside cleaning up welds with grinder.

- person inside fine

- hose was running over surface into underpan"

When asked about this in evidence Miss Russell was a little puzzled about the reference to "person inside fine" but said that she remembered being concerned as to how a person could work inside the underpan whilst it was at the same time being sprayed with water. She surmised that she must have raised this with Mr Smith and been told that it was not a problem. Her evidence about this was to some extent consistent with the passage from her January 2012 report that I have quoted above and with what she said on page 6 of her first report, which was as follows:

“Mr Percival then entered the underpan of primary screen A carrying two hand-held grinders, one fitted with a wire brush and the other with a grinding disc. He started to clean areas of the tubes in preparation for welding. Mr Smith started to remove 20 bolts, 12 on the drive side and 8 on the walkway side. I understand that he placed a shovel underneath each bolt and used an oxygen-propane cutting flame to remove the external portion of the bolts, catching the hot metal in the shovel. Mr Smith started on the driveside side of the primary screen, where he cut and replaced eight bolts. He then moved to the walkway side, where he cut and replaced four bolts. After removing the outer portions of the bolts he checked that the remaining lengths were cool before pushing them through to the underpan. Mr Smith then cut the outside sections from four further bolts but, since he did not have enough replacement bolts he did not remove the remaining lengths at that time. Mr Smith informed that he had a running water hose with which he played water over the rubber lining of the underpan in the area that he was cutting.

. . . Mr Smith sprayed some water into primary screen A, being careful not to wet Mr Percival, who was still inside the underpan, and then he sprayed water into a secondary screen A, where Mr Shaw had been last undertaking hot work. I understand that at approximately 09:15 hours Mr Percival completed the grinding work inside the underpan and then positioned a scaffolding board inside.”

Miss Russell's notes included a helpful diagram showing where the 20 bolts were that were cut by Mr Smith on the morning of 3 February 2008. 8 were half way along the length of the screen on the east side, and 8 were opposite these on the west side. A further 4 bolts were at point near the north west corner of the screen and must have been, I think, bolts securing the rosta block, which formed part of the suspension unit for the screen. These are the 4 bolts to which Mr Smith refers in his witness statement as being above a steel shelf. It is not clear whether or not they also secured one of the tubes.

I showed Miss Russell's diagram to Mr Smith at the conclusion of his evidence and he said that the height of the tubes above the base of the underpan that were secured by the 8 bolts that he had cut on each side of the screen was about 3 m or 10 ft. It is quite clear to me, and by the conclusion of the evidence I do not think that anyone was in a position to dispute this, that no-one could have carried out any grinding work on the flanges of these tubes whilst standing in the underpan, unless he had some form of platform (of which there is no evidence). So far as the 4 bolts at the rosta block are concerned, Mr Smith explained that at this point the sloping base of the underpan was very close to the tubes. It is reasonably clear that no one could use an angle grinder on those tubes from the underpan because there would have been no room in which to do so. In addition to these factors, Mr Smith said in evidence that all the cracks to the flanges of the tubes were at the top, not at the bottom. He said that the cracks propagated from the holes for the bolts, which is why it was essential to remove the bolts before the vee-ing out of the cracks using the angle grinder could take place.

In the light of these considerations, there is no reason why Mr Percival, or anyone else, should have been in the underpan of primary screen A at any stage during the morning of 3 February 2008. In my judgment, the only place from which Mr Percival could have used and did use the angle grinder to vee-out the cracks in the flanges was from the screen itself, either by lying on the tubes - particularly if the cracks were to the flange of a lower tube - or by sitting on the upper layer of tubes with his legs through the screen. The latter would probably have been easier with a board to sit on.

Quite apart from these considerations, it is inconceivable that Mr Percival would have been using an electric angle grinder which was plugged into a splitter box that was exposed to water. I would reject emphatically a suggestion that any electrical equipment was in the underpan at any time during the work that morning. However, by the end of the trial I do not think that anyone was suggesting that this was the case.

At this point, it is instructive to consider the notes taken by Mr Heath of his conversations with Mr Smith on 6 February 2008. I can find nothing in these notes that contradicts the account that Mr Smith gave in evidence, although Mr Heath told the court that when he first met Mr Marshall and Mr Smith he was shown a copy of the handwritten statement that Mr Smith had made three days before with its reference to Mr Percival being in the “screen & underpan”. It seems likely that because Mr Heath, unlike Miss Russell, had not discussed the incident with Mr Macdonald, he did not have the impression at any time that Mr Percival might have been carrying out work from within the underpan.

The only aspect in which Mr Heath's evidence conflicted with that of Mr Smith was in relation to the use of the shovel. But, as I have already explained, Mr Heath was unable to remember whether he derived that information by way of assumption from what he saw or from what he was told by Mr Smith. I am quite confident that it was the former. I have no hesitation in accepting Mr Smith's evidence that he held the shovel against the side of the screen just below the bolt that he was cutting.

For all these reasons I am quite satisfied that Miss Russell was simply mistaken when she thought that Mr Percival might have been carrying out work from within the underpan. I am also satisfied that her misunderstanding arose because she, unlike Mr Heath, had first discussed what had happened in some detail with Mr Macdonald who told her that one of the men had been working within the underpan. Whether or not that reflected Mr Macdonald’s genuine belief based on the contents of Mr Smith's hand written witness statement, or was something that it suited his purpose to emphasise, I cannot say. However, it is perhaps relevant to note that Mr Percival, who also made a statement on the day of the fire said that he was "inside screen with 2 x 4” grinders . . . veeing out cracks etc for welding". He said nothing about being in the underpan.

A further problem that bedevilled the whole investigation into the fire was the fact that immediately after the fire Mr Macdonald was saying that no hot work had taken place in the scalping house on the morning of the fire. Whether he just assumed this, or whether he accepted at face value what he had been told by Mr Clancy, probably does not matter. The significant thing, from the point of view of the investigation, was that, as a result, none of those who investigated the fire during the few days after its occurrence gave any serious consideration to and then followed up the possibility that the fire might have started in the scalping house, and not the screening house. This was almost certainly the reason why Miss Russell did not take a sample of the conveyor, CV6.

It is also unsatisfactory that Miss Russell was not given the opportunity to interview all the available witnesses, which is the usual practice when fire experts are called in to investigate a fire such as this. Had she been able to do so under suitable conditions, I suspect that some of the issues that have arisen in this case would not have seen the light of day.

As I have already said, I accept the evidence of Mr Smith that he started cutting the bolts at sometime shortly before 8 am and I find that he completed the hot cutting shortly before 8:15 am.

It is clear from the evidence that the first piece of hot work that was carried out in the screening house on the morning of Sunday, 3 February 2008, was the welding by Mr Shaw of the remaining 1 or 2 lugs to the A side secondary screen whilst Mr Garnham used the hose to spray the screen. I doubt if that work would have taken longer than about 10 minutes, with possibly another 10 minutes of setting up time and hosing down the area before the work started. Since Mr Shaw and Mr Garnham arrived in their place of work at about 7:30 am, I would have expected that welding to have finished shortly before 8 am, perhaps 5-10 minutes earlier.

During the period prior to starting the work I find that Mr Smith hosed down the entire area, as he described in paragraph 11 of his witness statement, and was then involved in setting up the gas bottles and cutting torch prior to carrying out his own work. I find that once Mr Shaw and Mr Garnham had finished with the hose, Mr Smith took it back and wedged it in the gap between the screen and the underpan at the southern end of the A side primary screen in readiness for his own work. I should explain that he was able to do this because, as he explained, there were rubber flaps at the end of the underpan between which the hose could be tightly wedged.

Up to this point it is not clear what there would have been for Mr Percival to do, since he could not start vee-ing out the cracks in the tube flanges until after Mr Smith had cut the nuts off the relevant bolts and they had cooled sufficiently for him to be able to remove the shanks of the bolts. This may explain why Mr Smith said in his witness statement, at paragraph 13, that Mr Percival went inside the primary screen to prepare the central rail using the angle grinder that was fitted with a wire brush. This reflected what was in the draft witness statement prepared by the loss adjuster shortly after the fire.

However, as I have already noted, in cross-examination Mr Smith said that cleaning the central rail was no part of Mr Percival’s job that morning. He thought that the reference to the central rail in paragraph 13 of his witness statement was a mistake, and that it should have referred to the tube flanges. He said also that Mr Percival did not start work straight away, but when he was giving this evidence he may have been referring to the task of vee-ing out the cracks rather than to any other work.

In his hand written witness statement, prepared on the day of fire, Mr Smith refers to Mr Percival going inside the screen "to clean pipework ready for weld repairs". This could be a reference to cleaning the central rail or a reference to preparing the flanges for welding.

If one looks at the inherent probabilities of the situation, it seems to me most unlikely that Mr Percival would have been allowed to stand around doing nothing for some 20-25 minutes until Mr Smith was ready to start the hot cutting work. It is far more likely that he had some other job to do, and in my view that job was wire brushing the central rail using the angle grinder in preparation for some work that was to be carried out on it later that day. I consider that Mr Smith's present recollection to the contrary is mistaken.

The issue that was really contested at the trial was whether Mr Percival continued doing that (or some other) work after Mr Smith had started to cut off the ends of the bolts. Mr Walker drew attention, correctly, to the fact that nowhere in any of the witness statements was there any specific reference to Mr Percival acting as a fire watch. In Mr Percival's handwritten statement he said that "I was vee-ing out cracks etc for welding while my colleague Phil Smith burned off bolts from outside, catching heads on a shovel to prevent damage". I consider that it is important to bear in mind that these statements were made in some haste at the end of what must have been a very traumatic day, and so one must guard against reading the words too literally. However, that said, Mr Percival's account is not a promising platform on which to base a case that he was carrying out a dedicated fire watch whilst Mr Smith was cutting the bolts.

Unfortunately, no other statement was ever taken from Mr Percival and he was not called as a witness. Mr Smith told the court that Mr Percival left the Defendants under a cloud, and Mr Marshall said that he had not been in touch with Mr Percival during the three or more years since he left the company. I have already mentioned Mr Marshall's unsatisfactory evidence when asked why Mr Percival had not been called as a witness.

Mr Walker has reminded me of the well-known dictum of Lord Diplock in Herrington v British Railways Board [1972] AC 877. He said, at page 930.

“The appellants, who are a public corporation, elected to call no witnesses, thus depriving the court of any positive evidence as to whether the condition of the fence and the adjacent terrain had been noticed by any particular servant of theirs or as to what he or any other of their servants either thought or did about it. This is a legitimate tactical move under our adversarial system of litigation. But a defendant who adopts it cannot complain if the court draws from the facts which have been disclosed all reasonable inferences as to what are the facts which the defendant has chosen to withhold."

This is not that case. Here the Defendants have called evidence. The point is that they have not called all the evidence that it is said they could have called. By letter dated 16 December 2011 the Defendants’ solicitors enclosed the statements of the witnesses whom they proposed to call, but these did not include a statement from Mr Percival. In relation to him, the solicitors said that he had been sent his statement and that they “hoped to receive his approval by close of business". In fact, no witness statement from Mr Percival was ever served. There is no evidence as to whether this was because Mr Percival simply failed to respond, perhaps because he had no wish to assist the Defendants, or whether he responded by indicating that he did not agree with the contents of the draft statement that he had been sent, or something else.

If matters had rested there, I would have difficulty in seeing how the court could properly draw any inference one way or the other. However, Mr Marshall's refusal to give any indication as to the reason why no witness statement from Mr Percival had been served lends some weight to Mr Walker's submission that the court should draw the inference that Mr Percival was never instructed to act as a dedicated fire watch.

The insurers have disclosed the preliminary report by Sigma, the loss adjusters they instructed to investigate the circumstances of the fire. Sigma’s Regional Manager, Mr Wookey, was the person who carried out the investigation. In his report dated 11 February 2008, he said that he had been assisted in his enquiries by Mr Marshall and Mr Smith. He said that he formed a favourable impression of the Defendants’ approach to health and safety and risk management matters.

In relation to the Burning and Welding Warranty, Mr Wookey wrote this:

“We have discussed this at great length with the Insured and believe the Insured have complied with Sections 1, 3 and 4. However, we do not believe that the Insured have complied with Section 2.”

In the light of this, it is perhaps a little surprising that in their defence the insurers alleged, for the first time in unambiguous terms, that the Defendants were in breach of paragraphs 1 and 3 of the warranty (although the latter had been raised tentatively in a letter dated 10 October 2008), as well as paragraph 2. However, what is clear is that Mr Wookey was reasonably satisfied at the time that the Defendants had carried out a fire check (what he called Section 3), both during and after the hot work, in accordance with the warranty.

I am troubled by the lack of any explanation for the absence of Mr Percival, but I do not consider that the circumstances justify my drawing the inference from this that Mr Walker asks me to draw because there is some evidence - from Mr Smith - that Mr Percival was looking out for signs of fire whilst he (Mr Smith) was carrying out the cutting, even though that is the only direct evidence that this happened. There is also the contemporaneous conclusion reached by Mr Wookey that there was no breach of the warranty in relation to the fire watch.

Nevertheless, I do not feel that I am in a position to make a positive finding on the basis of this evidence that Mr Percival was keeping a fire watch: Mr Smith's primary concern at the time would have been with his own work. Conversely, I find myself unable to conclude that, as a matter of probability, Mr Percival was not keeping a proper fire watch . The burden of proving that a fire watch was not kept in accordance with either UMA's Hot Work Procedure or the Burning and Welding Warranty, which is in some respects less stringent on this aspect, is on UMA and the insurers, respectively. I find that neither of them has discharged it. Each of them could have taken steps with a view to securing the attendance of Mr Percival as a witness, but as far as I am aware neither of them did so.

I am also unable to reach a conclusion as to precisely what work Mr Percival did carry out on that morning. I am satisfied that he did vee-out some of the cracks in the flanges using the angle grinder before he left the area to go to breakfast - because he said so in his handwritten statement - but I am unable to say whether he completed this work or simply stopped at a convenient moment before going to breakfast.

The experts were agreed that in terms of propensity to cause a fire the hot work activities can be rated in the following order of increasing severity: grinding, tack welding, seam welding and oxy-propane cutting. There is no evidence before the court that the sparks produced by angle grinding steelwork could set fire to either the Semperit rubber lining or the mastic that was used in the joints. Mr Heath said in his report that sparks from angle grinding have a temperature of 600-900°C, and that by the time that they have travelled about 1 m away from the grinder they have cooled sufficiently to have lost their orange glow.

The spatter produced by oxy-propane cutting consists of globules of molten steel with temperatures of the order of 1500°C. It is well known that they can travel distances of up to 10 m. If the technique being used is "washing", rather than "lancing", the globules will not travel so far - although Mr Heath said, and I accept, they may be larger.

Mr Heath and Miss Russell tested the flammability of a damaged section of lining, together with some mastic, that was removed from the B side underpan after the fire. When it was held vertically above a naked flame it ignited easily after a few seconds, and then continued to burn steadily with a yellow flame. There was a strong smell of burning rubber. The mastic sealant also ignited easily, and then burned almost without a flame but with a bright glowing area of combustion. The smell of rubber was less distinctive.

Ignition tests of this sort provide very little indication of the likelihood that the material under test could be ignited by, for example, sparks from angle grinding. I am very doubtful as to whether such sparks would have been capable of setting fire to either the rubber lining or the mastic joints in the underpan below the A side primary screen. Neither UMA nor the insurers have proved that sparks produced by an angle grinder cutting steel would be capable of setting fire to either the lining or the mastic. Accordingly, to the extent that either UMA or the insurers are still contending that sparks from angle grinding caused the fire, I reject that contention.

In case it matters, I find that the use of the angle grinder with the wire brush attachment did not amount to hot work within UMA’s Hot Work Procedure document or the insurers’ Burning and Welding Warranty. It was not “cutting equipment”.

There is also no direct evidence that molten globules of steel could ignite either the lining or the mastic. However, the experts have agreed that if the fire started in the screening house then it was most likely the result of the hot work being carried out by the Defendants. It is implicit in this conclusion that at least one of the types of hot work being carried out would have been capable of igniting either the lining or the mastic. Therefore it is a conclusion that hot cutting work, at least, would have been capable of igniting either the lining or the mastic and thereby starting the fire.

On this basis I am satisfied that UMA and the insurers have proved that the fire was probably caused by spatter from oxy-propane cutting igniting either the lining or the mastic in the underpan below the A side primary screen. Indeed, I did not understand the Defendants to dissent from the conclusion that spatter from oxy-propane cutting could have started a fire.

It must follow from these considerations that, so far as UMA's claim is concerned, any failure to carry out a fire watch whilst Mr Percival was using the angle grinder in the A side primary screen is irrelevant as a matter of causation in the light of my finding that it has not been proved that the use of an angle grinder was capable of causing this fire. However, so far as the Defendants’ claim against the insurers is concerned, the position may be different: this is because a breach of warranty does not have to be causative of the loss.

As I have indicated, I am satisfied that whilst Mr Smith was carrying out the hot cutting water was being directed continuously into the underpan towards the area where he was working. However, I am unable to decide whether the hose was directed at the central divider in the underpan so that a spray of water was redirected back to the area where he was working, or whether the hose was pointed directly to the side of the underpan where he was working. Immediately after the fire he appears to have been saying the former, whereas in evidence he said the latter. I suspect that on this point Mr Smith's memory has failed him. It may well be that it was his practice to employ one or other technique according to the precise area where the work was being carried out. I see no reason why directing a high-pressure hose at the central divider so as to create a spray of water, as opposed to a straight jet, might not have been a more effective technique in some situations. Since no-one has carried out any tests on this, it remains an area of uncertainty

However, having watched and heard Mr Smith being cross-examined for several hours I formed the clear impression that he is a conscientious and highly competent craftsman. I have no doubt that he would have deployed the hose in the most effective manner for the work that he was doing at the time. There was absolutely no point in doing otherwise when the protection afforded by the water was of such importance. I did not get the impression that Mr Smith was a man who was willing to cut corners, or that Mr Marshall would have retained his services for so long if that was the case. Whilst I have expressed reservations about some of the evidence given by Mr Marshall, I do not have any reservations about reaching the conclusion that he is a highly competent and safety conscious operator who runs, as he said, a tight ship. There was no challenge to his assertion that the Defendants had an unblemished safety record for the 20 years prior to the fire. That sort of record is not likely to have been achieved by accident, and of course Mr Smith has been working for the Defendants throughout that period.

In relation to Mr Smith's activities on the morning of the fire I make the following findings of fact. First, Mr Smith hosed down the whole working area around the A side primary and secondary screens before any hot work started in the screening house. Second, whilst he was carrying out his cutting work water from the hose was being continuously directed into the side of the underpan on which he was working. Third, he hosed down the primary screen underpan after the hot work was completed and before they left the area for breakfast.

In addition, I am quite satisfied that both Mr Smith and Mr Percival checked the area carefully, both visually and by using their sense of smell, before they left. The last check would have been about 1¼ hours after Mr Smith had completed the cutting.

In my judgment, the only candidate for the cause of the fire is ignition of the lining or the mastic in the underpan by a globule of molten steel. I accept the evidence of both Miss Russell and Mr Heath that the other potential causes of a fire in the screening house identified in their reports, such as the explosion of a light, a defect in one of the items of electrical equipment, faulty wiring, smoking, and so on, can be excluded with reasonable confidence. No one has seriously suggested otherwise. The three experts have agreed that hot work was the most likely cause of this fire (whenever it started).

There are only two ways in which spatter from the cutting could get into the underpan. The most obvious is through the gap between the screen and the underpan, but another possibility pointed out by Miss Russell is that spatter could go over the top of the screen and then drop down between the tubes into one or other side of the underpan. Whichever way it went, it is unlikely to have travelled as much as 10 m before coming to rest within the underpan and so it would still be extremely hot.

Spatter passing through the gap, either directly or by way of ricochet from the shovel, would be most likely to land in the nearer half of the underpan. However, it is not possible to tell from the photographs, and Miss Russell was not able to express a view with any confidence, whether or not spatter passing through the gap could travel over the central divider and land in the other half of the underpan which was not being sprayed with water. On the evidence as it stands, it is not possible to exclude this as a possibility.

As to the manner in which spatter might have entered the body of the underpan during the hot cutting and what happened when and after it arrived, I can only conceive of eight possible scenarios (assuming in each case that the underpan is properly hosed down at the end of the work):

As I have indicated, each of these scenarios assumes that the underpan is properly hosed down at the end of the work; however, if it was not hosed down properly any one of the scenarios could have caused the fire. But I have already found that Mr Smith did hose down the underpan properly after the hot work had been completed and that he and Mr Percival carried out a careful check of the area.

In their Memorandum of Agreement the experts agreed that:

“3.22

We agree that a competent fire watch would have noticed a flaming fire in the underpan and, on that basis, we consider that there was no flaming fire within the primary A underpan whilst the fire watch was present. Therefore, have a fire to have been related to the hot work then it must have started as a smouldering fire.

3.23

We agree that the tests revealed that the mastic-type sealant between the sheets of rubber lining had a propensity to smoulder and produced very little outward sign of the smoulder in relation to smoke or odour.

3.24

We agree that it is possible that the smoulder continued to develop underneath the surface of the combustible material such that it could not have been reached by the quenching water. Alternatively, the area in which smouldering might have occurred was not quenched. The smoulder could then have progressed in longer than the fire watch and subsequently undergone transition to flaming.

3.26

We agree that this could explain why a fire initiated by hot work at the underpan might not have been extinguished or discovered by the steps taken following the hot work."

The point made in the first sentence of paragraph 3.24 above is relevant to my scenario (7). For the reason that I have given I consider that this situation could arise even if the underpan was hosed down with reasonable care after the work was completed.

In my judgment what went wrong on this occasion was that no one appreciated that a globule of molten metal striking the band of mastic at the top edge of the lining could penetrate to an extent that could cause a smouldering fire at a point within the lining that would not be reached by the quenching water. To meet this rare situation that might arise after any hot cutting work (particularly of the type being carried out on the day of the fire), a fire watch is required for a period that was much longer than 30 minute minimum period prescribed by UMA's Hot Work Procedures. I have already found that Mr Smith remained in the area for about 1¼ hours after he completed the cutting work; in addition, I find that he remained there for more than 30 minutes after the bolts that he had cut had cooled.

I therefore find that the defect in the procedures adopted on this occasion was that a fire watch for one hour after the hot cutting work had been completed (the period usually specified - see paragraph 277 below) was insufficient to detect the transition from a hidden smouldering fire into a flaming fire if a hot particle of spatter acted in the manner described in my scenario (7). If someone had remained on fire watch in the area for, say, two hours following completion of the work, this transition would have been seen very shortly after it occurred and the fire would probably have been extinguished using a portable fire extinguisher.

Mr Walker submits that, since the experts have agreed that, if the fire started in the screening house as a result of the hot work being carried out by the Defendants, "either the procedures were inadequate or they were not followed", the Defendants must have been negligent. In my view this does not necessarily follow. But in any event, as finally developed (in the manner described at paragraphs 183-186 above), the submission confined the case to carelessness in the use of the water or in the nature of the fire watch.

In one sense, the first limb of this agreement reached by the experts could be regarded as self-evident. If the procedures were adequate then, by definition, they would prevent a fire. A statement that the procedures were inadequate does not compel the court to find that they were not carried out with reasonable care. The answer to that question must depend on the reasons why the procedures were inadequate.

Mr Walker submits also that this is a case of res ipsa loquitur - so that the onus is on the Defendants to show that the fire occurred without fault on their part. There is some force in this submission. The evidence suggests that the same method of working had been used without incident for almost 20 years, with the result that it might be reasonable to infer that the outbreak of fire on this occasion must have had something to do with the manner in which the system of work was implemented, rather than with the system itself.

However, in my judgment the position is not so clear cut. A factor that was present on this occasion was the presence of the different lining material that had been introduced in 2006. I assume that the new material was chosen by UMA: at any rate, no one has suggested otherwise. It is not clear from the evidence how often repairs of the type that were being carried out to the A side primary screen on the day of the fire had been carried out between the change of lining material in 2006 and the fire. The impression I have from the evidence as a whole is that it was probably not very often. There is the possibility that the new material is more combustible than the material that it replaced, so that procedures which appeared to have been tried and tested may no longer have been appropriate.

On the evidence before the court I am in no position to reach any conclusion as to whether or not the present lining material, Semperit, is more or less combustible than the material that it replaced. All that can be said is that the possibility that it is more combustible (as Mr Farla said) cannot be excluded, which goes some way to undermining any reliance on the doctrine of res ipsa loquitur.

The repairs that were being carried out to the A side primary screen on the day of the fire were, according to the evidence, repairs of a type that had been carried out with the same precautions many times before without incident. In these circumstances, it seems legitimate to conclude that, prima facie, these precautions - as taken by the Defendants on numerous previous occasions, and known to and approved by UMA - were reasonable.

It is of course possible that a globule of molten steel has never previously acted in the manner postulated in my scenario (7), so that what has happened on this occasion has simply disclosed a defect in the precautions of which everyone was unaware. Alternatively, it is possible that this fire occurred because the new lining material, Semperit, is more combustible than the material that it replaced. However, I do not consider that the existence of either possibility necessarily means that the Defendants were in breach of the duty of care that they owed to UMA. The duty owed to UMA was, as Mr Walker effectively accepted - consistently with the position taken in his oral closing submissions - to use reasonable care and to take certain specified precautions. I consider that what constitutes reasonable care has to be that which is reasonable as between the Defendants and UMA in the context of this particular operation: the concept of reasonable care is not some absolute standard that exists independently of its context. Conduct that is reasonable in one context may be unreasonable in another. It is in my judgment relevant that UMA, a knowledgeable client that knew its own plant and demanded high standards of safety, imposed its own hot work procedures and knew of and was satisfied with the precautions that the Defendants had been taking (and proposed to take on this occasion) when carrying out these repairs.

I have to confess that I do not find this question of breach of duty to be straightforward. It is too easy simply to say that the hot work being carried out by X caused a fire, therefore X cannot have been carrying out the work with reasonable care. At least two of the experts agreed that, whilst the risk of causing a fire when carrying out hot work can be reduced to a very low level, very often it can never be eliminated. With this I do not disagree.

With the knowledge and experience that was available to the Defendants at the time (and to UMA also) I consider that it was not reasonably foreseeable that a fire watch lasting for one hour after completion of hot cutting of the type carried out by Mr Smith on 3 February 2008 would be insufficient to detect a smouldering fire. Although UMA's Hot Work Procedures made provision for a fire watch for a period of up to 4 hours, depending on the risk, it seems that no-one appreciated that the work being carried out by Mr Smith involved a risk that called for an extended fire watch period of more than an hour. As a result, no special provision was made in either the method statement or the Permit to Work for a fire watch in excess of 30 minutes, or even 1 hour.

It is alleged in the Particulars of Claim that the Defendants should have left one man behind in the area of the A side primary screen to watch for any sign of fire whilst the others were away having breakfast. However, this allegation was not supported in those terms by the evidence of Miss Russell. In her report Miss Russell referred to the recommendation by the Fire Protection Agency, in a document known as RC7, that a fire watch should continue for at least 30 minutes after the hot work is completed, with further checks at regular intervals up to 60 minutes after completion. This is the recommendation that was effectively incorporated into the warranty. Her conclusion was that the Defendants seemed to have complied with this recommendation. Mr Heath expressed no view on this. On the basis of my finding that Mr Smith remained in the area for about 1¼ hours after he completed the cutting, Miss Russell's conclusion must be correct.

Although in his written closing submissions Mr Walker said that the defendant should have appointed a dedicated firewater "while each man carried out his hot works, and for a suitable period after the completion", he did not make any submissions as to what a suitable period would have been. In the light of Miss Russell's evidence, it seems to me that he could not have plausibly have contended for a period of longer than 1 hour. In these circumstances, I find that the Defendants were not in breach of contract or of any duty of care owed to UMA by failing to leave one man in the area to act as a fire watch for a period that was in excess of 1 hour.

In his closing oral submissions Mr Philip Shepherd QC, who appeared for the insurers, suggested that a probable mechanism for the start of the fire was a hot article of molten steel landing on the strip of mastic just below the top of the underpan: in effect, my scenario (7). He suggested that this potential mode of ignition could have been easily prevented by simply draping a welding blanket over the steel strip so that it hung down the inside of the underpan and protected the top edge of the rubber or mastic.

So far as I am aware, this was the first time that this particular precaution had been suggested. The precaution suggested by Mr Heath consisted of blocking the gap using a welding blanket clipped to the frame of the screen (a suggestion not adopted by Miss Russell). He did not suggest that this strip of mastic was so particularly vulnerable that it called for some alternative precaution of the type suggested by Mr Shepherd.

Whilst I consider that if a welding blanket had been draped over the top of the side of the underpan in this manner it would probably have prevented this fire, this was not a precaution that anyone had thought to take. However, since this particular precaution was not mentioned in any of the expert reports and was not raised with any witness, I consider that it would not be fair to the Defendants to entertain it at this stage. For example, I can see that there might have been difficulties in keeping such a blanket in place whilst the hose was being directed on to the very area where it would have been hanging, and there was also evidence that blankets can create an obstruction or tripping hazard. This is just the type of issue that should be explored in evidence.

In any event, it is clear from Mr Farla’s evidence that he did not regard the lining of the underpan as something that could (or should) be protected by the use of welding blankets. He knew, and I have already found that his knowledge is to be treated as that of UMA, that the Defendants would not be using welding blankets during this particular operation. His knowledge and approval of the methods that the Defendants were to employ would, in my judgment, have been sufficient to preclude UMA from relying on this allegation even if it had been raised at an appropriate time.

At the end of the day my conclusion on breach of duty must turn largely on my impression of Mr Smith. As I have already indicated, I formed the view that he was a craftsman of integrity. I can see no reason why he should have exercised any less care on this occasion than he did on any other. Everyone, including him, accepted that there was a risk that spatter from oxy-propane cutting could get into the underpan. I find that he exercised reasonable care and skill when carrying out the activities described at paragraphs 257 and 258 above. The Defendants, and presumably UMA also, considered that spraying the underpan with water from a high-pressure hose whilst the hot work was being carried out was the most appropriate precaution to prevent fire. Of course, as Mr Townsend said in his report, live water can, like other protection methods, sometimes be intermittent in coverage. However, not only does a spray of water protect the lining, like static screens, it also extinguishes or cools the hot metal particles that come into contact with it.

I consider that it was reasonably foreseeable that a molten globule of steel might avoid the water and hit and penetrate the horizontal mastic that protected the top of the lining just below the steel strip that was fixed to the top of the side of the underpan. However, I do not consider that the following matters were reasonably foreseeable. First, that a globule of molten metal might penetrate sufficiently far into the lining so as to start a smouldering fire that would not be extinguished when the underpan was hosed down on completion of the work and would not be detected by its smell when the area was being checked for signs of fire. Second, that it might cause a fire that could smoulder unseen and undetectable within the lining for about 1½ hours or more and then make the transition into a visible flaming fire.

If this was how the fire started, and in my judgment it probably was, I do not consider that the Defendants were negligent in not taking one or other of the precautions that might have prevented it, namely draping a welding blanket over the top of the steel strip so as to cover the horizontal line of mastic or maintaining a fire watch for more than 1¼ hours after the completion of the hot cutting work. As I have indicated, by the end of the trial Mr Walker did not suggest the contrary. As was explained in evidence, welding blankets can present a hazard and become a tripping obstruction and so one would not use them in this situation unless it was seen as being reasonably necessary. If, as I find, the sequence of events that caused this fire was not reasonably foreseeable, then I do not consider that the Defendants can be criticised for failing to take the precaution that would have prevented it.

Whatever Mr Percival may have been doing whilst he was in the screen, I am satisfied that Mr Smith carried out the cutting, the positioning of the hose and the subsequent hosing down of the underpan with reasonable care. In my judgment, UMA has failed to prove that Mr Smith was in breach of the duty that he and the Defendants owed to UMA. If there was any inadequacy, it was in the agreed precautions, not in their execution. But since UMA approved those precautions at management level, in the circumstances of this case I consider that it does not lie in its mouth to complain that they were inadequate, particularly when that criticism is made solely with the benefit of hindsight.

Further, as I have already indicated, whether or not Mr Percival was maintaining a dedicated fire watch whilst Mr Smith was carrying out his hot cutting work has no relevance to the cause of this fire. The fire did not break out in a manner that Mr Percival would or could have seen during this period. All that Mr Percival might have seen was that some of the spatter went into the underpan to be apparently extinguished. I find it as a fact that by the time that Mr Percival and Mr Smith left the area of the A side primary screen there was no fire of any sort to be seen in the underpan.

The purpose of having a fire watch, at least as I see it in this case, is to ensure that a fire will be spotted as soon as it breaks out, so that the person or persons present can take prompt and effective action to extinguish it. The lack of a proper fire watch whilst Mr Smith was carrying out the cutting would only have been relevant if the fire had broken out at that time and had not been seen until it was too late to put it out.

Accordingly, this is not a case in which the principle can be invoked that was identified by the Court of Appeal in Drake v Harbour [2008] EWCA Civ 25, an authority relied on by Mr Walker, namely that if there is an established breach of duty and the damage which has occurred is the sort of damage which one might expect to occur from the work being carried out, then it is legitimate to conclude that the damage was caused by the breach - in other words, to take a reasonably robust approach to causation. If in this case, the defendants, in breach of duty, failed to have a proper fire watch whilst Mr Smith was carrying out the hot cutting and the timings had shown that a fire must have broken out during the period of that fire watch, then I can see that causation would be established. But on my findings of fact there was nothing untoward that could have been seen by a person keeping a fire watch during the cutting operation carried out by Mr Smith. In my view it is the nature and extent of the fire watch that was kept after the cutting work was finished that is important.

Since I have now found that the Defendants are not liable for this fire, the question of the liability of the insurers has become academic. However, since the matter has been dealt with fully at the trial I should give my conclusion on the question of the Defendants compliance with the Burning and Welding Warranty.

So far as compliance with the precautions in paragraph 2 of the warranty are concerned, the position is clear. That required combustible materials which cannot be moved to be "covered and protected" by non-combustible material. These two requirements are cumulative. Assuming for this purpose that water is a non-combustible material, the fact that the fire occurred following the sequence of events described in my scenario (7), as I find it probably did, means inevitably that the strip of mastic at the top of the side of the underpan was not covered by the water at all times because a hot molten article got through.

Accordingly, there was a breach of the warranty and the Defendants’ claim for an indemnity would have failed.

The question of whether or not there was a breach of paragraph 3 of the warranty is more difficult. So far as Mr Smith's cutting work was concerned, I find that there was no breach because a fire watch was kept for a period in excess of 60 minutes after his work was completed. The question is more difficult when it comes to Mr Percival's work.

I have already found that Mr Percival could not start vee-ing out the cracks until Mr Smith completed his hot work, which was at 8:15 am. Assuming that by this time the first bolts that Mr Smith had cut had cooled sufficiently to be removed by hand, Mr Percival could straightaway have started vee-ing out the cracks around the holes in the flanges for those bolts. Having done this he might then have started to clean up those flanges using the wire brush attachment on the other angle grinder before changing position and starting to vee-out the cracks on the next set of flanges. Alternatively, he might have decided to complete all the vee-ing out before using the wire brush. One simply does not know: to form a view would be to speculate.

It would also be speculation for me to reach any view as to how far Mr Percival had progressed with his work before he and Mr Smith decided to leave for breakfast. For example, he might have completed all the vee-ing out on one side of the screen in 10-15 minutes - from the description of what was involved that was given in evidence, it would probably have taken only a minute or two per hole. He might then have started to clean up those flanges using the wire brush attachment. I do not know how long this would have taken, but I imagine that it might have taken longer than the vee-ing out.

It is possible that Mr Percival might have been using the disc cutter at a time which was less than 1 hour before he and Mr Smith went off to have breakfast, but this is speculation. It is also possible, but I have not been able to make a finding in relation to it, that Mr Percival was not keeping a dedicated fire watch whilst Mr Smith was carrying out the hot cutting.

The burden of proving that there was a breach of the warranty lies firmly on the insurers. It was open to them to take steps to find Mr Percival and to secure his attendance at court, but there is no evidence that they attempted to do this. Neither of those who investigated the fire on behalf of the insurers at the time, Mr Heath and Mr Wookey, expressed any view that supported an allegation of breach of warranty in relation to the fire watch.

In these circumstances I would not have been prepared to find that, as a matter of probability, there was a breach of the warranty in relation to the fire watch. The allegation of a breach in relation to the availability of fire extinguishers was expressly abandoned by Mr Shepherd during his oral closing submissions.

My findings can be summarised as follows:

In these circumstances, the Defendants are not liable, either in contract or in tort, to UMA in respect of fire.

The Defendants’ claim against the insurers for an indemnity in respect of their liability for the fire does not arise. However, if the Defendants had been found liable to UMA their claim for indemnity would have failed because they were in breach of the Burning and Welding Warranty in the policy. The only breach was in relation to the failure to protect and cover the lining of the underpan; no breach has been established of the requirements relating to keeping a fire watch or having fire extinguishers readily available.

UMA's claim against the Defendants fails and is dismissed. The Defendants’ claim against the insurers also fails and is dismissed.

I will hear the parties in relation to questions of costs if they cannot be agreed.