England and Wales High Court (Technology and Construction Court) Decisions

IN THE HIGH COURT OF JUSTICE
QUEEN'S BENCH DIVISION
TECHNOLOGY AND CONSTRUCTION COURT

		Royal Courts of Justice Strand, London, WC2A 2LL
		22nd July 2011

Mr Justice Coulson:

A: INTRODUCTION

1. The catastrophic fire which destroyed the claimants'^[1] large confectionery factory in Pontefract on the evening of 8 June 2005 came as something of a surprise to everyone involved. It was thought by those present, and by the Fire Brigade when they first arrived, that the earlier fire which had broken out in the 'oil pop' popcorn production area had been satisfactorily extinguished. But it soon became apparent that this was incorrect, and 40 minutes after the alarm was sounded, the fire was so well-set that the Fire Brigade decided that it could not be aggressively fought.
2. That the development of the fire was unexpected has one significant consequence for this litigation: it means that, as the fire experts agreed, each of the three different theories as to how the fire spread was neither obvious nor straightforward. Repeatedly in the evidence, each of the competing theories as to fire spread was described as no more that merely possible, and the experts spent much of their time arguing for one theory or another on the basis that their favoured theory was less improbable than the alternatives.
3. Ironically, an analysis of the events surrounding the original contract between the claimants and the defendant for the supply of a CO2 fire suppression system for the popcorn machinery (but not the 'oil pop' production area as a whole); the events at the time of the installation of that system and thereafter; and the wider evidence as to the health and safety regime at the factory, and the claimants' fire prevention measures in particular, leads inexorably to the conclusion that, whilst the occurrence of this particular fire on this particular evening might have been surprising, the fact that a large fire originated in the 'oil pop' popcorn production area, and went on to destroy most of the claimants' huge factory building in a very short time, was all too predictable. The size of the damages claim in these proceedings, put at over £100 million, only increases the sense of missed opportunity that hangs over the whole story.
4. The building that was largely destroyed in the fire was called the New Manufacturing Unit ('NMU') located at Ferrybridge Road, Pontefract, in West Yorkshire. These premises were regularly referred to as 'Monkhill', although that was also the name by which the second claimant was commonly known. At the site there were two production buildings, an old manufacturing unit dating from the 1920s, and the NMU. The NMU had originally been built in 1989 but, in 1997, it was the subject of a major extension. It is common ground that the fire started in the extension to the NMU, in a part which, when it was built, was called Production Area B but which, until 2003, was apparently just used for storage.
5. In 2003, the claimants decided to move their existing 'oil pop' popcorn production facility in Leeds to Production Area B in the extension of the NMU. Both budget and timetable were tight. The existing popcorn production unit was protected in Leeds by a wet chemical system over the cooking pans (where the popcorn was heated in oil so that it 'popped'), and a CO2 suppression system in the elevator, which carried the popcorn in buckets up and across part of the factory so that it could be dropped into a hopper (also covered by the CO2 suppression system), from which the sleeves of popcorn were then filled manually.
6. Because ADT, the defendant, already provided fire protection systems at the NMU, they were engaged by the second claimant (whom I shall call "Monkhill") to provide the wet chemical system and the CO2 suppression system for the two new 'oil pop' popcorn production lines at the NMU. The contract was agreed in September 2003. There is a dispute over the relevant terms and conditions. There is no dispute that the CO2 element of the contract was worth a total of about £9,000 for the two lines.
7. The CO2 suppression system was installed in October/November 2003. Production started before the system had been fully tested and commissioned. Commissioning was completed in December 2003. Although, in early 2004, an additional fire detection system supplied by the defendant, known as VESDA, was considered by the claimants as a way of minimising the risk that burning popcorn might get onto the conveyor from the pans, it was not ultimately proceeded with. Similarly, one significant recommendation made by the claimants' own Group Risk Department (both before and after the contract between the parties), for the installation of sprinklers into the NMU, was not followed, whilst a critical assumption about the existence of fire segregation turned out to be entirely misplaced. These errors and omissions form an important part of this case.
8. During 2004, changes were made to both the hopper and the elevator. The defendant had some limited involvement in those changes. There were also other events, including a fire in the hopper in June 2004, on which the defendant now relies in these proceedings, not least to support its allegations of a break in the chain of causation. Generally, very few of the operatives in the 'oil pop' popcorn production area were aware that the CO2 suppression system protecting the elevator and the hopper could be manually activated, and even fewer knew how that manual activation was to be triggered.
9. On the evening of 8 June 2005, an operative, Alan Hardcastle, was filling the large plastic containers, known as sleeves, at the C Line hopper. He suddenly became aware that the sleeve that he was filling was melting because of the burning popcorn within it. He raised the alarm. The shift manager, Dave Carter, ran across to the C Line hopper and went up the ladder to the gantry. He looked down into the hopper and saw that the popcorn was on fire. Flames were coming to the top of the hopper. He shouted for the fire hose and the conveyor was switched off. He doused the hopper with water and, within a few seconds, he put out the fire.
10. Meanwhile, following a procedure which the operatives had used before (and which had been approved by the managers), Mr Hardcastle released some of the burning popcorn from the bottom of the hopper. It is estimated that a total of about three sleeves of burning popcorn were deposited on the floor before the hatch jammed, which may have been as much as 180 litres of popcorn altogether. The operatives then endeavoured to stamp out the burning popcorn on the factory floor. When the fire alarm sounded, the operatives thought that the fire had been put out, and they left the 'oil pop' production area, went into the next door packaging hall, and exited the NMU.
11. When the Fire Brigade arrived, they were told that the fire had been put out. There was then a curious episode, explored in greater detail below, in which the Fire Brigade did not immediately begin the fire-fighting operation, but instead sought a key to a door that led into the back of the 'oil pop' area. This involved a security guard being called and, so it appears on the CCTV footage, potentially valuable time was lost. At about 8.52pm, it became apparent that there was a serious fire, and smoke could be seen venting at the eaves on the south wall of the NMU. Very shortly thereafter, at about 9.08pm, it became apparent that the fire had taken such a firm hold that the Fire Brigade would adopt only a defensive attitude towards it. In essence, this meant that the fire was fought from the outside only, allowing it to rage through the empty NMU but keeping it away from any adjoining buildings. In consequence, the whole building, and the machinery within, was destroyed.
12. There are three basic alternatives to explain the rapid spread of the fire. The defendant puts forward two different theories. One is that the operatives failed to put out the burning popcorn by stamping it out on the floor, and that pieces of burning popcorn, which had been released from the hopper and not extinguished, set fire to boxes and other combustible material in the 'oil pop' production area. The fire then spread to the south wall where combustible materials were tightly packed on racking. The other theory put forward by the defendant is that, unbeknownst to him, Mr Hardcastle had already put some smouldering popcorn into a sleeve which had then been put in a box and moved towards the south wall. This theory postulates that, just after the operatives left the building, the smouldering popcorn in the filled sleeve ignited and set fire to the box and subsequently the other combustible material on the south wall.
13. The claimants' theory as to the spread of fire is that, before Mr Carter reached the gantry and put the fire out with the hose, the flames in the hopper had extended upwards, beyond the top of the hopper, and ignited one of the buckets in the C Line elevator. It is then said that the fire spread through the plastic components of the C Line elevator, with molten plastic dripping out of the openings in the elevator (in particular the mesh at the boot of the vertical section) which created a pool fire. That pool fire spread then back towards the south wall and ignited the combustible materials there.
14. Accordingly, it can be seen that both sides' theories as to fire spread start at the oil pop cooking pans, and end at the south wall. The principal difference is that the claimants' case involves fire spread through the elevator (where the CO2 system was in place) and the defendant's theories involve a spread outside the hopper and elevator (where there was no fire prevention system).
15. I propose to deal with the issues that arise in this case in the following way. I make some general observations on the factual evidence in Section B. In Section C, I deal with the details of the construction of the NMU and, at Section D, with Project Kettle, the claimants' scheme to move the 'oil pop' production from Leeds to the NMU. At Section E, I address the formation of the contract between the claimants and the defendant and, at Section F, I set out the terms of that contract.
16. Thereafter, at Section G, I deal with the manufacturing process as it existed between 2003 and 2005 and, at Section H, I identify what appear to me to be the relevant post-contract events and the incidence of fire. At Section I, I set out my findings of fact in connection with the fire on 8 June 2005. At Section J, I make some general observations on the expert evidence before coming on to address the two matters on which expert evidence was particularly important: the spread of the fire (Section K), and the potential liability of the defendant for the fire (Section L). At Section M, I address the arguments about causation, and at Section N, the defendant's allegations of contributory negligence. At Section O there is a short section dealing with points raised by the defendant as to separate duties of care and the claimants' title to sue. There is a short summary of my conclusions at Section P below.
17. Before embarking on this process I ought to express my thanks to all counsel for the highly efficient way in which the trial was conducted and, in particular, the incisive cross-examination of the many witnesses on either side. It was a model of its kind, made more impressive still by the fact that there were many aspects of this potentially large case that had not been fully thought through by the time the trial started (and even by the time it finished). This general lack of readiness manifested itself particularly in the problems with the expert evidence (Section J below), and the unsuccessful mediation, arranged over two months after the end of the trial, which caused a delay in the finalisation of this Judgment.

B: GENERAL OBSERVATIONS ON THE FACTUAL EVIDENCE

B1: The Claimants' Witnesses

18. The claimants' witnesses fell into four categories. First, there were those concerned with the claimants' procurement systems, which was relevant to the claimants' case that such systems meant that the defendant had previously contracted on the claimants' terms and conditions. They were Robert Fowler, Michael Hearn, Natasha Jane Lee, and Wendy Parsons.
19. Secondly, there were those concerned with the moving of the 'oil pop' production from Leeds to Pontefract, an operation which was known as Project Kettle. They were principally Stephen Barker and Andrew Hamilton, although certain aspects of Project Kettle, and some subsequent events, were dealt with by Gerard McDonough, Andy Eames and Graham Webber. Thirdly, there were those concerned with questions of risk assessment and health and safety training. They were John Caldicott, Mandy Clough and Michelle Irving. Fourthly, there were the group of witnesses who gave evidence about the fire on 8 June, including Dave Carter, Graham Challinor, Kevin Holmes, Tony Norton, Pat Priestley, and Philip Shaw.
20. As to the witnesses in the first category, their evidence was general in nature and of limited assistance. Through no fault of their own, they had been unable to unearth any specific evidence which showed that the defendant had in fact previously contracted on the claimants' terms and conditions.
21. The witnesses in the second category were of particular importance. Mr Barker did his best to remember particular conversations and events, but it was quite clear that he had little recollection of a number of the key discussions and events in which he was involved. As we shall see in the more detailed analysis in Section D below, there were times when he exhibited a rather casual attitude to some of the important tasks with which he was entrusted.
22. The other principal witness in this category was Mr Hamilton. I am afraid that, of all the factual witnesses called by the claimants, I regarded Mr Hamilton as the least satisfactory. He regularly refused to answer the questions that were put to him, and gave the impression of being both defensive and evasive. Given the need to stick to the trial timetable, I found myself intervening more regularly than I would have wished during the course of his cross-examination, in order to try and get him to answer the questions that were being put to him. It was unfortunate for the claimants that Mr Hamilton made such a poor impression, given that his evidence was so important and played such a vital role in so many of the key events. In the round, I considered his evidence to be self-serving and unreliable.
23. The witnesses in the third category were honest and careful. They did, however, exhibit a trait that was noticeable across all categories of the claimants' witnesses, namely a reluctance to accept, even with hindsight, that a particular decision was their responsibility, and the tendency to identify the relevant decision-maker as "we", without ever being able to identify who those people were and when and how the relevant decision was actually taken. On a number of occasions, witness X said that a decision was the responsibility of Y, but when that was put to witness Y, they exhibited genuine surprise at such a suggestion (and sometimes indicated that responsibility actually lay with witness X).
24. The witnesses in the fourth category were, in many ways, the most impressive of the witnesses called by the claimants. I have concluded that Mr Carter and Mr Norton in particular demonstrated considerable bravery in attempting to deal with the fire on 8 June 2005. Although some of the actions that they took when the fire was first spotted could, with hindsight, be criticised, that stems, not from any failing on the part of the operatives themselves, but from omissions in their training as to the appropriate fire-fighting methodology to adopt.

B2: The Defendant's Witnesses

25. The defendant's witnesses also broke down into four categories. The first category (Hugh McQuaid, Brian Beaumont and Douglas Flavell) were those involved in the contract for and the design of the CO2 suppression system for the elevator and the hopper. The second category (Paul Taylor and Ian Walker) were involved in the installation and commissioning of the system and, in Mr Taylor's case, were also involved in the later reconfiguration of the elevator. The third category (Dean Kear, Kevin Arbuthnot, Carl Butler, and Kevin Kemp) were all witnesses giving evidence related to the cause and spread of the fire, although only one (Mr Kear) was actually present at the NMU on 8 June 2005. The fourth category (Adrian Dearden, David Doherty, Andrew Hunt, William Kerr, Stephen Laybourn, Christopher Shields, Brian Turner, Steve Vernon, and David Wilson) were all witnesses whose statements were taken as read, and who did not therefore give oral evidence. I agree with paragraph 40 of the claimants' written closing submissions that, with one or two exceptions noted in the subsequent Sections of this Judgment, the witnesses in this category added very little to the matters that I have to decide.
26. As to the witnesses who gave oral evidence, I found that those in categories one and two above were straightforward and honest witnesses who endeavoured to answer the questions put to them. One common theme of their evidence was that they often assumed that a particular aspect of the contract work would be checked or considered by somebody else when, on the evidence, this did not happen and was never likely to have happened. Accordingly, although Mr Dennys QC, on behalf of the defendant, correctly suggested that there were times when the evidence showed that the claimants' left hand did not know what its right hand was doing, the same criticism could be made of some elements of the defendant's own evidence.
27. As to the witnesses in the third category, I note that Mr Kear was a previous employee of the claimants and, although they descibed him as a disgruntled ex-employee, I found his evidence to be careful and honest. Mr Arbuthnot and Mr Kemp, who were involved as representatives of the Fire Service, were also honest witnesses but, in particular in relation to Mr Kemp, their recollection of the relevant matters was hazy. Dr Butler added very little, despite the fact that he had been retained on behalf of the defendant immediately after the fire and inspected the remains of the NMU.
28. With those general observations concerning the factual evidence in mind, I turn to deal with the first two narrative sections of the Judgment: that is to say, the evidence about the NMU itself, and the events surrounding Project Kettle, leading up to the making of the contract between the claimants and the defendant.

C: THE CONSTRUCTION OF THE NMU

C1: Original Fire Segregation/Compartmentation

29. The older part of the NMU was constructed in 1989, largely using materials, such as polyurethane and polystyrene, now recognised as being a fire risk. The extension to the NMU, which was where the fire started, was built in about 1997, generally with internal panels containing rockwool, a more fire-resistant material. The extension consisted of three parts. In the western part of the building there was an area known as Production Area A, where the 'air pop' popcorn production took place. Air pop was a process of production which involved heating the popcorn in ovens until it popped. It is clear from the relevant as-built drawings, and I find as a fact, that Production Area A was surrounded by specific fire-rated panels which were designed to provide a 1 hour fire-resistant box. All of the drawings show Production Area A as a self-contained compartment within the NMU with a 1 hour fire resisting wall around it.
30. To the east of the Production Area A was what was known as the Packaging Hall, where the popcorn was put into sleeves and boxes. It contained a large amount of combustible material. To the north of the Packaging Hall was Production Area B, which was originally used for storage. It was shown on the as-built drawings, produced by Turner Design and Build, as being divided from the Packaging Hall by what was described as a "hygienic partition". It is important to note that, contrary to the walls around Packaging Area A, the as-built drawings do not indicate that this dividing wall, or any other part of the structure around Production Area B, was intended to provide 1 hour – or any - fire resistance. Indeed, there are no as-built drawings or similar documents which indicate that any other fire resistance was provided in the NMU, apart from that surrounding Production Area A.
31. It was an important element of the defendant's case on both causation and contributory negligence that the construction of Production Area B (which was eventually the place to which the 'oil pop' production process was moved in 2003), did not involve any fire-resistant walls and did not offer any fire compartmentation. The as-built drawings which appeared to make that clear were therefore put to Mr Hamilton, the man who acted as project manager for Project Kettle. He was typically evasive, saying on more than one occasion that Production Area B "was designed with a popcorn process in mind". He went on to say that, because the 'air pop' (in Production Area A) was a 1 hour fire-rated process, "it would make sense for that [Production Area B] all to be 1 hour fire-rated as well." Obviously that would have made sense, but the issue was whether or not it was ever done. On all the evidence, I find that it was not done, either when the extension was built in 1997 or, as set out below, when the 'oil pop' process was moved to Pontefract in 2003. I find that the only part of the NMU which contained any attempt at fire segregation was Production Area A^[2].
32. It should also be noted that, when the 'oil pop' process was set up in Production Area B, it was convenient to make a hole in the wall between that area and the Packaging Hall, so that the new elevator could pass through the dividing wall. This allowed some of the popcorn, for the retail market, to be boxed up in the Packaging Hall, whilst the popcorn for the cinema trade was brought back into Production Area B, and dropped into the C and D Line hoppers. The evidence was that there was no seal or firestopping between the new elevator and the wall, so the hole in the wall around the elevator meant that there was no fire separation between the two areas. In fact, it appears that what set the fire alarm off on the night of 8 June was smoke or steam passing through that gap and into the Packaging Hall.
33. Mr Webber, who was the engineering manager at the NMU, gave evidence about the suspended ceiling above the 'oil pop' production area. He described it as being strong enough for a man to walk on. It had a coated steel sheet top surface and a hygienic white powder surface coated below. Thus, in the NMU, the operatives looked up to see a coated UPVC layer. Above, in the ceiling void, there were lights that were recessed into the suspended ceiling. The bulbs had to be changed from the void. His evidence in re-examination made clear that he could not help with the precise make-up of the panels, because he never saw the insides of the panel.
34. The evidence concerning fire-stopping in this suspended ceiling (in other words, the extent to which a fire, if it got into the ceiling void, would have been contained) was agreed by the experts. It was agreed that the compartmentation around Production Area A was "taken full height to the roof. Save for this, no cavity barriers are indicated in the roof space"^[3]. The evidence was that this was the only part of the NMU which was even possibly the subject of such fire-stopping. Indeed the evidence suggested that, in relation to the remainder of the NMU, there was no fire-stopping in the roof void at all.
35. There was certainly no fire-stopping in the ceiling above the 'oil pop' production area (the former Production Area B). This meant that, once fire got into that roof void, it could spread quickly at roof level right across the NMU. The cross-examination of Dr Mitcheson ended with this exchange:

"Q: So, if you have a fire in Production Area B, which we know we had, then a fairly obvious route for it to travel is upwards into the roof void, across the popcorn packaging hall at that level, and into the area of the NMU which pre-existed the extension?

A: Correct."

The absence of fire-stopping in the roof void above Production Area B (and its possible presence above Production Area A) was entirely consistent with the as-built drawings, and my findings based upon them, namely that when the NMU extension was built, Area A, which was going to be used imminently for production purposes, was therefore the subject of fire segregation works, whilst Area B, which was not at that point going to be used for anything other than storage, was not the subject of fire segregation works.

C2: Were Fire Segregation Works Carried Out in 2003?

36. No fire segregation works were carried out around Production Area B – or anywhere else – as part of Project Kettle. Indeed, that was ultimately common ground. This was a significant failure on the part of the claimants, not least in failing to follow their own guidance documents. For example, paragraph 6.2.2 of revision 4 of the claimants' 'Risk Control Guide to Property Conservation' recommended a minimum 4 hour fire separation between production and storage areas, with particular care needed "around more 'hazardous' processes…" It appears that this recommendation was simply ignored.

C3: Later Evidence About Fire Segregation

37. I should refer to two subsequent events directly concerned with the issue of fire segregation within the NMU. In 2004, after the 'oil pop' had been moved, and because the claimants' Group Risk department were concerned about the risks posed by the 'oil pop' production process, Mr Barker was asked to prepare a register showing the extent of polystyrene panels in the construction of the NMU. The eventual product of this process was a single, smudged sketch which showed both polystyrene and rockwool panels in the construction. The drawing indicated that polystyrene panels had been used for the internal and external walls of the old NMU, including the internal partitions, but that rockwool panels (that is to say panels with a rockwool core, which were regarded as better for fire-protection purposes) had been used in the 1997 extension. Amongst other things, this drawing indicated that the rockwool panels had been used for both the fire separation walls around Production Area A and for the construction of the wall between Area B and the Packaging Hall. That was not, of course, what was shown on the original as-built drawings.
38. Mr Barker was unable to explain how and why the subsequent sketch was different. He said that he had employed a company called Craig Engineering to carry out the survey and they had marked up a drawing (now lost), which he had copied in order to produce this one. He did not think that the actual results of the investigation had been recorded anywhere. In addition, he confirmed that Craig Engineering did not appear to have looked at the internal or external skins of the roof, agreeing that they had simply produced a plan of the walls. There was therefore no evidence of any kind to support the accuracy of the subsequent sketch. To the extent that it conflicted with the as-built drawings, I consider that the as-built drawings, which were properly drawn, to scale, and produced by those involved in the construction process, were more likely to be accurate.
39. The failure to segregate the 'oil pop' production area quickly became a subject of concern and investigation after the fire itself. On 28 June 2005, 3 weeks after the fire, Mr Alan Beard of the claimants' Group Risk Department asked those on site:

"Could you confirm whether a fire rating was assumed confirmed for the wall between oil pop and popcorn popping for the fire risk assessment?"

In other words, he wanted to know whether the 'oil pop' production area had ever been segregated. As we shall see, this query arose in part because such compartmentation had been assumed to exist by the Group Risk Department years before, and they were now (rather late in the day) asking to see the evidence that underpinned that assumption.

40. There was never a direct answer to this question, either at the time that it was raised or when the matter was explored with Mr Hamilton in cross-examination. His elliptical answer at the time (contained in an internal email) was in these terms:

"We knew that all the walls/ceiling construction within popcorn was rockwool, but we were not aware of differing insulation standards to create additional fire integrity.

We did know that the air pop area had to have fire shutters to the doors and realised that we had to install one additional fire shutter door to production area…to provide a similar level of protection for this production area."

In his oral evidence, Mr Hamilton was unable to do better than suggest that a 1 hour rating had been assumed, although he could not say how or why.

C4: The Fire Certificate

41. It is usual in fire cases for any issue as to the fire rating or fire prevention qualities of a building subsequently destroyed in a fire to be answered by reference to the Fire Certificate issued by the local Fire Service. Unhappily in this case, the Fire Certificate poses more questions than it answers.
42. The Fire Certificate itself is dated 28 March 1991. That Certificate therefore related only to the original NMU. In 1996, there was apparently a discussion between Fire Officer Sedgwick and the claimants about the extension, because the Fire Service retained a copy of a drawing, showing the extension, which contained his manuscript notes made at a meeting in November 1996. The plan and the notes are consistent with the intention to segregate Production Area A and nothing else (see paragraphs 31 above). There was certainly no note concerning the compartmentation of Production Area B, presumably because, in 1996, there was no immediate plan to use Production Area B for a manufacturing process involving fire (or if there was, it was not shared with Fire Officer Sedgwick). It appears that there was subsequently an inspection by the Fire Brigade, after the NMU extension had been completed, on 30 June 1999. Although this recorded that the building was not as shown on the certificate plan, it also said that the premises still met the criteria for Certification^[4].
43. Accordingly, in 2003, when the claimants were proposing to move the 'oil pop' production lines from Leeds to Pontefract, into Production Area B at the NMU, the Fire Certificate was something of a muddle. The claimants did not even have a copy of it. Moreover, as set out in Section D10 below, the situation was not improved when the Fire Officer visited in September 2003. And, although the NMU was visited occasionally after that, the Fire Certificate of 1991 was never subsequently amended.
44. However, whatever else was unclear, one thing can be deduced: that at the time of the amendment to the Fire Certificate in 1996 (if this is what the drawing, with Fire Officer Sedgwick's manuscript notes, amounted to), at the time of the inspection in 1999, and even in early 2003, there was nothing in the Fire Certificate documentation to indicate that Production Area B was shortly to be used for an industrial process involving naked flames.

C5: The Construction Generally

45. Save for the unequivocal agreements set out in the Joint Statement of 31.3.11, which are infrequent, the expert evidence was unable to provide very much assistance on the question of the construction of the NMU generally. That was as a result of the lack of evidence after the fire, and the lack of proper records. Dr Mitcheson, the claimants' fire expert, thought that the photographs of the panels after the fire did not show a residue typical of polyurethane or any other foam, and considered that the patterns showed the paint finish of the panels. This might have been consistent with the lack of plastic panels, although this evidence was rather vague and a little haphazard. Dr Mitcheson also said that any areas where panels joined the roof at an angle was a weakness and that the evidence was that there were gaps in the roof void that would allow an employee to walk from the old to the new section of the NMU. That was obviously a route, as Dr Mitcheson confirmed, for the transmission of fire.
46. In summary, I find that:

a) The older part of the NMU had an external envelope of steel cladding insulated with polyurethane, and internal partitions insulated with polystyrene. It was therefore highly combustible.

b) The extension of the NMU was constructed of external steel panels, in which the type of insulation is unknown^[5], and internal panels, sometimes known as 'hygienic panels' which generally used 'rockwool' insulation. There was no fire-rating for these panels. Only Production Area A was the subject of specific fire segregation works; all other internal walls were not compartmentalised, and had no particular fire rating.

c) Save for the roof void above Production Area A (which may have had the fire-stopping above it, at least in the sense that the compartment walls were taken up to the underside of the roof), the roof void was open and would allow fire to travel uninterrupted throughout the building.

47. It may be, as Mr Hamilton indicated, that it had long been the claimants' intention that Production Area B would become the site of the 'oil pop' production area, and that the same fire compartmentation would be required as for Production Area A. But, on the evidence, that intention was not realized. As to the detail, there were three matters which, I find, the claimants never addressed in 1997, or 2003, or thereafter. First, as shown on the as-built drawings, unlike Production Area A, Production Area B was not designed or built as a separate fire compartment with fire-rated walls. If the 'air pop' process necessitated a 1 hour fire resistant compartment, then the 'oil pop' process, which was much more dangerous, because it involved naked flames, necessitated at least the same level of protection, and probably something closer to the 4 hour rating recommended by the claimants' guidelines (paragraph 36 above). It never got it.
48. Secondly, when in 2003 the elevator was knocked through the wall between Production Area B and the popcorn Packaging Area, there was never any attempt made to seal up the hole around the conveyor. This provided a clear and obvious path for smoke and fire from the 'oil pop' area to the Packaging Hall, with its particularly combustible materials. No thought appears to have been given to the integrity of the 'oil pop' area at all. And thirdly, as Dr Mitcheson confirmed, there was no proper fire stopping in the roof void, again demonstrating the lack of any effective fire separation around the 'oil pop' area.
49. Accordingly, when the NMU was built, Production Area B was not fire-stopped or otherwise segregated. That omission was not spotted or rectified at the time of Project Kettle; on the contrary, as set out in Section D below, a wholly false assumption was made that such compartmentation already existed, and matters were made worse by the failure to put in fire-stopping in the roof, and the cavalier attitude to the effect of the conveyor passing through the wall into the Packaging Hall. No segregation or compartmentation works were carried out until after the fire in June 2005, when the investigators belatedly recommended rebuilding the destroyed NMU using 2 hour-rated fire partitions between manufacturing zones.

D: PROJECT KETTLE

D1: The Risks Involved In The Move From Leeds To Pontefract

50. Up to 2003, popcorn using the 'oil pop' process was manufactured at one of the claimants' subsidiary companies, Kernels, in Leeds. Kernels had moved to the factory in Leeds after their original popcorn factory had burnt down. Ms Priestley, one of the claimants' shift managers, explained that it was her understanding that there had been a big fire at Kernels which had been caused by burning popcorn from a pan fire spreading to the hopper.
51. Mr Hamilton was also aware in 2003 that the process with which he was concerned had previously caused a major fire, although he said that he could not remember the details. He also said that, when he first visited the Leeds factory in early 2003, he had been told that they had also had two more recent fires, one in August and one in December 2002.
52. This was important information. It meant that the claimants, and Mr Hamilton in particular, knew from the outset that the very operation which they proposed to install at the NMU was so dangerous that it had already accounted for one factory and caused fires in a second. They should have been keenly conscious of this, so as to undertake a comprehensive review of the fire risks inherent in moving the 'oil pop' production to Pontefract. As the following paragraphs make clear, that did not happen.
53. Mr Hamilton confirmed that the proposal was to move the 'oil pop' production from Leeds to Pontefract in order to increase the manufacturing output. At Leeds there were two rooms, with about 6 pans in the salt room, and about 12 pans in the sweet room, making a total of 18 pans. Although the documents are confused, I think it more likely than not that both processes had CO2 protection. The plan for Monkhill was that there would be two lines of oil pop popcorn production, known as the C Line and the D Line. There would be a total of 24 pans, 12 for each Line.
54. The timing was very tight. As Mr Barker confirmed, it was imperative that the new lines were up and running by Christmas 2003, in order to meet the surge in demand for popcorn, particularly for the cinema trade, over the Christmas period. This timetable could not be changed. As to the hierarchy, Mr Hamilton was to be the project manger in charge of the project, which was named Project Kettle. Mr Barker, an experienced mechanical engineer, was essentially his deputy. As set out later in this Section of the Judgment, they were the two people principally involved in the arrangements, including the making of the contract with the defendant.
55. I find that Mr Hamilton never fully appreciated the increased risk that the proposed transfer of the 'oil pop' production would bring to the NMU. As far as he was concerned, Project Kettle was a straightforward transfer of a particular process from one factory to another. He never sat down and worked out what the risks were in moving the 'oil pop' production into Production Area B of the NMU and/or how those risks could be best dealt with.
56. However, those increased risks were realised by representatives of the claimants' Group Risk Department. But a series of errors and misassumptions meant that those concerns were never properly addressed. Mr Caldicott, one of those representatives, agreed in evidence that there was an increased risk because the 'oil pop' process was being transferred to the NMU, a factory which undertook a lot of other operations. He said: "Worst case scenario, yes, there was a risk of cut-across from this line into the existing facilities, yes." He went on to say that the process:

"…was inherently risky. It had open flames. That was the first time that I had seen open flames within any of our production facilities. So immediately the flag was raised to a need to look carefully at the risk controls with that. We were picking it up and putting it onto an existing site. My view was that we were putting it into a segregated box."

57. His reference to a 'segregated box' was important because it was always Mr Caldicott's (mis)assumption that there was 1 hour fire segregation around Production Area B. When he was picked up on that answer, and asked if a proper investigation needed to be carried out as to the integrity of the alleged box, he agreed, saying:

"We needed certainly to be confident that we had a level of segregation within that box, certainly."

In fact, as noted in the preceding section of this Judgment, no such investigation was ever done, whether in 2003 or thereafter. As a result, the claimants put the 'oil pop' process into a location which they knew required fire segregation; which they assumed, without any investigation, had such segregation; and which, in fact, had no fire segregation or compartmentation at all. As previously noted, this was an important failure on the part of the claimants to act in accordance with their own published guidance and in the manner that Mr Caldicott said that he expected.

58. On 7 April 2003, Mr Hamilton produced his first attempt at estimating the costs of Project Kettle. He made plain that it had been costed out "on a requirement of minimum spend to introduce a safe system of work". On the attached spreadsheet, amongst other items listed, were new doors into the extension. It said that these would segregate the current packing area to "maintain conditions". There was said to be a "risk to popcorn packing without this."^[6] Mr Hamilton confirmed that this was a reference to the wall between the Packaging Hall and Production Area B where the 'oil pop' process would go. He said that the proposed doors in that wall were to provide fire segregation and also to protect the Packaging Area from moisture. Again, like Mr Caldicott, he worked on the groundless assumption that the wall was already a 1 hour fire separation wall, when one look at the as-built drawings would have shown him that that was not the case.
59. In addition, the spreadsheet included an item, "relocate fire systems", at a cost of £12,000. It was said that this "assumes suitability and capability ok". Mr Hamilton denied that he was simply assuming that the transfer of systems from Leeds to Pontefract would be straightforward, and said that he was sure that there would be a need to investigate the systems themselves in a bit more detail, and to consult with the claimants' Group Risk Department.
60. On 25 April 2003, in a further email from Mr Hamilton, the proposed work and costs were updated. This showed that the cost of a new fire suppression system was put at £38,000. It was unclear where this figure had come from, although Mr Hamilton said that it was his budget estimate for moving the existing Wormald system from Leeds to Pontefract. At Leeds, Wormald provided a system that involved wet chemical fire suppression above the pans, and a CO2 system within the existing elevators and hoppers. By this stage, the favoured option for Project Kettle was described in Mr Hamilton's contemporaneous documents as "Transfer existing Leeds plants plus add additional kettles from back-up plant".
61. In one such option document, under the heading 'Risks' Mr Hamilton had included these words:

"Introduce oil popping into Ponte

Skills/Culture issues

Increase congestion

Additional warehousing required."

He said in cross-examination that, in this document, he was identifying the risk of introducing the 'oil pop' process into Pontefract, which was principally the risk of introducing an open fire source and a hot process into a manufacturing unit which did not otherwise have one. The short note here does not demonstrate that he considered the matter in any detail. I find on the evidence that he did not.

62. Mr Hamilton was also asked who, within the claimants' organisation, owned that risk. He avoided the question, saying that "we took advice from Group Risk. I do not know whether they took advice from others". As to who he meant by "we", he said "myself and some of the operations team". He went on to say that it was "a complex and important area to ensure that we covered off the risk correctly". Eventually, he was driven to say that the responsibility for the risk identified in his option document was owned by Mr Peter Sanders, the operations director at the NMU, who did not give evidence at the trial and who, on the face of the documents, had very little involvement in the detail of Project Kettle. In my judgment, this was a typical bit of 'blame-shifting' by Mr Hamilton: he was the project manager for Project Kettle, and I find that he was therefore responsible for recognising and then minimising all the risks created by that project.

D2: The Meeting On 24 June 2003

63. Mr Sanders was, however, involved in one part of the story: he sent an email to Mr Caldicott, from the claimants' Group Risk Department, seeking advice on fire precautions. This led to an important meeting at the NMU on 24 June 2003 involving, amongst others, Mr Stephenson and Mr Caldicott from the claimants' Group Risk Department, and Mr Hamilton. Mr Caldicott confirmed that this comprised a partial review of the NMU site risk, although a full review was not subsequently carried out. He said that it was important to ensure that the claimants' own guidance documents were followed. The particular emphasis, he said, was on 'not having to come back to the facility to carry out additional work after the project had been completed'.
64. There are no notes of the meeting on 24 June. There is, however, a copy of a drawing of the NMU which was discussed at the meeting. That drawing, produced by Mr Barker, showed the existing NMU, but it was overlaid, in Production Area B, with a schematic drawing showing the location of the 24 pans, the 2 new elevators, and the 2 new hoppers. It also showed the elevator passing though the wall into the Popcorn Packaging Hall. The drawing was noted up with Mr Stephenson's manuscript amendments.
65. Mr Hamilton could not remember what, if any, other background detail he would have provided to Mr Stephenson and Mr Caldicott. He did not know if he explained to them the experience of popcorn manufacture at Leeds and the obvious fire risks from that process; there was no evidence to suggest that he had. He thought it was more a question of telling them what they were planning to do at the NMU: "what does it look like now, what is it going to look like at the NMU." Certainly that was what was shown on the drawing. As to fire prevention, Mr Hamilton arranged for Mr Eames, the maintenance manager, to be there in order to explain about the fire detection systems in place at the NMU. It appears that there was no discussion about the fire suppression systems that were then in place at Leeds.
66. The notes on the drawing refer to the CO2 suppression system in the two hoppers. Mr Hamilton said:

"As far as I can recall, we didn't discuss anything other than – well R102 systems, wet chemical systems to the process, blankets and CO2 system to hoppers and conveyors.

Q: So by this date it has been decided to use a wet chemical system and either a CO2 system or some water-based system for the hoppers and probably also the elevators, correct?

A: There was no probably. We only ever considered hoppers and elevators."

67. Mr Stephenson's notes referred to "CO2 or water head" which was agreed to be a reference to sprinklers in the hoppers. Mr Hamilton eventually agreed that "we were probably discussing both options at this time."
68. There was also a discussion on 24 June about the need for the new door between Production Area B and the Packaging Hall to be a fire door. That was noted in Mr Stephenson's handwriting. It was unclear whether there was any discussion about fire compartmentation or separation; in my view, based on the evidence I heard, that was very unlikely. The discussion was general. There seems simply to have been an erroneous assumption made at the meeting, confirmed by Mr Hamilton, that the wall between Production Area B and the Packaging Hall was at least 1 hour fire resistant. For that reason, it was agreed that the fire door was to have a similar rating. There was no cogent evidence as to how or why this erroneous assumption had been made.
69. Mr Caldicott could not recall discussing the detail of the existing construction, and had no recollection of being shown any drawings of the original construction of the extension to the NMU. He noted that Mr Stephenson's notes on the drawing indicated that there were rockwool panels in all areas of the extension whilst, in the older part of the NMU, they were polystyrene. However he confirmed that it was always difficult to check such an assumption because, as he put it, "you can't always tell what's behind the panels". He assumed that a view had been taken that the walls provided at least a 1 hour fire separation, but he did not know why that assumption had been made. Mr Hamilton was also unable to explain it. Both men accepted that the only construction information now available (ie the as-built drawings) did not show 1 hour fire separation.
70. The meeting on 24 June was the subject of a subsequent email report from Mr Stephenson dated 27 June 2003. Under the heading Risk Control Measures, there were a number of important entries:

"1. Separation

It is important to reduce the affect of a fire in any area of the NMU building by creating reliable fire separation between the main building areas. For this project this means installing a fire door between the new area and the existing popcorn packaging hall.

The door should be at least 1 hr fire rated. The door should operate manually and automatically. Ideally the automatic function would be tied to the site fire detection system, less preferably by a fusible link.

I understand the bucket elevator conveyor opening between rooms will be sealed as tightly as possible – to reduce potential fire/smoke spread through this opening…

6. Fire protection -Ansul./CO2

You plan to relocate the existing Ansul chemical fire suppression systems from Leeds to this site. These units are fairly good and will provide a useful first line fire control system. I understand a "fogtec" system was also proposed – in my view this doesn't offer good value.

The existing hoppers for finished cooked products are equipped with normally flooding CO2 systems. Presumably this system is in place because occasionally hot or burning popcorn enters the hoppers and results in a fire. We need some kind of control to prevent this occurrence and I think there are 2 options:

a. Continue with the existing CO2 units and relocate them. They are reasonable units for fire control but CO2 presents life safety concerns and you need good controls to prevent harm to people in or around the hoppers. Also annual servicing is needed.

Or

b. Remove the CO2 systems and run a water pipe off the internal hose system to a sprinkler at the top of the hopper. The sprinkler could be a closed head type with simple valve and flow alarm. This approach may be cheaper – and requires almost no annual maintenance. This assumes at least 1.5 bar pressure is normally available from a 32mm pipe.

The Ansul and CO2 systems are reasonably good, but we should not be under the impression that they will control all fires that could stem from these cookers. The only truly reliable way of controlling a fire at this site is through comprehensive sprinkler protection. Sprinkler protection is a significant investment and is beyond the scope of this project, it will be discussed again during the sites full loss prevention review."

71. Mr Caldicott was asked about Mr Stephenson's firm recommendation that sprinklers should be used. He said:

"… if you want to achieve highly protected risk status, then we really should have sprinklers in this site."

He said that it was recognised that it was unlikely that sprinklers could be installed for every single site that the claimants had. He said it was "raising a flag to the site operation management to say: look, you have got a risk here, a fire risk; if we had sprinklers we could feel a little more comfortable about it." There was then this exchange:

"Q: If I follow your answer, it is putting a flag to site to say: look, if we don't have sprinklers, then you must make sure that your controls are fully adequate to control the risk?

A: Yes, as best we can."

Later in his cross-examination, Mr Caldicott expressly accepted that the claimants – "Cadburys in the widest sense" - agreed to accept the risk of not having sprinklers (Transcript 5/181).

72. Mr Hamilton was also asked about Mr Stephenson's memo of 27 June. He confirmed that, following the advice, "we agreed to relocate or install CO2 units". He expressly accepted that he knew that this would not control all fires that might break out and that, as he put it "the only truly reliable way" of controlling fires was by way of sprinkler protection. He said that "our way of protecting against those instances was training and fire extinguishers."

D3: Mr Hamilton's Responses

73. Mr Hamilton confirmed that Mr Stephenson's memo was sent to him, and that the so-called 'site responses', shown in red on a subsequent copy of the document, were prepared by him personally.
74. In answer to Mr Stephenson's comments about the CO2 system, Mr Hamilton said that "site discussion has agreed that we will continue with the CO2 systems." He said in evidence that these site discussions involved him and Mr Eames and the operations team. It was therefore put to him that he must have satisfied himself that the CO2 system was adequate to control the risk of a fire in the hopper and the elevator. His answer was revealing:

"I think we recognised that there was a system installed at Leeds that was – appeared to be doing a job there, so was worthy of looking towards relocation. It is just – you have just got to be careful here about the insinuation that, you know, I have decided that this is a technically capable and competent system because – that is not in my remit. I don't have that knowledge to be able to make that judgment."

75. This answer, and the other evidence on this topic, highlighted in paragraphs 92 - 96 below, strongly suggested to me that, despite Mr Hamilton's protestations to the contrary, a decision was reached without any proper technical appraisal having been undertaken, and certainly without any reference to the defendant. It was suggested to him that he had simply proposed removing the existing CO2 systems from Leeds and putting them in Pontefract, and that was the extent of his consideration of the technical issues. Mr Hamilton agreed with that, but said that if somebody had told him that it was not going to work then that would have been worth discussing further. But he said that the Group Risk Department had advised in the memo that CO2 systems were "reasonably good", and so he had made a decision that he would continue with a CO2 system.
76. By reason of Mr Stephenson's advice, Mr Hamilton also knew, and accepted that he knew, that the CO2 system would not control all possible fires in the new 'oil pop' production area. He accepted that the only way that such protection could be provided was by the use of sprinklers. But his site response to Mr Stephenson's recommendation of sprinklers was: "This is not in the scope of Project Kettle". He confirmed that this meant that he was not going to consider sprinkling the production area or any other part of the NMU, because it was outside the scope of the immediate project that he was implementing.
77. I find that this was a wholly inadequate response; what was needed was an assessment of the risks before the project workscope was defined, rather than the preparation of a specification based on a 'minimum spend' (paragraph 58 above), and then a refusal to consider any recommendation which might add to the workscope, even if they provided the only way of truly controlling the risk. There was then this important exchange:

"Q: My suggestion to you is this: that that is completely the wrong way round of looking at it, isn't it? You should have assessed the risk, and I don't mean you personally for this purpose, but Monkhill should have assessed the risk and then decided what to do. Not decide what to do and then assess the risk?

A: Well we were able to assess the risk, because it wasn't a new operation to the business per se; we were already a manufacturer at Leeds.

Q: But it follows, doesn't it, that insofar as that was a risk, that was a risk that Cadbury knowingly decided to run?

A: That was a decision that wasn't made at my level."

78. I agree with the premise of Mr Dennys' questions. Mr Hamilton's approach was the wrong way round. And I disagree with Mr Hamilton's last answer: the evidence was that his contemporaneous site response meant that nobody gave any further thought to the possibility of sprinklers for another year.

D4: The Capex Application

79. The scope of Project Kettle was defined by the application that Mr Hamilton made to the claimants' Board, known as the Capex [Capital Expenditure] Application. The application was made on 10 June 2003. It appears that some of the relevant people, including Mr Hamilton, signed off the Capex Application as approved, before the documentation which made up the full application had even been completed. For example, Mr Hamilton signed it off on 13 June 2003, which was before the meeting with Messers Stephenson and Caldicott on 24 June and before Mr Stephenson's advice of 27 June (which was to be included in the final Capex Application documentation). Mr Hamilton indicated that it was not unusual for some of the key stakeholders to sign off before the documents were complete. This stuck me as a rather slack approach; it certainly means that the signatories cannot be taken to have approved every part of the documentation that made up the Capex Application, given that many of them would not have had all the documentation at the time that they signed.
80. The objective of the Capex Application was stated to be:

"Facilitate the transfer of Cinema Popcorn production from the current Leeds site into the Pontefract site. This will complete the transfer of all popcorn manufacturing from Leeds to Pontefract. This case is proposed as it delivers significant reduced overall costs in the manufacturing of cinema popcorn. Monkhill will have the best and most flexible volume popcorn manufacturing facility in the UK with the capability to provide both oil and air pop varieties in line with customer requirements."

81. Mr Hamilton prepared the document headed 'ACE Forms', which made up a part of the Capex Application, in early June. It stated:

"CS Group Risk Management have reviewed the proposed installation and fire prevention and containment measures. They have confirmed the investment and equipment included as part of this case fully meet the objective of minimising the fire risk that the introduction of oil popping to the Pontefract site undoubtedly brings. The local Fire Officers are also being consulted for the input and advice."

At the time that they were written, these statements were quite wrong, because there had not even been a meeting with Group Risk. The document appeared to be anticipating what Mr Hamilton hoped that the Group Risk Department would say. Furthermore, for the reasons already noted, even when they reported, Group Risk did not agree that the proposals "minimised the fire risk". Mr Stephenson's memo of 27 June, with the site responses, was however included in the Capex Application by the time that it was signed off by the other stakeholders on 8 July.

82. Also included in the Capex Application was a quotation from Wormald Fire Systems, the company who supplied the fire systems for the factory in Leeds, dated 6 June 2003. That quotation offered to redesign, modify and extend the existing wet chemical system above the pans for £21,800 and, in the alternative, offered the Fogtec water mist system for £32,000. That was the water-based alternative which Mr Stephenson decided against (paragraph 70 above).
83. The Wormald quotation noted that there was an existing CO2 system that protected the enclosed hopper and elevator which the claimants might wish to retain, and made plain that no additional work or materials was included in the quotation. Mr Hamilton confirmed that Wormald had not allowed anything in their quotation for the CO2 system, because the elevators were being relocated from Leeds to Pontefract and they already had CO2 systems fitted and configured within them. Mr Hamilton said it was his judgment at the time that "we will move all as is and install as is…the existing CO2 systems on the conveyors that were being relocated." That provides further support for the view that Mr Hamilton saw this as a straightforward replacement exercise and nothing more (see paragraph 75 above).
84. Mr Booth, one of the claimants' senior project engineering managers, gave more detail about the Capex process. He made clear that the decision-makers would not necessarily read every word of the application, but would be generally familiar with the documents. As to the Stephenson memo of 27 June 2003 (paragraph 70), he said that the risk from spread of fire was highlighted, at least in the round, and that meant that it was a concern. But he said that, provided that the Group Risk Department were happy with the installation, then the scheme would be approved. He said the decision-makers would have assumed that the 'good control measures' to which Mr Stephenson referred in his memo were somewhere within the plans of Project Kettle. As to the claimants' internal guidance documents (which were many and varied, and which included the recommendation of 4 hour fire segregation), he said that again the Board would rely on the Group Risk Department to have a far greater understanding of such documents and their implications.
85. On the specific issue of sprinklers, which the Cadbury's guidance documents said should be provided for all 'significant' projects, it was Mr Booth's view that Project Kettle would not count as significant, although he accepted that there was some fire risk. I find that this was an erroneous view, given the fact that the claim now, following the failure of a critical component of Project Kettle, is for £110 million odd; even the claimants could not absorb such losses lightly. The error appears to have arisen because, for the claimants, what determined 'significance' was the value of the proposed works, not the value of the building and equipment at risk if those works were not properly specified or designed. This equation between cost and significance was an elementary error, which Mr Caldicott expressly accepted in cross-examination (Transcript Day 5/174)^[7]. Mr Booth also referred to the importance of the Fire Officer's views, a matter which, as noted at paragraph 81 above, had been anticipated by Mr Hamilton, although not yet progressed.
86. Also on the subject of sprinklers, Mr Booth accepted that the Project Kettle Capex Application was signed off without there being any financial provision for the provision of sprinkler protection at the NMU. Again, he said that the Board was relying heavily on the Group Risk Department for their advice as to how the project should proceed, and the regular risk reviews of sites that were part of the Group Risk system. Thus, he said, the Board would expect the reference in Mr Stephenson's advice about sealing the bucket elevator opening (which had been responded to by Mr Hamilton with the promise that "the opening for the conveyor will be sealed to the edges of the conveyor and fully comply with this requirement"), would be addressed during the course of Project Kettle and checked (by way of risk review) thereafter. He agreed that those at the site, which meant in particular Mr Hamilton, Mr Sanders and Mr Eames, would have ownership of that obligation.
87. Finally, in relation to the CO2 system generally, Mr Booth confirmed his basic understanding that there was an existing CO2 system in Leeds which was simply going to be transferred over to Pontefract. He expressly confirmed that it was on that basis that Project Kettle was approved on 8 July by the Capex Board.
88. At the time that the Capex Application was signed off, Project Kettle was also discussed at the Pre-Board Meeting. The minutes refer to improved efficiency and proven technology. But there is also this warning: "H and S raised as a concern, not thought to be a problem given fire walls etc". Again, it seems to me that, even at this high level, the question of fire separation (or "fire walls", as they were here called), and the assumption that was made that there was effective fire-stopping and compartmentation around the area for which 'oil pop' production was proposed, was considered to be of critical importance to the proposals as a whole. This erroneous assumption gave rise to the conclusion that there was no health and safety issue; it follows that, if the claimants had bothered to check the as-built drawings, they would have seen that the assumption was wrong, and that therefore there was a major health and safety issue.

D5: The Commencement of Project Kettle and the Choice of the CO2 System

89. The chronology of Project Kettle was signposted by Mr Hamilton's regular notes, dated every two or three weeks. These were usually, but not always, notes of meetings at which Project Kettle was discussed with his team. The first such meeting was on 27 June 2003. The notes confirmed that a 1 hour fire rated roller shutter door was required for the wall into the Packaging Hall. Mr Barker, who attended the meeting, was asked whether the assumption was that the walls enclosing the proposed manufacturing area provided a 1 hour fire rated enclosure. He said "at the time of this meeting, I do not think it was known", an important admission, given the assumption that had been made to that effect in Mr Stephenson's memo of the very same day. If it was not known, why had Mr Stephenson been told at the meeting on 24 June that it was?
90. At the second Project Kettle meeting on 4 July 2003, there was a lengthy entry about the CO2 system. It read:

"CO2 can be used in local proximity to operators, but it does need to be well contained. So is an option for the buffer hopper on the packing hall gantry. Other options are the R102 chemical system or water. The R102 system best suits surface fires, so possibly not ideal and water is a mess!"

Mr Hamilton confirmed that this information could have either come from Wormald or from the claimants' Group Risk representatives. There had been no contact with the defendant at this stage. He said he knew that the CO2 system needed to be contained, in part so as to enable the gas to be retained for long enough within the elevator/hopper to put out the fire, and in part because it was a danger to other people.

91. At the meeting of the Project Kettle team on 14 July 2003, Mr Hamilton noted:

"2-off CO2 systems will be costed up to extinguish any fires as they are conveyed through the process. (Bucket Elevators and Hoppers)

Can we afford all of this?? Is it overkill??

Need to understand more about the current system."

Mr Hamilton's suggestion that the claimants might be over-spending, by installing a system costing a few thousand pounds, in order to protect building and equipment now said to be worth £110 million, is unfathomable. It betrays an absence of understanding of what the fire risks really were, and restates the false cost=significance proposition. At this point, in early July 2003, the decision appears to have been made to install a CO2 suppression system for the elevator and the hopper. Given that, as Mr Beaumont (one of the defendant's designers) said, CO2 systems were rare and made up about 5% of all fire suppression systems, and were less and less used, it is worth trying to identify how that decision had come to be made by the claimants (and Mr Hamilton in particular).

92. The first, if not primary, motivation seems to be the claimants' view that they already operated with an existing CO2 system in Leeds, and that all they wanted to do was to replicate that in Pontefract. That was what Mr Barker said in terms. It was also what Mr Hamilton said: "the system [in Leeds] had been installed and was operational. So on that basis, I could only have assumed that it was an appropriate system."
93. Secondly, there was the suggestion that, by early July, Mr Hamilton had satisfied himself that a CO2 system was appropriate. Some of his evidence suggested that that was why the CO2 system had been chosen, although when it came to it, in cross-examination, he indicated that the decision was because of the advice that he had received from others. At one point he said "I didn't choose any system. I took advice and installed what was recommended." When asked from whom he had taken advice, he referred to Wormald and the defendant, but of course the defendant had not even been approached at this stage. By the time the defendant was involved for the first time, the replacement of the CO2 system had been agreed and the limits on the expenditure on Project Kettle had been fixed by the Capex Application.
94. Thus, when Mr Hamilton said that he had taken the advice of others, the only people other than Wormald (whose quotation barely referred to the CO2 system at all) who could have given such advice were the Group Risk representatives, Mr Caldicott and Mr Stephenson. But as we saw from Mr Stephenson's memo of 27 June, they did not endorse the system in ringing tones; they were described simply as "reasonable units". It was also made plain that although they are "reasonably good", they will not control all fires.
95. It is difficult not to conclude, therefore, that Mr Hamilton was happy for the CO2 system to be replicated because that was what he had always envisaged; because neither Wormald nor the Group Risk Department had positively advised against it; and because any other system was going to be more expensive and therefore outside the Capex Application. In essence, his original answer, to the effect that the CO2 system was chosen because it was already working at Leeds, was, I find, the principal reason why it was chosen for Pontefract, and had been chosen before the defendant was first contacted.
96. If there was any doubt about this, it is confirmed by the evidence of Mr McQuaid, the representative of the defendant first involved in setting up the contract, who said:

"What we did is we replicated what they had in the Leeds factory, in terms of the way that the system operated on a single knock process with a heat probe in the duct itself…"

He went on to explain how, as a matter of detail, the system for Pontefract replaced that which had existed at Leeds, because that is what the claimants had asked for.

D6: The Involvement of the Defendant

97. Mr McQuaid became involved on 8 August 2003, when he met Mr Hamilton at the Leeds factory. It appears that, shortly before that date, the defendant had been contacted by Mr Hamilton to quote for the wet chemical system and the CO2 suppression system for the C and D Line elevators and hoppers. The reason why the defendant was being involved rather than Wormald seems to have been because the defendant had previously supplied the fire protection to the whole of the Pontefract site.
98. Mr Hamilton could not remember the specifics of his meeting at Leeds on 8 August. He said that he thought he would have showed Mr McQuaid the lines where the CO2 systems were installed, but that was about all that he could say. He could not remember the specifics of the conversation and he did not recall if there were any drawings or other information made available to Mr McQuaid on that occasion. In re-examination Mr Hamilton said that, in those circumstances, "you go on autopilot to some extent. You are exploring information that will take you towards a quotation."
99. Mr Barker explained that the reason that Mr McQuaid had been invited to Leeds was so that he could supply a similar quotation for a similar system to that which existed at Leeds. That only confirms the point I have made above, to the effect that all the claimants wanted to do was replicate what they had in Leeds; indeed, Mr Barker said in terms that what he was looking for was for the existing system to be replicated at Pontefract, and that was confirmed by Mr McQuaid.
100. Mr McQuaid recalled seeing just one conveyor system at Leeds, but he did not think that it was working at the time of his visit because it was Friday afternoon. He was not aware that it had plastic buckets held on plastic chains; he did not recall being shown the inside of the elevator or the hopper. He described some of the other machinery that he saw. He said he assumed that the elevator and hopper were enclosed "because to put the CO2 system in there, the enclosure has to be reasonably air tight to contain the gas to extinguish the fire." He was to tell Mr Flavell, one of the defendant's designers, that he thought the system was enclosed. He went on to say that the defendant then went ahead to replicate what the claimants had in Leeds. This encompassed a 'single knock' process (two sensors, either of which could trigger the suppression system) with a heat probe in the duct and another in the hopper. The probes were set at 100 °C at Leeds "so all we did was replicate what they had in Leeds on the conveyor in Pontefract. So the decision to put that probe in there, the type of probe and the operation temperature of that probe was my decision, based on what they had at Leeds."
101. Immediately following the meeting on 8 August, Mr Hamilton sent Mr McQuaid a layout drawing. This was a copy of the same drawing which Mr Barker had prepared 5 or 6 weeks before for the meeting on 24 June. It simply showed Production Area B with the new machinery in it, together with the elevator projecting through the wall out of Production Area B and into the Packaging Hall. There were no dimensions clearly shown. Mr McQuaid said that the drawing was "something to pass on to my colleagues, in terms of looking at it for the requirement for the extinguishing systems." He passed it on to Mr Potts, the defendants' design manager.
102. It seems that the drawing was also passed to Mr Flavell, a product manager for the defendant, based in Scotland. He was looking at the wet chemical side of the system. The drawing also found its way to Mr Beaumont of the defendant, who had worked as a designer in Leeds and had then moved to the Manchester office. Mr Flavell confirmed that his involvement at this point was in relation to the wet chemical system and in particular the Ansul R102 system which the defendant did not actually provide.
103. There was a further meeting of the Project Kettle team on 14 August 2003. This referred to the fact that the defendant had been contacted for an alternative quote. It also said that Mr Barker would be dealing with the defendant during Mr Hamilton's imminent holiday. As a result, it was Mr Barker who met Mr McQuaid when he visited the Pontefract site on 19 August. Mr Barker had no recollection of this meeting, or the drawing which he sent to the defendant after the meeting, which again showed the proposed layout for the 'oil pop' production process.
104. Mr McQuaid recalled discussing with Mr Barker the claimants' requirements for the installation and commissioning of the CO2 system at Pontefract. Although it was suggested to him that there was a discussion about smoke detectors in the ceiling of the 'oil pop' area, he thought that that was later in the story, and it did indeed appear that the discussion about smoke detectors occurred in September, as Mr McQuaid had recollected. Mr McQuaid also recalled that, on 19 August, there was some mechanical equipment in the area and he and Mr Barker looked at the detection system on the ceiling itself. He confirmed that Production Area B was 'a long way off' being equipped and there was nothing else there to assist as to what the elevators and hoppers would be like when they were installed.
105. On 20 August 2003, Mr Potts of the defendant sent Mr Flavell, the designer, copies of the drawings to which I have previously referred. Mr Flavell confirmed in his evidence that at this stage he was assisting with the design because of his specialist knowledge. He said he could not recall the specifics of any discussion that he might have had about the proposal. As for the drawings, he said there were enough to show where the location of the machinery was going to be and he might have scaled off them. Importantly, although he said expressly in cross-examination that these were insufficient drawings on which to base a design, it appears that this was precisely what the defendant went on to do.

D7: The Defendant's Initial Design and Specification

106. The next thing that happened was the production by the defendant of their initial design and specification to go with their quotation for the work. The specification was prepared by Mr Flavell. He explained that parts of the specification were in standard form, and parts were expressly referable to this particular project. He could not remember if, by the time he prepared the specification at the end of August, he had been told that the elevators had within them a system of polypropylene buckets and chains. He said that he based the design on an enclosed volume and would have assumed that the enclosure would be checked in due course by the installation and commissioning teams.
107. The specification contained the following opening passages:

"Client: Monkhill

Hazard: Elevator and Hopper arrangements

Date: 28/8/03

PROPOSAL

For the protection of the above hazard, we propose the design, supply, delivery, installation, testing and commissioning of a CO2 Fire Fighting System.

ADT Fire and Security CO2 fire fighting systems are designed, manufactured and installed to suit the specific requirements of the risks to be protected and comply generally with the requirements of BS5306 Part 4.

CO2 extinguishes fires by reducing the oxygen to a level that does not support combustion and by cooling."

It was agreed that these proposals were in the defendant's standard form of wording and typically used in the defendant's specifications.

108. There was then what was called a 'Summary of Calculations' identifying the volume of the areas to be protected and setting out the quantity of CO2 required to protect it. Thus, each elevator was said to be 10 cubic metres in size requiring 22.5 kg of CO2. Each of the hoppers was said to be 1 cubic metre in size. It seems that these volume figures came from scaling off the general arrangement drawings sent out by Mr Hamilton and Mr Barker (paragraphs 101 and 103 above).
109. The specification went on to say:

"THE SYSTEM

The CO2 Fire Fighting System is designed, manufactured and installed to suit the specific requirements of the hazard to be protected.

The complete system consists of one or more storage containers coupled to a system of pipework of discharge nozzles specifically sized using a hydraulic flow calculation program designed to ensure discharge of the design quantity within a specified period of time…

TOTAL FLOODING

Total Flooding systems are generally for the protection of enclosed hazards. Nozzles would be designed to flood the entire hazard enclosure with CO2 to develop an even concentration.

ANCILLARIES…

Discharge Nozzles

An appropriate number of discharge nozzles will be provided to ensure adequate distribution of CO2 within the hazard area…

HOLDING TIME – Total Flooding Systems

The design concentration of CO2 should be held long enough to ensure complete extinguishment. The holding time will vary with the gas tightness of the enclosure. The required holding time may vary between one and in excess of twenty minutes depending on the hazard involved.

Openings or ventilation ducts/forced air handling must be closed or shut down automatically before or in conjunction with the discharge.

All doors should open outwards and be fitted with self closing devices. Doors and windows should be in the closed position prior to the CO2 system discharging. Any penetrations through the walls, ceiling and floor of the protected area(s) for cables, pipes etc should be suitably sealed.

It is difficult to calculate accurately the leakage rate from any particular enclosure without full scale discharge tests, however a technique has been developed known as the Room Integrity or Door Fan Pressurisation Test which can provide an enclosure retention time prediction.

Where self-closing doors, dampers or shutters are required to reduce the loss of CO2, the client shall supply and install these items and ADT will install, at an additional cost if not included within our proposal, the necessary pressure-operated trips to control their release…"

110. Finally, the specification dealt with testing, commissioning and maintenance. Under the heading 'Tests' it said:

"On completion of the order, a test will be made to prove the correct function of the equipment installed. This would exclude any gas discharge. All parties concerned are invited to attend and witness the tests. Acceptance certifications will be signed as a record that the installation has met with your satisfaction. All such tests will be carried out during normal working hours."

111. Mr Flavell was cross-examined by Mr ter Haar, by reference to British Standard 5306 (the BS referred to in the Specification), on the basis that the defendant failed to comply with various aspects of it. Certainly, Mr Flavell's answers confirmed that working documents, including drawings of the distribution system, were not produced. Furthermore, drawings showing the location and construction of protecting enclosure walls and partitions were also not produced, although of course it is an issue for me as to who should have provided such drawings. Mr Flavell agreed that, in these circumstances, what was required was a detailed drawing of the elevator which he never saw. Neither did he see a full height or schematic diagram of either the elevator or the hopper. He also agreed that the working documents should have identified the relevant hazards, which included the plastic buckets and chains of the elevator.
112. In addition, Mr Flavell confirmed that he assumed that somebody at the defendant would check the enclosure to see that it conformed with the design assumptions. He had assumed that this check was carried out. In addition, at the commissioning stage, he said that he would have expected somebody to check the enclosure – in this case the elevator and the hopper – to seal any significant air leaks, if it was practical to do so. Finally, Mr Flavell accepted that for the holding times to be effective^[8], there had to be some form of enclosure, and that the importance of closing any openings was so great that he would have expected that to be drawn to the claimants' attention by whoever at the defendant was handling the project.
113. None of the things which Mr Flavell thought should be checked later, or which should be the subject of express warnings, actually materialised. Instead, a design was produced on very limited information, and was never the subject of review or subsequent checking. In this way, the claimants' haphazard approach at the outset was not picked up by the defendant, and was compounded by the defendant's own rather cavalier attitude to what they were being asked to do. There was, however, one important difference of scale: the claimants' failure to do things properly affected both the 'oil pop' production area and the whole of the NMU; the defendant's failure was limited to the small part of the manufacturing process for which they were asked to design a fire protection scheme.

D8: The Defendant's Quotation

114. On 28 August 2003, the defendant sent the claimants a quotation. The quotation was for £9,009 "to supply, install and commission an ADT extinguishing system" for the C and D Lines^[9]. It went on:

"This Quotation is based on the attached Proposal & General Specification Notes, and where appropriate Enquiry Specification and Tender Drawings provided. Unless and until other terms and conditions have been agreed in writing ADT Fire and Security Standard Terms and Conditions as detailed in our Commercial Agreement shall apply."

It is common ground that the specification referred to at paragraphs 107 - 110 above was sent with this quotation. It is also common ground that the defendant's standard terms and conditions – which were normally printed on the back of their Commercial Agreement - were neither enclosed with nor attached to the quotation. Neither was any Commercial Agreement sent with the quotation.

115. It is unnecessary to set out all of the standard terms and conditions which the defendant used at that time. The critical one for present purposes is the provision entitled 'Limitation of Liability' which was in these terms:

"Only you know the value of your premises, its property and contents. We are not and cannot be an insurer of your premises and its contents and our charges are in no way related to their value.

The fire and security industry is unique having regard to the relatively low cost of the services and the high values which can be at risk. For this reason, we limit the amount of our liability and the most we will be prepared to pay for any loss will be no more than twenty times the amount of the yearly service charge (excluding VAT) you are paying for the system at the time of the loss. You should also read condition 6 over the page.

If you would like to increase this limit, you will have to pay an extra charge so we can arrange the appropriate insurance cover for the extra liability, the level of which has to be agreed. In such case, you will have to give us any information our insurers may need."

116. Clauses 4 and 6 were lengthy clauses which, amongst other things, repeated the limitation of liability at 20 times the amount of the yearly service charge and also included a number of other exclusion provisions. They were in these terms:

"4 The purpose of the system

4.1 The system is designed to reduce the risks of loss or damage to your premises so far as this can be done by the use of this type of equipment. However we do not guarantee that the system cannot be removed, tampered with or made to stop working by you or by any unauthorised person. If this happens, we are not responsible for any losses you may suffer directly or indirectly.

4.2 We do not guarantee to you that:

a) particular losses or injuries will be prevented by using the system; or

b) that the system will work continuously without errors, in particular where interruptions or errors are due to something beyond our reasonable control.

4.3 Our products are designed and manufactured to high standards. However, even our products, like all mechanical and electronic devices, can develop faults.

4.4 We do not know the value of your premises or its contents and the purpose of this agreement is not to act as insurer of your premises or your contents…

6 Our liability to you

…

6.3 We accept that we must make sure that the system is of satisfactory quality, that it is suitable for the purpose in condition 4 and that the system will meet with the description provided before it was installed. We confirm that we are entitled to sell the system to you.

6.4

a) As well as the responsibility which we accept in condition 6.3, we accept responsibility for death or personal injury caused by our failure to take reasonable care or to use reasonable skill. For all other liability, the most we will pay for loss of any kind will be no more than 20 times the amount of the yearly service charge (excluding VAT) which you are paying at the time of the loss. We have worked out the amount of the yearly service charge based on this limit of liability.

b) If we have only sold and/or installed the system in circumstances where there is no yearly service charge to be paid by you, out liability in condition 6.4(a) will be no more than 20 times the amount of the yearly service charge (excluding VAT) that we would have secured for a system of a similar size.

c) If this agreement covers more than one system, the calculation of the limit of liability in condition 6.4(a) will be based on the yearly service charge for the system giving rise to the liability.

6.5 Apart from those responsibilities accepted by us under condition 6.3 and for death and personal injury under condition 6.4(a), in all other situations we do not accept responsibility for any indirect loss which depends on us having special knowledge of your affairs which we would not normally know, even if the loss is due to our fault."

At paragraph 6.6 there were 9 specific clauses purporting to exclude liability for various kinds of loss.

D9: The Claimants' Purchase Order

117. The evidence was that the quotation was provided to Mr Barker by Mr Hamilton, when Mr Hamilton asked Mr Barker to place an order for the CO2 suppression and wet chemical systems. The orders were entered on the claimants' SAP system. That is a matter explored in greater detail in Section E2 below. Mr Barker said that he did not read the specification before placing an order.
118. Mr Barker then sent the defendant a Purchase Order, dated 3 September 2003. It contained a vendor number, 107378. Mr Barker said that he created the Purchase Order by opening up a screen on his computer and entering the defendant's details. The vendor number then came up automatically on the screen. He said:

"When you place an order, I believe one of the first things they ask for is the vendor number. So once you put the vendor number in, if you have selected the right vendor number, then you will go straight into the ordering sequence."

119. The Purchase Order was divided into two parts, the first part dealing with the wet chemical system, and the second part dealing with the CO2 suppression system. That part of the Purchase Order said:

"CO2 Fire suppression system. As per your quotation dated 28/8/03.

Systems to be fitted to:-

2 Elevators

3 Hoppers 50% Payment"

The delivery date was 25 September 2003. The payments were to be made in two stages of £4,504.50 each. Finally, at the bottom of the Purchase Order, the following printed words could be found:

"This is subject to Monkhill standard terms & conditions already supplied, unless otherwise agreed. Additional copies available on request."

It is common ground that these terms and conditions were not included with the Purchase Order, which assumed that they had been "already supplied".

120. Again it is unnecessary to set out all of the second claimants' standard terms and conditions in use at that time. The following are of particular relevance:

"3. QUALITY AND DEFECTS

a) All Goods supplied and/or Services carried out shall be of good quality and subject to the Buyer's approval and in particular must meet the governing Specification and CTB Standards but without limitation also be as required by law in respect of title, quantity, quality, purpose or description…

4. INDEMNITY

The Seller will indemnify the Buyer against the following

a) Any loss or damage or injury cost expense or liability whatsoever and whensoever arising from or in connection with the supply of the Goods and/or Services for which the Buyer may be liable to third parties due to any negligence, breach of statutory duty, breach of contract or other act or omission of the Seller…

e) Any loss damage cost expense or liability sustained by the Buyer as a result of the failure of the Seller to perform the Services or supply the Goods in accordance with the Contract or due to any negligence, breach of statutory duty, breach of contract or other act or omission of the Seller…

13. LEGAL REQUIREMENTS

a) The Seller warrants that the design, construction and quality of any of the Goods or Services to be supplied under the Contract comply in all respects with all statutes, statutory rules, orders and regulations which may be in force at the time and further that the Services and the sale or use of the Goods by the Buyer will not infringe any Intellectual Property Rights of any third party. The Seller undertakes to indemnify the Buyer against any loss, damage, liability, costs or expenses which the Buyer may suffer or incur by reason of any breaches of the said warranties… "

D10: The Meeting With The Fire Officer

121. On 5 September 2003, the Fire Officer, Mr Kevin Kemp, visited Monkhill and met Mr Hamilton. The purpose of the visit was nowhere recorded, and neither man had any particular recollection of the meeting. It was surprising that there were no notes of this meeting, nor any follow-up letters from the claimants to the Fire Service in relation to the matters discussed, because the claimants had, in their Capex Application, already acknowledged the importance of obtaining the Fire Officer's agreement to what they proposed (paragraph 81 above). Moreover, as noted below, although the claimants, and Mr Hamilton in particular, were later to claim that Mr Kemp agreed to a whole range of particular matters which have a direct bearing on the fire two years later, and therefore on the issues in this case, the evidence suggests that, on the contrary, there was no detailed discussion of Project Kettle at the meeting on 5 September.
122. Mr Kemp said that the claimants had been asked for a copy of their Fire Certificate but they did not have a copy. Later, the out-dated Fire Certificate, to which I have referred in Section C4 above, was produced. Mr Kemp said that he made two or three inspections like this every day and he had no reason to remember this visit in particular. He could not recall any particular part of the interior of the NMU. It was common ground that there was very little equipment in the proposed 'oil pop' production area, although the canopies and extraction unit for the C and D Lines were apparently in place.
123. Mr Hamilton also had no specific recollection of the meeting with Mr Kemp. He said that he thought he would have explained to Mr Kemp what the process was going to be, but he accepted that the equipment had not been moved from Leeds so that there was nothing physical at the Monkhill premises to alert Mr Kemp to the fact that the 'oil pop' production was going to be moved there. Hence Mr Hamilton's repeated insistence that, although he had no recollection of it, he thought that he "would have explained" that process to Mr Kemp.
124. In my judgment, based on the material before me, I consider that it is most unlikely that Mr Hamilton explained the new process to Mr Kemp at all, or in anything other than the most general terms. I say that, not only because of the absence of any of the relevant production equipment at Monkhill that day, but also because of the nature of the documents which the claimants did generate after Mr Kemp's visit. Later that day, Michelle Irving, the claimants' health and safety manager, sent Mr Hamilton an email dealing with the missing Fire Certificate. It went on to record what Mr Hamilton had said to her about his discussion with Mr Kemp:

"From general observation, he was happy with our fire detection system, internal emergency lighting, testing procedures and employee smoking policy.

Areas which require attention are:

Improved signage

Appropriate extinguishers

Fire doors being wedged open

Travelling distances

Inner rooms (smoke detection)

External lighting survey"

There is nothing in that email which suggests that Mr Kemp had been asked to consider, or did consider, what the fire safety implications might be of a process – involving naked flames - that had yet to be installed at the NMU. All of the matters listed related to the existing operations.

125. That view is confirmed by Mr Hamilton's response on 8 September 2003. This recorded that Mr Kemp had found a copy of the Fire Certificate relating to the old NMU building. Unsurprisingly, therefore, Mr Hamilton noted that the certificate had been granted "without the compartmentalisation Kevin highlighted on the basis of an above-average fire detection system." He went on:

"Once we have the Certificate we can get into the detail of the proposed changes. This again will be done through Wakefield.

Kevin's view is that they will not be so sympathetic towards the compartments we have in the latest layout, but we will need to wait and see."

126. I find that Mr Hamilton's memo is once again only consistent with a situation in which, during the conversation with Mr Kemp on 5 September 2003, he had not dealt with the detail of the proposed changes involved in the transfer of the 'oil pop' production process (because he said he would not "get into the detail" of them until the claimants had the Fire Certificate), and which, even if there had been an outline discussion, led the Fire Officer to conclude that better compartmentalisation was going to be required.
127. Mr Hamilton's attempt to make the best of this exchange for the claimants led him to give a number of different explanations for these contemporaneous documents, (all based on a conversation that he does not remember). These included the suggestion that the compartmentalisation did not relate to the 'oil pop' area, but to the chocolate rooms in the other part of the NMU. It certainly appears from a comparison of the drawings that the internal layout of the old nut-packing area had been broken up into smaller areas which were not shown on the drawing that went with the Fire Certificate, but this explanation only confirms my conclusion that Mr Kemp was principally concerned with the building as it then existed, not what might be done in the future when the 'oil pop' process was brought in. It was therefore suggested to Mr Hamilton that Mr Kemp could not have been told anything about the major changes that were going to be made in the production area, because he was only concerned about rearrangements in the nut packing area. Although Mr Hamilton denied that, it seems to me to be the only plausible explanation.
128. Mr Kemp's evidence was that he would have been concerned about the fire risks if he had been told about the detail of the 'oil pop' process. In my view, that is overwhelmingly likely, given that all of the witnesses, including the claimants' own^[10], accepted that the 'oil pop' process, which relied on a naked flame, was a process that brought with it the clear and obvious risk of fire. The absence of any documents emanating from the claimants, and sent to the Fire Service dealing with the proposed changes, was further evidence that it was not a matter that was discussed (or not in any detail) on 5 September 2003.
129. There is another important part of this aspect of the story. Mr Hamilton accepted that his email of 8 September 2003, with its reference to 'getting into the detail of the proposed changes once the claimants had got the Fire Certificate', meant that, even on his case, the Fire Officer was still expecting to receive further details from the claimants, particularly in respect of the proposed changes in Production Area B. In other words, the ball was still firmly in the claimants' court. Moreover, that was common sense, given that the claimants were in control of their own oil pop system and they needed a Fire Certificate which gave it the necessary clean bill of health.
130. However, there was no further liaison at all between the claimants, and Mr Hamilton in particular, and the Fire Officer. Mr Hamilton made no further attempt to contact the West Yorkshire Fire and Rescue Service ("WYFRS") during Project Kettle. Following the completion of the works in 2003, there was no amendment to the Fire Certificate and nothing which indicated that the new arrangements had been specifically considered, let alone approved, by a Fire Officer. Again, the claimants failed to take even the most basic steps for ensuring a safe system of work at the NMU.
131. This is very important because, a year later, in June 2004, the claimants' Group Risk Department produced a Property Risk Assessment of the NMU (see paragraphs 352-366 below). This again recommended a sprinkler system. The site response – drafted by Mr Hamilton - was that a site decision had been taken to control the risk of fire through fire prevention, protection and segregation "as agreed with local fire service". The same Risk Assessment made various recommendations as to improving the protection and operation of the oil popping area, and Mr Hamilton's site response was again to say that the operating procedures "were shared with the local fire service."
132. Mr Hamilton confirmed that these were both references to his unminuted, unrecorded discussion with Mr Kemp on 5 September 2003. I find that it was a plain misrepresentation of the position to say that the operating procedures had been shared with Mr Kemp (they were not even in place at the time), and that it was quite wrong to say that the CO2 suppression system or any other part of the fire protection in the oil pop area had been discussed, let alone agreed, with Mr Kemp. These matters were, in my view, to have a profound effect on the course of events leading up to the fire.

D11: The Defendant's Work

133. On 11 September 2003, Mr Flavell visited the premises in order to take measurements of the equipment and its location, in so far as it was installed at that time. Mr Flavell said that he had virtually no recollection of the site visit, but it was again common ground that neither of the elevators had been installed by then. That meant that it was also impossible to deal with the question of openings. Mr Barker said he did not recall any discussion about the details of the elevator on that occasion. Mr McQuaid confirmed that he could not recall the elevators being in place on that date. He thought the visit was to look specifically at the electrical system and that anything to do with the CO2 system was a matter for his colleagues.
134. The elevators were designed by a company called Gough Engineering. There was a drawing of a Gough elevator which at some stage was provided to the claimants, although it is unclear whether that drawing was ever sent on to the defendant. The drawing showed that there were three principal openings in the elevator: the in-feed at the start; the outlet for the popcorn for retail packaging, and the outlet for the cinema popcorn. It now appears that in addition there was a large mesh opening at the foot of the elevator, and some inspection hatches, none of which were shown on the drawing.
135. Mr Barker said that the Gough elevator was standard and was simply a later version of the one in Leeds. He thought that the new elevator was identical in the size of the buckets, the width, and the length. He said that, because the system had worked at Leeds "alarm bells were not ringing at Pontefract to say that we should change it in any way." The C Line conveyor at Leeds was modified and reused at Pontefract. Mr Barker confirmed that he had not at this stage sent the defendant the drawings of the elevator, although he had provided them with the information as to the width and height of the conveyor. He confirmed that the only information that the defendant was given was that the Gough elevator did not have obvious openings, except where the feed went into the elevator. He said that the defendant had access to the equipment at the Leeds site, and they could, if they asked, view the equipment that the claimants were going to install at Pontefract. I consider that this was a rather unhelpful stance, and did not explain why Mr Barker failed to send all the necessary information to the defendant without waiting to be asked for it.
136. Mr Flavell carried out the initial design. It appears that he assumed that the risk in the elevator and hopper was of a deep-seated fire, as defined in the British Standard, and carried out some calculations on this basis. However, Mr Flavell subsequently contacted Mr Beaumont, another designer with the defendant, to undertake some further work on the design, and he apparently concluded that the risk was of a surface fire. A deep-seated fire risk required a higher concentration of CO2, a concentration that, in this case, would last about 20 minutes. Mr Beaumont said that this was "difficult with machinery. The rules are really set around rooms. Machinery is going to be leaky." He went on to say: "It would be a little bit difficult to maintain it for 20 minutes, I imagine, but it would depend on the machine."
137. It was not always clear what effect this difference of approach had on the detailed calculations, or the specification and proposed location of the equipment to be provided as part of the CO2 system, particularly as Mr Beaumont said that, for the purposes of design, "you do not take holding time into account". However, in the light of Mr Beaumont's view that "something would eventually burn", and that in this case, that 'something' was the popcorn, it was difficult to see why Mr Flavell had not been right to assume a deep-seated fire, at least in the hopper. It was put to Mr Beaumont that, on this basis, there was a discrepancy between the two approaches. Mr Beaumont denied that, but expressly agreed that all the hydraulic calculations which he produced were based on a surface risk. The upshot of his explanation was that the defendant was trying to cover both risks.
138. Mr Beaumont also agreed that a designer should survey the enclosure to be protected, or at least examine drawings of the enclosure. He said that if the design engineer worked from drawings, he would need the installing engineer to check that his design assumptions were well-founded. As part of the same process, he said that the designer would want to establish the dimensions, geometry and layout of the enclosure and its surroundings so as to determine the location of the detectors, nozzles and so on, although he said that it was unnecessary to carry out too detailed an investigation into the robustness of the construction of the enclosure that was being protected because "generally you have a feel for these things".
139. But on the facts, Mr Beaumont undertook no inspection at all, had no information as to the detail of the Gough elevator or the buckets and chains within it, advised no-one that various assumptions would have to be subsequently checked, and had general layout drawings which Mr Flavell had already agreed were inadequate for design purposes (paragraph 105 above). In fact, Mr Beaumont said that he used measurements taken not from the drawings (because the only drawings that the defendant had did not contain what Mr Beaumont described as "the kind of information that I would want to receive"), but from the specification (i.e. the 10 cubic metres for the elevator and the 1 cubic metre for the hopper), which measurements had never been verified by anyone.
140. In my view, the patently inadequate information which the defendant had, and the difference of approach between its two designers which was its probable result, demonstrates a fundamental flaw at the heart of the defendant's work on this contract. It is no answer to say (as the defendant's closing submissions attempt to do, at paragraph 76) that the surface-fire assumption related to the risk in the elevator, and the deep-seated assumption to the risk in the hopper. There was nothing to say that the design of the system was intended to be or actually was different in those two areas: on the contrary, the evidence suggests that it was the same for both because the two different parts of the machinery had never been separately considered. For reasons that I will come to, I consider that the defendant's failure to appreciate just how significant was the risk of a deep-seated fire in the hopper (as opposed to the elevator) was an important part of this story.
141. At about the same time as the design of the CO2 system was being finalised, there was some further consideration of the heat sensors in the ceiling of the old Production Area B. Mr Barker had realised that they were unlikely to be effective, because the area was now going to be used for popcorn production. The defendant advised that the ceiling sensors were to be set at 90°C. The quote for this further work was sent by Mr McQuaid to Mr Barker on 29 September 2003. Mr Barker expressly accepted that, in this process, he was exercising some independent judgment. It was this exchange that Mr McQuaid correctly recalled happening in late September, rather than being a matter that was discussed with Mr Barker at the meeting on 19 August.
142. The installation of the CO2 system was carried out by Mr Taylor, who was a qualified electrician. As to the information that he had for this purpose, he had five sets of documents, namely:

a) A specification for a fire alarm system, prepared by the defendant and dated 17 September 2003 setting out some general, rather than technical details as to the panels and heat detectors. This specification had never been sent to the claimants;

b) A schedule of equipment prepared by the defendant;

c) A specification for the wet chemical and CO2 system which looked to be in slightly different form, and without the detail about volumes, to that of 28 August;

d) Mr Beaumont's calculations, referred to above;

e) Drawings, which were the various layout drawings to which reference has already been made and which, as Mr Taylor confirmed, would not provide the dimensions of the elevator and the hopper.

This was something of a ragbag of material, which (amongst other things) did not include the same version of the specification sent out on 28 August. Again, this does not suggest that this low-cost job was high on the defendant's list of priorities.

143. Mr Taylor confirmed that the choice of the heat detectors to be located in the hopper and the elevator had been made by somebody else, and was not his responsibility. This was the first CO2 installation he had ever carried out. The installation of the sensors, nozzles and pipes were carried out according to what Mr Barker described as 'a tight timetable' in October 2003. There are no records of the detailed work carried out and surprisingly, no as-built drawings. It appears that, by November, the 'oil pop' production was up and running, even though the CO2 system had not yet been commissioned, so was not yet in operation.
144. During the pre-commissioning period, when the claimants were producing popcorn without the CO2 system in operation, there was a fire in the hopper. This demonstrated the fire risks inherent in the popcorn production process, which would have been exacerbated by the fact that, at this time, the process was operating without any CO2 system in place at all. It also demonstrated that burning popcorn could go from the pans, through the elevator, into the hopper and out again, which was precisely what was to happen on 8 June 2005. The fire was put out by emptying the hopper and using a CO2 extinguisher to put out the burning popcorn on the factory floor. As a result, the CO2 canisters needed to be refilled.
145. The commissioning of the CO2 system was carried out shortly after 9 December. Mr Taylor was taken by Mr ter Haar through the commissioning manual. He confirmed that he had not re-measured the protected area, or confirmed that the quantity of gas supplied was adequate for the measured volume. He said that he had checked the system had been installed in accordance with the drawings but, as we have already seen, there were no drawings that showed the CO2 suppression system (as designed or as installed), and there were no marked-up or scaled drawings of the whole 'oil pop' popcorn process with the CO2 suppression system marked onto them. There were therefore no design drawings against which any meaningful checks could be performed. In addition, as to leakage areas, Mr Taylor said he would have expected the design team to have carried out any necessary survey and would assume that they had done their job. He said he would not know if any openings were part of the design factor in the design calculations. He said he would assume that whatever openings there were had already been factored into the design. He said he did not know that Mr Flavell had assumed that the integrity of the enclosure would be checked by the installation team.
146. As to the British Standard, Mr Taylor said that he was aware of the requirement to produce working documents and agreed that there were no documents which included a description of occupancies and the hazards being protected. He said that was a matter for the design team. He confirmed that he had done no visual check of the enclosure.
147. Mr Walker, another electrician employed by the defendant, was also involved in the commissioning. He confirmed that he was not asked to carry out any inspection of the enclosure of the conveyor and the hopper for air tightness, and said that was not something he would expect to do. He confirmed that his job was to make sure that, when the system was turned on, the electrics worked. He explained the single knock system (i.e. if either the sensor in the hopper, or the sensor in the elevator, was triggered, the whole CO2 system would be activated). At the commissioning stage he thought that one of the lines might have been working and one of the lines was not. In his statement he had referred to the hopper lids which he noted were lent up against the side of the hoppers. He agreed that if he had thought that this was anything to get worried about he would have drawn it to Mr Barker's attention, and he did not do so. The system was commissioned by Christmas 2003.

E: THE FORMATION OF THE CONTRACT

E1: The Issues

148. The general contractual issue between the parties was whether they contracted on the claimants' terms and conditions or, alternatively, on the defendant's terms and conditions or, in the further alternative, on the basis of a contract which incorporated both or neither sets of conditions. The critical dispute was the defendant's term purporting to limit to £13,781.60 what would otherwise be, on the claimants' case, a liability of about £110 million. In many ways, this dispute could be categorised as a straightforward 'battle of the forms', albeit one that is worth rather more than usual.
149. I propose to set out shortly the relevant principles governing this sort of dispute (Section E2 below). I then deal with the general background of the relationship between the parties (Section E3 below) and the specific background to this contract (Section E4 below). I set out my primary analysis of the formation of the contract in Sections E5 and E6 below. There is a summary at Section E7.

E2: Applicable Principles

a) Performance

150. During the course of this trial there was the faint suggestion that, by reason of the uncertainty over its formation, no contract had ever been concluded between the parties. It seems to me that, in all the circumstances of this case, that proposition is unsustainable. The Court of Appeal have repeatedly stressed that where, as here, a contract has been fully performed, the court is overwhelmingly likely to conclude there was a completed contract: see Pagnan v Feed Products [1987] 2 Lloyd's Rep. 601 and G Percy Trentham v Archital Luxfer [1993] 1 Lloyd's Rep. 25. In the latter case, Steyn LJ (as he then was) said:

"The fact that the transaction was performed on both sides will often make it unrealistic to argue that there was no intention to enter into legal relations. It will often be difficult to submit that the contract is void for vagueness or uncertainty. Specifically, the fact that the transaction is executed makes it easier to imply a term resolving any uncertainty, or alternatively, it may make it possible to treat a matter not finalised as inessential."

The correctness of this approach has recently been restated by the Supreme Court in RTS Flexible Systems Ltd v Molkerei Alois Muller Gmbh [2010] 1 WLR 753.

151. I am in no doubt that, in the present case, there was a mutual intention to create legal relations, and that a binding contract was concluded. The evidence set out at paragraphs 97 to 120 above admits of no other interpretation.

b) The 'Battle of the Forms'

152. Where, as also happened here, one set of terms and conditions is identified in a tender offer, and another is identified in a purchase order, there may be what has been called the 'battle of the forms'. It is the court's task in such cases to try and construe which, if any, is the operative set of conditions. The best-known case on this topic is Butler Machine Tool Co Limited v Ex-Cell-O Corporation (England) Limited [1979] 1 All E.R. 965. In that case Lord Denning said that the court's task was to:

"…look at all the documents passing between the parties and glean from them, or from the conduct of the parties, whether they have reached agreement on all material points – even though there may be differences between the forms and conditions printed on the back of them."

153. Later on in his judgment he said:

"There are yet other cases where the battle depends on the shots fired on both sides. There is a concluded contract but the forms vary. The terms and conditions of both parties are to be construed together. If they can be reconciled so as to give a harmonious result, all well and good. If the differences are irreconcilable, so that they are mutually contradictory, then the conflicting terms may have to be scrapped and replaced by a reasonable implication."

I observed in J Murphy & Sons Limited v Johnston Precast Limited [2008] EWHC 3024 (TCC), that this passage in Lord Denning's judgment indicated an approach which was not very different to that outlined by Steyn LJ in Archital Luxfer.

154. Butler is often cited as authority for the proposition that, where there is a battle of forms, the critical act may be the so-called firing of the last shot, the last letter or communication in the series. But some care is needed with that analysis. The court has to construe all the relevant communications. Thus, in Butler, the seller quoted on his terms and conditions and the buyer sent out an acknowledgement of order which referred to the buyer's terms and conditions. The seller signed the acknowledgement. The buyer's terms and conditions were subsequently held to apply, even though the seller wrote again, after signing the acknowledgement, in a letter which referred to his own terms and conditions. That later letter was found to be irrelevant because it referred simply to the price and identity of the machine in question, and did not, as a matter of construction, operate to incorporate the seller's terms back into the contract.
155. The 'battle of the forms' has been the subject of a recent decision of the Court of Appeal in Tekdata Interconnections Limited v Amphenol Limited [2009] EWCA Civ 1209; [2010] Lloyd's Law Rep. 357. In that case, the sellers quoted on their terms and conditions; the buyers generated a purchase order which stated that the purchase was on their terms and conditions; and when the sellers acknowledged the purchase order, they repeated that their own terms and conditions applied. At first instance, the judge said that the traditional view would be that the contract terms were those of the sellers, since their acknowledgement was the last shot and had expressly purported to incorporate their terms and conditions. However, he found, for various reasons, that other circumstances indicated that the parties intended that the buyers' terms should apply. This was, at least in part, because there was a commercial history in which the parties had contracted on the buyers' terms.
156. The decision was overturned in the Court of Appeal. It was held that, in a 'battle of forms' case, the traditional analysis based upon offer and acceptance had to be adopted, unless the documents passing between the parties and their conduct showed that their common intention was that some other terms were intended to prevail. The Court of Appeal held that it would be difficult to displace the traditional analysis, unless it could be said that there was a clear course of dealing between the parties. No such course of dealing had been proved in Tekdata.
157. Dyson LJ (as he then was) said at paragraph 25:

"In my judgment, it is not possible to lay down a general rule that will apply in all cases where there is a battle of the forms. It always depends on an assessment of what the parties must objectively be taken to have intended. But where the facts are no more complicated than that A makes an offer on its conditions and B accepts, that offer on its conditions, and, without more, performance follows, it seems to me that the correct analysis is what Longmore LJ has described as the "traditional offer and acceptance analysis" i.e. that there is a contract on B's conditions. I accept that this analysis is not without its difficulties, in circumstances of the kind to which Professor Treitel refers in the passage quoted at paragraph 20 above. But in the next sentence of that passage Professor Treitel adds 'for this reasons the cases described above are best regarded as exceptions to a general requirement of offer and acceptance'. I also accept the force of the criticisms made in Anson's Law of Contract, 28^th Edition. But the rules which govern the formation of contracts have been long established and they are grounded in the concepts of offer and acceptance. So long as that continues to be the case, it seems to me that the general rule should be that the traditional offer and acceptance analysis is to be applied in battle of the forms cases. That has the great merit of providing a degree of certainty which is both desirable and necessary in order to promote effective commercial relationships."

This clear, concise reasoning in support of the 'last shot' requirement, when seen in the context of all the communications between the parties, is binding on me and of direct relevance to the outcome of the contractual dispute in this case.

c) Contract Interpretation

158. In a case like this, the background to the contract is relevant to the proper interpretation of the contract in two entirely separate ways. First, because the claimants seek to argue that there was a previous course of dealing between the parties which incorporated their own terms and conditions, it is necessary to examine the past relationship between the parties to see what, assistance, if any, that provides as to the terms which the parties intended to incorporate into this specific contract.
159. Secondly, of course, the background to the contract is important as an aid to identifying the formation and assisting in the interpretation of the contract. In his well-known speech in Investors Compensation Scheme v West Bromwich Building Society [1998] 1 WLR 896, Lord Hoffmann said that "interpretation is the ascertainment of the meaning which the documents would convey to a reasonable person having all the background knowledge which would reasonably have been available to the parties in the situation in which they were at the time of the contract" and that, with the exception of previous negotiations and declarations of subjective intent, and on the proviso that it should have been reasonably available to the parties, the background included "absolutely anything which would have effected the way in which the language of the document would have been understood by a reasonable man." In BCCI v Ali [2002] 1 AC 251 at 269, Lord Hoffmann qualified this by stressing that it meant anything which a reasonable man would have regarded as relevant.
160. I therefore turn to consider the background to this contract under these two separate headings.

E3: The Background/Previous Dealings

161. The claimants maintained that the defendant had previously contracted on their terms and conditions, such that there was a course of dealing between them on those terms. However, the principal difficulty with this argument is that none of the many witnesses from the claimants who gave evidence on or around this topic could demonstrate that such previous contracts had indeed been agreed on that basis.
162. Mr Hearn worked in the claimants' purchasing department from 1999. He was responsible for creating and managing a business strategy in relation to capital and asset care purchasing. This was principally due to the merger between Cadbury and Trebor Bassett, and the need to rationalise the two supplier bases for the purchase of capital and engineering equipment. However, although the purchasing department might occasionally get involved in the negotiations, Mr Hearn made plain that what he called the transactional issues, that is to say the raising of a purchasing order and the like, would not be something that they would ordinarily be involved with.
163. He said there was a computerised system which recorded these transactions. The vendors each had a vendor number within that system. Mr Hearn said:

"Within the system, in order to raise a purchase order, it is necessary to specify a vendor, and that vendor has a number associated with it. Under the situation that you have just described, there could be two entries in there: one, the number originating from the Trebor Bassett system; one the number originating from the Cadbury system."

He said that it was possible to negate a number within the database so that it was no longer available for use. As to how the factory, who would be placing the order, would obtain a vendor number, he said that this was controlled by the vendor payments department. He thought that there were a limited number of people who had access to the process which allowed a new vendor or vendor number to be put on to the system.

164. Miss Natasha Lee was also involved in the purchasing department, although she had been involved in setting up a procurement team for what was called 'commercial and central', which related to professional services, a different team to that run by Mr Hearn. Miss Lee confirmed that she had instructed people in her department to investigate whether they could find copies of any documents which suggested that, prior to 2003, the defendant had signed up to the claimants' standard terms and conditions. She confirmed that they had been unable to find anything at all. She had also not been able to find when the defendant, through its Leeds office, had first become a vendor to the claimants.
165. Miss Lee also said that, before the computerised system, the claimants used a manual process to set up suppliers. That involved the supplier returning a signed acceptance hard copy which was then scanned onto the system and filed. She confirmed that, if a copy had been scanned, there ought to be two copies of the acceptance in the archives: the original and the scanned. Despite this, no one had been able to find in the archives even one copy of an acceptance in which the defendant had signed up to the claimants' standard terms and conditions.
166. Mr Fowler was in the claimants' Procurement Business Services Department. He confirmed his understanding that, as long as a vendor number existed, then the local site or factory would be able to raise a purchase order with that vendor. He said that the SAP system changed in 2005, and that from then on all vendor numbers had a 7 digit code. Some 6 digit codes had been retained from that earlier time. As noted in paragraph 118 above, the vendor number in this case, ascribed to the defendant's Leeds office, was a 6 figure number^[11]. Mr Fowler also said that research showed that the defendant had a number of offices which had 7 digit vendor numbers and one – this one - that predated it. The different vendor numbers related to the different offices operated by the defendant.
167. Finally in this list of witnesses was Miss Parsons, who was the claimants' Vendor Payments Manager. She explained that a purchase order was necessary in order for payments to be made, and that the purchase orders were generated automatically by a computer at the site or factory. She was the only one of the claimants' witnesses to suggest that, from the material that she had been able to analyse on her computer, the defendant may have previously agreed to the claimants' standard terms and conditions. This assertion was based on the fact that, out of a range of options, she had identified that the defendant had previously agreed to a particular payment term ("before end of the next month, due net") which, according to Miss Parsons, was one of the claimants' standard terms and conditions. However, the difficulty with this evidence was that it was just as likely that such a term could have been agreed specifically for that particular contract; its agreement did not automatically mean that the defendant had signed up to the claimants' terms and conditions. In any event, the term in question could hardly have been described as unusual or exceptional.
168. The evidence was that, because Mr Barker was easily able to obtain a vendor number for the defendant's office in Leeds, the claimants had probably contracted with the defendant's Leeds office before (possibly in 2001). But there is no persuasive evidence that any earlier contract had necessarily incorporated the claimants' terms and conditions, or that the defendant had only been given a vendor number in the first place because they had previously agreed to the claimants' terms and conditions.
169. I find that the mere fact that there was a pre-existing vendor number cannot be taken as evidence that the defendant had ever previously agreed to the claimants' terms and conditions. Furthermore, if there was any doubt about it, my analysis of the subsequent maintenance contract (Section H2 below), makes plain that, on any view, that maintenance contract was let by the claimants to the defendant in 2004 on the defendant's terms and conditions, despite the fact that the same 6 digit vendor number had been used as a reference.
170. In those circumstances, I conclude that the claimants have not made out any course of prior dealing on their standard terms and conditions. Furthermore that view is unaltered by the defendant's email to the claimants of 23^rd September 2005 (after the contract and after the fire), written by a Mr Day (who did not give evidence) which suggested that, after asking an unnamed accounts manager for his view, "we work to your terms and conditions". Although Mr ter Haar relied heavily on this untested assertion in his closing submissions, it is of no probative value because it did not identify the basis of the alleged belief, or deal with the period during which it was said such contracts had been placed. As a statement of general application it was also demonstrably wrong: the maintenance contract was not placed on the claimants' terms and conditions.

E4: The Specific Background To This Contract

171. In accordance with the approach outlined in Investors Compensation Scheme, it seems to me that the specific background to this contract was as follows:

a) The claimants wanted to move their 'oil pop' popcorn production lines from Leeds to the Monkhill premises at Pontefract;

b) Those production lines were protected by a CO2 suppression system provided by Wormald;

c) Unless the claimants were specifically advised, either by the specialist fire prevention contractor or their own Group Risk Department, that a CO2 suppression system was inappropriate in all the circumstances, they were going to replicate the CO2 system at Leeds in the new facility at Monkhill;

d) The claimants had decided on a CO2 suppression system before making any contact with the defendant;

e) The defendant was invited to quote for this work because they were the suppliers of the fire systems at Monkhill;

f) The defendant was provided with very little information on which to quote. They had some generic drawings and they scaled from these and made various assumptions as to the size of the hoppers and elevators. Those assumptions were then set out in their specification of 28 August 2003.

g) The defendant quoted for the work and did not indicate that the use of a CO2 system in the elevator and the hopper was in any way inappropriate.

E5: Primary Analysis of the Formation of the Contract

a) The 'Last Shot'

172. The defendant's quotation was based on their terms and conditions. It was, on any view, an offer. The Purchase Order was based on the claimants' terms and conditions. As explained below, they were very different. The Purchase Order was therefore a counter-offer. If there was no further exchange of contractual documentation, and the work was done pursuant to the last document in the chain, then there can be little doubt that the work was carried out pursuant to the Purchase Order of 3 September 2003. In those circumstances, the conventional analysis would be that the claimants' terms and conditions applied.
173. It seems to me that this case is, on a proper analysis, indistinguishable from the traditional approach that was the subject of the decision of the Court of Appeal in Tekdata. As indicated above, the paragraph that I have cited from the judgment of Dyson LJ is directly applicable to this case. I consider that I am bound by Tekdata and that it provides a complete answer to the issue as to the appropriate terms and conditions.
174. There was a suggestion, pleaded by the defendants, that the Purchase Order of 3^rd September was somehow not 'the last shot'. They relied on the updated specification of 17 September 2003, which Mr Taylor had at the time of the installation (paragraph 142a) above). But there were a number of insurmountable difficulties with this argument, which eventually saw its effective abandonment at the time of the closing submissions. For completeness, I deal with them briefly below.
175. First, the evidence was that the revised specification, and the covering quotation, which were both dated 17 September 2003, were never sent by the defendant to the claimants. It cannot be the last shot in the battle if it was never fired. Secondly, there was no evidence that these documents were ever received by the claimants. Indeed, their positive evidence was to the effect that these documents were not received.
176. Thirdly, standing back from the detail, it seems certain that, on the facts, the critical work was carried out by the defendant in accordance with the exchanges at the end of August and the beginning of September. For example, Mr Flavell's visit on 11 September 2003, in order to obtain information to assist the design process on which he was working, was plainly taking place in accordance with that exchange of documentation. Accordingly, I accept Mr ter Haar's submission that the quotation and specification of 17 September were immaterial in any event, because the work was being undertaken pursuant to the earlier documents.

b) Acceptance/Agreement of Defendant's Terms?

177. In order to get round the traditional approach, as exemplified by Tekdata, Mr Dennys argued that the Purchase Order was not in law a counter-offer, but a simple acceptance of the defendant's terms and conditions. He raised three points in support of that contention: first, that the Purchase Order said that it was "as per your quotation" (in circumstances where the quotation incorporated the defendant's terms and conditions); second, that the claimants' standard terms and conditions would apply "unless otherwise agreed", and that it had been otherwise agreed that the defendant's terms and conditions would apply; and third, that the Purchase Order was not, and was not intended to be, a counter-offer.
178. There are, I think, a number of separate difficulties with these arguments. First and foremost, there was no reference whatsoever to the defendant's terms and conditions in the subsequent Purchase Order. The Purchase Order accepted the quotation for the work, but it said nothing that could be construed as an acceptance of the defendant's terms and conditions. Almost all purchase orders accept the quotation provided by the supplier. The critical question is the terms and conditions (if any) on which the quotation is being accepted and, in the absence of any reference to the defendant's terms and conditions in the Purchase Order, it does not seem to me that the reference to the defendant's quotation can be construed in the way that Mr Dennys suggests.
179. I also reject Mr Dennys' argument based on the words "unless otherwise agreed". Nothing on the face of the Purchase Order could be taken as being an agreement between the parties that the defendant's terms and conditions (which, it is common ground, the claimants had not even seen) were being 'otherwise agreed' so as to extinguish the claimants' standard terms and conditions, to which express reference was made in the Purchase Order.
180. In essence, this was a straightforward purchase order from the buyer which, in the time-honoured way, was an attempt by the buyer to replace the seller's terms and conditions with his own. The position would only have been different if somewhere on the face of the Purchase Order there was an express agreement by the claimants to be bound by the defendant's terms and conditions. But there was no such agreement, in part for the very good reason that those terms and conditions had not even been supplied by the defendant to the claimants.
181. This interpretation also provides a complete answer to the suggestion that the Purchase Order was not a counteroffer; in my view, for the reasons stated, that is exactly what it was. It was accepted by the defendant when, having received it in early September, they then began work on the design of the system. Mr Flavell's visit to the site on 11^th September 2003 is the clearest evidence that the Purchase Order had been accepted and that a contract had been made.

c) The Terms of the Purchase Order

182. The defendant next argued that the terms of the Purchase Order did not negate their terms and conditions and that, in some way, the contract was capable of being read as including both sets of terms and conditions.
183. It is certainly true that the Purchase Order did refer to the quotation without qualification, and did not contain, as many similar purchase orders do, a provision that said that the claimants' standard terms and conditions operated to exclude any other terms and conditions. But on the other hand, the Purchase Order said clearly that, in the absence of other agreement, 'the Monkhill conditions will apply'. Thus, it seems to me impossible to conclude that the acceptance of the quotation in the Purchase Order created some sort of hybrid contract, incorporating both sets of terms, if there was a significant difference between the claimants' terms and those of the defendant. In my judgment, there plainly was such a difference.
184. Let us take the question of liability as an example. On the defendant's case, their liability was limited to a maximum arrived by calculating 20 times the annual service charge. Leaving aside the difficulty with identifying that particular figure in these circumstances, it is clear that the defendant relies on this as a major limitation of their liability. They calculate that they would be liable for no more than £13,000, a calculation which refers to the subsequent maintenance contract. That provision is simply impossible to reconcile with the second claimants' terms and conditions which, although not draconian in themselves, included damages and liability provisions which would at least allow the claimants to recover the maximum damages that they could prove in the event of the defendants' breach of contract. As previously noted, that is currently put at £110 million.
185. There is either a contractual limitation on the extent of the defendant's liability or there is not. It is impossible to read these provisions as covering both eventualities: they are contradictory. One must prevail over the other, and in accordance with the traditional approach, I conclude that it is the terms and conditions referred to in the Purchase Order that prevail.

d) The Lack of Notification

186. The defendant complains that they received no notification of the claimants' terms and conditions. They say that, although the Purchase Order referred to them, the terms and conditions were not provided, so the defendant had never seen them. It is therefore said the terms were not incorporated. There are four answers to that submission.
187. First, it is right to point out that the same is equally true of the defendant's terms and conditions. Those were not attached to the defendant's quotation and were nowhere set out in the quotation or specification of 28 August.
188. Secondly, the reference to the claimants' standard terms and conditions was a reference to such terms and conditions "already supplied". That is consistent with the evidence from the claimants that they operated on at least the assumption that all their supply contracts were entered into on their own terms and conditions. Although, as Mr Dennys rightly said, that was an aspiration rather than an immutable rule, it was consistent with that aspiration that their standard Purchase Order would refer to those terms and conditions on the basis that they had indeed been previously supplied to the supplier.
189. Accordingly, the words on the Purchase Order would not only have alerted the defendant to the existence of the claimants' terms and conditions, but they would have also alerted them to at least the possibility, if not the probability, that they already had those terms and conditions, because they had already contracted on them in the past. It thus provided a double-edged warning to the defendant as to both the existence and the potential applicability of the claimants' terms and conditions. The ball was therefore firmly in the defendant's court if they did not have a copy (or challenged the applicability) of the claimant's terms and conditions.
190. Thirdly, I am relatively relaxed about the incorporation of the claimants' terms and conditions as such, even though they were not sent out with the Purchase Order, because, unlike the defendant's terms and conditions, they do not contain any particularly onerous clauses. The principal provisions, which I have summarised at paragraph 120 above, are, so it seems to me, unremarkable. They do not, for example, purport to impose a liability on the defendant which they would not otherwise have, or make them liable for types of loss and damage which the law would not ordinarily impose.
191. Fourthly, it is trite law that notice within a contractual document identifying and relying on standard trading terms is sufficient to permit incorporation of those terms: see Circle Freight International Limited v Medeast Golf Exports Limited [1988] 2 Lloyd's Rep 427 at 433. Further, given that none of the claimants' terms and conditions were unduly onerous, they did not require specific or particular notice: see Interfoto Picture Library Limited v Stiletto Visual Program Limited [1989] QB 433 and HIH v New Hampshire [2001] EWCA Civ 735.
192. For all of these reasons, therefore, I consider that the fact that the claimants' terms and conditions were not enclosed with the Purchase Order does not prevent their incorporation into the contract between the parties.

e) The Lack of Clarity

193. Mr Dennys' final argument was that the reference at the bottom of the Purchase Order was unclear because it referred to the "Monkhill standard terms and conditions" and he said that it was unclear who or what they were. The terms and conditions (to which reference has been made at paragraphs 120 above) were actually entitled "Cadbury Trebor Basset standard terms and conditions of purchase" and, said Mr Dennys, it was not clear that they were the same thing as the Monkhill terms.
194. I reject that argument. The full title of the second claimants is 'The Cadbury UK Partnership (formally known as The Cadbury Trebor Basset Partnership T/A Monkhill Confectionery)'. It is plain from all of the documents that the second claimants were referred to throughout by the shorthand title, 'Monkhill Confectionery.' Although the original documents prepared by the defendant not unreasonably referred to Kernels (a subsidiary of the second claimants at the time and the company manufacturing the popcorn at Leeds which the defendant had visited), this misapprehension was quickly altered. That can be seen from the defendant's quotation and specification of 17 September 2003, which were both made out to 'Monkhill Confectionary.'
195. On the facts, therefore, there can be no doubt that the defendant considered that it was contracting with an entity referred to, at least in shorthand, as Monkhill Confectionery. I find therefore that there was no lack of clarity in the reference to 'Monkhill standard terms and conditions' at the foot of the Purchase Order.

f) Summary

196. For these reasons, therefore, I have concluded that this contract incorporated the defendant's quotation and specification of 28 August 2003; the claimants' Purchase Order dated 3 September 2003; and the claimants' standard terms and conditions. As a result, the unusual circumstances in which a bare contract might be found (GHSP Inc v AB Electronic Ltd [2011] Lloyd's Rep Plus 21) do not arise.

E6: Other Matters

197. There is one further point that I should make. Even if I was wrong about the incorporation of the claimants' terms and conditions, and in some way the defendant's terms and conditions were incorporated instead, I would have had no hesitation in concluding that the defendant had failed to limit its liability to 20 times the amount of the yearly service charge. The reason is obvious: there was no yearly service charge. The defendant's terms and conditions appear to assume that there was a maintenance agreement either in existence, or coming into effect on installation, which specified an annual service charge.
198. But this contract was not a maintenance agreement, and made no reference to such a contract coming into existence. This was a contact for the provision of goods and services; there was no maintenance charge. Accordingly, the limitation of liability clause, as drafted, could not operate. I bear in mind that such clauses will usually be construed strictly, albeit not as strictly as clauses which purport to exclude liability altogether (see Ailsa Craig Fishing Co Ltd v Malvern Fishing Co Ltd [1983] 1 WLR 964). But the provision, as it stands, was meaningless, because the limitation was calculable only by reference to the value of a non-existent contract. That was too uncertain to be enforceable. And although a maintenance contract was agreed in August 2004, I accept Mr ter Haar's submission that "the price for a service agreement entered into nearly a year after the original agreement does nothing to make the limitation clause sufficiently clear at the time of contract, such that it can be said with any confidence that the parties clearly intended the defendant should be limited to a liability of £13,781.60 in the event it was in breach of that contract" (paragraph 110 of the claimants' closing submissions).
199. At one point, the defendant argued that, as a matter of interpretation, I should replace the words "the amount of the yearly service charge" with "the amount of the contract for the supply of goods and services". But that would take creative interpretation well beyond the words used by the defendant themselves in their own terms and conditions. It would effectively amount to rewriting the contract. That would be illegitimate. Moreover, it could not be said that such a provision should be implied. The contract works perfectly well without this clause; it simply means that the defendant remained liable for the damages caused by their own breach of contract, without any limitation.
200. For that reason, therefore, it seems to me that, even if I was wrong, and the defendant's terms and conditions applied instead, this provision would still not operate so as to limit the defendants' liability in the way that they maintained.
201. In the circumstances it is unnecessary for me to deal with any of the other contractual arguments that were raised in connection with the defendant's attempt to limit its liability. However, for completeness, I should say that, if, contrary to my view, the defendant's terms and conditions were incorporated into this contract and if, also contrary to my view, the condition that purported to limit liability was properly workable, then I would not have excluded it on the basis that it was particularly unusual. In that regard, I respectfully agree with the decision of Christopher Clarke J in Shepherd Homes Limited v Encia Remediation Limited [2007] EWHC 70 (TCC), in which he said that, in his experience, a clause limiting liability to the contract price was not particularly unusual, and he referred to other contracts identified in the evidence before him, and which contained similar clauses. Similarly, in the present case, I have been referred to the quotation of Wormald which, amongst other things, sought to limit liability to the total price payable by the claimants under the contract, or £10,000, whichever sum was the lesser.
202. Accordingly, so it seems to me, if I was wrong as to incorporation, and also wrong as to whether or not the clause could operate in these circumstances, I would not have excluded the clause on the ground that it was unusual.

E7: Summary

203. For the reasons set out above, I find that the contract between the parties was made in accordance with the second claimants' Purchase Order, and incorporated the second claimants' terms and conditions. I make plain that that finding is based on a traditional analysis of the battle of the forms; I find that there was no relevant prior course of dealing which could lead to the same conclusion. I further find that, even if I was wrong about the terms and conditions, the limitation of liability clause on which the defendant relies was incapable of operation in this context, because it was designed for a maintenance contract, not this kind of supply contract.

F: THE TERMS OF THE CONTRACT

F1: Introduction

204. As set out in Section D above, I have concluded that the contract was made when the defendant accepted the claimants' counter-offer and carried out the work between September and December 2003. The contract documents comprised:

a) The quotation and specification dated 28 August 2003;

b) The Purchase Order of 3 September 2003;

c) The claimants' standard terms and conditions.

205. The next set of issues concerns the express and implied terms of the contract. This is an important matter, not only for its own sake, but also because the nature and extent of the express and implied obligations on the part of the defendant will determine whether, as a matter of principle, their arguments as to contributory negligence are open to them. Because of the potential importance of that issue, the emphasis of the claimants' case changed during the course of the trial, and moved from a focus on the defendant's obligation to take reasonable skill and care (which would permit the defendant, if in breach, to argue contributory negligence), to a much greater emphasis on an alleged warranty or guarantee (which would not permit the defendant to raise allegations of contributory negligence, and would therefore require the defendant to raise such matters by way of causation and nothing else).
206. Accordingly, because of this change of emphasis, I have found it helpful to set out the claimants' case as originally pleaded (Section F2 below) and then as presented in their closing submissions (Section F3 below). I deal with the defendant's case in response at Section F4 below. I identify the relevant issues at Section F5 and set out my analysis at Section F6. There is a short summary at Section F7.

F2: The Claimants' Pleaded Case

207. The claimants' original pleaded case set out, at paragraph 28 of the Particulars of Claim, some of the claimants' own terms and conditions. Those are the ones which I have set out at paragraph 120 above. To those express terms was added an implied term, to the effect that the defendant "should exercise reasonable skill and care in the provision of its services. This should be implied either by virtue of the Supply of Goods and Services Act 1982 section 13 and/or necessity and/or business efficacy and/or the practice in the trade and/or otherwise."
208. In the autumn of last year, the claimants amended their pleading in relation to the terms of the contract. The principal amendments concerned their case as to the previous course of dealing (which are irrelevant for this purpose, because I have rejected that case on the facts). The other amendment, following on from the alleged implied term as to reasonable skill and care, alleged a further implied term:

"…that the CO2 suppression system supplied and installed by ADT, and the equipment that comprised it, would be of satisfactory quality and/or fit for its purpose as a CO2 suppression system and/or conformed to its description. This should be implied either by virtue of the Supply of Goods and Services Act 1982 sections 3 and 4 and/or necessity and/or business efficacy and/or the practice in the trade and/or otherwise."

209. Accordingly, the issues as to contract terms raised on the claimants' pleaded case were the incorporation of their own express terms (which I have accepted) and the implication of three further terms, principally by reference to the Supply of Goods and Services Act 1982, in relation to i) reasonable skill and care as to the services provided; ii) satisfactory quality; and iii) fitness for purpose.

F3: The Claimants' Case at Trial

210. The claimants' case by the end of the trial was very different. By then, the emphasis was on the specification provided by the defendant, and in particular, the first three paragraphs of that specification, which I have set out at paragraph 107 above. It was the claimants' case that these paragraphs amounted to a form of warranty or guarantee on the part of the defendant (which was not an obligation incorporating reasonable skill and care and therefore did not permit any arguments as to contributory negligence). At its highest, it was said that the warranty was to the effect that there would be no occurrence of fire at all. But that was refined during the course of argument and the emphasis shifted again, this time to the assertion that the defendant had warranted that its system would prevent any fire from escaping either the hopper or the elevator.

F4: The Defendant's Case

211. The defendant's principal case was based on their own terms and conditions, which is a case that I have rejected. As to the claimants' pleaded case:

i) The term as to reasonable skill and care in the provision of services was admitted;

ii) The implied terms as to satisfactory quality and fitness for purpose were denied. In particular, the defendant pleaded that the claimant had failed to identify:

"(a) any specific purpose said to have been communicated by an identified representative of the claimant to a representative of the defendant,

(b) the respect in which the claimant is alleged to have relied on the defendant and the extent of that reliance,

(c) the precise goods said to have been supplied."

212. In addition, at paragraph 22(4) of the amended defence, the defendant contended:

"(i) The evidence served by the claimant does not disclose any case [as to specific purpose and reliance] and is anyway denied. The basis on which the parties proceeded was that the existing system in the claimants' factory at Leeds should so far as possible be replicated in the new installation at Pontefract;

(ii) The system of fire suppression to the elevator and hopper would be a CO2 system using fixed probes at a pre-determined temperature (of 100°C) and would be capable of being triggered manually;

(iii) The defendant was quoting in competition and cost was an issue;

(iv) The sale was not a sale by description."

213. During the course of the trial, the defendant maintained these pleaded points. In addition, as to the alleged warranty or guarantee, the defendant denied any such warranty, arguing that the claimants' new case was based on a mis-reading of the specification.

F5: The Issues

214. It seems to me therefore that, in relation to the express and implied terms of the contract, the following issues arise for determination:

a) Was there a term of the contract that the defendant would exercise reasonable skill and care in the provision of its services?

b) Was there an implied term that the CO2 system would be of satisfactory quality?

c) Was there an implied term that the CO2 system would be fit for its purpose?

d) Was there some form of warranty or guarantee to be derived from the first three paragraphs of the specification and, if so, what was the nature, scope and extent of that warranty?

F6: Analysis

a) The Nature of the Contract

215. It is worth identifying at the outset the nature of the contract between the parties. It was a contract in which the defendant was to supply to the claimants a system designed to suppress fire in the elevator and/or hopper. That system was made up of various pieces of equipment, such as the sensors and the nozzles, as well as the fire alarm panels and the electrical wiring. But what made it a system (as opposed to a random selection of equipment) was the design: the pulling together of all the relevant information into a designed system that used CO2 to suppress fire in the elevator or hopper. Accordingly, the most important element of the workscope was that done by the defendant's designers: the decision to use a particular piece of equipment in a particular place, with the intention of achieving a particular effect.

b) Reasonable Skill and Care

216. In my view, there was an implied term of the contract that the defendant would exercise reasonable skill and care in the provision of its services. Those services were, as I have said, the critical element of the contract, namely the design of the CO2 system. That term was therefore implied by section 13 of the Supply of Goods and Services Act 1982.^[12] The implication of that term was not in dispute. It was an obligation to take reasonable skill and care; it was not a guarantee of success. It did, however, mean that, contrary to the defendant's closing submissions, the defendant had to do rather more than just provide a CO2 system, as requested.
217. This was an important term because it regulated the design of the suppression system. Moreover, it carried with it the usual inference that someone in the position of the defendant providing design services will not generally owe a higher obligation than the duty to use reasonable skill and care: see George Hawkins v Chrysler (UK) Ltd [1986] 38 BLR 36. That case is authority for the general proposition that, where the designer is obliged to exercise reasonable skill and care, in the absence of special circumstances, it is not open to the court to extend the responsibilities of the designer beyond that duty. Whilst that case arises in the context of professional negligence, I can see no justification for adopting a different approach to a specialist supplier providing a vital design service.

c) Satisfactory Quality

218. There was a good deal of argument about whether the contract contained an implied term, by reference to section 4 of the Supply of Goods and Services Act, as to satisfactory quality. In my view, this argument was entirely redundant, given that the claimants' terms and conditions (which I have found were incorporated into this contract) included at clause 3 an express term that the goods supplied would be of "good quality". That was a more onerous term, so it seems to me that that is sufficient to resolve this point. However, I deal briefly with the arguments about section 4 below.
219. Section 4(2) of the Supply of Goods and Services Act provides that:

"Where, under such a contract, the transferor transfers the property and goods in the course of a business, there is an implied condition that the goods supplied under the contract are of satisfactory quality.

(2A) for the purposes of this section and section 5 below, goods are of satisfactory quality if they meet the standard that a reasonable person would regard as satisfactory, taking account of any description of the goods, the price (if relevant) and all the other relevant circumstances."

Mr Dennys' first argument was that this term did not apply because this was primarily a contract for the supply of services, not the supply of goods. He relied on the test, set out by Greer LJ in Robinson v Graves [1935] QB 579 at 587, that what mattered was whether the substance of the contract was the production of something to be sold or the exercise of skill and labour. In my view, applying that somewhat old-fashioned distinction, this was a contract for the supply of skill and labour, such that section 4 would not apply at all. But applying a more modern test, looking at the reality of the parties' contract, I would reach the same conclusion, for the reasons already given at paragraph 215 above: what mattered most here was the design.

220. If, contrary to my primary view, this was at least in part a contract for the supply of goods as well as a contract for the supply of services, then the goods comprised the individual component parts of the system (the sensors, nozzles and so on). On this basis, the term set out in section 4 would otherwise be implied into the contract, but only insofar as it related to those individual components. In any event, there was no need for such a term, in the light of the express term as to "good quality".

d) Fitness for Purpose

221. The sections of the 1982 Act dealing with fitness for purpose are as follows:

"4(4) Subsection (5) below applies where, under a contract for the transfer of goods, the transferor transfers the property and goods in the course of a business and the transferee, expressly or by implication, makes known –

(a) to the transferor, or

(b) where the consideration or part of the consideration for the transfer is a sum payable by instalments and the goods were previously sold by a credit-broker to the transferor, to that credit-broker,

any particular purpose for which the goods are being acquired.

(5) In that case there is (subject to subsection (6) below) an implied condition that the goods supplied under the contact are reasonably fit for that purpose, whether or not that is a purpose for which such goods are commonly supplied.

(6) Subsection (5) above does not apply where the circumstances show that the transferee does not rely, or that it is unreasonable for him to rely, on the skill or judgment of the transferor of the credit-broker."

222. It is the claimants' case that, as a result of these provisions, there was a fitness for purpose obligation in the contract. The defendant denies that allegation, first because it maintains that this was not a contract for the supply of goods at all, and second because, even if it was, the claimants had no evidence and no pleaded case that at any stage they made known to the defendant "any particular purpose for which the goods were being acquired". The defendant also denies that there was any relevant reliance.
223. The claimants' closing submissions at paragraphs 95-97 skirt round these difficulties by consistently eliding questions of quality (which was covered by an express term) with an obligation as to fitness for purpose (which requires an implied term under the Act), saying that it was reasonable and necessary to imply a term that the CO2 system "should work". That begs the question as to what "work" means; it ignores the reasonable skill and care obligation (because the system might not 'work' in certain circumstances, but not be negligently designed); and ignores the statutory requirement that, for a fitness for purpose term, that purpose must have been made clear. For the reasons set out below, I do not consider that the claimants have established the implication of the fitness obligation.
224. First, for the reasons already set out in paragraph 219 above, I accept the submission that this was not, in substance, a contract for the supply of goods, so section 4 is not in play at all. But secondly, if I was wrong about that, it seems to me that, on the evidence, the defendant is indisputably right to deny the implication of a fitness for purpose obligation. At no stage did the claimants ever indicate any particular purpose for which the CO2 system was being acquired. Obviously it was being acquired to suppress fire, but that was not a particular purpose; it was simply the condition which the system was then designed to deal with.
225. In these circumstances, a "particular purpose" might have been, for example, a specific request that the fire suppression system ensured that no single piece of burning popcorn ever escaped from the elevator or the hopper. Had that particular purpose been made plain, it is likely that the defendant would have said that it was impossible to achieve, because one piece of burning popcorn would not get anywhere near to the necessary temperature required to trigger the system and that any lesser temperature would mean that the system would have been activated repeatedly.
226. I also accept that there was no reliance of the kind necessary to trigger the implication of the alleged fitness for purpose provision. By the time the claimants approached the defendant for the first time, they had already decided that they were going to replicate the existing system at Leeds in the NMU at Pontefract. Accordingly, not only did they fail to make any particular purpose known to the defendant, but, save in one general respect which is not relevant to this issue (and which is addressed in the next paragraph), the claimants did not rely on the defendant's skill or judgment at all: Rotherham MBC v Frank Haslam Milan & Co Ltd (1996) 78 BLR 1. As Mr Hamilton made plain, the existing CO2 system worked very well at Leeds; he was simply interested in making sure that the same system could operate at Monkhill. He had decided on a CO2 system before he had ever made any contact with the defendant.
227. I accept that, if the CO2 system had been wholly inappropriate, or was never going to work in the conditions at Monkhill, then the defendant was obliged to say so. If the system was completely unsuitable for the elevator and hopper, the defendant could not simply provide a replacement for the Wormald equipment without pointing out to the claimants that the system was, in truth, a complete waste of money. But there was no case – nor on the evidence could there have been - that the CO2 suppression system was itself inappropriate in concept, or misconceived in principle, or could never have worked. Accordingly, that issue does not arise.
228. Thus, in the absence of any evidence that a particular purpose was made known by the claimants to the defendant, and in the absence of anything other than general reliance, I conclude that the implied term as to fitness for purpose was not incorporated into this contract.

e) Warranty/Guarantee

229. The remaining issue is whether or not the first three paragraphs of the Specification (paragraph 107 above) amounted to a warranty or guarantee that the CO2 suppression system would meet the risk of fire^[13]. It is, I think, the claimants' case that this warranty or guarantee cuts across the implied terms, particularly the implied term as to reasonable skill and care, and imposed an absolute obligation on the defendant to design a system which prevented fire.
230. For a whole series of different reasons, I have concluded that this argument, although undoubtedly ingenious, does not stand up to proper scrutiny, and I find that there was no such warranty or guarantee. I reject it under two broad heads: the proper definition of the word 'risks' in the specification, having regard to the background to the contract, and generally, as a matter of construction.

i) 'Risks'

231. I take first the definition of the word 'risks' used in the specification (paragraph 107 above). That was not a word that the parties defined in their contractual or post-contractual exchanges. It was not really a matter to which they devoted any attention at all until shortly before the trial, when it suddenly loomed large as an important element of the claimants' warranty/guarantee case. However, the background to the contract provides some assistance in identifying and interpreting 'the risks' referred to in the specification.
232. It is the claimants' pleaded case that 'the risks' were the occurrence and spread of fire: any fire in the hopper and/or elevator, of any kind, from any source. That was exemplified by the second reply to the defendant's request for further information where, at reply B, in answer to that very question, the claimants said:

"The risk that the CO2 fire suppression system was supposed to guard against was the occurrence and spread of fire. The whole point was that the CO2 suppression system would automatically activate to extinguish or control a fire in either the hopper and/or Gough Elevator. It failed to either activate within a reasonable time and/or suppress/extinguish the fire sufficiently or at all."

233. In my view, the suggestion that the risk referred to in the specification was the occurrence and spread of any fire cannot be sustained. On the contrary, the CO2 system depended on the occurrence of a fire in order to operate at all: there had to be a fire in order to cause the sensors to be activated and the CO2 suppression machinery to be triggered. The CO2 suppression system did not prevent a fire from starting: it was instead designed to put it out. Thus the risk cannot have been the occurrence of fire; on the contrary, the CO2 system presupposed that there would be a fire before it could be activated.
234. In the claimants' closing submissions, Mr ter Haar refined his analysis still further, and instead argued that the 'risk' referred to was the risk of the fire getting out of the hopper or elevator. Although that case was not at the time pleaded, it seems to me to be a rather more thoughtful attempt to define the word 'risk', as used in the specification. But I have concluded that it is still not a persuasive submission.
235. First, I find that nobody, on either side, thought for a moment that if, for example, one piece of popcorn was passing through the system and caught fire somewhere between the pan and the hopper, and then dropped straight through the hopper and into the bagging hatch, the CO2 suppression system would or should have been triggered. One flaming or smoking piece of popcorn would not (and was not intended to) have triggered the system, because the flame from one piece of popcorn would not have reached the necessary temperature in order to activate the sensor.
236. Secondly, there was the clear evidence from Mr Hamilton that he was aware that the sensors could not necessarily respond to every fire. I refer in particular to the note at Bundle R/310 in which, during a post-fire interview, Mr Hamilton said that he knew that the sensors in the system at Leeds may not have been sensitive and that "we expect to have to put out small fires manually", a point he also accepted in cross-examination (Transcript Day 5/24).
237. Thirdly, there was the evidence (paragraphs 70-72 and 76 above) that Mr Hamilton (and everyone else at the claimants) was aware that the only truly reliable way to deal with the risk of fire in the 'oil pop' production area was with the use of sprinklers. He knew that the CO2 system would not necessarily deal with all fires, including fires escaping from the elevator and hopper, and was aware of that from the outset of his involvement (because he knew that there was a CO2 system at Leeds and that this did not prevent the outbreak of fires there).
238. Fourthly, and perhaps most significantly of all on this point, it must be remembered that this was a system that was designed to provide, in the alternative to an automatic operation, a manual system, whereby the release of the CO2 could be triggered manually. The immediate question that arises is: how would an operative have realised that the CO2 system needed to be triggered manually unless and until he or she was aware of a fire, and how would he or she become aware of a fire in the hopper or the elevator, in circumstances where both were wholly enclosed? The answer, of course, is that he or she would have become aware of the fire, and the need for the manual activation, only when the fire escaped either the hopper or the elevator. Thus the very existence of the manual activation system negated the suggestion that the defendant warranted that the system would suppress all fires before they escaped the hopper or the elevator.
239. Finally on this point, I note that the specification also uses the word "hazard" but this, too, cannot be read as a reference to the spread of fire. Indeed, the hazard appears to be defined by the specification itself as simply the area – the "elevator and hopper arrangements" – to be covered by the CO2 system.
240. Taking into account the background to this contract, and the information available to both parties at the time that it was made, it seems to me that the 'risks' which the system was designed to address was a developed fire (ie not just a burning piece of popcorn or two) in the hopper or the elevator which might escape the enclosure and, because of its developed nature, prove difficult to extinguish. The mere fact that fire escaped from the enclosure would not necessarily mean that the defendant had failed to comply with its obligations under the contract. But if the design failed to address the risk that a fire would build up in the enclosure and then escape, such that it might no longer be possible for the claimants to control it, it might point to a failure by the defendant to exercise reasonable skill and care in the design.

ii) Construction Generally

241. I also reject the warranty/guarantee argument as a matter of broader construction. It seems to me that the first three paragraphs have to be read in the context of the specification as a whole. That document, read as a whole, was plainly not providing any kind of warranty or guarantee, but explaining the various assumptions that had been made as to the design and the various matters, such as openings, which were in any event beyond the defendant's control. In such circumstances, it seems to me that the words in the first three paragraphs cannot be taken out of context and then transformed into some form of strict liability provision.
242. In addition, the fact that the system had been designed "to suit the specific requirements of the risks to be protected" (to use the words of the specification) does not mean that the system was guaranteed to eliminate those risks every time. It was designed, manufactured and installed with those risks in mind, but the obligation as to design was to take reasonable skill and care and there was nothing in these words to impose upon the defendant a more onerous obligation. A design intended 'to suit the specific requirements of the risks' may or may not be done with reasonable skill and care, but it was not a promise that the design would eliminate all risks. A promise that a particular risk will be eliminated must say so in clear terms. The specification did not do so.
243. If there was any doubt about that last point, I note that the words in the specification go on to say that the system was designed, manufactured and installed to "comply generally with the requirements of BS 5306 Part 4." That was a non-specific and general obligation to comply generally with requirements which are, in this British Standard, often set out by way of recommendation rather than obligation. That is again entirely consistent with the exercise by the defendant of reasonable skill and care and inconsistent with a guarantee or warranty.
244. For all these reasons, I do not consider that the words in the specification amounted to a warranty or a guarantee. What they do is to provide the benchmark against which the obligations as to reasonable skill and care and good quality are to be measured.

f) British Standard 5306 Part 4

245. There was an express obligation, as a result of the three paragraphs of the specification, that the design would "comply generally with the requirements of British Standard 5306 Part 4". Again, this was not a mandatory requirement to meet every part of the BS. It was an obligation to comply generally with the BS: it meant that the recommendations of British Standard 5306 were to apply generally, but if there was a good reason for departing from those recommendations then that would not be a breach of contract.

F7 Summary

246. Accordingly, for the reasons set out above, I consider that:

a) The contract included the express terms contained in the claimants' standard terms and conditions and set out at paragraph 120 above;

b) The contract included the implied term as to reasonable skill and care in the provision of the all-important design services and the express term as to good quality in relation to the individual components of the system;

c) The specification, with its reference to British Standard 5306 and other assumptions and requirements, formed the benchmark against which the defendant's obligations were to be performed. The real risk that the system should have been designed to meet was a developed fire in the hopper or the elevator which, if it escaped the enclosure, might no longer be capable of being controlled.

247. For the reasons set out above, I find that there was no obligation of fitness for purpose, and no warranty or guarantee, whether as alleged or at all.

G: THE MANUFACTURING PROCESS AND THE INCIDENCE OF FIRE

G1 The Process Itself

248. It is important to understand the 'oil pop' popcorn production process in order to make sense of the events after the system was installed and before the fire, and the events on the night of the fire itself.
249. There were two 'oil pop' production lines in the former Production Area B. They were known as the C Line and the D Line. Each consisted of 12 cooking pans, arranged in 6 pairs. The popcorn was heated over naked flames in these pans using soya oil. Once the corn reached a certain temperature it 'popped', turning itself inside out in the process.
250. Once the popcorn in the pan had popped, it was tipped onto a conveyor. If one or more pieces of popcorn were flaming on the conveyor at that point, there was no fire detection or suppression system in place to deal with it; the system relied on the operatives to see the burning piece, knock it off the conveyor, and stamp it out on the floor.
251. The popcorn was then carried along the conveyor to the rotary dosing drum, which was used to apply savoury coatings to the popcorn. There was some debate as to whether the dosing drum would cool the popcorn. It seems to me that it might have had that effect, but that was not its principal purpose. After the popcorn had been through the dosing drum, it passed along the conveyor and into what was known as the boot of the elevator.
252. Within the elevator were a series of plastic buckets, 151 in all, which were driven by a plastic chain. The popcorn was tipped off the conveyor and into these buckets. Almost immediately, the buckets swung vertically upwards to rise some 12 feet into the air. The elevator was enclosed in this vertical section. The buckets then turned horizontally and ran across the remainder of old Production Area B at ceiling level. The evidence was that the buckets, which had thin rims or lips, overlapped, so as to minimise the amount of popcorn falling out into the elevator itself.
253. As already noted, the Gough elevator carried on horizontally through the wall that separated old Production Area B from the Packaging Hall. The popcorn for the retail trade was dropped from the buckets into a retail bagging machine, which looked rather like a carousel, in the Packaging Hall. The popcorn for the cinema trade came back in the buckets along the elevator and returned into the 'oil pop' production area, and was then deposited into a hopper. When the buckets reached the relevant point in the conveyor, they were tipped into a position that was slightly greater than vertical, so that the contents of the bucket fell down the chute and into the hopper below. The chute was surrounded by a skirt, and there were also two inspection hatches, or lids. The evidence was that the buckets remained vertical for no more than 1 or 2 seconds. They then returned to the horizontal position and continued on their journey back down the horizontal section of the conveyor to collect a fresh load of popcorn.

G2 Oil Spillage

254. Once the C and D Lines were up and running in December 2003, oil spillage occurred onto the floor of the production area. Mr Barker confirmed that this was a problem at the outset, but he thought that it was dealt with effectively and quickly. In fact, it took rather longer to address than that, and the impression created was of a rather casual approach to the issue.
255. The first sign of a spillage problem was a reference in an email to Michelle Irving dated 26 November 2003, which expressed the hope that the amount of oil spillage would reduce following a change in the way in which the oil was added to the process. The slipping hazard was also noted generally in a meeting on 3 December 2003.
256. A system of slip resistant mats was tried, but the Project Kettle meeting notes of 9 December 2003 indicated that they were not effective. The deep cleaning of the floor worked for a while but was not a permanent solution. There was a suggestion that non-slip Wellington boots should be tried. By the time of the next Project Kettle meeting, on 23 December 2003, there was a suggestion in the manuscript notes that the problem was getting worse, and that oil was getting carried out into the corridors. Catch trays were needed under the end of the conveyors. At the next meeting on 27 January 2004, it was reported that the non-slip Wellingtons had not been successful, but that the catch trays had been fitted. Mr Barker explained that the catch trays were located where the conveyors went round the discharge roller.
257. Finally, in this sequence, at the health and safety review meeting on 27 January 2004, it was reported that Mr Barker was carrying out a full review of the current kit design which would incorporate a solution designed to reduce oil spillage onto the floor. Mr Barker confirmed that, at least at this point, the problem still persisted. However, this is the last reference to it in the documents and some of the operatives indicated that, after a while, the spillage problem was eased. Mr Webber said that there were problems "in the early stages when they were starting production" but that those were addressed as time went on. When asked how they were addressed he said:

"Steve Barker had equipment made which reduced the spillage of oil and they also fitted safety mats on the floor. They had specialist footwear, I recall, on the operators who worked in that area. So there were several measures introduced to minimise the effect of spilt oil."

258. Mr Carter also confirmed that there had been a spillage problem at the outset but that it got better, saying that every new venture had teething problems but as things progressed it had improved considerably. On the other hand, Mr Kear said that it was "always slippy" in that area and there was a limited time available to clean the floors.
259. Accordingly, the evidence was that, although the situation was improved by the rather ad hoc remedies created by Mr Barker, the problem of oil spillage did not cease entirely, and was a part of the 'oil pop' environment in June 2005.

G3: Pan Fires

260. Mr Carter, the shift manager on the night of the fire, told me that pan fires were a fairly common occurrence and that they usually led to a fire incident report. He said he always reported his pan fires. He said that sometimes the pan fire was a fairly minor event, when tiny bits of black popcorn stuck underneath the pan and would melt. He said "it wasn't even burning really, it was smouldering."
261. Mr Norton, another operative who was also there on the night of the fire, said that there were fires early on when the 'oil pop' process was relatively new, but by 2005 they had become much less regular. However, he said that 12 pan fires in a full year of working was, as he put it, "pretty good odds"; such a frequency was "not bad for a full year." In answer to a question from me he confirmed that one fire in one pan per month was 'good going'.
262. Mr Shaw said that there were two types of pan fire: one when the oil caught fire before the popcorn went in and secondly, a fire in the corn after it had been cooked. He said that one way of dealing with that was putting cold oil in the pan on top of the hot oil. Another way was to tip it onto the floor and extinguish it on the floor. He did not use a fire extinguisher. A third way was to put the lid over the pan and leave it to burn itself out.
263. Other witnesses gave similar evidence. Mr Holmes confirmed the various ways of putting the pan fire out, noted above. Mr Challinor was not aware of any pan fires on his shift. As far as he was aware his operators knew what to do in case of a pan fire but he himself had not had any training on that issue, even though he was a shift manager.
264. Accordingly, pan fires in the 'oil pop' production area were accepted as a fact of life by the men on the factory floor. It is not clear the extent to which that condition was appreciated by (or even known to) the more senior management, whether at the Monkhill site or elsewhere. Because there was no automatic detection system for such fires, their detection and extinction depended on chance, and the alertness of the operatives. That was not a satisfactory or safe working environment.

G4: Fires on the Conveyor

265. Mr Shaw confirmed that, sometimes, the overheated corn was tipped onto the conveyor and the fire was not noticed until it was on the conveyor line itself. Mr Norton said that the first line of defence involved patting the flaming pieces or piece of popcorn with gloves or hitting them with a spoon. Mr Shaw said he would use a fire extinguisher. Mr Holmes said that he put such fires out with a hand or sometimes a spoon and emphasised that he would not ordinarily stop the conveyor. However, if the fire was larger, he said he would have used an extinguisher and stopped the line, but only if the fire had spread, and he could not recall that ever happening. He said that, as people got used to the process, fires on the conveyor were rare. Mr Kear said it was quite easy to stamp out a fire by stamping on the popcorn or using a flat palm.
266. Again, therefore, the detection of fires on the conveyor depended on luck and the alertness of the operatives; there was no system in place which could detect fires on the conveyor, let alone put them out. That was again unsatisfactory, for the reasons that I have already given.

G5 Fire in the Elevator/Hopper

267. The third kind of fire was when burning popcorn entered the enclosed elevator, or caught fire somewhere that it could not be seen. That happened in November 2003 (paragraph 144 above). It happened again in June 2004, when a fire occurred in one of the hoppers. Both Mr Carter and Mr Norton were there on that occasion. Mr Norton said that the burning popcorn became apparent when it was emptied out of the hopper onto the floor. He thought that the operatives had put the burning popcorn out using two CO2 fire extinguishers and the fire hose, "just to make sure". He was asked whether standard procedure when faced with a hopper fire would be to empty the hopper. He said:

"It depends how bad it was because – and how full it was, because you have got to think if it were, say, quarter full, half full, there were a lot of popcorn there, so if we were chucking it on the floor, it would have just spread and gone everywhere. If it were only a little bit of popcorn you could just set it out and contain it."

268. In my judgment, this was an important event. It demonstrated beyond any doubt that the CO2 suppression system would not necessarily be triggered automatically by flaming popcorn entering the elevator or the hopper. Even more importantly, it demonstrated that popcorn could burn undetected in the hopper and only be discovered when it was discharged.
269. It does not appear that the claimants were particularly concerned about the hopper fire in 2004. There was very little report documentation relating to that fire. There were no corrective action notices or any health and safety follow-up at all. If nothing else, the absence of any such follow-up material demonstrates that no-one at the claimants believed that the CO2 suppression system would definitely, and in every case, ensure that fire could not escape from the elevator or the hopper^[14]. Had anyone have been of that view, then they would have expressed their surprise and concern after the fire in June 2004, and the defendant would have been asked to explain what had gone wrong.

G6: The Operation of the CO2 System

270. As to the CO2 system itself, it was plain that many of the operatives had little knowledge of how it worked. Mr Carter, as a shift manager, had a greater understanding than most. He said:

"I assume that it had a sensor inside the hopper and that if it got to a certain temperature, it would set it off."

He said that he did not know very much about the CO2 gas itself. As to openings, he indicated that he knew it was important to keep the inspection hatches closed, but sometimes they were open to clear blockages and obstructions. He knew that there would be a loud noise if the system went off.

271. The defendant's commissioning engineer, Mr Walker, gave Mr Barker a detailed demonstration of the manual system as part of the commissioning. But there was a good deal of evidence about the inadequate attempts by Mr Barker and the claimants to pass this information on to others, and to instruct the operatives in the manual use of the CO2 fire suppression system.
272. At a meeting on 3 December 2003, before the system had even been commissioned, a review of fire precautions said that "instruction in use how to manually activate the system will be included in the fire extinguisher training session." That was noted as being Mr Barker's responsibility. When he gave evidence, Mr Barker agreed that it was his job to instruct the operatives as to the manual activation of the CO2 system. However, when it was put to him that this was never in fact included in the fire extinguisher training session, he said that he did not know because he had not been involved in that training session. He went on to say:

"Other than informing the team leaders on how to activate manually, and eventually automatically the CO2 system, via the break glasses that were in place, I never got involved in the training process or participated in training anybody in that area. It wasn't my field of expertise."

273. At a further review meeting on 6 January 2004, it was noted that instructions as to how manually to activate the system had been provided by Mr Barker to particular individuals. Mr Barker's evidence was that he had only told the team leaders, Dave Carter and Pat Priestley, and the engineering factory manager, Graham Webber. He had not told anybody else. It was unclear when or how he had passed on this information to the team leaders. As to the decision that only the team leaders would be trained, rather than all the operatives, Ms Irving could not recall when that had been made and she was very anxious to tell me that she would not have made that sort of decision. It seems that the decision may have been made by Ms Priestley, who was herself a Line Manager.
274. It was noted at the meeting that Mr Barker was going to write up the instructions with illustrative photographs because "all operators need to be aware" of it. Mr Barker said that he was approached by one of the trainers to assist in putting information together on the CO2 system. He said that he took some photographs and got information from the defendants or the schematic drawings and gave it to the trainer, Heather Witherspoon, so she could put together a document to train the workforce on the shop floor. However, this seemed unlikely because, at the next review meeting on 27 January 2004, the same minute appeared, again requiring Mr Barker to write up the instruction with illustrative photographs and that all operators needed to be aware of them.
275. On 18 March 2004, Michelle Irving, the health and safety manger, sent Mr Barker an email noting that there were several outstanding actions for him to complete. The first was again the preparation of instructions as to manually activating the fire suppression system with illustrative photographs, which would then be trained in by the Training Assistant. Mr Barker repeated in evidence that he had passed on information to Heather Witherspoon and that nothing ever came back to him. He said that, as far as he was concerned, the trainers had had the information that he was asked to provide and he did not need to do anything else. I find that, again, this was not accurate: as the subsequent documents made plain, Mr Barker responded to the email by telling Michelle Irving in writing that he would issue those photographs to Heather "and go the details of how to operate next week." It was unclear why Mr Barker was not doing what he was being asked.
276. Again, contrary to Mr Barker's oral evidence, it does not seem that anything further happened in relation to this outstanding matter until April, when Ms Irving issued a formal Corrective Action Notice ("CAN"). The person responsible for the failure was identified as Mr Barker. The deficiency was described as "lack of instructions and training to operatives in oil pop as to how manually activate the fire suppression system." The recommended action was "provide written instructions with illustrative photographs to allow for display and training in".
277. Mr Barker had two answers to this apparent criticism, neither of which was in any way convincing. The first was that he has not seen the CAN. The second was that CANs were "commonplace at Pontefract, and they were just a vehicle for getting an issue on the table, and whenever somebody raised a corrective action, as part of that system, they had to put somebody's name against it. Now it didn't mean that if they put that person's name against it they took ownership of that issue." On this issue, at least, I prefer the evidence of Ms Irving who told me that, if somebody did not complete their designated action within an agreed timescale, and it was repeatedly coming up, she would formalise that into a CAN. She denied that such a Notice was just a way of 'getting an issue onto the table': she said that this matter had obviously been on the table at the earlier meetings and the Notice "was to ensure he actually completed it".
278. Ms Irving told me that, although Mr Barker had verbally told the team leaders about the manual activation (paragraph 273 above), she wanted to have something written down. Accordingly, she said that Mr Barker "was to formalise an instruction with illustrative photographs." Unfortunately, not even the paper trail for this CAN was satisfactorily resolved. There was supposed to be a completion email once the action had been taken. No such email has been provided. Ms Irving said in her statement that she had no way of knowing whether corrective action was completed or not which, so it seems to me, demonstrates a slack system of health and safety documentation.
279. At a meeting on site on 4 May 2004, this matter was again top of the agenda. On this occasion the minutes suggested that "instruction on display to be reviewed and check all those who need training have been." The action points were noted as being Isabel O'Mara, and Michelle Irving herself. But nobody could help as to what happened to this outstanding matter after that. In this way, something which was recognised as being of great importance – the training of the operators in the use of the manual CO2 system – was allowed to fizzle out without any active steps taken to ensure that it was undertaken.
280. As I have said, the evidence as to what the team leaders (or anyone else involved in the 'oil pop' process) actually knew about the manual activation of the CO2 system was equally unsatisfactory. Mr Webber was the engineering manager at Monkhill. He was asked about his knowledge of the manual operation. He said:

"I do have a vague recollection of standing at the CO2 panel with Steve Barker and another person, which could have been the ADT engineer or Pat Priestley or somebody like that, going through the system of how it worked. At that point I do vaguely recall receiving the re-set keys for the panels, which I then subsequently placed in the engineer's workshop…It was my understanding that Pat Priestley then arranged for the information to be got to the appropriate people in production."

Mr Webber confirmed that he did not take part in any formal training with any other member of staff in relation to the manual operation. He said he did not see any records of any individuals having been trained on the CO2 system.

281. Mr Carter was one of the few people to whom Mr Barker thought he had explained the manual system. Mr Carter said that Mr Barker had had a conversation with him in about December 2003 (which is consistent with the evidence recorded in paragraph 273 above), but that was the only occasion on which he was told anything about the manual operation of the CO2 system. He said he could not recall being told that it was a feature of the system that there was a break-glass panel. Mr Carter's answers to the subsequent questionnaire^[15], dated April 2005, revealed that it was his understanding that, in the case of a serious fire, he would break the glass to activate the overall fire alarm system (not the CO2 system) and that if there was a small fire, he would empty out the hopper contents using the chute at the bottom and put out the contents on the floor of the factory. In that event, he said, he would use the CO2 extinguisher. He later reiterated that he was not aware that it was dangerous to be around CO2.
282. Mr Norton, an operative in the 'oil pop' part of the NMU, said that he had not been trained in the manual CO2 system at all. Mr Holmes also confirmed that none of his training had been directed towards the manual operation of the CO2 system. He did not recall any plastic notice explaining the manual operation, and he was not aware that the CO2 system could be discharged manually by hitting the yellow glass box near the canisters at the exit to the production area. Mr Challinor, who was a Shift Manager, did not know how to manually operate the system, or indeed anything else about it. He said:

"I didn't know how to test or I didn't know how it was tested, and again, if a problem arose, I would have to seek advice or report it through to the engineering manager or his counterpart on the shift."

283. Despite this widespread ignorance, on 22 April 2004, the site management at Monkhill reached the conclusion, on 22 April 2004, that "people are happy with the documentation training done on the CO2 and chemical fire suppression systems in oil pop."^[16] This was another example of the claimants' right hand not knowing what the left hand was doing. A point had been raised about the absence of proper training, which would or could be highly relevant to the fire prevention systems at Monkhill; it had been discussed and was waiting to be actioned; and then it petered out without resolution.
284. The experts disagreed as to the importance of the team leaders and operatives understanding the manual activation system. Mr Stephens, the claimants' fire suppression systems expert, went as far as to say that he did not believe that the operatives should have been instructed to use the manual controls at all. He said:

"It was sold as an automatic system; it should have operated as an automatic system…I don't believe that they should operate it manually. I believe that when it is in operation it should be automatic."

On the other hand, Mr Jackman, the defendant's fire suppression expert, maintained that the manual part of it was important. He said:

"It has got to be a Pavlovian response. This is the object of training in this situation. It has got to be something that everybody is willing to just do in the event…The minute you get burning material coming out the bottom, I think everybody is going to be aware that there is something going on there shouldn't be, and it should be a Pavlovian response to go over and just literally hit the manual button."

285. In my judgment, Mr Jackman was plainly right on this issue. The system had been sold as providing both an automatic and a manual function. It was therefore axiomatic that everyone in the 'oil pop' production area should have been trained in the manual discharge of the CO2 suppression system. I find that they were not properly trained and that this was a glaring failure of the claimants' health and safety management system. This default was all the more egregious since Mr Hamilton had said, in response to the Group Risk recommendation of 27 June 2003 that sprinklers were the only reliable way of controlling fires, that the risk would be dealt with by way of "training" (paragraph 72 above). In this instance, at least, it manifestly was not.
286. From a practical point of view, there was another difficulty with the manual system which nobody seemed to think through: since the CO2 suppression system was located within the elevator and the hopper, and since those were enclosed, how would an operative ever have been able to see that there was a fire which required the manual activation of the CO2 system? That is a matter to which I return when I deal with the fire on 8 June 2005.

G7: Bagging

287. The popcorn, having passed through the elevator, was then discharged into the hopper. The hopper was about 1m square and rather more in height, with a tapered lower section running down to the release hatch. One operative – the bagger – would stand by the hatch, filling the sleeves with popcorn by hand. The sleeves were plastic and were about a metre high. Each contained about 60 litres of popcorn. Four such sleeves fitted into one tall cardboard box.
288. The evidence was that the bagger would fill the sleeves at the hopper with about 6 made-up boxes around him, waiting to be filled with completed sleeves. Once he had filled each of the 6 boxes with 4 sleeves, they would be collected together, shrink wrapped by hand, and put on a pallet. A pallet truck would then take them away. The boxes themselves had to be made up by hand. Mr Carter said that it was quite common, at the end of a shift, for 12 boxes to be made up so as "to get the next shift started". Thereafter, when each set of 6 was finished, a further set of boxes had to be made up. In this way, on the floor around the hopper would be a number of unmade up and made-up boxes. In addition, the various items used for this packaging, such as the plastic sleeves, the unmade up boxes and the shrink-wrap, were stored in greater quantities about 7 metres from the bottom of the hopper in racks against the south wall of old Production Area B.
289. There was no real dispute about any of this. One of the operatives who did not give evidence, did a drawing at the time of the fire which showed 50 or 60 boxes which (so he said) had already been made up for the purposes for shift use during the evening. Mr Shaw said that he did not recall that being the case. It is a number far in excess of the number identified by Mr Carter in his evidence. It seems to me unlikely that so many boxes would have been made up in advance of being filled with the completed sleeves. There was simply not enough room. However, there may well have been 50 or 60 boxes, flat packed, waiting to be made up and to be filled with sleeves, on the racks on the south wall.

G8: Cleaning

290. Mr Carter was asked about the regularity of cleaning. He said that the operatives cleaned between shifts, and did some cleaning as they went along. At the end of the shift, there was a clean using the power hose on the floor, and the floor was swept. The pans were cleaned out using the fire hose. The hoppers were also cleaned out using the fire hose but, as Mr Carter explained, that usually happened once a week, on a Friday, when there was a big clean down. Sometimes the hopper was also cleaned out if the products were being changed (i.e. a switch from salty to sweet popcorn).
291. For the avoidance of doubt, it seems to me that, on the evidence, the oil pop area was kept as clean as possible by the operatives, and nothing of significance turns on any alleged lack of cleanliness or oil residue.

G9: Conditions Generally

292. A number of the claimants' witnesses were asked questions on the assumption that there was a hot environment within the oil pop production areas. They denied that. Mr Barker said it was not particularly hot because there was an air extract system. He thought the conditions in that department were not a lot different to other areas in the factory. Mr Carter said something similar, describing the environment merely as 'warm'. He said it was not smoky because of the extract fans. It was reasonably clear of smoke and not really noisy.

H: PARTICULAR EVENTS POST-CONTRACT

H1: The VESDA System

293. As noted in paragraph 144 above, there was a fire in the hopper prior to the commissioning of the CO2 system. This fire, which occurred on 25 November 2003, had apparently started with a smouldering batch of popcorn entering the elevator and catching fire in the hopper. As a result, thought was given to the possibility of a system which would detect smouldering popcorn on the conveyor before it went into the elevator, it being common ground that the CO2 suppression system did not provide such protection.
294. The starting point for this process was the written Risk Assessment dated 26 November (the day after the hopper fire) prepared by Mr Bywater, the claimants' health and safety officer, and reviewed by Ms Irving. The hazard that was identified was "the possibility and frequency with which fires could occur due to burning corn in the cooking pans." The current control measures were said to include the installation of an automatic suppression system. That was a reference to the defendant's CO2 system. But amongst the control measures that were described as being "Under Investigation/Being Actioned" was the investigation of an "early fire detection system". This was the VESDA System, also supplied by the defendant.
295. At the meeting to review fire precautions on 16 December 2003, it was recorded that there was an ongoing investigation by Mr McDonough of the claimants into early detection options. The minutes referred to a demonstration of a possible system by the defendant, to take place on 19 December. The demonstration actually took place on 22 or 23 December. Mr McQuaid of the defendant showed how the VESDA equipment worked by passing a match under the detector. As a result, it was decided that the option would be taken forward and that there would be further trials in the New Year, as confirmed in the Project Kettle meeting notes of 23 December 2003.
296. On 2 February 2004, Mr McDonough told Ms Irving that he was waiting for the go-ahead to order a stainless steel hood so that the online trial could be carried out. Mr Hamilton immediately noted that this was outside the scope of the project work. Mr Barker said that, from this point on, he had no involvement in the investigation into VESDA; he did not witness the online test; and he was not involved in the decision whether or not to go ahead with the VESDA System.
297. The principal demonstration took place in March 2004. The results were set out in a document, which attached various graphs. Mr McDonough said that whilst he did not have the document, he saw the material, including the graphs, on the computer screen at the time of the demonstration. Mr McDonough confirmed that the purpose of the test was to see whether the VESDA System would do what it was designed to do, which was to detect and sound an alarm if there was smouldering/smoking/flaming popcorn on the conveyor. Mr McDonough agreed that, at the trial, the VESDA system was indeed able quickly to detect smoking/smouldering/flaming popcorn. He accepted the encouraging results of the test. He said that the VESDA System was very fast and 'was a fully competent system.'
298. On 23 March 2004, the defendant provided a quotation for the VESDA system in the sum of £6,960 plus VAT. There was no response to the quotation. At the claimants' health and safety review meeting on 4 May 2004, there was this entry:

"Early detection demo by ADT – arrangements for on-line trial demonstrated on 16/3/04 – cost in the region of £20k. Quote with AH [Mr Hamilton], to raise Ops Team. Need to clarify position with AH."

And there, the paperwork trail goes cold. The VESDA system was never installed. The defendant heard no more about it. Why not?

299. The claimants' evidence on that question appeared to be designed to suggest that the VESDA system was not pursued for technical reasons. The first difficulty with that argument was the absence of any record of any discussion as to the technical merits or defaults of the system; the absence of anything which suggested that the VESDA system was ever considered by reference to the risks that it was designed to minimise and the benefit that it might provide.
300. The second difficulty is the very thin nature of the alleged technical deficiencies in the claimants' witness statements. Mr Barker suggested that the problem with the VESDA system was that it would not deal with popcorn that self-ignited after it had passed the system, but that was hardly the point, since the VESDA system would at least provide detection at the point where it was installed, which was not presently covered by any fire prevention system at all. Furthermore, if there was a concern about a failure to protect the whole of the conveyor, then the answer lay in the sprinklers which the claimants' Group Risk Department had already recommended but which the individuals on site had apparently set their face against. And finally, given that Mr Barker said that he was expressly not involved with any consideration of the VESDA system, the factual basis of his evidence on this topic was unclear.
301. Mr McDonough, who was expressly dealing with the VESDA system on behalf of the claimants, had no explanation for why the system was not pursued. He appeared to suggest that it was Ms Irving who was dealing with this from the health and safety point of view, another attempt to pass the responsibility to somebody else in the management chain which was such a feature of the claimants' collective evidence. Mr McDonough accepted that it was "a business decision" not to proceed with the VESDA system, although who made the "business decision" was unknown. Although at one point he sought to argue that the VESDA system did not enhance the protection provided, he was forced to accept that "it could provide an early warning" and that, depending on the speed of reaction, such a system could be an enhancement to the fire protection to the factory.
302. Typically, Ms Irving was surprised when it was suggested to her that she had been identified as the person responsible for turning down the VESDA system. She said that it was a technical matter for Mr McDonough, Mr Eames and Mr Hamilton. She thought that it would have been discussed with the operations team, but it was not something that she recalled. The buck moved on.
303. In one sense, the most alarming evidence of all about the VESDA system came from Mr Hamilton himself. He too expressed real surprise at the suggestion that the decision not to go ahead with the VESDA scheme was his responsibility (despite the clear terms of the minute of the meeting of 4 May 2004 referred to at paragraph 298 above). The best Mr Hamilton could do was to say that VESDA was discussed but that it was decided – by whom he did not know – that the "spend was not going to give us a significant advantage and it wasn't to be progressed". The remainder of Mr Hamilton's evidence on this point consisted of an attempt to belittle the results of the VESDA trial: he repeatedly said that the system had worked "under the trial conditions", as if something very different had been expected or required. Since the trial of the VESDA system was just that, it was a little difficult to know under what other conditions Mr Hamilton thought the system might work. He then persisted in saying that there was no certainty that the VESDA system would have operated in practice, "no guarantee that it would have worked" at all. Again, since the claimants were dealing with the minimisation (rather than the guaranteed elimination) of risk, the references to certainty and guarantee were wide of the mark.
304. The fire suppression experts were also notably partisan on this topic. Mr Stephens suggested that the VESDA system was not particularly good because it would not detect all smouldering popcorn (even if, which he did not agree, popcorn could smoulder). He was also obliged to argue that the system was not installed because the defendant did not press the benefits of VESDA sufficiently with the claimants. Mr Jackman, on the other hand, went so far as to say that, without VESDA, the whole fire detection system installed by the defendant was potentially flawed, which immediately begged the question why that had not been pointed out to the claimants at the time. I do not accept either of these partisan views.
305. Based on the material set out above, I make the following findings:

a) The VESDA System would obviously have enhanced the fire protection provided by the other systems already in place. In particular, it would have provided a way of warning the operatives if smouldering/smoking/flaming popcorn had got onto the conveyor without them noticing. There was no system in place, other than a reliance on the operatives' alertness, to deal with that possibility.

b) The VESDA System was demonstrated to work satisfactorily. There was therefore no technical reason why the system should not have been installed. Neither was there any detailed discussion of the technical pros and cons of the system by the claimants following the successful trial in March 2004.

c) The VESDA System was not progressed either because the claimants had decided not to spend any more money on the NMU or because, as happened with a number of other matters, it just got lost in the claimants' usual blizzard of email exchanges, questionnaires, review meeting minutes, and corrective action notices.

306. The VESDA story only confirms the view that I have already reached about the claimants' attitude to health and safety generally, and the fire risks posed by the 'oil pop' process in particular. They were really not very bothered about the practical details of what risks the process created and what might be done to address them. The main imperative seemed to be to spend as little money as possible.
307. The fire in the hopper in November 2003, which began the process of enquiry into the VESDA system, was started by smouldering/smoking/flaming popcorn getting onto the conveyor and into the hopper. It seems very likely that the fire in the hopper in June 2004 (paragraphs 267-269 above) started in the same way. The catastrophic fire in June 2005 was also likely to have started there. There must have been a reasonable chance that, if the VESDA system had been installed in 2004, the fire the following year would have been detected and put out before it caused widespread damage.
308. But it is not possible to allow the defendant the full benefit (as it were) of the claimants' failure to install the VESDA system. Mr Jackman's evidence, that the VESDA System was a crucial part of a proper protection system and that without it, the CO2 suppression system was potentially flawed, was a view which, even if exaggerated, should have been shared with the claimants. It was not. Of course, it was not for the defendant to put pressure on the claimants to enter into a further sub-contract with them, but both parties should have realised that, since the risk of burning popcorn getting onto the conveyor was not even going to be addressed by the existing system, the system would have been significantly improved if a VESDA system was in place, and the defendant should have been clearer in its dealings with the claimants on that issue.

H2: The Maintenance Contract

309. The demonstration of the VESDA system was not the only ongoing contact between the claimants and the defendants in the months after the contract works had been completed in December 2003. The evidence was that the claimants were relying on the defendant in relation to the maintenance of the existing systems at Monkhill and that, eventually, the parties entered into a maintenance contract in writing which incorporated the defendant's terms and conditions, and which covered the CO2 system
310. Mr Webber, the engineering manager, claimed that he was not responsible for the maintenance of the CO2 system. He said that this function was sub-contracted to the defendants. He was, he said, simply the liaison person if there were any faults with the system. As a result, he suggested that there was nobody at Monkhill who had responsibility for maintenance issues in relation to the CO2 suppression system. Mr Webber also said that information relating to the system would have come to him, in stages, as the project was completed. He said that he might have read the maintenance manuals provided by the defendant but he could not recall them.
311. He was taken through a number of the extracts from the defendant's manual and users' handbook. He agreed that he did not carry out weekly checks as set out in the manual, although there was a maintenance procedure in place for the main items of engineering equipment. Thus any damage to the equipment associated with the elevator would have been recorded on the claimants' planned maintenance checks. He was unclear as to what the hazards and the enclosures were that were referred to in the maintenance manual. As to checking that personnel had been properly trained, he thought that would have been done, but not through him. He said that this was a matter for the production manager. He was not involved in the setting up of the maintenance agreement. In addition, Mr Webber was also asked questions about the claimants' own guidance documents but he made it clear that he had not seen those documents either.
312. Mr Webber's evidence was that the whole question of the placing of the maintenance contract with the defendant was taken out of his hands because there had always been a maintenance contact in place with the defendant which had been agreed by Mr Eames. That was wrong. The documents made clear that there was no maintenance contract at all until July 2004. Then, the omission was addressed, not by anybody on site, but by the claimants' Group Risk Department. They sent a Property Risk Assessment to Ms Irving on 19 July 2004. That sought confirmation that, amongst other things, there was a routine maintenance contract for the CO2 and wet chemical systems. At that stage, there was no such contract. How that had come about was never explained. Accordingly, following the involvement of Group Risk, a maintenance contract was let by the claimants to the defendant in August 2004.
313. The defendant sent the claimants a maintenance quotation on 4 August 2004. It was in similar form to the quotation of 28 August 2003 with the critical difference that, on this occasion, a copy of the Commercial Service Agreement, which contained the defendant's terms and conditions, was actually sent with the quotation. That Agreement contained a place for the claimants to sign and, on 12 August 2004, Mr Eames signed the Agreement, accepting the quotation. There was no purchase order. And when the defendant sent in an invoice containing their old 6 digit vendor number, the invoice was processed and paid. Accordingly, as I have already found, the mere fact that there was a vendor number did not mean that the defendant had agreed to the claimants' terms and conditions; on the contrary, Mr Eames' signature on the Commercial Service Agreement made clear beyond doubt that the maintenance contract was let by the claimants to the defendant on the defendant's terms and conditions.

H3: The Hopper Modifications

314. The original hopper was thought to be about 1 cubic metre in size (actually, it was probably slightly bigger than that). In about June 2004, modifications were made to the hopper. Mr Barker said:

"One of the reasons why we decided to change both hoppers was because the metal detectors on top of both of the existing hoppers didn't work. So it meant we couldn't effectively detect metal when we were sending packs out. So it was unacceptable to continue working like that. And the hoppers, I believe one of them was a mild steel construction, which again was unacceptable in a modern food factory. So we took the decision to upgrade the hoppers, the metal detector system, to make it compliant."

315. It appears that there was also a problem of bridging, in that sometimes there was a blockage at the outlet at the foot of the hopper when the popcorn formed a bridge between the sides of the hopper and needed to be broken up, in order to be released. Again, the modifications to the hopper were designed to deal with this problem. Those modifications resulted in a bigger hopper, about 1500mm wide by 1450mm high, tapering to the conical section. They had been completed by the end of 2004. There was no suggestion, and no evidence, that these modifications had any effect on the CO2 suppression system, or the fire in June 2005.

H4: The New Elevator

316. In the autumn of 2004, the second claimants made another Capex Application in relation to an upgrade to the C line in the 'oil pop' production area. Mr Barker was named as the project manager. The project was designed to improve efficiency. Amongst other things, it meant that the elevator would be larger in size, although not significantly.
317. The involvement of the defendant in these modifications was very limited. As the costing made plain, their additional work was estimated at £1,200. It was envisaged that they would simply disconnect the existing CO2 system in the Gough elevator and then reinstall the same system once the modified Gough elevator had been installed. The basis for the £1,200 came from an email from Mr Taylor of the defendant to Mr Barker dated 9 September 2004. That was in these terms:

"Further to our recent discussion regarding the above we confirm our quotation in the sum of £1,200.00 to refix the existing detector and nozzle into 1no conveyor. Should the conveyor be increasing in size we would recommend a redesign to ensure that there will be sufficient gas to extinguish a fire. When you have one, please forward a copy of a drawing of the new conveyor detailing all dimensions and we will check this and advise accordingly. We trust that as before the access equipment will be supplied by yourselves."

318. Mr Barker alleged that he had concerns about installing a bigger elevator and he was asking Mr Taylor for advice about whether the claimants should improve or upgrade the CO2 suppression system. That of course is the opposite of what was said in the defendant's email. Mr Taylor said that, on the contrary, when Mr Barker called him and explained he was increasing the size of the conveyor, it was he who recommended that a redesign of the CO2 system should be considered, not Mr Barker.
319. Mr Barker said that he confirmed the dimensions of the new conveyor over the telephone, although there was no record of such conversation. Mr Taylor said that he did not recall any such conversation. I am inclined to agree with Mr Taylor: in the absence of any documents, I find that this conversation did not take place. However, it seems that, at some point, Mr Barker did send a drawing of the new elevator to the defendants.
320. On 26 November 2004, Mr Taylor sent Mr Barker a further email in these terms:

"The additional cost per visit to isolate the fire alarm to allow the installation of a new conveyor will be carried out for the total sum of £350.00 per visit.

We have sent the drawing of the new conveyor to our design department who will verify if the existing cylinders are adequately sized to accommodate the new conveyor. Any cost implications will be advised under separate cover asap along with additional costs to allow for 4 additional nozzles for the popcorn pans."

It appears that this email was resent on 22 December.

321. Mr Barker's evidence was that, not only had he asked Mr Taylor for an assessment as to whether the claimants needed to upgrade the existing CO2 system, he could not understand why Mr Taylor never came back to him to say whether or not the existing system was adequate. Mr Taylor thought that he would have given the drawing to the defendant's design manager for assessment. He would not get involved further: he would expect the design department to liaise directly with the claimants if anything else was necessary. There was no evidence that this ever happened.
322. Accordingly, not for the first or last time in this case, it would appear that a matter of potential importance had been raised between the parties, and then allowed (by both of them) to wither away. The claimants ought to have followed up on Mr Taylor's emails to get confirmation that the old CO2 system would work satisfactorily in the new elevator. As often happened, they failed to do so. But, as Mr Taylor accepted, somebody at the defendant's Leeds office – most likely somebody in their design department – should have got back to the claimants to tell them whether or not the old system was adequate for the new elevator, and failed to do so. Happily, at least in this instance, this failure had little effect; there was no cogent evidence that the changes would or should have had any significant effect on the design or installation of the CO2 system.
323. However, it should be noted that consequences of the installation of a new elevator were not limited to the unsatisfactory petering out of the issue as to whether the CO2 system needed to be modified. The evidence was plain that, although the sensors were disconnected in December 2004 to allow the works to the elevator to be carried out, they were not reconnected until May 2005, just a month before the fire. This time, the responsibility for the default lay entirely at the door of the claimants.
324. Mr Barker said that one of the reasons for the non-reconnection of the sensors was that the work had to be carried out when the production lines were not operational and it was difficult to find a time for that to happen. But the result was that, for over 5 months from December 2004, the CO2 system could only be operated manually (if anyone had known how to do that properly: see Section G6 above). In his evidence, Mr Barker sought to say that he had made every effort to get the system reconnected and, at least at one point, he suggested that Mr Webber was to blame for the failure. On analysis, that again turned out to be wrong.
325. Mr Webber confirmed that between December 2004 and May 2005 the sensors were disconnected. He did not know whether the claimants' Group Risk Department had been notified: such notification seems unlikely. Mr Webber said it was not his responsibility to report it; he said that it was Mr Barker's responsibility to deal with the information regarding the outstanding works after Christmas 2004. He said: "It was his remit…it was not my responsibility to communicate any information as far as I was aware, no." Although Mr Webber sought to say that the CO2 system could not be described as being out of action because it could still be operated manually, this overlooked the point that there were no proper instructions as to the manual operation and very few people knew how to operate it.
326. As to the detail of this ongoing failure, it appears that the existing carbon dioxide sensors were disconnected in December 2004. On 3 February 2005, Mr Webber sent Mr Barker an email saying that he had been in the roof space and had noticed that some of the defendant's wires were exposed, as though they were waiting to be reconnected. He asked Mr Barker if he knew about this. Mr Barker replied to say that the wires were part of the CO2 fire suppression system which was to be reinstated next week. However, that reinstallation did not happen. Mr Webber said that this was because the second claimants would not stop the production line so that the sensors could be reconnected. He said that there was a large amount of work that needed to be done when production was not operating, and that it was hoped to do all the outstanding work in a two day window in the future. It was pointed out that the reconnection of the sensors only required a shutdown for an hour or two.
327. The following week, on 10 February 2005, Mr Webber sent Mr Eames an email which made it plain that, for completely different reasons, he was stopping production to carry out various items of maintenance work. It was put to him, therefore, that only a week after his concerns about the sensors, he had booked a production shutdown for other matters, and that this would have been an appropriate time to reconnect the sensors. He agreed that it may have been, but he said that this was something that was being organised by Mr Barker, and not for him to arrange. He said: "at this point I had no communication link set up with any of the people involved in the project work, so at this stage it was within Steve's remit to organise the work, not myself." That answer, of course, was one of many which demonstrated a lack of basic coordination between the claimants' employees.
328. On 2 March 2005, Mr Barker sent an email to Ms Carol Morgan, who was in charge of production, to say "we really need to complete the engineering work on C line oil pop which has been scheduled and postponed twice now." He suggested 17/18 March. That was refused but instead dates of 14 and 15 April were provided. This was therefore a 2 day shutdown which, as Mr Webber confirmed, involved much more than the reconnection of the sensors. Although it appears that the 2 day shutdown took place on 14 and 15 April, the sensors were not reconnected, because nobody at the claimants contacted the defendant to ask them to go to Monkhill to undertake that work. Again, there was no explanation for that omission.
329. On 21 April 2005, Ms Priestley sent Mr Webber an email, copied to Mr Barker, asking for a date for making the fire suppression system operational. It seems that this email finally spurred Mr Webber into action, and he contacted the defendant. It is noteworthy, as Ms Rawley pointed out to Mr Webber during his cross-examination, that Mr Webber did not contradict Ms Priestley's observation that "we have not got a working system" by suggesting that the manual system was still in place.
330. The defendant was therefore booked to come on 29 April, but the engineer was unable to make himself available on that date and instead visited on 6 May 2005. By that date, there had been a false activation of the fire alarm because the CO2 control panel was cracked. It appeared that water had got into it. Mr Webber had thought it possible that the panel had been cracked because somebody had trodden on it whilst they were climbing up or down to the gantry. He said the panel had pulled away from the backing plate on the wall and the water had got in that crack. It does not appear that this latest problem was passed onto the defendant so, when the defendant's engineer arrived on 6 May, he had to ask for assistance from a service engineer to come and look at the panel.
331. Mr Vernon, the engineer who attended on 6 May, said that not one but both of the CO2 panels were lit up "like Christmas trees". In other words both the C Line panel (which was not cracked) and the D Line panel (which was) were lit up and he discovered that the hopper sensor connection wires had been pulled out of the sensor and just left dangling on the C Line hopper. He said that this was why the C Line panel was showing so many fault lights. Mr Webber confirmed that. He was unable to explain why that had not been reported to the defendant earlier.
332. As noted above therefore, the reconnection of the sensors was a lengthy story of muddle and inefficiency on the part of the claimants. It again demonstrated that, on site, the claimants appeared blithely unconcerned that their only fire prevention system in the 'oil pop' production area was out of action for so long. That immediately raises a wider concern: for all the paper that they generated, just how good were the claimants' health and safety arrangements? The answer, as explained in greater detail in the next section of this Judgment, was: not very good at all.

H5 Heath and Safety Issues

a) Popcorn Fire Safety Advice

333. The only written guidance dealing specifically with fire risk in the popcorn process was a document prepared by Ms Irving in December 2003 entitled 'Popcorn Fire Safety Advice and Evacuation Procedure'. The relevant parts of this document read as follows:

"Minor fires

If a minor fire (e.g. burning piece of corn) is discovered, turn the gas off to the individual pan

If it is in a pan, close the lid and leave it for 10 minutes – removal of oxygen will allow it to burn out

If it is on a table, belt or conveyor, stop it and tackle the fire immediately with a foam extinguisher, as per your training

If it is an electrical fire a CO2 extinguisher should be used

In all cases inform a team leader/manager immediately

An Incident Report will be completed by a Team Leader or Manager. This will allow for monitoring of all incidents

Disposal of product: affected product should be disposed of in the Pig Food skip in line with Clean Up procedure/work instruction

Serious Fires

The area is fitted with heat sensors set at 90°C. If they are activated (either automatically or manually) they will set off the alarms in the whole of NMU but will NOT set off the line fire suppressant systems. However if the line fire suppressant systems are activated by any means, then these will activate the building fire alarms.

In the event of a serious fire (e.g. the canopy is ablaze) the gas and power to all kit will automatically become isolated when the sensors are activated, apart from extraction of the air handling system which will clear smoke. This will activate the main NMU fire alarm system."

334. Ms Irving was asked about this document (whose distribution and publication history were unclear). It was put to her that the procedures noted here were not in accordance with those which were in fact adopted by the operators (such as using cold oil in the pans or dumping the popcorn on the floor and extinguishing it or stamping it out). She said that she was unaware of that, asserting that this was the document that would have been 'trained out' and that she would have expected the operatives to have followed this. Although Mr Webber told me that he had seen operatives sweeping the burning corn off the conveyor and extinguishing it on the floor, Ms Irving maintained that she did not know that that was going on. She accepted that this procedure represented a risk of fire taking hold in the combustible materials. She said that it was unfortunate that employees were "acting outside their training".
335. The other important thing to note about the Procedure document is that it only addressed pan fires. It made no reference to fires in the hopper or elevator, and it was not subsequently updated after the hopper fire in 2004. And even in respect of pan fires, it does not appear that the procedures identified by Ms Irving were followed, or that this document received a wide circulation. Instead, a series of ad hoc general procedures were adopted, detailed below.

b) Approved Fire Procedures

336. Ms Clough, who was a trainer at Monkhill, gave evidence about the approved fire procedures actually in place for the different kinds of fire. She said that if there was a small hopper fire, the approved procedure was to open the chute, let the burning popcorn drop to the floor and stamp out the fire. If it was more serious, it was left to the CO2 system to put it out. The judgment as to whether the hopper fire was serious or not was to be made by the team leader or the supervisor. She said that she had not been involved in training the operatives in the event that the CO2 system failed to discharge.
337. As to other types of fire, she said that if there was a fire in the pan, the procedure was to turn off the gas and either leave the pan lid down or tip cold oil on it. It was also possible to empty the pan onto the floor and stamp out the burning pieces, but that was not the preferable course. If there was a fire on the conveyor, Ms Clough said that, if it were possible to tap it out with a spoon, before it got any further, then that was the approved procedure. Otherwise it would again be knocked onto the floor and stamped out there.
338. Although these were the procedures which Ms Clough told me were the approved procedures for fires, and they more or less reflected what the operatives actually did (as set out in Sections G3 and G4 above), nowhere were these approved procedures set down in writing. This was a curious omission. It was one which Ms Clough endeavoured to make good in April 2005, by means of a questionnaire that was put to the operatives, to check and, if necessary, correct their answers. In fact, this process raised more questions than it answered.

c) The Test Questionnaire

339. The questionnaire had a crib sheet. The relevant section of the crib (ie the approved policy) read as follows:

"PAN FIRES:

Q: What is the procedure if a pan fire occurs?

A: Turn gas off, turn table chute to face outwards, tip pan and allow the corn to fall into the pig food bin.

Q: What extinguisher is user for a pan fire?

A: Red = foam

HOPPER FIRE:

Q) What is the procedure if a hopper fire occurs?

A) The hopper has an automatic alarm, but the Team Leader or Manager must be alerted"

340. These short answers were not comprehensive. For example, as Ms Clough agreed, there was no procedure for dealing with a fire on the conveyor. Furthermore, Ms Clough said that, if there was a hopper fire, the purpose of alerting the team manager or team leader was to let them know what was happening. If the alarm went off, then the operative had to follow the fire alarm procedure. She said that, although she was aware that the system went off automatically (which is not referred to in the crib) she readily agreed that she did not really know anything else about how the CO2 system operated. There was no reference in the crib to the manual activation of the system.
341. Some of the operatives' answers to these test questions were looked at during the evidence. They displayed a range of answers, which were not always in accordance with the limited crib. Thus a Mr Lloyd said that, if there was a hopper fire, the answer was "automatically work. Open chute and contain on floor". Ms Clough said that was a satisfactory answer because it showed that he knew that it worked automatically and that he knew not to try to tackle the fire himself. That last answer is a little difficult to follow because it not only appeared to negate the utility of the manual system, but it was also inconsistent with the remaining answer, that the operative would open the chute and contain the fire on the floor (presumably by stamping out the burning popcorn).
342. Another answer, which Ms Clough accepted was unsatisfactory, was from a Mr Pierce who said that if a hopper fire occurred the answer was to "set the fire alarm off". Similarly, the answers of a Mr Shaw appeared to reveal that he did not know some of the approved procedures, and so was then told about them by Ms Clough. Perhaps most tellingly of all, Mr Carter's sheet suggested that he was not aware of the manual activation system. Ms Clough agreed that that would be a matter of concern, although it does not appear that, at the time, it was regarded as such, or that anything was done about it.
343. All of this evidence had one thing in common: all of those involved with the 'oil pop' production process, and the fire risk that it represented, did not believe (and did not operate on the understanding) that the CO2 suppression system would automatically put out every fire that may occur in the hopper or elevator. On the contrary, it was correctly assumed that the fire might escape from the elevator or hopper without having been put out by the CO2 suppression system. In my judgment that was a sensible and realistic assumption. It confirms my rejection of the claimants' case that the defendant offered some sort of warranty or guarantee that this would not happen^[17].
344. The other witness to give evidence about the test questions was Ms Priestley, one of the shift leaders. She was also asked about the crib and the significance of informing the team leader or manger if there was a hopper fire. She appeared to think that this was because, if there was such a fire, there would be a certain clean-up operation and an investigation, and the team leader would "normally be the person that would fill that form in and pass information on to me." She agreed that the team leader had to know how to react to a fire in the hopper. But she also accepted that she was not in a position to tell the team leader how to activate manually the CO2 system, and that she did not investigate with the team leader what that team leader's knowledge of the operation of the CO2 system actually might be. She said she assumed that the team leaders were aware of how to activate the system manually. As noted in Section G6 above, that was not a safe assumption.
345. In addition, Ms Priestley said that she was always aware that there was a significant risk arising out of the 'oil pop' operation, in part because she knew that the previous popcorn factory in Leeds had burnt down. She said:

"It was open gas flames. We didn't have any process similar to that, so the very process was probably a little bit daunting, because it was an open gas flame with an oil pan on."

She agreed that the process represented a significant risk.

346. The questionnaire process was the subject of an email from Ms Irving after the fire in June 2005. She noted that, out of the 18 employees who completed the questionnaire, the majority answered the question about the fire in the hopper to the effect that the CO2 system automatically went off. The email was (rightly) concerned with how and why it was that the operatives were unaware of the manual element of the system. Accordingly, Miss Irving referred to the Corrective Action Notice identified at paragraph 276 above. The report on this point ended:

"It was reported by Pat Priestley that a decision was made only to show the Team Leaders this, and David Carter confirmed that Steve Barker demonstrated how to manually activate the system."

d) Fire Incident Reports and Similar Documents

347. A number of witnesses said that, if there was a fire, it would generate an incident report form. There were 16 such forms between December 2003 and March 2005. On the basis of the evidence about frequency, referred to above, it therefore seems that not every fire was the subject of such a form. For example, there were no reports between December 2003 and June 2004, although the evidence was that there were pan fires during that period. The reports, which were looked at during the evidence of Ms Priestley, demonstrated that the most common source of the fire was in the pans. The report of 18 December 2003 ascribed the fire to the operator being distracted; the report of 14 June 2004 stated that the pan had overheated because the operative had gone for a break not realising that the pan was still on. Indeed, it appeared from these reports that leaving the pans on when going for a break was a reasonably common cause of fire. It was, for example, the cause of another fire on 20 January 2005, and the reports show that the relevant operative on that occasion had caused a number of similar fires in the past.
348. The other kind of formal report created was the Risk Assessment document originally produced by Mr Bywater of the Health and Safety department. The only such documents in the bundle were those generated by the fire on 26 November 2003 (see paragraphs 267-269 above). As Ms Irving explained, these documents identified a number of risks, not limited to fire. It was these documents which said that one of the control measures to be put in place was the installation of the VESDA System. Although the document indicated that that item was completed, as we know, it did not ever happen.
349. It was put to Ms Irving that the documents referred to above comprised the entirety of the Fire Risk Assessment material relating to the 'oil pop' process. Her quibble with that question was that, in her view, these documents did not comprise a Fire Risk Assessment as such. Ms Irving said that, from her perspective, there was no formal Fire Risk Assessment process and there was no formal HAZOP^[18] undertaken in relation to the 'oil pop' process. She said that, as a result, she accepted that there was no structured or systematic evaluation of the 'oil pop' process to identify risks to personnel or equipment in a HAZOP format. This omission was the subject of internal criticism after the fire in June 2005.

e) Fire Risk Assessments

350. I consider that Ms Irving was right: there were no formal Fire Risk Assessments in the relevant period. She had carried out one such Fire Risk Assessment at the Monkhill site, but that was dated 12 May 2003, months before the oil pop process was moved there. It was in the claimants' pro-forma. But as she confirmed, there was no subsequent Fire Risk Assessment of the new 'oil pop' process at Monkhill. Although I had wondered whether Mr Stephenson's report of 27 June 2003 (paragraph 70 above) was intended to be such a document, Ms Irving was at pains to say that it was not a Fire Risk Assessment. She had no involvement in it, and it was produced before the 'oil pop' process was moved to Monkhill, so it was irrelevant anyway.
351. Ms Irving was aware that it was the claimants' legal obligation to carry out a formal Fire Risk Assessment. In order to do that they had to appoint a competent person to carry out that process. She said that it was not formally stated anywhere that she was the competent person to undertake such an assessment, and she accepted the limitations of her knowledge and expertise. She said that this was why the claimants sought expertise from elsewhere, effectively from their own Group Risk Department. Accordingly, it was in the documents generated by the Department that something at least approaching a Fire Risk Assessment might be found.

f) The Group Risk Report of June 2004

352. I have already referred to the claimants' policy, by which their Group Risk Department carried out periodic risk assessments of their premises. In June 2004, the Monkhill premises were the subject of such a report. It was sent to site and, following a discussion between Ms Irving, Mr Eames and others, some of the items were then the subject of what was called a 'Site Response'; prepared by the Monkhill staff.
353. The June 2004 report summarised the construction of the NMU in these terms:

"The NMU is 70% foamed plastic insulated metal panel and 30% rockwool insulated metal panel. It appears that a significant amount of the foamed plastic is polystyrene, with some usage of fire rated Kingspan reported."

Ms Irving confirmed that this information did not come from her, because she did not know what the construction was. This broad 70%/30% division was probably based on the (unsupported) assumption that the old part of the NMU had been built using foam panels, whilst the extension had been built using rockwool panels.

354. The June 2004 Report also dealt with the fire risk posed by the plastic panels, under the heading 'Combustible Construction'. The recommendation was in these terms:

"Manage the fire risk associated with polystyrene insulated hygienic finish panels as follows:

a) establish a register of hygienic finish panels with polystyrene (or polyurethane) insulation

b) ensure that all new or replacement panelling installations use fully non-combustible materials

c) where feasible replace polystyrene insulated panels on a phased basis or as part of future projects

d) ensure that as far as possible permanent or temporary (e.g. hot work) ignition sources are kept away from areas with combustible construction and that protective procedures are fully implemented when work is necessary in these areas

Comments: it is estimated that > 60% of the NMU is constructed using polystyrene insulated composite panels."

355. Although it is clear that the Group Risk Department was concerned about the fire risk posed by the construction of the NMU, the Site Response was, in the circumstances, somewhat lethargic. It proposed that a Corrective Action Notice would be raised "for a register/plan to be established by end of 2004." In the event, no such register was provided. Instead, Mr Barker engaged Craig Engineering and subsequently produced the unsatisfactory sketch referred to in paragraphs 37-38 above. Ms Irving said that she did not chase up afterwards to ensure that there was a register of external or roof panels, although she accepted that that would have been useful. She thought that the reason why it had not been done was possibly because Mr Barker had interpreted the question as relating only to internal panels.
356. It is not unfair to observe that the sketch could have been produced in 5 minutes, by somebody who was working on the simple assumption that the old part of the NMU was built of plastic panels, whilst the new part of the NMU had rockwool inserted in the panel cores. In my view, even if the sketch was accurate (which I do not accept, for the reasons noted in paragraphs 37-38 above), it fell well short of the Group Risk recommendation to establish a register of hygienic finish panels. Furthermore, given the reference in the report to the dangers of hot work, nobody could seriously have thought that the request did not go to both internal and external panels, and roof panels as well. Another opportunity to discover that there was, in fact, no 'fire box' around the 'oil pop' production area, as the claimants had always assumed, was provided by Group Risk, but not taken by the Monkhill management.
357. The Fire Risk Assessment then went on to address the Fixed Automatic Fire Suppression System. That was described in these terms:

"Recommendation: provide an independent, reliable fire water supply and sprinkler system for the site in accordance with the Cadburys Schweppes standards…

The loss expectancy with what was called "the deficiency" in place (ie not having sprinklers) was estimated at £16.9 million by way of property damage and £12.2 million by way of business interruption, a total of £29.1 million. With sprinklers, the report estimated that the loss was reduced to less than £1 million. It went on:

Supporting notes: without automatic fire suppression installed a fire may be expected to spread throughout the NMU due to the continuity of combustible materials of polystyrene panelling in the building construction.

A fire with a correctly operating, reliable fire water supply and automatic sprinkler system would be expected to be limited to 300 square metres. Loss estimates based on a value of £1500/square metre and loss of the air popping operation for 1 month.

Estimated cost to complete the recommendation is £500,000."

358. This was the second time in a year that Group Risk had positively recommended that sprinklers be installed at the NMU. Unfortunately, the recommendation was again not acted on. The response was in these terms:

"Site decision to control the risk of fire through fire prevention, protection, and segregation, as agreed with local fire service. The area of highest risk, oil pop, is covered with a fire suppression system."

I have already pointed out, at paragraph 132 above, that the fire detection and prevention systems in the old Production Area B had not been agreed with the local Fire Service. This was therefore a wholly misleading and inaccurate response.

359. Even on the most basic level, it was also wrong to say that the 'oil pop' process was "covered with a fire suppression system". Although Ms Irving maintained when asked about it that "the areas where the fire may be caused were covered with suppression systems which we were advised would meet our requirements", that was incorrect. The most obvious example of something that was not covered at all was the risk of burning popcorn getting undetected onto the conveyor, which was the cause of at least two fires in the past and would prove to be the likely cause of the hopper fire in June 2005. There was no detection system in place to address that risk.
360. Notwithstanding these errors, Ms Irving said that this particular site response was hers, albeit based on the input of others. She maintained, however, that it was not her decision not to put in sprinklers. She said that that "would be a business decision". Furthermore, although Ms Irving thought there would have been a consideration of the cost of installing a sprinkler system, measured against the potential benefits that it might bring, she could not recall such a decision being discussed or reached. There is no document or other evidence to suggest that there was any such consideration.
361. Accordingly, I find that, just as in 2003, those on site at Monkhill did not consider (properly or at all) the Group Risk Department's recommendation that a sprinkler system be installed. I consider that was partly for budgetary reasons and partly because, since they had never thought through all of the risks involved in the 'oil pop' process, they had persuaded themselves of the erroneous belief that the CO2 suppression system was all that they needed.
362. During the evidence, the issue of sprinklers was also taken up with Mr Caldicott, from the Group Risk Department. I have already referred to some of his evidence as to the inherently risky nature of the decision to move the 'oil pop' process to Pontefract^[19]. He was asked about Mr Stephenson's email to the Monkhill staff, sent just after the report of June 2004, which said that the recommendations could usually be split into three basic categories. The third category was described in these terms:

"Risks that the BU [Business Unit i.e. Monkhill] has investigated, discussed and agreed to 'accept the risk'.

For example – often smaller sites decide not to install sprinkler protection. In this case the risk of fire and impact on BU is weighed against the cost of implementation (e.g. sprinkler installation) and a decision is made. This is documented and reviewed regularly for change and latest management teams continuing acceptance. In this case the recommendation remains on the report, as the risk to CS and our insurers remains."

363. A number of points arise out of this email. First, Mr Stephenson envisaged a process whereby the risks of not installing sprinklers would be weighed against the cost of installation. That balancing exercise was never attempted, beyond the report itself, which showed that a £500,000 spend would reduce a £29.1 million projected loss to less than £1 million. The email also recommended that the balancing exercise and the decision should be "documented"; but obviously that did not happen either. Thirdly, the email said that the decision was to be reviewed regularly. That also did not happen. As a result of everything that happened (or failed to happen), and as Mr Caldicott expressly agreed in cross-examination, I find that the claimants agreed "to accept the risk" of not having a sprinkler system at the NMU.
364. Also on the subject of sprinklers, there was some expert evidence from Mr Stephens and Mr Jackman. Mr Stephens said that "one would always appreciate sprinkler systems in a building such as that" and he said that he would probably have advised the claimants, as he would have advised every other industrial factory owner, to put in a sprinkler system back in 2003. He went on to say:

"Well, as someone who specialises in active protection measures, it is always preferable to see industrial buildings sprinkler protected and, indeed, sprinkler protection is one of my speciality areas."

Mr Jackman said there was a need for sprinklers, although he said that the risk identified by the claimants' Group Risk Department was "more to do with the polystyrene structure which was the old NMU before the extension was put on. I think that was the main reason for the sprinkler system being recommended."

365. In the Group Risk report of June 2004, the 'oil pop' production area was the subject of additional express recommendations, as follows:

"Improve the protection and operation of the oil popping area as follows:

a) Connect the carbon dioxide and wet chemical extinguishing systems to the site alarm systems to indicate system discharge and protection isolation

b) Confirm that there is a routine maintenance contract on the carbon dioxide and wet chemical extinguishing systems

c) Ensure the gas piping, gas installation valves and emergency stop controls are clearly labelled."

The Site Response to this was in these terms:

"a) During the process of isolation, the operating procedures where [sic] shared with the local fire service, and a local decision made to operate systems separately to minimise unnecessary business interruption. This has operated successfully to date

b) CA raised to ensure appropriate maintenance contract is in place by end of 2004

c) CA raised to ensure labeeling [sic] is in place by the end of AP9"

366. I have already made the point that the suggestion that the operating procedures were 'shared' with the local Fire Service was inaccurate^[20]. I have also noted that it was this exchange which led to the maintenance contract being placed with the defendant.

g) Summary on Health and Safety Issues

367. As a result of this analysis, I make the following findings about the claimants' health and safety regime at Monkhill prior to the fire:

a) The claimants had never sat down and identified precisely the risks and hazards posed by the very unstable process of cooking popcorn using naked flames. They had therefore never undertaken any establishment of standard operating procedures.

b) The closest that the claimants come to a comprehensive review of the fire risks inherent in the process was the Group Risk Fire Assessment Report of June 2004. That made a number of recommendations which were deliberately not followed by those at Monkhill. The stated reasons given for not following that clear advice were based on erroneous assumptions and were technically unjustified; the real reasons were a failure to consider the advice with sufficient care and a desire to spend the minimum.

c) Because of the absence of a comprehensive review of fire safety, and because of the slack management to which I have previously referred (which meant, amongst other things, that there was no proper system of reporting of fires), there was a general absence of clarity in the safety procedures and in the training of operatives. This was evident in the different answers given in answer to the questionnaires, and the fact that the actual practice in dealing with the fires differed, at times markedly, from the approved procedures.

d) There were particular deficiencies in relation to the training of the operatives in the operation of the CO2 system. Very few people knew how it worked: even Ms Priestley accepted that she had never had any training in its operation.

e) Such training as had been provided was given to some team leaders only, and in an ad hoc way, so that even Mr Carter, who had had some training, did not feel confident that he understood precisely how the system worked. There was never any instruction manual or photographs showing the manual operation of the system, despite the recognition that this was required.

368. A number of these deficiencies were identified by the investigators appointed to look into the fire in June 2005. In short, as is so often the way with health and safety matters, there was a gap, at times a gulf, between the reality on the production factory floor, and the bureaucratic computer-generation of questionnaires, forms, memos and notices. There was an illuminating passage in Ms Priestley's cross-examination, highlighting this gap, when she was asked how Ms Irving and Mr Bywater, the health and safety managers, would actually have known about the risks of fire occurring in the production line. Ms Priestley answered:

"I can't answer for what they knew really. And I can't remember any specific conversation that I had with him on that."

In my view, this gap between reality and form-filling, had a part to play in the events of 8 June 2005.

I: THE FIRE ON 8 JUNE 2005

I1: The Evidence

369. There were a number of different sources of evidence relating to the fire. The principal ones were:

a) The written and oral evidence of Messrs. Carter, Norton, Shaw, Challinor and Kear. Those employees were all present in the 'oil pop' production area at the time that the fire was discovered. Their evidence about what happened was broadly consistent and I consider it to be the most reliable evidence about the fire that I received.

b) The notes made by those investigating the fire, such as the West Yorkshire Fire Service and Dr Mitcheson, including records of interviews with those present. Many of those interviewed subsequently provided witness statements and gave oral evidence. Accordingly, the notes of the interviews are really only of assistance where they concern interviews with individuals who did not give oral evidence, and the major drawback with that evidence is that it was never tested in cross-examination. It must therefore be treated with a certain amount of caution.

c) The evidence from those who investigated the fire in the days and weeks after it had happened. This culminated in the report from the West Yorkshire Fire and Rescue Service ("WYFRS") dated 22 July 2005. I consider that this material is helpful but, again, to the extent that it was not tested in cross-examination, its probative value must be regarded as limited.

370. Accordingly, my primary findings, set out below, come from the oral evidence of the witnesses, and the agreed timeline, prepared by Dr Mitcheson, as supplemented by other witnesses.

I2: My Primary Findings

371. On the evening of 8 June 2005, the 'oil pop' production area was working normally. Mr Norton was working six pans on the C Line. Mr Hardcastle was at the hatch of the C Line hopper, bagging the popcorn into sleeves. There were a number of made-up boxes close to the bottom of the hopper, but not more than about half a dozen. There was a pallet truck in the vicinity on which there were boxes containing completed sleeves. Mr Carter was doing some paperwork close to the south wall of the 'oil pop' production area.
372. There was nothing particularly unusual in the condition of the elevator and hopper. Although, since the elevator had been enlarged, the old lid for the hopper could no longer be used, there was a skirt arrangement at the top of the hopper which operated as a rough and ready enclosure. I also find on the evidence of Mr Carter that at least one of the two inspection hatches was in place, and probably both. The top of the hopper was therefore enclosed, save for the opening into the elevator.
373. Mr Carter was the fire marshal. He told me that he felt qualified to tackle any fire that might arise. Mr Challinor, who was the senior fire marshal in the NMU at the time, was not based in the 'oil pop' production area. Furthermore, he was unlikely to have been of very much assistance in the event of a fire, because on his own evidence he had had no training in evacuation procedures, and had also had no training in detailed matters such as the manual activation of the CO2 suppression system. To that extent, therefore, Mr Challinor was the senior fire marshal in name only.
374. At about 8:30 pm, there was a commotion at the C Line hopper and it seems that somebody shouted 'Fire!' and whilst filling a sleeve with popcorn from the hatch at the bottom of the hopper, Mr Hardcastle became aware that the sleeve in his hands was melting. Mr Carter looked up and saw what he described as 'smouldering' popcorn in the 'melting' bag. Mr Carter was adamant that the popcorn in Mr Hardcastle's sleeve was not flaming. At most, it appears that the popcorn might have been smoking. I again accept Mr Carter's evidence.
375. Mr Carter got up from where he was sitting and ran towards the hopper. Mr Norton thought that he or someone else shouted "the hopper's gone up". As Mr Carter climbed the gantry, Mr Norton pressed the emergency stop button and the C Line conveyor came to a halt. Mr Norton looked towards the top of the hopper, where he could see orange flames reflected in the underside of either the elevator or the ceiling panels (he was not sure which). Those orange flames were the flames in the hopper.
376. Mr Carter had noticed, as he moved towards the gantry, a wisp of smoke at the bottom of the hopper, and also at the top of the hopper. When he climbed the gantry and looked inside, he could see the same flames which were in the reflection that Mr Norton had seen. Mr Carter told me, and I find, that he saw flames across the surface of the corn in the hopper, which he said was about a quarter full. The flames were coming up to the top of the hopper but, at the time that he was looking, those flames were not rising above the hopper or reaching up into the elevator. He confirmed that, as he leaned in to look, the flames 'were not singeing his eyebrows'. Mr Norton also said that he did not see flames coming out of the top of the hopper.
377. Mr Carter shouted for the fire hose. He said that he was used to using the fire hose in the hopper because that was the way in which the hopper was cleaned out on a Friday during its weekly clean. Mr Norton heard the shout for the hose. Mr Carter told me that he switched the hose on when it got to the top of the gantry and he put the hose on full to drown the fire in the hopper. He said that within a few seconds, the fire had gone out. A lot of water came out of the fire hose, and there was a good deal of steam which was drawn up into the elevator.
378. Shortly after this, the fire alarm went off. There is an issue as to whether the fire alarm was activated as a result of steam and/or smoke passing through the C Line elevator into the Packaging Hall, or around the gap in the wall between old Production Area B and the Packaging Hall. Either way, it appears that it was the fire alarm in the Packaging Hall that was first activated.
379. Whilst Mr Carter was dealing with the fire in the hopper, the other operatives were dealing with the burning popcorn at the dispatch chute. They endeavoured to stamp out the smouldering/smoking popcorn from the sleeve that had begun to melt in Mr Hardcastle's hands. In addition, in accordance with the approved procedure, and as had happened during the fire the previous year, the operators opened the chute to release further burning/smouldering popcorn from the hopper in order to put that material out on the floor in the same way. The evidence as to this came from Mr Norton and the various interview notes with Mr Hardcastle. Although there is a lack of clarity as to the precise quantity of popcorn that was emptied out onto the floor, the most reliable estimate would appear to be a total of 3 sleeves' worth of popcorn (the one that was burning in Mr Hardcastle's hands, and two more). Mr Hardcastle said that, after dropping the first sleeve, he managed to press the handle (ie to release the popcorn) twice more. On that basis, a large amount of popcorn, possibly as much as 180 litres (3 x 60 litres), almost all of which would have been burning, was discharged over the floor of the production area. Even on the claimants' case, it was 150 litres of popcorn. A fire extinguisher was also used to try and put out this popcorn.
380. The operatives moved the made-up boxes away from the bottom of the hopper. Mr Carter thought they had been moved a few feet out of the way. Mr Norton said that they were moved away from the hopper towards the east wall, "just into open space, so they weren't near anything…away from the other cardboard." Again, his evidence would appear to suggest that the boxes were moved only a few feet. That was also the recollection of Mr Kear, who thought the boxes were moved 4 or 5 feet.
381. Generally, it would seem that Mr Carter and the other operatives acted coolly on discovering the fire and, particularly in Mr Carter's case, with a good deal of personal bravery. But, as Mr Kear said, everything happened very fast. He said, and I accept, that it was 'quite a confusing scene'. He memorably described it as "just a big dash and a big panic".
382. Mr Carter and the other operatives then left the 'oil pop' production area. They left through the plastic door in the wall and went into the Packaging Hall. That was immediately next to the roller shutter door which, on the evidence, had been hit by a forklift truck some weeks earlier and was jammed in a half closed position. As they left, the men thought that they had put out the fire in the hopper (using the hose) and on the floor of the production area (by stamping). Whilst in evidence Mr Carter was confident that he had put the fire out, it seems that there was a lingering doubt about it. I note that in a number of the interviews, those who were there queried whether or not the fire had in fact been extinguished. Thus Mr Bray told WYFRS that the operatives "were not sure if the fire was out"; he told Ms Irving that Mr Carter said that he did not think that he had put it out; and he told Mr Siddons that "the oil pop lads were flustered…and weren't quite sure they had put it out".
383. The fire alarm had gone off at 8:33 pm. Just over a minute later, the operatives left the NMU. At 8.35 pm, the defendant notified the Fire Service of the fire alarm activation. At 8:36 pm, when the operatives from the 'oil pop' production area were already outside at their muster station, Mr Bray went into the NMU building and looked at the fire alarm display board. He left the building immediately afterwards. It seems that he went back again a couple of minutes later but returned within 30 seconds. At 8:41 pm, he went into the NMU building for a third time but was again outside by 8:43 pm, when the first fire appliance arrived. It seems that these repeated visits were to check the main control panel. The second fire appliance arrived at 8:44 pm and the fire crews entered the building. At that stage, there was no external sign of any fire at all, and Mr Carter explained to the fire fighters that he thought that the fire had been put out.
384. At 8:46 pm, smoke became visible for the first time above the roof of the NMU. It seems that that smoke was blown from the 'oil pop' canopy extraction fans. At about the same time, Mr Bray and some of the fire men went into the Packaging Hall. Much more smoke was apparent. A point arises as to the type of smoke it was.
385. There was then a rather puzzling delay. It appears that the fire fighters wanted to enter the 'oil pop' area through the small door in the northern section of the east wall of the NMU extension, tucked away in the corner of old Production Area B, furthest from the seat of the fire. That door was locked, and there was a delay whilst a security guard was summoned to unlock the door. During the same period, there was also some time lost whilst the firemen offloaded and laid out the hoses.
386. At 8:52 pm, heavy white smoke was seen venting from the eaves on the south elevation. This is the south wall of the 'oil pop' production area. I find that the presence of this smoke at that location meant that, just 22 minutes after the fire had started, the integrity of Mr Caldicott's fondly-imagined 'fire box' had already been breached.
387. For the next few minutes, the fire fighters continued to lay out their hoses. Further fire fighting vehicles were requested. However, it is not clear whether any attempt had been made at this point to try and fight the fire. Although one report produced by WYFRS suggested that the first water onto the building was at 8:55 pm, that time is contradicted by other evidence, also from WYFRS, which suggests that, as late as 9:08 pm, no such equipment was yet in use. There is no CCTV footage of water going onto the fire before that time. Taking into account all the circumstances, it seems to me unlikely that any water was used to fight the fire until 9:13 pm, when the WYFRS incident log finally reports "1 large jet and 2 BA use".
388. By the time the water was on to the building at 9:13 pm, the WYFRS had already decided that they would operate in defensive mode only. That decision had been taken at 9:08 pm, once it became clear that no lives were at risk. It meant that the fire would be fought from outside the building only, with the priority on controlling the fire and ensuring that it did not spread to other buildings. It appears that this decision was taken because of the hold that the fire had taken on the NMU in the 38 minutes since the sleeve melted in Mr Hardcastle's hand. How and why the fire had spread so quickly is addressed in detail in Section J below.

I3: The Fire Investigation

389. WYFRS were involved in an extensive investigation into the fire after the event. In this they were assisted by Dr Mitcheson, who sat in on many of the interviews with those involved. It appears that Dr Butler was also instructed early on by the defendant, and he too was involved in some of the interviews, but his evidence at the trial was entirely peripheral.
390. In the aftermath of the fire, on 13 June 2005, there was a meeting between various representatives of the claimants and Mr Arbuthnot of WYFRS. It appears that the mood of this meeting was somewhat hostile, with Mr Arbuthnot seeking to explain how and why WYFRS had adopted such a defensive strategy within a short time of arriving on site. Mr Arbuthnot said:

"I think the senior representative of Trebor on site, Mr Sanders, I think it is correct, had expressed some dissatisfaction with Fire Service tactics and obviously because I was still involved in managing the operation, I hadn't had a chance to (a) fully explain our tactics and (b) I was unaware of some of the history leading up to the point of my attendance."

He said that he was there to readdress that misconception and explain why it was that the defensive strategy had been adopted.

391. It appears that, at that meeting and subsequently, Mr Arbuthnot and WYFRS considered that there had been a critical delay on the part of the claimants. He said:

"It is from the point that the ignition first commences to the point that an effective fire fighting operation commenced and actually, on this particular occasion, that was a very long time."

However, during his cross-examination by Mr Quiney, it became apparent that Mr Arbuthnot had a number of misconceptions about how long various events had lasted and that, at the time of the meeting and subsequently, he had not appreciated the true chronology, which I have set out above. Mr Arbuthnot also criticised the claimants for leading the firemen to the wrong location and misinforming them that the fire had been put out, so that WYFRS did not utilise their normal start-up procedures. I am bound to say that each of these criticisms of the claimants seem to me to be wholly misplaced.

392. First, on the agreed timeline, the fire appliances arrived less than 15 minutes after the fire had been discovered and 10 minutes after the fire alarm had been activated. There can be no question of delay, given such short response times. Thereafter, I agree with Mr Arbuthnot that there was "a very long time" before WYFRS commenced an effective fire-fighting operation, but that, so it seems to me, was wholly their responsibility. The claimants' operatives cannot be criticised for saying that they thought they had put the fire out; that was their honest opinion. The critical delay, between the arrival of WYFRS at 8:43 pm, and the start of an effective fire fighting process (and even then only in defensive mode) at 9:13 pm, remains largely inexplicable. In those circumstances, it seems to me that it was unsurprising that the claimants were somewhat hostile when Mr Arbuthnot attended the meeting on 13 June 2005.
393. WYFRS produced a fire investigation report dated 22 July 2005. The investigating officer was Mr R Hemingway. The report is not entirely satisfactory; for example the stated timeline ends at 8:55 pm, and that is with the incorrect notation that that was when water was first applied. As I have said, that time is contradicted by WYFRS' own incident log. It is curious that the report did not go on and deal with the timeline relating to the decision to adopt a defensive strategy, or to offer any detailed explanation for that decision.
394. The last three paragraphs of the report read as follows:

"It is known that the numbers on the particular shift were low and that Tony Norton who was also on C Line was doubling up and fetching raw materials for the oil pop operators. It is possible that either smouldering or flaming popcorn may have been allowed on to the conveyor unnoticed. This had then been allowed to enter the hopper and it either dropped as a smouldering piece on to other popcorn or has flamed almost immediately when the air was allowed to get to it through the dropping process. This will have readily ignited the popcorn already within the hopper, the temperature of the product in the hopper being approximately 50-60°C.

It is also likely that smouldering or flaming popcorn had been allowed to enter a number of the plastic sleeves having been filled and placed in the boxes ready for dispatch prior to the one that was subsequently noticed as being on fire. These boxes were moved away from the base of the hopper at the time of the hopper fire and relocated around the high racking of packaging material on the south wall. These boxes and plastic sleeves of popcorn are likely to have readily ignited due to the temperature of the popcorn and availability of air within the plastic sleeve and box to support a flaming fire.

With the evidence available so far this is the most likely cause of the fire. Further witness statements from Mr Dave Carter indicated that there were flaming popcorn dropping out of the base of the hopper when being bagged off and also flames reaching the top of the hopper of C Line which would allow flames to enter the opening to the plastic buckets and horizontal conveyor system. This is likely to have ignited the popcorn and possibly the plastic buckets that were moving allowing the fire to spread into the packaging hall. Further fire spread will have occurred as indicated above in the packaging material located on the high racking on the south wall."

These paragraphs are of little assistance to the court, because they cover most of the possibilities as to the spread of the fire, but do not indicate which option is preferred, or why.

395. Accordingly, it is necessary for me to go on and consider the various possibilities raised in the WYFRS report, so as to reach a conclusion as to the most likely way in which this fire spread. Because that is principally a matter of expert evidence, it means that, before I come on to analyse the evidence relating to fire spread, it is necessary to set out my (often uncomplimentary) views about the expert evidence generally in this case.

J: EXPERT EVIDENCE

J1: The Failures In The Process

396. At the first CMC in this case on 26 February 2010, Ramsey J gave permission to the parties to call two experts, one in fire suppression systems, and one in relation to forensic fire investigations. He ordered that, by 15 October 2010, the parties would agree preliminary lists of liability issues to be considered and discussed by those experts. The parties failed to comply with that direction and no list of issues was identified. At the second CMC on 22 October 2010, Akenhead J ordered that the list be agreed so that the experts could meet without prejudice not later than 3 November 2010. He then ordered that they should prepare a joint statement in accordance with CPR 35.12, setting out the matters on which they agreed and disagreed, and that their subsequent reports would be limited to those matters on which they were not agreed^[21]. He ordered that those reports had to be served by 22 December 2010, with supplemental reports on 21 January 2010. He also gave permission to the claimants to serve additional expert evidence from an architect specialising in requirements for fire protection and safety in building construction and the spread of fire. The report was to be provided no later than 14 January 2010.
397. Again, the parties and/or the experts failed to agree a list of issues. Moreover, it appears that there were difficulties with the conduct of the experts' meetings. Of course, since those meetings were without prejudice, it is impossible for me to say why that was. I was merely told that the experts "fell out" and that, as a result, no r35.12 statement was prepared. Bluntly, I have to say that experts appointed in civil litigation have no business to "fall out" and to fail to comply with the orders of the court. Experts are there to provide evidence on technical matters in order to assist the court, and for no other purpose. If they take matters of personal disagreement to such a level, they are failing to provide that service.
398. The parties' solicitors then abandoned the joint statements altogether, and instead got the experts to concentrate on their first and second round reports. Both sets of reports were provided late. It appears that the first round reports were provided in January, and the supplemental reports were provided, after the pre-trial review, in mid-February, just a month before the start of the trial. The order as to the joint statements was varied and a new date of 22 February 2011 was ordered for their provision.
399. Yet again, no joint statements were provided either in accordance with that order or at all. At the commencement of the trial, there were still no joint statements. Instead, the parties jointly sought permission to allow each of the four principal experts (i.e. all of the experts except Mr Jowett, the claimants' architectural expert) to provide an oral presentation to the court. It quickly became apparent to me that these presentations were designed to allow into evidence a variety of disparate matters, some of which were not even in the extensive reports produced by the four experts in question. These presentations amounted, therefore, to a third attempt on the part of the experts to express their opinions.
400. The joint statements were not produced until part-way through the trial, following an intervention from me. Unfortunately, they were of little or no use, because they were not focused on the issues between the parties. Instead, they operated as a sort of summary of some of the technical differences between the experts, often unlinked to the particular matters of importance which I have to resolve.
401. The directions given in this case by other judges in the TCC were, if I may say so, sensible and appropriate. In particular, they envisaged that the issues would be identified first, and would be the subject of the r35.12 statement. The reports could then focus on the issues on which the experts failed to agree. In that way, both the statements and the reports would assist the court in arriving at the answers to the relevant technical issues. Instead, in breach of the court orders, the experts adopted a completely back-to-front process. They produced a first round of reports which bear all the hallmarks of having been prepared in a hurry, and which do not address many of the relevant issues. The second round of supplemental reports was used principally as a vehicle for another attempt at setting out the author's conclusions. There was then a court presentation, which was a third attempt at such a process (with some criticisms of the other side thrown in), and finally a joint statement which added nothing to the overall debate. As a result of all this floundering, the expert evidence was unfocused, and often unrelated to the real issues between the parties. In a claim allegedly worth £110 million, I found this approach to expert evidence unsatisfactory and unhelpful.
402. I set out below some of the particular criticisms that I have of the approach and substantial evidence of the four principal experts. This is designed to demonstrate how and why I consider that large swathes of the expert evidence are, for a variety of different reasons, unreliable and unhelpful. I should make plain that none of these criticisms apply to Mr Jowett who was, in any event, recovering from an operation at the time of the trial and was therefore unable to participate in the run-up to the trial or to give oral evidence.

J2: Dr Mitcheson (Claimants' Fire Expert)

403. Dr Mitcheson was called in to deal with the fire on behalf of the claimants' insurers within days of the fire itself. He sat in with the interviews with some of the operatives and made notes of those interviews. He produced a contemporaneous report which has not been disclosed. His notes of the interviews were eventually disclosed shortly before the trial, but for a long time disclosure was refused on the grounds that the notes – like the original report – were privileged.
404. For what it is worth, I reject that suggestion. It seems to me plain that Dr Mitcheson was appointed by the claimants' insurers to investigate the cause of the fire. In those circumstances, it could not be said that litigation was the dominant purpose of his investigation: see Seabrook v British Transport Commission [1959] 2 All ER 15, and Waugh v British Railways Board [1980] AC 521. Thus I consider that the report and notes were disclosable. Moreover, since Dr Mitcheson had access to people and things which the defendants' experts did not, I consider that disclosure was required pursuant to the overriding objective.
405. I am wholly unconvinced by Dr Mitcheson's own explanation as to why his original report has not been produced. He was plainly uncomfortable about it when he said:

"We were now preparing for trial, so certain matters that I might have included – well, certain things became more important than others, so the report that I wrote in 2005, that was for the insurers, through their solicitors; might have contained material that was not suitable for the court."

It was therefore put to him by Mr Dennys that 'the certain things' to which he referred were matters which would not now advance the claimants' case. Although Dr Mitcheson denied that, he gave no examples of the "unsuitable" material he had in mind. If his early opinions had been overtaken by events or subsequent information, they might now be irrelevant, but they would not be 'unsuitable'. Accordingly, I have concluded that the original report should have been provided and I am troubled as to the explanation as to why it was not. It would doubtless have been of assistance.

406. In July 2005, Dr Mitcheson carried out tests with WYFRS. He accepted that those were not sophisticated tests, and were limited to exploring the burning characteristics of popcorn and how easy it might be to extinguish a fire in such materials. At this point, Dr Mitcheson confirmed that it was not part of his remit to look at the spread of fire which, given the speed of its development and the extensive nature of the damage, seems a surprising restriction on his brief.
407. With one exception, it is not possible to say what, if anything, Dr Mitcheson did on this case between June 2005 and the production of his report in January 2011. He certainly undertook no further tests. But there was a curious incident in June 2010 when he went to see the writer of the WYFRS report, Station Manager Hemingway, in order to get him to "contextualise" his report of July 2005. This seems to have been nothing less than an unsubtle attempt to get SM Hemmingway to soften those parts of his report which were, on their face, contrary to the claimants' interests in this litigation. This was not an appropriate task for an expert; Dr Mitcheson would have been much better off carrying out tests which showed whether his theory could possibly work or not, rather than trying to rewrite a report that was already 5 years old.
408. But Dr Mitcheson did not carry out any further tests before he produced his first report in these proceedings. That first report, dated 12 January 2011, referred to smouldering popcorn being conveyed into the elevator and making the transition to flaming either as it entered the elevator or in transit within the elevator "or more likely when tipped from the Gough elevator bucket into the C Line hopper". He thought the fire had become established in the C Line hopper some time before burning material was discharged into a plastic sleeve at the bottom of the hopper by one of the operatives. He also thought it was unlikely that burning material had previously been discharged from the C Line hopper and concealed within other sleeves. He concluded that, although he had done no tests relevant to this theory, the fire got into the Gough elevator and that this had (in a way not explained in the report) produced a running liquid fire.
409. Dr Mitcheson's supplemental report was dated 24 February 2011. This largely concentrated on criticising Dr Lygate's theory that the fire had spread because smouldering popcorn had made the transition to flaming in sleeves and boxes which had already been packed. Dr Mitcheson referred in his supplemental report to two further tests that he had by then carried out. One involved burning popcorn in a bucket, which was apparently designed to show the height of the flames that might be produced. The second was to hold a lighter against a plastic bucket in order to demonstrate that, after seven seconds, the applied flame set fire to the rim of the bucket. Neither of these tests were carried out with or in the presence of Dr Lygate, or any other expert: they were entirely unilateral. No tests were carried out by Dr Mitcheson that endeavoured in any way to replicate the conditions in the elevator and the hopper on the night of 8 June 2005.
410. In his oral presentation at the trial, and in the accompanying notes, Dr Mitcheson gave new evidence in support of his theory that melting plastic would have dripped out of the Gough elevator and created a pool fire. Further elaboration of that theory could then be found in his answers in cross-examination. As I have already said, the joint statement that he signed took matters no further.
411. Accordingly, the principal problems with Dr Mitcheson's evidence were that:

a) He failed to disclose his original report or notes which would have been prepared at a time when he had investigated the physical remains of the NMU, and interviewed those who were there, advantages which no other expert ever had;

b) He relied on unsophisticated tests which made no attempt to replicate the conditions in the oil pop production area on the night of 8 June 2005;

c) He adopted an unacceptably partisan approach, particularly in his attempt to "contextualise" the WYFRS report;

d) He failed to agree and/or carry out any joint testing with Dr Lygate or any of the other experts;

e) He failed to set out his theory as to fire spread in a coherent and comprehensive fashion.

412. In addition, the point also needs to be made that, whilst in his first report Dr Mitcheson referred to and relied on the presence of smouldering popcorn, by the time of his supplemental report, just a month later, he was adamant that popcorn could not smoulder in any circumstances. Whilst it is of course right that an expert is always entitled to change his mind, if he does so, he must explain why he originally thought X, and how he has now come to conclude that the answer is Y. He also needs to explain the significance of his change of mind. Dr Mitcheson, in keeping with the slapdash nature of the expert evidence generally in this case, did neither.

J3: Dr Lygate (Defendant's Fire Expert)

413. Dr Lygate's first report was dated 19 January 2011. It was much too long, consisting of 75 pages, despite the fact that Dr Lygate had not carried out any tests at all and was really commenting on material that had long been generally available. The report concluded that the most likely cause of the fire was that smouldering popcorn, which had already been sleeved and put in boxes, made the transition to flaming after the operatives had left the 'oil pop' production area, and set fire to the combustible materials on the south wall. There was no mention at this stage of his alternative theory, which was not postulated until his supplemental report, to the effect that the fire spread because the burning popcorn that was released from the bottom of the hopper was not fully extinguished by the operatives.
414. Prior to his supplemental report, Dr Lygate arranged the carrying out of "cube tests" in which 100 mm cubes of popcorn were placed in an insulated box to see how long a sleeve of smouldering popcorn would take to make the transition to flaming. Those tests were curiously described in sections 3.5 and 3.6 of his supplemental report.
415. The difficulties with those tests, and the propositions to which they were said to give rise, were many and varied. First, they were not carried out by Dr Lygate, who did not even witness them. They were carried out by students at Edinburgh University, and for some time the only information that the claimants had about the tests (other than the short summary in Dr Lygate's supplemental report) came from a blog, taken from the internet, and written by one of the students, under the heading "Today, We Have Been Mostly Burning Popcorn…"^[22]. Secondly, although there was apparently a video of the tests, the video of the complete tests has not, even now, been made available to the claimants or the court. Thirdly, of course, it goes almost without saying that these tests were carried out unilaterally, so that those who were not there, including Dr Mitcheson, the claimants, and the court, have had to make the best of the very limited information available.

J4: Mr Stephens (Claimants' Fire Suppression Systems Expert)

416. Mr Stephens' first report was dated 14 January 2011. It was again inordinately long, running to 76 closely-typed pages, and yet it failed to deal in a cogent way with the only issue on which Mr Stephens' evidence could be relevant, namely what was wrong with the design or installation of the CO2 fire suppression system, and how a proper design/installation could have been achieved. His supplemental report on 21 February 2011 is a paragraph-by-paragraph critique of Mr Jackman's report, in which any semblance of the wood has been completely obliterated by the trees.
417. Mr Stephens' reports, and his subsequent oral evidence, graphically illustrated the failure on the part of the experts to get a grip with the real issues in this case at the outset. Instead, his reports seemed to operate on the basis that if the defendant had said X, then Mr Stephens had to try to demonstrate that the answer must be Y, no matter how peripheral the issue. It is an unfocused and unhelpful approach.
418. Perhaps as a result of this pervasive deficiency in his written material, Mr Stephens' oral evidence degenerated into bad-tempered bickering. He was an extremely difficult witness, repeatedly failing to answer the questions put to him and unwilling to make even the most basic assumptions in order to answer the questions being asked. His simplistic comments (such as the suggestion, referred to already at paragraph 284 above, that the system was sold as an automatic system and should therefore have operated as an automatic system, regardless of the manual operation), were either wrong or self-fulfilling. I regret that I derived little real assistance from much of his evidence.

J5: Mr Jackman (Defendant's Fire Suppression Systems Expert)

419. Mr Jackman's written evidence suffered from some of the same deficiencies as that of Mr Stephens. He too was extremely partisan; he too had failed to get to grips with the principal issues. Much of his initial report, dated January 2011, was taken up with criticisms of the claimants which did not form part of the pleaded allegations of contributory negligence, and were therefore irrelevant and inadmissible.
420. The other important element of Mr Jackman's first report was the CFD modelling which had been carried out, not by Mr Jackman himself, but by a colleague. It was said that this modelling demonstrated that the CO2 suppression system as designed and installed would have worked and put out the fire. The difficulty was that the CFD modelling system was not fully explained in Mr Jackman's report and the claimants had to spend a good deal of time trying to obtain information from Mr Jackman about various elements of the modelling exercise. That information was still outstanding when Mr Stephens had completed his evidence, so he was unable to deal with it.
421. The modelling was itself flawed for the following reasons:

a) It failed to take account of some of the openings, including the mesh at the foot of the vertical section of the elevator. That was a large opening. Although that was only discovered shortly before the trial, it does render the whole modelling exercise somewhat questionable.

b) It failed to take account of accurate dimensions and specifications. For example, it did not take account of the true RTI of the Olsen heat detector nor the true dimension of the buckets.

c) The figures appear to have been rounded but it is not clear precisely how this has been done (see paragraph 64.4 of the claimants' written closing submissions).

422. Mr Jackman's supplemental report contained various additional calculations which purported to show that, even assuming against the claimants that there were openings at the top of the hopper, the CO2 suppression system would have been very likely to have been triggered within 2 and a half minutes from the time that smouldering popcorn made the transition to flame. Although this was said to demonstrate that the system would have worked as intended, I agree with Mr ter Haar that, on a proper analysis, this calculation revealed that there was a fundamental flaw in the design of the CO2 suppression system: if the calculation was correct, given that on any view the popcorn in the hopper had been burning long enough to have flames at the top and burning popcorn at the bottom, why had the CO2 not been released long before Mr Carter got to work with the fire hose? I note that this relatively obvious conclusion from Mr Jackman's supplemental report was not apparent to Mr Jackman himself until his cross-examination.
423. Mr Jackman had a rather dismissive manner as a witness. However, there were two particular passages of his cross-examination (one of which was based on the findings to which I have referred in the previous paragraph) during which he seemed unable to refute the suggestion that the CO2 suppression system fail to operate as the defendant had intended^[23]. In relation to the allegations of liability therefore, I consider that Mr Jackman's concessions were of the highest importance.

J6: What Should Have Happened

424. What should have happened in relation to the expert evidence was this. First, a list of issues should have been agreed between the solicitors in the summer of last year. That list should then have formed the agenda for the experts meetings in the autumn. Secondly, following those meetings, a r35.12 statement ought to have been prepared so that the experts' reports could have been limited to those issues on which the experts had failed to agree. All of that was, of course, in accordance with the original court orders. If there had been a problem, the parties should immediately have come back to court for help. The meetings between the experts, if properly conducted, would also have revealed that further tests/experiments were necessary. A programme for those tests should have been drawn up and they should have been planned and carried out jointly or, at the very least, in the presence of all the experts.
425. In this way, the parties would have ensured that the correct issues were dealt with by the experts, and the same basic test material would have been common to each. The court would then have been provided with focused expert evidence on those issues which they could not agree, and could have worked its way through the competing opinions, issue by issue, in order to arrive at conclusions as to the spread of the fire, and the allegations of liability.
426. Instead, for the reasons that I have outlined, the court has had to struggle with unsatisfactory and disparate expert evidence, often unrelated to the real issues, prepared and delivered in a variety of places and in an unacceptably partisan way. Unsurprisingly, perhaps, this has created real difficulties in the preparation of parts of this Judgment. It has also led me, very unusually, to be dubious about the reliability of all of the expert evidence that has been presented to me. This is emphatically not a case where the court is able to prefer one expert over another and let that approach dictate the result.

K: THE SPREAD OF THE FIRE

K1: The Beginning And The End

427. Although, because of the difficulties to which I have previously referred, the experts did not appear to realise it until part way through the trial, it was apparent that they were in fact agreed about both the origin of the fire, and where, within the oil pop production area, the fire ended up within about half an hour or so of breaking out. It was agreed that the fire started on the C Line in the 'oil pop' production area. It was agreed that smouldering or burning popcorn got into the elevator, and subsequently the hopper, and caused the fire. And although they disagreed how the fire got from the hopper to the south wall, they were agreed that, in one way or another, the fire did find its way to the south wall. As Dr Mitcheson put it: "I have no doubt in my mind that the packaging material that was stored in the racking on the south wall was involved in the fire at an early stage." It was that fire which Dr Lygate said caused the energised or active smoke venting from the south eaves at 8:52 pm, the breaching of the so-called 'fire box'. Dr Mitcheson accepted that that energised or active smoke was likely to be a sign of a fire in the racking material at the southern elevation.
428. This agreement, that the fire reached the south wall very quickly, was important. That was confirmed in Dr Mitcheson's cross-examination:

"So, at this point, the theory of spread of fire from the hopper or elevator converged. You have a fire on the south wall, that fire can spread into the roof void, through weaknesses in the top hat section and the joints between the rockwool panels; and once the fire is in the roof, the building is lost?

A: Essentially, yes…

Q: So we have two hypotheses then at this point as to the spread of fire. One is that it emanated from the material which was discharged from the bottom of the hopper and the other is that it resulted from the elevator fire spreading in the way that we just looked at. Correct?

A: Yes.

Q: Right. Now these are two viable hypotheses, aren't they?

A: Yes they are.

Q: Neither is obviously right, neither is necessarily wrong?

A: That is correct."

It is the issue as to which – if any - of these viable hypotheses is the probable explanation for the spread of the fire with which the remainder of this Section of the Judgment is concerned. In undertaking that exercise, I bear in mind that the correct approach to causation in a case like this has been recently re-stated by Akenhead J in Harooni v Rustins Ltd [2011] EWHC 1632 (TCC): I do not have to accept any possibility, even that which might be described as the highest-ranked possibility, unless I consider that it is the probable cause of the spread of the fire.

429. Accordingly, I have endeavoured to follow the guidance of Thomas LJ in Ide v ATB Sales Ltd [2009] RTR 8 when he said:

"In the vast majority of cases where the judge has before him the issue of causation of a particular event, the parties will put before the judge two or more competing explanations as to how the event occurred, which though they may be uncommon, are not improbable. In such cases, it is, as was accepted before us by the appellants, a permissible and logical train of reasoning for a judge, having eliminated all the causes of the loss but one, to ask himself whether, on the balance of probabilities, that one cause was the cause of the event."

K2: From Pan To Hopper

430. Dr Mitcheson's original view as to what happened between the pan and the hopper, as set out in his first report, can be summarised in this passage of cross-examination:

"Q:…popcorn is overheated at the pan. That is stage one.

A: Yes.

Q: Smouldering popcorn enters the elevator. Stage two.

A: Yes…

Q: Popcorn makes the transition to flaming before it reaches the bed of popcorn in the hopper?

A: That's right, as it's discharged from the elevator bucket…I thought it was less likely [to make the transition to flaming in the bucket journey] than when the bucket was tipped on its side in order to discharge the smouldering popcorn."

The evidence was that the journey time within the elevator was about 1 minute and 22 seconds.

431. Dr Mitcheson confirmed that his original view had been that it was less likely that the popcorn would be flaming on the journey from pan to hopper, and that it made the transition to flaming on discharge into the hopper. But he now said that he believed that the popcorn must have been flaming before it entered the elevator. It appears that the principal, if not only, reason for this change of mind was that Dr Mitcheson no longer believed that popcorn was capable of smouldering. He accepted in cross-examination that if the popcorn had been flaming from the moment it left the pan then it would have been flaming as it passed through the dosing drum and the rotary sieve, and that such flames would have been unobserved by any of the operatives. Dr Mitcheson confirmed that, even on this theory, the flaming popcorn would not have set fire to the bucket on the journey of 1 minute and 22 seconds. He confirmed that it was unlikely that the popcorn would burn through a bucket.
432. As a result of Dr Mitcheson's change of mind, a good deal of time and effort was spent in this trial arguing about whether or not popcorn could smoulder. As I made plain during the course of the trial, it seemed to me that that was an issue of entirely academic interest, because there was no part of this case which turned on a finding that popcorn either could or could not smoulder. However, for what it is worth, I have concluded that, on balance, it is more likely than not that popcorn is indeed capable of smouldering (as Dr Mitcheson originally assumed and Dr Lygate maintained throughout). My reasons for this conclusion are set out below.
433. First, the operatives referred repeatedly in their evidence to 'smouldering popcorn'. If any group of people in this case are going to know whether or not popcorn actually smoulders, it is going to be the operatives who worked with it day in and day out. Secondly, for the reasons that I have already given, I have found that the popcorn in the sleeve that Mr Hardcastle was bagging at the time of the fire was not flaming, even if it was about to make the transition to flaming. It was therefore smouldering. Thirdly, everyone, including Dr Mitcheson, originally assumed that popcorn could smoulder and it seems to me that this was a commonsensical and logical assumption to make.
434. I accept that Dr Mitcheson undertook some unilateral tests in January/February which, on his view, demonstrated that he could not get popcorn to smoulder. However, those tests were not designed with that factor in mind and were very rudimentary. Moreover, when cross-examined about it, Dr Mitcheson accepted that his tests had generated 'glowing' popcorn and, for present purposes, I can see no reason to distinguish between 'glowing' and 'smouldering' popcorn. Dr Lygate made no such distinction. Dr Mitcheson confirmed that, on one test, he got what he described as "a very slight glow".
435. Dr Lygate explained how and why the tests done by the post-graduate students at Edinburgh University confirmed his view that popcorn could smoulder. He reiterated that 'glowing was an essential stage in smouldering'. He said that in two of the four tests which were undertaken, smouldering was established and the time to flaming was between 5 to 11 minutes. The potential difficulty, of course, with this evidence was that Dr Lygate had judged smouldering from the video recording, having not been present for the tests himself. However, he also referred to the dustbin tests carried out by Dr Mitcheson and Mr Stephens and maintained – I think rightly - that those also demonstrated smouldering.
436. Accordingly, to the extent that it matters, I have concluded that popcorn can smoulder. I have also concluded that, on the night of 8 June 2005, the popcorn was smouldering but not flaming when it passed into the elevator. This was one of the reasons why it was not seen by the operatives and removed before it entered the elevator: there were no flames to alert them. In my judgement the popcorn smouldered and did not make the transition to flaming until, at the earliest, it dropped into the hopper.

K3: In The Hopper

437. Dr Mitcheson said that, at the latest, the popcorn flamed as it fell into the hopper. He had not set out in his first report any estimate as to how long the flaming popcorn would have stayed in the hopper before it was removed via the discharge mechanism at the bottom. In his supplemental report, he estimates a figure of about 8 minutes. He went on to say:

"The time between deposition on the surface and emergence at the discharge may be less than that [8 minutes] because of the downward spread of flame through the bed. You're not just relying on the discharge of popcorn from the base of the hopper to effect downward movement of flame.

Q: Right. So you have what we might call the sugar mix effect?

A: Yes."

However, Dr Mitcheson confirmed that he had not done any experiment to see whether the liquid being exuded from the kernels of popcorn when they were burning – the so-called sugar mix – shortened the residence time.

438. Dr Mitcheson was asked about the effect on the next load of popcorn being dropped into the hopper on a single flaming kernel. He did not accept that the next load would subdue the flame. However he did agree that it was important for his theory (that the fire went back into the elevator) that flames of 2 metres in height were achieved and that, for that theory to work, whatever flaming/burning popcorn had been deposited in the hopper must remain alight after the next batch of popcorn had been deposited on top of it. Dr Mitcheson said that "the scenario that I have is that once flaming has been introduced into the popcorn in the hopper bed it continues." And, as noted, the fire had to last somewhere between 5 and 8 minutes.
439. Dr Mitcheson was asked about the lateral spread of the fire. The dustbin tests appeared to show that, for the fire to travel across the whole of the surface area of the hopper, it would take four times 106 seconds, assuming it was linear. This was nowhere set out in Dr Mitcheson's reports. As to the downward spread of fire, the tests appeared to show that, in 1 minute and 46 seconds, the fire had travelled 150 mm in a downwards direction. However, Dr Mitcheson was very reluctant to commit himself as to the speed of any downward spread of fire, saying that there were so many variables. He said this was particularly so given the fact that, although one additional load of popcorn was added during the tests, it was unclear at what interval(s) this popcorn was added. It did not appear that the dustbin tests made any attempt to replicate what was happening in the C Line elevator and hopper.
440. Dr Lygate confirmed that, by the time Mr Hardcastle was bagging the hot popcorn at the dispatch chute, there was a flaming fire at the top of the popcorn in the hopper. However, he disagreed with various elements of Dr Mitcheson's evidence as to how that had been created. First, he rejected the relevance of the sugar mix, saying that as the burning liquid was exuded from the popcorn at a high level, it flowed down through the bed, and thus cooled and was extinguished.
441. Secondly, Dr Lygate was emphatic that there was smouldering popcorn in the hopper, because on his tests, he could not get the fire from the top down fast enough if the popcorn was not smouldering. He thought the fire in the hopper was smouldering, and then made the transition to flaming, and burnt from the bottom up, so that Mr Hardcastle discharged burning popcorn at the same time as there was a flaming fire at the top. There was this exchange:

"Q: It is not going to start from the bottom and work its way up is it?

A: That is exactly what happens in a smouldering fire, because it is starting within the body of the hopper and spreading up and down at the same time. And that was what I sought to explain in my presentation on Thursday."

442. As to the time the fire would take to spread, Dr Lygate said that, taking an average figure, it would have taken about 7 minutes to get the fire spread from the centre of the hopper to the edges of the hopper. He thought it might take less because the fire spread would accelerate. He therefore said we were 'probably talking in the order of 5 minutes'. He went on:

"Q: So if we have flames across the bed of the hopper, you are talking about something that had been burning for five minutes. The question is, then, how far does that flame of fire spread in a downward direction in that same period?

A: Fires spread more quickly upwards than they do downwards. Downward rates of fire spread are roughly of the order of lateral rates of fire spread.

We have a bed which is of the order of 1.3 metres deep from the upper point of the discharge chute to a quarter full…so this fire then has to travel downwards through 1.3 metres in the same time that it spreads across the hopper. It doesn't work."

This was the principal reason why Dr Lygate said that the fire was not flaming at the top and the bottom simultaneously.

443. Dr Lygate's preferred analogy for the mass of smouldering popcorn in the hopper was with a tennis ball, increasing to the size of a football. The increase was radial. He said that because heat rises, the popcorn would smoulder in the upward direction faster. It therefore began to take on the appearance of a balloon with a pointed edge at the bottom and bigger and wider at the top.
444. I have concluded that, to the extent that they differed significantly, Dr Lygate's theory as to what was happening in the hopper was to be preferred to Dr Mitcheson's. The first reason for that is because it is consistent with the findings of fact I have made, and in particular my finding that the popcorn that Mr Hardcastle was bagging was smouldering, but it was not flaming. Secondly, I consider that Dr Lygate's analysis of the time of fire spread is the most reliable available; his calculations demonstrate that, on a balance of probabilities, there was a smouldering block of popcorn which spread more quickly up than down, and which, for the reasons he explained, resulted in flaming popcorn on the surface of the hopper at the same time as smouldering popcorn in the sleeve being bagged by Mr Hardcastle.
445. By contrast to Dr Lygate's evidence as to what happened in the hopper, Dr Mitcheson's explanation was confused and contradictory. I consider that his original view, as to when the popcorn made transition to flaming, was correct, and that this aspect of his subsequent evidence was destroyed by his dogged assertion that popcorn could not smoulder.

K4: The Release of Smouldering Popcorn from the Hopper (Lygate 2)

446. We now come to the three alternative theories as to fire spread. I start with the theory advanced, originally as an alternative, by Dr Lygate, which was referred to during the trial as 'Lygate 2'. This theory postulated that there was one sleeve of smouldering popcorn which was dropped or fell out of Mr Hardcastle's hand, spreading burning popcorn onto the floor. The operatives then added to that by discharging two further sleeves' worth of popcorn (a total of up to 180 litres) onto the floor, in order to put out the burning popcorn by stamping on it. This theory suggests that, with this large volume of popcorn on the floor, spread and scattered across a wide area, it was not all stamped out and that, when and shortly after the men left the 'oil pop' production area, the smouldering/flaming popcorn on the floor set fire to other combustible materials and thereby caused the fire to spread.
447. My initial instinct was that this was the most realistic of the three competing theories as to fire spread, because it was the only one that was consistent with all the facts, and the only one that did not rely on a series of coincidences. Instead, it involved the deliberate (if unwitting) spread of burning material across a wide area, which material – the popcorn – was, on any view, smouldering and/or beginning to flame. After all, that was why it was being discharged from the hopper in the first place: in order that it could be extinguished. Furthermore, although Dr Mitcheson suggested that he did not investigate this theory because it was inconsistent with the accounts of the eye-witnesses (who generally thought the fire had been extinguished), that explanation simply does not stand up to scrutiny. First, it must be at least possible that not all of the operatives saw every last piece of popcorn being extinguished. Secondly, this was consistent with the lingering doubt that not all the popcorn had been extinguished (see paragraph 382 above).
448. Dr Mitcheson's other reason for rejecting the theory that the fire spread occurred outside the hopper (the "primary plank" of his position, as he put it) was that the popcorn was already flaming when it came into Mr Hardcastle's hand. I have already rejected that as a matter of fact (paragraph 374 above). Accordingly, in the absence of that primary plank, Dr Mitcheson was always struggling to refute the theory that the fire spread through the deliberate discharge of the burning popcorn.
449. In the end, Dr Mitcheson's main focus of attack on Lygate 2 was the impossibility of the fire spreading from the hopper to the south wall, a distance of about 7 metres. For this purpose, he relied on the evidence that the boxes were only moved a few feet and that there was no combustible material between the boxes and the south wall. But this ignored what I find to be the likely presence of a pallet truck, apparently loaded with boxes, about half way between the C Line hopper and the racking. That would plainly have provided a stepping stone for the fire as it spread.
450. That was certainly Dr Lygate's view. He said that this alternative theory had come "out of a better understanding of what happened at the bottom of the hopper". In particular, he stressed the fact that there was almost three sleeves' worth of burning popcorn which had been released from the hopper. He said that the stamping itself could have spread the burning popcorn a good distance away from the hopper to some other form of combustible substance and/or, by a series of stepping stones, to the racking in the south wall. As to the particular stepping stones relied on, he referred to the boxes that were thrown or moved away from the foot of the hopper and the pallet truck itself.
451. It was put to Dr Lygate that there were insufficient stepping stones for the 7 metres distance to be bridged. He disagreed:

"I think there is boxes between the pallet truck and the racking which are made up ready to receive the sleeves. So we have only got, to my mind, to get a distance of about 3.5 metres depending on where this pallet truck is, but I think Dr Mitcheson gave evidence to it being about half way between the hopper and the racking. So that means we have got to get it from the hopper to the pallet truck because if we get it to the pallet truck we have pretty much got it to the racking because we have got the combustible materials there.

Now I don't know what there was between the hopper and the pallet trucks, nor do I know if the pallet truck is fully loaded with boxes or partially loaded with boxes. So we could have boxes round the hopper, boxes in the pallet truck, empty boxes ready to be filled and then the racking."

Accordingly, Dr Lygate refuted the suggestion that there were insufficient stepping stones for the fire to spread quickly to the south wall, and said that it was entirely possible that the fire could have spread in this way between 8:33 pm and 8:52 pm.

452. As to the effect of the actions of the operatives in dealing with the smouldering popcorn on the floor, Dr Lygate said that, when they stood on it and stamped on it, "that popcorn pile is not going to remain in place, it is going to be displaced. The question is, how far is it displaced?" Dr Lygate thought that it was at least possible that, not only did the stamping not put out all the smouldering popcorn, but that the stamping itself displaced the popcorn over a wide area (including, of course, towards the south wall). He refuted the suggestion that the flames would die in transit.
453. He accepted that this theory was not entirely straightforward. In particular, he agreed that the flame from a single piece of popcorn might only be 2 or 3 inches high and might therefore struggle to find any combustible material to ignite. He also accepted that the chance that a single piece of burning popcorn causing this damage was a possibility but he would not say that it was probable. In his view, there was a higher probability that there was a clump of burning popcorn which was displaced by the stamping. He said in clear terms that he thought it more likely than Dr Mitcheson's theory (Mitcheson Final)^[24].
454. I consider that the theory encompassed in Lygate 2 is a realistic explanation for the spread of fire. It is the only one that involves and relies on a known action, namely the stamping out of a large quantity of burning popcorn on the factory floor. It is consistent with the factual evidence, in particular with the doubt at the time as to whether the fire had been fully put out. And I find that there was sufficient combustible material to carry the fire to the south wall (including the boxes and the pallet truck), all within the necessary time frame, particularly given the widespread displacement caused by stamping. Accordingly, it seems to me on all the evidence to be the probable cause of the fire spread. The only remaining question then becomes whether there is any other theory which is more likely than this one.

K5: The Subsequent Flaming of Sleeved Popcorn (Lygate 1)

455. Dr Lygate's first theory (Lygate 1) postulated that smouldering popcorn was put into a sleeve or sleeves by Mr Hardcastle, boxed and shrink wrapped and put on the shelves near the south wall. Shortly after the men vacated the production area, the smouldering popcorn on the shelf made the transition to flaming and started the fire^[25]. Dr Mitcheson's principal reason for rejecting this theory as to fire spread was his view (not in his first report but in his second report) that popcorn did not smoulder. I have rejected that view for the reasons already given.
456. There is no doubt that, on the face of it, this theory explains how the fire got so quickly to the south wall. However, even that is not entirely straightforward; as noted above, the evidence was that the boxes were only moved a few feet from the bottom of the hopper and were not particularly close to the south wall. Furthermore, unlike Lygate 2, this theory does not involve the possible displacement of burning popcorn over a wide area, caused by the stamping, which would have reduced (and possibly reduced significantly) the distance over which the fire had to travel to the racking.
457. In truth, I consider that there are a number of other insurmountable obstacles which prevent any acceptance of this theory. The first such obstacle is that, if this theory were correct, everything that happened in the production area involving Mr Carter, Mr Hardcastle's melting sleeve, the fire hose in the hopper and so on, was all completely irrelevant to the spread of the fire. It means that, by some extraordinary coincidence, the fire that burst out of the sleeved popcorn was not detected because the production area had already been evacuated as a result of another, completely unrelated, outbreak of fire. That seems to me to stretch credibility beyond breaking point.
458. Secondly, as Mr ter Haar suggested in cross-examination, the timing on this hypothesis becomes very tight. After all, there was no sign of flame in the sleeves in the boxes when the operatives left the production area. That was at 8:33 pm. 18 or 19 minutes later, we know that the fire was well-developed on the south wall. So the smouldering popcorn in the sleeve would have had to have made the transition to flaming almost immediately the men left the production area, which was short compared to the range of 5 to 11 minutes which Dr Lygate said would be the relevant timescale for smouldering.
459. Thirdly, the smouldering popcorn would have had to have been sleeved and boxed by Mr Hardcastle without him being aware that the popcorn was smouldering. Dr Lygate accepted that that was a logical consequence of this theory. I find it impossible to believe that Mr Hardcastle could have put smouldering popcorn into a sleeve without becoming aware of it. After all, we know that Mr Hardcastle did attempt to put smouldering popcorn into a sleeve and the result was the sleeve melting in his hand and the first raising of the fire alert. Accordingly, this theory of fire spread requires the smouldering popcorn to be sufficiently hidden from Mr Hardcastle during the bagging process for it to be undetectable. I consider that that is most unlikely.
460. Fourthly, there is absolutely no evidence to support the proposition that smouldering popcorn inside a sleeve, without any moving air to assist in combustion, would make the transition to flaming. I appreciate that Dr Lygate said that there would have been a volume of air in the sleeve and an extremely porous mixture, so that the heat in the bag was going to cause air to flow. But I note that he went on to accept that there had been no tests, of any kind, to see whether a smouldering fire could be turned into flames inside a hermetically sealed sleeve. It seems to me that such tests would have been an essential ingredient if this theory were to be seriously advanced.
461. Dr Lygate accepted that this theory of fire spread was improbable. Unlike Lygate 2, it is not supported by any of the factual evidence. For all the reasons that I have indicated, it seems to me to be wholly unlikely. It gives rise to too many coincidences. In this regard, the final exchange during Dr Lygate's cross-examination on this topic demonstrates in summary why I reject Lygate 1:

"Q: So it has got to be really rather a perfect distance doesn't it? It has to be so far into the popcorn that it can't be felt by the operative but so close either to the top or to the edge that it reaches the side of the sleeve sufficiently early to be able to cause a fire which breaks out by 8:46?

A: Yes.

Q: So we end up, don't we, with a number of extraordinary coincidences if you are right. We end up with popcorn that smoulders but doesn't reach flaming before it's got to the bottom of the hopper; that is essential?

A: Yes

Q: It's got to be not noticed by the operative, that is right, otherwise as a matter of common sense, this incident wouldn't occur?

A: Yes, they don't notice the smouldering in the hopper.

Q: It has to be towards the top or towards the edge of the sleeve but not so close that the operative notices it?

A: That is correct.

Q: It has to have sufficient oxygen within a hermetically sealed sleeve to be able to continue to combust?

A: That's a given.

Q: And it has to move from non-flaming to flaming at the 5 minute end of the spectrum rather than the 11 minute end of the spectrum?

A: It does."

462. I therefore reject Lygate 1 as being considerably less likely as a reason for the spread of fire than the discharge of burning popcorn all over the floor of the factory (Lygate 2). It is not a plausible explanation for the fire spread.

K6: Fire Getting Back Into The Elevator (Mitcheson Final)

463. As I have already noted, the claimants' theory as to fire spread, referred to during the trial as Mitcheson Final, is nowhere noted in one single document. It is an amalgamation of ideas arising out of both of Dr Mitcheson's reports, his presentation on Day 9 of the trial, and his cross-examination. This final explanation involves the following stages:

a) The flames in the hopper reached such a height that they extended beyond the top of the hopper and up into the elevator.

b) When the flames reached up into the elevator, they set fire to a bucket in the vertical, tipped position. The theory accepts that the bottom of the buckets in the horizontal position would not catch fire. Thus it is a critical element of Mitcheson Final that it was the rim of the bucket in the vertical position that caught fire first.

c) By the time that the rim and then the bucket had caught fire, it was in the horizontal position. The fire spread to the next bucket as it moved along the conveyor back towards the pans.

d) The molten plastic would have dripped out of the various openings in the Gough elevator, including the opening at the bottom, the existence of which the claimants only discovered shortly before the trial. The liquid plastic created a pool fire which, in some way, took the fire from the floor around the elevator to the south wall.

464. Not only was this theory not fully set out anywhere in Dr Mitcheson's reports, but it was contrary (at least in part) to his original approach. Moreover, the focus of his reports was an attack on Lygate 1 and Lygate 2, rather than an attempt to explain and bolster his own theory. Dr Mitcheson accepted that "I probably spend more time looking at Dr Lygate's theory than I do on mine". In my view, these were all signs that Dr Mitcheson was not very confident in the plausibility of his own theory, and had decided that attack was the best form of defence. For the reasons set out below I share that scepticism: on the material before me, I considered that the Mitcheson Final is, on a proper analysis, the least likely theory of fire spread of any of those presented.

a) Tests

465. The first (and in many ways most fundamental) problem with Dr Mitcheson's explanation is that it was entirely untested. When that was put to him in cross-examination, he denied it, saying that "we've explored how easy it is to ignite buckets with very small flames." But that was a reference to a simple experiment whereby someone hold a gas lighter against the rim of a bucket for 7 seconds until the bucket ignited, which in no way replicated the situation within the elevator. In operation, the bucket was in the vertical position above the hopper (the only position, on this basis, in which the rim could have been ignited) for no more than one or two seconds. In addition, there were no other tests carried out that were of any relevance to the detail of Mitcheson Final.
466. This failure to carry out proper or comprehensive tests was explored in Dr Mitcheson's cross-examination:

"Q: If you were going to carry out some experiment to demonstrate this effect, I would suggest that it would be sensible to construct some containing vessel of the size of the hopper, (1); (2) to fill it with popcorn to the assumed depth; (3) to ignite it; (4) to see what happens when you put bucket loads of popcorn at 1 second intervals on top of it. You can repeat it for a bit of smouldering popcorn, and you might get an answer.

A: That would be an ideal situation. I have reservations about the smouldering popcorn.

Q: I know you have reservations about smouldering popcorn, but that's a test which will actually show you something isn't it?

A: In an ideal world, yes.

Q: Not burning popcorn in a mop bucket or a dustbin?

A: They give valuable guidance on the behaviour of burning popcorn.

Q: Yes. And of course – and you may think this is a comment but I will ask you it anyway – in the context of litigation involving a claim of £100 million, spending a bit of money on testing is hardly money wasted is it?

A: That's probably correct."

467. In my view, Dr Mitcheson was right to accept that much fuller tests should have been carried out, particularly given the alleged value of the claim. In particular, an experiment should have been carried out which sought to see whether the rim of a bucket, hanging in a vertical position for no more than 1-2 seconds, could ever have been ignited in the way postulated by the theory. That was such a small window of opportunity that at the very least, it required was some sort of experiment to see if it was ever going to be realistic. Holding a lighter against the rim of a stationary bucket for 7 seconds in no way replicated the mechanism postulated by Dr Mitcheson. I address this further below.

b) The Necessary Flames

468. Dr Mitcheson accepted that the only way he could get the fire into the elevator was if the flame had come up from the hopper below. This required a flame that reached higher than the top of the hopper and into that part of the elevator where the buckets were moving across the opening.
469. The principal difficulty with this element of the theory is that there was no evidence of any sort that the flames reached that height. Mr Carter put out the fire in the hopper. He was clear that the flames did not reach up beyond the top of the hopper when he saw them. Mr Norton did not see flames of that height, either, although he was not in as good a position to see as Mr Carter. Thus, not only is there no evidence that the flames were of the height necessary for this theory to be correct, but there is positive evidence that the flames were not of the necessary height.
470. Accordingly, for Mitcheson Final to be right, it means that the flames which Mr Carter saw, and which he subsequently doused with the fire hose, were not as strong and/or as high as the flames that set fire to the bucket. Moreover, if the theory was right, these higher flames must have existed before Mr Carter's intervention (because he put the fire out in the hopper altogether). But there was no suggestion that the fire was in some way slowing down or slackening off by the time Mr Carter intervened; indeed all of the evidence was that this fire was developing all the time. Thus, because Mr Carter did not see flames coming out of the top of the hopper, and because Mr Carter said that he was not aware of any fire above him in the elevator, then it seems to me most unlikely that the necessary flames were ever created in order to achieve this effect.
471. There was some cross-examination about the likely temperature in the flames. Huge ranges of temperatures, from 320°C to 1250°C, were referred to. But there was no common ground and no comprehensive testing that could give even a guide as to what the temperature in the flames might have been. One thing however, does appear clear: the flames – at least in the centre of the hopper - were hotter than 100°C, which was the temperature to which the sensors in both the hopper and the elevator were set.

c) The Moment of Ignition

472. The Mitcheson Final theory is also very specific as to when the moment of ignition would have been: on this hypothesis, the key moment of ignition would have been when a bucket was in the tipped vertical position, and at no other time. This was because, as Dr Mitcheson accepted in cross-examination, the flames would not have set fire to the bottom of the buckets (i.e. the buckets in the horizontal position). He said that it was only the thin rim on the side of the bucket that was vulnerable, and that would only be exposed to the flame when the bucket was vertical.
473. How likely was it that the rim of the bucket was set alight by the flames in the hopper? Even assuming (which I do not) that the flames reached above the top of the hopper so as to reach the bucket rim, the flames only had between one and two seconds to set fire to the rim of the bucket. That is because Mr Stephens calculated that the buckets remained in the vertical/tipped position for that short period as they moved across the top of the hopper. After that, the buckets went back into the horizontal position, where it is accepted that they would not have been ignited.
474. There was no evidence to suggest that the bucket in the tipped position could have been ignited in the relevant period of one to two seconds. As noted above, the only tests that were carried out showed that the rim of a stationary bucket could not be ignited unless a flame was held against the rim for as long as 7 seconds. Thus, even if the flames were reaching up into the elevator, the limited tests carried out demonstrate that there was simply insufficient opportunity for the rim to be set alight when the bucket was in the vertical position. If it takes 7 seconds to ignite the rim of a stationary bucket, then the rim of the moving bucket could not have been ignited in the one to two seconds that was available. That is a further reason why I consider this theory to be wholly implausible.

d) The Spread of the Fire in the Elevator

475. It is next said that, once one of the buckets was on fire then, having moved back into the horizontal position, it was touching the next bucket and the fire could have spread. But the real question was how long that might have taken. Again, as Dr Mitcheson accepted, there were no tests to demonstrate how long it might take for fire to pass from one bucket to another. Although Dr Mitcheson said he thought it might take 30 seconds for fire to go from bucket to bucket, that was complete conjecture and, as he accepted, he had not made any attempt to transmit flame from one bucket to another bucket (despite the fact that this was an essential part of his theory).
476. Further, Dr Mitcheson had overlooked the fact that, although the buckets were touching as they moved back towards the pans in the upper section of the elevator, they very soon dropped down the vertical section of the elevator. During that part of the journey, the buckets would no longer have been touching. Bucket to bucket fire transmission would not have taken place during that part of the process. And thereafter, the buckets would have been back on to the conveyors and, if they had been ignited at that point, they would have almost certainly been seen by the operators.
477. Accordingly, it seems to me that the Mitcheson Final theory also fails at this point too. It has simply not been tested in any relevant way at all.

e) Smoke

478. The only part of the contemporaneous evidence which Dr Mitcheson was able to point to in support of his theory was the fact that there was some evidence of dark smoke in the packaging hall when Mr Bray and the fire fighters went into the Packaging Hall at 8:43pm. Dr Mitcheson's opinion was that this smoke was referred to in the contemporaneous notes as being 'dark' with an 'acrid' smell and taste. He therefore concluded that this must have come from the burning plastic in the elevator. However, there are a number of reasons why I consider that this part of Dr Mitcheson's evidence was over-stated.
479. First, there was no evidence that the smoke in the Packaging Hall had come from inside the elevator. Indeed, the contrary is the case. Nobody said that they saw energised smoke coming out of the opening from the C Line elevator where it came to feed the retail carousel in the Packaging Hall. Instead, the general assumption was that this smoke had come through from the popcorn production area in the gap around the elevator, where it passed through the wall. That was Mr Bray's positive evidence as to what he saw, repeated in a number of different places in the contemporaneous interviews. That therefore suggests that there was not a fire in the elevator, but instead a fire in the 'oil pop' production area.
480. Secondly, as to the appearance of the smoke, it cannot be said that those who saw the smoke thought that it was black or very dark, as it would have to be if it was produced by burning plastic. Mr Roberts did not refer to dark smoke. Mr Widdowson, one of the fire officers, referred to the smoke as "darkly grey" which does not necessarily suggest smoke produced by plastic. Furthermore, it was common ground that there were plenty of plastic products in the popcorn production area which could have caught fire, such as the shrink wrap and the cellotape. Dr Mitcheson accepted that there were many plastic materials in the production area which could have contributed to the 'darkly grey' smoke. No dark or black smoke can be seen on the video, either.
481. Dr Mitcheson said that the eye-witnesses also referred to the acrid or plastic smell in the smoke. But again that was an overstatement. Mr Roberts did not refer to any such smell. Mr Carter said expressly that the smoke did not have such a smell. And although My Bray referred to the smoke as "more acrid", it was unclear with what he was comparing it. Mr Widdowson did refer to a plastic smell but, for the reasons already noted, that smell could well have derived from the other plastic materials in the popcorn production area.
482. Accordingly, it seems to me that the evidence about the smoke does not point irresistibly to there being a fire in the elevator. Indeed, since the evidence was that the smoke was not coming out of the end of the elevator, the evidence about smoke supports the alternative view that I have reached, that there was not a fire in the elevator, and that the smoke in the Packaging Hall had come through the gap in the wall.

f) Pool fires

483. Mitcheson Final also required the molten plastic to drip out of openings in the elevator onto the floor of the factory and thus in some way spread the fire towards the south wall. That part of the theory was not developed in either of Dr Mitcheson's written reports. Furthermore, because this theory had only occurred to Dr Mitcheson late, it was not a possibility for which he had made any attempt to look for evidence in the days after the fire, when he was examining the ruins of the NMU. This is important. The evidence was that there would have been – or probably would have been – physical signs of such pool fires, if they had occurred. For example, Dr Lygate was asked whether there could be visible signs of the extent of spread of the pool. He said:

"Yes, there can be. Not in every case but generally one would expect to find evidence of the size of the pool because as I indicated, the material that is falling first on the floor is cooled, re-solidifies, and burning material falls on top of it, and as the polypropylene burns, you get the char formation which protects the material underneath. So often you will find areas of re-solidified plastic under the area where the pool has actually been."

484. Dr Mitcheson identified no such evidence, even though he had inspected the remains of the building after the fire. There was therefore no physical evidence that such pools had occurred. Furthermore, I also accept Dr Lygate's evidence that the rate at which any such pool fire would have formed, and the rate at which such a fire would have grown, would have been very slow. This was not a point that Dr Mitcheson addressed, but Dr Lygate was plain that a pool fire underneath the elevator would not develop rapidly enough and flow rapidly enough to reach the racking on the south wall in the time period required.

g) Conclusions on Mitcheson Final

485. Dr Lygate said that he thought Mitcheson Final was less probable than his own theories and subsequently described it as having a "very low probability". For all the separate reasons which I have given, I accept that description. In my view, Mitcheson Final was the least likely of all of the three theories that were put forward. It was certainly less likely than Lygate 2.

K7: Summary

486. For the reasons set out above, I consider that, of the three theories put forward as to fire spread, the most likely was Lygate 2, the fire spreading as a result of the discharge from the hopper onto the floor of a large quantity of burning popcorn, which was then displaced by stamping. Part of that discharge was inadvertent, namely the smouldering popcorn that melted the sleeve in Mr Hardcastle's hand, but a larger proportion of that discharge was deliberate, because it was the way in which the operatives had been taught to deal with a hopper fire.
487. In my view, Lygate 2 was the only one of the three theories that was consistent with, and actually supported by, the contemporaneous factual evidence. Lygate 1 required too may coincidences to be plausible and Mitcheson Final was both wholly untested and contrary to much of the factual evidence.
488. I acknowledged at the outset of this Judgment, and repeat here, that all three of the competing theories as to fire spread contained at least some improbabilities. This was in part because the operatives generally thought that the fire had been dealt with at the time (so its spread came as a surprise) and in part because, when he was first involved, Dr Mitcheson's remit apparently did not include any consideration of the spread of the fire. By the time that it became apparent that spread was a critical element in the case^[26], it was too late for any of the experts to consider the physical evidence that might have gone either to support or disprove any of their theories. Accordingly, I have to work with the (limited) material available to me and, for the reasons I have given, I am in no doubt that – on the evidence - Lygate 2 represents not only the most likely explanation for the spread of the fire of the three on offer, but was also, on the balance of probabilities, the way in which the fire spread that night.

L: LIABILITY

L1: Two Preliminary Matters

489. Two preliminary matters need to be addressed: whether, given the failure of the claimants' case as to fire spread, there is any liability claim open to them at all, and, if so, the relevance (if any) of the maxim res ipsa loquitur.
490. In opening, the claimants' case on liability seemed to be inextricably linked to the success of the Mitcheson Final theory as to fire spread. Superficially, this was unsurprising, because it was the only theory which involved the fire going into the elevator (where the defendant's CO2 suppression system was installed) as opposed to those open areas of the premises which were not covered by the defendant's system. It was therefore unclear to me whether the claimants had any other, simpler case, to the effect that the defendant was in breach of contract because the CO2 system had failed to suppress this particular fire in the hopper, wherever it went thereafter. On Day 6, I raised this question with Mr ter Haar, and he confirmed that the essence of the claimants' case was indeed based on the proposition that the fire passed from the hopper up into the elevator, the case on fire spread which I have rejected.
491. On one view, therefore, that might be the end of the case. That was certainly the submission advanced by Mr Dennys (see, for example, paragraph 112 of the defendant's closing submissions). But by the end of the trial, Mr ter Haar had refined his position significantly and he asserted, at paragraph 228 of his own closing submissions, that "whichever hypothesis as to fire spread the court adopted, the system failed and ADT are liable for the losses". This volte face was, I find, not due to any oversight by counsel; it was again the result of the experts' failure to think through the issues, and their evidence on those issues, at a much earlier stage and advise accordingly (see Section J above).
492. The question then becomes: despite its late entry into the case, and the failure to articulate it in answer to my specific question on Day 6, is this modified case open to the claimants? I have concluded that it is. It is within the claimants' pleadings, and the contrary was not suggested. It is the product of evidence of which the defendant was aware and on which the claimants' witnesses were extensively cross-examined. It is a claim which it is therefore open to the claimants to make. It is, however, a matter of regret that what is, on my analysis, their only sustainable way of recovering £110 million odd should emerge in this rather roundabout fashion.
493. Having found that the claimants' modified case as to liability is open to them, it is necessary to consider the overall shape of such a case. There was a debate on the pleadings, albeit not one raised directly in the oral submissions, as to whether or not the maxim res ipsa loquitur might apply in this case. I have assumed that the absence of any reference to it in the claimants' closing submissions indicates a decision on their part not to put their case by reference to the proposition that, just because there was a fire, there must have been a breach of contract on the part of the defendant. In my view, that was a sensible approach, because in cases like this, it is almost always necessary to try and explain what happened by reference to the terms of the contract and the possible breaches thereof. The failure of the claimants' warranty/guarantee argument also makes that a necessity.
494. That said, it seems to me that I have to approach the allegations of breach in the knowledge that, at the time of Mr Carter's intervention, there was smouldering, burning popcorn at the bottom of the hopper, and flaming popcorn all across the top of the hopper, which flames were large enough to extend to the top of that hopper. That was not only a potentially very dangerous situation, which required immediate intervention, but it was also something which, it might be thought, a properly-designed, automatic fire suppression system in the hopper ought to have dealt with. Instinctively, therefore, it seems to me that there is, and always was, a case for the defendant to answer. I stress that that cannot affect the burden of proof, which remains on the claimants throughout. But the critical question remains: if the automatic suppression system had been properly designed, commissioned, and installed, how was it that the hopper contained a large mass of burning and flaming popcorn which had not been extinguished?

L2 Design/Generally

495. As set out in paragraphs 216-217 above, the defendant was obliged to take reasonable skill and care in designing this system. I have already said that I consider this design obligation to be central, because it was that which required the defendant to sit down and work out how a fire in the hopper might be dealt with by a CO2 suppression system. It seems to me that, given the size of the hopper, the designer would have thought that there was at least a risk of a deep-seated fire in the hopper and one which needed to be suppressed before quantities of burning popcorn were discharged from the bottom and/or before the fire spread, either to the elevator, or in some other way, outside the hopper.
496. I have concluded that the defendant failed to comply with this design obligation. I have concluded that the highly developed fire in the hopper at 8:30 pm on 8 June 2005 was not one which a properly-designed automatic suppression system should have allowed to develop to such an extent. The system should have been designed to ensure that the risk of burning popcorn being discharged from the dispatch chute was minimised (even if it could not be eliminated altogether), and that any fire with flames that were rising to the top of the hopper, and had spread across its whole surface area, together with burning popcorn at the bottom of the hopper, was a fire which should have been extinguished. The defendant was in breach of contract for failing to design a system which met those reasonable requirements.
497. The principal reason for my conclusion that the defendant failed to use reasonable skill and care in the design of the system in the hopper is based on the evidence of Mr Jackman, the defendant's expert in fire suppression systems. In two lengthy passages of cross-examination, one at the end of Day 12 and one at the start of Day 13, Mr Jackman made a number of important admissions about the general deficiencies in the design of the system^[27]. Mr Dennys dismissed these parts of the cross-examination as being unfairly premised, which inevitably led to the answers provided by Mr Jackman, and there is something in that criticism. But, on a proper analysis, it cannot explain away the admissions which, in my view, were both inevitable and significant. I consider that Mr ter Haar's cross-examination on these occasions was generally fair and focussed, and he obtained a comprehensive set of answers from Mr Jackman which lead to the unavoidable conclusion that the defendant failed to design the CO2 fire suppression system using reasonable skill and care.
498. In those circumstances it is necessary to set out, at some length, the relevant passages of cross-examination. The cross-examination occurred on Day 12 (starting at page 170 of the transcript) and dealt with Lygate 1 (the theory that the fire was in the hermetically sealed sleeve within a box). Mr Jackman agreed that that fire had come via the hopper:

"Q: On that basis a fire, albeit a small smouldering fire on this hypothesis, had passed through the hopper, had not been detected by the CO2 system and the result was that the building burnt down?

A: Yes…

Q: You would accept that if the CO2 system had activated and extinguished the smouldering fire in the hopper, this property damage incident would not have occurred, would it?

A: If the operators had actually operated the manual release mechanism, that was there provided specifically for this sort of situation, I believe it would have extinguished the fire.

Q: That doesn't work, does it? That theory completely doesn't work because the first time they would have operated the manual release system would have been after they realised there was a problem. Do you agree with that?

A: Yes.

Q: By that time, on the theory that is put forward on behalf of ADT, this smouldering popcorn was already in a box waiting to erupt. So the operation of the manual system would not have made any difference, would it?

A: In that respect, no, it was already outside…

Q: The original hypothesis which Mr Lygate put forward, still one of his two hypotheses, namely that there was a latent source of smouldering popcorn, which had not been detected by the CO2 system and had not activated the CO2 system, if the lids were on, you accept this system failed?

A: I think there is a more fundamental question, and I am going to get told off I'm sure but…

Mr Justice Coulson: Well answer that one, and then tell us the more fundamental point.

A: The answer is yes."

499. As to Lygate 2, which I have found was the most likely explanation for the spread of the fire, the same questions were put. Mr Jackman agreed that "the system has failed to extinguish a developed or developing fire." Later on there was this exchange:

"Q: So is the answer to my question that if the CO2 system did not have VESDA protection, then the fire spread, as explained by Dr Lygate in his more recent report, would not have been prevented? The answer to that is yes, isn't it?

A: Probably."

500. As to Dr Mitcheson's theory, there was this exchange on Day 13 (page 7 of the transcript)^[28]:

"Q: Just so you can understand what I am putting to you. As I understand Dr Mitcheson's theory it is this, that in the mechanism that he described, the combustion had already started in the materials within the conveyor/elevator before the operatives left the premises. Take that as being the premise. Obviously we are going to argue before my Lord as to whether that's a safe assumption.

If that be right, you would accept, wouldn't you, that this system failed to respond as it should have done?

A: In as much that the CO2 wasn't ejected, yes."

501. Accordingly, whichever theory as to the spread of fire was right, in these passages of evidence Mr Jackman accepted that the general design or specification of the system had failed to come up to the necessary standard. That was confirmed in this exchange:

"Q: Now let me come back to the hypothesis of the fire in the hopper and, as we know, on whichever theory we are talking about, Lygate 1, Lygate 2 or Mitcheson, we have to start with either a smouldering or flaming fire in the hopper. It was against that risk that the heat detector in the hopper was intended to protect the building?

A: Indeed."

This answer then led on to Mr Jackman's confirmation that, according to the tests that had been carried out, the residence time (that is to say, the time that this fire had existed in some form or another in the hopper), was somewhere between 8 and 15 minutes. In other words, between the time when some form of burning material (as I have found, smouldering popcorn) entered the hopper, until the time that the smouldering popcorn emerged into Mr Hardcastle's hand, there was a period of between 8 and 15 minutes. Mr Jackman was then taken to the analysis in his supplemental report. On the assumption, against the claimants, that (contrary to my findings) all of the openings at the top of the hopper were open, the actual residence time was much longer than the time which Mr Jackman was saying it would have taken the CO2 system to detect the fire. He had two figures for such detection time: 153 seconds if the lids were open and 101 seconds if they were closed.

502. Accordingly, on the figures, the fire in the hopper on 8 June 2005 would have been burning for at least 8 minutes before the popcorn emerged into Mr Hardcastle's hand, which time exceeded, by a considerable margin, the appropriate detection times calculated by Mr Jackman. It was therefore put to Mr Jackman that this was exactly the risk that the system was designed (but had failed) to suppress. Mr Jackman sought to argue that the lids were a critical factor, but he was reminded that even the slower response time, assuming that the lids were off, was still far shorter than the length of time that this fire had actually been burning in the hopper.
503. There were then these critical exchanges with Mr Jackman:

"Q: You see, what I am looking at is your conclusion - scenario 1 actual, with the lids off. What this table appears to show is that, with the lids off, the fire would be detected in 153 seconds. Is that right or not?

A: That is what our model showed, yes.

Q: But I think you are now saying that, with the lids off, the fire might not be detected in 153 seconds. Which is the position?

A: I think in all probability it would be detected but as I said, with charges coming on, and we are not able to separate that radiation component, we already had a factor of safety in there.

Q: Alright, I am willing to take your answer, that in all probability a fire which had been burning for at least 8 minutes would have been detected by this system?

A: Yes…

Q: On the assumption that the pink line takes, that is to say that the openings are open, do you agree with me that if a fire had been burning for at least 8 minutes, the probability is that the heat detector should have detected that fire?

A: Yes.

Q: Just so that there is no possible ambiguity, that is on the basis that the lids and the other openings are open; you understand that?

A: Yes.

Q: And that probably would be even greater if the residence time was 15 minutes rather than 8 minutes, wouldn't it?

A: In as much more heat is being given off under those circumstances, yes."

504. It seems to me that these passages in Mr Jackman's cross-examination provide complete support for my initial reaction to the factual evidence in this case; namely that, if a fire was flaming at the top of the hopper and smouldering at the bottom of the hopper, a developed fire had taken hold in the hopper as a whole, and a properly-designed CO2 suppression system should have been automatically triggered some time before Mr Hardcastle's sleeve began to melt and Mr Carter intervened with the fire hose. The defendant was therefore in breach of its design obligations. On the basis of Mr Jackman's admissions, no other conclusion is possible. However (and for completeness), as noted in the next section of this Judgment, that conclusion is confirmed by a consideration of the individual elements of the design.

L3 Design/Specifics

a) Hopper, Not Elevator

505. It is important at the outset to stress that the particular element of the design with which I am concerned is the design of the CO2 system as it related to the hopper. There was a lot of evidence about the elevator and the openings in the elevator and their potential effect on the design. All of that would only have been relevant if the Mitcheson Final theory as to fire spread was upheld, because it is only that theory which involved a possible fire in the elevator. However, for the reasons set out in Section K6 above, I have rejected Dr Mitcheson's theory as to fire spread. Accordingly, the CO2 suppression system in the elevator is irrelevant for present purposes. Nobody suggested that the CO2 suppression system in the elevator should have been triggered by the smouldering popcorn as it passed on its way towards the hopper. The problems began when the popcorn was tipped into the hopper. It is therefore the design of the CO2 system in the hopper that matters.
506. I make this clear for four reasons. First, because it means that there is a good deal of evidence which I do not need to address, because that was concerned with the adequacy or otherwise of the design of the suppression system in the elevator. There is also much of the closing submissions, by both sides, which is similarly irrelevant.
507. Secondly, I make the point because it seems to me that, on the evidence, the defendant failed adequately to distinguish between the suppression requirements of the elevator, and the suppression requirements of the hopper. I consider that different design considerations apply to the elevator, where the popcorn was in relatively small quantities in the buckets, and passing along the conveyor, and the hopper, where a large volume of popcorn might sit for some time and where, so it seems to me, the fire risk was much more pronounced. And this links to the third point, which is that it again becomes relevant to note (paragraphs 136-140 above) that the defendant was confused as to the nature of the possible fire risk (deep-seated or surface) at least in part because of the failure properly to differentiate between the risk in the elevator and the much greater risk (due to the volume of popcorn) in the hopper.
508. The final reason why I make it clear at the outset that I am dealing with the liability allegations concerning the hopper only goes back to the ongoing complications caused by the change in emphasis in the claimants' case to which I referred at paragraphs 490-492 above. Thus, by way of example, at paragraph 102 of their closing submissions, the defendant maintains that "if the fire had spread beyond the hopper/elevator enclosure before the system discharged, then the design of the system is a causal irrelevance". That is just not right; the fire escaped the hopper in this case before the system discharged, because the system never did discharge, despite the developed nature of the fire. But the design issue remains fundamental: would a reasonably-designed system have extinguished that fire before it escaped (ie before there was burning popcorn being discharged at the bottom of the hopper, and leaping flames at the top)?

b) Going About The Design

509. Mr Stephens said that, at the outset, the defendant should have sought as much information about the elevator and hopper as it could. He said that the defendant should not have designed on assumptions, but should have gone out to find out the facts for itself. The defendant's suggestion appeared to be that this exercise was impossible for the defendant because the new production lines were not up and running at the time of the design, but that is a non-point, certainly in relation to the hopper, which was a basic piece of equipment, designed to store a large quantity of popcorn which had been freshly cooked and which was then to be put into sleeves. For the purposes of the fire suppression system, all that the defendant needed to know was how big the hopper was, and what the configurations were at the top and at the bottom. Moreover, the defendant produced a design on the assumption that the hopper was about 1 cubic metre in size and it does not appear that in the event, this assumption was significantly inaccurate.
510. In cross-examination on Day 13, starting at page 19 of the transcript, Mr Jackman agreed that the first step for a designer was to identify the risks which were to be protected and to design a system that suited the specific requirements of the risks to be protected. He agreed that it was good practice for the designer to make sure that he understood the production processes that were being carried out, what the combustible products were, and what the risks were of those products catching fire. He said that a designer would need to consider whether he was dealing with smouldering or flaming fires, and to be aware that one could turn into the other. In my judgment, as the factual evidence clearly shows, the defendant did not do any of this.
511. Importantly, Mr Jackman thought that the designer should work from detailed drawings and produce his own working drawings. He also accepted that where, as here, a specialist was providing a design, and then manufacturing and installing the product so designed, it was important to ensure that there was proper communication between the designer and the installer. If the designer was making any assumptions, particularly an assumption that something would be subsequently checked by the installer (such as the air tightness of the enclosure), Mr Jackman agreed that it was important that the assumption was communicated to the installer.
512. Again, I find on the evidence that none of this happened as Mr Jackman said that it should. There were no working drawings or design drawings here, which I think is one of the reasons why it has, in the past, proved so difficult to pinpoint precisely what went wrong with the system on 8 June 2005. Mr Jackman appeared to suggest that such documents were unnecessary because "this is a very simple system. It is two detectors, it is two valves and some piping." But he was forced to agree that the simpler the system, the simpler it would have been to produce working documents. None of the assumptions made by the designers were identified or subsequently checked.
513. Thus the way in which the defendant went about the design was not the way recommended by their own expert Mr Jackman. The position was further complicated by the fact that there were two designers, Mr Beaumont and Mr Flavell, who reached different design assumptions as to the nature of the risk being protected against, perhaps because, although they agreed that they did not have sufficient information on which to do a design (paragraphs 105 and 139 above), they went ahead with it anyway. The assumptions that they made were not passed on, either to the claimants or to their own internal staff, and this omission was exacerbated by the lack of design or as-built drawings. The inescapable impression was, I am afraid, that this was always viewed by the defendant as a quick job which could be done on the cheap.

c) Type of Fire

514. There was a good deal of debate between the experts about the type of fire – whether deep-seated or surface – that should have been envisaged at the time of the design. As a matter of fact, as set out in paragraphs 136-140 above, we know that Mr Flavell went for one, and Mr Beaumont the other. This mattered because a deep-seated fire required a longer period for the CO2 to be retained and therefore a more extensive and robust system.
515. Mr Stephens said that the design was defective because it appeared to be based on the premise that any possible fire would have been a surface fire, and he maintained that a fire in either the hopper or the elevator would in fact have been a deep-seated fire. There was a good deal of cross-examination about his view as to whether or not a fire in the elevator could be described as a deep-seated fire but, for the reasons that I have already given, that is now irrelevant.
516. The definition in British Standard 5306, Part 4 of a deep-seated fire was "a fire involving solids subject to smouldering." I have already found, contrary to the claimants' case, that the material in the hopper was indeed smouldering on 8 June 2005. It seems to me that, in accordance with that definition, and in accordance with common sense, the design of the system in the hopper should have been based on the assumption that any fire in the hopper would be a deep-seated fire. That was confirmed by Mr Jackman in his evidence on day 13:

"Q: Now you have accepted, haven't you, that it is a reasonable approach that the fire in the hopper could be regarded as a deep-seated fire?

A: I have accepted that, yes.

Q: [British Standard page 194 paras 3.6] A deep-seated fire is a fire involving a solid subject to smouldering?

A: Yes."

517. Notwithstanding Mr Jackman's agreement, that a fire in the hopper would be a deep-seated fire, and his express acceptance that smouldering fires were one of the risks that might be experienced in the hopper (Day 13/page 30), he sought on a number of occasions to slide away from this agreement, principally, I think, on the basis that popcorn was not solid but a granular product. However, it seems to me that that was an attempt by Mr Jackman to avoid the consequences of his agreement that the defendant should have – but failed to - design for a deep-seated fire in the hopper. I do not believe that either Mr Beaumont or Mr Flavell adequately distinguished between the fire risk in the hopper and the fire risk in the elevator. The system did not suppress a smouldering fire in the hopper, as it should have done, because the system had not been specifically designed to deal with a deep-seated fire in the hopper.

d) Type and Location of Sensor

518. The choice of the particular sensor and the location of that sensor in the hopper were both fundamental matters of design. It seems to me that, probably following on from the decision to do a design on admittedly inadequate information, and the failure to design for a deep-seated fire in the hopper, the defendant did not properly consider the type and the appropriate location of the sensor in the hopper. The sensor was the critical piece of equipment if a fire in the hopper was to be satisfactorily detected and extinguished by the CO2 suppression system. But, on the evidence, I find that insufficient consideration was given to choosing the most appropriate sensor and placing it in the best location within the hopper.
519. That both the type and the location of the sensor were inadequate can be demonstrated by the fact that, despite the flames which Mr Carter could see coming up to the top of the hopper when he stood on the gantry, and which Mr Norton could seen reflected in the elevator or the ceiling, the automatic system was not triggered. There cannot be any doubt that the temperature in those flames was more than 100°C (the temperature that the sensor was set at). So why had the system not been activated? There seemed to be two particular reasons for that.
520. First, the type of sensor that was chosen did not have a particularly fast reaction time (RTI) which, given the risk of a deep-seated fire in the hopper (as opposed to the elevator), I consider that it should have had. Secondly, the sensor was located very much to one side of the hopper. Thus the flames, which naturally concentrated in the middle of the hopper, and which, up to a point, may have given rise to some form of chimney effect into the opening of the elevator above, would have been slow to be detected by the single sensor off to one side of the hopper. I expand on those two design failures below.
521. The type of sensor was an Olsen 54B. Although this was a detector that had been used for this kind of system in the past, the evidence was that this had a relatively slow RTI. It was thermally insensitive. Although Mr Stephens would not necessarily have used one of these kinds of detectors at all, he was critical of the decision to use the Olsen sensor in the hopper^[29]. When other kinds of sensors were put to Mr Stephens, he made plain that he would have in fact used an infra-red detector to stop the elevator (which was not in fact very different from the VESDA system). However, that was not to be taken (as the defendant maintains in its closing submissions) that Mr Stephens was in some way giving the Olsen sensor a clean bill of health; on the contrary, he was critical of all such sensors, and therefore would certainly not have countenanced using a sensor that was thermally insensitive.
522. It seemed to me that the biggest single difficulty was in relation to the location of the sensor. Mr Stephens said:

"I think the problem with the detector in the hopper is it is actually almost in the corner, which is not going to be one of the locations where you will get the highest velocities and highest gas temperatures."

I find that this evidence was not only unchallenged but unchallengeable (indeed, I note that the issue as to location was not really addressed in the defendant's closing submissions, despite the fact that it featured early on in the claimants' closing submissions, at paragraph 10). It was perfectly possible that a deep-seated fire could start in the hopper and not be detected by the sensor because it was not located in a centralised position, where the flames and heat would be at their highest. That seems to me to be a singular deficiency in the design of this system.

523. This was also a point that was revisited with Mr Stephens in his re-examination. Although he said that he thought that the position of the sensor in the hopper was better than the position of the sensor in the elevator, he thought that the problem with the sensor in the hopper was that it was away from the place where the hot gases would build up. He said "that if you had a perfectly sealed hopper, the fire is still going to burn because you will get the hot gases going up through the elevator. And my opinion is that it will draw in air from the elevator down into the hopper at the same time. So you will have hot fire gases going up from the hopper and at the same time you will have air coming back down into the hopper from the elevator." The difficulty was that the sensor was located away from the central area where the hot gases rose and the air came down.
524. I accept Mr Stephens' evidence as to the inadequate location of the sensor. It seems to me that the combination of a sensor that was itself slow to react, with a location to the side of the hopper, out of the way of the likely area of flaming, created the problem on 8 June 2005: a flaming fire at the top of the hopper which the single sensor simply did not pick up so as to trigger the CO2 suppression system. That failure can also be linked back to the defendant's failure to go about the design in the way that Mr Jackman said that he would have expected (paragraphs 509-513), and the different approaches as to the type of fire (paragraphs 514-517). The failure to locate a more reactive sensor closer to the central part of the hopper was the principal cause the failure of the automatic suppression system on 8 June 2005^[30].
525. The claimants maintained a separate case that there should have been more than one sensor. However, it seems to me that, on analysis, this is only a version of the previous point as to location. On the evidence, if a sensor had been placed in the right position, it would not have mattered if there was only one. On the other hand, if there was a difficulty in precisely where the sensor should go or would fit, then more than one sensor would have been a good idea.
526. In addition, I should say that the deficiency of type and location of sensor provides a complete answer to the defendant's oft-repeated argument (paragraph 94(iv) of their closing submissions, for example) that there cannot have been anything wrong with the system because, if there had been a substantial fire in the hopper, the sensor would have been triggered. That is manifestly the wrong way round: there was a substantial fire which did not trigger the sensor. Why not? Principally because of the type and (particularly) the choice of location of the single sensor. In my view, that gives rise to a breach of the implied term as to reasonable skill and care. It does not give rise to a breach of the defendant's quality obligation because there was nothing wrong with the sensor itself. There was nothing to suggest, for example, that it was broken or badly made.

e) Concentration and Distribution of Gas

527. It was agreed that other critical elements of the design included the concentration of gas, the distribution of gas, and the retention time. In this case, the design concentration was 65% and Mr Stephens said he would not take issue with that as a design concentration. Mr Stephens made the point that he did not know what the extinguishing concentration of the CO2 would be for popcorn because that is not a material identified in the British Standard and there were no tests done to identify it, but it is therefore difficult to see what criticism can now be made of the defendant in consequence.
528. That brings us back to retention time, which is itself linked to the type of fire envisaged at the design stage. This is because the British Standard does not deal with any sort of suggested retention time for surface fires. For a deep-seated fire, which is what I have concluded the defendant should have designed for, the British Standard requires a retention time of 20 minutes. Mr Stephens thought that, if the enclosures were reasonably tight, there was more than enough gas in the design to fill the enclosure, although this seemed to be an answer directed at the elevator rather than the hopper. Mr Jackman's evidence on this topic was even harder to follow. In the transcript for Day 13, at page 47 he suggested that there was enough gas in the design to meet this retention time, whereas, at pages 48 and 49, he indicated that the amount of gas in the design may be insufficient. At page 53, he said that the gas may have been sufficient, after all.
529. For these reasons, I am not able to reach a concluded view as to whether there was sufficient gas capacity in the hopper as designed to extinguish the fire on the night of 8 June 2005. In any event, much of the debate about the gas was academic and hypothetical, because the real issue is why the gas was not triggered at all.

f) Nozzles

530. Although there was a dispute about the design of the nozzles, and the number of nozzles, I consider that it was as academic as the debate about the amount of CO2. Much of it was concerned with the number of nozzles in the elevator, which is now immaterial. And although Mr Stephens thought there should have been two nozzles in the elevator rather than one, he accepted that there were problems with the precise location of the nozzles because of the risk of fouling the bucket machinery. It is unnecessary for me to comment further on this allegation.

g) Summary on Design

531. I have already set out in Section L2 the general evidence from Mr Jackman which has persuaded me that my initial impression was correct, and that a fire in the hopper that produced large flames at the top, and smouldering popcorn at the bottom, was a fire which a reasonable CO2 system should have been designed to avoid. Although the expert evidence as to the precise deficiencies in the design which led to that situation was somewhat haphazard^[31], I have concluded that particular problems lay with the way in which the defendant approached the design (and the inadequate information on which it acted); the failure to identify the nature of the fire risk being designed against; and (particularly) the type and the location of the sensor in the hopper. The evidence was that the temperature of the flames at the top of the hopper, being in excess of 100°C, should have activated the system within, at most, 2 ½ minutes. In fact, due to the deficiencies that I have noted, the fire in the hopper burned for much longer than that without automatically triggering the system.

L4: Installation

532. Despite one or two vague suggestions during the evidence, there was really no sustainable case that the installation of the CO2 equipment was in any way deficient. I take the point, urged on me by Mr ter Haar, that it appears that the installation was carried out by electricians, who had little or no understanding of the CO2 suppression system. But that only points up the lack of design drawings (a criticism that I do accept); in the absence of any specific deficiency in the installation, I conclude that it is irrelevant to the fire.

L5: Testing and Commissioning

533. This aspect of the work can also be dealt with shortly. The defendant's specification made it plain that the defendant would not be carrying out a commissioning test which involved the discharge of CO2. That specification was accepted by the claimants. They therefore had no contractual right to a gas discharge test. Thus, although Mr Stephens purported to argue that the defendant "should not have said that: they should have made it clear that that [a discharge test] was what was necessary", I am bound to say that I consider that was both unrealistic and contractually irrelevant.
534. Both Mr Stephens and Mr Jackman were critical of the defendant's commissioning, but their criticisms related to the question of openings in the elevator, the defendant's failure to point out the openings to the claimants, and the effect that those openings had, or might have had, on their design assumptions. Again, however, because that issue related to the elevator (and therefore Mitcheson Final), it is now irrelevant.

L6: Conclusions

535. On the night of 8 June 2005 there was a fire in the popcorn in the hopper which radiated out from the centre. In consequence, smouldering popcorn was being discharged from the bottom of the hopper and flaming popcorn could be seen across the surface of the hopper.
536. My initial impression was that this fire was just the sort of developed, extensive fire which the automatic CO2 suppression system in the hopper was designed to put out. That view was confirmed by the answers that Mr Jackman, the defendant's expert on fire suppression systems, provided in his cross-examination by Mr ter Haar. In the light of those answers, it seems to me clear that, on any view, the defendant was in breach of contract.
537. The specific breaches of contract did not occur during the latter stages of the contract, namely at the time of the installation of the equipment and its commissioning. What went wrong in this case was the original design and specification of the CO2 suppression system insofar as it related to the hopper. No or insufficient thought was given to the risk that there could be a deep-seated fire in the hopper which, the longer it burned, could spread both downwards, towards the dispatch chute, and upwards, to the top of the hopper and over into the elevator, as well as threatening the integrity of the hopper itself. The system was dependant upon one single sensor, which did not provide a quick RTI, and which was located away from the central part of the hopper where, on the expert evidence, the flames were likely to be strongest. A different sort of sensor, in a more centralised position, was vital if the large volume of popcorn in the hopper was not to create a very real risk of a significant and deep-seated fire.
538. Accordingly, for these reasons, I consider that the defendant fell below the standard to be expected of an ordinarily competent fire prevention sub-contractor, and failed to design the CO2 suppression system in the hopper with reasonable skill and care. In addition, I conclude that there was a breach of the express obligation to comply generally with the British Standard, because, contrary to the clear recommendations there, the design did not meet the risk of a deep-seated fire. There was no other breach of contract on the part of the defendant, whether as alleged or at all.

M: CAUSATION

M1: The Issues

539. It is the claimants' case that the defendant's breach(es) of contract caused the catastrophic fire in the NMU. The defendant maintains that there were a series of failures on the part of the claimants which, they say, broke the chain of causation and rendered claimants wholly or principally liable for the fire and its consequences. These matters are put, first, by way of causation, and secondly by way of contributory negligence. In this Section of the Judgment I deal with the issues as to causation and leave the arguments as to contributory negligence to Section N below.

M2: General Principles

a) Evidential Burden

540. Whilst a defendant who asserts a break in the chain of causation must discharge the evidential burden of proving such a break on the balance of probabilities, the legal burden of proof rests throughout on the claimants to demonstrate that the defendant's breach of contract caused its loss: see paragraph 43 of the judgment of Gross LJ in Borealis AB v Geogas Trading SA [2010] EWHC 2789 (Comm); [2011] Lloyd's Law Rep, Volume 1, page 34.

b) The Significant Causal Effect of the Intervening Act/Omission/Conduct

541. At paragraph 2-78 of the 19^th Edition of Clerk and Lindsell on Torts, the learned editors said that, in order to comprise a novus actus interveniens, so breaking the chain of causation, the conduct of the claimant "must constitute an event of such impact that it obliterates the wrongdoing" of the defendant. The word "obliterates" is, I think, taken from the judgment of Buxton LJ in Roberts v Bettany [2001] EWCA Civ 109, although it appears that this wording has been deliberately omitted from the 20^th Edition. However, as Gross LJ noted in Borealis, "for there to be a break in the chain of causation, the true cause of the loss must be the conduct of the claimant rather than the breach of contract on the part of the defendant; if the breach of contract by the defendant and the claimant's subsequent conduct are concurrent causes, it must be unlikely that the chain of causation will be broken." In circumstances where the defendant's breach of contract remains an effective cause of the loss then in ordinary circumstances, the chain of causation will not be broken: see County Limited v Girozentrale Securities [1996] 3 All ER 834.

c) The Unreasonable/Unfair Nature of the Intervening Act/Omissions/Conduct

542. The 19^th Edition of Clerk and Lindsell had also suggested that "for the claimant's subsequent conduct to be regarded as a novus actus interveniens it should be such as can be characterised as reckless or deliberate" and that "unreasonable conduct can be dealt with by a finding of contributory negligence". This formulation is not in the 20^th edition because it was not accepted by the Court of Appeal in Spencer v Wincanton Holdings Limited [2009] EWCA Civ 1404; [2010] P.I.Q.R.P8. Sedley LJ said that he was uneasy about the importation of a formula ('recklessly or deliberately') from the field of criminal law where recklessness was commonly equated with intent.
543. As to unreasonableness, Sedley LJ noted that unreasonable was "a protean adjective. Its nuances run from irrationality to simple incaution or un-wisdom." He preferred a formulation based on "fairness" which suggested that:

"…a succession of consequences which in fact and in logic is infinite will be halted by the law when it becomes unfair to let it continue. In relation to tortious liability for personal injury, this point is reached when (though not only when) the claimant suffers a further injury which, while it would not have happened without the initial injury, has been in substance brought about by the claimant and not the tortfeasor."

544. Mere unreasonable conduct on the part of a claimant will not necessarily break the chain of causation. That is particularly so where the defendant's breach remains an effective cause of the loss, albeit in combination with the claimant's failure to take reasonable precautions in its own interest: see both Borealis at paragraph 45 and Girozentrale.

d) The Relevance Of The Claimants' Knowledge

545. The claimants' state of knowledge at the time of and following the defendants' breach of contract may, in some cases, be a factor of particular significance in any consideration of unreasonableness or fairness. The more the claimant has actual knowledge of the breach, the dangerousness of the situation which has thus arisen, and of the need to take appropriate remedial measures, the greater the likelihood that the chain of causation will be broken: see paragraph 46 of the judgment of Gross LJ in Borealis.
546. The case of Schering Agrochemicals Limited v Resibel NV SA [1992] CA (transcript number 1298, otherwise unreported) may be said to have some similarities with this case. There, the defendants provided safety devices to guard against the known risk of fire. They were defectively designed and did not provide the required protection. On 30 September 1987 there was a serious fire which gave rise to the claims. However, 3 weeks earlier, there had been a small fire in the same place, which had disclosed to the plaintiffs that the safety devices did not work. In consequence of that earlier fire, it was held, both at first instance and in the Court of Appeal, that the defendants were not liable in respect of the loss occasioned by the fire on 30 September. In the judgments of the court, the reasons for the result differed, ranging from a break in the chain of causation to a failure to mitigate and remoteness of damage. But the key to the case was that the smaller fire was held to have been sufficient to disclose to the plaintiffs that a breach of contract had occurred, and made it reasonable to expect the plaintiffs to take appropriate steps to minimise the consequences. A highly dangerous situation had been revealed and the production line ought to have been stopped so that a proper investigation could take place.
547. By contrast, in Girozentrale, the plaintiff did not know the defendant's breach of contract, which involved representations to potential investors in breach of the terms of the engagement letter. There, Hobhouse LJ said:

"Where a plaintiff does not know of a defendant's breach of contract and where he is entitled to rely upon the defendant having performed his contract, it will only be in the most exceptional circumstances that conduct of the plaintiff suffers to break the causal relationship between the defendant's breach and the plaintiff's loss. The plaintiff's conduct was not voluntary in the sense of being undertaken with a knowledge of its significance. Conduct which is undertaken without an appreciation of the existence of the earlier causal factor will normally only suffice to break the causal relationship if the conduct was reckless. It is the character of reckless conduct that it makes the actual state of knowledge of that party immaterial."

I should add that, at paragraph 46 of his judgment in Borealis, Gross LJ confirmed that he did not read the observations of Hobhouse LJ in Girozentrale as furnishing support for any rule of law that only reckless conduct would serve to break the chain of causation. Instead, he said, the observations were authority for a more limited proposition that, in circumstances where the claimant was unaware of the defendant's breach of contract, normally, but not invariably, only recklessness would suffice.

548. With those principles in mind I turn to the causation issues in this case. I do so under the following sub-headings:

a) Did the fire in June 2004 break the chain of causation? (Section M3 below);

b) Did the failure to operate the CO2 system manually break the chain of causation? (Sections M4 below);

c) Did the conduct of the operatives in emptying out and stamping on the burning popcorn break the chain of causation? (Section M5 below);

d) Did the failure to segregate the 'oil pop' area and/or the failure to install sprinklers break the chain of causation? (Section M6 below);

e) Summary (Section M7 below).

549. I consider all the allegations of a break in the chain of causation against the background of these findings:

a) The initial cause of the fire which destroyed the NMU was the existence in the hopper of a developed fire which had been allowed to burn for much too long, such that there were large flames across the surface of the popcorn, and smouldering popcorn at the bottom of the hopper;

b) Whilst the defendant provided no guarantee that smouldering or burning popcorn would not escape from the hopper, the fact that, on this occasion, burning popcorn was discharged from the bottom of the hopper was due to the failure by the defendant to design the CO2 suppression system exercising reasonable skill and care;

c) The burning popcorn that was discharged was scattered over a wide area of the floor. It set fire to other combustible materials and created the fire in the 'oil pop' production area which then spread to the rest of the NMU.

M3: The Fire In June 2004

550. It was the defendant's primary case on causation that, in accordance with the Court of Appeal decision in Schering, they were not liable for any of the damage caused by the fire. The argument ran as follows. There had been a fire in the hopper in June 2004. As a result, the claimants knew that the defendant's automatic fire suppression system did not work but failed to do anything about it. Accordingly, there was a break in the chain of causation which meant that, a year later, the defendant was not liable for the consequences of the fire.
551. It seems to me that, if I had upheld the claimants' argument as to the warranty/guarantee, and found that there was a term of this contract that the defendant had warranted that no fire would ever escape from the hopper, then the defendant's argument of a complete break in the chain of causation would have had considerable force. Indeed, it would then be impossible to distinguish this case from that of Schering because, on this assumption, the claimants would have known in June 2004 that the CO2 suppression system was not providing the guaranteed situation (whereby no burning popcorn escaped the hopper), and they would have had plenty of time (far longer than the plaintiffs in Schering) to do something about the breach before the fire in June 2005.
552. In his closing submissions, Mr ter Haar complained about this argument and said that, had it been properly pleaded, then further investigation would have been carried out in relation to the fire in June 2004. I do not accept that submission. Evidence was adduced from the claimants' own witnesses about the fire in 2004^[32], and evidence and records relating to all the fires in the NMU following the moving of the 'oil pop' process to that facility were disclosed. No additional documents or other material was identified as being relevant to the fire in June 2004 but somehow missing or otherwise not disclosed. I am therefore satisfied that the argument is open to the defendant. I am also satisfied that, if the defendant had provided the warranty or guarantee alleged by the claimants, the fire in June 2004 would have demonstrated to the claimants that the defendant was manifestly in breach, and their failure to act would have broken the chain of causation in its entirety.
553. But, for the reasons given in paragraphs 229-244 above, I have concluded that there was no such warranty or guarantee. In my judgment, the defendant was obliged to exercise reasonable skill and care in the design of the fire suppression system, but provided no warranty or guarantee that burning popcorn would never escape from the hopper. If that analysis is right, then the fact that there was a hopper fire in June 2004 would not necessarily have alerted the claimants to the possibility that the defendant was in breach of contract. I accept that the claimants ought to have investigated that hopper fire more thoroughly than they did, and this may have led to modifications to the system, but I cannot find that, in circumstances where there was no guarantee that burning popcorn would not escape from the hopper in certain circumstances, the claimants' failure to do anything about the earlier fire broke the chain of causation.
554. Indeed, I consider that the dearth of contemporaneous material indicates that the claimants were unfazed by the hopper fire in June 2004, and that only demonstrates still further my conclusion that at no time did the claimants ever believe that the CO2 suppression system provided any sort of guarantee or warranty at all. If anyone within the claimants' management at Monkhill had thought for a moment that there was such a warranty, then they would have reacted very differently to the June 2004 hopper fire. As it was, it appears that they considered the hopper fire to be unremarkable, which was only consistent with the conclusion that the contract between the parties contained no warranty or guarantee of the sort so vigorously contended for by the claimants in their closing submissions. Thus, in the light of the contractual obligations that I have found, this contention that the fire in June 2004 broke the chain of causation must fail.

M4: The Failure To Train Out The Manual Operation Of The System

555. The defendant's next contention in support of their case that they were not liable for the fire at all revolved around the claimants' failure to 'train out' the manual operation of the CO2 suppression system. The defendant contends that, if the operatives had been properly trained, at the first sign of the fire they should have manually activated the system. That would have put the fire out and there would have been no risk to the rest of the building.
556. Although, in my judgment, the claimants' failure to train out the operatives in the manual activation of the system was negligent (for the reasons summarised in Section N4 below), I do not consider that it broke the chain of causation. There are three separate reasons for that.
557. First, I consider that it would be wrong as a matter of principle to criticise the operatives as they strove to deal with the emergency that had been created. It is trite law that the claimants' conduct which could be described as a panic or reflex reaction, will not amount to a novus actus: see, for example, Brandon v Osborne Garrett Co Limited [1924] 1 KB 548. The same principle applies to the acts of rescuers: see Cutler v United Dairies (London) Limited [1933] 2KB 297. Thus the operatives' failure to activate the manual system was irrelevant from a causation perspective.
558. Secondly, although very few of the operatives knew how manually to activate the system, one of those operatives was Mr Carter. It is true that his knowledge of the manual activation was limited at best, but he had at least had a rudimentary discussion with Mr Barker when the system was first installed. Accordingly, Mr Carter may have been able manually to activate the system on 8 June 2005. He chose not to do so. He explained in his evidence that, when he was alerted to the fire, his immediate thought was that the CO2 suppression system "has obviously failed because it didn't go off automatically…something like that should go off on its own". He therefore had to put the fire out in another way. That explains his use of the fire hose. In what Mr Kear called 'the rush and the panic' of the moment, it seems to me entirely reasonable that, whatever training Mr Carter might have received about the manual activation, he concluded that the CO2 system had failed completely, and that he had to adopt another means of putting the fire out. In other words, it is not possible to say that, even on this basis, the failure to ensure that the operatives were trained in the manual activation of the system had any causal effect.
559. Thirdly, it must be remembered that the CO2 suppression system could not be manually activated unless and until one of the operatives had become alerted to the fire. Thus, with a fire burning inside the hopper, it was always going to remain unseen unless and until burning popcorn was discharged or the hopper itself caught fire. Accordingly, the fire would have already escaped the hopper before a decision would have had to have been made as to whether or not to manually activate the system. In those circumstances, any manual activation would have come too late in the causal chain to have had any effect.
560. For these reasons, therefore, I reject the defendant's second attempt to argue that the chain of causation was broken at the outset and rendered them not liable for any damage at all caused by the fire.

M5: The Stamping Out Of The Popcorn

561. The defendant's third causation argument concerns the stamping out of the popcorn by the operatives. They maintain in their closing submissions that it was not foreseeable that once burning popcorn had been detected, the operatives would empty out more burning popcorn from the hopper on to the floor in an attempt to stamp it out. Again, it is said that this broke the chain of causation.
562. I reject that submission for a number of reasons. First, it assumes that the fire was caused by the burning popcorn deliberately discharged from the hopper (which, on this hypothesis, was unforeseeable), as opposed to the burning popcorn that was discharged into Mr Hardcastle's sleeve (which I find was the foreseeable consequence of the defendant's breach of contract). I cannot make that distinction; on the evidence, the fire might have been caused by either or both, so the defendant cannot show a break in the chain of causation anyway.
563. Secondly, I reject that submission for much the same reasons as I have rejected the previous submission in relation to the absence of training as to the manual operation of the system. To the extent that the discharge of further popcorn was simply a manifestation of the failure to train the men in the proper use of the CO2 system, the reasoning above applies again. To the extent that the defendant now maintains that they could not have foreseen that, with burning popcorn released inadvertently from the chute, the operatives would deliberately let out more, it seems to me that the argument is unsustainable. It must have been foreseeable that, if burning popcorn was released inadvertently from the hopper, so as to alert the operatives to the existence of a developed fire in the hopper without the CO2 system having gone off, burning popcorn might then be deliberately released in order that it could be extinguished.
564. Further and in any event, for the reasons noted in paragraphs 557 and 558 above, it is not appropriate to criticise the operatives in the 'rush and the panic' of the situation, particularly in circumstances where, as here, that situation had been created by the defendant's breach of contract. It is not therefore possible to say in the circumstances that the stamping out of the burning popcorn broke the chain of causation. Accordingly the defendant's third causation argument is also rejected.

M6: Fire Segregation and Sprinklers

565. Of all the defendant's causation arguments, it was the arguments in respect of the absence of fire segregation and the absence of sprinklers which gave me the greatest pause for thought when considering the issue of causation. That is because I am in no doubt, for the reasons dealt with in much greater detail in Section N3 below, that the claimants were woefully at fault for the absence of both of these elements of construction in the NMU. I explain in that section how and why I consider that the claimants were unreasonable, even reckless, in their failure to take the advice or act on the recommendations of the Group Risk Department and that, in consequence, a very significant reduction for contributory negligence is required.
566. Those findings, therefore, fulfil that part of the necessary test (unreasonable or unfair conduct) by which the defendant has to demonstrate a break in the chain of causation (paragraphs 542-544 above). But the other element of that test, noted in paragraph 541 above, is the need to show a genuine break in the chain of causation: the happening of an event which became the true cause of the loss. As noted there, concurrent causes will not generally do; if the defendant's breach remains an effective cause of loss, the chain of causation will not usually be broken.
567. It seems to me that, despite the validity of their criticisms about segregation and sprinklers, the defendant cannot get over that second hurdle. I have found that the fire was their responsibility and stemmed from their breach of contract. That fire ran its course, starting in the 'oil pop' production area, and then running into the rest of the extension in the NMU and then the older part of the NMU. There are obviously arguments available to the defendant to the effect that the fire would have been slowed down, and could even have been stopped altogether, if the segregation had been present or if there had been a sprinkler system in place (although there was precious little expert evidence on the point). But, so it seems to me, such findings would simply mean that the absence of segregation and/or the absence of sprinklers became a concurrent cause of the damage in the further reaches of the NMU. Those omissions cannot, as a matter of causation, remove or render irrelevant the original cause of the loss.
568. Accordingly, I have concluded that, whatever the degree of default or unreasonableness on the part of the claimants, in relation to both segregation and sprinklers, the defendant is unable to demonstrate that these matters broke the chain of causation in accordance with the test as outlined most recently by Gross LJ in Borealis.

M7: Summary

569. For the reasons set out above, I am not persuaded that any of the four matters relied on by the defendant broke the chain of causation. The fire in 2004 (Section M3 above) would only have broken the chain of causation if the defendant had in some way warranted that no fire would ever escape from the hopper and for the reasons which I have given, there was no such warranty. The fire in June 2004 merely indicated that fire escaping the hopper was a real possibility and not something which the CO2 system was guaranteed to stop. In my judgment, that was already known to the claimants.
570. The failure to provide proper training in the manual activation of the system was a default on the part of the claimants but manifestly did not break the chain of causation: by the time manual operation became an option, the fire had escaped the hopper (Section M4 above). And for the reasons set out in Section M5 above, I also reject the suggestion that the stamping out of the popcorn on the floor broke the chain of causation, partly because I consider it was foreseeable, and partly because it would be wrong to criticise the operatives, given the circumstances which the defendant's default had created.
571. For the reasons set out in Section M6 above, I consider that, whilst the unreasonableness/unfairness element of the claimants' failure to provide fire segregation and/or sprinklers has been emphatically made out, and is therefore very relevant to the arguments as to contributory negligence, the defendant has not shown that the absence of segregation or sprinklers broke the chain of causation to the extent required by the authorities. Accordingly, all of the causation arguments fail.

N: CONTRIBUTORY NEGLIGENCE

N1: General Principles

572. Section 1 of the Law Reform (Contributory Negligence) Act 1945 provides:

"Where any person suffers damage as the result partly of his own fault and partly the fault of any other person or persons, a claim in respect of that damage shall not be defeated by reason of the fault of the person suffering the damage, but the damages recoverable in respect thereof shall be reduced to such extent the court thinks just and equitable having regard to the claimant's share in the responsibility for the damage."

Although originally designed to relate only to actions in tort, contributory negligence is a relevant concept in claims for damages for breach of contract, but with a very important limitation, discussed below. As to the question of "fault", the general approach has been to consider whether the claimant's act or omission was made against its own interest. Inevitably, questions of reasonableness or fairness become relevant.

573. In Forsitkrings Vesta v Butcher [1989] AC 852, the Court of Appeal made clear that there were three categories of contractual obligation relevant for this purpose:

i) Where a party's liability arises from breach of a contractual provision which does not depend on a failure to take reasonable care;

ii) Where the liability arises from an express contractual obligation to take care which does not correspond to any duty which would exist independently of the contract;

iii) Where the liability for breach of contract is the same as, and coextensive with, a liability in tort independently of the existence of a contract.

Only if the breach is in category iii) can a defence of contributory negligence arise.

574. This issue arose rather starkly in Barclays Bank PLC v Fairclough Building Limited [1995] QB 214, where the Court of Appeal made plain that, under a contract where the obligations undertaken by (in that case) a skilled contractor required strict performance, and did not impose any duty on the claimant to prevent breaches from being committed by the defendant, the claimant was entitled to recover in full damages in respect of the breaches, without any reduction for contributory negligence. Beldam LJ said that "contributory negligence is not a defence to a claim for damages founded on breach of a strict contractual obligation."
575. However, in the present case, the allegations of contributory negligence are open to the defendant because I have concluded that the critical breach of contract on their part was a breach of an obligation to design the system with reasonable skill and care. A co-extensive duty would have been owed to the second claimants in tort, and (subject to Section O below) to the first claimants (with whom there was no contract). This was not a breach of a strict liability obligation. Indeed, if it had been, for the reasons noted at paragraphs 553-554 above, I consider that the claimants' claim would have failed the necessary causation test entirely.

N2: The Background To The Allegations Of Contributory Negligence

576. There are a number of pleaded allegations of contributory negligence, of which the three most important were the failures to put in fire segregation and sprinklers, and the failure to train in the manual operation of the system. Mr ter Haar complained in closing that these matters were not properly pleaded, although he subsequently accepted that the pleading point had been resolved by agreed amendments immediately after the trial. And although he maintained that these matters had not been fully explored in evidence, I consider that this was no more a matter for legitimate complaint than Mr Dennys' criticism of the claimants' change of case on liability (paragraphs 490-492 above). The relevant witnesses were cross-examined and the allegations of contributory negligence were fully explored in evidence, much of which has been set out in the narrative sections of this Judgment. Indeed, the key material emanated from the claimants' own documents. No prejudice has therefore resulted from this (yet further) manifestation of the experts' failure to get to grips with the issues and the concomitant lag in the pleaded allegations.
577. Before looking at the specific allegations of contributory negligence, it is necessary to consider the background, because it is important to stress that these were not isolated errors made by a claimant organisation which, otherwise, properly dealt with all aspects of fire safety at the NMU. On the contrary, for the reasons set out at some length in Sections C, D, G and H above, these three particular failures were part of a much wider, indeed systemic, failure on the part of the claimants to analyse the risks that they were creating when they put the 'oil pop' production process into the NMU.
578. In my view, the events as they unfolded in this case make pretty sorry reading. There is an all-pervasive feeling that, time after time, the claimants simply failed to measure up to their responsibilities and failed to take the repeated opportunities that they were given to address the problems. In those circumstances, the three specific alleged defaults in respect of fire segregation, sprinklers and manual training are not isolated or out-of-character events, but merely three aspects of an overall and repeated systemic failure. In reaching that conclusion, I have regard in particular to the following failures, referred to in the earlier parts of this Judgment:

i) Prior to completion of the defendant's works

a) The failure to act on the knowledge that a previous factory had been destroyed by fire due to the popcorn process and that fires were commonplace at the factory in Leeds (paragraphs 50-52 above);

b) The failure to undertake a proper or comprehensive review of the fire risks inherent in the 'oil pop' process before the scope of Project Kettle was finalised and before the Capex application was made (paragraphs 55 and 77 above);

c) The failure to operate a proper management system so that it was clear who was taking responsibility for which decision and which risk (paragraph 62 above);

d) The failure to provide the Group Risk Department with full information and the failure to follow the advice that was provided in June 2003 (paragraphs 63-78 above);

e) The failure to take proper advice about or undertake a technical appraisal of the CO2 system and to work out what such system protected and what it did not (paragraphs 91-96 above);

f) The failure to provide proper drawings or other documents showing the proposed 'oil pop' process and the machinery and equipment that were to be protected by the CO2 system (paragraphs 98-105 above);

g) The failure to seek an updated Fire Certificate or carry out with the Fire Officer a comprehensive review of the 'oil pop' process either on 5 September 2003 or thereafter (paragraphs 121-132 above);

h) The failure to analyse how the fire in November 2003 had started and whether the CO2 system would in fact have made any difference (paragraph 144 above);

ii) Post-Contract Events

i) The failure to deal promptly with the oil spillage problem and the somewhat 'Heath Robinson' nature of the attempts to resolve it (paragraphs 254-259 above);

j) The failure to assess the quantity and cause of both pan fires and conveyor fires and the failure to appreciate that it was only chance and the alertness of the operatives that prevented the fire from passing into the elevator and the hopper (paragraphs 260-266 above);

k) The failure to appreciate that the VESDA system would detect such fires and thus add significantly to the protection provided by the CO2 system; the failure to install the VESDA system after its successful trial; and the attempts subsequently to denigrate the VESDA system on spurious technical grounds (paragraphs 293-308 above);

l) The failure to enter into a maintenance contract until prompted to do so by the Group Risk Department (paragraphs 309-313 above);

m) The making of modifications to the hopper and the elevator without having a proper record of precisely what was being changed and why, and without expressly seeking advice from the defendant on the potential effect of these modifications (paragraphs 314-322 above);

n) The failure to reconnect the sensors for over 5 months following the modification works, allowing the process to operate without any fire protection at all for the whole period (paragraphs 323-332 above);

o) The failure to consider or address the significance of the fire in June 2004, the failure to seek advice on the risk that the CO2 system would not necessarily prevent a fire in the hopper and, due to the operatives' subsequent action, that the fire in the hopper was being dispersed on the floor of the production area (paragraphs 267-269 above);

iii) Health and Safety

p) The failure to ensure that the operatives complied with the written safety guidance or alternatively that the written safety guidance correlated to what happened in practice (paragraphs 333-349 above);

q) The failure to use a comprehensive and consistent system of written fire reports (paragraph 347 above);

u) The failure to appreciate that some of the procedures used in practice, such as letting out burning popcorn from the bottom of the hopper and then stamping it all over the floor, carried with it a significant fire risk;

r) The failure to obtain a proper and up-to-date Fire Certificate after the 'oil pop' process was installed at Monkhill and the failure even to go back to WYFRS after early September 2003 (paragraphs 129-130 above);

s) The failure to carry out a Fire Risk Assessment after the 'oil pop' process was installed at Monkhill (paragraphs 349-351 above);

t) The failure to act on the recommendations of the Group Risk Department in their June 2004 report, including the failure to carry out a proper survey or investigation into the nature on the construction of the NMU (paragraphs 352-366 above);

u) The failure to address the matters set out in the health and safety summary at paragraph 367 above.

579. It is in the context of those various failures on the part of the claimants, extending over a two year period before the fire, which form the necessary background to a consideration of the allegations in respect of the fire segregation, the sprinklers and the failure to train out the manual system.

N3: Segregation and Sprinklers

a) Introduction

580. These two items were dealt with together in the evidence at trial, and have indeed been dealt with together in the other reported cases. It is therefore convenient to deal with them together in this Judgment. However, as is explained below, these were two separate failures on the part of the claimants and I consider that, on the particular evidence of this case, each of them had a major effect on the damage and destruction beyond the 'oil pop' production area.

b) Particular Authorities

581. In J Sainsbury PLC v Broadway Malyan [1999] PNLR.286, HHJ Humphrey Lloyd QC ruled that Sainsbury had not been contributorily negligent in failing to put in sprinklers. He said:

"Sainsbury was in my judgment free to decide whether it was in its own commercial interest to install sprinklers. Others might have done so but I do not consider Sainsbury was at fault in not doing so. BM knew that there would be no sprinklers, as did EGP. They must be taken to have accepted the risk that any damage caused by negligence might therefore be greater than it would have been. I see no ground for holding that there was contributory negligence in not having sprinklers."

As to the absence of fire barriers in the ceiling void, the judge reached the same conclusion, pointing out that no one from the relevant professionals had suggested that there was anything wrong in omitting cavity barriers. However the judge did go on to make a finding of contributory negligence on the part of Sainsbury in relation to one of the fire walls which had been built to a defective design which the Sainsbury representative had failed to pick up.

582. In Fosse Motor Engineers Limited v Conde Nast and National Magazine Distributors Limited [2008] EWHC 2037 (TCC) Akenhead J dismissed the claim against the defendants for reasons concerned with causation. His findings in relation to contributory negligence where therefore obiter. However, on the topics of compartmentation and sprinklers, he said:

"111(a) It was asserted that much of the fire damage was attributable to the absence both of a suitable system of sprinklers and of compartmentation at the warehouse.

(b) There was no statutory requirement (such as Building Regulations) for a building of this size built in 1976 to have sprinklers or compartmentation.

(c) None of the literature at the time of the fire makes it clear that for an existing building of this age these are required.

(d) There are numerous warehouses which do not have such systems in. Put another way, there seems to be a respectable body of opinion among warehouse owners that they are not necessary.

e) Fosse was not required by the Fire Brigade who checked the building over some time before the fire or their insurers to install sprinklers or compartmentation. Fosse had taken some steps to provide an expensive new fire alarm system and other safety measures.

f) The evidence on compartmentation was not satisfactory. There was no attempt to explain what compartmentation was needed as a reasonable minimum or the extent to which it would have limited fire damage. If it was extensively compartmentalised the damage would have been much less; if it was limited to means of escape for people little of the building would have been saved."

583. There is an immediate and obvious part point of difference between the present case and the two authorities that I have cited. In these cases, the premises were, respectfully, a large superstore and a commercial warehouse; here, the NMU was a large manufacturing unit and the 'oil pop' process involved the use of naked flames.

c) Statutory Requirements

584. It was not entirely clear whether, for a building of this kind, housing a production process that involved naked flames, compartmentation and/or sprinklers were mandatory requirements or not. The Building Regulations are not always very helpful in determining such matters, because they are usually couched in such general terms. Moreover, they envisage close co-operation between the building owner and the local fire service, which did not happen here.
585. Although it was agreed by the experts that, if it was designated as an ordinary factory, the NMU was not required to have compartmentation or sprinklers, it was also agreed that this might very well have been different, and additional fire protection and enhanced compartmentation might have been imposed, if the NMU had been designated a place of special hazard. It seems to me that, because of the use of naked flames in its manufacturing process, and the mixed use of the building, the NMU would probably have been so designated; that was, I think, the effect of Mr Kemp's evidence. Thus, if anybody had turned their mind to the question of Building Regulations at the time, they may have concluded that, in the particular circumstances here, compartmentation and sprinklers were required.
586. However, all of this is immaterial for present purposes because, on the plain evidence in this case, and whatever the Building Regulations said, the claimants' management at Monkhill knew that fire segregation was critical^[33] (and wrongly assumed that they had it), and were twice advised to put in sprinklers and twice failed to follow that advice, both times on a false basis. These points are expanded below.

d) Fire Segregation

587. I have set out at some length at paragraphs 31-40, 56-57 and 68-69 above the assumptions made by the claimants' Group Risk Department, and the management at the Monkhill site, relating to the segregation of the 'oil pop' production area. All that evidence went one way. The claimants knew that they were putting a potentially dangerous process into an area that had been previously used for storage. As Mr Caldicott explained, they were only happy to do that on the assumption that they were putting the process into a 'fire box', that is to say an area that had been properly segregated from the rest of the building. If Mr Caldicott or Mr Hamilton had discovered for themselves, or if they had been told, in the summer of 2003, that the former Production Area B was not a 'fire box' and was not segregated in any way from the rest of the NMU, then I am in no doubt at all that, on their own evidence, they would have taken steps to provide the necessary fire segregation. They would not have allowed the move to go ahead without it.
588. The claimants were acutely aware of the importance of fire segregation in this instance. That was why they wanted a 1 hour roller shutter door put in the wall: it was because they believed that the wall itself provided 1 hour fire separation. Even that was far less than the 4 hour rating provided by the claimants' own internal guidance (paragraph 36 above). It was on that express basis that the move to Monkhill went ahead, and on that express basis that the subsequent references to fire segregation were addressed. The claimants thought they had done what they knew, as an absolute minimum, they needed to do.
589. In fact, the claimants wholly failed to carry out a proper survey in 2003 and repeated that failure a year later. They did not even bother to do a 5 minute check on the as-built drawings. Instead the evidence to which I have referred above makes clear that the claimants persuaded themselves that the necessary fire separation was already in place, without a shred of evidence to support such a conclusion.
590. As a result of the claimants' failure, they put a dangerous process, which carried with it the plain and obvious risk of fire^[34], into a part of the building that was not fire segregated at all, and was right next door to the Packaging Hall, where there was a large amount of combustible material. To that extent, at any rate, I accept Mr Dennys' description that the fire on 8 June 2005 was indeed "an accident waiting to happen".
591. Accordingly, contrary to the views expressed on the facts in the cases that I have cited above, I have concluded that the claimants were in obvious default by not segregating the 'oil pop' production area. I reach that conclusion on the facts. Unlike the two cases to which I have referred, the claimants here were not only aware of the importance of fire segregation, they fully intended to achieve such fire segregation. Unhappily, they went about realising that intention in such an inept way that, although the absence of such fire segregation would have been obvious to anyone who bothered to pull out the as-built drawings, the claimants merely persuaded themselves of the contrary position, and left it at that. That was wholly unreasonable conduct; if necessary, given how easy it would have been to check the construction records, and the number of express opportunities the claimants were given to do just that, I would hold that their failure was reckless.
592. What was the causal effect of this? In my judgment, it was very significant. The fire brigade were on the scene within 15 minutes of the alarm. If the 'oil pop' production Area was in the condition which the claimants erroneously believed it to be, then that was only a quarter of a way through the 1 hour fire protection provided by the (non-existent) fire box. The fire brigade would have had plenty of time to tackle the fire in that area and put it out over the next 45 minutes, so as to ensure that, even if the 'oil pop' area was destroyed, together with the machinery inside it, the fire might well have been contained within the area of the fire box.
593. Accordingly, the concurrent cause of the fire in the other parts of the NMU was the claimants' failure to do what they knew they had to do, and what they erroneously thought that they had done, namely to provide a fire box around the 'oil pop' production area. For these reasons, I consider that a significant reduction by way of contributory negligence is required because of the absence of segregation.

e) Sprinklers

594. During the course of the evidence, I pointed out that I have never tried a fire case in which it was not suggested by the defendant that the claimants should, long before the fire, have put in sprinklers. Nor have I ever tried a fire case in which the claimant's claim for the cost of rebuilding did not include the cost of a full sprinkler system. Both of those elements feature in this case. But what is unique, in my experience, is that here the claimants' management at Monkhill were advised, not once, but twice (and by their own Group Risk Department, on whom they said they always relied) to put in a full sprinkler system and, on both occasions, they failed to follow those recommendations.
595. That is the most significant reason to distinguish this case from Sainsbury v Broadway Malyan and Fosse Motors. Unlike there, these claimants were positively advised that sprinklers should be put in; that sprinklers were the only certain way of eliminating the risk of fire. It is therefore necessary to look at the reasons why they were not installed, to see whether or not the conduct of the claimants can properly be criticised in refusing to follow the recommendations and, if so, what the casual significance of this default might be.
596. In June 2003, the Group Risk Department advised for the first time that sprinklers were the only truly reliable way of ensuring that there would be no fire^[35]. On the evidence, this advice was simply ignored. No proper consideration was given to putting in sprinklers, despite this recommendation. That was apparently because the Capex Application for Project Kettle had not included for sprinklers and, although the recommendation came before the Capex Application had been finally signed off, the limited workscope had been signed off by some of the stakeholders already and was clearly regarded as a "done deal" by the others. Because Project Kettle did not include for sprinklers, and because it was on a tight timetable anyway, Mr Hamilton was able to dismiss even the suggestion of sprinklers on the basis that it was 'outside the scope of Project Kettle'.
597. This was completely the wrong way round (paragraphs 77-78 above). Instead of working out what they needed to do in order to ensure the safe transition of the oil pop production process from Leeds to Pontefract, and then make an application for the costs of doing that work, the claimants put together a specification which was expressly based on minimum spend and disruption and the minimum in terms of safety, getting approval for that application, and then, when additional matters were inconveniently recommended, they were able to reject them on the basis that they were outside the approved workscope. This cavalier attitude to the recommendations of their own Group Risk Department was all the more surprising given the repeated evidence from those involved that those on site relied on the Group Risk Department for all technical matters.
598. The extent and scope of the claimants' failure to take on board the recommendations of the Group Risk Department were exacerbated by the events in the summer of 2004 when, for the second time in a year, there was a clear recommendation to Monkhill to put in sprinklers^[36]. This time, the report even spelt out the financial advantages of sprinklers; for a spend of £500,000, a projected loss of just under £30 million would be reduced to just under £1 million (see paragraph 357 above). The advantages are not therefore only apparent in hindsight; they were, or should have been, clear at the time.
599. Why was that recommendation ignored for a second time? Again, there was no obvious answer to that, because at no time did the claimants' staff at the NMU sit down with the report, consider the costs/benefit analysis, or give any thought at all to the pros and cons of putting in sprinklers as recommended. I am sorry to say that the inescapable conclusion is that the claimants' management at Monkhill were desperate to avoid doing anything which involved them spending money.
600. What is more troubling is the reason publicly put forward to the Group Risk Department as to why the recommendation was not being followed. The Monkhill management claimed that there was no need for sprinklers because the Fire Service had approved the existing fire protection mechanism. As I have pointed out at paragraphs 132 and 358 above, that was quite wrong on the facts. The Fire Service had reached no such agreement, and the contemporaneous documents and the oral evidence made that clear. This was a significant misrepresentation, although whether it was deliberate or inadvertent I am unable to say. It is yet another important difference between this case and Fosse Motor Engineers.
601. But even giving Mr Hamilton the benefit of the doubt, and even assuming that he did genuinely (but erroneously) believe that he had reached some sort of understanding with the Fire Officer on 5 September 2003, the very least he should have done in 2004, faced with the repeated recommendation of sprinklers, would have been to go back to his files and check his records of the alleged agreement with Mr Kemp of the previous year. He would have found that there were no such records; that there was no updated Fire Certificate; and that there was nothing to say that there had ever been any sort of detailed agreement with the Fire Service approving the existing arrangement. He ought then to have gone back to the relevant Fire Officer, arranged a visit, gone through the proper process, and obtained, in writing, an agreement that the fire protection systems, without sprinklers, were indeed adequate. In all probability, on the available evidence, that agreement would not have been forthcoming, and WYFRS would have agreed with Group Risk about the installation of sprinklers. But either way, even if Mr Hamilton was right and his original discussion with Mr Kemp had ranged over these matters, the very least he should have done was to ensure that the matter was properly followed up and recorded.
602. For these reasons, I conclude that the failure to install sprinklers was highly unreasonable. If necessary, I would categorise it as reckless. Moreover, I am in no doubt that the absence of sprinklers had a very significant causal effect. If there had been sprinklers, they may well have provided full protection for at least the rest of the NMU, beyond the 'oil pop' production area. Whilst I cannot on the evidence say that sprinklers would definitely have put out the fire beyond the 'oil pop' production area, such an outcome must have been at least likely. Thus the absence of sprinklers was a concurrent cause of the damage beyond the 'oil pop' area.
603. The significant causal effect of the failure to put in sprinklers can be demonstrated by reference to a number of elements of the claimants' own evidence. When Mr Caldicott accepted in cross-examination that the claimants had accepted the risk that there were no sprinklers (Transcript Day 5/181), he was effectively accepting that, without sprinklers, there was a risk that the fire would take hold and then burst out of the oil pop area and spread, unchecked, throughout the rest of the NMU.
604. It can also be seen in the June 2004 report from the Group Risk Department. That warned that the damage to the claimants with the 'deficiency' (i.e. without sprinkers) would be just under £30 million^[37] and advised that, if sprinklers were put in, the figure would be under £1 million. The thirty fold increase is plain enough from the report: it is the result of a comparison between a small part of the NMU being damaged, with a concomitant small area of machinery also lost, against the loss of the whole building. The claimants knowingly ran that risk, and they lost.

f) Appropriate Percentage Reduction

605. What, then, should be the appropriate reductions to reflect the findings of contributory negligence that I have made in relation to the absence of fire segregation and the absence of sprinklers? The parties provided very little assistance in relation to the appropriate percentages, although Mr Dennys argued that if, as I have done, I upheld the sprinkler argument, then a reduction of close to 100% was justified. Mr ter Haar did not really address the percentage argument at all.
606. In many ways, the best evidence as to the appropriate percentage reduction in respect of sprinklers can be seen in the Group Risk Report of June 2004, which demonstrates that the losses were likely to be thirty-fold greater without sprinklers. On its face, this might be said to justify a reduction of 95% or more on the sprinkler allegation alone. But I have concluded that it would be inappropriate to make a reduction quite that large. I have to reflect the competing causes of the loss and the chances that, even with sprinklers and/or segregation, the fire damage might still have been extensive. There is also the possibility that the sprinklers and/or the segregation may not have been as effective in practice as they should have been in theory.
607. Doing my best, therefore, on the basis of the information before me, I consider that it is appropriate to make a reduction of 75% to reflect the absence of sprinklers and the absence of fire segregation. Although more generous to the claimants than their own June 2004 report might suggest, it seems to me that this properly reflects the state of the evidence and takes due allowance, to the extent possible, of all the competing uncertainties. I also take some comfort from the fact that this sort of reduction is common in those personal injury cases where, although liability is established against the defendant, the claimant can be shown, in the round, to bear significantly more than half the responsibility for the accident.
608. Accordingly, I find that the appropriate percentage reduction to reflect the absence of sprinklers and the absence of fire segregation was 75%.

N4: The Failure to Train

609. For the reasons set out in paragraphs 270-286 above, I consider that the claimants were additionally negligent and at fault in failing to train all the operatives in the manual use of the system. However, for the reasons set out in paragraphs 555-560 above, I do not consider that that failure had any sort of causal relevance and, accordingly, was nowhere near as significant as the failure in respect of segregation and sprinklers.
610. The remaining question for me is whether or not I ought to make a further reduction in the claimants' recovery to reflect this default. I have concluded that, in the round, there should be no further reduction. It seems to me that, in the circumstances of this case, a result which renders the defendant liable for 25% of the losses, but renders the claimants liable for 75%, is fair and equitable. I do not consider that it would be appropriate to make a further reduction to reflect this element of contributory negligence. The breach/failure on the part of the claimants in relation to training is, however, relevant generally because it seems to me to demonstrate neatly the claimants' ongoing failure to put in place a proper system of health and safety.
611. Further, whilst this separate failure does not, on my primary analysis, add to the percentage reduction for contributory negligence, it could operate to justify the same result if, for whatever reason, I was wrong to conclude that the claimants were at fault for either the lack of segregation or the lack of sprinklers, or if it was said that the 75% for sprinklers and segregation on their own was too high.

N5: Other Matters

612. A number of other matters were alleged by way of contributory negligence. They can be dealt with shortly.
613. The defendant alleged that the hopper lids were left off. I have found that, as a matter of fact, this allegation has not been made out. In any event, for the reasons explored above, even if the lids had been left off, it had little or no effect on the fire and the analysis on causation. This allegation therefore had no relevant causative effect.
614. It is alleged that the employees threw boxes containing smouldering popcorn towards the south wall. This allegation has not been made out on the facts; on the contrary, I have found that there was no smouldering popcorn in the boxes. In any event, even if, contrary to my findings, there was smouldering popcorn in the sealed sleeves, the operatives did not know that and cannot therefore be criticised: see paragraph 557 above.
615. There is an allegation that the operatives delayed in alerting the fire service. I do not understand that allegation. The fire was noted at 8:30 pm and the fire alarm sounded, thereby automatically alerting the defendant, at 8:33 pm. I do not believe that the operatives can be criticised because they spent 3 minutes trying to deal with the fire themselves. Indeed, in my view, the operatives, and Mr Carter in particular, should be commended for the brave and sensible steps that they took in trying to deal with the fire. It was not their fault that they had not been properly trained in the manual use of the CO2 suppression system.
616. The final allegation concerns VESDA. I have dealt with that at paragraphs 293-308 above. Whilst I am critical of the claimants for failing to think through the implications of the VESDA system, and for letting the proposal simply wither away, I do not believe that, in the round, the failure to install the VESDA system amounted to contributory negligence. The critical difference between the alleged failure to install the VESDA system, on the one hand, and the lack of compartmentation and sprinklers on the other, is that the compartmentation and sprinklers were either assumed to be in place or positively recommended by the claimants' own Group Risk Department. VESDA, on the other hand, was touted merely as a potential add-on to the CO2 suppression system by the defendant, and was never recommended by anyone else. With hindsight, VESDA should perhaps have been added to the CO2 system, and it might have made a difference on the night of the fire (although no-one can say for sure) but, given the absence of positive advice or recommendation from the defendant, contributory negligence does not run.

O. DUTIES OF CARE/TITLE TO SUE

617. As noted above, I have not so far found it necessary to distinguish between the two claimants. There is, however, a very clear distinction between them. The first claimant ("Trebor Basset") owns the NMU, and therefore claims the cost of replacing the building. The second claimant ("Monkhill") owned the business and the machinery and equipment within the NMU.
618. The claims that I have analysed above are essentially Monkhill's claims against the defendant in contract. However, to the extent that it makes any difference, I find that the defendant also owed Monkhill a duty of care at common law. I consider that this duty, which was not ultimately disputed by Mr Dennys, would be co-extensive with the contractual duty, and would not give rise to any different liability on the part of the defendant.
619. Trebor Basset, on the other hand, had no contractual link with the defendant. In those circumstances, of course, they would only have a right to recover if they were owed a duty of care by the defendant. As I understand it, subject to the next point, Mr Dennys does not dispute that such a duty of care was owed to Trebor Basset, although he properly reserves the defendant's position in respect of all claims for economic loss.
620. The defendant also raises points as to the ability of either claimant to recover some of the specific heads of loss. Thus, by way of example, the defendant accepts that Trebor Bassett had title to sue for the cost of rebuilding, because there was evidence to show that they owned the building, but the ownership of plant and equipment has not yet been established in the same way. There is also an issue about the cost of 'keeping the business going' during the rebuild. It has been expressly agreed that all such disputes can be left over to any hearing on quantum.
621. There is a suggestion, at paragraph 118 of the claimants' closing submissions, that the reduction for contributory negligence that I have made should only affect the claim by Monkhill (because it is based on the acts and omissions of their employees) and should not affect the claim by Trebor Bassett. Although I am unable to reach a concluded view on this point (because it was not developed in argument, and I have heard no evidence relating to it), it seems to me to be very likely to be wrong in principle and unsustainable on the facts. Following Cadburys' takeover of Trebor Bassett, it may well be artificial to distinguish between the two companies or their employees; indeed, it is unclear whether Trebor Bassett had any employees at all after the takeover. If the management at Monkhill behaved in a negligent and reckless fashion, it seems to me that that should adversely affect the claims of both companies, particularly as the defaults related to the condition of the buildings which Trebor Bassett owned, and which they had therefore failed to put into a proper and safe condition. Any other result would not reflect the merits or justice of the position as I have found it to be.

P: CONCLUSIONS

622. For the reasons set out in Section F above, I have concluded that the contract between the parties did not incorporate the defendant's terms and conditions. There was therefore no relevant limitation of liability. I have also concluded that there was no warranty or guarantee provided by the defendant to the claimants and that their principal obligation was the one originally pleaded by the claimants, namely that the defendant owed an implied obligation to carry out the design of the system exercising reasonable skill and care.
623. For the reasons set out in Section L above, I have concluded that the defendant was in breach of that obligation. The developed nature of the fire in the hopper at 8:30 pm on 8 June 2005, with flaming popcorn at the top and smouldering popcorn at the bottom, was the sort of fire which the CO2 suppression system should have been designed to extinguish. The principal design failure concerned the type of fire assumed in the design and the type and location of the sensor in the hopper, which was slow to react and out of the way of the likely concentration of flame.
624. For the reasons set out in Section K above, I have concluded that the fire spread as a result of the attempts to stamp out the burning popcorn on the factory floor (Lygate 2). It appears that there were at least 3 sleeves worth of popcorn discharged, much of which may have been smouldering or burning, and the act of stamping out that quantity of material inevitably caused a scattering of burning material across a wide area. That was a more plausible theory as to fire spread than any of the others, and I consider it, on the balance of probabilities, to be the probable cause of the spread of the fire. Indeed, I consider that both of the other two theories (smouldering popcorn bursting into flames in a bagged sleeve or the fire getting back into the elevator) to be wholly unlikely, for the reasons given.
625. For the reasons set out in Section M above, I consider that the defendant's breach of contract caused the fire in the 'oil pop' production area. The defendant is therefore liable for the cost of rebuilding the 'oil pop' production area. Although the claimants' defaults in respect of the absence of segregation and sprinklers were serious and had a significant causal effect, they did not break the chain of causation when the fire moved on to destroy the rest of the NMU. They were a concurrent cause of the losses beyond the 'oil pop' area, but they did not become the single true cause of those losses.
626. For the reasons set out in Section N above, I find that the claimants were contributorily negligent in failing to segregate the 'oil pop' production area and/or failing to put in sprinklers. The correct percentage reduction for those allegations of contributory negligence, to reflect the seriousness of the claimants' default and their causal significance, is 75%. The claimants were also negligent in failing to ensure that the operatives were fully trained in the manual activation of the CO2 system. Although the 75% figure is technically subject to the argument identified at paragraph 621 above (that the reduction should only affect Monkhill, not Trebor Bassett), I consider that, for the reasons given there, the argument is most unlikely to succeed.
627. I would ask counsel to draw up an order accordingly. I will deal separately with all issues as to costs, although it may be that those issues will have to await any subsequent trial on quantum.

Note 1   For the purposes of all but Section O of this Judgment, I do not generally make any distinction between the two claimant companies, and refer to them collectively as “the claimants”.    [Back]

Note 2   I derived little assistance on this important topic from the Experts’ Joint Statement, Item 12. Mr Jowett thought that the Packaging Hall wall may have been constructed from similar panels as those enclosing Production Area A, but Mr Jackman did not agree and there was no evidence to support that supposition. Mr Jackman was of the view (Item 19) that the enclosure around the ‘oil pop’ area would not have provided a 1 hour fire resistance; Mr Jowett’s view was equivocal.    [Back]

Note 3   Item 11 of the Experts’ Joint Statement of 31.3.11    [Back]

Note 4   A drawing dated 15th March 1999 was included in the court bundles, and there was a suggestion by the claimants that this formed part of an amended Fire Certificate. There was, however, no evidence of that, and the drawing itself contained no reference to it. Likewise, the drawing contained no references to fire exits, extinguishers and the other sort of information one would expect to see in a Fire Certificate. Accordingly, in the absence of any other evidence, I do not accept that this unmarked and unstamped drawing formed any part of the Fire Certificate.     [Back]

Note 5   It might have been fibre glass, but it is not possible to reach a conclusion, even on the balance of probabilities.    [Back]

Note 6   This was a rare acknowledgment by Mr Hamilton of the risk to the combustible materials in the Packaging Hall posed by the naked flames involved in the ‘oil pop’ process, which was not, in the event, subsequently considered in any detail.    [Back]

Note 7   If ‘significance’ equated to cost, then the defendant’s work here, which cost £9,000, was very insignificant, a proposition which, of course, runs counter to the whole basis of the claimants’ claim.    [Back]

Note 8   On the issue of holding times, it is worth noting that Mr Flavell suggested that the British Standard envisaged flooding a room with CO2, and then holding it, rather than enclosed machinery.     [Back]

Note 9   In other words, the CO2 system on the C Line that allegedly failed cost the claimants £4,504.50 to install, yet gives rise to a claim for around £110 million.    [Back]

Note 10   See, for example, Mr Caldicott’s concern about open flames, which he had never seen before at any of the claimants’ production facilities, noted at paragraph 56 above.    [Back]

Note 11   The number was 107378    [Back]

Note 12   “In a contract for the supply of a service where the supplier is acting in the course of a business, there is an implied term that the supplier will carry out the service with reasonable care and skill.”    [Back]

Note 13   At the time of the closing submissions, there was a dispute about whether this point had been properly pleaded. However, I allowed both sides to make amendments following the close of the case, in order to deal with this, and other, pleadings points which had been taken. At the hearing on 5th July 2011, both sides confirmed that they were content that all issues – including this one – had been properly raised, and no further pleading issue now arises.    [Back]

Note 14   That was also consistent with Mr Hamilton’s understanding that the heat detectors could not pick up every fire (paragraphs 236-237 above) and everyone’s awareness that the only way in which every fire could be dealt with was by the installation of sprinklers (paragraphs 70-72 and 76 above).    [Back]

Note 15   See paragraph 342 below    [Back]

Note 16   Taken from an email of that date from Mr Hamilton to Ms Irving.    [Back]

Note 17   See paragraphs 229-244 above.    [Back]

Note 18   Hazard and Operability Study    [Back]

Note 19   See paragraphs 56 and 74 above.    [Back]

Note 20   See paragraph 132 above    [Back]

Note 21   Although the practice of ordering joint statements before the exchange of reports is not universal, it is common practice in the TCC, because it usually provides the best way of ensuring that the experts focus their reports on the real issues between the parties.    [Back]

Note 22   As I said to the parties during closing submissions, the fact that the best evidence about a potentially important element of the expert evidence was reduced to a catchphrase from The Fast Show provides an eloquent testament to the often shambolic nature of the expert evidence in this case.    [Back]

Note 23   Set out in full at paragraphs 498-503 below.    [Back]

Note 24   He did at one point describe Lygate 2 as ‘somewhat improbable’, but that was simply a reflection of the fact that, as all the experts agreed, the spread of fire was not a straightforward issue in this case, and each of the three theories had their difficulties.    [Back]

Note 25   The claimants’ closing submissions suggest at paragraph 229 that Dr Lygate’s two theories are simple variations of each other. In truth they are very different, for the reasons explained in this part of the Judgment.    [Back]

Note 26   As I have said, in my view, fire spread was always critical: it is in any large scale fire.    [Back]

Note 27   The importance of these passages of evidence was apparent at the time. It was unsurprising that they featured so heavily in the claimants’ closing submissions (paragraph 133, pages 39 to 41).    [Back]

Note 28   The remainder of these passages in the evidence of Mr Jackman come from day 13    [Back]

Note 29   On this point, I agree with paragraph 207 of the claimants’ closing submissions, that no thought was given by the defendant to the choice of sensor; it was simply replacing the sensors it thought were in use in Leeds.    [Back]

Note 30   It was at one stage rather faintly argued that the sensor could not have been located anywhere else. That was manifestly not the case. And even if it had been then, as all the experts agreed, the defendant should have advised the claimants that the sensor could not be located in a place that made the system effective.    [Back]

Note 31   The deficiencies in the expert evidence as to design were, in my view, a direct result of the failures detailed in Section J above.    [Back]

Note 32   See Section G5 above    [Back]

Note 33   See for example paragraph 88 above    [Back]

Note 34   As noted, it had already burnt down one factory, and caused numerous other fires.    [Back]

Note 35   Paragraphs 63-78 above    [Back]

Note 36   Paragraphs 352-368    [Back]

Note 37   It seems clear that this was the estimated value of the complete building, the machinery and any business interruption in 2004.    [Back]