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

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

		St. Dunstan's House 133-137 Fetter Lane London EC4A 1HD
		29/10/2009

His Honour Judge Toulmin :

1. At approximately 5.00 pm on Saturday 9^th July 2005, the Hard Rock Café ("the Café") at 148-150 Old Park Lane, London, W1, on the corner of Old Park Lane and Piccadilly in Mayfair, was open and trading when a fire broke out in the ground floor kitchen. The premises were evacuated without incident. The fire brigade arrived promptly but, despite all its efforts, it was unable to bring the fire under control until 11.35 pm that evening. The last fire appliance did not leave the scene until 8.18 am the following morning.
2. In the intervening time the fire spread throughout the ground floor of the Café causing considerable damage to that part of the building and heat and smoke damage to the remainder of the building.
3. The parties have co-operated effectively in the case management of the litigation. As a result, subject to liability, the damages are agreed at £721,984.86. Interest of £138,000 has also been agreed, making a total of £859,984.86.
4. The claim is brought by Amsprop Ltd ("Amsprop") as the freeholders of the property. The Defendant, I.T.W. Ltd trading as Hobart UK ("Hobart"), is a firm of kitchen service engineers which was called out from time to time to the Café and attended the Café on the 7^th July 2005 to service and repair what is described as the cheese-melter or Salamander grill (the Salamander) in the kitchen on the ground floor of the Café. The fire experts are agreed that the fire started as a result of the ignition of grease residues in the hood/duct above the flue vent of the Salamander.
5. The allegations of negligence are set out in the Particulars of Claim. The substance of the allegation is that Mr Parkinson, the Defendant's engineer, attended the premises on the 6^th and 7^th July 2005. It is alleged that according to his own note, the extractor was pulling gas off one of the burners of the Salamander. It is alleged that Mr Parkinson permitted the staff at the Café to continue to use the Salamander when it was in a potentially dangerous condition and unsafe to do so. Amsprop contend that Mr Parkinson should have given instructions that the Salamander was not to be used. They contend that had Mr Parkinson given such instructions the fire would not have occurred.
6. The allegation is that the source of ignition of the fire was that a flame from the burners on the Salamander was drawn into the duct above caused by the malfunctioning of the grill which had been identified by Mr Parkinson. It caused the deposits in the duct to ignite into a major fire.
7. The allegations of negligence in the Particulars of Claim are (paragraph 6) that Hobart owed a duty to take reasonable care not to cause loss and damage to the Claimant and other owners and occupiers of the property, to observe and comply with the Gas Safety (Installation) and Use Regulations 1998 and to observe and comply with the Health and Safety Commission Approved Code of Practice and with the Gas Industry Unsafe Situations Procedure ("the Corgi Guidance").
8. In its Defence (paragraph 13) Hobart put Amsprop to proof as to the cause of the ignition. It claims that the Salamander was not connected to the main extract ducting and that flames from the Salamander could not directly ignite deposits of grease in the duct in the manner alleged by Amsprop since any flame would not come into contact with the side of the extraction hood where the deposits may have collected. Hobart joins issue with Amsprop both over whether it has established the probable cause of the fire and also whether their employee, Mr Parkinson, was negligent.
9. In response, on the issue of causation, Amsprop contend that the fire experts are agreed that there are only two possible causes of the fire. In paragraph 9 of their first joint statement dated 18^th June 2009, the experts, Dr Foster for the Claimant and Mr Christie for the Defendant, agreed:

"that there were only two possible causes of the fire in this instance:

(a) a flare up from food on the grill and

(b) flame from a burner being drawn into the duct"

10. Amsprop invites the Court to consider the two alternatives and to conclude that (b) is the more likely cause and therefore, on the balance of probabilities, the cause of the fire. The appliance in question, has two burners and is used for toasting bread and melting cheese, on nachos and potato skins.
11. Proceedings were not commenced until 22^nd May 2008. The witness statements are dated June 2009. However, Dr Foster, in the course of investigating the fire, took some statements a few days after the fire. Ms Jillian Raw of Barlow Lyde & Gilbert, solicitors, also took some written statements on the 17^th June 2006. Mr Parkinson was not aware of the fire until almost two years after it had occurred.
12. In the event, Mr Brompton, Mr Ogunnaike, Mr Dodsworth and Mr Douglas were called to give factual evidence by Amsprop. Mr Parkinson and Mr Grundon gave factual evidence for Hobart.
13. Dr Foster was due to give factual evidence in addition to his expert evidence but in the event was not called to do so and paragraph 4(b) of the Reply was deleted.
14. Paragraph 4(b) of the Reply alleged:

"(b) Dr Foster, the expert retained by the Claimant who inspected the Salamander grill and the extraction ducting a couple of days after the fire, has the same recollection, namely that the hood was connected to the top of the appliance."

15. "The same recollection" referred to paragraph 4(a) of the Reply:

"(a) Kim Brompton, the UK Facilities Manager of the Hard Rock Café (HRC) at all material times, has confirmed that to the best of his recollection the extractor duct work was "integral" to the top of the Salamander grill."

16. In cross-examination Mr Brompton accepted that he did not have any clear recollection of the configuration of the vent/hood. This is important because if there was a substantial gap admitting air then this materially affects the degree of likelihood of the fire being caused in the manner contended for by the Claimant.
17. Otherwise Mr Brompton was not in London at the time of the fire but travelled to London after the fire and from 11^th July 2005 assisted Dr Foster in his investigations. As part of his responsibility, in October 2005 Mr Brompton searched for, but was unable to find, the Maintenance Diary and Hobart log sheets. These important documents are missing. An employee who had custody of them immediately after the fire told Mr Brompton that they had been handed over either to Hard Rock's insurance solicitors or to Dr Foster. The result is that this crucial evidence of what happened immediately before the fire and on the day of the fire is not available to the Court.
18. Mr Ogunnaike is still employed at the Hard Rock Café. He was working at the Salamander at the time of the fire. He was, understandably, very nervous and defensive in answering any question which might imply the possibility of any criticism in the way in which he carried out his duties. I take this into account in considering his evidence.
19. Mr Dodsworth worked at the Hard Rock Café until 2007. At the time of the fire he was the kitchen Assistant Manager working next to Mr Ogunnaike about an arm's length away from him. In his statement, taken on 13^th July 2005, he said that he normally stood with his back to the grill facing the restaurant. In the same statement he said that he did not see any flare up that produced flames that extended to the filters. He did not see any flare up in the Salamander.
20. Mr Douglas was employed by the Hard Rock Café as the only full-time maintenance engineer. He was not in the Café at the time of the fire because he was off sick. He explained the process of switching on the Salamander, a task to be undertaken each day by the engineers rather than by the kitchen staff. He also gave evidence of earlier problems relating to the Salamander.
21. Mr Parkinson is the engineer from Hobart who was called out to the Salamander on 6^th July 2005 immediately before the fire. Aged close to 60, he has worked on servicing commercial kitchen equipment since 1975. Although he has no formal qualifications, he has had extensive training on the maintenance of a number of appliances including, on 10^th November 2001, the Wolf gas product Salamander – char broiler-cheese maker-solid top griddle fryer. He was an impressive witness clearly on top of his job. I bear in mind, of course, that the most competent can make mistakes for which, if they are unlucky enough that the matter comes to court, they must be held accountable.
22. Mr Grundon is currently the Assistant General Manager of the Hard Rock Café in Amsterdam. In July 2005 he had been working as an Operations Manager at the Hard Rock Café in London for about 16 months. He gave evidence that he was on duty on the day of the fire and was responsible for the evacuation of the Café after the fire had started. He was also present in the ground floor kitchen talking to Mr Dodsworth immediately before the fire started. He was, in general, a reliable witness.
23. Mr Galvin was the Rank Watch Manager at Acton Fire Station. His statement, which is agreed, explained that the fire brigade received a call at 5.10 pm on 9^th July 2005 and attended at 5.15 pm. He attended personally at 5.45 pm and also assumed overall control of the investigation.
24. Amsprop's solicitors served Civil Evidence Act notices in relation to Mr Konicki and Mr Diaz. Mr Konicki was a dishwasher at the Hard Rock Café at the time of the fire. Mr Diaz was a table cleaner. Their evidence was served by way of Civil Evidence Act statements and forms part of the evidence in the case although in the event neither party placed any significant reliance on it.
25. There is one aspect of the case on which I should comment at the outset. There is no expert evidence in relation to Mr Parkinson's standard of performance on his maintenance visits to the Hard Rock Café on 6^th and 7^th July 2005. The case for the Claimant is put on the basis, either that Mr Parkinson, in effect, acknowledged that the appliance was unsafe and he should have advised against its use or that it would be obvious to anyone that Mr Parkinson was negligent in not isolating the Salamander on that day. At paragraph 52 of the Claimant's closing submission there is a suggestion that I should apply the test of "the risks which ought to have been obvious to any reasonable competent engineer". I have to treat this approach with the greatest caution since I have no expert witness on the standard to be applied to a reasonably competent engineer undertaking this work.

The experts' joint statement and the BRE investigation

26. The experts are to be commended on the constructive and conscientious way in which they have approached their task as experts. They have been able to reach a level of agreement which has meant that they did not need to give oral expert evidence thus saving significant time and cost. I should add that the two joint statements which they prepared are very carefully and professionally written with appropriate reservations, i.e. conclusions depending in some cases, very properly, on my findings on the evidence.
27. The experts carried out tests on the burners in the Salamander on 19^th February 2009 at BRE Global ("the BRE test"). They agreed a report dated 25^th February 2009 and also agreed a commentary on the test. The video of the test was shown to me at the hearing. The experts also agreed the two joint statements on 18^th June 2009 and 18^th July 2009.
28. The experts agreed that the fire originated in the extract system associated with the Salamander as a result of the ignition of grease residues in the hood/duct above the flue vent of the Salamander.
29. Unfortunately Dr Foster did not take a photograph that showed the inter-relationship of the bottom edge of the hood/duct and the appliance flue vent but, on analysis of the photograph, the experts agreed in the first joint statement that a gap may have been present between the bottom of the hood and the top of the grill. On Dr Foster's analysis in the first joint statement, the experts are agreed that this cannot have extended for more than 45 mm from the top surface of the grill. The experts concluded that the front edge of the hood/duct is likely to have been in contact with and below the top of the front edge of the appliance flue, although there may have been a gap between the bottom edge of the hood/duct and the top of the grill.
30. In the second joint statement this view is modified. The experts agree that the depth of the open area at the rear of the grill was not 45 mm but 75 mm or about 3 inches. There are two segregated sections of the grill vent. The rear section communicates with the grill compartment in which the cooking took place. The source of the ignition could only have emerged from the rear section of the grill vent.
31. The experts have been unable to draw any conclusion as to whether or not the hood/duct was drawing from the flue vent alone, from the flue plus the centre airspace at the rear of the grill or from the flue vent and a portion of the airspace at the rear of the grill.
32. Importantly, the experts agree that if the duct was drawing from the flue plus the centre airspace at the rear of the grill, the test indicates that the burner flame is most unlikely to have been drawn into the hood/duct unless the velocity of air in the duct was very high – in excess of the maximum velocity that would be provided by the test fans and in excess of what the experts might expect to find in a conventional fume extraction system.
33. The experts also agree that if there was a gap between the bottom of the hood/duct and the top surface of the grill or a gap between the back of the grill vent and the back edge of the hood/duct then the likelihood of the flame from the burner being drawn into the duct decreases as the size of the gap increases.
34. The experts also agree that the BRE tests show that under certain duct/hood of the flue configurations, as the ventilation rate increases a point can be reached at which burner flame lift off can occur.
35. The experts also comment helpfully and appropriately on the evidence. If I accept that the e-mail sent by Mr Brompton to Mr Inwood on 26^th November 2005 is factually correct, this would be consistent with the ventilation conditions being sufficiently high to cause the burner flame to be drawn into the hood/duct. The conditions had not changed since Mr Parkinson's visit.
36. On the other hand, the experts agree that Mr Parkinson's evidence is consistent with the burners being left in a safe condition. The experts say that "rolling on the burner" looked like waves and "moving visibly" indicate a very minor disturbance to the flame.
37. Paragraph 14 of the joint statement explains:

"The BRE test showed that such flame behaviour existed from the moment when the increased ventilation rate caused a blue flame to establish beneath the burner plaque extending through the range of velocities up to the situation close to the flame being drawn into the hood/duct. However, CCC (Mr Christie) and CDF (Dr Foster) also note that Mr Parkinson says that the movement of the flame was "not extreme". CCC and CDF agree that, on the basis of the BRE test, if the flame movement was not "extreme" then flame lift off would not occur as a result of any changes they might expect to occur in the ventilation system."

38. The experts conclude unequivocally that if I find that Mr Parkinson's description of the condition of the burner at the time when he completed his work, equated to minimal disruption to the burner flame, there is no available evidence to explain why an ostensibly correctly operating burner flame deteriorated. Cleaning would have been a potential cause but it is to be noted that the most recent cleaning had occurred a few days before Mr Parkinson's visit. The experts agree that, in these circumstances, it is unlikely that flames from the burner entered the duct to cause the fire.
39. The second joint statement considered the witness statements of Mr Grundon, Mr Ogunnaike, Mr Douglas and Mr Dodsworth.
40. In relation to Mr Grundon's evidence, they comment that Mr Grundon observed that the grill was "working properly". His correct recognition in two photographs in the BRE test between one in which the burners were burning properly and one in which they were not, indicates that he understands what a properly operating burner should look like.
41. Mr Douglas describes observing an event in which he saw "a very large flame shooting out of the pilot across the burner". The experts agree that this description cannot be technically correct as they could envisage no circumstances in which a large flame might be produced from the pilot. If such an event occurred, it must have been produced by the ignition of gas from the main burner.
42. The experts note the conflict between Mr Grundon's evidence that during peak times the Salamander was turned up on full power so that it was hot and ready to cook at all times and Mr Ogunnaike's evidence that when it was not in use the Salamander was left on pilot. (This is also related to an issue as to how full the Café was at 5.00 pm on that Saturday afternoon).
43. Mr Ogunnaike asserts that: "The type of food cooking under the Salamander at any one time would not have resulted in a flare up on the grill". The experts, having read this, do not resile from the alternative possibility that the fire could have been caused by a flare up of food left on the grill. In relation to Mr Dodsworth's evidence, the experts emphasise that the food stuffs being cooked under the Salamander are capable of catching fire if heated for sufficiently long under the grill, although there would be a lower tendency to produce a transient flare up than with greasy food such as burgers.
44. Clearly if I accept Mr Ogunnaike's evidence that there was no food on the grill immediately before the start of the fire, the possibility of the flare up being caused by the alternative theory is eliminated.
45. The experts also agree that whichever of the alternative causes was responsible for the fire, there would have been a delay between ignition and the discovery of the fire. They also agree that the extent of the delay would be influenced by the extent and pattern of the grease contamination within the system. They agree that a delay of around five minutes would not be unreasonable.
46. Finally, the experts considered Mr Brompton's e-mail dated 26^th October 2005 on the basis that the evidence which had been reported to him and which is set out in the e-mail was correct. They concluded that the effect which "David" said he observed, namely that the flame from the Salamander was drawn into the duct/hood when the duct was connected, would indicate that the duct was drawing from the grill vent or the grill vent and a relatively small opening between the back of the vent and the hood/duct but, most probably, not from the centre airspace at the back of the appliance.
47. The experts express no view as to whether the term "connected" and "disconnected" are apt to describe the following processes but the effect said to have been observed could have been caused by:

"(a) unscrewing the top cover of the Salamander and lifting it a little to prevent the extract system from creating a negative pressure in the top burner compartment;

(b) unscrewing and removing the front cover of the top burner compartment to prevent the extract system from creating a negative pressure in the top burning compartment or (which the Claimant concedes is the least likely)

(c) unscrewing the inspection cover of the Salamander duct beneath the main extract canopy hood which would enable a path of lower resistance through which the extracted air could be drawn and prevent the extract system from creating a negative pressure in the top burner compartment".

The Facts

48. The Salamander/cheese melter was installed at the Café in about 2001. The Hobart call out log shows that there were problems from time to time. There were problems in 2001. In 2002 there is a note that Hobart was called out because the burners were not lighting. New burners were fitted.
49. On 20^th May 2003 there was a complaint that the pilot light kept going out. A new valve was fitted and the Salamander was tested "ok".
50. On 17^th September 2003 Hobart again attended. The complaint was that the Salamander was not working. The record states that the equipment was checked and stripped down. The valve was removed and a new display valve was installed.
51. Of much greater relevance, Hobart was called out on 8^th April 2005. The fault reported was "Leaking gas urgent". Hobart's service sheet on 8^th April 2005 noted that "Burner no good. Ordered burner to be sent to site for standby tech".
52. On 13^th April 2005 the service sheet recorded "New burner fitted and tested". Hobart's report was signed by Mr Douglas. The technical report noted "Complaint unit cheese melter not working. Check, strip down, remove valve and thermo – fit new LD spray valve. Check and tested ok".
53. On 5^th June 2005 the Salamander is recorded as "Making a noise". Hobart was called out. Mr Parkinson attended. The complaint was that the burner was "popping". Hobart found that the left hand burner was not burning correctly. It was removed and a new burner was fitted from the service support vehicle.
54. Records of the cleaning of the ducts by Duct Clean Services have been disclosed. On 3^rd July 2005 there is a Report that deep cleaning of the kitchen extract systems took place. It would appear to have been done every two months. The Report refers to very heavy contamination in some areas and to the "inaccessible duct which opens out before the filters". The Report also refers to years of neglect in some areas. There is no evidence that this Report was seen by Mr Parkinson.
55. Mr Douglas, in evidence, thought that there had not been cleaning in the duct for a long time before the fire.
56. This area of the evidence was not investigated in great detail at the trial. The agreed conclusion of the experts is that there were deposits in the ducts which caught fire. Mr Parkinson is a tall man and he was able to give evidence of the configuration of the hood of the Salamander and the vent. He was sure that a gap existed at the front face. In oral evidence he clearly marked a gap between the appliance flue and the extractor hood.
57. The crucial report relates to Hobart's visits on 6^th and 7^th July 2005. The job sheet of Mr Parkinson records visits from 10.30 to 11.30 am on 6^th July 2005 and from 7.30 to 9.45 am on 7^th July 2005. Mr Parkinson notes: "Complaint cheese melter not lighting. Check cleaning. Fit and tested left working ok. I returned to site remove faulty burner and fit new. Check burner working ok, box unit up. Check found extractor pulling gas off burner unit. Need unit check for air flow. Left working, explain to engineer."
58. Mr Parkinson was, as one would expect, cross-examined at length by Mr Moxon-Browne QC. In the report Mr Parkinson records the times when he was on site. He said in evidence that he "Returned to site on the following day".
59. He explained in oral evidence that:

"On the first day there were all the checks and stuff and the burner was not burning correctly so I ordered a new burner which was delivered the following day".

60. He said that the rolling movement which he observed meant that the burner was working to 95% of capacity and was not dangerous. He said that the call out complaint was no more than an indication of the problem and that it was the responsibility of the engineer to find the fault.
61. He was asked searching questions about the smoke test and the note that he thought that the extractor was pulling too much air which was causing the burner flame to roll slightly. Mr Parkinson explained that he told the engineer that there was slight movement on the burner:

"It was not pulling the gas off the burner because, if it was so, the pilot would automatically go out. It has a fail safe device on it. So if you are pulling that much gas off the burner also you are pulling the pilot flame off as well, that would go out, the main burner would go out."

62. Mr Parkinson agreed that he had the power to say that the Salamander must not be used but said that he did not see any danger regarding the machine.
63. He also agreed that the Salamander was not working to his satisfaction. He explained:

"I go through a machine and I have done everything I can to make sure this machine is working efficiently and I find something out of the ordinary, not in my control, that belongs to the Hard Rock engineer and the maintenance contracting, then that is the people I talk to".

64. On 26^th October 2005 Mr Brompton sent the e-mail to which I have already referred. It was an e-mail containing what he regarded as significant evidence. He quoted the note from the Maintenance Diary of 7^th July 2005 (diary since lost).
65. The note is written by David Havlicek and says:

"Technician from Hobart is here and working on the Salamander, the reason right half of Salamander is not working as well as left one is very strong air flowing which causes that flame is literally pulled in. He said he had fixed it and changed gas pressure a bit."

66. The suggestion is that this in Mr Havlicek's handwriting. This is somewhat surprising since Mr Roger Pierre was the Hard Rock Café engineer who countersigned Mr Parkinson's work sheet. It is not entirely clear what Mr Havlicek's responsibilities were but it would appear that his main responsibility was clearing tables.
67. The email went on to say that Mr Havlicek remembered the incident and the Hobart engineer demonstrated how the flame from the Salamander was drawn into the duct when connected and when disconnected it stayed within the grill area. The experts in their second joint statement express reservations about what was meant by connected and disconnected. The duct work was a permanent installation which was not capable of being connected and disconnected. They speculate about what this might have meant. Nevertheless, if I accept Mr Havlicek's evidence, the experts are agreed that the fire would be capable of starting in the way which it is contended for by the Claimant.
68. Mr Parkinson was cross-examined about the email and said that he adjusted the gas pressure slightly which was normal procedure for ensuring that there was not a fault on the machine.

"If the burner is burning and you see that it is not burning correctly you make sure that the actual appliance is not over-gassed, so that is why you do a gas check."

69. Mr Parkinson was also asked about Mr Havlicek's recollection and he replied "No, I showed him the movement of the burner flame but I did not show him a flame going into the duct work". I note that Mr Roger Pierre signed off Mr Parkinson's work sheet confirming that the Salamander was left working to his complete satisfaction.
70. Mr Havlicek did not give evidence and there is no statement from him although I understand that he was still working at the Hard Rock Café when the importance of his evidence had become clear to the Claimant. His status and qualifications are by no means clear. Obviously, since he did not give oral evidence he could not be cross-examined about what he had reported to Mr Brompton over three months after the fire. There is a suggestion that he is someone employed to clean tables and was standing in for Mr Douglas who was off sick, but this does not explain why Mr Pierre, who I understand was an engineer working part-time, signed off Mr Parkinson's report.
71. I should refer finally to a summary of the incident on the 9^th May 2007, nearly two years after the event set out by Mr Davies, Mr Parkinson's supervisor. Mr Davies says that his evaluation of Mr Parkinson's service visit indicates that the burner which was not functioning was replaced.

"There was then evidence of flame left from the main burner due to the excessive extraction rate of the system. It was also noted by the technician that the extract system required cleaning due to the level of grease he had witnessed. The high rate of extraction coupled with the need for cleaning was reported to the site engineer who duly advised that this would be addressed."

This note does not deal directly with the question as to whether or not Mr Parkinson thought that the Salamander was in a dangerous or potentially dangerous condition such that he should have ordered that the Salamander be shut down. It does set out for the words on Mr Parkinson's customer service sheet. He was emphasising to the site engineer that the extraction system needed cleaning. He was the only witness who was tall enough to look at the extraction system and see the deposits which had accumulated.

72. I conclude that Mr Parkinson was an honest and truthful witness whose evidence I can accept. It is not possible on a fair reading of his evidence to conclude that he was in any way admitting that he should have closed the Salamander down or that when he left it, it was in any actual or potentially dangerous condition. Further, I have no evidence before me to conclude that any reasonably competent engineer, faced with the same situation on the 7^th July 2005, would have closed down the Salamander.
73. I find that in Mr Parkinson's note on the work sheet he meant in relation to the burner:

"pulling gas off burner by a fraction of an inch sufficient to cause a visible blue shimmering below the red radiance, i.e. not quite all the gas was being burned in the correct place at the surface of the ceramic plaque where it produced radiant heat, but a small proportion of it was burning just beyond the surface."

74. I find that, as he said truthfully in his oral evidence, if he had thought that there was a risk from the Salamander he would have isolated it.
75. I also find that the burners were working to 95% efficiency and the state in which Mr Parkinson left them was not one of danger or potential danger of fire. I note that there was no complaint that the Salamander was working inefficiently between the time of Mr Parkinson's visit and the time of the fire.
76. I am satisfied that Mr Parkinson had an overriding duty, if he thought that there was a risk that the Salamander may be unsafe, to close it down. Mr Parkinson was well aware of this obligation. In the circumstances, he reasonably concluded that there was a fault which needed to be corrected by the Hard Rock Café maintenance engineers which did not involve any risk to the safety of the Salamander when it was being used. In reaching this conclusion I take into account Mr Davies' note of the 9^th May 2007.
77. In the period between Mr Parkinson's visit on the 7^th July 2005 and the fire at around 5pm on the 9^th July 2005, the Salamander appears to have been working effectively. There was no complaint to Hobart that Mr Parkinson had not dealt with the problem.
78. The sequence of events on the 9^th July 2005 is that those in the kitchen heard "a whooshing" noise from the Salamander followed later by the start of the fire. Mr Dodsworth and Mr Grundon gave evidence that they saw the burners on the grill in their usual radiant condition. The "whooshing" did not cause the cooking to be stopped. Some minutes later the fire started.
79. There are a number of factual issues in relation to the fire. These impinge on the operation of the Salamander in the ground floor kitchen. There is an issue as to whether or not the Salamander remained lighted at all times ready to receive cheese and nachos or whether this was a less busy time when it was on pilot only. This issue is related to the question of how busy the kitchen was.
80. In Mr Galvin's incident report on the 10^th July 2005 he notes that "fire is directing from ground floor to roof. 50 customers and 10 staff evacuated. This was a protracted incident".
81. Mr Dodsworth said in oral evidence that "the café was not full up". Mr Grundon, who gave oral evidence, said that this was the tourist season and it was very busy. He said that there would have been a queue of people waiting outside. He said in his witness statement that he was clear how full the restaurant was because he managed the evacuation of the restaurant once the fire was discovered and he closed the restaurant. In cross-examination he was asked about Mr Galvin's figures. He maintained that the restaurant was full. He noted that Mr Galvin must have obtained the figures from another officer (not called to give evidence) since he had to come from Acton and did not arrive on the scene until 5.45pm, by which time the evacuation had been largely completed.
82. I was told that the ground floor seating area, served by the ground floor kitchen, would be likely to fill up before the basement area. There was, indeed, a suggestion that the ground floor would fill up first before the basement was open. I have no direct evidence of this and I can go no further than to assume that, in general, customers would prefer to be on the ground floor rather than in the basement.
83. After considering all the evidence on this issue I conclude on the balance of probabilities that at 5.00 pm on the day of the fire the upstairs was at least close to being full. This would mean that there were frequent orders requiring the use of the Salamander.
84. In the kitchen at the time of the fire were Mr Ogunniake and Mr Dodsworth. On the day of the fire Mr Ogunniake had started work at 9.00 am. He had a 15 minute break at 1.45 pm. He had been working continuously since then. Mr Grundon gave evidence that 5pm was at the end of the shift. Mr Ogunniake must have had a long day.
85. Mr Ogunniake said that the procedure was that, when an order required the use of the Salamander for melting cheese on nachos, melting cheese on potato skins, toasting bread or what he described as "some kiddies stuff" he would be given a signal, about 2 minutes before the other items were ready, to start cooking the items under the Salamander. He said that the Salamander had two burners and, depending on the number of orders, both could be in use at the same time. He explained that first thing in the morning the burners were lit by the engineers.
86. Mr Ogunniake said that the Salamander was on pilot when not in use and that, when he wanted to use it, he adjusted the Salamander knob from pilot mode to high. It is common sense that if the Salamander was turned to pilot mode the burners would cool down and this would increase the length of cooking time over the cooking time when the food was placed on the burners at maximum heat.
87. The experts have agreed that it would take 10-15 minutes to go from pilot heat to maximum heat and 2-3 minutes for the burners to get hot. This would suggest that, in order to ensure that all food was ready together, normally the Salamander would remain with the burners on high. If it was on pilot mode it would be difficult to ensure that all the food would be ready to serve at the same time.
88. Mr Dodsworth said in oral evidence that the Salamander would be used on a constant basis and would not be turned down to pilot during the day. I accept his evidence on this.
89. The alternative theory that the fire was caused by food cooking on the Salamander is dismissed by the Claimant. Their case is that the food which was put on the Salamander would not flare up in the way that occurred in the fire and further that I should accept Mr Ogunniake's evidence that there was no food on the Salamander cooking at the relevant time.
90. It is clear that there was food under the grill when Dr Foster inspected the kitchen on the 10^th July 2005 very soon after the fire. This was shown clearly in two photographs. Dr Foster did not apparently examine the food or the kitchen bins beside the Salamander.
91. Mr Ogunniake's evidence is surprising. In his earlier statement he said that there was no food on the Salamander immediately before the fire. Later in January 2006 he said that there was no food burning on the Salamander. He resisted the idea that food could catch fire.
92. Finally, he conceded that this was possible in these terms:

"A: When you are working there you have to be next to the Salamander. In that case it cannot catch fire. Once you are working there, there is no way you can leave the section and go somewhere else. Your attention will be on the Salamander.

Q: But if somebody's attention was not on the Salamander they (the food) could catch fire, could they not?

A: That is somebody, but not me sir."

93. Mr Ogunniake agreed in cross-examination that the position was that he could not accept that there could not be a flare up of food left to cook on the Salamander because whilst he was cooking on the Salamander he had to watch it all the time and did so.
94. In the course of the following evidence Mr Bartlett QC put to Mr Ogunniake the process by which the food was prepared. It was suggested to Mr Ogunniake that it was possible for the person operating the Salamander to be distracted by other tasks which he had to perform. Mr Ogunniake replied that you had to watch the grill. Equally he evaded the question of whether or not he had seen a flare up because someone had left food on the grill for too long by saying that it had happened "probably in my absence but not in my presence".
95. Mr Dodsworth agreed that work in the kitchen was "pretty relentless work". He agreed that burning fat could flare up.
96. Mr Grundon said in his witness statement that the chefs were under enormous pressure to get orders out as quickly as possible. This is common sense in a busy café where throughput of customers is important. Mr Grundon also said in oral evidence that if the Salamander chef had several orders to prepare at the same time, he would work on the orders concurrently. He said that if the chefs were preparing nachos "they would prepare that, put it under the grill but they would not be watching it for the entire time it was under the grill." In accepting this evidence, I reject Mr Ogunniake's evidence on these matters.
97. I find the following as facts:

i) I accept the evidence of the experts that food left cooking on the Salamander was capable of causing the fire. No test was carried out to explain how long the cheese would take to catch fire but the experts' conclusion based on the evidence that has been given is that the fire could have occurred in this way.

ii) The café was busy in the minutes immediately before the fire, although not necessarily completely full. In these circumstances, the burners were not on pilot light but were fired up ready for cooking.

iii) The Salamander chef would not be watching the grill for the entire time that food was cooking if he had a number of orders to process.

iv) Mr Ogunniake was at the end of a very long and busy day. At the time of the fire he had not had a break for three hours.

98. Considering all Mr Ogunniake's evidence I reject his assertions (a) that he watched the Salamander all the time the food was cooking, (b) the Salamander was left on pilot when the food was not cooking, (c) that he was incapable of inattention to the Salamander, particularly at the end of a long and tiring shift and (d) that no food was cooking on the Salamander immediately before the fire.
99. In these circumstances, I find that of the two possible causes of the fire the possibility of it being caused by a flare up from food on the grill is rather more likely than the alternative explanation that it was caused by a flame from a burner being drawn into the duct. I accept the Defendant's contention that if Mr Ogunniake had been paying full attention to the grill he would have seen the flame being drawn into the duct if that had happened.

The Law

100. Amsprop contends that where there are only two possible causes, as here (on the agreed finding of the experts), a cause which had seemed improbable may become probable where it is the more probable of the two causes. In that case it may become, on the balance of probabilities, the probable cause.
101. In Kiani v Land Rover Limited [2006] ECWA Civ 880 Waller LJ said at paragraph 30:

"I do not myself think it is false logic to reason where only two possibilities are under consideration, both of which seem unlikely, if one seems much less likely than the other the less likely can be discounted thus making the first likely to happen on the balance on probabilities."

102. This provides a helpful gloss on Rhesa Shipping v Edmonds (The Popi N) [1985] 1 WLR 948 where the House of Lords concluded that if a Claimant could not prove a probable cause, the claim must fail.
103. These and other authorities were considered by Akenhead J in Fosse Motor Engineers v Conde Nast Distributors Limited [2008] EWHC 2037 (TCC). I respectfully agree with his analysis. If I conclude that if out of only two possible causes of the fire, the flare up of food on the grill seems much less likely, I should discount that possibility and conclude that the fire was caused by a flame on the burner being drawn into the duct. I would then need to go on to consider the allegation of negligence against Mr Parkinson.

Conclusion

104. In order to succeed the Claimant must prove, on the balance of probabilities, (a) that the fire was caused by a flame from one or both the burners on the Salamander being drawn into the duct causing the grease in the duct to catch fire and (b) that Mr Parkinson was negligent in not ordering the Hard Rock Café to shut down the Salamander at the end of his visit on 7^th July 2005 at least until the duct had been cleaned professionally and the deposits of grease had been removed.
105. In relation to the cause of the fire, I cannot, on the evidence, conclude that it was less likely, let alone much less likely, that the fire was caused by a flare up of food on the grill rather than by a flame from one of the burners being drawn into the duct. On the evidence as presented to me, it would be speculation to say which seems the more possible cause. As I have indicated, if I was forced to speculate, I should conclude that the flare up of food on the grill as a result of inattention by Mr Ogunniake was the more likely cause. I emphasise that this is speculation and is not a finding of fact against Mr Ogunniake.
106. If I had concluded that the fire was caused, on the balance of probabilities, by the flame from the burners being drawn into the Salamander, I should have had to consider whether or not Mr Parkinson was negligent in not requiring the Hard Rock Café to shut down the Salamander at the end of his visit. I have accepted Mr Parkinson as a truthful witness. I have also accepted his explanation of his conduct in noting, as I have already found, a slight disturbance to the burners. On the evidence, I conclude that he was not negligent in giving the advice which he gave. I reach this conclusion on the evidence before me. I have, of course, not heard any expert evidence on this issue.
107. For these reasons, I find for the Defendant.