HEMSLEY v LINDSAY

Luke John HEMSLEY (Plaintiff) PARTIES : Graham James LINDSAY and David Francis LINDSAY (Defendants)

D Wheelahan QC with P Woods (Plaintiff) COUNSEL : R Stitt QC with D Wilson

Mitchell Playford & Radburn by their agents Orr & Company (Plaintiff) SOLICITORS : Hunt & Hunt (Defendants)

IN THE SUPREME COURT OF NEW SOUTH WALES COMMON LAW DIVISION

JUDGMENT 1 ADAMS J: Introduction: The plaintiff, Luke John Hemsley (now known as Reikan Sebastian Hemsley), was appallingly injured by burns that he suffered when petrol fumes were ignited by the burner in the gas hot water system in the laundry of his aunt's rented premises. He sues the then owners of the property, Graham and David Lindsay. There is no substantial dispute about the extent and consequences of the plaintiff's injuries. The question of substance in this case is whether the defendants are liable in negligence for them because, as they knew or ought to have known, the plate covering the pilot light cavity in the side of the heater could not be attached. 2 The water heater in question was a conventional device found in many homes. It used a gas burner to heat the water, controlled by a temperature device that switched the burner on when the water cooled to a pre-set temperature. The gas in the burner was ignited by a pilot light. Access to the pilot light and burner was obtained through an opening in the side of the heater near the base. A removable cover protected the pilot light from draughts that might extinguish it and from accidental intrusion, as from a child's inquisitive actions. The plaintiff's case is that the clips on the protective cover were so damaged by corrosion that it would not remain attached to the side of the heater and that, although it leaned against the opening, it only partially covered it. This meant, it was contended, that when the petrol fumes were released into the air of the laundry by the plaintiff's innocent but foolish emptying of a petrol container onto the floor, the fumes were ignited a short, but nevertheless significant, time before this would otherwise have happened had the cover been properly in place. The result was that the plaintiff's father, who had seen what had happened and had attempted to rescue his son, got to him some seconds too late to prevent his being seriously burnt. The plaintiff's case is that, had the cover been correctly in place, contact between the fumes and the heater flame would have been delayed by the few seconds that would have made all the difference. 3 It is, as it seems to me, crucial to appreciate that it is no part of the plaintiff's case that the defective cover plate caused or contributed to the explosion; nor is it alleged that the defendants knew or ought to have known that the inability to attach the cover plate was capable of affecting the speed with which any flammable vapour introduced into the laundry might ignite or that this was a reasonably foreseeable consequence of such non-attachment. On the other hand, the plaintiff does have to prove, as I understand it, that the defendants ought to have foreseen the liberation of petrol or similar flammable fumes in the laundry that might be ignited by the heater flame. 4 The defendants argue that, even if the cover plate were defective as claimed, it was not designed or attached for the purpose of delaying ignition in the circumstances of this (or, indeed, any other) case and, at all events, its misplacement had no material effect on the time taken for the petrol fumes to ignite. The defendants also contend that, even if the fumes ignited more quickly than otherwise would have occurred, as claimed by the plaintiff, this was an occurrence that was neither reasonably foreseeable nor within any general class of foreseeable events that might have been caused by the fact that the cover plate was defective and that rendered the defendants liable for the plaintiff's injuries.

The laundry explodes 5 The following account is largely taken from the evidence of the plaintiff's aunt, Mrs Craft, who I thought was an honest witness whose account was, by and large, reliable. On the day of the accident, Mrs Craft had borrowed a lawn mower, intending to mow the lawn, and had purchased some petrol from the local service station. The fuel was placed in a plastic two-litre bottle of the sort used for milk or soft drink with a plastic screw top which, I infer, was easy to remove. The bottle contained about a litre and a half of conventional two-stroke fuel, a mixture of oil and petrol. Shortly after Mrs Craft returned home from the petrol station she went out to the rear door, intending to fill up the mower and start cutting the grass. She was forestalled by meeting her mother (Mrs Hemsley) at the rear door, who had brought the plaintiff with her. Mrs Craft put the petrol down in the laundry, naturally enough on the floor, although this was to have tragic consequences. 6 The two women went into the kitchen/dining room area and, shortly after, Kevin Hemsley the plaintiff's father (and Mrs Craft's brother) arrived and joined them in the kitchen. Mrs Craft's three children, aged from about eighteen months' to five years' old, were playing with the plaintiff and went into the toilet area, which was between the kitchen/dining room and the laundry. Mrs Craft said that she heard a tap running and issued a general direction to turn the tap off and "hop outside and play". Her daughter replied that she was washing her hands. Mrs Craft, then sitting at the end of the table, heard what she took to be the sound of water coming from a tap once more. She jumped up and looked through the toilet into the laundry where the plaintiff stood with a bottle and realised that the noise she had heard was the petrol pouring out of it. She called out, "Kevin, he has got the petrol." Mrs Craft thought the plaintiff was about two and a half to three metres in front of her with her son, Joshua, between them. The plaintiff was standing in front of the tubs pouring the petrol out of the bottle all over his legs and the floor where it could be seen spreading. In the meantime, Kevin Hemsley raced towards the children, pushing Joshua out of the way with his left hand and grabbing for his son. It was about this time that the fumes exploded. This occurred, I think, either just before or as Mr Hemsley pulled him away. 7 Mr Hemsley's account is much the same but differs in several details. He said that he was sitting at the table and heard what he thought was water running. He said to his sister, "If that's water, the children will have it everywhere". He said that Mrs Craft stood up from her chair and, as he was slowly getting up himself, she said to him something like, "They have got the petrol". He just stood up and "I pushed her and boom and as I stood there, that's how it happened". Mr Hemsley said that, as he moved towards the entrance area of the laundry, he saw the plaintiff in the middle of the laundry, pushed Josh out of the way with his left hand, and grabbed his son. 8 Having regard to the suddenness and speed with which these events occurred and their shocking aftermath, there must be some real doubt about the estimates of time in Mrs Craft's and Mr Hemsley's accounts. It is notorious that estimates of time and the sequence of events are recalled with significant uncertainty in emergencies or emotionally fraught situations. Subject to this problem, I am satisfied that their evidence, in substance, is reliable. However, the problem is very significant because of the relatively short delay in ignition ultimately proposed by the plaintiff's expert evidence in the event that the cover plate was attached. 9 The crucial period is the lapse between the plaintiff commencing to spill the petrol onto the floor and when he was grabbed by his father. Mrs Craft was obviously not listening out for any pouring noise. It is clear that the risk that the children might play with the fuel container was not present in her consciousness, otherwise she would immediately have removed it from the laundry. She was involved in her interaction with the other members of her family though, no doubt, having what I might describe as half an ear out for what the children might be doing. The initial pouring of petrol may well have been a quiet spill. As the plaintiff lifted the container it became louder and when Mrs Craft and Mr Hemsley heard it there must have been a further delay whilst the possible significance of the sounds came to their minds, though both thought at first that it was simply running water and Mrs Craft (on her account) did not realize the truth until, having taken a couple of steps, she saw the plaintiff with the container. There was a delay while she got up and took some steps, then realized what was happening and called out, then whilst Kevin comprehended her concern and then whilst he moved to rescue the plaintiff. I would accept that these delays are short, but their presence is inevitable and their cumulative effect significant. The problem that they pose for the plaintiff is that he must prove more probably than not both that the vapour ignited before it would otherwise have done had the cover plate been properly attached and that he would probably have been rescued during that window of opportunity. In this context, the time that the spill-carried vapour would have taken to reach the heater is probably between 12 and 15 seconds. However, this does not mean that the vapour ignited at this time. When it reached the heater flame is a much more difficult question to answer. I deal with this issue below. 10 The defendants submit that, having regard to the fact that Joshua was nearer to the laundry door (and further away from the heater) than the plaintiff and was himself severely burnt before Mr Hemsley pushed him out of the way, the explosion probably occurred almost as soon as Mrs Craft called out and some seconds before Mr Hemsley intervened. When Mr Hemsley's evidence is taken into account, I think that this submission is well made. 11 What I have called an explosion comprehensively describes two connected, quickly occurring, but distinct events: the first was the combustion of the fuel and air mixture in the laundry; the second was the "flashback" in which the flame ignites the petrol vapour back along the fuel trail and to the plaintiff (because he was soaked in fuel) where the liquid was very rapidly heated by radiation and convection from the flame, replacing the burning vapour with burning volatiles. Joshua was badly burnt although he was pushed out of the way very quickly. His burns were caused merely by combustion of the vapour since, unlike the plaintiff, he had not been splashed with fuel. 12 I accept that the flashback had engulfed the plaintiff before Mr Hemsley got to him. Even if Mr Hemsley had grabbed the plaintiff a second or two earlier and started to move back out of the laundry, it is almost certain that, given the distances and speed of combustion, the plaintiff still would have been very seriously injured by the flashback and its ignition of the volatile fuel on his body. 13 The defendants contend that the evidence does not permit the Court to infer, without inappropriate speculation, that a delay in combustion of the vapour of the order posited even by the plaintiff's expert evidence would have saved the plaintiff from serious burns more or less as extensive as those he, in fact, suffered. As will shortly become obvious, there is a real dispute between the plaintiff's experts and the defendants' as to whether the non-attachment of the cover plate (assuming this to be so) affected the speed with which the fuel ignited. 14 Making a common sense assessment of the evidence, I have concluded that, had Mr Hemsley seized the plaintiff about four or five seconds before the flashback reached his son, it is probable that he would not have been seriously burnt. The evidence does not permit me to say what more probably than not would have been the position had the time frame been less. Nor can I decide how much less seriously the plaintiff would have been burned had he been grabbed in that time but his injuries would probably have been relatively trivial and short-lived. 15 The specification of particular times for each supervening event would give an impression of precision that the evidence simply does not permit. Overall, I am satisfied that about 15 seconds elapsed between the commencement of the spill and when the plaintiff was seized by his father. However, I cannot say more probably than not that this then occurred. I have been unable to reach this level of confidence in respect of a time lapse of less than about 20 seconds. Of course, even these numbers give a somewhat misleading impression of certainty. 16 The scientific controversy in this case may be stated thus: if the cover plate was properly attached to the side of the heater would that have made any difference to the speed of ignition of the petrol fumes? I deal first, however, with the preceding question, whether the cover plate was in fact defective.

The state of the cover plate 17 The aperture or pilot light inlet is shaped roughly like a trapezium with a base of about 20 cm, a top of about 10 cm and inclined sides of about 20.5 cm. The cover is also for all practical purposes a trapezium with a base of 20.5 cm, a top 8 cm wide and sides 27 cm long. A semi-circular "bite" in the top accommodates part of the gas inlet mechanism. Two edges run down the sides of the plate, keeping it proud of the surface of the heater by 4 cm at the top (substantially filled by the gas inlet mechanism) and 2.5 cm at the bottom, where it forms a vent. The vent is about 20 mm from the ground. The cover is attached to the heater by two lugs which protrude from it and fit into corresponding holes near the base. About 16.5 cm from the bottom of the cover and attached to its back is a metallic spring clip which goes over the gas inlet pipe and helps to hold the cover in place. There was a control knob for adjusting the temperature of the water unit. The pilot light is 35 mm from the ground. 18 The heater was raised on dimples and had three air vents let into the base. These vents were about 10 mm off the ground, so that around the bottom perimeter of the heater there is a 10 mm high peripheral area which gave air access to the heater cavity. 19 In 1974 Mr Smith moved into the premises at 48 Barker Street, then occupied by a Mr Barry Cook and his daughter, Glenda. Mr Smith and Ms Cook married about two years later and in due course they left the premises in about mid-1984. Mr Smith said that, when the burner ignited to commence heating, there was a sound described by him as "pouf", I take it much like the sound of gas lighting in a barbeque or on a stove top when there has been a short delay in applying the flame after the gas has been switched on. He said that when that happened the cover would fall off. As I understand Mr Smith's evidence, the cover had become progressively worse until, by the time he left the premises in mid-1984, it had deteriorated to the point that it could not be attached at all. He said that the lugs were worn and that the central clip was missing. Mr Smith said that he simply lent the cover up against the aperture. Mr Smith said that he complained to Mrs Lindsay (the now-deceased mother of the defendants) that this was dangerous because his children might crawl close to it or put their hands into it. He said that his complaints were rejected by Mrs Lindsay since he was not the tenant but merely shared the premises with Mr Cook who was. Mr Smith said that he attempted to raise this matter at least three or four times but was rebuffed. He said that, whilst he was in the premises, the problem was never attended to. 20 Mr Smith was cross-examined by Mr Stitt QC for the defendants as to several features of his evidence and to his statement to an investigator that the cover would fall off the heater from time to time and that it was "always possible to replace" it. It was not put to him that his description of the cover's physical condition was wrong, nor that he had not made any complaints to Mrs Lindsay. The earlier point may not matter all that much, since the design of the cover is such that if it had been properly attached, it could not accidentally fall off and the fact that it did so - as asserted to the investigator - indicated that it was in all likelihood defective. Mr Smith, I should add, did not accept that the account of his conversation with the investigator put to him in cross-examination was accurate and the investigator was not called, nor was there any attempt to tender his report. Of course, over the years Mr Smith's memory may well have varied from the fact but he struck me as an honest and uncalculating witness who believed what he said was the truth. He said that he had discussed this matter with his father-in-law but that as it happened, Mr Cook did not have the heater fixed though Mr Smith believed that he had made ineffectual complaints. 21 In about mid-1984 Mrs Craft moved in to the premises and said that Mrs Lindsay, who acted for the defendants as property manager, occasionally visited the house to collect the rent. Mrs Craft said that, shortly after she moved in, she noticed that the cover was not connected to the heater and that when she tried to do so it kept falling off. Mrs Craft said that the cover plate was rusted and it would not fit in. She lent it up against the hot water system to conceal the flame from her young children. She said that when she asked Mrs Lindsay about it her response was to not to worry because, as I understand the sense of the evidence, the previous tenant was unconcerned about it. Mrs Craft said that she asked Mrs Lindsay to fix the cover because she was worried that her children might burn themselves if they put their hands into the aperture. Mrs Craft said that she was constantly making sure that her children did not go near the heater because of her concerns. 22 The defendants rely on evidence, largely coming from one of the defendants, that Mrs Lindsay, as the matron for many years at the Casino Base Hospital, was a responsible and careful person. Mr Lindsay said that his mother was involved in management of the property, that she often went to the premises and took a keen interest in their maintenance. She herself even mowed the lawns. The photographs of the house appear to show that it was well maintained. It was submitted that it would be highly improbable that a person such as Mrs Lindsay with that interest in the premises would have permitted dangerous equipment to have remained there, especially with young children in the house. Furthermore, Mr Cook, who was Mr Smith's father-in-law and had been the science master at Casino High School, was also a mature and responsible person. If he had thought the heater was dangerous or defective it seems most unlikely, it was submitted, that he would have permitted this state of affairs to continue, especially when his young grandchildren were living in the premises with him. Furthermore, he refused the defendants' offers to replace the gas heater with a new electric heater. 23 Mr Lindsay also gave evidence that, when the gas supply changed to bottled gas from the town grid, the heater was inspected by the gas fitter. It was submitted that it was highly improbable that the gas fitter would have changed the gas over if he thought that the hot water system was defective or presented a danger to persons occupying the premises. 24 Mrs Lindsay also employed a licensed builder, Mr Unicomb, as a handyman to perform maintenance work on the property. He could not be called as he is dead. If a complaint had been made about the defective water heater, it was submitted that it was virtually certain that it would have been brought to Mr Unicomb's attention and that something would have been done about it. Mr Lindsay himself inspected the premises from time to time and occasionally undertook maintenance work. He said that he was never aware of any problem with the water heater, nor was it brought to his attention by the tenants or anyone else. 25 On the face of it, the defendants' arguments have some weight. However, they are contradicted by direct evidence from witnesses who I judged to be truthful. The condition of the cover was not a matter, it seems to me, about which they might have been mistaken and, hence, wrongly believed that they were telling the truth. The matters relied on by the defendants are suppositious. Encomiums about Mrs Lindsay by her son, though understandable, could not be regarded as independent. I am also unable to assess Mr Cook's attitude or the reasons for it. Whether the gas fitter whose job was merely to change over the supply noticed that the cover was defective or thought that this might be significant is too speculative to have any significant weight nor can I safely draw any conclusions from what Mr Unicomb might or might not have done. For all I know, he may have made an unsuccessful enquiry about replacement of the cover. My understanding of Mr Lindsay's evidence was that the matter of the heater was not brought to his attention but Mrs Craft was not asked whether she had done so, and if not, why not. 26 The final witness relied on by the defendants is Mr Stokes, the fire officer who attended the scene shortly after the fire was extinguished. His job was to ensure that the fire was in fact out and to make a brief assessment as to how it had been caused. He said that he walked into the laundry and noticed a gas water heater which was intact. He was not asked to recollect this particular call out, however, until sometime in 2000, namely almost sixteen years afterwards. He agreed with the occurrence book entry that it was described as a "small fire which was extinguished by means of a garden hose" and agreed that it was not "an overly memorable event". Although the occurrence book report was produced, Mr Stokes said that it was not usual for it to contain any statement as to the possible cause or causes of the fire. Such a statement would be contained in a report prepared by the Coroner but that he was unaware either that any report was actually prepared in this case or of its contents. (Mr Stokes gave evidence that he saw a fuel container in the laundry but that it was a four litre metal container with the lid off, lying on its side. In the end, however, the possible significance of this was not the subject of submissions and I note it merely for completeness.) Mr Stokes said that the heater appeared to be in its normal condition with "nothing missing. It all appeared to be connected correctly". He noticed no corrosion. He said that the cover plate was on the heater, fixed in place, although he did not inspect it. He said that this particular fact had "stuck in the back of my mind ever since the day it occurred" because of the injury to the boy. On testing this evidence, however, it appeared that what Mr Stokes was really saying was that the aperture in the side of the heater was completely covered by the plate and he agreed that because he could not see the internal mechanism of the heater, he concluded that the plate was clipped into position. 27 It is clear to me, and I think it was clear to the witness, that the dimensions of the relevant items are such that, even with the cover plate on the ground leaning up against the side of the heater, it would almost completely, if not completely, cover the aperture. 28 The evidence given by Mr Stokes, at face value, strongly supports the defendants' contention that the cover plate was not damaged and, at all events, was in place at the time of the fire. After so many years, I would be surprised if Mr Stokes's recollection about this matter was reliable although I do not doubt his honesty. Giving Mr Stokes's evidence due weight, I have come to the conclusion that, more probably than not, the cover plate was indeed defective and could not be attached to the heater in its designated position. However, it is not possible for me to say that, more probably than not, the cover plate was not leaning against the side of the heater, in which event it would have virtually covered it.

The effect of the absence of the cover plate 29 It is essential for the plaintiff to establish not only that the cover plate was not attached to the heater because it was defective, but also that, had it been attached, the ignition of the petrol fumes would have been retarded or delayed sufficiently to have permitted Mr Hemsley to have rescued his son. The scientific evidence concerning the possible effect of the cover plate on the movement of the flammable vapour into the heater cavity is complicated and difficult. This is demonstrated by the fact, if by nothing else, that highly qualified experts were called on both sides who came to quite different conclusions about this question even though their expertise was considerable and, I am satisfied, they were all doing their honest best to consider it objectively. Although, to an extent, each expert sought to justify and defend his thesis, this went no further than a genuine attempt to deal with the criticisms levelled at their reasoning. I did not think that they became advocates for the party in whose case they were called. Indeed, it is because of their professionalism that I have been able to come, with some confidence, to a conclusion at least on the balance of probabilities on the scientific issues in the case. 30 An initial question arises which it is convenient to dispose of first. The plaintiff's expert (Dr Green) proposed different ignition delay times depending on whether the burner was on or the only source of ignition was the pilot light. The defendants' experts (principally Dr Casey) thought that there was no, or no significant, difference in time lapse whether the plate was attached or unattached and whether the burner was alight or not. As his evidence developed, Dr Green conceded that, when the burner was alight, ignition would have been delayed only by a second or so if the cover was attached. Both sides were therefore agreed - though for quite different reasons - that the attachment of the cover plate made no significant difference if the burner, as distinct from the pilot light alone, were alight. Whether the burner was on or off at the time of the explosion is very difficult to determine. On the one hand, it appears that no domestic activities such as bathing or showering or clothes or dishwashing were being undertaken at the time. Although one of the children had washed her hands shortly before the petrol spill, this was probably with cold water and, if not, the hot water used would not have been substantial. On the other hand, it is not possible to assess how long it had been since the burner last operated and thus whether the water in the heater had anyway cooled to the point that the automatic switch would have commenced reheating. Commonsense suggests that the burner is alight for only a very small proportion of the daylight hours. Thus, in the absence of actual use of hot water at the time as by washing dishes or clothes or the like, the burner is far more less to be on than off at any particular time. It seems to me that this makes it more probable than not that the burner was not alight at the time of the petrol spill. Accordingly, the discussion of the scientific evidence focuses on the suggested consequences of attachment of the cover plate with the burner on and the pilot light providing the flame source for ignition. I should mention that, if this finding is wrong and the burner was alight at the time of ignition, the evidence of the plaintiff's expert and that of the defendants is to the effect - though for differing reasons - that attachment or otherwise of the cover plate would have made no difference to the time of ignition.

The physical environment 31 Crucial to the fundamental issue in this case, namely whether there would have been a significant delay in ignition if the cover plate were attached to the heater, is not only the structure of the heater and the size and location of the holes in its cylindrical shell that gave access to the air and petrol mixture that ultimately combusted but also the immediate environment of the laundry itself. In this respect, the most important factor was that the floor of the laundry, as is conventional, sloped towards a drainage hole which, as it happened, was in the bathroom floor behind the heater and somewhat to one side. It was agreed that this had the effect of moving the liquid petrol towards the heater. As I have previously implied, the combustible material is not the petrol itself but the mixture of air and petrol (usually called the "vapour" in this judgment) that is formed when the petrol, being volatile, fumes. Close to the liquid fuel the fumes will be too rich to ignite. A concentration gradient is created as the petrol fumes undergo a process of mixing with the air in the room, with the mixture becoming leaner, broadly speaking, as the distance from the liquid increases. It should also be borne in mind that, as the fuel moves, the evaporated fumes lag somewhat in the same way that steam from the funnel from a steam train plumes out behind the train as it moves forward. Of course, the speeds involved are enormously less but, even so, the effect of the movement is to create a combustible mixture along much of the leading edge of the fumes, the available mixture for combustion being increased by the drag effect of the air on the fumes as the fuel itself moved forward. It was agreed that that the fuel was moving towards the heater at about 0.15 metres/second, with the "plume" necessarily moving more slowly than the liquid fuel. (It is important to note that this number is an approximate estimation and not a measurement. It is obvious that the speed depends on, amongst other things, the angle of the slope, the acceleration rate and the distance travelled, none of which were measured. As will appear, the margin of error necessarily involved in this estimate has significant consequences.) These speeds are unaffected by the presence or absence of the cover plate. It seems very probable that the moving liquid was a very significant influence on ambient (or background) convection in the room - leaving aside the possible effect of the heater. 32 The flammable concentration of vapour created along most, if not all, of the length of the plume from a very short distance above the liquid fuel almost certainly rose at least to the height of the pilot light inside the heater. Although, at the height of the inlet, the plume caused by flow of fuel along the floor must have been of negligible volume and unlikely to have much of an effect on the ambient convection at that point, there is nonetheless a real likelihood that the initial ignition was of vapour entering through the slots at the bottom of the heater, as distinct from the pilot light cavity. In short, if the vapour ignited because the flammable plume arrived at the pilot light through the floor vents before the vapour entering the pilot light cavity reached the flame, the presence or absence of the cover was irrelevant. (I should mention that, whilst the plaintiff's expert, Dr Green, accepted that somewhere in the leading edge of the plume as it moved under the heater, the fuel/air mixture would be within the flammable zone, he thought at the height of the burner, namely 20 mm, most of the plume would be between 20 and 40% petrol vapour, which would be too rich to burn. Accordingly, the leading edge would ignite but it would be quickly followed by the over-rich mixture. However, the crucial height for present purposes is the height of the pilot light, so this possibility can be ignored.) 33 The other, and directly relevant, parameter is what is called the "chimney effect" created by the design of the heater. It is this effect upon which the plaintiff's case partly depended, at least at first. Air heated by either the pilot light or the burner is drawn into the heater flue, creating a suction that pulls air into the heater which, together with the gas, creates a flammable vapour ignited by the flame. It is obvious that when the burner is on the volume of heated air and vapour is all the greater and suction may be expected to increase. When the pilot light is the only source of heat the suction and thus the chimney effect is very much reduced, by perhaps as much as 97% or even 99% (both these numbers, however, being unmeasured estimates). It is the "chimney effect" at the pilot light cavity that is crucial here and it cannot be assumed that this effect is at a scale that is even measurable, even if the burner is on. As the expert evidence unfolded it appeared to me, that a crucial (though not the only important) difference between the expert witnesses called for the plaintiff and those called for the defendants, was the significance afforded by them to the results of attempts to measure these effects. 34 I have mentioned the ambient convection that naturally occurs in any room. As will be seen, measurements of air velocities taken close to the heater - and presumed to be caused by the chimney effect - showed very slow speeds. These speeds must be considered in the context that airflow in a closed room, say the convection set up by one wall being warmer than the others, would be expected to generate speeds of 0.02 metres/second or greater. Open windows, movement by persons in the room and the like would all support air speeds greater than this. If ambient speeds exceeded those with which the air moved into the access cavity, then the potential effect of the cover plate on chimney flow is substantially lessened, if not entirely removed. At the relevant time, at least two children were in the laundry and one was pouring liquid onto the floor. The liquid was moving towards the heater. A door was open and people were moving just beyond it. There was, as I understand it, no open window in the laundry. I think it follows that it was probable that ambient air flow was at least at 0.02 m/s and may well have been more.

The plaintiff's expert evidence 35 The plaintiff relied principally upon the evidence of Dr A R Green, Senior Lecturer in the School of Safety Science at the University of New South Wales. Dr Green has extensive relevant experience and a long list of publications to his credit. It was not suggested that Dr Green was not fully qualified to give the opinions stated in his reports and in evidence and it is therefore unnecessary to deal further with his expertise. 36 Dr Green's opinions changed somewhat in the course both of his investigation and his evidence in response to experimental data reported by the defendants' experts and as further explanations were forthcoming that clarified particular aspects of their reports. Dr Green placed great significance upon what he called a "zone of influence", in lay terms being the effect that the heat generated within the heater by the burner or the pilot light - and the chimney effect thereby created - had on the flow of air and petrol vapour from the room into the heater. Dr Green concluded that the time taken by the vapour to reach the flame would have been greater when the cover plate was attached because (as I understood his evidence) although the smaller access area to the flue (and hence the flame) available via the vent increased the velocity of the air at that point (because the same volume of gas is moving through a smaller opening) the larger area of the open cavity created a significantly greater zone of influence that had the effect of increasing the velocity of the vapour over the space between the point of the spill to the inlet. The existence of such a zone of influence was strongly disputed by the defendants' experts. They accepted that the vent created by attachment of the plate would cause some increase in velocity of the vapour at or close to that point but considered that the theoretical reasoning underlying Dr Green's view of the position that applied when the cover plate was off was inapplicable for a number of reasons that related principally to the distinctions between suction and jet effects and laminar and turbulent flows. Whatever the merits of this debate were as a matter of theory (perhaps applied theory is a fairer term), the defendants' experts considered that their experiments decisively proved that Dr Green's model was inapplicable. 37 I think that the significance of Dr Green's opinion about the extent and velocity-effect of the zone of influence is very much reduced by my finding that the burner was not alight at the crucial time. Even on his theory, the suction or chimney effect produced by the pilot light is very greatly less - agreed at 1%-5% - than that produced by the burner. This is one reason for my having had very great difficulty in accepting Dr Green's early conclusion (later departed from) that, with "the pilot light only being on the difference between cover plate on, cover plate off is something of the order of 20 to 30 seconds…[whilst with] the main burner on, though, that difference comes down to about 3 to 5 seconds". 38 Turning to the significance of the petrol flow on the floor carrying the vapour towards the heater, Dr Green said that from the point of spill to ignition (agreed at about 2 metres), the time lapse would have been in the order of 12 seconds, expressing the opinion that the "dominant effect is the convection of that vapour [caused by the flow]". He added - "That is with the plate off. With the plate on, you have still got that effect of flow but you are now, your velocity is high enough to, where the zone of influence becomes important and not a direct flow. Q. What effect does that have? A. It basically lengthens the time." 39 I understand Dr Green also to be saying that the effect of the plate would have been to prevent the movement of some of the vapour into the cavity by operating as a barrier and slowing the movement of other air into the cavity by requiring it to take a longer path to the flame (though this is not a zone-of-influence effect). Since this is an effect at, or almost at, the vent, where (as I understand him) Dr Green says there is an increased velocity by virtue of the smaller area of access to the cavity, it must be considered as a countervailing influence to the delay caused by the longer path to be taken. I have difficulty in understanding why the net effect is to slow rather than to speed up the ingress of the vapour (but see below for further discussion of this point). More decisive an obstacle to my determining this matter in favour of the plaintiff is that there was no real attempt to measure or assess the extent of the posited net change, assuming it to involve a delay in ignition and the imprecision of the theoretical result, even if slight, renders the problems of time estimation to which I have referred above all the more significant and, I think, adverse to the plaintiff's case. 40 It is obvious that the difference of 20-30 seconds stated above is a substantial departure from the conclusion expressed in Dr Green's first report that, whether the burner was on or only the pilot light, the absence of the cover plate would have advanced the ignition time by "more that 100 seconds". Making full allowance to the evidence of Dr Green about the difference between the assumptions he made at the time of that report and when he gave evidence, the variation indicates how correct his observation was to the effect that the variables that affected the crucial question were such as to introduce a significant degree of uncertainty into any conclusions about it. 41 Before moving to Dr Green's final position - and that urged on me by counsel for the plaintiff - I should summarise the defendants' scientific case.

The defendants' expert evidence 42 It is, I hope, already clear that the crucial scientific question in this case was whether the speed with which the flammable vapour reached the pilot light was significantly affected according to whether the cover plate was attached in its proper position. (It was agreed that the possibility that the cover plate was leaning against the side of the heater does not need to be taken into account. Accordingly, this complication can be put aside.) The defendants' case developed as investigations proceeded and responses to the plaintiff's expert reports were prepared. 43 It is sufficient to deal with this evidence in its ultimate form and to focus on the evidence of Dr Casey. Dr Casey is a consultant engineer and a lecturer in the School of Mechanical and Manufacturing Engineering at the University of New South Wales. He describes himself as an "experimentalist". Although his experience has been less concentrated than Dr Green's on issues directly concerned with heat and combustion, it was not suggested that he was not qualified to express the opinions contained in his evidence. In addition, he worked with Dr Barrett, also a highly qualified engineer with a great deal of experience in fire and explosion investigation. 44 Before moving to experiments conducted by Dr Casey, I should mention some experiments performed by Dr Barrett to confirm that there would not be any noticeable difference on air flow into the heater when only the pilot light was burning depending on whether the cover plate was attached or not. Direct measurements were made at the access hole using both an anemometer of the rotary vane type and a Pitot tube. (No measurements were taken at the base.) No measurable readings were recorded, whether the plate was on or off. It is obvious that some air was being burned, else the pilot light would not have remained alight and that the air in the cavity must have come both from the access hole and the apertures in the base of the heater. The limitations in the instruments used - velocities less than 0.02 metres/second could not be measured and the directional limits of the Pitot tubes - qualified the reliance that could be placed on these results. Nevertheless, the fact that no variation could be detected was far from a trivial result, especially in light of later experiments conducted by Dr Casey, to which I now turn. (The fact that similar results were obtained when the burner was lit gives rise to serious questions about the correctness of the assumption to which I referred above that, so far as the ingress of air through the inlet hole is concerned, the effect of the burner is much greater than that of the pilot light.) 45 In part because of the limitations of the instruments used by Dr Barrett in the experiments I have briefly mentioned above, Dr Casey constructed a number of experiments designed to measure actual airflows into the pilot light cavity as well as the vents in the base, using a hotwire anemometer as well as a Pitot probe. The former instrument has an ultra fine wire, through which an electrical current is passed, suspended between the tips of two probes. As it heats, its resistance changes. Because air flowing over the wire tends to cool it, its resistance is lowered. Any flow across the wire in any direction will lead to a cooling effect and therefore a change in the resistance. These variations in resistance permit the velocity of the airflow to be determined. Unlike the Pitot probe, therefore, the hotwire anemometer is not directionally sensitive. It is also far more sensitive than the Pitot probe in its speed of response and the slowest velocity capable of detection. 46 Great care was taken to choose a day with little or no wind and to ensure that the laboratory was sealed. All moving objects (including people) that might affect ambient air velocity were removed. The hotwire anemometer was used to record ambient air velocity in such a way as to confirm that it did not distort the experiment. Of course, these precautions meant that the laboratory was very different from the laundry in which the explosion occurred but this does not affect the reliability and applicability of the experiments. The actual situation, as a whole, was more complicated in the laundry than in the laboratory so that, if anything, the movement of air in the laundry would have been greater than that in the laboratory. From the plaintiff's point of view, therefore, the approach taken by Dr Casey was a conservative one. At all events, it was necessary because otherwise the effect of the cover plate could not be isolated and, hence, properly tested. 47 The air flow both at the access and the base was measured by probes placed in positions which enabled them to measure the air speed at various points moving out from the wall of the heater. Measurements were taken at heights of 10 mm and 20 mm above floor level. When the cover plate was removed, the inlet was probed to determine the air speed at a number of locations along a symmetrical 5 x 5 grid over the aperture at the plane of the heater shell. These experiments were conducted both with the burner on and off. 48 The results indicated that, whether the burner was on or off, at distances of less than 60 mm radially out from the heater, the air speed fell below the test ambient levels (a level, I repeat, considerably less than those which would be expected in actual circumstances). Accordingly, at distances exceeding 60 mm from the heater, the effects of the heater did not significantly influence the air flow since the ambient air velocity, even in the sealed laboratory, was greater than that developed by the suction or chimney flow to the heater. 49 With the burner on and the cover off, the airspeed results at the inlet itself varied between zero (ie, below detectable limits) and 0.16 metres/second for the top half of the opening and between 0.12 m/s and 0.26 m/s for the bottom half. Integrating these measurements yields an overall airspeed, perhaps, of about 0.07 m/s, less than ambient convection. With the cover on, measurements could not be taken at the plane of the inlet but, at 20mm out and the height of the vent, they averaged close to 0.2 m/s, whilst increasing speed. Once through the vent and into the cavity - close to the plane of the heater shell - the expansion of the vapour must slow its velocity. Dr Casey thought, and I accept, that these measurements demonstrate there was no significance difference in airspeeds at the heater, even with the burner on, whether the cover was on or off. With the burner off, all measurements were less than 0.1 m/s. When considering the effect of the pilot light alone, this finding is the crucial result from Dr Casey's tests. 50 The results obtained with the hot wire anemometer differed somewhat from those obtained with a Pitot probe. Dr Casey's opinion is - and I accept - that, having regard to the directional sensitivity of these probes, the results obtained from the former instrument are to be preferred. The highly diffusive nature of the flow to the heater creates a multi-directional flow pattern that is insufficiently straight over the length of the Pitot tube to enable it to function efficiently. 51 The hotwire results show that the suction flow with the burner on could only have an effect on any flammable vapour over the last 60 mm (of the approximately 1830 mm) of travel from the place where the fuel was spilt to the heater, whether the cover plate was on or off. In other words, the path of the vapour would be influenced by factors other than the heater for at least 97% of its journey, upon the (unlikely) assumption that ambient convection was less than 0.2m/s. With only the pilot light on, there is no, or virtually no, effect on airflow velocities outside the heater cavity if the cover plate is attached. 52 I should add that these conclusions, essentially those of Dr Casey, are agreed by Dr Barrett, a consultant engineer retained for the defendants. Dr Barrett's report is both objective and persuasive. He concluded - "There are a lot of uncertainties attached to this incident and the actual presence or absence of a cover plate is a matter for speculation. Similarly, the distance between a spillage and its orientation with respect to the heater is not known. Other factors such as room ventilation and floor slope also enter into consideration. The experimental work performed by myself and Dr Casey was simply to discover under what conditions the presence or absence of a cover plate would affect the rate of flow to the heater. The accumulated evidence would support the belief that the rate of air flow to the heater does not increase when the cover plate is absent and that other factors are dominant." 53 As is already clear, I have determined that the cover plate was not attached at the time of the explosion. However, the other uncertainties to which Dr Barrett refers remain uncertain although the direction and extent of the floor slope are not now the subject of significant dispute.

Dr Green's ultimate position and Dr Casey's critique 54 Dr Green agreed that the measurements made by Dr Casey were of crucial importance in determining the convection flows at the heater. He appeared, I think, to accept that the zone of influence that he had posited (by deduction from jet flows) was far greater than Dr Casey's measurements justified and concluded that that the most significant contributor to the movement of vapour was imposed by the flowing petrol carrying it in its train. (I interpolate that, since ambient convection was reckoned at 0.2 m/s or greater, this conclusion is questionable.) He then pointed out that, when the cover plate was off, there was no obstacle to the horizontal flow of the vapour straight into the heating cavity but that, with the cover on, the vapour had to enter (for practical purposes) through the bottom of the vent and thus to change direction. The flow at the vent, he said, was vertical in order to negotiate slight wall of the heater shell to enter the heater cavity and "you tend to find in these types of flow that the vertical velocity component is an order of magnitude less or at least an order of magnitude less" than vapour travelling horizontally. 55 The time difference to ignition, if any, between the cover plate being on and off - with the burner off - was principally determined, according to Dr Green, by the reduction in the velocity of the vapour at the vent when the cover plate was attached caused when it changed direction and its ensuing velocity over the distance to the pilot light. Upon the basis that the relevant airflow was vertical into the cavity, Dr Green surmised that it would have been moving at 0.02 m/s (that is to say, an order of magnitude less than its horizontal velocity of 0.2 m/s before it changed direction to enter the vent). But this seems to confuse Dr Casey's results when the burner was on and the results when the burner was off. In the latter case - the case under consideration - those results showed an airspeed of less than 0.1 m/s. Perhaps Dr Green meant that he assumed an ambient convection speed of 0.2 m/s but this is not his evidence and I do not think that I should infer that this is what he meant. Moreover, as Dr Casey pointed out, since the hot-wire anemometer is omnidirectional, there is no basis for assuming that the velocity he measured was horizontal and, accordingly, that the approximation of an order of magnitude difference when the vapour changed direction is appropriate, even if one accepted that the vapour changed to a vertical flow, which is also subject to doubt. In other words, when Dr Green said (as I understood him) that Dr Casey's 0.2 m/s was a measure of horizontal speed, this was not what the instruments were measuring: they were measuring airspeed in all directions over the heated wire. I should add that the bottom of the vent, as I understand, it did not require, as a matter of its physical relation to the inlet, that vapour moving from outside through the vent had to move vertically, though no doubt some was deflected - perhaps vertically - as ambient convection carried it into the cavity. 56 Consequently, whilst it is reasonable to infer that some vapour was moving horizontally, it cannot be inferred that it was moving at 0.2 m/s, which is the starting point of Dr Green's analysis. Secondly, the order of magnitude proposed as applicable to a change of direction from horizontal to vertical will not apply to vapour that did not change direction to that extent in order to negotiate the vent and enter the heater cavity. These problems represent further uncertainties which Dr Green's hypothesis does not, as it seems to me, take into account. 57 In addition to the slowing effect caused by the change of direction of the airflow as it arrived at the attached cover plate, Dr Green pointed to the longer path that the vapour needed to negotiate in order to reach the pilot light as distinct from the unimpeded flow to the light in the absence of the cover plate. One difficulty (which I found convincing) with this analysis, as Dr Casey said, is that the "unimpeded" flow must slow once it enters the cavity and expands into that space and, thus, it ought not to be assumed that the velocity of the vapour remains constant to the flame. Another difficulty is that, on my examination of the photographs, I think that the path to the pilot light required to be taken by vapour entering the cavity around the cover plate is only very slightly longer than the path it would need to take if the cover plate was unattached. 58 With the cover plate off and assuming that the distance to the pilot light was 150 mm (although the only evidence was that it was between 100 mm and 150 mm), if the vapour was moving at a velocity of 0.02 m/s, over the whole of the distance to the pilot light, the time to ignition would have been about 7.5 seconds whilst, if that velocity were 0.2m/s, it would have been 0.75 of a second. The difference here suggested is less than 6 seconds and it is obviously subject to a significant margin of error. The actual size of that margin is very difficult to assess, since the calculation depends, at the least, on an assumption (albeit educated) about the order of magnitude of the difference between horizontal and vertical velocities, although it is obvious - as I have pointed out - that not all the vapour travelling through the vent (the cover being attached) is travelling vertically, nor is all the vapour passing through the open inlet travelling horizontally. Dr Green considered that the limited zone of influence that operated when the cover plate was attached would have an additional slowing effect on the plume entering the cavity but this seems to me to be quite wrong by reference to Dr Casey's measurements. 59 It will be seen that the plaintiff can only succeed if I accept more probably than not both the theoretical approach and the precise calculations of Dr Green. One major problem with the former element is that I am satisfied that his application of jet force dynamics to the suction force relevant here is flawed in principle for the reasons given by Dr Casey and led Dr Green to what was demonstrated (as he, I think, ultimately conceded) to be a significantly mistaken opinion about the scale of his posited zone of influence; another problem is his mistaken assumption that there would be a difference of about one order of magnitude in the suction force at the inlet caused by the burner as distinct from the pilot light. Even if these criticisms be dismissed, so far as precision is concerned, the plaintiff cannot succeed unless I accept that the margin of error in Dr Green's calculations is so small as to be insignificant. His own language describing the difference between horizontal and vertical movement (an "order of magnitude") is general rather precise and imports a further element into the margin of error. In short, although Dr Green assayed a precise calculation, that precision - when his assumptions are examined - is very much more apparent than real. 60 An additional problem arises, to my mind, from Dr Green's treatment of the significance of the spill. As I have mentioned, although Dr Green's calculations depended on a horizontal airspeed of 0.2 m/s, he also stated that the most significant influence on the movement of the vapour towards the heater was the spill itself, estimated (as mentioned above) at 0.15 m/s. Dr Green assumed that this would have the effect of pulling the plume along at the same speed, which must be wrong, as Dr Casey pointed out. Whatever the velocity of the plume just above the liquid, it is difficult indeed to accept that it was moving any faster than ambient convection at the height of the inlet aperture (20 mm), let alone at the height of the pilot light (35 mm). The assumption of Dr Green that, at these distances from the floor, it was moving at anything like the speed of the liquid strikes me as plainly wrong. Another obvious point is that the height of the plume above the spill is unknown, although it is certain that it plumes back at an angle. This leads to the reasonable inference that the spill would reach the underside of the heater before the plume reached the inlet aperture some 20 mm above it, in all likelihood significantly before the plume. Dr Casey showed that there was significant airflow through the base vents into the heater cavity. The distance between the pilot light and the base is 25 mm (assuming that the distance stated by Dr Casey is measured from the ground - which is what I have inferred), very much less than the distance between the inlet aperture and the light. An elementary calculation shows (assuming Dr Green's order of magnitude reduction in vertical velocity to be correct) that the vapour would reach the pilot light from the vents in slightly less than 2 seconds. Accordingly, even if the vapour reached the open inlet aperture at the same time as the liquid, the vapour entering through the base vents would ignite well before the vapour that accessed the pilot light through that aperture. This hypothesis was not put precisely to the experts but it seems to me to be implicit in or follow from Dr Casey's evidence and certain elements - using the distance of 20 mm rather than the correct distance - were also put to Dr Green by Mr Wheelahan QC. At all events, I do not rely on this calculation to reach my conclusion in respect of Dr Green's evidence but simply to indicate that there are troubling uncertainties about the reliability and precision of his conclusions. 61 Dr Casey's analysis of Dr Green's solution for the potential velocity changes occurring by the attachment of the cover plate involves a number of further significant criticisms of Dr Green's position. The two most significant of these relate, firstly, to the volumetric calculations performed by Dr Green in which he compared the volume of vapour going through the pilot light cavity with the volume of vapour going through the base vents. Dr Casey has convincingly demonstrated to my mind that Dr Green's analysis proceeded upon a mistaken understanding of the design of the heater and the way in which the measurements performed by Dr Casey were undertaken. Furthermore, Dr Green's calculation is so inconsistent with Dr Barrett's actual measurements at the pilot light cavity that I feel bound to reject them. Although this criticism is directly related to the situation when the burner was on, it indicates a troubling error in Dr Green's consideration of the issues in the case. The second significant criticism, also persuasive and more fundamental, I have referred to already in a slightly different context. That is, that it is wrong to think of there being anything like a horizontal airflow moving towards the inlet aperture. There is no, or next to no, suction effect with only the pilot light burning (as Dr Green agrees). And the spill is most unlikely to be pulling the vapour towards the heater at the height of the inlet at all, let alone horizontally. If the air at the inlet is moving in all directions, a basic assumption of Dr Green's ultimate position is substantially undermined. I have concluded, in accordance to Dr Casey's opinion (though making some allowance for uncertainties) that, if there was any delay in ignition likely to have been caused by the cover plate had it been attached, it would only have been of the order of 2 or 3 seconds at the most.

Conclusion on the scientific evidence 62 It would be sufficient to state that, at the end, I am not persuaded more probably than not that the attachment of the cover plate would have significantly delayed ignition. However, I think I should say, as a matter of fairness to the defendants, that I have concluded that the overwhelming probability is that the non-attachment of the plate had little or no significance in terms of causing a delay to the explosion caused by the spillage of the fuel: the attachment of the plate would not have enabled the plaintiff to escape his dreadful injuries.

DUTY OF CARE AND FORESEEABILITY 63 My conclusion on the facts means that there must be judgment for the defendants. However, against the possibility that this conclusion may be held to be in error, I should deal with the second limb of the defendants' case, namely that there was in the circumstances no duty of care to ensure that the cover plate would attach to the heater and that the injury suffered by the plaintiff was not foreseeable as a consequence of a failure to ensure that the cover plate could be attached. 64 It is, of course, possible that a householder could bring any explosive material into a house so that any appliance with a naked flame could potentially ignite it. The fact that a container of petrol might be brought into the laundry is foreseeable in this sense. It is also foreseeable that, if this happened, petrol might spill and a flammable vapour released into the room that might be ignited by the heater flame. It is not contended by the plaintiff that any injury caused by a fire or explosion in this event could be laid at the defendants' door. It is not suggested that the non-attachment of the cover plate made the explosion - or, indeed, any explosion - any more likely. It is not contended that it was foreseeable that the absence of the cover plate might advance the time of explosion by a few seconds. 65 On the question of causation, counsel for the plaintiff submit that, whilst there was an extreme risk of injury to the plaintiff without "the defendants' contribution…by omitting to repair…the cover plate on the heater, the risk of injury to the plaintiff was significantly magnified" because, "in delaying the explosion the explosion, it would have averted the harm to the plaintiff", citing the judgment of Gaudron J in Bennett v Minister of Community Welfare (1992) 176 CLR 408 at 420-422. It is, I think, significant to note that the passage to which counsel refers contains the following (at 422) - "…[There] is usually no reason to separate or distinguish the question of breach of a common law duty from that of causation…That is because a duty is imposed by the common law by reason that it is a precaution which a reasonable person in the position of the person sued would have taken to prevent a foreseeable risk of harm of the kind suffered. Thus questions of the sufficiency of the precaution to avert the harm are inevitably subsumed in the finding that there was a duty: a precaution is not classified as "reasonable" unless it can be said that its performance would, in the ordinary course of events, avert the risk that called it into existence ." (Emphasis added.)

Applying this principle, it seems to me that, since the possibility that the harm constituted by the posited delay in ignition might be avoided by attachment of the cover plate was not reasonably foreseeable, no reasonable person in the position of the defendants would have taken the precaution of ensuring the cover plate was attachable to avert it. 66 The defendants concede that they owed a duty of care to the plaintiff but that was limited to the avoidance of a defect that would have been manifest on a simple inspection of the premises and which rendered them unsafe: Northern Sand Blasting Pty Limited v Harris (1996) 188 CLR 313. It is trite that the duty on the defendants was to take reasonable care to avoid a foreseeable risk of harm to the plaintiff. This requires consideration of "the question whether it is reasonable to require a person to have in contemplation the risk of injury that has eventuated": Game v NSW (2002) 211 CLR 317 per Gleeson CJ at [12]. 67 It seems to me that the question, "Is it reasonably foreseeable that petrol might be spilled on the laundry floor and release a flammable vapour which might be ignited by the heater flame?" must be answered, "Yes". However, the question, "Would the attachment of the cover plate make any difference?" must be answered, "No". The question, "Might the attachment of the cover plate make any difference to the time lapse between the spill and the ignition?" would simply not occur to any reasonable person. Accordingly, I would answer the question posed by Gleeson CJ and quoted above, "No". 68 I have already mentioned the purpose that the cover was designed to serve. No one suggested and, indeed, it would be absurd to suggest, that it was designed to delay any ignition that might occur in the circumstances of this case. If the cover, when attached, caused any material delay in ignition in these circumstances (and this would only have arisen if the burner were off) this would have only been an accidental consequence of the design. The mere fact that it might have been foreseeable that the flame, either of the pilot light or of the burner, could ignite any flammable vapour with which it came in contact is not material. After all, the heater functioned by ignition of a flammable vapour, albeit the gas component came from a purpose-built pipe, and required the ingress of air to permit combustion. Even if this were a crucial fact, it has not been suggested by the plaintiff that there were any means by which, consistent with the reasonable functioning of the heater, ignition in the circumstances of this case could have been avoided. The real question is not whether this occurrence was foreseeable but whether it was reasonably foreseeable that damage or injury might result from the possibility that non-attachment of the cover plate would, in some circumstances, cause ignition to occur a few seconds before it might happen if the cover plate were attached in the event that petrol vapour had been released in the laundry. 69 In reliance on Glass JA's analysis of what his Honour called "the foreseeability enquiry at the duty, breach and remoteness stages" (in Minister for Environmental Planning v San Sebastian Pty Ltd and ors [1983] 2 NSWLR 268 at 295-6), Mr Wheelahan QC for the plaintiff submitted, first, that the defendants, acting reasonably, were not called on to exercise exceptional perspicuity or possess super-human qualities of imagination or have infinite resource and capacity (pace NSW Land & Housing Corporation v Watkins (2002) ATR 81-641, paras 71-75) in order to foresee the damage of some kind to the plaintiff. He contended that the naked flame, must and did constitute a risk of injury (by "naked flame" I apprehend that Mr Wheelahan was referring to the fact that the pilot light aperture was open) and that "if it (I think he meant the injury that such a flame might cause) came to pass, it was capable of having very serious consequences". 70 To my mind, the mere fact that the aperture was open did not create in any sensible way a significant risk of injury, at least if only the pilot light was burning. The evidence did not disclose precisely where the pilot light was situated within the cavity, but commonsense suggests that, if necessary, it was able to be relit without great difficulty. It is possible to envisage that an inquisitive child might poke some object into the heater and cause it to catch fire and that burns might result but this possibility, as a matter of commonsense, appears to be so remote and so different from what happened in this case as to require it to be dismissed as a relevant danger for the purpose of the principle contended for by Mr Wheelahan. Analogies must be used with care but it seems to me that his argument would require all matchboxes and cigarette lighters to be childproof to guard against the risk of catastrophic fires. It should be also be borne in mind that the cover plate was, at all events, easy to remove for the very obvious reason that, if the pilot light should go out it was necessary that the householder could easily relight it. 71 It seems to me that Mr Wheelhan's contention must also fail because it does not take into account the fundamentally different character of the danger, submitted to be foreseeable, of burning arising merely from the non-attachment of the cover plate and the danger, conceded to be not foreseeable, that ignition of a flammable vapour coming into contact with the pilot light might occur some seconds earlier than if the cover plate were attached, where the burning - in the absence of delay - could not be said to have been caused by the "defective" cover plate. 72 In Hughes v Lord Advocate (1963) AC 837 a young child was severely burned when he entered an unguarded tent covering a manhole that had been opened to repair underground telephone equipment and knocked or lowered one of the warning paraffin lamps into the hole, where it exploded. Particular reliance is placed by the Mr Wheelahan on the speech of Lord Reid in which his Lordship said (agreeing with Lord Guest) - "[At 845-6] The explanation of the accident which has been accepted…is that, when the lamp fell down the manhole and was broken, some paraffin escaped, and enough was vaporised to create an explosive mixture which was detonated by the naked light of the lamp. The experts agree that no one would have expected that to happen: it was so unlikely as to be unforeseeable. The explosion caused the…injuries… [At 847]…This accident was caused by a known source of danger, but caused in a way which could not have been foreseen and, in my judgment, that affords no defence…"

It seems to me that the last statement should be understood for present purposes with the addition after the word "danger" of the words: for which the defendant would have been held liable had that particular danger caused the injury in a foreseeable way .

Lord Jenkins came to the same conclusion as Lord Reid, essentially for the reason that "the distinction drawn between burning and explosion is too fine to warrant acceptance" giving as an example the possibility that the child may have been very severely burnt by taking one of the lamps and upsetting it over himself, in which event the defendant would have been liable but, if the lamp exploded instead, its argument would require that it be held not liable: (1983) AC at 850. Lord Morris of Borth-y-Gest pointed out ( ibid at 851) that it should have been foreseen that children might be allured by the lights and were likely to play with them and thus to be burned, even though, as it happened, the pursuer was more severely burned by an explosion that could not have been reasonably foreseen. His Lordship concluded ( ibid at 853) that - "…there was a duty owed by the defenders to safeguard the pursuer against the type or kind of occurrence which in fact happened and which resulted in his injuries, and the defenders are not absolved from liability because they did not envisage 'the precise concatenation of circumstances which led up to the accident'…" 73 Lord Guest said (at 856) -- "…Concentration has been placed in the courts below on the explosion which, it was said, could not have been foreseen because it was caused in a unique fashion by the paraffin forming into vapour and being ignited by the naked flame of the wick. But this, in my opinion is to concentrate on what is really a non-essential element in the dangerous situation created by the allurement. The test might be better put thus: was the igniting of paraffin outside the lamp by the flame a foreseeable consequence of the breach of duty? In the circumstances there was a combination of potentially dangerous circumstances against which the Post Office had to protect the appellant. If these formed an allurement to children it might have been foreseen that they would play with the lamp, that it might tip over, that it might be broken, and that when broken the paraffin might spill and be ignited by the flame. All these steps in the chain of causation seem to have been accepted…in the courts below. But because explosion was the agent which caused the burning and was unforeseeable, therefore the accident, according to them, was not reasonably foreseeable. In my opinion this reasoning is fallacious. An explosion is only one way in which burning can be caused. Burning can also be caused by the contact between liquid paraffin and a naked flame. In the one case paraffin vapour and in the other case liquid paraffin is ignited by fire. I cannot see that these are two different types of accident. They are both burning accidents and in both cases the injuries would be burning injuries. Upon this view the explosion was an immaterial event in the chain of causation. It was simply one way in which burning might be caused by the potentially dangerous paraffin lamp…" 74 Lord Pearce considered that the circumstances posed an obvious risk of burning and conflagration and a fall, all of which occurred. The mere fact that the mishandled lamp produced a violent explosion instead of an ordinary fire did not "create an accident and damage of a different type from the misadventure and damage that could be foreseen" so that the "accident was but a variant of the foreseeable": [1983] AC at 858. 75 It seems to me that the circumstances in the present case are very different. In Hughes the event that occurred was held to be not significantly different to that which was foreseeable and for which the defendant would have been liable. Here, it could not be said that it was reasonably foreseeable that an injury might be caused by any delay of a few seconds in ignition of an explosive vapour. Allowing for the sake of argument that it was reasonably foreseeable that an explosive vapour might be liberated in the room where the heater was placed and thus that there might be an explosion caused by its contact with the pilot light with a consequent serious injury to anyone caught up in it, this would have been the inevitable result of the conventional construction of the heater and would have occurred whether the cover was on or off. The danger here was not the explosion, for which the defendants were in no sense responsible: it was the danger of a reduced time lapse by virtue of the non-attachment of the cover plate. 76 Mr Wheelahan QC also points to Chapman v Hearse (1961) 103 CLR 112 at 120, where the following passage from Haynes v Harwood [1935] I KB 146 at 156 was approved - "It is not necessary to show that this particular accident and this particular damage were probable; it is sufficient if the accident is of a class that might well be anticipated as one of the reasonable and probable results of the wrongful act."

It is clear that the general class of which the "particular accident and…[the] particular damage…might well be anticipated" is that containing the occurrence caused by the wrongful act of the defendant. Here, the explosion was not caused by any wrongful act of the defendants. The "accident" is the accelerated explosion time that, it is alleged, prevented rescue or, perhaps more precisely, the prevention of rescue because the explosion occurred some seconds before it would have happened had the cover plate been attached. I am simply unable to see how such an event could have been anticipated in any meaningful way "as one of the reasonable and probable results" of a defective cover plate. 77 In Hughes the foreseeable result was the burning by fire and the sufficiently similar result was the burning by explosion. Here, the foreseeable result (as submitted) was the burning by fire or explosion. I do not agree that the acceleration by a few seconds of that result could be said to be a similar result in any sense: the two results are wholly different in character. Accepting (as I do not) that the plaintiff would probably have been rescued had the cover plate been attached, and thus that the accelerated explosion by virtue of its absence was a substantial cause of his injury, it is obvious that this contribution is entirely different from the contribution made by the explosion itself. The unforeseeable circumstance that an injury might have been caused by the absence of the cover plate because an explosion took place a few seconds earlier than it would have done had the plate been attached is a wholly different cause of injury than the explosion itself. Moreover, in Hughes the foreseeable injury from which it was agreed on all hands the defendant was obliged to protect the injured child was burning from ignited paraffin. In this case it is not submitted - and could not be submitted - that the defendants had a duty to protect the plaintiff from being burned by the introduction of an explosive vapour into the laundry. The wrong alleged against them is that the non-attachment of the cover by virtue of its rusted condition caused that explosion to occur a short time before it otherwise would have. 78 The problem facing the plaintiff may be put in a somewhat different way. As I have said, the relevant "danger" (to use the language of Lord Reid above quoted), which the plaintiff must prove here, is that of hastened explosion in the event of the introduction at some distance from the heater of an explosive vapour likely to come into contact with the pilot light flame. In no sense (as is conceded) was the absence of the cover a cause of the explosion. For the reasons I have given, I do not consider that there was any material hastening of the explosion caused by the non-attachment of the cover. Even if there were, the evidence certainly does not permit me to conclude that this was known or ought to have been known. Indeed, as I understand the plaintiff's case, it is conceded that the hastening of the explosion by virtue of the non-attachment of the cover was not foreseeable by the defendants in any sense, precisely or otherwise. 79 Counsel for the plaintiff have cited many cases demonstrating that the principle for which Hughes v Lord Advocate stands has become firmly embedded in our law. Of course, I mean no disrespect in declining to mention or discuss them. But, for the reasons that I have given, it seems to me that that principle does not assist the plaintiff in the circumstances of this case. 80 CONCLUSION

I have every sympathy for the plaintiff but it seems to me that I have no choice but to give judgment for the defendants. **********

DISCLAIMER - Every effort has been made to comply with suppression orders or statutory provisions prohibiting publication that may apply to this judgment or decision. The onus remains on any person using material in the judgment or decision to ensure that the intended use of that material does not breach any such order or provision. Further enquiries may be directed to the Registry of the Court or Tribunal in which it was generated.