England and Wales Court of Appeal (Civil Division) Decisions

You are here: BAILII >> Databases >> England and Wales Court of Appeal (Civil Division) Decisions >> Stannard (t/a Wyvern Tyres) v Gore [2012] EWCA Civ 1248 (04 October 2012)
URL: http://www.bailii.org/ew/cases/EWCA/Civ/2012/1248.html
Cite as: [2012] WLR(D) 266, [2013] 3 WLR 623, [2013] Env LR 10, [2013] 1 All ER 694, [2014] 1 QB 1, [2012] 42 EG 133, [2012] 3 EGLR 129, [2014] QB 1, [2012] EWCA Civ 1248

[New search] [Printable RTF version] [Buy ICLR report: [2014] 1 QB 1] [Buy ICLR report: [2013] 3 WLR 623] [View ICLR summary: [2012] WLR(D) 266] [Help]

		Neutral Citation Number: [2012] EWCA Civ 1248
		Case No: B2/2011/2377

IN THE COURT OF APPEAL (CIVIL DIVISION)
ON APPEAL FROM WORCESTER COUNTY COURT
MR RECORDER POTTS
0HR00229

Royal Courts of Justice
Strand, London, WC2A 2LL

04/10/2012

B e f o r e :

LORD JUSTICE WARD
LORD JUSTICE ETHERTON
and
LORD JUSTICE LEWISON
____________________

Between:

	Mark Stannard (t/a Wyvern Tyres)	Appellant
	- and -
	Robert Raymond Harvey Gore	Respondent

____________________

(Transcript of the Handed Down Judgment of
WordWave International Limited
A Merrill Communications Company
165 Fleet Street, London EC4A 2DY
Tel No: 020 7404 1400, Fax No: 020 7404 1424
Official Shorthand Writers to the Court)
Jonathan Waite QC and Michele de Gregorio (instructed by DAC Beachcroft LLP) for the appellant
Philip Rainey QC and Nicholas Isaac (instructed by Beaumonts Solicitors) for the respondent
Hearing date: 26th April 2012

____________________

Judgment
____________________

HTML VERSION OF JUDGMENT
____________________

Lord Justice Ward:

Every law student is likely to remember how Mrs Donoghue suffered when a snail in a state of decomposition floated into her glass from the bottle of ginger beer manufactured by Mr Stevenson. Almost as memorable must be the case of Mr Ryland's reservoir bursting and flooding Mr Fletcher's colliery which gave rise to the rule we now know as Rylands v. Fletcher. Lest we cannot recall the iconic ruling by Blackburn J. on behalf of the Exchequer Chamber, here it is:

"We think that the true rule of law is, that the person who for his own purposes brings onto his lands and collects and keeps there anything likely to do mischief if it escapes, must keep it in at his peril and, if he does not do so, is prima facie answerable for all the damage which is the natural consequence of its escape. He can excuse himself by shewing that the escape was owing to the plaintiff's default; or perhaps that the escape was the consequence of vis major or the act of God; but as nothing of this sort is here, it is unnecessary to enquire what excuse would be sufficient. The general rule, as above stated, seems on principle just. The person whose grass or corn is beaten down by the escaping cattle of his neighbour, or whose mine is flooded by the water from his neighbour's reservoir, or whose cellar is invaded by the filth of his neighbour's privy, or whose habitation is made unhealthy by the fumes and noisesome vapours of his neighbour's alkali works, is damnified without any default of his own; but it seems reasonable and just that the neighbour, who has brought something on his own property which was not naturally there, harmless to others so long as it is confined to his own property, but which he knows to be mischievous if it gets on his neighbour's, should be obliged to make good the damage which ensues if he does not succeed in confining it to his own property. But for his act in bringing it there no mischief could have accrued, and it seems but just that he should at his peril keep it there so that no mischief may accrue, or answer for the natural and anticipated consequences. And upon authority, this we think is established to be the law whether the things so brought be beasts, water or filth, or stench," see Fletcher v Rylands (1866) LR 1 Ex 265, 279/280.

Whilst entirely concurring with that opinion, Lord Cairns, the Lord Chancellor, added the gloss that the land should not be used "for any purpose which I may term non-natural use" see Rylands v Fletcher (1868) LR 3H.L. 330, 339.

Although Donoghue v Stevenson [1932] AC 562 has withstood the test of time, Lord Bingham of Cornhill was to say:

"Few cases in the law of tort or perhaps any other field are more familiar, or have attracted more academic and judicial discussion, than Rylands v Fletcher", see Transco v Stockport MDC [2004] 2 AC 1 at [3].

It was subjected to even more withering criticism by Mason C.J. and the majority of the High Court of Australia in Burnie Port Authority v General Jones Pty Ltd (1994) 179 CLR 520 at [18] (though it is fair to say that the criticisms were not entirely accepted by the House of Lords in Transco):

"Notwithstanding the many accolades which have been, and continue to be, lavished on Blackburn J's judgment ((see, e.g., Wigmore, (1894) 7 Harvard Law Review 315, 383, 441 at 454: "the master-mind of Mr Justice Blackburn"; Newark, "The Boundaries of Nuisance", (1949) 65 LQR 480 at 487: "his great judgment"; Salmond and Heuston on the Law of Torts, 20th ed. (1992) at 314: "always been recognised as one of the masterpieces of the Law Reports"), that brief exposition
of "the true rule of law" is largely bereft of current authority or
validity if it be viewed, as it ordinarily is, as a statement of a
comprehensive rule (see, e.g., Jones v. Festiniog Railway Co.
(1868) LR 3 QB at 736 per Blackburn J: "the general rule of common law".) Indeed, it has been all but obliterated by subsequent judicial explanations and qualifications. …"

In Transco itself Lord Hoffmann did say at [39]:

"It is hard to escape the conclusion that the intellectual effort devoted to the rule by judges and writers over many years has brought forth a mouse."

Now we have another occasion to labour. How, if at all, does the rule apply where the damage to the claimant's land is caused by the "escape" of a fire which raged through the defendant's premises, its ferocity fed by the eventual ignition of the large stack of tyres which he had brought onto his land on which he carried out the business of a motor vehicle tyre supplier and fitter. On 22nd August 2011 Mr Recorder Potts sitting in the Worcester County Court held that all the requirements of Rylands v Fletcher were satisfied and accordingly judgment for the claimant was entered with damages to be assessed. The defendant now appeals.

The facts in a little more detail

The defendant carried on business as Wyvern Tyres supplying, fitting and balancing car and van tyres. The business was conducted at Unit 111 and 107a on the Holmer Trading Estate in Hereford. In the front part of the premises was the ramp and other pieces of equipment necessary for changing and fitting the tyres, vulcanising them and balancing them. To the side were two small offices. Along the length of the building to the rear were some six specially constructed racks for the storage of tyres. The Recorder found that the defendant "squeezed stock" into the room, some tyres being "haphazardly and untidily above the racks as well as on the racks themselves" and others were "piled high in 'chimneys'". The claimant's premises were behind the defendant's. Between them was a space used as a further area for the storage of new and part-worn tyres. There were no racks in this area and the tyres were simply piled high vertically in these "chimneys", some of them being located along the wall which divided the claimant's premises from the defendant's. It was difficult to know how many tyres were held by the defendant but the Recorder was satisfied that at the material time about 3,000 tyres were stored in and outside the building.

Having heard the evidence of experienced former fire officers, the Recorder concluded that tyres have recognised particular characteristics so far as concerns fires and fire fighting as follows:

"(a) They are not in themselves flammable, and in their normal state will not ignite unless there is sufficient flame or heat from another source.

(b) Once however a primary fire has developed, and intensified it can produce sufficient heat or flame to ignite rubber composite tyres.

(c) If tyres catch fire then combustion develops rapidly depending on the quantity of tyres present and how they are stored.

(d) Once fire takes hold of tyres they are difficult to put out."

A fire did break out at about 6.15 pm on 4th February 2008 some time after the defendant had secured the premises, set the alarms at 4.50 pm and left the site. It seems that the fire had started in the front workshop section and had quickly intensified. Such was its severity that ten pumping fire appliances attended at the scene with three other special vehicles, a command unit, Ariel ladder platform and water carrier. Once the fire took hold its intensity grew rapidly. By 19.55, just over an hour after the original attendance, the entire Wyvern Tyres unit was fully ablaze; it was already unlikely that the claimant's premises could be saved and priority was given, in vain it seems, to saving the adjoining units. Alas they were also totally destroyed and for safety reasons were reduced to rubble the following day. The Recorder made these findings:

"4. I am … satisfied on the balance of probabilities that the fire originated in Wyvern's premises, and probably towards the front of the workshop area. From there it spread back into the workshop and tyre storage area, and to the side into the office and reception, and thence to the unit next door occupied by the double glazing business, and to the rear into Mr Gore's (the claimant's) premises which adjoined directly onto Wyvern's premises, the two being separated by what appears from Mr Gore's plan to be simply a single skin of brick."

The Recorder accepted the evidence of the fire officer who attended the fire that the cause of the fire was "something occurring in the wiring and appliances within Wyvern's premises". Thus he accepted that the primary source was "electrical".

The Recorder asked whether the tyres made a difference. He held:

"15. In my judgment on the balance of probabilities they did. Something that stands out in this case in the descriptions of the fire by the witnesses, including fire officers, is its intensity and severity. … This was a fire which spread with great rapidity and intensity. Indeed it was so aggressive that the Fire Service plainly had difficulty in bringing it under control, and could only do so some hours after they had first been called out and attended and after they had been reinforced with further equipment and personnel.

16. While it is clear that tyres per se are not readily combustible and will not constitute a primary source of combustion, it is plain that they can catch fire from some other source as I have found to have existed in this case. Once the tyres were alight there was a huge problem for all concerned, including adjoining and neighbouring occupiers, because in the words of Mr Denton [the fire officer], "Once a fire takes hold of tyres they are difficult to put out". That is what, on the balance of probabilities, I find happened in the instant case with consequences that are of course only too apparent."

The judgment under appeal

The claim was brought both in negligence and in strict liability. The claim in negligence was rejected. The Recorder found, first, that the primary cause of the fire lay in the wiring or electrical appliances in Wyvern's premises but there was nothing to show that such a state of affairs was the result of a failure to maintain or keep in good order the electrical system itself or all those electrical appliances that were located within the premises, as opposed to something that might have arisen entirely by accident. He was satisfied that the machinery was regularly checked and that the electrical work had been competently carried out and checked a reasonable time before this fire started. Secondly he found that it was not reasonable to expect the defendant to invest in an automatic alarm system or sprinkler system and in any event even if he should have done so, there was no evidence to suggest that such systems would have been likely to have prevented the fire from spreading to the claimant's premises. The third allegation of storing the tyres without any adequate measures in place to prevent them catching fire or to minimise the chances of them catching fire was so vague and so lacking in specificity as to make it in practice impossible even to respond to it by way of defence other than by bare denial. There was no allegation and no evidence at all of what the defendant should or could have done to prevent the spread of fire that had undeniably broken out from reaching the tyres and thereby involving them in the conflagration. The claimant failed to prove that the defendant could have done something in terms of the storage of the tyres that would have prevented their igniting at all or would have reduced the chances of them doing so. Any complaint of a breach of the Regulatory Reform (Fire Safety) Order 2005 for the failure to take such general fire precautions as would ensure so far as was reasonably practicable the safety of any of the defendant's employees could not take the claimant's case further because there were no employees and in any event nothing in that statutory duty conferred any right of action in civil proceedings. The Recorder accordingly held that:

"… subject to the question of any Rylands v Fletcher liability, the failure of Mr Gore to establish to the satisfaction of the court any negligence on the part of Mr Stannard means that, apart from any Rylands v. Fletcher liability, he has the benefit of a defence under section 86 of the Fires Prevention (Metropolis) Act 1774 on the basis that the fire was accidental."

Strict liability under Rylands v Fletcher

In a careful judgment, the Recorder analysed the authorities presented to him. Dealing specifically with fires in connection with the rule in Rylands v Fletcher he observed how the Fires Prevention (Metropolis) Act 1774 limited a defendant's liability to non-accidental fires but he added:

"Even then of course where a fire arose from something dangerous that the defendant had brought onto his land there could be nothing accidental about any fire that arose as a result, and what became the rule in Rylands v Fletcher continued to apply with full force, although of course it is the fire itself that is the dangerous thing that escapes the defendant's land, rather than whatever caused the fire to arise in the first place."

As for Rylands v Fletcher he noted:

"The following well-known requirements to the rule in order that liability may be established:

(i) the defendant must bring onto his land something that is dangerous;

(ii) the danger must escape from the defendant's land to the claimant's land; and

(iii) the use to which the defendant had put his land must be "non-natural"."

He concluded that "there was plainly an escape within the meaning of the Rylands v. Fletcher rule" and "liability therefore turns on whether or not Mr Stannard's activities on Wyvern's premises were dangerous and a non-natural use of his land." The only relevant activity was the storage of tyres. They were not in themselves flammable and would not ignite unless there was a sufficient flame or heat source. It is, however, not impossible for tyres to catch fire as obviously happened here and if they do ignite they have a special fire risk quality. This is that once alight they may burn rapidly and intensively such that they are difficult to put out. Given the haphazard way a large number of tyres were stored there was an exceptionally high risk of damage to the claimant's premises if fire broke out. That was a foreseeable risk with the result that the defendant's activities in storing tyres in the numbers and ways that he did were dangerous within the meaning of the rule. As for non-natural use, the Recorder asked himself whether what the defendant did in storing and enlarging the tyre storage area for that purpose was out of the ordinary. Here the state of affairs created by the defendant was out of the ordinary. It was not normal and not routine, the standard of normality and routine being that of a tyre business storing tyres in an orderly fashion and in such numbers that are well within the capacity of its dedicated storage facilities. This was non-natural use. Consequently the requirements of the rule in Rylands v. Fletcher were established and judgment was entered for the claimant accordingly.

Discussion: the only question in this appeal

The appellant does not seek to challenge the findings of fact nor the dismissal of the claim in negligence. It is, however, submitted on the appellant's behalf that the Recorder erred in his application of the test for strict liability under the rule in Rylands v. Fletcher as applied to fire cases. I propose to deal first with what seems to me to be the established formulation of the rule in classic Rylands v. Fletcher cases, leaving for later consideration the question of whether there is a special or different rule of a Rylands v. Fletcher kind to deal with damage caused by fire.

Classic Rylands v. Fletcher

I need not repeat Blackburn J's formulation and the gloss added to it by Lord Cairns as set out at the beginning of this judgment. The first significant clarification was made nearly 40 years later in Rickards v Lothian [1913] AC 263, 280. There the Privy Council was concerned with property on the second floor of the premises which were damaged by the continuous overflow of water from a lavatory basin on the top floor caused by the water tank having been turned on full and the waste-pipe plugged through the malicious act of some person. The Privy Council were of the opinion that the case did not come within the principle of Rylands v Fletcher because the matters complained of took place through no fault or breach of duty on the defendants' part but were caused by a stranger over whom and at a spot where they had no control. Lord Moulton added at p. 279:

"But there is another ground upon which their Lordships are of opinion that the present case does not come within the principle laid down in Fletcher v. Rylands. It is not every use to which land is put that brings into play that principle. It must be some special use bringing with it increased danger to others, and must not merely be the ordinary use of the land or such a use as is proper for the general benefit of the community."

The Privy Council were of the opinion that the provision of a proper supply of water to the various parts of a house is not only reasonable, but has become, in accordance with prevailing sanitary views, an almost central feature of town life and it would be wholly unreasonable to hold an occupier responsible for the consequences of acts which he is powerless to prevent.

Rainham Chemical Works Ltd v Belvedere Fish Guano Company Ltd [1921] 2 AC 465 is often cited but I confess that I find the jurisprudence on Rylands v Fletcher is hardly advanced by this case. At a time of war a process was invented whereby picric acid could be manufactured from dinitrophenol (DNP) and nitrate of soda. DNP had been used mainly for the manufacture of dyes, and it was a stable compound which did not explode easily. It was not in itself dangerous. Nitrate of soda was not an explosive but wood or bags impregnated with moist nitrate of soda will, when dry, burn fiercely if ignited. A hot flame is needed to ignite it and when ignited, large quantities of DNP become a dangerous explosive. While neither DNP nor nitrate of soda was, in itself, dangerous, they became a source of danger if stored in quantities and in close proximity to one another. It was proved that it was due to that cause that a massive explosion took place which caused damage to neighbouring property. On the evidence the manufacture of picric acid from DNP and nitrate of soda might or might not be dangerous in its character, but in that case it was being manufactured under dangerous conditions, and those dangerous conditions caused the accident. Accordingly the principle of Rylands v. Fletcher became applicable. It was not, per Lord Carson, "seriously argued" that the defendant company was not liable for the damages caused by the explosion. Indeed it seems that before Scrutton LJ, the trial judge, it was admitted that the person in possession of the DNP was liable under the rule in Rylands v. Fletcher for the consequences of the explosion. Thus the real issue in the case was stated by Lord Buckmaster at p. 471 in these terms.

"Now, the foundation of the action was a claim based upon the familiar doctrine established by the case of Fletcher v. Rylands, which depends upon this - that even apart from negligence the use of land by one person in an exceptional manner that causes damage to another, and not necessarily an adjacent owner, is actionable: … In the present case the use complained of was that for the purpose of making munitions, which was certainly not the common and ordinary use of the land, two substances, namely, nitrate of soda and dinitrophenol, were stored in close proximity, with the result that on a fire breaking out they exploded with terrific violence. It may be accepted that it was not known to either of the defendants that this danger existed, but that in itself affords no excuse, and the result is that the plaintiffs' cause of action is well founded and the only matter for determination is against whom the action should be brought."

The disputed question was whether responsibility lay at the door of the defendant company or the personal defendants who had a licence from the inventor to manufacture the required picric acid. Whilst, therefore, that case bears some similarly to the case before us (large quantities of materials which when once alight will cause damage to neighbouring property), there is little help on the principles to apply.

Another explosion at a time of war gave rise to Read v. J. Lyons & Co Ltd [1947] AC 156. This was a claim by an inspector of munitions who was injured in the course of her duties while in a munitions factory operated by the respondents by the explosion of a high explosive shell during the process of manufacture. She failed because an essential condition of the application of the rule in Rylands v. Fletcher was the "escape" from the land of something likely to do mischief if it escaped. Here the injuries were caused on the premises of the defendant. Doubts were expressed, obiter, as to whether the action lay where the claim was for damages for personal injury as distinguished from damage to property. Once again, this case gives us little help.

Cambridge Water Co v. Eastern Counties Leather Plc [1994] 2 AC 264 is a much more important authority. There a chlorinated solvent used by the leather manufacturer seeped into the water company's borehole resulting in the water becoming unfit for human consumption. Although the water company had abandoned its claim in nuisance and could only succeed on the basis of the rule in Rylands v. Fletcher, Lord Goff of Chieveley nonetheless considered it desirable to look at the nature of liability in both nuisance and Rylands v. Fletcher.

Comparing nuisance and the rule in Rylands v. Fletcher he said at p. 299:

"We are not concerned in the present case with the problem of personal injuries, but we are concerned with the scope of liability in nuisance and in Rylands v. Fletcher. In my opinion it is right to take as our starting point the fact that, as Professor Newark considered [in his seminal article on "The boundaries of nuisance" (1949) 65 L.Q.R. 480], Rylands v. Fletcher was indeed not regarded by Blackburn J. as a revolutionary decision: see, e.g., his observations in Ross v. Fedden (1872) 26 L.T. 966, 968. He believed himself not to be creating new law, but to be stating existing law, on the basis of existing authority; and, as is apparent from his judgment, he was concerned in particular with the situation where the defendant collects things upon his land which are likely to do mischief if they escape, in which event the defendant will be strictly liable for damage resulting from any such escape. It follows that the essential basis of liability was the collection by the defendant of such things upon his land; and the consequence was a strict liability in the event of damage caused by their escape, even if the escape was an isolated event. Seen in its context, there is no reason to suppose that Blackburn J. intended to create a liability any more strict than that created by the law of nuisance; but even so he must have intended that, in the circumstances specified by him, there should be liability for damage resulting from an isolated escape."

He questioned whether foreseeability of damage was a prerequisite for recovery of damages and said at p.302:

"… Blackburn J spoke of "anything likely to do mischief if it escapes", and later he spoke of something "which he knows to be mischievous if it gets on his neighbour's [property]," and the liability to "answer for the natural and anticipated consequences." Furthermore, time and time again he spoke of the strict liability imposed upon the defendant as being that he must keep the thing in at his peril; and, when referring to liability in actions for damage occasioned by animals, he referred, at p.282, to the established principle that "it is quite immaterial whether the escape is by negligence or not." The general tenor of his statement of principle is therefore that knowledge, or at least the foreseeability of the risk, is a prerequisite of the recovery of damages under the principle; but that the principle is one of strict liability in the sense that the defendant may be held liable notwithstanding that he has exercised all due care to prevent the escape occurring."

And at p. 306 he concluded:

"… it appears to me to be appropriate now to take the view that foreseeability of damage of the relevant type should be regarded as a prerequisite of liability in damages under the rule."

In the result, since the plaintiffs were unable to establish that the pollution of their water supply by the solvent was foreseeable, the claim failed.

He considered that there was a striking similarity of function between the principle of reasonable user in nuisance (the principle of give and take as between neighbouring occupiers of land) and the principle of natural use in Rylands v Fletcher. As for the latter, he pointed out Lord Moulton's gloss in Rickards v Lothian and added at p. 308:

"Blackburn J.'s statement of the law was limited to things which are brought by the defendant onto his land, and so did not apply to things that were naturally upon the land. Furthermore, it is doubtful whether in the House of Lords in the same case Lord Cairns, to whom we owe the expression "non-natural use" of the land, was intending to expand the concept of natural use beyond that envisaged by Blackburn J. Even so, the law has long since departed from any such simple idea, redolent of a different age; and, at least since the advice of the Privy Council delivered by Lord Moulton in Rickards v. Lothian …, natural use has been extended to embrace the ordinary use of land. …

Rickards v. Lothian itself was concerned with a use of a domestic kind, viz. the overflow of water from a basin whose runaway had become blocked. But over the years the concept of natural use, in the sense of ordinary use, has been extended to embrace a wide variety of uses, including not only domestic uses but also recreational uses and even some industrial uses.

It is obvious that the expression "ordinary use of land" in Lord Moulton's statement of the law is one which is lacking in precision … A particular doubt is introduced by Lord Moulton's alternative criterion - "or such a use as is proper for the general benefit of the community." If these words are understood to refer to a local community, they can be given some content as intended to refer to such matters as, for example, the provision of services; indeed the same idea can, without too much difficulty, be extended to, for example, the provision of services to industrial premises, as in a business park or an industrial estate. But if the words are extended to embrace the wider interests of the local community or the general benefit of the community at large, it is difficult to see how the exception can be kept within reasonable bounds. … I myself, however, do not feel able to accept that the creation of employment as such, even in a small industrial complex, is sufficient of itself to establish a particular use as constituting a natural or ordinary use of land."

He did not, however, think it necessary to attempt any redefinition of the concept of natural or ordinary use because the storage and use of chemicals by the defendant was "an almost classic case of non-natural use", explaining at p. 309 that:

"It may well be that, now that it is recognised that foreseeability of harm of the relevant type is a prerequisite of liability in damages under the rule, the courts may feel less pressure to extend the concept of natural use to circumstances such as those in the present case; and in due course it may become easier to control this exception, and to ensure that it has a more recognisable basis of principle."

Of even greater importance for the understanding of the current state of the law is Transco Plc v Stockport MBC [2003] UKHL 61, [2004] 2 AC 1. Here water pipes laid by the council from their tower block of flats to the water main fractured and the resulting flood caused the collapse of an embankment on part of their land through which the claimant had the right to run its gas main. The collapse of the embankment caused damage to the gas main. Transco's main claim was that the council was liable without proof of negligence under the rule in Rylands v. Fletcher. It succeeded before the judge but was reversed by the Court of Appeal. Their Lordships were being asked to review and, if they thought it right to do so, to hold the rule in Rylands v. Fletcher no longer to be good law. They were being invited "to follow the trail blazed by the majority of the High Court of Australia in Burnie Port Authority (supra) by treating the rule in Rylands v Fletcher as absorbed by the principles of ordinary negligence." Although Lord Bingham of Cornhill saw that there was "a theoretical attraction in bringing this somewhat anomalous ground of liability within the broad and familiar rules governing liability in negligence" he was unwilling "to suppress an instinctive resistance to treating a nuisance-based tort as if it were governed by the law of negligence." So the rule remains the law of the land. But the speech of Lord Bingham is the latest and last word on the subject and bears close study.

Like Lord Goff he recognised that Blackburn J did not conceive himself to be laying down any new principle of law and at [3] he said:

"It seems likely, as persuasively contended by Professor Newark ("The Boundaries of Nuisance" (1949) 65 LQR 480, 487-488), that those who decided the case regarded it as one of nuisance, novel only to the extent that it sanctioned recovery where the interference by one occupier of land with the right or enjoyment of another was isolated and not persistent."

Lord Bingham favoured retaining the rule but restating it so as to achieve as much certainty and clarity as is attainable, recognising that new factual situations are bound to arise posing difficult questions on the boundary of the rule, wherever that is drawn. His restatement was set out in paragraphs [9]-[11] of his speech. I extract these important observations:

"9. The rule in Rylands v Fletcher is a sub-species of nuisance, which is itself a tort based on the interference by one occupier of land with the right in or enjoyment of land by another occupier of land as such. From this simple proposition two consequences at once flow. First, as very clearly decided by the House in Read v J Lyons & Co Ltd … no claim in nuisance or under the rule can arise if the events complained of take place wholly on the land of a single occupier. There must, in other words, be an escape from one tenement to another. Second, the claim cannot include a claim for death or personal injury, since such a claim does not relate to any right in or enjoyment of land. …

10. It has from the beginning been a necessary condition of liability under the rule in Rylands v Fletcher that the thing which the defendant has brought on his land should be "something which ... will naturally do mischief if it escape out of his land" (LR 1 Ex 265, 279 per Blackburn J), "something dangerous ...", "anything likely to do mischief if it escapes", "something ... harmless to others so long as it is confined to his own property, but which he knows to be mischievous if it gets on his neighbour's" (p 280), "anything which, if it should escape, may cause damage to his neighbour" (LR 3 HL 330, 340, per Lord Cranworth). The practical problem is of course to decide whether in any given case the thing which has escaped satisfies this mischief or danger test, a problem exacerbated by the fact that many things not ordinarily regarded as sources of mischief or danger may none the less be capable of proving to be such if they escape. I do not think this condition can be viewed in complete isolation from the non-natural user condition to which I shall shortly turn, but I think the cases decided by the House give a valuable pointer. … Bearing in mind the historical origin of the rule, and also that its effect is to impose liability in the absence of negligence for an isolated occurrence, I do not think the mischief or danger test should be at all easily satisfied. It must be shown that the defendant has done something which he recognised, or judged by the standards appropriate at the relevant place and time, he ought reasonably to have recognised, as giving rise to an exceptionally high risk of danger or mischief if there should be an escape, however unlikely an escape may have been thought to be.

11. No ingredient of Rylands v Fletcher liability has provoked more discussion than the requirement of Blackburn J (LR 1 Ex 265, 280) that the thing brought on to the defendant's land should be something "not naturally there", an expression elaborated by Lord Cairns (LR 3 HL 330, 339) when he referred to the putting of land to a "non-natural use" … Read literally, the expressions used by Blackburn J and Lord Cairns might be thought to exclude nothing which has reached the land otherwise than through operation of the laws of nature. But such an interpretation has been fairly described as "redolent of a different age" (Cambridge Water [1994] 2 AC 264, 308), and in Read v J Lyons & Co Ltd … and Cambridge Water, at p 308, the House gave its imprimatur to Lord Moulton's statement, giving the advice of the Privy Council in Rickards v Lothian …

"… It must be some special use bringing with it increased danger to others, and must not merely be the ordinary use of land …"

I think it clear that ordinary user is a preferable test to natural user, making it clear that the rule in Rylands v Fletcher is engaged only where the defendant's use is shown to be extraordinary and unusual. This is not a test to be inflexibly applied: a use may be extraordinary and unusual at one time or in one place but not so at another time or in another place (although I would question whether, even in wartime, the manufacture of explosives could ever be regarded as an ordinary user of land, as contemplated … in Read v J Lyons & Co Ltd …) I also doubt whether a test of reasonable user is helpful, since a user may well be quite out of the ordinary but not unreasonable, as was that of Rylands, Rainham Chemical Works or the tannery in Cambridge Water. Again, as it seems to me, the question is whether the defendant has done something which he recognises, or ought to recognise, as being quite out of the ordinary in the place and at the time when he does it. In answering that question, I respectfully think that little help is gained (and unnecessary confusion perhaps caused) by considering whether the use is proper for the general benefit of the community. … An occupier of land who can show that another occupier of land has brought or kept on his land an exceptionally dangerous or mischievous thing in extraordinary or unusual circumstances is in my opinion entitled to recover compensation from that occupier for any damage caused to his property interest by the escape of that thing, subject to defences of Act of God or of a stranger, without the need to prove negligence."

Lord Hoffmann agreed that the criterion of exceptional risk must be taken seriously and creates a high threshold for a claimant to surmount. Although giving reasons of their own, Lord Scott of Foscote and Lord Walker of Gestingthorpe agreed with Lord Bingham.

This is, therefore, the seminal authority for the test to be applied in a classic case of Rylands v Fletcher.

The proper approach I extract from those compelling authorities is this.

(1) The defendant must be the owner or occupier of land.

(2) He must bring or keep or collect an exceptionally dangerous or mischievous thing on his land.

(3) He must have recognised or ought reasonably to have recognised, judged by the standards appropriate at the relevant place and time, that there is an exceptionally high risk of danger or mischief if that thing should escape, however unlikely an escape may have been thought to be.

(4) His use of his land must, having regard to all the circumstances of time and place, be extraordinary and unusual.

(5) The thing must escape from his property into or onto the property of another.

(6) The escape must cause damage of a relevant kind to the rights and enjoyment of the claimant's land.

(7) Damages for death or personal injury are not recoverable.

(8) It is not necessary to establish the defendant's negligence but an Act of God or the act of a stranger will provide a defence.

If that is how the rule is to operate in any classic Rylands v Fletcher case, the next question is whether there is any other or some special rule for cases involving damage caused by the spread of fire.

Liability for damage caused by the spread of fire

Writing in the Cambridge Law Journal in 1931 that great master of the law of tort, Professor Percy Winfield, was driven to observe that there was a "somewhat confusing medley of remedies for the same wrong, which is traceable in the historical development of liability for the escape of fire," see P.H. Winfield: Nuisance as a Tort (1931) 4 C.L.J. 189. He concluded at p. 206:

"The result seems to be that, at the present day, damage resulting from the unintentional escape of fire may be redressed by any of the following remedies:

"(i) the old action of trespass on the case;

(ii) an action of the Rylands v Fletcher type: query whether this has totally absorbed (i);

(iii) an action for nuisance;

(iv) an action for negligence."

…"

This appeal concerns head (ii) only and although any views I may express about the others will be treated as obiter, it is necessary to have some overview of the whole confusing picture. I propose, however, to do no more than take a quick canter over what is undoubtedly tricky terrain. The bye-line to A.S. Ogus' article "Vagaries in Liability for the Escape of Fire", (1969) C.L.J. 104 says it all:

"How great a matter a little fire kindleth." (General epistle of James iii.5)"

The old action of trespass on the case

The custom of the realm is that a person is liable for the damage caused by the escape of his fire – the ignis suus rule. By general custom of the realm the remedy for damage caused by the spread of fire was an action on the case pur negligent garder son few. The declaration was in the form "whereas by the custom of England a man is bound to keep his fire ignem suum safe and secure lest by default of custody of it loss should fall upon his neighbours in any way, the defendant tam negligenter ac improvide apud custodivit, quod pro defectu debitae custodiae the fire spread and did damage …".

The essential features are that (1) fire spreading from a person's land is his fire; (2) that person is under a duty to contain the fire: (3) if it escapes, he is liable for the damage that results; (4) "negligenter" was a pleader's adverb: it did not mean negligently in the sense we now use the word. To quote Winfield again (Winfield: The Myth of Absolute Liability), (1926) 42 L.Q.R. 37, 49:

"Exactly what it did mean must remain a matter of conjecture but it excluded liability when the fire spread or occurred (i) by the act of a stranger – a man was not liable for that, though he was for the act of his servant, his wife, his guest, or one entering his house with his leave or knowledge; (ii) by misadventure which, to anticipate a modern term, seems to be equivalent to inevitable accident, or something which by no care reasonable in the circumstances could have been avoided."

The earliest case on the subject is Beaulieu v Finglam (1401) B. & M. 557 where, as reported by Fifoot History and Sources of the Common Law, 1949, Markham J. said:

"A man is held to answer for the act of his servant or of his guest in such a case; for if my servant or my guest puts a candle by a wall and the candle falls into the straw and burns all my house and the house of my neighbour also, in this case I shall answer to my neighbour for his damage.

…

I shall answer to my neighbour for each person who enters my house by my leave or my knowledge, or is my guest through me or through my servant, if he does any act, as with a candle or aught else, whereby my neighbour's house is burnt. But if a man from outside my house and against my will starts a fire in the thatch of my house or elsewhere, whereby my house is burned and my neighbours' houses are burned as well, for this I shall not be held bound to them; for this cannot be said to be done by wrong on my part, but is against my will."

Moving forward nearly three centuries to Turberville v Stamp (1702) 12 Mod. 152, a fire started by the defendant in his field was held to be as much ignis suus as one lit in the house. Sir John Holt C.J. said:

"Every man must so use his own as not to injure another. The law is general; the fire which a man makes in the fields is as much his fire as his fire in the house; if it is made on the ground, with his materials, and by his order; and he must at his peril take care that he does not, through his neglect, injure his neighbour: if he kindle it at a proper time and place, and the violence of the wind carry it into his neighbour's ground, and prejudice him, this is fit to be given in evidence. But now here it is found to have been by his negligence; and it is the same as if it had been his house."

These were times when fires were a scourge. The Great Fire of London in 1666 had had a devastating impact on crowded urban living. Something had to be done. In 1707 the Fires Prevention (Metropolis) Act (6 and 31) was passed "for the better preventing the mischiefs that may happen by fire". Provision was made for improved safety, for the control of fires and for penal sanctions on servants carelessly setting fire to houses. The Act was repealed but re-enacted, eventually as the Fires Prevention (Metropolis) Act 1774, section 86 providing that:

"No action, suit, or process whatsoever shall be had, maintained, or prosecuted against any person in whose house, chamber, stable, barn or other building, or on whose estate any fire shall accidentally begin, nor shall any recompense be made by such person for any damage suffered thereby, any law, usage or custom to the contrary notwithstanding."

It may be that, as Sir William Holdsworth in his History of England vol. XI p. 607 speculates, by this time lawyers were beginning to think it anomalous that a man should be liable for fire damage not caused by his negligence. It may have been that Parliament was simply resolving any doubt there was as to whether fire spreading from a person's house was ignis suus. Another explanation is implicit in Lord Tenterden's judgment in Becquet v MacCarthy [1831] 2 B. & A.D. 951, 958:

"… By the law of this country before it was altered by the statute of 6 an. c.31, s.6, if a fire began on a man's own premises, by which those of his neighbour were injured, the latter, in an action brought for such an injury would not be bound in the first instance to show how the fire began but the presumption would be (unless it were shown to have originated from some external cause) that it arose from the neglect of some person in the house."

Thus one can surmise that the purpose of this statute was to remove the presumption of negligence on the defendant's part and cast the burden of proving negligence on the plaintiff.

Curiously there seems to be little contemporaneous authority on what was meant by "accidentally". Lord Lyndhurst flirted with the idea that accidentally was used as the antithesis of wilfully: see his obiter remarks in Canterbury v A.-G. (1843) 1 Phil. 306. The more informed view is expressed by Lord Denman C.J. in Filliter v Phippard (1847) 11 QB 347, 357:

"It is true that in strictness, the word accidental may be employed in contradistinction to wilful, and so the same fire might both begin accidentally and be the result of negligence. But it may equally mean that a fire produced by mere chance or incapable of being traced to any cause, and so would stand opposed to negligence of either servant or masters. And, when we find it used in statutes which do not speak of wilful fires but make important provision with respect to such as are accidental, and consider how great a change in the law would be effected, and how great encouragement would be given to that carelessness of which masters may be guilty as well as servants, we must say that we think the plaintiff's construction [accidental as opposed to negligent] much the most reasonable of the two."

An action for negligence

By the nineteenth century a prototype of our modern claim in negligence was emerging. In Vaughan v Menlove (1837) 3 Bing N.C. 468 the defendant's haystack spontaneously combusted and it was alleged that he had "wrongfully negligently and improperly kept his haystack so that it became liable to ignite" and so be a danger to the claimant's property. The jury were left to consider the question of negligence and that direction was upheld. Tindal C.J. said:

"… But there is a rule of law which says you must so employ your own property as not to injure that of another; and according to that rule the defendant is liable for the consequences of its own neglect; and though the defendant did not himself light the fire yet mediately, he is as much the cause of it as if he had himself put a candle to the rick; for it is well known that hay will ferment and take fire if it be not carefully stacked."

This was the age of the railways and much litigation followed. In Aldridge v The Great Western Railway Company (1841) 3 Man & G 515, Tindal C.J. held that Mr Aldridge had to show some carelessness on the railway company's part for the emission of the sparks from their engine which set fire to his crop. Piggot v The Eastern Counties Railway Co (1846) 3 CB 229 was another case of sparks setting fire to the claimant's property. Tindal C.J. held that:

"The defendants are a company entrusted by the legislature with an agent of an extremely dangerous and unruly character for their own private and particular advantage. And the law requires of them that they shall, in the exercise of the rights and powers so conferred upon them, that on such precautions as may reasonably prevent damage to the property of third persons through or near which their railway passes. The evidence in this case was abundantly sufficient to show that the injury of which the plaintiff complains was caused by the emission of sparks, or particles of ignited coke, coming from one of the defendants' engines: and there was no proof of any precaution adopted by the company to avoid such a mischief. I therefore think the jury came to a right conclusion, in finding that the company was guilty of negligence, and that the injury complained of was the result of such negligence."

We are seeing the emergence of a claim for negligence, the culmination of which in the nineteenth century was the well-known case in 1883 of Heaven v Pender (1882-1883) 11 Q.B.D. 503 which had nothing to do with fire, but which is often thought to be the progenitor of the modern law.

Turning to more modern times, an example of the claim in negligence for fire damage is Balfour v Barty-King [1957] 1 Q.B. 496 where an independent contractor negligently used his blow lamp to thaw frozen pipes and set fire to the combustible materials nearby. The fire spread to the claimant's property and destroyed it. The Court of Appeal held that the defendant was responsible for the negligence of the independent contractor. In H&N Emanuel v GLC [1971] 2 All E.R. 835 the Court of Appeal held that an occupier was liable for the escape of fire caused by the negligence not only of his servant but also of his independent contractor and anyone else who was on his land with his leave and licence. The only occasion when the occupier would not be liable for negligence when the negligence is the negligence of a stranger.

An action in nuisance

The modern example of this is Spicer v Smee [1946] 1 All E.R. 489 where the plaintiff's bungalow was completely destroyed by a fire which originated in the defendant's bungalow owing to a defect in the electrical wiring. The wiring had been negligently installed by an independent contractor. Atkinson J. held that the state of wiring in the defendant's bungalow constituted a nuisance on her property for which she was liable. The Fire Prevention (Metropolis) Act, 1774, did not afford her a defence since the fire was due to the negligence or nuisance created by the defendant or those for whom he was responsible.

An action of a Rylands v Fletcher type

One simply cannot say that there is no liability for fire damage under the Rylands v Fletcher rule because the very man who established the rule said there was: see Jones v Festiniog Railway Co (1868) L.R. 3 Q.B. 733. Lush J. who was also a member of the Exchequer Chamber in Rylands v Fletcher agreed with him. This was another case where the plaintiff's haystack had been fired by sparks from a railway engine. Blackburn J. held at p. 736:

"The general rule of common law is correctly given in Fletcher v Rylands, that when a man brings or uses a thing of a dangerous nature on his own land, he must keep it in at his own peril; for he is liable for the consequences if it escapes and does injury to his neighbour. Here the defendants were using a locomotive engine with no express Parliamentary powers making lawful that use, and they are therefore at common law bound to keep the engine from doing injury, and if sparks escape and cause damage, the defendants are liable for the consequences though no actual negligence be shown on their part."

I can readily understand this decision as an example of Rylands v Fletcher. The dangerous thing which the defendant railway company brought onto their land was a steam engine which depended for its locomotion on the burning of coal, particles of which would be belched forth from its maw onto the haystack adjoining the railway line. Although the engine itself remained on the defendant's land the sparks, which were an essential part of the machine, escaped and the danger posed by such an escape was high and it was foreseeable.

It is not as easy to rationalise the decision in Musgrove v Pandelis [1919] 2 KB 43 which remains for me, as for others, a troubling case. The defendant kept a car in a garage below the rooms occupied by the plaintiff. The defendant's servant, one Coumis, set about moving the car which necessitated applying the pressure pump to the carburettor to start the engine. For some unexplained reason there was an explosion in the carburettor and the petrol in the carburettor took fire. Instead of immediately turning the tap off which, if done, would have led to the fire burning itself out harmlessly, Coumis delayed and the continued supply of petrol extended the fire until it enveloped the car and ultimately the plaintiff's premises and furniture.

It is far from clear how the claim was brought. From the report of the proceedings at first instance - see 1919 1 KB 314 - it would seem that the claim was brought to recover damages as a result of negligence. At p. 44/45 of the Court of Appeal hearing, we are told that:

"The action was tried before Lush J., who held that this enactment had no application to a case which fell, as in his view this case fell, within the principle of Rylands v Fletcher. Further he held that the fire which caused the damage did not begin accidentally, but as a result of the negligence of Coumis. He therefore gave judgment for the plaintiff."

Bankes L.J. after stating the facts, proceeded as follows at p. 46:

"The plaintiff brought this action alleging that the fire was the result of Coumis's negligence. The negligence finally relied on was that he did not instantly turn off the petrol tap and so stop the further flow of petrol into the carburettor … The defendant's main defence, apart from disputing the negligence, was founded on s. 86 of the Fires Prevention (Metropolis) Act 1774, and the argument has been chiefly directed to the construction of that enactment. Lush J. took the view that the statute did not apply at all; and I agree. He also held that if that view was not correct, the fire which caused the accident did not accidentally begin within the meaning of the Act and there also I agree."

If the claim was indeed a claim in negligence, then the finding that the defendant was vicariously liable for the negligence of his servant would be the end of the matter. The defence under s. 86 could not prevail in the light of that finding. There would be no need to consider Rylands v Fletcher and references to the rule could be treated as obiter. That would be one view of this case. Bankes L.J. went on however to consider whether s. 86 provided a defence where the case fell within Rylands v Fletcher. He said at p. 47:

"Rylands v Fletcher is merely an illustration of that old principle [that a man must so use his own property as not to injure that of others], and in my opinion Lush J. was right in saying that this case, if it falls within that principle, is not within the protection of the statute."

So Bankes L.J. continued:

"The question then, is whether this motor car, with its petrol tank full or partially filled with petrol, was a dangerous thing to bring into the garage within the principle of Rylands v Fletcher. … I agree with Lush J. that this motor car was dangerous within that principle. The defendant brought it or caused it to be brought upon his premises and he is responsible for the fire which resulted, and is not within the protection of the statute."

Warrington L.J. was of the same opinion. He held:

"If this motor car with the petrol in its tank was potentially dangerous, such as a man's own fire, then it was the defendant's duty to see that the potential danger did not become an actual danger causing damage to his neighbour. The Act of Geo. III is no protection against that liability."

Duke L.J. also agreed. He held:

"I do not see how this case can be taken out of the principle of Rylands v Fletcher … in the present case there is petrol which is easily convertible into an inflammable vapour; there was the apparatus for producing a spark, and added to those there was a person supposed to control the combustion but inexperienced and unequal to the task. Taken together the presence of the petrol, and the production of the inflammable gas, all those combustibles together with the inexperience of the person placed in charge of them, it is impossible to say that this is not an instance of the principle laid down by Blackburn J."

He agreed that the case was outside any possible protection of the Act of Geo. III.

Musgrove had not escaped judicial criticism. In Collingwood v Home & Colonial Stores [1936] 3 All E.R. 200, 205, Lord Wright M.R. said:

"It was sufficient ground for the particular decision that the fire, that is to say the substantial fire, which actually caused the damage, was not caused without negligence. The fire which caused the damage was that which flowed from the original innocuous fire spreading through the fault of the chauffeur to the petrol in the tank, and that was clearly due to an act of negligence, and therefore, the protection of the statute did not apply. That was enough for the decision of the case. … That having been laid down, the Court of Appeal proceeded to hold that the principle of Rylands v Fletcher would apply here. … Well, I certainly have no desire to criticise in any way the actual decision in that case so far as it is based on the view that the real and substantial and destructive fire was the result of negligence. I confess, however, I find some difficulty about the other ground on which the decision was based, though, if it were necessary, I should follow the ruling of the Court of Appeal and apply it here if the case came within the scope of that ruling. But I do not think it does."

Romer L.J. said at p. 208:

"I think at sometime it will be desirable if the House of Lords would consider the case of Musgrove v Pandelis, so far as the decision in the case was based upon Rylands v Fletcher. Of course the rule in Rylands v Fletcher, as it is well known, is a rule which relates to the escape from somebody's premises of a dangerous animal or thing brought by the owner upon those premises, and does not relate to a case like the present, or a case like the Court of Appeal had to deal with in Musgrove v Pandelis where there had been an escape of nothing from the defendant's premises … But what will have to be considered is whether Musgrove v Pandelis … can be supported seeing that the decision involves these two propositions, (i) that a motor car is – I am quoting from the judgment of Bankes L.J., a dangerous thing to bring into a garage and (ii) that the use of one's land for the purpose of erecting a garage and keeping a motor car there is not an ordinary and proper use of the land – two propositions which, but for that authority, I should myself respectfully have doubted."

And I respectfully agree with them.

In Balfour v Barty-King [1957] 1 Q.B. 496, the Court of Appeal did not think it necessary to consider the doctrine of Rylands v Fletcher as a separate head of liability though Lord Goddard C.J., giving the judgment of the Court, did add at p. 505:

"No doubt the doctrine of that case applies to fire, and is subject to the exception of the damage being caused by a stranger."

McKenna J. was clearly perplexed by Musgrove's case pointing out in Mason v Levy Auto Parts of England Ltd [1967] 2 Q.B. 530, 542 that the car had not escaped from the land neither had the petrol in it but he felt it his duty to follow the case unless it had been overruled or unless the principle did not apply to the facts of his case.

So we come back to Transco. This was not a fire case and the crucial question is whether, when their Lordships were re-stating the Rylands v Fletcher rule, they were leaving aside fire damage as some special category outside the general rule they were propounding. I do not see how that can be inferred. The House was being invited to follow the High Court of Australia in Burnie Port Authority and treat the rule in Rylands v Fletcher as absorbed by the principle of ordinary negligence. Burnie Port Authority was a fire case. The High Court of Australia necessarily had to consider the escape of fire as a case within Rylands v Fletcher. In the course of work being carried out to the Port Authority's premises by independent contractors employed by them, their welding activity resulted in sparks or molten metal falling on cartons containing highly inflammable polystyrene which caught fire. The ensuing conflagration damaged the claimant's premises and the adjoining premises. Their conclusion, as expressed by Mason C.J. giving the majority judgment at [43]:

"In these circumstances and subject only to the above-mentioned possible qualification in relation to liability in nuisance, the rule in Rylands v Fletcher with all its difficulties, uncertainties, qualifications and exceptions, should now be seen, for the purposes of the common law of this country, as absorbed by the principles of ordinary negligence."

The House of Lords must have been aware that fire damage could be thought to be within the scope of Rylands v Fletcher. Moreover, Musgrove was cited to their Lordships though only Lord Walker of Gestingthorpe referred to it in his speech saying:

"107. The majority in Burnie Port Authority v General Jones Pty Ltd 120 ALR 42 commented that the scope of the Rylands v Fletcher principle has been progressively restricted from within and without. Both those observations are correct up to a point, but the process has not been entirely one-way traffic. Since the middle of the 19th century many activities which were once regarded as unusually dangerous (such as running railways, which no longer use steam locomotives fuelled by coal manually shovelled into the firebox) have become commonplace. Other activities unknown in the 19th century (including all those connected with the internal combustion engine) have come on the scene, being regarded first as dangerous innovations (see Musgrove v Pandelis …) but now as basic necessities."

I am bound to conclude, therefore, that when their Lordships laid down their guidance for the application of Rylands v Fletcher, they did not exclude cases of the escape of fire and the principles they espoused should be applied in fire cases as well as in other more classic examples of escaping dangerous things.

My conclusions on fire and the rule in Rylands v Fletcher

I feel bound for the reasons I have given to hold that in an appropriate case damage caused by fire emanating from an adjoining property can fall within the Rylands v Fletcher rule. The appropriate case is likely to be very rare, as Lord Bingham said in Transco:

"[5] … Consideration of the reported English case law over the past 60 years suggests that few if any claimants have succeeded in reliance on the rule in Rylands v Fletcher alone."

Lord Hoffmann said:

"[39] … It is perhaps not surprising that counsel could not find a reported case since the second world war in which anyone had succeeded in a claim under the rule. It is hard to escape the conclusion that the intellectual effort devoted to the rule by judges and writers over many years has brought forth a mouse."

Cases of fire damage are likely to be very difficult to bring within the rule because (1) it is the "thing" which had been brought onto the land which must escape, not the fire which was started or increased by the "thing". (2) While fire may be a dangerous thing, the occasions when fire as such is brought onto the land may be limited to cases where the fire has been deliberately or negligently started by the occupier or one for whom he is responsible. Is this not a relic of the ignis suus rule? (3) In any event starting a fire on one's land may well be an ordinary use of the land.

As for Musgrove, I reluctantly cannot conclude that it was decided per incuriam. I might be prepared to say, if forced to do so, that since the core finding was of negligence, then the findings based on Rylands v Fletcher were obiter. I prefer, however, simply to treat Musgrove as a fact sensitive case. Liability must, as Lord Bingham has explained to us, always be judged in the light of prevailing conditions of time and place and Musgrove would most definitely not be decided today as it was nearly a century ago. Musgrove should, in my judgment, therefore simply be relegated to a footnote in the history of Rylands v Fletcher.

Fire cases and this appeal must be judged in accordance with the test to be derived from Transco which I have set out at [22] above. Applying those principles I reach the following conclusions:

(1) The "thing" brought onto Wyvern's premises was a large stock of tyres.

(2) Tyres, as such, are not exceptionally dangerous or mischievous.

(3) There is no evidence that Mr Stannard recognised nor ought he reasonably to have recognised that there was an exceptionally high risk of danger or mischief if the tyres, as such, should escape.

(4) The tyres did not escape. What escaped was the fire, the ferocity of which was stoked by the tyres which were burning on, and remained burning on, Wyvern's premises. The Recorder was wrong to conclude it was the escape of fire that brought the case within Rylands v Fletcher principles.

(5) In any event, keeping a stock of tyres on the premises of a tyre-fitting business, even a very large stock, was not for the time and place an extraordinary or unusual use of the land. Here again the Recorder erred.

(6) Therefore Rylands v Fletcher liability is not established and, no negligence having been proved, the claim must fail.

(7) The moral of the story is taken from the speech of Lord Hoffman : make sure you have insurance cover for losses occasioned by fire on your premises.

I would therefore allow the appeal and dismiss the claimant's claim.

Lord Justice Etherton:

I pay tribute to the careful and comprehensive judgment of the Recorder. I nevertheless agree that this appeal should be allowed. It raises important points, but I can express my views quite briefly.

Firstly, in the light of the comprehensive review of the Rylands v Fletcher principle in Transco plc v Stockport MBC [2003] UKHL 61, [2004] 2 AC 251, I do not consider that the facts of the present case satisfy the basic requirement of the Rylands v Fletcher principle that there must have been an escape of something which the defendant has brought onto his or her land. Secondly, if there is a different requirement for so-called "fire" cases, where what has escaped from the defendant's land is fire generated from something that the defendant has brought onto his or her land, the principle still does not apply in the present case because tyres are not easily set alight and so do not pose any inherent danger of catching or causing fire. Thirdly, in any event, Mr Stannard's use of his property was not a non-natural use of his land for the purposes of the Rylands v Fletcher principle.

The classic statement of the strict liability principle by the Court of Exchequer Chamber in Rylands v Fletcher itself (at (1866) LR 1 Ex 265, 279) (approved by the House on Lords on appeal at (1868) LR 3HL 330 with the addition of the requirement of non-natural user) requires an escape from the defendant's land of something which he or she has brought there and which is likely to cause harm to the claimant if it escapes:

"We think that the true rule of law is, that the person who for his own purposes brings on his lands and collects and keeps there anything likely to do mischief if it escapes, must keep it in at his peril, and, if he does not do so, is prima facie answerable for all the damage which is the natural consequence of its escape."

It is obvious that the principle of strict liability, as stated there, requires modification if it is to apply where the escape is not of something which the defendant has brought onto his or her land, but is fire which has spread from things which were brought onto the land, and the fire was not deliberately or negligently started by the defendant.

As the cases show, it is necessary to modify the classic statement of the Rylands v Fletcher principle in a radical way in order to encompass such a situation, but there is no clear consensus as to how the principle should be restated. In some of the fire cases it was regarded as sufficient that the defendant brought onto his land something which was (at the time) considered inherently dangerous, such as a train (Jones v The Festiniog Railway Company 1868) LR QB 733) or a car (Musgrove v Pandelis [1919] 2 KB 43). In other cases, such as Mason v Levy Auto Parts of England Ltd [1967] 2 QB 530, it was the particular danger presented by the flammable nature of the materials brought onto the land by the defendant which was regarded as critical. The point was addressed by MacKenna J in Mason at page 542 as follows:

"If, for the rule in Musgrove's case to apply, there need be no escape of anything brought onto the defendant's land, what must be proved against him? There is, it seems to me, a choice of alternatives. The first would require the plaintiff to prove (1) that the defendant had brought something onto his land likely to do mischief if it escaped; (2) that he had done so in the course of a non-natural user of the land; and (3) that the thing had ignited and that the fire had spread. The second would be to hold the defendant liable if (1) he brought onto his land things likely to catch fire, and kept them there in such conditions that if they did ignite the fire would be likely to spread to the plaintiff's land; (2) he did so in the course of some non-natural use; and (3) the things ignited and the fire spread. The second test is, I think, the more reasonable one. To make the likelihood of damage if the thing escapes a criterion of liability, when the thing has not in fact escaped but has caught fire, would not be very sensible."

That was the approach taken by HH Judge Coulson QC, sitting as a High Court Judge, in LMS International Limited v Styrene Packaging and Insulation Limited [2005] EWHC 2065 (TCC) 30 September 2005, in which he undertook an extensive analysis of the Rylands v Fletcher principle, with particular regard to the review by the House of Lords in Transco plc v Metropolitan Borough Council [2004] 2 AC1 and the fire cases. He said as follows at [33]:

"In cases concerned with fire, the rule in Rylands v Fletcher requires two things. First, the defendant must have brought onto his land things which were likely to cause and/or catch fire, and kept them in such a condition that, if they ignited, the fire would be likely to spread to the claimant's land. To put it another way, those things must represent a recognisable risk to the owners of the adjoining land. Secondly, the actions on the part of the defendant must arise from a non-natural user of the defendant's land …"

A different formulation is put forward by Mr Philip Rainey QC and Mr Nicholas Isaac on behalf of Mr Gore, namely that the relevant danger which must be established is that the things collected on the defendant's land are likely to cause or facilitate the escape of the fire. They approve the formulation of Judge Coulson in LMS International if his expression "cause fire" is understood as meaning "cause the fire which escapes and damages the neighbouring land". In other words, they say the emphasis should not be on the danger posed by the flammability of the materials brought onto the defendant's land but on the danger posed by the role of the material in the escape of the fire from the defendant's land.

The lack of clarity and consensus as to the proper formulation of the Rylands v Fletcher principle in fire cases does not merely reflect poorly on a legal principle of strict liability 146 years after its first published judicial articulation. It calls into question whether the principle applies at all to fire cases. The Rylands v Fletcher principle was comprehensively reviewed by the House of Lords in Transco. That was not a fire case, but I consider that the reasoning and analysis of the House of Lords leave no scope for the application of the principle on the facts of the present case where there has been no escape of anything brought onto the defendant's land, but the escape is of fire, which was not created by the defendant.

The House of Lords in Transco had all the national, international and academic tools for a thorough review of the Rylands v Fletcher principle. They decided as a matter of policy to retain the principle, while restating it to achieve as much certainty and clarity as possible. As Lord Bingham said:

"8 There remains a third option, which I would myself favour: to retain the rule, while insisting upon its essential nature and purpose; and to restate it so as to achieve as much certainty and clarity as is attainable, recognising that new factual situations are bound to arise posing difficult questions on the boundary of the rule, wherever that is drawn."

Lord Bingham, with whom Lord Scott and Lord Walker agreed, described the core aspects of the principle in paragraphs [10] and [11]. The latter paragraph addressed the issue of non-natural user. The former identified, at the heart of the principle, the exceptional danger or mischief if there is an escape from the defendant's land. In describing that aspect, Lord Bingham said as follows:

"10 It has from the beginning been a necessary condition of liability under the rule in Rylands v Fletcher that the thing which the defendant has brought on his land should be "something which ... will naturally do mischief if it escape out of his land" (LR 1 Ex 265, 279 per Blackburn J), "something dangerous ...", "anything likely to do mischief if it escapes", "something ... harmless to others so long as it is confined to his own property, but which he knows to be mischievous if it gets on his neighbour's" (p 280), "anything which, if it should escape, may cause damage to his neighbour" (LR 3 HL 330, 340, per Lord Cranworth). The practical problem is of course to decide whether in any given case the thing which has escaped satisfies this mischief or danger test, a problem exacerbated by the fact that many things not ordinarily regarded as sources of mischief or danger may none the less be capable of proving to be such if they escape. I do not think this condition can be viewed in complete isolation from the non-natural user condition to which I shall shortly turn, but I think the cases decided by the House give a valuable pointer. In Rylands v Fletcher itself the courts were dealing with what Lord Cranworth (LR 3 HL 330, 342) called "a large accumulated mass of water" stored up in a reservoir, and I have touched on the historical context of the decision in paragraph 3(3) above. Rainham Chemical Works [1921] 2 AC 465, 471, involved the storage of chemicals, for the purpose of making munitions, which "exploded with terrific violence". In Attorney General v Cory Bros & Co Ltd [1921] 1 AC 521, 525, 530, 534, 536, the landslide in question was of what counsel described as an "enormous mass of rubbish", some 500,000 tons of mineral waste tipped on a steep hillside. In Cambridge Water [1994] 2 AC 264 the industrial solvents being used by the tannery were bound to cause mischief in the event, unforeseen on the facts, that they percolated down to the water table. These cases are in sharp contrast with those arising out of escape from a domestic water supply (such as Carstairs v Taylor (1871) LR 6 Ex 217, Ross v Fedden (1872) 26 LT 966 or Anderson v Oppenheimer (1880) 5 QBD 602) which, although decided on other grounds, would seem to me to fail the mischief or danger test. Bearing in mind the historical origin of the rule, and also that its effect is to impose liability in the absence of negligence for an isolated occurrence, I do not think the mischief or danger test should be at all easily satisfied. It must be shown that the defendant has done something which he recognised, or judged by the standards appropriate at the relevant place and time, he ought reasonably to have recognised, as giving rise to an exceptionally high risk of danger or mischief if there should be an escape, however unlikely an escape may have been thought to be. "

Lord Bingham summarised the Rylands v Fletcher principle most succinctly in paragraph [11] as follows:

"An occupier of land who can show that another occupier of land has brought or kept on his land an exceptionally dangerous or mischievous thing in extraordinary or unusual circumstances is in my opinion entitled to recover compensation from that occupier for any damage caused to his property interest by the escape of that thing, subject to defences of Act of God or of a stranger, without the need to prove negligence."

Lord Hobhouse stated in paragraph [66] that "[t]he content of the rule has been clearly spelled out by Blackburn J [in Rylands v Fletcher]". Lord Scott similarly said in paragraph [75] that the classical exposition of the rule was to be found in Blackburn J's judgment.

The various analyses in Transco all endorse the original formulation of the principle in the Court of Exchequer Chamber and in the House of Lords in Rylands v Fletcher. They make clear that the relevant mischief or danger is not something inherent in what is brought onto the defendant's land, for the principle may apply where what is brought onto the land is in itself harmless (such as water). The speeches in Transco make clear that the relevant mischief or danger lies in what will happen if there is an escape from the defendant's land of what has been brought onto the land. Accordingly, they leave no scope for the formulations in the previous fire cases, which required, as a condition of liability, that there was some inherent danger (irrespective of escape) in what the defendant has brought onto his or her land, whether by virtue of flammability (eg. Mason) or otherwise (eg. Festiniog Railway Company and Musgrove).

I do not consider that Transco can properly be regarded as leaving entirely open the possibility of a radically different formulation of the principle in fire cases. While it is true that not all the fire cases were cited in the speeches or, apparently, in submissions in the House of Lords in Transco, at least one of the leading fire cases, Musgrove, was cited (and indeed mentioned by Lord Walker at [107]). In any event, it is clear that the House of Lords in Transco was undertaking a comprehensive and definitive review of the Rylands v Fletcher principle in order to provide legal certainty. Leaving out of account a significant area of activity, without even mentioning that it was the intention to do so, would have been entirely inconsistent with that objective.

In those circumstances, and in the light of the emphasis in Transco and elsewhere that the Rylands v Fletcher principle should not be expanded further, I can see no scope for the application of the principle in the present case where there was no escape of the tyres which caught fire and the fire itself was not created by Mr Stannard.

If I am wrong about that, both principle and the weight of authority support the conclusion that the Rylands v Fletcher principle only applies to fire cases, where the fire has not been created by the defendant, if the fire has spread from materials brought onto the defendant's land and those materials pose a particularly high risk because they are likely to cause or catch fire and, if they do, such fire will be likely to spread to the claimant's property. If I am wrong on my first point about Transco, that was a correct summary of the law by Judge Coulson in LMS International, following the approach of MacKenna J in Mason. I, therefore, reject the gloss put on Judge Coulson's summary on behalf of Mr Gore. In a standard non-fire Rylands v Fletcher case the culpability of the defendant is the act of bringing something onto his or her land which it is foreseeable poses an exceptional risk of harm to the claimant if it escapes. The foreseeable risk is causally related to (1) the defendant's act in bringing something onto his or her land, and (2) the likely consequences if that same thing escapes. In a fire case, where the fire has not been created by the defendant, the culpability must, by analogy, be the act of the defendant in bringing onto his or her land something which poses a foreseeable risk to the defendant involving the creation of fire and its escape. There must be a causal connection between the defendant's act in bringing something onto his or her land and the foreseeable risk of harm to the defendant in the event of an escape of fire. That causal connection can only be the inherent likelihood that what the defendant has brought onto his or her land will catch or cause fire and the fire will spread to the claimant's property. In the present case, the evidence was that tyres are not in themselves flammable and they are not easily set alight. The necessary link between the deliberate act of Mr Stannard (in bringing tyres onto his land) and the foreseeability of an exceptional risk of harm to Mr Gore in the event of an escape did not exist.

If that too is wrong, then I do not consider that it was open to the Recorder, in the light of the evidence, to conclude that Mr Stannard's use of his property was a non-natural use for the purposes of the Rylands v Fletcher principle. A non-natural use, in this context, is a use which is "extraordinary and unusual" and is a high threshold for a claimant to surmount: Transco at [11] (Lord Bingham) and [49] (Lord Hoffmann). It is a mixed question of fact and law. It would not be right to interfere with the Recorder's view unless it exceeded the generous ambit within which reasonable disagreement is possible. The Recorder reached his conclusion on non-natural use on the basis of the number of tyres relative to the size of Mr Stannard's premises and also the manner in which they were stored. That is, with respect, very far short of what would be required for a finding of extraordinary use such as to engage the Rylands v Fletcher principle. The commercial activity carried on by Mr Stannard as a motor vehicle tyre supplier was a perfectly ordinary and reasonable activity to be carried on in a light industrial estate. There was no evidence that the number of tyres and the method of their storage was out of the ordinary for similar premises carrying on that type of activity or that they posed a foreseeable or recognised danger. Indeed, as has been pointed out on behalf of Mr Stannard, any other conclusion would be inconsistent with the Recorder's rejection of Mr Gore's case in negligence. In those circumstances, it was not open to the Recorder to reach the conclusion that he did on this aspect.

For the sake of completeness, if I were wrong on all those points, I would uphold the Recorder's judgment in favour of Mr Gore. I would reject, both as a matter of precedent and principle, Mr Stannard's argument that section 86 of the Fires Prevention (Metropolis) Act 1774 Act excludes liability under the Rylands v Fletcher principle. That point was directly addressed and decided by the Court of Appeal in Musgrove and so is binding on this Court. The alternative basis for the decision in that case was that the relevant fire, for the purposes of section 86 of the 1774 Act, was begun negligently and was outside the section for that reason. Nevertheless, the point that liability under Rylands v Fletcher is not exonerated by section 86 forms part of ratio of the Court of Appeal.

That also seems to have been the view of the Court of Appeal in Collingwood v Home and Colonial Stores Ltd [1936] 3 All ER 200. In that case Lord Wright MR and Romer LJ, with both of whom Macnaghten J agreed, questioned the correctness of the decision in Musgrove insofar as it held that there was liability under the Rylands v Fletcher principle on the facts of that case, but neither of them said that they also doubted the conclusion in Musgrove that Rylands v Fletcher liability falls outside section 86 of the 1774 Act.

In any event, that conclusion seems to me to be right in principle. It still remains a principle of statutory interpretation that, in cases of doubt, there is a presumption against interference with the common law: see generally Bennion on Statutory Interpretation (5^th ed) section 269; Craies on Legislation (9^th ed) 14.1.7 and 14.1.11. It is generally accepted that the word "accidentally" in section 86 cannot be read without some qualification and that it does not, for example, embrace a fire begun or continued negligently: Filliter v Phippart (1847) 11 QB 437 at 357, Balfour v Barty-King [1957] 1 QB 496, Goldman v Hargrave [1967] 1 AC 645 (PC). In this limited context, in view of the classical declaratory theory of the common law (viz. judges merely expound and declare the existing law rather than create it), I do not consider that a particular problem to this approach is posed by the fact that Rylands v Fletcher was not decided until 1866, just as it has not been regarded as a problem for excluding negligently caused fires from section 86 of the 1774 Act that it was not until the nineteenth century that negligence began to emerge as the substantive allegation in an action on the case. In any event, there are perfectly sound and rational grounds to conclude that section 86 was enacted to clarify some limited areas of uncertainty in the customary law of ignis suus.

The factual and legal background to the enactment of section 86 was examined by A.I. Ogus in the excellent and well-known scholarly article "Vagaries in Liability for The Escape of Fire" [1969] CLJ 104. He there set out three hypothetical situations in relation to ignis suus. He concludes that section 86 was enacted to resolve the uncertainty in the second and third of those situations, namely: (II) where the defendant has lit a candle in his house, and an exceptionally strong wind, or the act of a stranger, knocks the candle over setting the house alight, and the fire then spreads to the plaintiff's land; and (III) where the defendant has lit a bonfire in his field, and, as a result of an exceptionally high wind or the act of a stranger, sparks are blown onto a pile of leaves, which ignite, the fire spreading to the neighbour's property. I see no reason to disagree with Mr Ogus' analysis that the only object of section 86 of the 1774 Act (and the equivalent provisions in its statutory predecessor (1707) 6 Anne c.31 x.6) was to resolve the doubts as to whether the defendant would be liable in those two particular situations.

Lord Justice Lewison:

I agree with Lord Justice Ward that the appeal must be allowed. However, I would go further in limiting the scope of strict liability in relation to fire. Let me explain.

We have all heard of the Great Fire of London in 1666. But it was one in a long series of devastating fires which were common in mediaeval and Tudor England. London was ravaged by fire many times during the twelfth and thirteenth centuries. Building codes dealing with fire prevention were promulgated in London as early as 1189. William Fitzstephen, writing in the twelfth century, wrote that "the only plagues of London are the immoderate drinking of fools and the frequency of fires."

Fire was an equally serious hazard in other towns. The loss had to fall somewhere. These were the days before insurance. The custom of the common law of England laid the loss at the door of the person on whose property the fire had started. In Beaulieu v Finglam (1401) YB 2 Hen 4 f 18 pl 6 the custom was pleaded thus: "quilibet de eodem regno ignem suum salvo et secure custodiat et custodire teneatur ne per ignem suum dampnum aliquod vicinis suis ullo modo eveniat". The plea went on to allege that "the defendant "ignem suum improvide et negligenter custodivit". The custom is translated in Baker & Milsom Sources of English Legal History: Private Law to 1750 p. 610 as follows: "everyone in the same realm should keep and is bound to keep his fire safely and securely so that no damage befall any of his neighbours in any wise through his fire". The plea was that the defendant kept his fire "so negligently that for the want of due keeping of the aforesaid fire" the plaintiff's house was burned. Thirning CJ said "that a man should answer for his fire which by misfortune burns another's goods". Markham J said:

"A man is bound to answer for his servant's act, as for his lodger's act, in such a case. For if my servant or lodger puts a candle on a wall and the candle falls into the straw and burns the whole house and also my neighbour's house, in this case I shall answer to my neighbour for the damage which he has suffered."

He added:

"I shall answer to my neighbour for anyone who enters my house by my leave or with my knowledge, or is my guest or my servant's guest, if he does something with a candle (or whatever) by which my neighbour's house is burned. But if a man outside my house, against my will, sets fire to the thatch of my house or elsewhere, so that my house is burned and my neighbours' houses are burned as well, I shall not be held bound to answer to them for it; since it cannot be said to be ill-doing on my part when it is against my will."

In Crogate v Morris (1617) 1 B & G 197 the principle was pithily stated: "if my friend come and lie in my house, and set my neighbour's house on fire, the action lieth against me." Thus Rolle's Abridgment was able to state:

"If fire (I know nothing of it) suddenly break out in my house and burn my goods, and also the house of my neighbour he shall have an action on the case against me."

It is noticeable that the custom is that a man should keep "his fire" (ignis suus) safe and secure; and that the judges' discussion is all about fires deliberately kindled (even in the case of a candle which is deliberately lit). The plea was that the fire had been kept "negligently". It is clear that "negligently" was not used in the way in which that adverb is understood in the modern law of tort. What it signified was that the defendant had failed to observe his legal duty to prevent his fire escaping and damaging others. Some of the subsequent cases are concerned with what counted as a man's fire or "ignis suus". That the custom was concerned with fires that had been deliberately kindled is borne out by the statement of the law in Comyn's Digest: Action on the Case for Negligence (A6):

"An action lies, upon the general custom of the realm, against the master of a house if a fire be kindled there and consume the goods of another."

In a case of 1582 (Anon (1582) Cro Eliz 10) the defendant fired a gun at a fowl. In so doing he set fire to his own and his neighbour's house. The action was brought by way of action on the case. The court was of the view if the plaintiff "had counted on the custom of the realm as in [Beaulieu v Finglam] the action had not been well brought". Although the court did not give reasons for this view, it was presumably because the fire was not deliberately kindled and was never within the defendant's control (i.e. it was not his fire or "ignis suus").

In Turberville v Stamp (1702) 12 Mod 152 the real question was whether the defendant was liable for a fire deliberately kindled in a field; in other words whether such a fire counted as his own ("ignis suus"). Sir John Holt justified the principle as follows:

"Every man must so use his own as not to injure another. The law is general; the fire which a man makes in the fields is as much his fire as his fire in his house; it is made on his ground, with his materials, and by his order; and he must at his peril take care that it does not, through his neglect, injure his neighbour: if he kindle it at a proper time and place, and the violence of the wind carry it into his neighbour's ground, and prejudice him, this is fit to be given in evidence. But now here it is found to have been by his negligence; and it is the same as if it had been in his house."

In a different report of the case in (1702) 1 Comyns 32 the judgment is summarised thus:

"The case was afterwards adjudged in favour of the plaintiff by the whole Court; for the action is as well for a fire kindled in the fields of the defendant as in his house, for it is the defendant's fire and kindled in his ground, and he ought to have the same care of a fire which he kindles in his field as of that which is made in his house, for the duty to take care of both is founded upon this maxim, sic utere tuo ut non lædas alienum; but if the fire of the defendant by inevitable accident, by impetuous and sudden winds, and without the negligence of the defendant or his servants, (for whom he ought to be answerable) did set fire to the clothes of the plaintiff in his ground adjoining; the defendant shall have the advantage of this in evidence, and ought to be found not guilty. But here the verdict hath found negligence in the defendant. Therefore judgment for the plaintiff."

The report in 1 Salk 13 is even more abbreviated. There are five important points to be made about this case. First, liability is based on the general principle that a man "must so use his own as not to injure another" ("sic utere tuo ut alienum non laedas"). Second, on the facts liability was established because of the finding of negligence. Third, liability would have been avoided if the fire had been kindled at a proper time and place. Thus fourth, it is at least possible that even the common law required a finding of negligence. But if that is wrong, then it is clear that a defendant had a good defence if the fire was caused by inevitable accident or by act of a stranger. Fifth, the custom was directed towards a fire that a person had deliberately kindled, but which had subsequently got out of control. This, in my judgment, is implicit in the requirement to plead "ignis suus".

Even where liability is based on the principle sic utere tuo ut alienum non laedas a requirement of negligence is not necessarily excluded. In Black v Christchurch Finance Co Ltd [1894] AC 48 Lord Shand, delivering the advice of the Privy Council said:

"The lighting of a fire on open bush land, where it may readily spread to adjoining property and cause serious damage, is an operation necessarily attended with great danger, and a proprietor who executes such an operation is bound to use all reasonable precautions to prevent the fire extending to his neighbour's property (sic utere tuo at alienum non laedas)."

In other words even where liability was founded on (or at least justified by) the maxim, liability was not absolute, but required the taking of all reasonable precautions.

In H & N Emanuel Ltd v Greater London Council [1971] 2 All ER 835 (another case of a deliberately kindled fire) Lord Denning MR considered the scope of liability at common law. He said:

"After considering the cases, it is my opinion that the occupier of a house or land is liable for the escape of fire which is due to the negligence not only of his servants, but also of his independent contractors and of his guests, and of anyone who is there with his leave or licence. The only circumstances when the occupier is not liable for the negligence is when it is the negligence of a stranger. It was so held in a case in the Year Books 570 years ago, Beaulieu v Finglam, which is well translated by Mr Fifoot in his book on the History and Sources of the Common Law. The occupier is, therefore, liable for the negligence of an independent contractor, such as the man who comes in to repair the pipes and uses a blowlamp: see Balfour v Barty-King; and of a guest who negligently drops a lighted match: see Boulcott Golf Club Inc v Engelbrecht. The occupier is liable because he is the occupier and responsible in that capacity for those who come by his leave and licence: see Sturges v Hackett. But the occupier is not liable for the escape of fire which is not due to the negligence of anyone. Sir John Holt himself said in Tuberville v Stampe that if a man is properly burning up weeds or stubble and, owing to an unforeseen wind-storm, without negligence, the fire is carried into his neighbour's ground, he is not liable. Again, if a haystack is properly built at a safe distance, and yet bursts into flames by spontaneous combustion, without negligence, the occupier is not liable. That is to be inferred from Vaughan v Menlove. So also if a fire starts without negligence owing to an unknown defect in the electric wiring: Collingwood v Home and Colonial Stores Ltd; or a spark leaps out of the fireplace without negligence: Sochacki v Sas." (Emphasis added)

Whatever the precise basis for liability analysed in modern terms, historically it was simply based on custom of the realm. As Lord Denning put it in H & N Emanuel Ltd v Greater London Council:

"There has been much discussion about the exact legal basis of liability for fire. The liability of the occupier can be said to be a strict liability in this sense that he is liable for the negligence not only of his servants but also of independent contractors and, indeed, of anyone except a 'stranger'. By the same token it can be said to be a 'vicarious liability', because he is liable for the defaults of others as well as his own. It can also be said to be a liability under the principle of Rylands v Fletcher, because fire is undoubtedly a dangerous thing which is likely to do damage if it escapes. But I do not think it necessary to put it into any one of these three categories. It goes back to the time when no such categories were thought of. Suffice it to say that the extent of the liability is now well defined as I have stated it. The occupier is liable for the escape of fire which is due to the negligence of anyone other than a stranger." (Emphasis added)

This is an important point. Liability in mediaeval law (at least when it was concerned with liability by custom) did not depend on proving that the defendant was negligent in the modern sense. Nor did it depend on proving that he was engaged in any particularly dangerous activity. As Sir William Holdsworth explains (History of English Law vol VIII p. 469):

"The form which this stricter liability took was not the form taken by the rule in Rylands v Fletcher, for that rule was then the general rule of civil liability."

Apart from fire cases, others were also subject to strict liability by virtue of custom. One such category of people was the innkeeper who was liable by custom of the realm for the safe custody of the goods of his guests. This analogy was drawn on by counsel in Turberville v Stamp who is recorded as having argued:

"that this action ought not to be grounded upon the common customs of the realm; for this fire in the field cannot be called iguis suus, for a man has no power over a fire in the field, as he has over a fire in his house. And therefore this resembles the case of an inn-keeper, who must answer for any ill that happens to the goods of his guest, so long as they are in his house; but he is not answerable, if a horse be stolen out of his close."

It was against this background that the first of the statutes restricting liability for fire damage was passed. This was the Act of 1707, s.6, which was originally temporary and applied to London only. According to Lord Denman CJ in Filliter v Phippard (1847) 11 QB 347, a decision of the Court of King's Bench:

"The Act contemplates the probability of fires in cities and towns arising from three causes, the want of water, the imperfection of party walls, and the negligence of servants. The Act provided some means for supplying these material defects: but the third section was directed against the moral one, the carelessness or negligence of servants, which (it observes) often causes fires: and it imposes on the servant by whose negligence the fire may have been occasioned a fine of 100l., to be distributed among the sufferers at the discretion of the churchwardens, or imprisonment for eighteen months in case of nonpayment. … The most usual cause of fires was assumed to be the negligence of servants: and the enactment might operate to induce habits of caution in that important class. The same statute, in the sixth section, enacts that, after a day named, no action shall be maintained against any person in whose house or chamber any fire shall accidentally begin, nor shall any recompence be made by such person for any damage suffered or occasioned thereby."

This was then re-enacted in 1772 and 1774, the last of the enactments remaining in force as section 86 of the Fires Prevention Metropolis Act 1774, which applies to the whole of England and Wales: Richards v Easto (1846) 15 M & W 246, 251. Section 86 provides:

"And . . . no action, suit or process whatever shall be had, maintained or prosecuted against any person in whose house, chamber, stable, barn or other building, or on whose estate any fire shall . . . accidentally begin, nor shall any recompence be made by such person for any damage suffered thereby, any law, usage or custom to the contrary notwithstanding: . . . provided that no contract or agreement made between landlord and tenant shall be hereby defeated or made void."

It is important to emphasise the scope of the change effected by section 86. Not only does it apply notwithstanding any custom to the contrary (which would include liability under the ignis suus principle), it also applies notwithstanding any law to the contrary. This would, as a matter of ordinary statutory construction, include any other route at common law to liability: (compare McKillen v Maybourne Finance Ltd [2012] EWCA Civ 864).

Contemporaneous commentators (Blackstone Commentaries on the Laws of England (10th edn, 1787) p 431 and Bacon A New Abridgment of the Law (5th edn, 1789) p 85 thought that the Act had so altered the law that even where a fire had been caused or allowed to spread by negligence no liability arose. This treated the word "accidentally" as having been used in contrast to "wilfully". In Viscount Canterbury v Attorney-General (1843) 1 Phillips 306 Lord Lyndhurst LC at least flirted with this view, although it was not necessary for him to reach a decision.

Vaughan v Menlove (1837) 3 Bing NC 468 was a case of a haystack that spontaneously combusted. The allegation against Menlove was that he had "wrongfully, negligently and improperly" kept his haystack "so likely and liable to ignite and take fire, and in a state and condition dangerous" to Vaughan's cottages. It is important to note that this was not a case of a fire that had been deliberately kindled. At trial Patteson J directed the jury to consider whether the fire had been occasioned by "gross negligence" on the part of Menlove. The argument for Menlove on appeal was that the jury should have been asked to consider whether Menlove had "acted bona fide to the best of his judgment". The court rejected that argument and held that the case had been rightly left to the jury on the question of negligence. Liability thus depended on negligence.

Viscount Canterbury v Attorney-General (1843) 1 Phillips 306 was a petition of right brought against the Crown by a former Speaker of the House of Commons whose possessions had been destroyed in the fire that swept though the Houses of Parliament in 1834 (and to which we owe Barry and Pugin's building). The argument for the Speaker was that the statute did not relieve against fires caused by negligence, contrary to the opinion that Blackstone had expressed in his Commentaries. The Attorney does not appear to have argued this point; because he succeeded on the ground that a petition of right did not lie for negligence on the part of the Sovereign or his servants.

In Filliter v Phippard Lord Denman CJ confirmed that Blackstone's view was wrong. The question was: what did the word "accidentally" mean in the context of section 86? He said:

"It is true that in strictness, the word accidental may be employed in contradistinction to wilful, and so the same fire might both begin accidentally and be the result of negligence. But it may equally mean a fire produced by mere chance, or incapable of being traced to any cause, and so would stand opposed to the negligence of either servants or masters. And, when we find it used in statutes which do not speak of wilful fires but make an important provision with respect to such as are accidental, and consider how great a change in the law would be effected, and how great encouragement would be given to that carelessness of which masters may be guilty as well as servants, we must say that we think the plaintiff's construction much the most reasonable of the two."

Thus the court held that a fire that had begun by negligence did not attract the protection of section 86. In Burnie Port Authority v General Jones Pty Ltd (1994) 179 CLR 520 Mason CJ writing for the majority of the High Court of Australia described this decision as "surprising". But it is too late to change that.

The railway age brought a new spate of litigation. Early locomotives were steam powered; and the steam was generated by means of fires in the fire box of the locomotive, deliberately kindled. Sometimes, however, the fire in the locomotive could not be kept fully under control; or sparks from the fire were carried up the locomotive's chimney together with the steam.

In Aldrige v The Great Western Railway Co (1841) 3 Man & G 515 the allegation was that sparks from the company's railway engine set light to the plaintiff's beans because the company carelessly managed their engine. Tindal CJ said:

"It is contended on the part of the defendants, that the plaintiff should be nonsuited; but I am not prepared to say that the fact of the engine emitting sparks may not amount to negligence. On the other hand I cannot say that a verdict ought to be entered for the plaintiff. I think that the special case should be withdrawn, and that the parties should go on to trial. To entitle the plaintiff to recover, he must either shew some carelessness by the defendants, or lay facts before the jury from which it may be inferred."

Clearly liability was not strict but depended on negligence. In Piggot v The Eastern Counties Railway Company (1846) 3 CB 229 sparks from the engine of a passing mail train set fire to the plaintiff's cart lodge. The claim against the railway company was that they "so carelessly, negligently, and unskilfully managed and conducted their said steam-carriage and steam-engine" that the plaintiff's cart house was set on fire. In other words the claim was framed in negligence. The point in the case was the admissibility of evidence; but Tindal CJ again described the legal principle as follows:

"The defendants are a company intrusted by the legislature with an agent of an extremely dangerous and unruly character, for their own private and particular advantage: and the law requires of them that they shall, in the exercise of the rights and powers so conferred upon them, adopt such precautions as may reasonably prevent damage to the property of third persons through or near which their railway passes. The evidence in this case was abundantly sufficient to shew that the injury of which the plaintiff complains was caused by the emission of sparks, or particles of ignited coke, coming from one of the defendants' engines; and there was no proof of any precaution adopted by the company to avoid such a mischance. I therefore think the jury came to a right conclusion, in finding that the company were guilty of negligence, and that the injury complained of was the result of such negligence."

Thus although the locomotive was regarded as "dangerous", liability still turned on negligence. The underlying reason may be that the use of steam engines was authorised by statute, and that therefore negligence was the only route to liability, although this reason is not explicitly stated. The ground began to shift in Vaughan v The Taff Vale Railway Company (1858) 3 H & N 743. This was another case of fire caused by sparks from the fire of a passing locomotive. Although the railway company had done everything that was practicable to the locomotives to make them safe, the railway embankment was covered with inflammable grass. It is probable that it was that which caught fire and from there the fire spread to the plaintiff's wood. The argument for the railway company was that they should not be liable if they had taken all reasonable care to prevent the fire. It was also argued that if the fire had started on the embankment and had spread from there to the plaintiff's wood, section 86 protected them in the absence of negligence. As often happened in the nineteenth century the ultimate result turned on the pleadings. There were two counts alleged against the railway company. The first count alleged that they had negligently managed their locomotive and had failed to provide "the proper means for retaining the fire and igneous matter in the said locomotive whilst the same was being propelled along the railway". The second count alleged that the railway company knew that there was a danger that the combustible material on the bank would ignite; and that they therefore had a duty "to preserve and keep the bank in such a state and condition that fire should not be occasioned by reason of the ashes, &c., falling and settling thereon from and out of the locomotives, and to take all necessary precautions to prevent any fire which might be occasioned from extending to, and burning the wood of the plaintiff". The trial judge directed the jury that:

"… if, to serve his own purposes, a man does a dangerous thing, whether he takes precautions or not, and mischief ensues, he must bear the consequences: that running engines which cast forth sparks is a thing intrinsically dangerous, and that if a railway engine is used, which in spite of the utmost care and skill on the part of the Company and their servants is dangerous, the owners must pay for any damage occasioned thereby".

He also pointed out to the jury that the railway company could have kept the grass cut, or could have used gravel or sand to make a non-inflammable belt along the railway; and he asked them:

"… whether they did not think that there was inevitable negligence in the use of a dangerous thing calculated to do, and which did cause, mischief."

The jury found for the plaintiff; and the railway company appealed. The judgment of the Court of Exchequer was given by Bramwell B (who had also been the trial judge). He said:

"The first question then is: Was there evidence for the jury? And, as they may have found on either count, was there evidence in support of each? Next: Was the evidence such as to warrant the strong opinion of the learned Judge?

We are of opinion, on both these questions, in favour of the plaintiff. Here is confessedly the use of an instrument likely to produce damage, and producing it. This, according to general rules, would make the defendants liable. But two answers were suggested on their behalf. The first was, that if the fire originated on their own land they were protected by the 14 Geo. 3, c. 78, s. 86. But we are of opinion that the statute does not apply where the fire originates in the use of a dangerous instrument, knowingly used by the owner of the land in which the fire breaks out. It is impossible to suppose that the engine driver is liable to eighteen months imprisonment under section 84, and equally impossible to suppose there is no remedy against either master or servant, for what is a wrong by one or both. We are of opinion therefore that this answer fails."

One difficulty with this case is that the jury may have found that there was negligence on the part of the railway company (which went to the second count); in which case it would have been clear that section 86 would not have given the railway company a defence. But the reason that Bramwell B gave for the conclusion that the protection of the 1774 Act was excluded was that the fire was caused by the use of a dangerous instrument, knowingly used by the owner of the land. The dangerous instrument was the locomotive in which a fire was deliberately kindled and was burning in order to propel it. It is also noticeable that one of the ingredients of liability was that the instrument was "likely to produce damage". The judgement was reversed by Court of Exchequer Chamber (1860) 5 H & N 679 on the ground that the use of the locomotive was authorised by Parliament and therefore that so far as the first count was concerned negligence had to be established. Since it was not clear whether the jury had found for the plaintiff on the first count or the second count, there would have to be a new trial. The point under section 86 did not arise for decision.

In Smith v The London and South Western Railway Company (1869-70) LR 5 CP 98 workmen, employed by a railway company in cutting the grass and trimming the hedges bordering the railway, placed the trimmings in heaps near the line, and allowed them to remain there for fourteen days, during very hot weather in the month of August. Fire from a passing engine ignited one of these heaps, and burned the hedge, and was carried by a high wind across a stubblefield and a public road, and burned the plaintiff's goods in a cottage about 200 yards away. The question was whether there was evidence of negligence to go before the jury. No one argued that the railway company was strictly liable. Bovill CJ said:

"I agree that the mere circumstance of the fire being caused by an engine of the company, is not enough to give a cause of action against them; but the plaintiff must shew some breach of duty on their part which occasioned the injury he complains of."

This decision was upheld by the Court of Exchequer Chamber (1870-71) LR 6 CP 14, among whom was Blackburn J. But as in the case of Vaughan v The Taff Vale Railway Company the decision was based on the fact that the company had been authorised by statute to use steam locomotives on the railway. Blackburn J said:

"I take it that, since the case of Vaughan v Taff Vale Ry Co, which was expressly affirmed in Brand v Hammersmith Ry Co, it is clear that when a railway company is authorized by their Act of parliament to run engines on their line, and that cannot be done without their emitting sparks, the company are not responsible for injuries arising therefrom, unless there is some evidence of negligence on their part."

So we come to the famous case of Rylands v Fletcher (sub nom Fletcher v Rylands (1865-66) LR 1 Ex 265). In the Court of Exchequer Chamber the judges were referred in argument to section 86 of the 1774 Act and also to Filliter v Phippard. Neither found a place in the judgment. The well-known statement of principle formulated by Blackburn J was as follows:

"We think that the true rule of law is, that the person who for his own purposes brings on his lands and collects and keeps there anything likely to do mischief if it escapes, must keep it in at his peril, and, if he does not do so, is primâ facie answerable for all the damage which is the natural consequence of its escape. He can excuse himself by shewing that the escape was owing to the plaintiff's default; or perhaps that the escape was the consequence of vis major, or the act of God; but as nothing of this sort exists here, it is unnecessary to inquire what excuse would be sufficient. The general rule, as above stated, seems on principle just. The person whose grass or corn is eaten down by the escaping cattle of his neighbour, or whose mine is flooded by the water from his neighbour's reservoir, or whose cellar is invaded by the filth of his neighbour's privy, or whose habitation is made unhealthy by the fumes and noisome vapours of his neighbour's alkali works, is damnified without any fault of his own; and it seems but reasonable and just that the neighbour, who has brought something on his own property which was not naturally there, harmless to others so long as it is confined to his own property, but which he knows to be mischievous if it gets on his neighbour's, should be obliged to make good the damage which ensues if he does not succeed in confining it to his own property. But for his act in bringing it there no mischief could have accrued, and it seems but just that he should at his peril keep it there so that no mischief may accrue, or answer for the natural and anticipated consequences. And upon authority, this we think is established to be the law whether the things so brought be beasts, or water, or filth, or stenches."

It is, to my mind, striking that despite all the examples that Blackburn J gave fire is not among them. He had already decided in Smith v The London and South Western Railway Company that strict liability at common law could be displaced by statute. In the light of the fact that the court had been referred in argument to 86 of the 1774 Act and also to Filliter v Phippard, he must have taken the view that section 86 had displaced strict liability; and accepted counsel's submission that:

"… the liability for fire is restricted to mischief arising from negligence…"

Let it be supposed, however, that he thought that fire was a dangerous thing, and intended that the principle he formulated would apply to it. We can then restate the principle thus:

"a person who for his own purposes brings on his lands and collects and keeps there [fire] must keep it in at his peril, and, if he does not do so, is primâ facie answerable for all the damage which is the natural consequence of its escape."

The principle thus formulated is consistent only with a fire deliberately kindled. A fire that starts accidentally and which damages the place where it begins can hardly be said to have brought on to the land by the occupier "for his own purposes". Nor in my judgment can he be said without straining language to have brought the fire onto the land at all. In Read v J Lyons Ltd [1947] AC 156 Viscount Simon said of Blackburn J's formulation:

"Blackburn J several times refers to the defendant's duty as being the duty of "keeping a thing in" at the defendant's peril and by "keeping in" he does not mean preventing an explosive substance from exploding but preventing a thing which may inflict mischief from escaping from the area which the defendant occupies or controls."

As is well-known the case went to the House of Lords (sub nom Rylands v Fletcher (1868) LR 3 HL 330). Neither section 86 nor any fire case was referred to in the course of argument. Lord Cairns formulated the principle thus:

"On the other hand if the Defendants, not stopping at the natural use of their close, had desired to use it for any purpose which I may term a non-natural use, for the purpose of introducing into the close that which in its natural condition was not in or upon it, for the purpose of introducing water either above or below ground in quantities and in a manner not the result of any work or operation on or under the land,—and if in consequence of their doing so, or in consequence of any imperfection in the mode of their doing so, the water came to escape and to pass off into the close of the Plaintiff, then it appears to me that that which the Defendants were doing they were doing at their own peril; and, if in the course of their doing it, the evil arose to which I have referred, the evil, namely, of the escape of the water and its passing away to the close of the Plaintiff and injuring the Plaintiff, then for the consequence of that, in my opinion, the Defendants would be liable."

He then approved the formulation of Blackburn J in the court below. Lord Cranworth said:

"If a person brings, or accumulates, on his land anything which, if it should escape, may cause damage to his neighbour, he does so at his peril. If it does escape, and cause damage, he is responsible, however careful he may have been, and whatever precautions he may have taken to prevent the damage."

Neither of their Lordships referred to fire. Nor does the way in which the principle is framed readily apply to fire unless the fire itself has deliberately been brought on to the defendant's land. Lord Cairns referred to the use of the land "for the purpose of" introducing something dangerous. Likewise Lord Cranworth was clearly contemplating something dangerous deliberately brought onto or accumulated on the land. No consideration was given to how this principle might sit with section 86 of the 1774 Act.

Jones v Festiniog Railway (1867-68) LR 3 QB 733 was decided after Fletcher v Rylands in the Court of Exchequer Chamber and only a week or two before the appeal was heard in the House of Lords. The railway company ran steam locomotives on its railway. Although it had taken all reasonable precautions against the emission of sparks from the engine, nevertheless sparks from the engine set the plaintiff's haystack alight and burned down his barn. Blackburn and Lush JJ held that he was entitled to succeed. Blackburn J said:

"The general rule of common law is correctly given in Fletcher v Rylands, that when a man brings or uses a thing of a dangerous nature on his own land, he must keep it in at his own peril; and is liable for the consequences if it escapes and does injury to his neighbour. Here the defendants were using a locomotive engine with no express parliamentary powers making lawful that use, and they are therefore at common law bound to keep the engines from doing injury, and if the sparks escape and cause damage, the defendants are liable for the consequences, though no actual negligence be shewn on their part."

The thing of the dangerous nature that the railway company had brought onto its land was the locomotive engine with the deliberately kindled fire. Lush J said:

"I can see nothing in this statute to licence the company to use locomotive engines. In the absence of this licence the company are left to their liabilities at common law: that is, if they use a highly dangerous machine, they must do so at the peril of the consequences if it cause injury to others."

Again it is clear that the highly dangerous machine was the locomotive, and it was that dangerous thing that the railway company had deliberately brought onto its land. No argument was addressed to the court about the possible effect of section 86. Jones v Festiniog Railway was followed and applied in Powell v Fall (1879-80) LR 5 QBD 597, a case of a steam powered traction engine on the highway. Sparks from the engine set fire to the plaintiff's haystack. Since the fire started on the highway, section 86 could not have applied. The question for the court was:

"whether the owner of a locomotive engine propelled by steam along a public highway using a fire for the purpose of generating the steam required to propel such engine, … and was managed and conducted with all reasonable care and without negligence, was liable to the plaintiffs for injury occasioning damage to a rick of hay standing on land adjoining the highway by sparks proceeding from such engine and firing the hay." (Emphasis added)

Mellor J answered that question in the affirmative and his decision was upheld by the Court of Appeal. In the Court of Appeal it appears to have been conceded that an action lay at common law, and the only question was whether there was a statutory exemption. As Bramwell LJ put it:

"The plaintiffs are protected by the common law, and nothing adverse to their right to sue can be drawn from the statutes: the statutes do not make it lawful to damage property without paying for the injury."

Modern formulations of the principle in Rylands v Fletcher also emphasise the voluntary act of the defendant in bringing the dangerous thing onto his land. They also emphasise that it is that dangerous thing that must escape before strict liability is engaged. In Transco plc v Stockport MBC [2003] UKHL 61 [2004] 2 AC 1, 12 Lord Bingham ended his discussion by saying:

"An occupier of land who can show that another occupier of land has brought or kept on his land an exceptionally dangerous or mischievous thing in extraordinary or unusual circumstances is in my opinion entitled to recover compensation from that occupier for any damage caused to his property interest by the escape of that thing, subject to defences of Act of God or of a stranger, without the need to prove negligence."

Mr Rainey relied on the decision of the House of Lords in Rainham Chemical Works Ltd v Belvedere Fish Guano Co Ltd [1921] 2 AC 465. The case concerned an explosion in a munitions factory. The facts are by no means clear, but what seems to have happened is that a fire broke out in the factory and ignited sodium nitrate stored there. The sodium nitrate in turn produced a hot flame which caused dinitrophenol to explode. The claim was tried by Scrutton LJ who held that liability under the principle in Rylands v Fletcher had been established. His decision was upheld by both the Court of Appeal and the House of Lords. However, what was argued on appeal was not, in substance, the question of liability, but whether liability attached to the directors of the company personally. In Read v J Lyons Ltd Viscount Simon (169), Lord Macmillan (174) and Lord Uthwatt (187) all pointed this out; and Viscount Simon also said that liability had been conceded before Scrutton LJ sitting at first instance as the trial judge. In Cambridge Water Co v Eastern Counties Leather Ltd [1994] AC 264, 303-304 for much the same reasons Lord Goff also regarded Rainham Chemical Works as "fragile authority" on the scope of the principle. On the other hand in Transco plc v Stockport MBC Rainham Chemical Works Ltd v Belvedere Fish Guano Co Ltd was regarded by the Law Lords as being a case that demonstrated the utility of retaining the principle in Rylands v Fletcher, although the implications of that case were not discussed. Since there was no discussion of the principle either in the Court of Appeal or in the House of Lords in Rainham Chemical Works Ltd v Belvedere Fish Guano Co Ltd I do not consider that it lays down any binding principle of law about the scope of the principle in Rylands v Fletcher, even though it may be regarded as a case which justifies retention of the principle. The manufacture of explosives may well be considered to be an inherently dangerous activity (even in wartime) sufficient to bring the principle into play. Thus the ingredients themselves are the dangerous things which have been deliberately brought onto the land. Section 86 cannot apply, because it applies only to fires; not to explosions.

I come now to Musgrove v Pandelis [1919] 2 KB 43. Mr Musgrove rented rooms above a domestic garage, in which Mr Pandelis kept a car. Mr Pandelis sent his chauffeur, Mr Coumis, to clean the car. Mr Coumis had to move the car within the garage. For that purpose he went to the bonnet and turned on the petrol tap to allow the flow of petrol from the tank to the carburettor, and started the engine, when suddenly there was an explosion, and flames were seen to be coming from the carburettor. There was no woodwork within eighteen inches of the carburettor, and if Mr Coumis had immediately turned off the tap of the pipe leading from the petrol tank the petrol in the carburettor would have soon burnt out, and the fire would have been prevented from spreading. But instead of doing so Mr Coumis wasted his time in looking for a cloth which he failed to find. He then went to the bonnet to turn off the tap, but was too late, for owing to the continued flow of the petrol into the carburettor the fire had spread to the body of the car. The garage itself then caught fire and the whole building was burnt, including Mr Musgrove's rooms overhead, together with a quantity of furniture belonging to him. Mr Pandelis relied, among other defences, on section 86. At first instance [1919] 1 KB 314 Lush J said:

"But, nevertheless, I am of opinion that the statute affords the defendant no protection; for though the fire in the carburettor was accidental in a popular sense, I do not think it was accidental in the sense in which that term is used in the statute. If a man brings on to his premises a dangerous thing which is liable to cause fire, such as a motor car with petrol in it, the carburettor of which is not unlikely to get on fire when the engine is started, and a fire results, though without any negligence on his part, he must be held liable, the statute notwithstanding, for the rule is that he must keep such a thing under control at his peril."

The key point for him, therefore, was that the petrol was "liable to cause a fire" and "not unlikely to get on fire". He also held that Mr Coumis was negligent in not immediately turning off the petrol tap. Lush J was upheld on appeal to this court. Each of the three Lords Justices gave judgments. Bankes LJ began by considering the state of the common law before liability for fire was restricted by statute. He said:

"A man was liable at common law for damage done by fire originating on his own property (1) for the mere escape of the fire; (2) if the fire was caused by the negligence of himself or his servants, or by his own wilful act; (3) upon the principle of Rylands v Fletcher. This principle was not then known by that name, because Rylands v Fletcher was not then decided; but it was an existing principle of the common law as I shall show presently."

Bankes LJ then referred to Filliter v Phippard. He said (correctly) that that case decided that a fire negligently begun was not protected by the statute; and posed the question:

"Why, if that is the law as to the second head of liability, should it be otherwise as to the third head, the liability on the principle of Rylands v Fletcher? If that liability existed, there is no reason why the statute should alter it and yet leave untouched the liability for fire caused by negligence or design. That the principle of Rylands v Fletcher existed long before that case was decided is plain. In Vaughan v Menlove Tindal CJ says: "There is a rule of law which says you must so enjoy your own property as not to injure that of another." Park J says: "Although the facts in this case are new in specie, they fall within a principle long established, that a man must so use his own property as not to injure that of others." Rylands v Fletcher is merely an illustration of that old principle, and in my opinion Lush J was right in saying that this case, if it falls within that principle, is not within the protection of the statute."

In my judgment this reasoning is historically unsound. As McKenna J pointed out in Mason v Levy Autoparts of England Ltd [1967] 2 QB 530 at common law there were not three separate routes to liability: there was only one. A householder was liable for the escape of his fire (ignis suus): no additional danger was needed to be proved. The liability was based on a custom of the realm and on no other principle. If the case was brought otherwise than on the custom of the realm (i.e. by action on the case) then negligence had to be proved. This view was shared by Holdsworth (whom I have already quoted) and by Lord Denning in Emanuel (whom I have also quoted). McKenna J said:

"There were not three heads of liability at common law but only one. A person from whose land a fire escaped was held liable to his neighbour unless he could prove that it had started or spread by the act of a stranger or of God. Filliter's case had given a special meaning to the words "accidental fire" used in the statute, holding that they did not include fires due to negligence, but covered only cases of "a fire produced by mere chance, or incapable of being traced to any cause." But it does not follow, because that meaning may be given to "accidental," that the statute does not cover cases of the Rylands v Fletcher kind where the occupier is held liable for the escape though no fault is proved against him. In such cases the fire may be "produced by mere chance" or may be "incapable of being traced to any cause." Bankes LJ was making a distinction unknown to the common law, between "the mere escape of fire" (which was his first head) and its escape under Rylands v Fletcher conditions (which was his third), and was imputing an intention to the legislature of exempting from liability in the former case and not in the latter."

Likewise in his article on Vagaries in Liability for the Escape of Fire (1969) CLJ 104 AI Ogus says of Bankes LJ's analysis:

"This is, without doubt, a distortion of the position at common law. It has already been seen that the modern concept of negligence did not exist before the nineteenth century. The argument on Rylands v Fletcher fares no better. Its origins are not to be found in the action on the case pur negligent garder son few. Though similar in result, their bases are substantially different. The old form of action took as its starting point the escape of fire from the defendant's land. The plaintiff need then only show that it was ignis suus which caused damage to his property. Rylands v Fletcher, on the other hand, requires the accumulation on the defendant's land of "something likely to do mischief if it escapes. The former is an example of tortious liability based on "causal responsibility". The latter is an example of strict liability attaching to the use of dangerous things. Nor is the quotation from Vaughan v Menlove of assistance. It was simply part of a discourse on the emerging concept of negligence, and in any case could apply equally well to the tort of nuisance."

The Law Commission has also pointed out the fallacy in the reasoning. In their report on Civil Liability for Dangerous Things and Activities (1970) (Law Com No 32), Appendix para 18 they said:

"Indeed, it is difficult to escape a dilemma: if a liability of the Rylands v Fletcher type existed before its classic enunciation in the case from which it takes its name (as the Court of Appeal in fact argued in Musgrove v Pandelis) then the broad language of the statute, to which MacKenna, J, drew attention can hardy be ignored; if on the other hand it were suggested that the doctrine of Rylands v Fletcher grew up after the 1774 Act then its development was necessarily limited by the mandate of Parliament."

Nor does the reference to the maxim that "a man must so use his own property as not to injure that of others" make good the reasoning. First, as shown by Black v Christchurch Finance (and as Ogus says) even where the maxim does apply it may require proof of negligence. Indeed that was the case in Vaughan v Menlove itself; where the debate was between an objective test of negligence on the one hand, and a more subjective test of acting in good faith. No question of strict liability arose in that case, despite reliance on the maxim. Thus in my judgment Bankes LJ misinterpreted what the court had decided in Vaughan v Menlove. Second, in so far as the custom was based on that maxim it was overridden by the statute. Even if it was not part of the ignis suus custom it was a "law" which the statute expressly disapplied. Third, later authority is inconsistent with Bankes LJ's proposition based on the maxim. Thus in Emanuel Phillimore LJ said:

"The LCC were undoubtedly occupiers of this land. They were sued in that capacity and they did not call any evidence to suggest otherwise. As such they owed a duty to their neighbours which is best described in the old Latin maxim: sic utere tuo ut alienum non laedas. As Markham J put it in Beaulieu v Finglam:

'I shall answer to my neighbour for him who enters my house by my leave or knowledge whether he is guest to me or my servant, if either of them acts in such a way with a candle or other things that my neighbour's house is burned.'

Since the Fire Prevention (Metropolis) Act 1774 it is I think necessary to insert the word 'negligently' after the word 'acts'.

In my judgment what Filliter v Phippard decided was that "accidentally" meant "without negligence". The legislation must have been intended to change the common law, whether or not it was based on custom, otherwise Parliament would not have used the wide words that it did. On the footing that fires started by negligence did not attract the protection of section 86, what change must Parliament have intended? The answer, in my judgment, must be that Parliament intended to abolish liability based simply on the proposition that a man "must so use his own as not to injure another", without proof of negligence. Whether liability arose because of custom or because of some other principle of common law does not matter. Both were overridden by section 86. This is confirmed by Emanuel.

It follows from this that I disagree with the view expressed by Judge Thornton QC in Johnson v BJW Property Developments Ltd [2002] EWHC 1131 (TCC) [2002] 3 All ER 574 and by Judge Coulson QC in LMS International Ltd v Styrene Packaging and Insulation Ltd [2005] EWHC 2065 (TCC) that McKenna J was wrong in his appreciation of the common law, or in the effect of the 1774 Act.

As I have said, liability for fire was not the only form of strict liability imposed by custom of the realm. The liability of innkeepers was another such liability. In Williams v Owen [1955] 1 WLR 1293 Mr Williams left his car overnight in the hotel garage. A fire broke out and destroyed his car. Finnemore J held that the strict liability of an innkeeper was limited to loss of his guest's goods rather than to their destruction. But he went on to deal with section 86. He said:

"I think that this liability of the innkeeper was a custom of the realm. It is true it is embodied in common law rules, but then common law is the legal expression of custom, and it seems to me that that also would be an answer in this particular case. I suppose that by 1774 the legislature had appreciated what Lord Goddard CJ laid down in Sochacki v Sas: "Everybody knows fires occur through accidents which happen without negligence on anybody's part." Parliament in 1774 apparently thought it right that they should make it plain that whatever customs or usages there were to the contrary, in this country a man should not be held to be responsible for a fire which occurred accidentally — which I take to mean "without negligence on his part."

Those are two points which I think would be enough to decide that there is in this case no absolute liability on the part of the innkeeper, first, because there was injury to the car and not theft or loss; and, secondly, because, in any event, as it was a fire the Act of 1774 would limit the liability of the innkeeper, so far as a fire is concerned."

In other words section 86 overrode the custom. In my judgment this is plainly right because the Act expressly exempts a person from liability for accidental fires "any law, usage or custom to the contrary notwithstanding". This would include the custom of the realm upon which liability for "ignem suum" was based. The only question, then, is what Parliament meant by the word "accidentally"; and that had been decided by Filliter v Phippard.

In Musgrove v Pandelis Bankes LJ continued:

"The question then, is whether this motor car, with its petrol tank full or partially filled with petrol, was a dangerous thing to bring into the garage within the principle of Rylands v Fletcher? Mr. Hawke says a motor car is not a dangerous thing unless it is in such a condition that an accident is to be apprehended. But the expectation of danger is not the basis of the principle of Rylands v Fletcher. A thing may be dangerous although the danger is unexpected. I agree with Lush J that this motor car was dangerous within that principle. The defendant brought it, or caused it to be brought upon his premises, and he is responsible for the fire which resulted, and is not within the protection of the statute."

It will be recalled that the reason that Lush J thought the car was dangerous was that the petrol in its tank was "liable to cause a fire" and "not unlikely to get on fire". Warrington LJ agreed on that point; and said:

"If this motor car with the petrol in its tank was potentially dangerous, such as a man's own fire, then it was the defendant's duty to see that the potential danger did not become an actual danger causing damage to his neighbour. The Act of Geo. 3 is no protection against that liability."

However, both Bankes LJ and Warrington LJ went on to consider the question of negligence and both held that Lush J was right on that issue. That, as has subsequently been pointed out, is sufficient to support the decision itself. Duke LJ seems to me to have adopted a different line of reasoning. He said:

"I do not see how this case can be taken out of the principle of Rylands v Fletcher, which was thus stated by Lord Cairns LC in the very words of Blackburn J: "The true rule of law is, that the person who, for his own purposes, brings on his land and collects and keeps there anything likely to do mischief if it escapes, must keep it in at his peril." He can excuse himself by showing that the escape was owing to the plaintiff's default or perhaps that it was the consequence of vis major or the act of God. In the present case there was petrol which was easily convertible into an inflammable vapour; there was the apparatus for producing a spark; and added to those there was a person supposed to control the combustion but inexperienced and unequal to the task. Taking together the presence of the petrol, and the production of the inflammable gas, or those combustibles together with the inexperience of the person placed in charge of them, it is impossible to say that this is not an instance of the principle laid down by Blackburn J."

However, although he held that the principle of Rylands v Fletcher applied, he nevertheless went on to consider whether the fire was accidental for the purposes of section 86. If he had thought that the application of Rylands v Fletcher was a complete answer that consideration would have been unnecessary. On that question he said:

"That would dispose of this case but for the defendant's contention that he is excused by s. 86 of the Fires Prevention (Metropolis) Act, 1774. In my opinion the terms of that enactment fall far short of showing a definite intention to relieve a defendant in such a case as this. The actions against which the statute gives protection are in respect of fires which shall accidentally begin. I have the greatest doubt whether this fire began accidentally at any stage. If it was all one fire, it was begun not accidentally but intentionally. If progressive stages may be regarded it was not a fire which began accidentally without negligence at the stage when it became a conflagration involving goods and premises. The question may some day be discussed whether a fire, spreading from a domestic hearth, accidentally begins within the meaning of the Act, if such a fire should extend so as to involve the destruction of property or premises. I do not covet the task of the advocate who has to contend that it does. In the present case the fire, so far as it was a means of mischief, resulted from the negligent omission to turn off the petrol tap, an act which would have stopped the flow of petrol. All the witnesses who had any experience of such matters drew a distinction between fire in a carburettor, where the vapour can be instantly out off, and such a fire as occurred in this case. The learned judge has found that this fire was due to negligence. I cannot disagree with him. Whatever may be the effect of the Act of Geo. 3 upon the nice questions that have been discussed, this case is outside any possible protection of that statute."

This is, with all respect, a rather confused passage; but in my judgment the key finding here is that the spread of the fire was attributable to negligence.

Musgrove v Pandelis has been subsequently considered on a number of occasions. In Job Edwards Ltd v Birmingham Navigation Proprietors [1924] 1 KB 341 rubbish was tipped on land belonging to a canal company and on adjoining land belonging to mine owners. The rubbish on the mine owners' land was found to be on fire, and the canal company feared that the fire might spread to their own land. Having called on the mine owners to extinguish the fire, the canal company entered the mine owners' land (by agreement) and put out the fire. The question was whether the mine owners were liable to contribute to the cost. Bailhache J held that the mine owners had no duty to prevent the spread of the fire. He held that:

"… where a fire occurs through no fault of the landowner, without his knowledge, and, as in this case, on matter brought on his land without his knowledge and against his will, he is not responsible for the spreading of such a fire to the adjoining land, but the neighbour is entitled to go upon his land and prevent the fire from spreading."

Thus the mine owners were not liable to contribute to the cost. The canal company appealed. Mr Vachell QC, for the canal company argued as follows:

"No doubt a person on whose estate any fire shall accidentally begin is exempted from liability by s. 86 of the Fires Prevention (Metropolis) Act, 1774, for damage caused by that fire; and it may be admitted that a fire accidentally began upon land belonging to the respondents. But the fire which caused or threatened to cause damage to the appellants' property was not the fire which accidentally began on the respondents' land, when it might have been extinguished quickly and easily, but the fire which was from May to October, 1920, allowed to increase and become formidable: Musgrove v. Pandelis."

Bankes LJ gave the leading judgment. He began by considering the distinction between a public nuisance (which he said a land owner had a duty to abate) and a private nuisance (which he said gave rise to no such duty). He then turned to the custom relating to fire, and explained it thus:

"The case of fire has always been looked upon in our law as a somewhat exceptional case. It was no doubt the ancient law or custom of England that a person in whose house a fire originated which afterwards spread to his neighbour's property and destroyed it must make good the loss, but I do not consider that rule as opposed to the view I am putting forward in regard to liability for injury done by a private nuisance, as the ancient law no doubt considered a fire as a public nuisance owing to the danger of its spreading. The view of the law which I am taking does not touch a case where the private nuisance has been caused, or allowed to continue, by any act or default on the part of the occupier of the land on which it exists."

This is a quite different explanation of the custom to that which he gave in Musgrove v Pandelis. The other point that he considered was the defence under section 86. As to that he said:

"Mr Vachell contended that whatever may have been the cause of the original fire it ceased to be an accidental fire within the meaning of the statute when the plaintiffs were informed of it, and that within the reasoning of the decision in Musgrove v Pandelis the fire as from that date must be treated as a second and independent fire. I cannot draw any such inference from the facts of the present case. In Musgrove v Pandelis Lush J drew from the facts the inference that there were in substance either two fires, the first an accidental one which did no damage, and the second which was due to negligence and did the damage; or alternatively that there was only one fire within the meaning of the statute, and that was the one due to negligence. This Court agreed with the view of the learned judge, but the facts of that case are very special, and have in my opinion no bearing upon the case we are now dealing with."

It seems clear from this passage that Bankes LJ himself did not regard Musgrove v Pandelis as applying to what might be called an ordinary case of spread of fire. Bankes LJ (with whom Astbury J agreed) held that the mine owners owed no duty to the canal company and hence dismissed the appeal. Scrutton LJ dissented. On the question of common law duty he said:

"There is a great deal to be said for the view that if a man finds a dangerous and artificial thing on his land, which he and those for whom he is responsible did not put there; if he knows that if left alone it will damage other persons; if by reasonable care he can render it harmless, as if by stamping on a fire just beginning from a trespasser's match he can extinguish it; that then if he does nothing, he has "permitted it to continue," and become responsible for it. This would base the liability on negligence, and not on the duty of insuring damage from a dangerous thing under Rylands v Fletcher. I appreciate that to get negligence you must have a duty to be careful, but I think on principle that a landowner has a duty to take reasonable care not to allow his land to remain a receptacle for a thing which may, if not rendered harmless, cause damage to his neighbours."

This is a critical passage because it was subsequently approved by the House of Lords in Sedleigh-Denfield v O'Callaghan [1940] AC 880 who disapproved the reasoning of the majority. As far as section 86 was concerned, Scrutton LJ said:

"That statute (14 Geo. 3, c. 78, s. 86) provides that no action shall lie against any person in whose house or on whose estate "any fire shall accidentally begin." This fire undoubtedly began accidentally so far as the landowner and his agents were concerned. It has been decided that the statutory restriction of the previous common law liability does not exclude liability for fires caused by negligence of the owner or persons for whom he is responsible, or by dangerous things for which the owner is responsible under the doctrine of Rylands v Fletcher. This leaves the difficult question - suppose the fire is caused by a trespasser, as if he throws down a match; and suppose the owner comes by immediately afterwards, sees the small fire, and could with no trouble extinguish it by stamping on it, but does not do so, so that the fire spreads and damages his neighbour, is he freed by the statute? He is then aware of a dangerous thing on his land which may damage his neighbour, and which by reasonable care he can prevent from damaging his neighbour, and he does nothing. I agree he is not an absolute insurer of that dangerous thing, for he did not himself create it, but I think on principle he is bound to take reasonable care of a dangerous thing which he knows to exist. Take the case of an ordinary house fire, where a coal leaps from the grate. If no one knows of the fire caused by the coal till it cannot be stopped, that fire may be within the protection of the statute, though Duke LJ doubted it in Musgrove v Pandelis. But suppose the owner sees it jump out, could extinguish it with a moment's trouble, and does not trouble to do so, could he plead the statute to protect him? In Musgrove v Pandelis, where the real danger arose from the fact that the defendant's servant negligently did not turn a tap to stop a supply of petrol to a fire, the Court treated the fire as two fires; I should respectfully have thought that it was safer to say that the fire was continued by negligence, and that the cause of action was not for a fire accidentally begun, but for negligence in increasing such a fire."

In Collingwood v Home and Colonial Stores Ltd [1936] 3 All ER 200 Lord Wright MR also considered Musgrove v Pandelis. He said that the ground for decision, in so far as it was not based on negligence was one about which he had "some difficulty"; and Romer LJ pointed out that in that case there had not been an escape of anything such as to bring the principle of Rylands v Fletcher into play. Although he said that Rylands v Fletcher was based on a broader principle, he doubted the propositions on which the alternative ground rested. Musgrove v Pandelis was doubted again by Lord Porter in Read v J Lyons & Co Ltd [1947] AC 156, 176, referring to Romer LJ's observations in Collingwood v Home and Colonial Stores Ltd. It was also regarded with scepticism by Lord Walker of Gestingthorpe in Transco plc v Stockport MBC (at 39).

In H & N Emanuel Ltd v GLC [1971] 2 All ER 835 the fire in question had been deliberately started by workmen in order to burn rubbish on a demolition site. It spread to neighbouring land. The Court of Appeal based its decision entirely on the question of negligence. Bearing in mind that the fire had been deliberately started, Rylands v Fletcher would have provided a very short answer if the principle applied to a fire that spread accidentally. But instead the Court of Appeal discussed negligence extensively in the context of the statutory defence under section 86. That approach is in my judgment quite inconsistent with Musgrove v Pandelis.

Charlesworth & Percy on Negligence say (para 13-112):

"An alternative ground for the decision in Musgrove v Pandelis was that the 1774 Act was no defence when the fire originated from a dangerous thing. This proposition has been criticised, but it was adopted and followed in Mulholland & Tedd Ltd v Baker. Since it is the fire which is the dangerous thing, whether it is caused by petrol, paraffin or anything else, and the object of the statute is to give protection against accidental fires, it is difficult to understand why the statute should not protect as much in one case as in the other. The presence of inflammable matter on premises is important when the question of negligence is being considered, so that if there is no negligence and it is found that the fire is accidental, it is submitted that the statute is a defence to the occupier of the land on which it begins, whatever may be the origin of the fire."

In my judgment Musgrove v Pandelis is wrong in so far as it describes the basis of the common law before the earliest of the fire statutes. It invents an unhistorical justification for the basis of the rule. That justification has been criticised by judges, by scholars and by the Law Commission. It is inconsistent with a subsequent decision of this court in Emanuel. For good measure, Bankes LJ himself said that it was decided on special facts; and, on the face of it, it is inconsistent with his own subsequent decision in Job Edwards Ltd v Birmingham Navigation Proprietors. In Collingwood v Home and Colonial Stores Ltd Lord Wright MR said that whatever it decided Musgrove v Pandelis did not apply to the case of a fire caused by an unknown defect in electrical wiring. This was also the view of the Court of Appeal in Stockport MBC v British Gas Plc [2001] EWCA Civ 212 (whose decision was affirmed sub nom Transco). It misinterprets both Vaughan v Menlove and Filliter v Phippard, and wrongly distinguishes them. Moreover it misstates the principle in Rylands v Fletcher, unless it is confined to fires deliberately kindled (or deliberately brought on to land). That, too, was in effect decided by Collingwood v Home and Colonial Stores Ltd where Lord Wright considered whether the electrical wiring (which had been brought onto the land deliberately and where the fire started) rather than the fire fell within the principle in Rylands v Fletcher. Musgrove v Pandelis is clearly an extension of the principle in Rylands v Fletcher because the occupier did not bring the fire onto his land. This was also the view of the High Court of Australia in Goldman v Hargrove (1963) 110 CLR 40 (affirmed [1967] 1 AC 645) in which Windeyer J said that the principle in Rylands v Fletcher did not apply to the fire that started on Mr Goldman's land "simply because the respondent did not bring the fire upon his land, nor did he keep it there for any purpose of his own. It came there from the skies. And he did nothing to make its presence there more dangerous to his neighbours." Lastly, having regard to the findings of fact made by Lush J at trial, the principle that Bankes LJ formulated was unnecessary to the decision, and wider than the facts of the case warranted.

In Cambridge Water Co v Eastern Counties Leather plc Lord Goff warned against developing the principle in Rylands v Fletcher into a rule of strict liability for damage cause by ultra-hazardous operations; and that warning was repeated by Lord Bingham in Transco plc v Stockport MBC. In Transco plc v Stockport MBC the House of Lords were at pains to confine the scope of the principle to its proper ambit. Lord Bingham pointed out that it was a necessary condition of liability that the defendant had brought onto his land something which would be a source of mischief or danger if it escaped. That is not the case with a fire that is not deliberately kindled, because the defendant has not brought it onto his land. As I have already said he summarised his discussion by referring to a case in which an occupier of land "has brought or kept on his land an exceptionally dangerous or mischievous thing in extraordinary or unusual circumstances." A fire that accidentally ignites cannot be described as that; and nothing else escapes. Lord Scott also said that the principle applied to:

"… cases where something or other, potentially dangerous, that the defendant has brought onto his land has escaped onto the plaintiff's land and there caused damage."

He described the element of escape as "essential". Lord Bingham also said that the use that the defendant made of his land would have to be "extraordinary and unusual" by reference to the time and place at which it was carried on before the principle would apply. This was plainly intended as a narrowing of the principle expounded in Rylands v Fletcher. In my judgment no extension of the principle in Rylands v Fletcher can now be justified in the light of the decisions of the House of Lords in Cambridge Water Co v Eastern Counties Leather plc and Transco plc v Stockport MBC. In my judgment Musgrove v Pandelis cannot stand with the subsequent decisions of the House of Lords, and it is in conflict with the subsequent decision of this court in Emanuel. It wrongly distinguishes two previous decisions of courts of co-ordinate jurisdiction. In those circumstances, in my judgment we are free to choose the correct path to follow: Young v Bristol Aeroplane Co Ltd [1944] KB 718, 729; Starmark Enterprises Ltd v CPL Holdings Ltd [2001] EWCA Civ 1252 [2002] Ch 306 (§ 64, § 97). In my judgment Musgrove v Pandelis is unsound authority and should no longer be followed.

I turn then to the true scope of the defence under section 86 of the Act. In Collingwood v Home and Colonial Stores Ltd [1936] 3 All ER 200 Lord Wright MR said of the Act:

"That Act, as is well known, changed the law, because before that Act if a fire spread from a man's premises and did damage to adjoining premises, he was liable in damage on the broad ground that it was his duty at his own peril to keep any fire that originated on his premises from spreading to and damaging his neighbour's premises. The protection of the section is limited by the word "accidentally." The meaning of that is discussed in Filliter v Phippard, where it was held that the section did not apply to a fire due to the negligence of the defendant or his servants."

The fire in that case was caused by an unknown defect in the electrical wiring system. There was no negligence established in relation either to the installation or the maintenance of the electrical system. The initial ignition was thus accidental. Nor, it seems, was there any negligence in controlling the fire once it had started, with the result that the statutory defence succeeded. Indeed the case was all the more striking because the damage was caused by water which had been deliberately applied in an attempt to extinguish the fire. Spicer v Smee [1946] 1 All ER 489 was another case of a fire starting in the electrical system. However, Atkinson J held that the wiring had been negligently installed; and that in that state the wiring was a nuisance. It was for that reason that the statutory defence failed. The decision itself can be justified on the basis that the wiring had been negligently installed. The finding of nuisance is, in my judgment, more questionable.

In Emanuel, immediately after the passage I have quoted from Lord Denning's judgment in which he said that at common law a person was not liable for fire which escaped without negligence, he added:

"All those cases are covered, if not by the common law, at any rate by the Fire Prevention (Metropolis) Act 1774, which covers all cases where a fire began or spreads by accident without negligence. But that Act does not cover a fire which begins or is spread by negligence: see Filliter v Phippard, Musgrove v Pandelis and Goldman v Hargrave."

In the same case Phillimore LJ said:

"Since the Fire Prevention (Metropolis) Act 1774 it is I think necessary to insert the word 'negligently' after the word 'acts'." (Emphasis added)

In Sochacki v Sas [1947] 1 All ER 344 the plaintiff lit a fire in the domestic grate in his room and went out one afternoon leaving the fire alight. A spark jumped out of the fire and set fire to the floorboards. The fire spread to an adjoining room where it caused damage. Although there was neither a fireguard nor a fender, Lord Goddard said that no negligence was involved. He held that the principle in Rylands v Fletcher did not apply, because although the fire had been deliberately kindled, the kindling of the fire was an ordinary use of land. He continued:

"It is not the case of a fire starting on one owner's premises and spreading to the premises of an adjoining owner. If a fire is negligently or improperly started by a person on his land, as for instance, lighting a bonfire which spreads, he may be liable, not merely to an adjoining owner who suffers damage, but to any other person who suffers damage. If I happen to be on somebody else's land at a time when a fire spreads to that land and my motor car or property is destroyed, I have just as much right against the person who improperly allows the fire to escape from his land as the owner of the land on which I happen to be. I do not doubt that for a moment, but here the fire was being used by a man in a fireplace in his own room. There was an ordinary, natural, proper, everyday use of a fireplace in a room."

There was no express discussion of section 86, but it is plain that Lord Goddard decided the case on the basis of negligence alone.

In Solomons v R. Gertzenstein Ltd [1954] 1 QB 565 a fire started as a result of an electrical short circuit. That set fire to some wood and in due course to a stack of paper. Lord Goddard held that section 86 applied. He said:

"In my opinion it was a short circuit that set fire to the wood in the neighbourhood of the ventilator and that in turn set fire to the stack of paper. Pausing here, it does not appear that this paper ever burnt freely; no doubt it caused a lot of smoke, and I accept the evidence that there was some flame seen, but it was not that stack apparently that caused the sudden sheet of flame which caused the real damage here. The cause of that is obscure; the only explanation was that offered by the fire officers, that the heating of the paint and varnish caused an accumulation of gas which suddenly ignited and rushed upwards. However, I do not propose to deal further with this because I cannot hold that placing packing paper and cardboard cartons in this recess behind the balustrade was negligent. Business of the sort carried on by the first defendants necessitates having a stack of packing and wrapping material at hand. This material is not highly inflammable like loose tissue paper or shavings would be. It is common knowledge that it takes a good deal to get closely packed thick paper well alight, though it will smoulder. But in any case I cannot see how it can be negligent to store this paper in what was a convenient recess any more than it would be to store it in one of the rooms occupied by the first defendants. They had no reason to suppose that there was likely to be a short circuit which would fire the panelling in the immediate neighbourhood of the stack, which I may mention was never burnt through. On the evidence before me I am not prepared to find that the fire was caused by the negligence of any of the defendants, and I hold that it was accidental and need only refer on this matter to Collingwood v Home and Colonial Stores Ltd."

Thus Lord Goddard held that section 86 applied where a fire accidentally began and set fire to material that was not readily inflammable; and that to store that material was not negligent. His process of reasoning was the same as that in Sochacki v Sas. The case went to the Court of Appeal where Lord Goddard's decision was reversed; but on a completely different point.

However, where there is negligence liability attaches, even if the negligence in question is that of an independent contractor. This is shown by another decision of Lord Goddard, this time sitting in the Court of Appeal, in Balfour v Barty-King [1957] 1 QB 496. In that case the fire started as the result of the negligent use of a blow torch by an independent contractor. The use of fire had, therefore, been deliberate. The argument for the successful plaintiff was that "If negligence be shown, it matters not against whom, the fire is not accidental." Lord Goddard said:

"The precise meaning to be attached to "accidentally" has not been determined, but it is clear from these last two cited cases that where the fire is caused by negligence it is not to be regarded as accidental. Although there is a difference of opinion among eminent text writers whether at common law the liability was absolute or depended on negligence, at the present day it can safely be said that a person in whose house a fire is caused by negligence is liable if it spreads to that of his neighbour, and this is true whether the negligence is his own or that of his servant or his guest, but he is not liable if the fire is caused by a stranger.

Who, then, is a stranger? Clearly a trespasser would be in that category, but if a man is liable for the negligent act of his guest, it is, indeed, difficult to see why he is not liable for the act of a contractor whom he has invited to his house to do work on it, and who does the work in a negligent manner."

I come now to the important case of Goldman v Hargrave [1967] 1 AC 645. A tree on Mr Goldman's land was struck by lightning and caught fire. He cleared a space round the tree of combustible material, and arranged for the tree to be cut down. However, once it had been cut down instead of putting out the fire, he left it to burn itself out. Unfortunately it did not, and it spread to Mr Hargrave's land. The question was whether Mr Goldman was liable. The Privy Council held that that an occupier of land was under a general duty of care, in relation to hazards, whether natural or man-made, occurring on his land, to remove or reduce such hazards to his neighbour. The existence of the duty must be based on knowledge of the hazard, ability to foresee the consequences of not checking or removing it, and the ability to abate it; and the standard of care required of the occupier is founded on what it was reasonable to expect of him in his circumstances. There are a number of important points that arise out of this case. First, the Privy Council rejected the argument that Mr Goldman had adopted the fire as his own "as suus ignis - and had made use of it for his own purpose or advantage." In my judgment this encapsulates the principle of ignis suus that I have already discussed. It reinforces the proposition which I believe to be correct, namely that a fire that was never deliberately kindled fell outside the principle of ignis suus. In order to count as "ignis suus" the fire must have been made for the purpose or advantage of the occupier. Second, the case was not one in which the occupier had brought a source of danger (i.e. the fire) onto his land. This is equally consistent with the rationale of Rylands v Fletcher, viz. that an occupier of land has brought something dangerous onto his land for his own purposes. Third, the Privy Council held that in the case of fire there was no difference between a fire that started from natural causes and one that had been started by human agency. Lord Wilberforce put it thus:

"Their Lordships would first observe, with regard to the suggested distinction, that it is well designed to introduce confusion into the law. As regards many hazardous conditions arising on land, it is impossible to determine how they arose - particularly is this the case as regards fires. If they are caused by human agency, the agent, unless detected in flagrante delicto, is hardly likely to confess his fault. And is the occupier, when faced with the initial stages of a fire, to ask himself whether the fire is accidental or man-made before he can decide upon his duty? Is the neighbour whose property is damaged bound to prove the human origin of the fire? The proposition involves that if he cannot do so, however irresponsibly the occupier has acted, he must fail. But the distinction is not only inconvenient, it lacks, in their Lordships' view, any logical foundation.

Within the class of situations in which the occupier is himself without responsibility for the origin of the fire, one may ask in vain what relevant difference there is between a fire caused by a human agency, such as a trespasser, and one caused by act of God or nature. A difference in degree - as to the potency of the agency - one can see but none that is in principle relevant to the occupier's duty to act. It was suggested as a logical basis for the distinction that in the case of a hazard originating in an act of man, an occupier who fails to deal with it can be said to be using his land in a manner detrimental to his neighbour and so to be within the classical field of responsibility in nuisance, whereas this cannot be said when the hazard originates without human action so long at least as the occupier merely abstains. The fallacy of this argument is that, as already explained, the basis of the occupier's liability lies not in the use of his land: in the absence of "adoption" there is no such use; but in the neglect of action in the face of something which may damage his neighbour. To this, the suggested distinction is irrelevant."

It is important to note that this passage is directed to a case in which the occupier is without responsibility for the origin of the fire: that is to say the initial ignition. This is clear from the subsequent structure of the Privy Council's advice because Lord Wilberforce went on to consider the separate question of liability for the spread of the fire. Thus the Privy Council went on to consider whether, once the fire had started, there was a duty; and held that there was a duty to do what was reasonable in the circumstances to prevent the spread of the fire. This was precisely the approach of Scrutton LJ in the Job Edwards case. On the facts Mr Goldman was negligent in having left the fire to burn out, rather than extinguishing it. The last point in the case was whether Mr Goldman was entitled to rely on the statutory defence under section 86 (which applies in some Australian states). On that question Lord Wilberforce said:

"The words "shall accidentally begin" are simple enough, but the simplicity is deceptive. Read literally they suggest that account need be taken of nothing except the origin of the fire and that given an accidental beginning, no supervening negligence or even deliberate act can deprive a defendant of the benefit of the statute. But further reflection suggests a doubt both because such a result seems capable of producing absurdity and injustice, and because of the inherent difficulty of saying what the expression "any fire" is intended to mean. A fire is an elusive entity; it is not a substance, but a changing state. The words "any fire" may refer to the whole continuous process of combustion from birth to death, in an Olympic sense, or reference may be to a particular stage in that process - when it passes from controlled combustion to uncontrolled conflagration. Fortunately, the Act has been considered judicially and, as one would expect, the process of interpretation has taken account of these considerations. In Filliter v Phippard Lord Denman explained the purpose of the earlier Act (6 Anne, c. 31, s. 6) as being to remove the supposed common liability of a person "in whose house a fire originated which afterwards spread to his neighbour's property" and held that it did not apply to a fire caused deliberately or negligently. This was carried further in Musgrove v Pandelis, where a fire started accidentally in the carburettor of a car, but spread because the chauffeur negligently failed to turn off the petrol tap. The Court of Appeal held that the Act did not apply. Bankes LJ put it that the Act relieved an owner for a mere escape of fire from his premises but did not relieve him against a claim for damages for negligence. The fire which caused the damage was, he thought, not the spark which caused the initial ignition, but the raging fire which arose from the act of negligence. Their Lordships accept this interpretation: it makes sense of the statute, it accords with its antecedents, and it makes possible a reasonable application of it to the facts of the present case, that is to say, that the fire which damaged the respondents' property was that which arose on March 1 as the result of the negligence of the appellant. The statutory defence therefore fails."

It will be noted that Lord Wilberforce treated Musgrove v Pandelis as having turned on negligence and refrained from endorsing any wider basis for defeating the statutory defence. The decision in Goldman v Hargrave can in my judgment be summarised thus:

i) An occupier of land is not liable for the initial outbreak of fire, whether due to natural causes or human agency, unless he himself has brought the fire onto the land;

ii) He has a duty to do what is reasonable to prevent the spread of the fire. If he fails to do what is reasonable to prevent the spread of fire he is negligent;

iii) If he is negligent in preventing the spread of the fire the statutory defence under section 86 will fail. If not, it will succeed.

The law as expounded in Goldman v Hargrave has since been accepted as being the law of England and Wales too: Leakey v National Trust [1980] QB 485; Bybrook Barn Centre v Kent CC [2001] BLR 55.

I turn now to consider a short line of cases about fire that have developed since Musgrove v Pandelis. E Hobbs (Farms) Ltd v The Baxenden Chemical Co Ltd [1992] 1 Lloyd's Rep 54 was a case in which the use of a grinding machine in a barn emitted sparks that set combustible material alight. The fire spread to adjoining property. Sir Michael Ogden QC found that the contractor who operated the grinding machine had been negligent, with the result that the statutory defence under section 86 failed. Since negligence was established, the decision is plainly correct. What he said about Rylands v Fletcher was therefore obiter. In Johnson v BJW Property Developments Ltd [2002] 3 All ER 574 a fire was deliberately kindled in a domestic grate. It spread to adjoining property because a fire brick lining to the chimney had been negligently removed. As Judge Thornton QC put it:

"It follows that the defendant in this case is to be held vicariously liable for the damage caused by the escape of fire into and onto the claimant's premises since the fire only escaped and caused damage because of the negligent workmanship of the defendant's independent contractor for whom the defendant is to be held separately liable."

I agree that on the facts found in that case, section 86 was no defence. I do not, however, agree with Judge Thornton's more general observations on liability for the spread of fire.

I have already mentioned Mason v Levy Autoparts of England Ltd. McKenna J followed Musgrove v Pandelis with obvious reluctance. He recognised correctly that the principle of Rylands v Fletcher could not be directly applied to a case where the occupier had not brought the fire onto the land. Nothing that the occupier had brought on to the land had escaped. He therefore had to find a modified and extended version of the principle. The principle that he formulated was that an occupier of land was liable if:

"(1) he brought onto his land things likely to catch fire, and kept them there in such conditions that if they did ignite the fire would be likely to spread to the plaintiff's land; (2) he did so in the course of some non-natural use; and (3) the things ignited and the fire spread."

This formulation has been followed in subsequent cases at first instance: see, for example LMS International Ltd v Styrene Packaging and Insulation Ltd [2005] EWHC 2065 (TCC); Harooni v Rustins Ltd [2011] EWHC 1632 (TCC).

The unfortunate result of Musgrove v Pandelis compelled McKenna J reluctantly to extend the principle of Rylands v Fletcher. It is equally unfortunate that in neither Cambridge Water Co v Eastern County Leather Ltd nor Transco plc v Stockport MBC did the House of Lords discuss the modified principle that had been applied to fire cases. But in the light of Transco plc v Stockport MBC the extension of the principle in Mason v Levy Autoparts of England Ltd was, in my judgment, a wrong turning in the law. No extension of the principle in Rylands v Fletcher can be justified. I would therefore overrule Mason v Levy Autoparts of England Ltd, a result that McKenna J would no doubt have welcomed.

Where a person brings combustible materials onto his land, he may well owe his neighbour a duty to take reasonable precautions to prevent their combustion or to take reasonable precautions to prevent the spread of fire. The principle was well put by Lord Macmillan in Read v J Lyons Ltd:

"Every activity in which man engages is fraught with some possible element of danger to others. Experience shows that even from acts apparently innocuous injury to others may result. The more dangerous the act the greater is the care that must be taken in performing it. This relates itself to the principle in the modern law of torts that liability exists only for consequences which a reasonable man would have foreseen. One who engages in obviously dangerous operations must be taken to know that if he does not take special precautions injury to others may very well result…. The sound view, in my opinion, is that the law in all cases exacts a degree of care commensurate with the risk created."

In my judgment the law is as stated in Goldman v Hargrave at least as regards fires that have not been deliberately kindled. An occupier of land will not be liable to his neighbour for a fire that begins accidentally unless he is negligent in failing to prevent its spread. The general test of negligence may entail the taking of special precautions where the use in question involves the accumulation or storage of inflammable or readily combustible materials. But that is a question of fact to be decided on a case by case basis. This conclusion is, I believe, also consistent with the subsequent decision of the Privy Council in Long Bee & Co v Ling Nam Rubber Works [1970] 2 Lloyd's Rep 247.

It may be said that the retention of a special rule of liability in the case of fire is justified as a matter of policy. But in my judgment that route is barred by Cambridge Water Co and Transco.

In the present case the Recorder found that the primary source of the fire lay in the wiring of electrical appliances in Mr Stannard's unit. However, he also found that:

"… there is nothing to show that such a state of affairs was the result of failing to maintain or keep in good order the electrical system itself, of those electrical appliances that were located within Wyvern's premises, as opposed to something that may have arisen entirely by accident."

He also found that there was no allegation or evidence about what (if anything) Mr Stannard should have done to prevent the spread of the fire from reaching the tyres, and thereby involving them in the conflagration. He thus concluded that, subject to the principle in Rylands v Fletcher, Mr Stannard made good his defence under section 86. Having referred to a number of case both on the principle in Rylands v Fletcher and the modified principle applied to fire cases, he concluded (§ 64):

"The only relevant activity was the storage of tyres. Was this dangerous within the Rylands v Fletcher rule, i.e. did it constitute a foreseeable and exceptional high risk of damage to Mr Gore's Premises? Tyres are not in themselves flammable, and will not ignite unless there is a sufficient flame or heat source. It is, however, not impossible for tyres to catch fire (as obviously happened here), and if they do ignite, they have a special fire risk quality. This is that once alight they may burn rapidly and intensively, such that they are difficult to put out…. If fire broke out there was an exceptionally high risk of damage to Mr Gore's Premises because of the rapidity and intensity of the fire that would be created by the tyres that Mr Stannard had stored on Wyvern's Premises."

Although the Recorder said that he drew support from Mason v Levy Autoparts of England Ltd, in fact he went much further than that case. One of McKenna J's conditions for liability was that the occupier brought onto his land things that were likely to catch fire. This was also the view of Judge Coulson QC in LMS International Ltd v Styrene Packaging and Insulation Ltd and of Akenhead J in Harooni v Rustins Ltd. On the Recorder's findings the tyres were not likely to catch fire. Thus although the scope of Rylands v Fletcher has been narrowed each time the highest courts have considered it, the Recorder in fact extended it beyond any previous expression of the principle. He imposed strict liability where it had not existed before. Accordingly, even if I am wrong in thinking that the extended principle in Rylands v Fletcher does not defeat a defence under section 86 against liability for the consequences of a fire that starts and spreads without anyone's negligence, in my judgment the Recorder was wrong to hold Mr Stannard liable on the facts that he found.