High Court of Australia Taylor, Windeyer and Owen JJ. Hargrave v Goldman [1963] HCA 56 ORDER Appeal allowed with costs. Order and judgment of the Supreme Court of Western Australia of 9th January 1963 set aside. Case remitted to the Supreme Court. Costs of the parties up to the time of the entry of the judgment appealed from to abide the order of the Supreme Court. Cur. adv. vult. The following written judgments were delivered:— Nov. 22 Taylor and Owen JJ. This is an appeal by two parties in a consolidated action from an order of the Supreme Court by which the claim of the appellants in respect of fire damage to their property was dismissed: Hargrave v. Goldman [9] . The facts are that the respondent was the owner and occupier of a lightly developed grazing property, some six hundred acres in extent, near Gidgegannup in Western Australia. On Saturday 25th February 1961 there was an electrical storm in this area and a tall tree with a branchy-top, which stood about the centre of the respondent's property and relatively close to his dwelling, was struck by lightning between 5 p.m. and 6 p.m. It was observed shortly afterwards that a fork of the tree, more than eighty feet above the ground, was on fire. The tree was about two hundred and fifty yards from the western boundary of the respondent's property and a somewhat lesser distance from the eastern boundary. On either side the paddocks were sparsely timbered but they contained a quantity of dry grass and dead tree tops. It was impossible for the respondent to extinguish the fire whilst the tree was standing and early on the following morning he telephoned the fire control officer for the district—a local farmer appointed pursuant to the Bush Fires Act, 1954-1958 W.A.—and asked that a "tree faller" be sent out to cut the tree down. This was done about midday but, in the meantime, the respondent had by means of a tractor and dozer blade cleared the area in the vicinity of the tree of all readily combustible material. In addition, he used a mounted six hundred gallon tank of water to spray the surrounding area so as to minimize the risk of the fire escaping. 1. [1963] W.A.R. 102. When Coombes, the tree faller, arrived the tree was burning fiercely in the fork and the bark from the ground up was on fire. The fire near the base of the tree was damped out and the tree was cut down. It is unnecessary to elaborate the details of what then happened for the learned trial judge found—and his finding was not challenged—that up to this point the respondent's conduct in relation to the fire was not open to question. But his Honour also found that if the respondent had exercised reasonable care he could, on the Sunday evening or on the following morning, have put out the fire by the use of water. This, the respondent claimed, he did on the Monday morning. His evidence was to the effect that he doused the fire thoroughly, rolled the logs over and then inspected them on the Monday night, on the Tuesday morning and night, and again on the Wednesday morning. In particular, he said that he spent two hours extinguishing the fire immediately after the departure of two visitors to his property on Monday morning—Mr. and Mrs. Jones. But Mrs. Jones' evidence was that the respondent left the property in his car immediately behind the car in which she was travelling and this evidence was accepted by the learned trial judge. How long the respondent was away from the vicinity of the fallen tree that day does not appear. But it is clear that there was an abundance of evidence fully justifying the finding that he did not, as he alleged, spend any time immediately after the departure of the Jones in extinguishing the fire. It was also established that the respondent was away from the property for a substantial part of the following day, Tuesday. On Wednesday morning the respondent went to work on a part of his property about a quarter of a mile or so distant from the vicinity of the fallen tree and about 12.30 p.m. Jones, who was working with him, drew his attention to smoke visible in a westerly direction. The respondent at once drove to his home and found that fire had burnt out most of the paddock to the west of where the tree had stood and that it had then extended two and one-half miles further to the west. In the course of the afternoon it extended further to the west and in the course of so doing it caused damage to the appellants' property. It is, we think, unnecessary to traverse the evidence in great detail but it is of some importance to notice the weather conditions which prevailed over the relevant period. An officer from the Perth weather bureau was called to give evidence of observations made at Guildford airport which, it was said, would reflect the conditions as they prevailed in the vicinity of the respondent's property. Particulars of these observations were as follows: Date Wind Direction Force Temperature Sunday 26th February N.W. till noon then a little 94 deg. S. of W. till 6 p.m.—then calm Monday 27th February 8 a.m. S. 8 a.m. — 4-6 m.p.h. 86 deg. 10 a.m. S.W. 12 noon — 10-13 m.p.h. 3 p.m. S.W. 8 p.m. — 7-8 m.p.h. 8 p.m. S. Tuesday 28th February Between S.E. to E. all day 10 m.p.h. at 9 a.m., freshening to 20 m.p.h. at 3 p.m., then easing. 97 deg. Wednesday 1st March E.N.E. 9 a.m. — 19 m.p.h. 105.2 deg. 12 noon — 21-22 m.p.h. 1 p.m. — 24 m.p.h. 1-2 p.m. — 25-26 m.p.h. The fire risk on these days was said to be severe and at some times dangerous. In particular it seems to have become dangerous on Wednesday 1st March as both the temperature and the velocity of the wind increased. It will be seen, therefore, that the fire caused by the striking of the tree created a very considerable risk not only to the respondent's property but also to the surrounding countryside. But it is also clear that the fire in the fork of the tree could not be extinguished without felling the tree and there is no question that the respondent was guilty of any lack of care in causing this to be done. Nor, as we have already said, did the learned trial judge find any fault in any of the steps which the respondent took on Sunday 26th February and which were discussed in his Honour's reasons. However he has found explicitly that on the Monday morning he could, by the exercise of reasonable care, have extinguished the fire and he rejected the respondent's evidence that he had taken adequate or reasonable steps to accomplish this. On the contrary he referred to evidence which indicated that the respondent's method of extinguishing a fire of this character was "to burn it out" and that this was inconsistent with his evidence that he had used water to extinguish it immediately after the departure of Jones. We think that there was abundant evidence to justify the finding of the learned trial judge that the respondent might by the exercise of reasonable care have extinguished the fire by the morning of Monday 27th February, and that he did not attempt to do so. Further we are of the opinion that the evidence clearly demonstrates that he did not, at any time thereafter, take any steps which could be regarded as reasonable in the circumstances then prevailing to prevent the fire from spreading. Before proceeding to consider the questions of law which were debated we should mention that the respondent challenged the finding of the learned trial judge that the fire which caused the damage had spread from or found its origin in the fallen tree. Upon this point counsel for the respondent said all that could be said but we think that the evidence leaves only one conclusion open. It is, in our view, beyond question that the fire spread from the fallen tree and we shall proceed to consider the case on that basis. We add that in reaching this conclusion we have, because of its unsatisfactory character in some respects, disregarded the evidence to the effect that early on the morning of Thursday 2nd March the respondent lit a number of fires on land immediately to the north of his property. In the circumstances to which we have briefly referred his Honour held that the respondent was under no liability to compensate the appellants for the damage which they sustained. In particular he held that there was no liability in nuisance because the fire had been caused by lightning and the respondent could not be said to have thereafter "adopted" or "continued" it. In his opinion the appellants' claim was not supported by the views expressed in Sedleigh-Denfield v. O'Callaghan [1] or by the dissenting judgment of Scrutton L.J. in Job Edwards Ltd. v. Birmingham Navigations [1] which received the approval of a number of members of the House of Lords in the former case. Further, he was of the opinion, that the occurrence of the fire, caused, as it was, by lightning, did not impose upon the respondent any duty of care with respect to his neighbours. In stating this proposition his Honour referred to Batchelor v. Smith [2] and Havelberg v. Brown [3] and he rejected the decision in the New Zealand case of Boatswain v. Crawford [4] . The proposition, his Honour said, accorded with the "broader rule that a landowner is under no liability for anything which happens to or spreads from his land in the natural course of affairs, if the land is used naturally" [5] . 1. [1940] A.C. 880. 2. [1924] 1 K.B. 341. 3. (1879) 5 V.L.R. (L.) 176. 4. [1905] S.A.L.R. 1 5. [1943] N.Z.L.R. 109. 6. [1963] W.A.R., at p. 108. The case of Batchelor v. Smith [2] was decided upon demurrer and the precise question which it decided was that the law did not impose upon the occupier of land any duty to extinguish a fire on the land which had been caused by spontaneous combustion. "No duty" it was said, "is cast on the defendant; he does nothing; he remains passive" [6] . But it was added "Had he interfered in any way, he might possibly have rendered himself liable" [7] . The decision in Havelberg v. Brown [3] expressly followed that in the earlier case but we doubt whether the broad proposition upon which these cases rest can stand consistently with the relatively modern development of the concept of negligence. In particular, it is inconsistent in principle with the dissenting observations of Scrutton L.J. in Job Edwards Ltd. v. Birmingham Navigations [1] . He said [8] : "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 [9] . 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" [1] . And at a later stage he expressed his agreement with a passage from the 5th ed. (1920) of Salmond on Law of Torts—"When a nuisance has been created by the act of a trespasser, or otherwise without the act, authority, or permission of the occupier, the occupier is not responsible for that nuisance unless, with knowledge or means of knowledge of its existence, he suffers it to continue without taking reasonably prompt and efficient means for its abatement" [2] . This passage and the observations of Scrutton L.J. received the express approval of Viscount Maugham, Lord Wright and Lord Romer in the Sedleigh-Denfield Case [3] . Further, the proposition advanced in that case—namely, that if a trespasser comes on to land and creates a nuisance the occupier of the land is not liable unless he either adopts the act of the trespasser or does something in the nature of ratification after he becomes aware of its existence—was unanimously rejected. On the contrary, the effect of their Lordships' reasons was that an occupier, with knowledge or presumed knowledge of the existence of a state of affairs on his land which is a potential nuisance but which has been created by a trespasser, is, nevertheless, liable in the event of damage resulting therefrom to the lands of his neighbours if by the exercise of reasonable care the damage would have been avoided. This proposition, stated as it is, relates, in terms, only to potential nuisances brought into existence by a trespasser and no doubt it was so stated in order to deal with the particular facts of that case. But we can see no distinction relevant to the question of liability between potential nuisances created by trespassers and potential nuisances coming into existence "otherwise without the act, authority, or permission of the occupier". Indeed, in the passage already cited from Salmond the test of liability is propounded as a common one for nuisances of either character and the same notion is apparent in the final proposition as stated by Rowlatt J. in Noble v. Harrison [4] when he said: "The result is that a person is liable for a nuisance constituted by the state of his property: (a) if he causes it; (b) if by the neglect of some duty he allowed it to arise; and (c) if, when it has arisen without his own act or default, he omits to remedy it within a reasonable time after he did or ought to have become aware of it" [5] . These propositions were referred to with evident approval by Dixon J., as he then was, in Torette House Pty. Ltd. v. Berkman [6] . Again, the test of liability propounded in the Sedleigh-Denfield Case [1] was applied in Slater v. Worthington's Cash Stores (1930) Ltd. [2] where the Court of Appeal held the occupier of premises liable for damage caused by a heavy fall of snow which had accumulated on the roof of premises during a severe snow storm which had come to an end some four days previously. The same test has also been applied by the Supreme Court of New Zealand in two cases concerning fires on country properties—Boatswain v. Crawford [3] and Landon v. Rutherford [4] . In neither case was it alleged or proved that the defendant had originated the fire; the basis upon which the occupier in each case was held liable was that the damage complained of by the plaintiff could have been avoided by the exercise of reasonable care on the former's part. These later decisions were in accordance with the test of liability which we think has been authoritatively established and which is correctly stated in the brief passage we have quoted from Salmond and, accordingly, we think the learned trial judge erred on this branch of the case. We notice in passing, however, that in the lastmentioned case the occupier admitted that he made no attempt to contain the fire and that Fell J. held that, in those circumstances, it was for him "to prove that it was impossible to do anything by taking reasonably prompt and efficient means to stop it spreading" [5] . We do not agree with this observation for in order to establish liability for negligence the plaintiff must always prove that the damage of which he complains was caused by the breach of duty alleged. 1. (1879) 5 V.L.R. (L.) 176. 2. (1879) 5 V.L.R. (L.), at pp. 178, 179. 3. (1879) 5 V.L.R. (L.), at p. 179. 4. [1905] S.A.L.R. 1 5. [1924] 1 K.B. 341. 6. [1923] 1 K.B., at p. 357. 7. (1868) L.R. 3 H.L. 330. 8. [1924] 1 K.B., at pp. 357, 358. 9. [1924] 1 K.B., at p. 360. 10. [1940] A.C., at pp. 893, 894, 910, 913. 11. [1926] 2 K.B. 332. 12. [1926] 2 K.B., at p. 338. 13. (1940) 62 C.L.R. 637, at p. 657. 14. [1940] A.C. 880. 15. [1941] 1 K.B. 488; [1941] 3 All E.R. 28. 16. [1943] N.Z.L.R. 109. 17. [1951] N.Z.L.R. 975. 18. [1951] N.Z.L.R., at p. 978. In the present case the learned trial judge referred to a number of cases relating to such things as the natural growth and spread of wild thistles (Giles v. Walker [6] ) and prickly pear (Sparke v. Osborne [7] ) and the spread of rabbits (Anderson v. Lockyer [8] ) but we do not think that these cases throw any light on the problem in this case. No principle was enunciated in the first case, in the second the plaintiff's claim rested, not upon any allegation of negligence, but upon an assertion of strict liability whilst in the third the claim, which failed, was that there had been a breach of a statutory duty on the part of the defendant giving rise to private right of action. 1. (1890) 24 Q.B.D. 656. 2. (1908) 7 C.L.R. 51. 3. (1950) 52 W.A.L.R. 60. This is enough to dispose of the case but it should be observed that the claim of the appellants does not rest merely upon the allegation that there was on the part of the respondent a failure to take reasonable steps to extinguish or prevent the spread of the fire in its original location in the fork of the tree. The respondent did, in fact, take some steps and these were initially taken as much for the preservation of his own property as of that of his neighbours. Indeed the taking of these steps was a measure which any prudent occupier would have adopted in the ordinary management of his property. It is, of course, a matter of general knowledge that trees in country areas are not infrequently set on fire by lightning and that, when observed, steps are taken to extinguish them or to contain them where possible as a matter of course. But when the tree in question here was cut down a hazard of a different character was created and it is beyond doubt that the respondent was under a duty to use reasonable care to prevent it causing damage to his neighbours in the countryside. The finding that, in the circumstances prevailing, he failed to discharge this duty with the result that the appellants sustained the damage of which they complain is we think unassailable. We add that on this view it is of no consequence whether his liability rests in negligence or nuisance. For these reasons the appeal should, in our opinion, be allowed and the case remitted to the Supreme Court for the purpose of assessing damages. Having reached this conclusion it is unnecessary for us to consider the further ground upon which the appellants based their claim, that is to say, liability for the breach of a statutory duty imposed upon the respondent by s. 28 of the Bush Fires Act, 1954-1958. We are, however, inclined to the view that the decision of the learned trial judge on this point was correct. It seems to us that it is impossible to regard a breach of s. 28 as giving rise to a cause of action for damages. It will be observed that the obligation imposed by that sub-section upon the occupier of land to "take all possible measures at his own expense to extinguish" a bush fire burning on his land applies only to fires burning on any land during "the restricted burning times" and during the "prohibited burning times" and only where the "bush fire is not part of the burning operations being carried on upon the lands in accordance with the provisions" of the Act. "Restricted burning times," by definition, means the period of time from the first day of October in any year to the next following thirty-first day of May and "prohibited burning times" are those times declared by the Governor by notice published in the Gazette (s. 17). However, the operation of a declaration under this section may be suspended by the Minister so far as it extends to any particular land and the suspension may be subject to any conditions specified by the Minister. Further, it should be observed that many classes of burning operations may during the prohibited burning times be carried on "in accordance with the provisions of the Act". We mention as instances the provisions of ss. 22 (2), 23, 24, 24A, 25 and 26. Accordingly, the suggestion that it was intended that a breach of s. 28 resulting in damage should give rise to a private right of action involves the notion that the right of action should, in part, be seasonal in character, in part, dependent upon the existence of a declaration under s. 17 and no relevant suspension thereof, and upon the fire in question not constituting "burning operations in accordance with the provisions of the Act". We would find it difficult to discover in legislation of this character an intention that s. 28 was intended to create rights inter partes in relation to the control of fires or that a breach of its provisions should give rise to a private right of action. After these reasons were prepared the respondent sought leave to reopen the appeal on the ground that fresh evidence had been discovered relating to the condition of the fallen tree and the immediate surroundings on the evening of Monday 27th February 1961. It was submitted that this evidence could not with reasonable diligence have been available at the trial and that the principles which, in the circumstances, we should apply, are those which would guide an appellate court in a motion for a new trial on the ground of the discovery of fresh evidence. Accordingly, we were asked to direct a general new trial should the judgment in favour of the respondent be set aside. On a mere reading of the affidavits filed in support of the application we are not greatly impressed by the evidence said to have been discovered or the reason advanced why it was not forthcoming at the trial and think there is a great deal of substance in the submissions made on behalf of the appellants. Nevertheless, since the existing judgment must be set aside and the issue of damages remains to be determined, we think that in the circumstance the appropriate course for us to follow is to leave it to the trial judge to determine whether at this late stage—issues of fact relating to liability having been litigated and pronounced upon—he should permit the respondent to reopen his case in order to adduce the fresh evidence. The reason why it was not available at the trial and the character and cogency of the evidence will, of course, be material matters for his consideration. That being so we think that we should set aside the order and judgment of the Supreme Court and remit the case to the Supreme Court for the assessment of damages if an application made to the trial judge by the respondent to reopen his case on the issue of liability be rejected or, if, notwithstanding the reception of the fresh evidence in question, or otherwise, the decision of the Court on the issues of fact relating to liability remains unchanged. Windeyer J. I agree in the conclusions of my brothers Taylor and Owen. But, as the well-considered arguments that we heard raised some fundamental questions, I shall state my reasons for myself. The respondent is an elderly man who lived alone on his grazing property of some six hundred acres of lightly timbered country near Gidgegannup. On Saturday, 21st February 1961, a tall red gum tree on his land was struck by lightning, and set on fire in a fork some eighty feet or more up from the ground. The respondent became aware of this next morning. Appreciating the danger of the fire spreading, he took prompt action. He telephoned the fire control officer under the Bush Fires Act, 1954-1958 W.A., the Road Board secretary and others. Through them he obtained assistance to fell the tree, so that the fire could be brought under control. While awaiting the arrival of the tree feller he cleared the ground near the tree with a bulldozer. The tree was cut down. As it fell sparks ignited another tree, which then fell down. As a result, there were two logs burning on the ground. Action was taken to contain both fires. Inflammable material, branches and debris lying nearby, was thrown upon them. This no doubt increased them temporarily; but it diminished the risk of their spreading, and was done for that purpose. For that purpose too, the respondent, with the help of a visitor, watered the ground near the fires from a six hundred gallon tank mounted on a trailer until, unfortunately, the stopcock became broken, rendering this equipment no longer serviceable as a sprinkler. The respondent apparently thought that the logs could then safely be left to burn themselves out. But on the next day they were still burning. The respondent said that he then put water on them and put them out. But his Honour did not accept his evidence that he did this. There was evidence from which his Honour could infer that in fact the logs were, to the knowledge of the respondent, still alight on the Tuesday, when he went off to Perth for the day, apparently thinking that there was no risk in leaving, or taking the risk of doing so. A Western Australian red gum, once alight, may burn or smoulder for a long time by reason of the resinous gum from which it derives its name. Tuesday and Wednesday were very hot days. On Wednesday the temperature reached 105 degrees, and a strong easterly wind was blowing. In the afternoon of that day the respondent, while away from his homestead on another part of his property, had his attention drawn to smoke. He returned to find a large bush fire. Part of his land was burnt out; and the fire had already gone about a mile and a half towards the west, and was travelling fast. In the ultimate result hundreds of acres, over many miles of the countryside, were devastated. The appellants are landowners, whose house and other property were destroyed. His Honour found, and on the evidence the finding was clearly justified, that the bush fire began from the burning logs on the respondent's land. He found too that until Sunday afternoon, and probably until late on that day, the measures that the respondent took to control the fire that the lightning had started were "unexceptionable, taking into account all the circumstances and the fire-fighting equipment available". But his Honour considered that thereafter the respondent was at fault. He said: "I am satisfied that had he taken reasonable care he could, on the Sunday evening, or at latest early on the next morning, have put out the fires by using water on them". Nevertheless, he held that he had committed no breach of duty to the plaintiffs, the appellants, and dismissed the action. The appellants' case is that on his Honour's findings they are entitled in law to damages. From their point of view it matters not under what rubric of the law of torts their claim should be placed. But the case is not one which can be decided without regard to legal categories and classifications of wrongdoing. In the argument the case for the plaintiffs was discussed as depending, alternatively or cumulatively, upon the common law as modified by the Fires Act of 1774, upon nuisance, negligence, the rule in Rylands v. Fletcher [1] , and breach of a statutory duty under the Bush Fires Act. The several theorems of law thus propounded may be considered separately. But, in considering each, it is necessary to bear in mind that the law of torts is developing to-day, as the common law has developed in the past: new situations are being subsumed under rules and principles that are proving extensive rather than restricted. This is largely the result of the expansive scope of the tort of negligence to-day. 1. (1866) L.R. 1 Ex. 265; (1868) L.R. 3 H.L. 330. (i) The common law and the Act of 1774: The early common law, or custom of the realm, made a man responsible in an action of case if his fire spread and burnt his neighbour's house. Much that appears in the cases collected in Comyn's Digest under the heading "Action upon the Case for Negligence: in keeping his fire" is now obsolete, but the main principles of the common law concerning fire still stand in the background of the law to-day. The earliest case, and the one often referred to in later reports, is Beaulieu v. Finglam [1] . That it still has vitality appears from the quotation of the Year Book made by Lord Goddard C.J. in Balfour v. Barty-King [2] , and recently by McGregor J. in Eriksen v. Clifton [3] . The averment was that the defendant had so negligently kept his fire (ignem suum tam negligenter custodivit) that the goods of the plaintiff were burnt. What weight should be put upon the word negligenter there, and whether it was traversable, are questions that have been debated by very learned writers. We do not have to decide the issue between Wigmore and Winfield; for, whatever it meant, the word did not import the modern idea of tortious negligence. It is therefore enough to say that, as Sir Percy Winfield showed, it is not correct that the spread of a fire created at common law an absolute liability altogether irrespective of any fault of the man from whose land it spread. The rule seems rather to have been that a householder was responsible for his fire—and that meant any fire lighted by an inmate of his house: but he was not responsible for a fire started by a trespasser. Although not absolute, this liability was rigorous; and counsel feelingly protested in Beaulieu v. Finglam [1] : "the defendant will be undone and impoverished all his days if this action is to be maintained against him, for then twenty other such suits will be brought against him for the same matter". To which Thirning C.J. replied: "What is that to us. It is better that he be utterly undone than that the law be changed for him". And for three hundred years it continued virtually unchanged, as can be seen from Turberville v. Stampe (1697). That case, important in the development of the law of vicarious liability for the acts of a servant as well as in relation to fire, is reported in many places: by Lord Raymond, Salkeld, Comyns, Comberbach, Carthew, Skinner, and elsewhere. Lord Raymond [4] gives the best report of the argument; but the record is set out in full by Salkeld [5] . A fire had been lit to burn off stubble in a field. The majority of the court said "a man ought to keep the fire in his field, as well from the doing of damage to his neighbour, as if it were in his house, and it may as well be called suus the one as the other": but it would be relevant to prove that "a wind and tempest arose and drove it into his neighbour's field". The ideas of remoteness of damage, and of unforeseen occurrences breaking the sequence of cause and consequence, were coming into the law. But the general principle remained: every man was liable for damage caused by his fire whether it was lit by him or by his servant. Parliament at last took a hand. The Act 6 Anne c. 31 (1707), continued by 10 Anne c. 14, provided that no action should be had against any person "in whose house or chamber any fire shall accidentally begin, or any recompense be made by such person for any damage suffered or occasioned thereby, any law usage or custom to the contrary notwithstanding". Blackstone stated the reason and policy of this as "for their own loss is sufficient punishment for their own or their servants' carelessness". The provision was continued in later enactments, culminating in the Fires Act, 1774, s. 86, which provided that no action shall be against any person "in whose house, chamber, stable, barn or other building or on whose estate any fire shall accidentally begin ". This Act, it has generally been accepted, became part of the law of Western Australia on the foundation of the Colony; and it has not been repealed there. In terms it might seem to apply to this case, as it has been construed as applying to country lands as well as to houses in cities and towns. But, although some reliance was put upon it in the argument, I do not think it directly affects the question here. True, the fire in the tree did "accidentally begin"; for that phrase has been held to mean a fire that begins by inevitable accident, as distinct from one caused intentionally or by the negligence of someone for whom the landowner was responsible: Filliter v. Phippard [1] . But the effect of the statute is narrowed by the decisions that it does not apply when a fire, although beginning without negligence, spreads as the result of negligence: Musgrove v. Pandelis [2] , and see Job Edwards Ltd. v. Birmingham Navigations [3] and Eastern Asia Navigation Co. Ltd. v. Fremantle Harbour Trust Commissioners [4] per Fullagar J. [5] . And that, according to the finding of the learned trial judge was what happened here. But putting the statute aside does not mean that we are thrown back to the rigorous rule of the mediaeval common law. This Court has held that the old rules have been absorbed into the principle of Rylands v. Fletcher [6] ; and that the strict liability of the common law is subject to the qualifications of and exceptions to that principle: Bugge v. Brown [7] ; Hazelwood v. Webber [8] . 1. (1401) Y.B. 2 Hen. 4 f. 18. p. 16. 2. [1957] 1 Q.B. 496. 3. [1963] N.Z.L.R. 705. 4. (1401) Y.B. 2 Hen. 4 f. 18. p. 16. 5. (1697) 1 Ld. Raym. 264 [91 E.R. 1072]. 6. (1697) 2 Salk 726 [91 E.R. 607]. 7. (1847) 11 Q.B. 347 [116 E.R. 506]. 8. [1919] 2 K.B. 43. 9. [1924] 1 K.B. 341. 10. (1951) 83 C.L.R. 353. 11. (1951) 83 C.L.R., at pp. 393, 394. 12. (1866) L.R. 1 Ex. 265; (1868) L.R. 3 H.L. 330. 13. (1919) 26 C.L.R. 110, at pp. 114, 115. 14. (1934) 52 C.L.R. 268. (ii) Rylands v. Fletcher: The attempt made at the trial to base the plaintiffs' claim on the rule in Rylands v. Fletcher [1] failed; and the argument before us virtually conceded that it rightly failed. Fire is a thing likely to do mischief if it escapes. Therefore, fire can come within the rule in the form in which it was enunciated by Blackburn J.; although the complexity of the distinctions between a natural and non-natural user of land, that has resulted from the words Lord Cairns used, and between dangerous and non-dangerous things, makes the application of the rule uncertain in some cases of fire and explosion: see Read v. J. Lyons & Co. Ltd. [2] , and Wise Bros. Pty. Ltd. v. Commissioner for Railways (N.S.W.) [3] . In the present case none of those uncertainties arises directly; but Rylands v. Fletcher [1] is excluded 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. Therefore the appellants could only use the principle of Rylands v. Fletcher [1] as somewhat distantly akin to this case, and by way of an approach to the proposition that their cause of action was in nuisance. To this I turn. 1. (1866) L.R. I Ex. 265; (1868) L.R. 3 H.L. 330. 2. [1947] A.C. 156. 3. (1947) 75 C.L.R. 59. 4. (1866) L.R. I Ex. 265; (1868) L.R. 3 H.L. 330. 5. (1866) L.R. I Ex. 265; (1868) L.R. 3 H.L. 330. (iii) Nuisance: A nuisance has been defined as an "unlawful interference with a person's use or enjoyment of land, or of some right over, or in connexion with it". This compendious description from Winfield on Tort 6th ed. (1954) p. 536 states the essence of nuisance as a tort. But some particularity is required to give content to the phrase "unlawful interference". Generally speaking the term "nuisance" denotes a state of affairs that is either continuous or recurrent. It is, therefore, somewhat misleading to use the word "nuisance" of a situation from which harm may occur if care be not exercised, but from which no actual harm is currently occurring. A thing that dangerously overhangs a highway, and which may fall at any moment, is however commonly called a nuisance. It currently and continuously interferes with the safe enjoyment of a public right of way and is thus a public nuisance. But in the present case what the appellants relied upon was the law of private nuisance. And a fire that is presently harmless is not a nuisance, although it may be fraught with danger and arouse apprehensions of harm. It is not that the law ignores prospective nuisances or threatening dangers. It does not, for their existence may be a ground for an injunction: Attorney-General v. Corporation of Manchester [4] . And there can be no objection to speaking of a "potential nuisance", as was done in this case, provided that it be remembered that the invasion of the common law rights of an owner or occupier of land does not occur until he suffers harm: cf. Torette House Pty. Ltd. v. Berkman [1] , per Dixon J. [2] . The matter may seem to be one of classification and terminology, rather than of substance; but the boundaries of the law of nuisance are indefinite enough without allowing the word to beg the question. It is nearly a hundred years since Erle C.J. said of nuisance, in a judgment which, because of his resignation, was never delivered: "This cause of action is immersed in undefined uncertainty The maxim, " sic utere tuo ut alienum non laedas ", is no help to decision, as it cannot be applied till the decision is made; and the use of the word "nuisance" in the discussion prolongs the dispute, because it means both annoyance that is actionable, and also that which is not actionable; and where the question is whether an annoyance is actionable, the word "nuisance" introduces an equivocation which is fatal to any hope of a clear settlement.": Brand v. Hammersmith and City Railway Co. [3] . 1. [1893] 2 Ch. 87. 2. (1940) 62 C.L.R. 637. 3. (1940) 62 C.L.R., at pp. 657, 658. 4. (1867) L.R. 2 Q.B. 223, at p. 247. One argument addressed to the learned trial judge, as it was to us, was that, whether the fire be considered as a present annoyance or as prospectively harmful, this case was altogether outside the law of nuisance, because, it was said, an action of nuisance arises out of some active use that a man makes of his land, liability being commonly attributed to the maxim sic utere tuo : and, it was said, the spread of the fire in this case was not the result of any use by the respondent of his land. This is too narrow a view. An occupier of land who passively suffers a nuisance to continue may be liable although he did not originally create it. Moreover it is not an essential element in liability for a nuisance that it should emanate from land belonging to the defendant, although commonly it does: Esso Petroleum Co. Ltd. v. Southport Corporation [4] ; affirmed [5] . 1. [1953] 2 All E.R. 1204, at p. 1207. 2. [1956] A.C. 218. The respondent had, however, a stronger answer to the case in nuisance. It was that the fire was not something for which he could be held responsible: he did not start it: he did not increase the danger of it: he did nothing to make himself responsible for it: all that he did was done with a view to making it harmless. The appellants sought to meet this by saying that, although the respondent had not created the nuisance, or potential nuisance, he had continued it. They relied upon the well-known statement by Viscount Maugham in Sedleigh-Denfield v. O'Callaghan [1] , that "an occupier of land "continues" a nuisance if with knowledge or presumed knowledge of its existence he fails to take any reasonable means to bring it to an end though with ample time to do so" [2] . There an artificial structure, a drain which was in a defective state, gave rise to a nuisance when it rained. The defendant had not constructed the drain. But he suffered it to remain defective. He did not take any steps to remedy it. He thus adopted or continued it. But here the respondent did take steps to eliminate the potential nuisance. They proved ineffectual it is true. But does that mean that he adopted or continued the fire so as thereby to become responsible for it and liable for the harm it might do? The appellants rely heavily upon the remarks of Scrutton L.J. in Job Edwards Ltd. v. Birmingham Navigations [3] , in his judgment which was approved by the House of Lords in Sedleigh-Denfield's Case [1] . 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; and then if he does nothing, he has "permitted it to continue", and become responsible for it" [4] . The Supreme Court of South Australia has suggested that the rationale of this is that "the risk might be so plain, and the remedy so easy and so obvious, that anyone would say that the failure to deal with the situation was equivalent to approval, and that, by failing to take this step, the landowner had continued or adopted the fire": How v. Jones [5] . That may explain a case where nothing of any significance was done. But it seems artificial in the case of a man who takes steps, although in the result ineffectual, to eliminate the danger. Trying to get rid of a thing can hardly be evidence of approval of it. Instead of imputing to the respondent an intention contrary to his real intent, the straightforward approach, in a case such as this, seems to me to be to ask: was he not liable in negligence? The essential question then is not: did the respondent continue the fire as a nuisance? It is: was he negligent in not rendering it harmless? 1. [1940] A.C. 880. 2. [1940] A.C., at p. 894. 3. [1924] 1 K.B. 341. 4. [1940] A.C. 880. 5. [1924] 1 K.B., at p. 357. 6. [1953] S.A.S.R. 82, at p. 87. (iv) Negligence: The distinction between nuisance and negligence is not altogether clear cut. Until the recognition in modern times of negligence as a tort in itself, many actions of case which we would to-day say were based on negligence were described as being for nuisances. The cases collected under "nuisance" in the third edition (1868) of Bullen and Leake show this. Negligence is not a necessary element in nuisance, although it may be an ancillary element in some forms of nuisance: see Jacobs v. London County Council [1] per Lord Simonds [2] and Sedleigh-Denfield's Case [3] , per Lord Wright [4] . The distinction between nuisance and negligence as separate torts may be of little, if any, importance for the ultimate decision of this case. But it is of some significance in considering the decisions relied upon in the argument. At the present day, and for present purposes, it may, I think, be stated as follows. 1. [1950] A.C. 361. 2. [1950] A.C., at p. 374. 3. [1940] A.C. 880. 4. [1940] A.C., at p. 904. In nuisance liability is founded upon a state of affairs, created, adopted or continued by one person (otherwise than in the reasonable and convenient use by him of his own land) which, to a substantial degree, harms another person (an owner or occupier of land) in his enjoyment of his land. In negligence liability is founded upon the negligent conduct of one person causing, to any degree, foreseeable harm to the person or property of another person (not necessarily an owner or occupier of land) to whom a duty of care was owed. (v) Duty of care: In the present case the learned trial judge found expressly that, had the respondent taken reasonable care, he could have put out the burning logs. I take it that his Honour meant by this that the respondent did not act as a reasonably careful man, who had a duty to extinguish the fires, would have acted in the circumstances. That, the appellants say, is a finding of negligence on which they are entitled to judgment, and again they refer to the illustration that Scrutton L.J. gave, in the passage I have quoted above, of stamping out a fire. His Lordship there recognized that, although the case had been debated as one of the duty to abate a nuisance, his proposition made liability depend on negligence. And he said: "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" [5] . 1. [1924] 1 K.B., at p. 358. Counsel for the respondent challenged the validity of this proposition, or at least its application in this case. The respondent, he urged, had no legal duty to the appellants to extinguish the burning logs or render them harmless. His argument led him to some observations concerning the concept of duty of care, as an element in the tort of negligence. This is a subject on which there is now a large body of learned academic literature. We were referred to some of the articles and text-books. I have read them and others. But it seems to me unnecessary to go far into the matter here. It may be that insistence upon a duty of care as a separate element in liability for negligence is, in theory, unnecessary; for it may be comprehended in the idea of negligence itself, an act or omission being careless only when a reasonable man would appreciate, if he thought about the matter, that it could have harmful consequences. As long ago as 1897 Holmes J. suggested that the idea of a duty of care was a superfluous addition to the requirement of reasonable care: (1897) 10 Harvard Law Review p. 47. And Professor Buckland, fittingly enough as a Roman lawyer, thought the duty of care "an unnecessary fifth wheel on the coach, incapable of sound analysis and possibly productive of injustice". He realized, however, that it was "certainly a part of our law": (1935) 51 Law Quarterly Review 637. Sir Percy Winfield took the same view in an article in the (1934) 34 Columbia Law Review, pp. 41-66 reprinted in his Select Legal Essays (1952) pp. 70-95. The matter is now beyond purposeful debate, except as an exercise in juristic philosophy. The concept of a duty of care, as a prerequisite of liability in negligence, is embedded in our law by compulsive pronouncements of the highest authority. And it may well be that it could not be otherwise, if the law of negligence is to have symmetry, consistency and defined bounds, and its application in particular cases is to be reasonably predictable. It is worth noting that, although the duty of care has no place as a separate element in the civil law of fault, Continental courts have had to meet the same problem as the common law courts have; and they have dealt with it in somewhat the same way. Thus it has been said that "French lawyers have been brought to the point of acknowledging the need for something not very far removed from the English duty of care": Lawson, Negligence in the Civil Law (1950) p. 31; and see Millner, Contrasts in Contract and Tort, in Current Legal Problems 1963, at p. 85 cf. Ryan, An Introduction to the Civil Law (1962), pp. 114, 115. In the recent, and most important, case of Hedley Byrne & Co. Ltd. v. Heller & Partners Ltd. [1] , Lord Pearce said: "The law of negligence has been deliberately limited in its range by the courts' insistence that there can be no actionable negligence in vacuo without the existence of some duty to the plaintiff" [1] . But it would, I consider, be wrong to conclude from that, and from descriptions, such as a "control device", that appear in textbooks that the controlling element, duty of care, was imposed upon the law of negligence in order to confine its twentieth century expansiveness. Rather, it seems to me, it had an earlier origin and grew up almost inevitably as negligence grew to be a separate tort. For example, in Bacon's Abridgment 6th ed. (1807) it is said under "Action on the Case" that: "In some cases an injury happens to a man in his property, by the neglect of another; yet if by law he was not obliged to be more careful no action will lie". And throughout the nineteenth century the courts held in numerous cases that on particular facts there was a duty on which an action of case could be founded—whether it was then classified as in nuisance or negligence is immaterial. As an illustration, it is enough to refer to Brown v. Mallett [2] . When Lord Esher, then Brett M.R., made his famous generalization in Heaven v. Pender [3] , his purpose was to state the circumstances in which "a duty arises to use ordinary care and skill " [4] . But we cannot, having regard to what has been decided in other cases, decide whether in a given case there is a duty of care simply by resorting to Lord Esher's generalization, even when qualified by the notion of proximity, not in the sense of physical nearness but in the metaphysical sense defined by Lord Atkin in Donoghue v. Stevenson [5] . 1. [1964] A.C. 465. 2. [1964] A.C. 465. 3. (1848) 5 C.B. 599 [136 E.R. 1013]. 4. (1883) 11 Q.B.D. 503. 5. (1883) 11 Q.B.D., at p. 509. 6. [1932] A.C. 562, at pp. 580, 581. How then are we to decide whether the respondent was under a duty of care in this case? His counsel, having raised the question, answered by quoting du Parcq L.J. in Deyong v. Shenburn [6] : "There has to be a breach of a duty which the law recognizes, and to ascertain what the law recognizes regard must be had to the decisions of the courts" [7] . Thus, the argument ran, if no court has said that there is a duty in a case such as this, then we cannot now say there is such a duty. But that extreme view of the "wisdom of our ancestors" cannot be accepted to-day. Lord Macmillan's words will bear quoting once again: "The grounds of action may be as various and manifold as human errancy; and the conception of legal responsibility may develop in adaptation to altering social conditions and standards. The criterion of judgment must adjust and adapt itself to the changing circumstances of life. The categories of negligence are never closed": Donoghue v. Stevenson [8] . And I would respectfully add a reference to Lord Devlin's speech in Hedley Byrne's Case [1] , in particular to two passages. One, the sentence: "English law is wide enough to embrace any new category or proposition that exemplifies the principle of proximity" [1] . The other: "Now, it is not, in my opinion, a sensible application of what Lord Atkin was saying for a judge to be invited on the facts of any particular case to say whether or not there was "proximity" between the plaintiff and the defendant. That would be a misuse of a general conception and it is not the way in which English law develops" [1] . The warning that is implicit in this is important. We are concerned with categories, not with the special facts of a particular case. 1. [1946] 1 K.B. 227. 2. [1946] 1 K.B., at p. 233. 3. [1932] A.C., at p. 619. 4. [1964] A.C. 465. 5. [1964] A.C. 465. 6. [1964] A.C. 465. This case is not one in which the obligation to use care and skill arises from an undertaking to do some work for the benefit of another. In a case of that kind an obligation to exercise due care and skill arises from the entering upon the work, whether for reward or gratuitously. But here what the respondent did in relation to the fire was not done pursuant to any undertaking to the appellants, nor was it done specifically for their benefit. It did not increase the danger of the fire spreading. Probably it diminished it. It seems to me impossible to say that, because the respondent did something to control the fire, he incurred a liability that he would not have incurred had he done nothing. If that were the law, a man might be reluctant to try to stop a bush fire lest, if he failed in his endeavours, he should incur a liability that he would not incur if he remained passive. The question comes to this: In a case such as this has the occupier of land a duty at common law—I put statutory obligations aside for the moment—to act at all? It was said that we must go to Donoghue v. Stevenson [2] , and that the principle of proximity would supply the answer. Fullagar J. wrote of Donoghue v. Stevenson [2] (in a paper published in the Australian Law Journal (1951) Vol. 25 p. 278): "It was not, of course, intended to make, and it does not make, everything nice and easy". 1. [1932] A.C. 562. 2. [1932] A.C. 562. Lord Atkin's well-known generalization explains the scope of a duty of care, that is to say it states who can complain of a lack of care when an obligation of care exists. But I venture to think that it is a mistake to treat it as providing always a complete and conclusive test of whether, in a given situation, one person has a legal duty either to act or to refrain from acting in the interests of others. The very allusion shows that it has not this universal application. The priest and the Levite, when they saw the wounded man by the road, passed by on the other side. He obviously was a person whom they had in contemplation and who was closely and directly affected by their action. Yet the common law does not require a man to act as the Samaritan did. The lawyer's question must therefore be given a more restricted reply than is provided by asking simply who was, or ought to have been, in contemplation when something is done. The dictates of charity and of compassion do not constitute a duty of care. The law casts no duty upon a man to go to the aid of another who is in peril or distress, not caused by him. The call of common humanity may lead him to the rescue. This the law recognizes, for it gives the rescuer its protection when he answers that call. But it does not require that he do so. There is no general duty to help a neighbour whose house is on fire. The question in this case, however, is not whether a man must aid another who is in distress or rescue him from a peril. It is whether he must try to forestall and prevent a peril. A man who, while travelling along a highway, sees a fire starting on the adjacent land is not, as far as I am aware, under any common law duty to stop and try to put it out or to warn those whom it may harm. He may pass on, if not with a quiet conscience at least without a fear of legal consequence. Has the occupier of land a legal duty to his neighbour in respect of a fire that he finds on his side of the boundary fence, but none in respect of a fire that he sees on his neighbour's land just across the boundary, assuming in each case that he realized what might be the consequences to his neighbour of his own inaction? If so, on what principle of the law of negligence does the distinction depend? I do not find such questions easy. The doctrine of proximity does not give the answer, because the question assumes both physical proximity and the metaphysical proximity of Lord Atkin's doctrine. But we may, I think, push such troublesome problems into the background. The trend of judicial development of the law of negligence has been, I think, to found a duty of care either in some task undertaken, or in the ownership, occupation, or use of land or chattels. The occupier of land has long been liable at common law, in one form of action or another, for consequences flowing from the state of his land and of happenings there, not only to neighbouring occupiers, but also to those persons who come upon his land and those who pass by. And, as I have remarked elsewhere, the tendency of the law in recent times has been to lessen the immunities and privileges of landowners and occupiers and to increase their responsibilities to others for what happens upon their land. To hold that the respondent had a duty to his neighbours to take reasonable care to prevent the fire on his land spreading would be in accordance with modern concepts of a land occupier's obligations. If it be a new step in the march of the law—and I do not think that really it is—then it is not a step which we need hesitate to take if nothing stands in the way. New precedents must accord with old principles: but as Lord Abinger C.B. once said, of an action for which no precedent was adduced, "We are therefore to decide the question upon general principles, and in doing so we are at liberty to look at the consequences of a decision the one way or the other": Priestley v. Fowler [1] . 1. (1837) 3 M. & W. 1, at p. 5 [150 E.R. 1030, at p. 1032]. But this is not a case that is bare of all authority. The learned trial judge based his conclusion on certain earlier decisions. The one most directly in point is Batchelor v. Smith [2] , a judgment of the Supreme Court of Victoria (Stawell C.J. and Stephen J.) allowing a demurrer to a declaration alleging damage by spread of fire from the defendant's land. The Chief Justice in giving his reasons said: "It is the duty of any person who originates or brings any matter, animate or inanimate, attended with danger, on his ground, to keep it within due bounds; but there is no authority for the proposition for which the plaintiff contends, that, not having brought it, he must remove it" [3] . Stephen J. concurred, saying: "The foundation of the whole case is that no duty was cast on the defendant to extinguish the fire" [4] . But the declaration had expressly alleged that the defendant, although aware of the danger to his neighbour, allowed the fire to remain burning on his land for the purpose and with the intention of burning and destroying certain stubble, reeds, sawdust, and refuse. In the face of that, it is hard to see why the demurrer was allowed or what answer there was to counsel's argument that the defendant had adopted the fire as his own, and become responsible for any injury resulting from it, just as if he had lighted it himself. The case seems to have been argued on the basis of Rylands v. Fletcher [5] and strict liability, and the decision, when analysed, cannot be regarded as of much weight in the present case. But its dogmatic denial of a duty has not been without effect. It was relied upon in the Supreme Court of New Zealand in Hunter v. Walker [6] , where it was held that the defendant was not liable for the spread of a fire that he had not lighted, although he could have put it out or checked it had he taken timely action. In 1905 it was again relied upon, this time in the Supreme Court of South Australia in Havelberg v. Brown [1] . But when the judgments in that case are studied in the light of the facts, it appears that all that was decided was that there was no absolute duty upon the defendant there to extinguish or control a fire of unknown origin that he had discovered on his land; and that there was no evidence that, in doing what he did in regard to it, he had acted otherwise than as a prudent man would act. That decision really carries the present matter no further. Neither does Black v. Christchurch Finance Co. [2] . That case and also McInnes v. Wardle [3] and the very recent case in New Zealand of Eriksen v. Clifton [4] all turned upon the responsibility of a landowner for the acts of an independent contractor who lit a fire. None of them was concerned with a negligent failure to extinguish or render harmless a fire of unknown origin. However, that question arose directly in Boatswain v. Crawford [5] . There a landowner, although told of a fire on his land, negligently failed to take reasonable steps to extinguish it, as in its early stages he could have done. Johnston J. held that he was liable for the consequences of its spreading beyond his land. He based his conclusion on Job Edwards Ltd. v. Birmingham Navigations [6] and on the approval of it in Sedleigh-Denfield's Case [7] . 1. (1879) 5 V.L.R. (L.) 176. 2. (1879) 5 V.L.R. (L.), at p. 178. 3. (1879) 5 V.L.R. (L.), at p. 179. 4. (1866) L.R. 1 Ex. 265; (1868) L.R. 3 H.L. 330. 5. (1888) 6 N.Z.L.R. 690. 6. [1905] S.A.L.R. 1. 7. [1894] A.C. 48. 8. (1931) 45 C.L.R. 548. 9. [1963] N.Z.L.R. 705. 10. [1943] N.Z.L.R. 109. 11. [1924] 1 K.B. 341. 12. [1940] A.C. 880. The learned trial judge, after a careful review of the cases, came to the conclusion that he should not follow Boatswain v. Crawford [5] . He considered that the correct rule was laid down in Batchelor v. Smith [8] . He was influenced in this view because he said it "accords with the broader rule that a landowner is under no liability for anything which happens to, or spreads from, his land in the natural course of affairs if the land is used naturally" [9] . To that proposition I now turn. 1. [1943] N.Z.L.R. 109. 2. (1879) 5 V.L.R. (L.) 176. 3. [1963] W.A.R., at p. 108. (vi) Things naturally on land: His Honour's statement echoes, but adds some words to, what Lord Goddard C.J. said in Neath Rural District Council v. Williams [10] , where a miscellany of illustration appears. But, like all propositions of a general character, the difficulty is not in its statement but in its application. Is country land in Australia "used naturally" if the occupier, aware of the risk of a bush fire that may cause a disaster to himself and his neighbours, does not act as a reasonably prudent man would act with a view to preventing this? Speaking generally, it is no doubt true that the law does not impose a duty upon anyone to arrest the processes of nature. But we are not concerned with generalities, but with the question whether the occupier of land must take care in the interests of his neighbours to prevent, if by reasonable measures he can, a small fire upon his land spreading and becoming a bush fire. That an answer to that question, arising in Australia to-day, should be sought for in a case about thistledown in England would surely surprise anyone who was not a lawyer. Are we—by examining what courts have said in cases about thistles, prickly pear, the roots of trees and the branches of trees, trees deliberately planted and trees growing naturally, rolling rocks, rabbits, weeds in watercourses, silt in streams, seaweed, snow and surface water—to abstract some general principle, to add qualifications to it, and then to try to apply it to a fire which lightning lit? I do not think so. If this were the way by which to proceed, I would be content to say that I see more resemblance between snow—see Slater v. Worthington's Cash Stores (1930) Ltd. [1] —and fire than I do between fire and thistledown; and that I cannot choose between growing prickly pear and dead seaweed as analogies with fire, and am prepared to discard both. But I do not think this is the way by which we must proceed. Therefore, although I shall refer to some of the cases that were cited, I shall not examine all of them. 1. [1951] 1 K.B. 115, at p. 122. 2. [1941] 1 K.B. 488; [1941] 3 All. E.R. 28. In some of the cases concerning things naturally on land the plaintiff's claim was based on nuisance: in some on negligence: in some on the doctrine of Rylands v. Fletcher [2] . The foundationstone of the doctrinal edifice appears to be Giles v. Walker [3] , the case of the thistles. The action was in negligence. Lord Coleridge C.J. disposed of it by saying: "I never heard of such an action as this. There can be no duty as between adjoining occupiers to cut the thistles, which are the natural growth of the soil. The appeal must be allowed" [4] . Lord Esher agreed. Recently the decision has come in for some criticism. The thistles, although no doubt a natural growth, had only grown on the defendant's land after he had turned it from forest into ploughed land. And in Davey v. Harrow Corporation [5] , Lord Goddard C.J., in delivering the judgment of the Court of Appeal, quoted [6] a remark that Lord Esher had made during the argument, as reported in the Law Times: "This damage is not caused by any act of the defendant. Can you show us any case which goes so far as to say that, if something comes on a man's land for which he is in no way responsible, that he is bound to remove it, or else prevent its causing injury to any of his neighbours?" [1] : Lord Goddard's judgment, in which he acknowledged his indebtedness to an article by Doctor Goodhart (Liability for Things Naturally on the Land (1930) 4 Cambridge Law Journal 13) , went on, quoting directly from that article: "Apparently counsel did not reply, but had he known of Margate Pier and Harbour Proprietors v. Margate Town Council [2] , it would have been a complete answer" [3] . In that case seaweed had been cast ashore by the sea. Left to lie, it became a nuisance to the neighbourhood. It was held that the landowner on whose land it was could be compelled to remove it. 1. (1866) L.R. 1 Ex. 265; (1868) L.R. 3 H.L. 330. 2. (1890) 24 Q.B.D. 656. 3. (1890) 24 Q.B.D., at p. 657. 4. [1958] 1 Q.B. 60. 5. [1958] 1 Q.B., at p. 71. 6. (1890) 62 L.T. 933, at p. 934. 7. (1869) 20 L.T. 564. 8. [1958] 1 Q.B., at p. 72. The only other case to which I need make particular reference is Sparke v. Osborne [4] , the prickly pear case. The decision influenced the American Restatement of the Law of Torts. The facts are well known. An injunction was granted by the Supreme Court of New South Wales to restrain the defendant from allowing prickly pear growing on his land to overhang, and in parts to break down, nine miles of dog-proof fence, thus allowing dingoes to get at the plaintiff's sheep and also causing prickly pear to spread in his land. On the appeal to this Court counsel for the respondent sought to uphold the injunction by a contention that "every owner of land on which there is prickly pear is bound at his peril to prevent its growing on his boundary in such a way as to overhang his neighbour's land": see [5] . It was this absolute proposition that the Court rejected. Griffith C.J. said: "Anyone who has seen prickly pear growing as it grows in some parts of Queensland, for instance, knows that it would be casting an intolerable burden upon the owner of the land if he were compelled to warrant all his neighbours from its spreading into their land" [6] : Doctor Goodhart seems to have thought that this statement of Sir Samuel Griffith, who knew more about Queensland than most men, was inconsistent with his later reference to the Prickly Pear Acts, which require a person to take precautions against the spread of the pest; for in his article he said: "Apparently the burden is not so intolerable when imposed by legislation". It is not, for the legislature recognized that what must be done in a given case depends upon what is practicable, and provided an elaborate administrative control with discretionary power "to endeavour to ensure the common benefit without causing special injustice to the individual", as Griffith C.J. expressed it [1] . The legislative requirements were inconsistent with the absolute common law duty contended for. The case occurred in 1908 when many millions of acres were infested by the rapidly-spreading pear, in many places so heavily infested as to be quite useless. The pest could only be eradicated at a cost which made the task unprofitable. It was not until later that the cactoblastus recovered this "lost province". A learned writer in the Harvard Law Review (1943), vol. 56, p. 772, recognized the ground of the decision: "On ordinary nuisance principles the practical basis for the decision in Sparke v. Osborne would rest on the fact that it would be an "intolerable burden" on the landowner to require him to check this particular pest, so that failure to do so would not constitute an unreasonable use of land, even though considerable injury resulted to the plaintiff and the prickly pear lacked any utility". All the members of the Court in their judgments gave, as reasons for not imposing this burden on the defendant, the facts that the pickly pear had not been brought on to his land by him; that its presence there, and its spread therefrom were the work of nature; that he could not in the circumstances be held liable for a mere non-feasance. They put some reliance on Giles v. Walker [2] and they distinguished the cases of trees overhanging a boundary. But the observations in the judgments concerning exoneration for the consequences of things coming naturally on land should be read in relation to the topic under discussion, that is growing things, trees and noxious plants. Bush fires were not in the mind of the Court at all. And the question of a duty of care did not arise, for the plaintiff did not base his claim on negligence, but on an allegation of strict liability. I therefore put the prickly pear case aside. 1. (1908) 7 C.L.R. 51. 2. (1908) 7 C.L.R., at p. 66. 3. (1908) 7 C.L.R., at p. 59. 4. (1908) 7 C.L.R., at p. 60. 5. (1890) 24 Q.B.D. 656. In the result no more, I think, emerges from the cases than one would have expected, namely that liability for negligence depends ultimately upon a concept of fault and that no man can be held at fault, morally or legally, simply for a happening not caused by any human agency: and that often the law does not hold a man at fault because he does not take any steps to arrest the consequences of such a happening, although he knows they may be harmful to other persons: but that sometimes it does. (vii) Conclusions: In my opinion a man has a duty to exercise reasonable care when there is a fire upon his land (although not started or continued by him or for him), of which he knows or ought to know, if by the exercise of reasonable care it can be rendered harmless or its danger to his neighbours diminished. Of course, if the fire were brought by him upon his land—in the sense of being started or intentionally kept alight there by him or anyone for whose acts he was responsible—his duty would not be merely to take reasonable care: it would be the strict duty of Rylands v. Fletcher [1] . 1. (1866) L.R. 1 Ex. 265; (1868) L.R. 3 H.L. 330. Strong support for the existence of a duty of care to prevent the spread of fire is to be found in the House of Lords' approval in Sedleigh-Denfield's Case [2] of the judgment of Scrutton L.J. in Job Edwards' Case [3] . We do not have to consider what things other than fire might come within his Lordship's general words "a thing which may, if not rendered harmless, cause damage to his neighbours" [4] . The dangers of fire have, from the earliest days of the common law, given rise to special responsibilities; and not only in the common law. In Roman law negligence in watching a fire lit by another was an exception, or apparent exception, to the general rule that mere omissions were outside the Lex Aquilia: see Digest IX, 2, 27. Coming back to modern times: In the United States, although the rule does not seem to be uniform, it is well established in some jurisdictions that a person on whose premises an accidental fire starts must exercise reasonable care to prevent it from spreading after he has notice of the fact, although he has no connexion with its origin: see American Law Reports Annotated, vol. 42, p. 821, vol. 111, p. 1149, vol. 18, 2nd, p. 1097. And that a negligent failure to prevent the spread of a fire of unknown origin creates liability seems to be the rule in Canada also: see Des Brisay v. Canadian Government Merchant Marine Ltd. [5] ; Mainella v. Wilding [6] . 1. [1940] A.C. 880. 2. [1924] 1 K.B. 341. 3. [1924] 1 K.B., at p. 358. 4. [1941] 2 D.L.R. 209. 5. [1946] 2 D.L.R. 749. The New Zealand decision in Boatswain v. Crawford [7] is, I respectfully think, correct. But I would not myself treat the liability which arises in a case such as this as involved in any way with nuisance. One way of stating the ground of liability is that a land occupier is liable if, by his negligence, a potential nuisance is permitted to become an actual nuisance. But I do not think that the liability arising from a negligent failure to extinguish or confine a fire is a liability only to neighbouring landowners or occupiers. Liability in negligence extends to other persons who may be harmed, that is to say, to those who are neighbours in the lawyer's sense as well as those who dwell in the neighbourhood. The grave and widespread consequences of a bush fire may make the liability of a careless individual ruinous for him; but this only emphasizes the seriousness of the duty of care. 1. [1943] N.Z.L.R. 109. (viii) The Bush Fires Act: The Bush Fires Act of Western Australia, s. 28 (1) (a), provides that, where a bush fire is burning on any land in the circumstances set out (and these would, it seems, include the fire in this case): "the occupier of the land shall forthwith, upon becoming aware of the bush fire, whether he has lit or caused the same to be lit or not, take all possible measures at his own expense to extinguish the fire". Failure to do so is punishable by a fine not exceeding £100. It may be that "all possible measures" means all reasonably practicable measures; but, whatever it means, I agree with the learned trial judge in thinking that this provision does not of itself create any civil right. But neither in my opinion does it supplant or limit the common law duty: cf. Edwards v. Blue Mountains City Council [1] . The bush fire legislation takes different forms in the different States. But the general effect in all States is, I think, that as it was put by Gavan Duffy C.J. and Starke J. in McInnes v. Wardle [2] , it "brings into relief the dangers to be foreseen and provided against" [3] . Here the respondent foresaw the dangers. He took some measures to provide against them, and notified the fire control officers. But his Honour held that he negligently left the fires when he could have extinguished them. 1. (1961) 78 W.N. (N.S.W.) 864; 6 L.G.R.A. 263. 2. (1931) 45 C.L.R. 548. 3. (1931) 45 C.L.R., at p. 550. I would allow the appeal. Since I wrote what appears above, affidavits have been filed on behalf of the respondent to the effect that there is evidence, not called at the trial, tending to show that the respondent did in fact extinguish the fires in the logs. The parties have forwarded to us their written submissions in relation to the admissibility of this material. I need say no more of it than that I entirely agree with what has been said by Taylor and Owen JJ. and with the order they propose.