Rubis Bahamas Ltd v Lillian Antoinette Russell (The Bahamas)

At the junction of Robinson Road and Old Trail Road in Nassau in New Providence, The Bahamas, there is a petrol service station. The appellant, Rubis Bahamas Ltd (“Rubis”), has been its leasehold owner since 2012, when Rubis acquired all the assets of Texaco Bahamas Ltd (“Texaco”), including the service station, and let the service station to Fiorente Management and Investments Ltd (“Fiorente”) by a lease dated 9 November 2012. Lillian Russell, who has brought this action and is the respondent to the appeal, owns a property opposite the south-east corner of the service station on the other side of Old Trail Road.

In 1994, there was a leak of fuel from the underground storage tanks which are situated in the south-east corner of the service station, opposite the Russell property. The leak caused contamination of the groundwater and soil in an area which included that property. There is a private well on the Russell property which at that time supplied the water used by the family for drinking. As a result of the contamination, the well water became unfit to drink. According to evidence given by Ms Russell and her brother at the trial, fruit-bearing trees on the property were also affected. After the 1994 leak, plums, cherries and jujus produced by these trees decreased tremendously in quality and were potentially unsafe to eat so that the family could no longer enjoy the fruits from their yard.

On or about 25 November 2012—just after Rubis had acquired the service station from Texaco and leased it to Fiorente—another leak of fuel occurred. This time the source of the leak was a riser pipe to one of the fuel dispensers situated in the middle of the service station. The equipment was repaired on 21 January 2013, but by then around 24,000 gallons of unleaded gasoline had escaped. A central issue of fact in this case, which remains in dispute on this appeal, is whether fuel from this leak (“the 2012 leak”) migrated through the subsoil and caused further contamination of the Russell property. There is an issue of law as to whether, if it did, Rubis is liable for the damage.

Ms Russell commenced this claim against Rubis in March 2015. Her statement of claim alleged that both the 1994 leak and the 2012 leak caused contamination of the soil and water table on her property through migration of petroleum products from the service station. In relation to the 1994 leak, her primary case was: (1) that Texaco was negligent in failing to clean up the contamination of the Russell property caused by this leak; and (2) that, when Rubis acquired Texaco’s assets, Rubis also assumed Texaco’s existing obligations and liabilities, including any liability for loss caused by failing to clean up the contamination from the 1994 leak. In relation to the 2012 leak, Ms Russell claimed that this was caused by negligence of Rubis in failing adequately to inspect, maintain and repair the fuel equipment at the service station; alternatively, that Rubis was liable for damage caused to her property under the law of trespass, nuisance and/or the rule in Rylands v Fletcher.

In its defence Rubis denied that it had assumed any liability of Texaco relating to the 1994 leak and also averred that any claim relating to that leak was in any case barred by the statute of limitation. In relation to the 2012 leak, Rubis denied that this leak had resulted in any contamination of Ms Russell’s property and also that Rubis had any legal liability for that leak and its consequences.

The trial of the claim was heard over three days in November 2019 and July 2020. Judgment was reserved and was not given until over 21 months later. In the meantime, the trial judge, Thompson J, had retired. His judgment, dated 14 April 2022, was delivered on 13 May 2022 by the then Chief Justice.

The inordinate delay in producing the judgment was matched by its inordinate length: the judgment is 187 pages long. Disappointingly, despite its prolixity, the judgment contains hardly any findings of fact or reasons for the decision. Much of it comprises documents (including the pleadings and witness statements) which have simply been photocopied and incorporated in the judgment in their original format, together with long passages from law reports and other documents which have been reproduced verbatim.

What the judge actually decided is not always clear. But there is no doubt about the result. The judge “accede[d] to the claims” of Ms Russell in relation to both the 1994 leak and the 2012 leak and awarded damages in a sum of $692,825 (comprising $250,000 for “loss of amenity value”, $439,375 for “cost of remedial works” and $3,450 for “cost of testing and appraisal”).

Rubis appealed from this decision. For reasons given in a judgment delivered by the President, Sir Michael Barnett, with which Crane-Scott JA and Evans JA agreed, the Court of Appeal allowed the appeal in part. The Court of Appeal held that the judge was wrong to find that Rubis was liable for damage caused by the 1994 leak—both because any claim relating to that leak lay against Texaco and not Rubis (and was not affected by the contractual arrangements between Rubis and Texaco) and because any such claim was time-barred. There is no appeal from that part of the decision.

But in relation to the 2012 leak, the Court of Appeal upheld the judge’s finding of liability, based on the rule in Rylands v Fletcher. The judge’s award of damages—which did not distinguish between harm caused by the 1994 leak and harm caused by the 2012 leak—was set aside. The Court of Appeal substituted an award of $159,450, consisting of $25,000 for loss of amenity value, $131,000 for diminution in value of the property and special damages of $3,450 (the cost of testing and appraisal).

Rubis appeals from that decision. The Board decided, as a preliminary issue, that an appeal lay as of right: see Rubis Bahamas Ltd v Russell [2025] UKPC 13; [2025] 1 WLR 2162.

The appeal raises three issues:

The migration issue

At the trial Rubis did not seriously dispute that the soil and groundwater at the Russell property had been contaminated by fuel from the 1994 leak. But Rubis denied that any further contamination was caused by the 2012 leak.

The evidence on this issue came, in the first place, from Ms Russell herself. She testified that, after the 2012 leak, “there was an appreciable difference in the level of fumes in the home. The odour was stronger and much more consistent and persistent”.

A valuable source of evidence because it is entirely independent of the parties to this litigation is a report dated 20 February 2014 prepared for the Bahamas Environmental Science and Technology Commission by the US-based engineering and consulting company Black & Veatch. Black & Veatch was retained to investigate and assess the release of fuel caused by the 2012 leak. The main aim of the assessment was to ensure that the actions taken by Rubis to address the release were appropriate to protect public health and adequately remediate the environment.

The Black & Veatch report is a detailed document which describes their extensive investigations into the 2012 leak, the extent of the environmental impact and risk to health caused by the leak, the actions taken by Rubis and further actions planned to clean up the fuel spill, and recommendations for further remedial actions. The investigations undertaken included installing some 58 monitoring and gasoline recovery wells in a range of locations on and off the service station site. One of these monitoring wells from which water samples were collected and analysed was located (with her permission) on Ms Russell’s property.

The Black & Veatch report records that the area most affected by the 2012 leak was to the north of the service station. A commercial building owned by Cable Bahamas Ltd to the west of the site was also affected. The residential properties found to be at greatest risk from exposure to contaminants were two properties on the far side of Robinson Road to the north. Although the report includes an account of an interview with Ms Russell discussing the effect on her property of the 1994 leak, there is no suggestion in the report that any further contamination had been caused to her property by the 2012 leak. Maps annexed to the report depict the “plume” of groundwater contamination at dates in March, May and October 2013, showing in each case that it did not extend as far as the Russell property. At the same time the report refers to a groundwater contamination plume located in the south-east corner of the service station (which is the part closest to the Russell property) “that appears to be separate from the contamination caused by the fuel dispenser release” (ie the 2012 leak). Elsewhere, the report refers to “[t]he isolated groundwater contamination at the south-east corner of the Rubis property that is likely the result of a historic release”. The implication is that any contaminants found in groundwater to the south-east of the service station were probably a continuing consequence of the historic 1994 leak. Ms Russell relied on this evidence to support her case that Texaco had failed to remediate the 1994 leak and that this failure continued up to at least the date of the Black & Veatch report.

Rubis retained another US consultancy firm, Arcadis, to advise it in relation to the 2012 leak and the remediation work required. An employee of Arcadis, Mr Jorge Ramirez, gave expert evidence for Rubis at the trial. Mr Ramirez exhibited to his report the results of analyses of samples taken from monitoring wells in October 2013, March 2014 and February 2015. These included analyses of samples taken from the monitoring well on the Russell property. The samples were tested for chemicals found in motor fuel—principally, Benzene, Toluene, Ethylbenzene and Xylenes (abbreviated to “BTEX”). Each time, Benzene was either not detected in the sample taken from Ms Russell’s property or was detected in a concentration so low as to be below the reportable limit; Toluene was not detected; and Xylenes were either not detected or detected at levels below the reportable limit. Only Ethylbenzene was detected in concentrations which were above the reportable limit, although still negligible: 0.0047 mg/L (milligrams per litre) in October 2013; 0.0028 mg/L in March 2014; and 0.0013 mg/L in February 2015. The first and highest of these numbers is equivalent to less than one part per two hundred million. This is far below the level at which, according to Mr Ramirez, groundwater would be considered contaminated and requiring clean-up.

Based on the maps and data from the monitoring wells, Mr Ramirez drew the conclusion that Ms Russell’s property was “not, and has never been, within the impacted zone” and that petroleum products from the 2012 leak “never came close” to entering her property.

Ms Russell also adduced expert evidence at the trial, given by a consultant engineer, Mr John Bowleg. As explained in his report, his firm Adarie Engineering and Environmental Services undertook to collect a groundwater sample from the monitoring well on the Russell property, to arrange for the sample to be analysed by a certified laboratory, and to review and interpret the results. In accordance with these instructions, Mr Bowleg collected a sample in August 2014 and sent it to a laboratory in Florida for analysis. The results of this analysis were that Benzene was not detected at all and the other three BTEX chemicals were detected in the following concentrations: Toluene – 0.00656 mg/L; Ethylbenzene – 0.00476 mg/L; Xylenes – 0.01318 mg/L. Although higher than the concentrations reported by Arcadis, these levels are still all well below the levels said by Mr Bowleg himself to be “‘conservative’ clean-up numbers to protect groundwater”. (Those numbers are 0.04 mg/L for Benzene, 20 mg/L for Toluene, 15 mg/L for Ethylbenzene and 167 mg/L for Xylenes.)

Mr Bowleg also reported, for comparison, results of analyses of samples taken from two monitoring wells at one of the residential properties to the north of Robinson Road. These showed concentrations of BTEX which were several orders of magnitude greater than those found in the sample collected from the Russell property. The concentrations detected in these samples were: Benzene – 2.68 mg/L and 12.4 mg/L; Toluene – 24.8 mg/L and 50 mg/L; Ethylbenzene – 1.45 mg/L and 0.78 mg/L; and Xylenes – 5.03 mg/L and 19.46 mg/L.

The conclusions stated in Mr Bowleg’s report dated 29 August 2014 included the following:

“The suspected source of the contamination is the ‘Rubis/Texaco’ facility at the intersection of Robinson Road and Old Trail Road. The source of the fuel spill must be isolated. The contamination has migrated from the source site.

We continue to detect traces of BTEX towards the east of the suspected site. The typical flow of groundwater in the area is towards the north or northeast, with some minor variations.”

Mr Bowleg was not asked to express an opinion about whether the minute traces of BTEX detected in the sample taken from the Russell property were the result of the 2012 leak or were residues persisting from the contamination caused by the 1994 leak. But in his oral evidence he said that, in the absence of remediation, contaminants will remain in the groundwater and soil for “many, many years”.

Although the judge recorded in his judgment most of the evidence summarised above, he did not evaluate it. He accepted an argument made by counsel for Ms Russell that this was unnecessary for a procedural reason. The nub of this argument was that, because it was not put to Ms Russell’s expert, Mr Bowleg, in cross-examination that the 2012 leak did not impact her property, both Rubis and the court were bound to accept that the 2012 leak did impact Ms Russell’s property. This was said to follow from “the rule in Browne v Dunn”.

This argument was developed in a “supplemental skeleton argument” submitted on behalf of Ms Russell at the end of the trial. Employing a technique not to be emulated, the judge copied out verbatim almost the whole of this document over some 28 pages of his judgment. After expressing his gratitude to counsel for having “meticulously set out the above”, he reproduced, over some five pages, long passages from the speeches of the House of Lords in Browne v Dunn (1893) 6 R 67. Then, at paras 104–106 of the judgment, he stated his conclusion as follows:

“104.

Plaintiff’s counsel makes the point that by the application of the rule in Brown v Dunn (sic) it is now not open to the defendant to invite the court to accept a contrary position to that of the witnesses who were in the witness box but not cross-examined on certain evidence highlighted above. The further position is that it is now not open to the court to accept any more such contrary proposition or position in circumstances where the defendant chose not to cross-examine on certain issues and/or allowed the witness to respond to the issues.

105.

As it relates to the above, I would be hard pressed not to accede to the plaintiff's position as put forward above. In this regard I do agree with the argument of the plaintiff.

106.

Having acceded to the above position as put by the plaintiff it means that each and every failure to cross-examine as set out by the plaintiff above is accepted as evidence which goes to the plaintiff’s credit.”

Of the witnesses alluded to in this passage, only Mr Bowleg is relevant for present purposes. In their “supplemental skeleton argument” counsel for Ms Russell had relied on Mr Bowleg’s evidence which the Board has quoted at para 22 above, highlighting the sentence “The contamination has migrated from the source site”. They submitted:

“… significantly, it was never put to Mr Bowleg on cross-examination that the 2012 leak did not impact [Ms Russell’s] property.”

The submission was then made that:

“It is therefore now not open to [Rubis] to invite the court to make the finding that the 2012 leak did not impact [Ms Russell’s] property in circumstances where that position, which is contrary to his evidence, was never put to Mr Bowleg on cross-examination.” (Emphasis in original)

Against this background, although the final words of para 106 (referring to “evidence which goes to the plaintiff’s credit”) make little or no sense, the overall conclusion reached at paras 104–106 of the judgment is clear enough. The judge accepted the argument made by Ms Russell’s counsel that he was bound, as a matter of law, to find that her property had been contaminated by fuel from the 2012 leak because the contrary proposition had not been put to Mr Bowleg in cross-examination. Having reached that conclusion, the judge did not consider whether the evidence adduced at the trial supported the conclusion or was inconsistent with it. On the view of the law which he accepted, the omission of counsel for Rubis to cross-examine Mr Bowleg on the point was decisive and foreclosed the need to assess whether, in fact, fuel from the 2012 leak had migrated to Ms Russell’s property.

The argument advanced by counsel for Ms Russell, which the judge accepted, was misconceived. It stood “the rule in Browne v Dunn” on its head. It sought to turn a flexible rule based on considerations of procedural fairness into a rigid weapon which can be used to wrongfoot or ambush an opponent who makes a technical misstep.

The facts of Browne v Dunn were that in closing argument at the trial of a libel action counsel for the defendant invited the jury to disbelieve witnesses called by the plaintiff whose evidence had not been challenged by cross-examination. The jury returned a verdict for the defendant which could be justified only if the evidence of those witnesses was disbelieved. The verdict was set aside on appeal and the House of Lords unanimously affirmed that decision. Their reasoning was based on what Lord Herschell LC described as “fair play and fair dealing with witnesses”: (1893) 6 R 67, 71. He said that “it will not do to impeach the credibility of a witness upon a matter on which he has not had any opportunity of giving an explanation by reason of there having been no suggestion whatever in the course of the case that his story is not accepted.” Lord Halsbury, concurring, said, at pp 76–77, that:

“nothing would be more absolutely unjust than not to cross-examine witnesses upon evidence which they have given, so as to give them notice, and to give them an opportunity of explanation, and an opportunity very often to defend their own character, and, not having given them such an opportunity, to ask the jury afterwards to disbelieve what they have said, although not one question has been directed either to their credit or to the accuracy of the facts they have deposed to.”

Although Browne v Dunn was concerned with a challenge to the credibility of witnesses, the rule that it illustrates is wider. The status and scope of the rule were recently considered by the UK Supreme Court in Griffiths v TUI (UK) Ltd [2023] UKSC 48; [2025] AC 374. Lord Hodge (with whom the other Justices agreed) stated the rule, at para 70(i), as follows:

“The general rule in civil cases … is that a party is required to challenge by cross-examination the evidence of any witness of the opposing party on a material point which he or she wishes to submit to the court should not be accepted. That rule extends to both witnesses as to fact and expert witnesses.”

The Griffiths case itself involved evidence given by an expert witness in a report which had not been challenged by cross-examination. The Supreme Court held that, in the absence of such a challenge, it had not been fair for the other party to advance detailed criticisms of the expert’s report or for the judge to accept those criticisms (para 75).

As explained in Griffiths, at para 70, the purpose of the rule in Browne v Dunn is to make sure that the trial is fair. This involves fairness to the witness whose evidence is impugned (and whose character may be publicly slighted if the evidence is rejected) and fairness to the party who has adduced the evidence of the impugned witness. This has particular importance where specific criticism of the evidence is made to which the witness has not been given an opportunity to respond. If in such circumstances the evidence is rejected, the party relying on it (as well as the witness) is prejudiced because, had the criticism been raised in cross-examination, the witness might have answered it and explained or clarified the impugned evidence.

Lord Hodge emphasised, at para 70(vii), that the rule is not inflexible and should not be applied rigidly. He gave various examples of circumstances in which the rule may not apply. One is a situation where an expert’s report contains a bald assertion of opinion without any reasoning to support it: see para 63. The reason why a bald assertion of opinion, described in one case as “a bare ipse dixit”, need not be challenged is that, by itself, it is worthless. What carries weight is not the mere fact that the opinion is held but the strength of the reasons for it: see paras 37–38.

Here, although Mr Bowleg’s expert report contained a statement that the “contamination has migrated from the source site” (ie the “Rubis/Texaco fuel facility” identified as “[t]he suspected source of the contamination”), the report did not express any view about whether fuel from the 2012 leak had reached the Russell property or whether the traces of BTEX detected in the groundwater sample which Mr Bowleg had sent for analysis were attributable to that leak or to the 1994 leak. Mr Bowleg could have been, but was not, asked by Ms Russell’s legal representatives to express an opinion on this point. The omission was no doubt deliberate. As mentioned earlier, it was part of Ms Russell’s case that the “historic release” in 1994 referred to in the Black & Veatch report had resulted in the continuing presence of petroleum products on her property. It was also her case that petroleum products from the 2012 leak had migrated to her property. It is understandable that Ms Russell’s legal representatives chose to keep both possibilities open and not to ask Mr Bowleg to say which he thought more likely. The references in his report to “Rubis/Texaco” seem designed to maintain this ambiguity.

It might have been preferable for counsel for Rubis to put to Mr Bowleg in cross-examination the positive suggestion that the traces of BTEX detected in the groundwater sample were more likely to be residues of the 1994 leak than a result of the 2012 leak. But the burden of proof was on Ms Russell, and Mr Bowleg had not asserted the contrary. Indeed, his evidence that, in the absence of remediation, contaminants will remain in the groundwater and soil for “many, many years” (see para 23 above) was consistent with that explanation.

In these circumstances, even if the rule in Browne v Dunn operated in the absolute and mechanical way suggested by counsel for Ms Russell, the omission to cross-examine Mr Bowleg on this point could not assist them and their reliance on the rule was completely misplaced. The rule would only have been relevant if Mr Bowleg had given evidence that the contaminants detected on Ms Russell’s property resulted from the 2012 leak. But Mr Bowleg had not given such evidence. Thus, the contention that accepting Mr Bowleg’s unchallenged evidence made it unnecessary to decide whether fuel from the 2012 leak had contaminated Ms Russell’s property was founded on a false premise.

The argument fares no better even if the statement in Mr Bowleg’s report that “[t]he contamination has migrated from the source site” is taken to have impliedly suggested that the source of the contamination of Ms Russell’s property was the 2012 leak. At its highest any such assertion was a “bare ipse dixit”, unsupported by any reasoning and therefore worthless as evidence. There was no requirement to challenge the validity or adequacy of any reasons for supposing that the traces of BTEX found on Ms Russell’s property derived from the 2012 leak rather than the 1994 leak as Mr Bowleg had not given any such reasons. On any view, therefore, the fact that Mr Bowleg was not cross-examined on the point involved no unfairness either to him or to Ms Russell.

It should be added that, even if there had been any merit in the argument based on the rule in Browne v Dunn, this would not have absolved the judge from his duty to assess the evidence as a whole. The significance of failure to cross-examine on a matter can only be gauged by considering how the evidence that was not challenged relates to the issues in dispute and the other evidence bearing on those issues. An omission to put a relevant matter to a witness does not place the trial judge “into a straitjacket, dictating what evidence must be accepted and what must be rejected”: MBR Acres Ltd v McGivern [2022] EWHC 2072 (QB), para 90 (Nicklin J). That is because, as a rule of fairness, the rule is not inflexible and “a more nuanced judgment is called for”: Griffiths, para 69.

In this case the trial judge failed altogether to make the required assessment, so that no finding of fact based on evidence was made on the question whether fuel from the 2012 leak migrated to the Russell property.

This error was not corrected by the Court of Appeal. Before the Board counsel for Ms Russell asserted that the Court of Appeal was not asked to decide whether the judge’s reliance on the rule in Browne v Dunn was justified, but only whether the judge had accepted Ms Russell’s case that fuel from the 2012 leak had migrated to her property. It appears that this was indeed how the Court of Appeal approached the matter. But it is not a tenable view of the case put forward by Rubis on its appeal to the Court of Appeal.

Ground 1(c) of its grounds of appeal was that:

“the Learned Judge failed to find as an issue of fact whether the release of petroleum products from the service station in or about 2012/2013 … migrated onto/escaped to the [Russell] property at all”.

Ground 8 was that:

“the Learned Justice … misdirected himself at paragraph 106 of the Judgment in concluding that the Judge was in any way excused from his duty to independently weigh and critically analyse all conflicting evidence before him …”

On its own, ground 1(c) was ambiguous. It could have meant that the judge failed to make any finding on the question whether petroleum products from the 2012 leak migrated to Ms Russell’s property; or it could have meant that the judge failed to decide this question as an issue of fact, and instead wrongly treated the matter as one of law which could be decided by applying the rule in Browne v Dunn. But any doubt about whether Rubis was making the latter complaint was dispelled by ground 8. This made it quite clear that Rubis was challenging the judge’s conclusion in the passage ending at para 106 of the judgment (quoted at para 25 above) that, because of failure by counsel for Rubis to cross-examine on the issue, the judge did not need to weigh and evaluate the evidence bearing on whether petroleum products from the 2012 leak had migrated to the Russell property and was excused from his duty to do so.

This argument was elaborated by counsel for Rubis in their skeleton argument for the appeal, which gave detailed reasons for submitting that the rule in Browne v Dunn did not apply and that the judge was wrong to invoke it.

Counsel for Ms Russell clearly understood the case they had to meet. In their skeleton argument and in the oral submissions made by Ms Rolle KC to the Court of Appeal, they repeated, in answer to ground 8 of the grounds of appeal, the argument they had made at the trial based on the rule in Browne v Dunn.

Regrettably, the Court of Appeal overlooked ground 8 when dealing with the migration issue. Having quoted ground 1(c), but not ground 8, Barnett P said, at para 49 of the judgment, that the complaint could be dealt with quickly. He dealt with it, at para 50, simply by saying that the judge clearly found as a fact that there was “a 2012/2013 release” and that it was in any event an agreed fact that there was a leak of fuel from the service station in 2012.

This missed the point in two ways. First, the Court of Appeal appears to have understood the question raised to be whether there was a release of petroleum products from the service station in 2012/2013. But that was not the issue. As the Court of Appeal itself noted, the fact that there was a release of fuel in 2012/2013 was not in dispute. The issue was whether fuel from this release had migrated to the Russell property. Second, the complaint on that score was not that the judge had failed to reach a conclusion; it was that he had failed to evaluate the evidence and make a finding of fact on the question, having wrongly held that he was excused from doing so by the rule in Browne v Dunn.

The result of this oversight was that the Court of Appeal did not consider, as it was required to do, whether counsel for Ms Russell and the judge were justified in relying on the rule in Browne v Dunn and, if not, how the question whether fuel from the 2012 leak had migrated to Ms Russell’s property should be decided.

This leaves the Board in the unusual position that neither of the courts below made a factual finding or undertook any evaluation of the evidence on a key factual issue. Counsel for Rubis have invited the Board to carry out this exercise itself and to find as a fact that the 2012 leak did not result in the migration of petroleum products to Ms Russell’s property; alternatively, to remit the case to the Court of Appeal with directions for it to consider the migration issue afresh.

Desirable as it is to bring this long-running litigation to an end as soon as possible, the Board considers that the appropriate tribunal to resolve what is a pure question of fact, if it is necessary to do so, is the court of first instance. Whether the question needs to be resolved depends on how the issues of law raised on this appeal are answered. Those answers will determine whether, if there was migration of petroleum products to Ms Russell’s property as a result of the 2012 leak, Rubis is liable for any damage caused by the contamination. The Board must therefore decide these legal issues.

Liability for the 2012 leak

The judge stated his conclusion on the issue of liability, at para 139 of his judgment, as follows:

“… I accede to the claims of the plaintiff as it relates to:

1.

The 1994 release;

2.

The 2012/2013 release;

3.

The claim for negligence;

4.

The nuisance and trespass claims;

5.

The claim under the rule in Rylands v Fletcher and

6.

The rule in Brown v Dunn [sic].”

On its face this appears to be an acceptance of Ms Russell’s entire case on liability. A bald statement of this kind, however, cannot stand as a judicial decision in the absence of any reasons or relevant findings. Although Ms Russell’s statement of claim alleged that Rubis was liable for damage caused by the 2012 leak on the four legal bases mentioned by the judge, the judge made no relevant findings and gave no reasons for acceding to the claims for negligence, nuisance and trespass.

The pleaded case of negligence was that Rubis failed adequately to inspect, maintain and repair the fuel equipment. Yet the judge did not address that question and made no finding to that effect. Trespass requires an intentional act which directly intrudes on land in the possession of another person. On the evidence no such act was committed here, and the judge made no finding that any such act was committed. The judge also made no finding of liability for nuisance separate from the rule in Rylands v Fletcher. He did, however, quote at length from the decisions of the House of Lords in Rylands v Fletcher (1868) LR 3 HL 330 and Transco plc v Stockport Metropolitan Borough Council [2003] UKHL 61; [2004] 2 AC 1, where the House of Lords affirmed the continued existence of the Rylands v Fletcher rule and considered its restriction to “non-natural use” of land. At para 113 of his judgment, the judge said that, in his opinion, “it cannot be refuted that petroleum tanks placed under the ground is not a natural use of land”; and at para 126 he said that Ms Russell “has satisfied the legal requirement inclusive of non-natural use of the defendant’s land and should succeed under this head”. Exiguous as it is, this amounts to a finding that Rubis was liable under the rule in Rylands v Fletcher for any damage caused by the 2012 leak; and it was treated as such by the Court of Appeal, which upheld the judge’s finding of liability on this basis.

In 1860, John Rylands arranged for the construction of a reservoir to supply water to his mill. Beneath the reservoir were some old mine shafts blocked with soil. When the reservoir was filled, one of these shafts burst downwards and water escaped, flowed through old mine workings under adjoining land and flooded Thomas Fletcher’s colliery. Although Rylands (and the manager of his mill who was also sued) were found not to have been negligent, the Court of Exchequer Chamber held on appeal that they were liable to compensate Fletcher in damages: (1866) LR 1 Ex 265. Blackburn J, who gave the judgment of the court, said, at p 279:

“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.”

Blackburn J went on, at p 280, to explain the justice of the rule:

“[I]t 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.”

The judgment of the Court of Exchequer Chamber was affirmed on appeal by the House of Lords: (1868) LR 3 HL 330. Lord Cairns LC, however, added a gloss, at pp 338–339, which has since given rise to much discussion and controversy, that the rule applied only to “non-natural use” of land.

Rubis contends that the courts below were wrong in law to hold that Rubis is liable under the rule in Rylands v Fletcher for any damage caused by petroleum products which escaped from the service station as a result of the 2012 leak. Two arguments are made. First, Rubis submits that it is not liable for damage caused by the escape of fuel because it was not the owner and controller of the fuel: that person was the lessee of the service station, Fiorente. Second, Rubis argues that the storage of petroleum products in underground storage tanks is an ordinary use of land in present-day Nassau in The Bahamas and, for that reason, does not satisfy the requirement of “non-natural” use which is a condition of liability under the rule in Rylands v Fletcher.

Rubis says that, as it has numerous service stations throughout The Bahamas and the wider Caribbean, the decision in this case will have a significant impact on its business and dealings with the wider community in The Bahamas and the Caribbean at large.

As with the migration issue, counsel for Ms Russell assert that the grounds on which Rubis disputes liability under the rule in Rylands v Fletcher were not raised before the Court of Appeal, with the result that Rubis should not be allowed to advance these “new grounds” before the Board. Such was the reliance placed in Ms Russell’s written case on this objection that counsel did not even condescend to deal with the merits of the two issues referred to at para 54 above. That was an error of judgment because the objection is misconceived.

On the appeal of Rubis to the Court of Appeal, ground 1(d) was that:

“the Learned Justice failed to find as a matter of law and an issue of fact … whether the operator of the service station (Fiorente Management Limited) was liable in whole or in part in relation to the [2012 leak].”

In their skeleton argument in the Court of Appeal counsel for Rubis made it clear that its case was not—as counsel for Ms Russell have sought to suggest based on a literal reading of this ground of appeal—that the judge failed to make any finding one way or the other as to whether Fiorente was the person liable in relation to the 2012 leak. It was that the judge erred in failing to find that the person liable was Fiorente rather than Rubis.

Ground 7 of the grounds of appeal was that:

“The Learned Justice erred in law and in fact … by finding that the operation of the service station for the purpose of storage, sale and supply of petroleum products was a non-natural use of land despite the proper use of the service station for the clear, commonplace and general benefit of the Marathon and wider community”.

The question whether the storage of petroleum products was a non-natural use of land was therefore squarely raised.

Although counsel for Ms Russell attempted to argue before the Court of Appeal, as they have before the Board, that it was not open to Rubis to advance its case on these issues, the Court of Appeal rightly did not give that submission the time of day and decided the issues on their merits. On the first, the Court of Appeal found that, under the terms of the lease between Rubis and Fiorente, the equipment used to store and dispense petroleum products was not leased to Fiorente and remained the responsibility of Rubis. As Rubis was still the owner and controller of the equipment from which the leak occurred, Rubis was liable for damage caused by the leak: see paras 52–53 of the judgment. On the second issue, after referring to several authorities, Barnett P concluded that “the trial judge could not be faulted for finding that the storage of fuel on property is not a natural use of the land” (para 58).

In these circumstances the contention that the relevant grounds of appeal were not advanced in the Court of Appeal is hopeless. They were advanced, contested and decided on their merits. Yet it was only when requested to do so by the Board that counsel for Ms Russell on this appeal addressed the merits of these issues in a supplemental written case. They should have been addressed from the outset.

In the century and a half since it was first articulated, the rule in Rylands v Fletcher has been the subject of many authorities, both in England and Wales and in other common law jurisdictions. The rule has been discarded in Australia (see Burnie Port Authority v General Jones Pty Ltd (1994) 179 CLR 520). But it remains part of New Zealand’s common law (see Nottingham Forest Trustee Ltd v Unison Networks Ltd [2021] NZCA 227; [2021] 3 NZLR 823) and is “alive and well in Canada as a 21st century tort” (see Linden, Feldthusen, Hall, Knutsen, Young (eds), Canadian Tort Law, 13th ed (2026), para 10.03(6)). In the United States it has developed into a general rule of strict liability for physical harm resulting from an abnormally dangerous activity (see Restatement of the Law Third, Torts: Liability for Physical and Emotional Harm (2010), § 20). In England and Wales, the continued existence of the rule has been affirmed and its scope considered by the House of Lords in two leading modern authorities: Cambridge Water Co v Eastern Counties Leather plc [1994] 2 AC 264 and Transco.

In Cambridge Water the defendants were leather manufacturers who used a chemical solvent in their tanning process. Small quantities of solvent spilled regularly onto the floor and seeped into the soil below the premises. From there the chemical gradually migrated through percolating underground water to a borehole about 1.3 miles away which the claimant water company used as a source for the local municipal water supply. As a result of the contamination, the water was rendered unfit for human consumption. A claim by the water company for damages based on the rule in Rylands v Fletcher failed at trial on the grounds that: (1) it was not reasonably foreseeable that, if solvent escaped from the premises, damage of the relevant type would result; and (2) the use of the solvent in the defendant’s business constituted a natural use of land. On appeal the House of Lords affirmed the result on the first of these grounds but declined to hold that the defendants should be exempt from liability on the basis of the exception for natural use. The reasons for the decision were given by Lord Goff of Chieveley, with whom the other law lords agreed.

In Transco the facts were that water escaped from an underground pipe which connected a block of flats to the water main. The connecting pipe and the land across which it ran were owned by the defendant council. The leak remained undetected for a long period during which a large quantity of water escaped and percolated some distance, causing an embankment which supported the claimant Transco’s gas main to collapse. Transco sued the council for the costs of repairs. The judge found that, while the damage had occurred without negligence on the part of the council, the council was liable under the rule in Rylands v Fletcher. The Court of Appeal reversed that decision, and Transco’s appeal to the House of Lords was dismissed.

Identifying the legal propositions for which Transco is authority is not straightforward as each of the law lords gave a separate reasoned speech. Lord Bingham of Cornhill and Lord Hoffmann each gave their own reasons for the decision, which are not in all respects consistent, as did Lord Hobhouse of Woodborough. Lord Walker of Gestingthorpe agreed with the reasons given by Lord Bingham and those given by Lord Hoffmann but also gave further reasons of his own. Lord Scott of Foscote agreed with the reasons given by Lord Bingham, Lord Hoffmann and Lord Walker, and also gave some further reasons of his own. It will be necessary to consider some of the differences between the various opinions expressed when addressing the question of what constitutes “non-natural use” of land.

But first it is important to identify the rationale for the rule in Rylands v Fletcher. One way to approach this, which the Board will adopt, is by considering the relationship in general terms between the rule and the torts of private nuisance and negligence.

In both Cambridge Water and Transco the House of Lords in obiter dicta characterised the rule in Rylands v Fletcher as closely related to the law of nuisance. In Cambridge Water, at pp 297–299, Lord Goff accepted the view put forward in a scholarly article by FH Newark, “The Boundaries of Nuisance” (1949) 65 LQR 480 that the judges who decided Rylands v Fletcher regarded it as a simple case of nuisance (despite the fact that, as Professor Newark acknowledged, they never once used the word “nuisance”). Professor Newark based this interpretation mainly on the fact that Blackburn J presented his statement of “the true rule of law” as established by existing authority rather than as new and revolutionary.

It is difficult to attach much weight to this point given that a classic method of developing the common law is to derive a new general rule from existing cases. Lord Atkin’s speech in Donoghue v Stevenson [1932] AC 562 is another famous example of this method. It may also be noted that the precedents on which Blackburn J drew were primarily cases of trespass, such as cases involving escaping cattle, and that he relied on considerations of justice as well as authority (see para 52 above). As Professor AWB Simpson has commented, Professor Newark’s interpretation of Rylands v Fletcher “signally fails to make sense of the case’s status as a leading case”: AWB Simpson, Leading Cases in the Common Law (1995), p 199. But whatever its historical accuracy, the attraction of Professor Newark’s view for Lord Goff was that:

“It would … lead to a more coherent body of common law principles if the rule [in Rylands v Fletcher] were to be regarded essentially as an extension of the law of nuisance to cases of isolated escapes from land, even though the rule as established is not limited to escapes which are in fact isolated.” See Cambridge Water, at p 306.

This view was accepted by members of the House of Lords in Transco, with Lord Bingham even going so far as to describe the rule in Rylands v Fletcheras a “sub-species of nuisance” (para 9). At the same time the House of Lords declined to abolish the rule as a distinct legal doctrine, both because they considered it too well established and that this would be too radical a step to take, and because (with the exception perhaps of Lord Hoffmann) they regarded its retention as in the interests of justice. Lord Bingham thought that there is “a category of case, however small it may be, in which it seems just to impose liability even in the absence of fault” (para 6). Lord Hobhouse was the most enthusiastic supporter of the rule, making a strong argument that, when properly understood, it comprises “a useful and soundly based component of the law of tort” (para 52) and that “[t]he rationale for it was and remains valid” (para 66).

Certainly, liability under the rule in Rylands v Fletcher and in the tort of private nuisance have much in common. Both are centred on land and frequently (although in each case not necessarily) involve disputes between neighbouring landowners. As under Rylands v Fletcher, liability for a nuisance created by the defendant is strict in the sense that it does not depend upon whether the defendant exercised reasonable care and skill: see eg Cambridge Water, at pp 299–300. And under both doctrines, as the House of Lords held in Cambridge Water, damages are limited to harm of a type which is reasonably foreseeable.

Yet the analogy should not be pressed too far. As Lord Wright observed, giving the judgment of the Board in Northwestern Utilities Ltd v London Guarantee and Accident Co Ltd [1936] AC 108, 119, liability under the rule in Rylands v Fletcher “is in many ways analogous to a liability for nuisance, though nuisance is not only different in its historical origin but in its legal character and many of its incidents and applications”. There are, as Lord Wright later said in Sedleigh-Denfield v O’Callaghan [1940] AC 880, 903, “well marked differences between the two juristic concepts”.

A significant difference lies in the fact that private nuisance is a “tort against land” which protects the right of a person with an interest in land to use and enjoy their land free from undue interference. Since the defendant is typically a neighbouring landowner who has a corresponding right, determining what interferences are unlawful characteristically involves striking a balance between conflicting rights—between “the right of the occupier to do what he likes with his own, and the right of his neighbour not to be interfered with”: Sedleigh-Denfield, 903 (Lord Wright). How the balance is struck by the law of nuisance is context-specific and takes account of matters such as the nature of the conflicting activities and the character of the locality. Typically, where the defendant’s activity is found to constitute a nuisance, the court will grant an injunction to prohibit the defendant from carrying it on.

By contrast, the rule in Rylands v Fletcher is not concerned with activities or states of affairs which cause an ongoing interference with the use and enjoyment of land. The harm from which the claimant is protected is physical damage caused by a tangible thing which escapes from land. In establishing the conditions for liability, the law strikes a different balance between conflicting interests. The rule does not restrict the defendant’s freedom to use land by prohibiting a particular use or activity. It recognises the defendant’s freedom to use the land as it chooses, even though that choice involves bringing something dangerous onto the land which carries a high risk of causing physical harm to others if it escapes. But the rule protects those exposed to this risk by requiring the person who creates it to pay for the harmful consequences if the risk materialises. As Roderick Bagshaw noted when commenting on Transco in “Rylands Confined” (2004) 120 LQR 388, 388–389:

“One of the functions of private nuisance is to decide what activities people should be permitted to pursue in a particular locality, given the effects of those activities on neighbours. But it is no part of the rule in Rylands v Fletcher to forbid particular activities. Rather, as Lord Hoffmann noted (at para 29), it is a rule which requires those who pursue particular activities to internalise the costs of escapes. It is a rule about who pays when things go wrong rather than about whether the defendant’s activity is wrongful.”

One consequence of recognising these differences between the law of private nuisance and the rule in Rylands v Fletcheris to cast doubt on the correctness of obiter dicta in Transco (and earlier in Read v J Lyons & Co Ltd [1947] AC 156, 170–171, 180–181) suggesting that compensation under Rylands v Fletcher cannot be recovered for personal injury, despite considerable authority that it can: see eg Miles v Forest Rock Granite Co(Leicestershire) (Ltd) (1918) 34 TLR 500; Shiffman v Order of the Hospital of St John of Jerusalem [1936] 1 All ER 557; Hale v Jennings Brothers [1938] 1 All ER 579; Perry v Kendricks Transport Ltd [1956] 1 WLR 85, 92; and see John Murphy, “The Merits of Rylands v Fletcher” (2004) 24 OJLS 643, 652–654; Donal Nolan, “The distinctiveness of Rylands v Fletcher” (2005) 121 LQR 421, 432–434. In Transco both Lord Bingham (at para 9) and Lord Hoffmann (at para 35) reasoned that, because the rule in Rylands v Fletcheris a “sub-species” or “special form” of nuisance, and because nuisance is a tort concerned only with protecting interests in land, it follows that a claim under Rylands v Fletchercannot include a claim for death or personal injury, since such a claim does not relate to any right to land. It might be thought that the logic of this argument operates the other way round. The rationale for holding a person who chooses to keep on land something dangerous if it escapes responsible for harmful consequences of its escape applies just as much where the consequence is death or personal injury as it does where the harm consists of property damage. Indeed, the former type of injury is generally considered more deserving of compensation. The implication is that, unlike private nuisance, the rule in Rylands v Fletcheris not concerned solely with protecting rights to land.

That is reinforced by one of the examples given by Lord Bingham in Transco, at para 6, of a category of case for which discarding the rule in Rylands v Fletcher would not serve the interests of justice. Lord Bingham referred in that context to “the tragedy at Aberfan”. That was a notorious disaster which occurred in 1966 when a tip of colliery spoil on the mountainside above the village of Aberfan in Wales collapsed, causing a landslide which buried a school and a row of houses. In total, 144 people were killed, 116 of them children aged mostly between seven and ten. As Nicholas McBride and Roderick Bagshaw point out in their book on Tort Law, 7th ed (2024), pp 340–341, the implication of treating liability under Rylands v Fletcher as limited to interference with land rights is that those responsible for the accumulation of colliery spoil would be strictly liable for damage caused to the school buildings by the landslide, but only liable to pay compensation to the families of the deceased and to injured survivors if negligence could be proved. Justice is not achieved by an interpretation of the common law which places more value on real property than on human life.

As in Transco, the question whether damages can be recovered for harm to human beings as well as land does not arise for decision on this appeal. But the Black & Veatch report explains the serious risks to human health that can arise from exposure to groundwater or soil vapours contaminated by petroleum products; and, according to allegations made by Cable Bahamas Ltd in another action brought against Rubis arising out of the 2012 leak, some 43 of its employees and one member of the pubic received medical treatment at clinics and hospitals as a result of inhaling hydrocarbon vapours. The Board would wish to reserve its opinion on whether compensation for such injuries could in principle be recovered under Rylands v Fletcher.

The tort of negligence is of far wider scope than the rule in Rylands v Fletcher and encompasses the territory in which the rule operates. Thus, it seems safe to assume that in any situation where liability exists under Rylands v Fletcher the defendant will also be liable in the tort of negligence if it can be shown that the escape of the dangerous thing was a result of carelessness of the defendant (or someone for whose acts or omissions the defendant was responsible). But the point of Rylands v Fletcher is that, as already mentioned, the exercise of care is irrelevant to liability. The defences available go not to fault but to the question of causation. The defendant is not liable if the cause of the injury is shown to have been an act of the claimant or of a third party or an act of God.

There is no inconsistency between liability under the rule and liability for negligence because they are founded on different principles. As Lord Hobhouse explained in Transco, liability under Rylands v Fletcher is not based on a concept of fault but on a concept of risk. Someone who brings onto and keeps on land something abnormally dangerous exposes others in the vicinity to an exceptional risk of harm if an escape occurs which is not in any way of their making. It is both just and economically efficient to hold such a person liable to compensate those others for damage inflicted in that event, irrespective of whether negligence can be established. In the words of Lord Hobhouse, at para 57:

“He who creates the relevant risk and has, to the exclusion of the other, the control of how he uses his land, should bear the risk. It would be unjust to deny the other a risk based remedy and introduce a requirement of proving fault.”

The difference between the concept of risk on which the rule in Rylands v Fletcher is based and the concept of liability for negligence has long been recognised. It was lucidly explained by CK Allen in Legal Duties and Other Essays in Jurisprudence (1931), p 194:

“There is of course a duty of careful management of any material thing, whether it belongs to the family of ‘Rylands v Fletcher objects’ or not; but it is difficult to see that there is any ‘duty’ to prevent a dangerous thing escaping through causes which have nothing to do with the maintainer’s fault … The true situation seems to be that he who maintains for his own advantage a peculiarly dangerous thing in proximity to others, necessarily imposes upon those others a risk of injury … greater than is to be reasonably expected in the ordinary circumstances of social life; and it is therefore just and expedient that he himself should bear the risk of making good any damage to others which results from the maintenance of the object.”

The principle is closely aligned with the “polluter pays” principle for allocating responsibility for environmental damage. Indeed, most cases where the rule in Rylands v Fletcher applies are cases of environmental damage. Admittedly, the cases in which the rule is needed have been reduced over time by greater statutory regulation. But such regulation is not exhaustive and, although the Board was not referred to any relevant legislation in The Bahamas, it is notable that where statutory remedies have been created in the United Kingdom they have typically adopted a similar principle of strict liability. Examples are statutory provisions dealing with escapes of gas (section 14 of the Gas Act 1965), radioactive material (section 7 of the Nuclear Installations Act 1965), water (section 209 of the Water Industry Act 1991) and waste (section 73(6) of the Environmental Protection Act 1990). The fact that the legislature has taken the view that strict liability is apt in these cases confirms that the principle underpinning the rule in Rylands v Fletcher remains relevant and consistent with society’s expectations. While it may now be desirable, as Lord Goff suggested in Cambridge Water, at p 305, to leave any further expansion of strict liability for operations of high risk to the legislature, there is no reason for courts to remove or restrict such residual protection as the common law already provides through the rule in Rylands v Fletcher.

As Lord Walker said of the rule in Transco, at para 99:

“Its scope for operation has no doubt been restricted … by the growth of statutory regulation of hazardous activities, on the one hand, and the continuing development of the law of negligence, on the other hand. But it would be premature to conclude that the principle is for practical purposes obsolete.”

Lord Walker also made the important point that the imposition of strict liability is not shown to be unnecessary or undesirable if a claim based on negligence would lead to the same outcome. To make that assumption is “to overlook the practical implications, in a case of this sort, of bringing a claim in negligence, perhaps against a powerful corporate opponent” (para 110). Proving a failure to exercise reasonable care typically requires expert evidence and raises questions about compliance with specialised industry standards and other matters in relation to which there is likely to be a significant imbalance in terms of knowledge and resources between the potential defendant and an individual citizen who has suffered damage caused by the escape of a dangerous thing. The present case illustrates this. As noted earlier, Ms Russell did not discharge the burden of proving that Rubis was negligent. A requirement to do so would have substantially increased the difficulty and cost of pursuing a claim and might well have put the possibility of a remedy beyond her reach.

The significant differences between the torts of private nuisance and negligence and the rule in Rylands v Fletcher are relevant in deciding the issues about the scope of the rule raised on this appeal, to which the Board now turns.

Rubis argues that it cannot be liable under Rylands v Fletcher because it was not the occupier of the service station and did not own or control the fuel, which was the dangerous thing. The fuel was owned by, and in the possession of, Fiorente, which had purchased it from Rubis. Fiorente was also the occupier of the service station, which was leased to Fiorente (including the storage tanks and other equipment). Rubis argues that, in these circumstances, any liability under the rule in Rylands v Fletcher lies with Fiorente, and not Rubis.

Although the cases sometimes refer to the person liable where the rule applies as the “owner” or “occupier” of the land, and such a description is often correct on the facts, both on principle and authority neither ownership nor occupation is necessary or sufficient to give rise to liability under the rule. In accordance with the rationale discussed above, the person responsible for damage caused by the escape of the dangerous thing is the person who created the risk of such damage by bringing the dangerous thing onto and keeping it on the land from which the escape occurred. That person need not own the land. Nor is there any reason why liability should turn on whether that person is classified as an “occupier” for the purposes of the law governing the liability of “occupiers” to persons present on land.

In Rainham Chemical Works Ltd v Belvedere Fish Guano Co Ltd [1921] 2 AC 465 a company carried on the manufacture of explosives on land which two of its directors had leased from a third party. An explosion occurred which caused damage to the plaintiff’s neighbouring property. Both the company and the directors were held liable under the rule in Rylands v Fletcher. In explaining why the company’s appeal to the House of Lords should be dismissed, Lord Sumner said, at p 479, that “[t]hey cannot escape any liability which otherwise attaches to them on storing [explosives] there merely because they have no tenancy or independent occupation of the land but use it thus by permission of the tenants or occupiers”. He cited several cases to illustrate that proposition. Lord Sumner might have included (though in fact he did not) Rylands v Fletcher itself. The reservoir from which water escaped was constructed on land owned by a third party, Lord Wilton, “in pursuance of an arrangement with Lord Wilton”: see Rylands v Fletcher (1866) LR 1 Ex 265, 267. It is unclear whether this “arrangement” was a lease or a licence. What is clear from all the judgments is that the nature of the arrangement was not seen as material. The connection with the land which gave rise to liability was simply that the defendants had caused something dangerous to be brought onto and kept on the land. That was sufficient to treat the land as “theirs” for this purpose. This is explicit in the speeches in the House of Lords. For example, Lord Cranworth said, at p 342:

“The defendants, in order to effect an object of their own, brought on to their land, or on to land which for this purpose may be treated as being theirs, a large accumulated mass of water, and stored it up in a reservoir. The consequence of this was damage to the plaintiff …” (Emphasis added)

In Northwestern Utilities Ltd v London Guarantee and Accident Co Ltd [1936] AC 108, 118, Lord Wright, giving the judgment of the Board on an appeal from Canada, observed:

“The rule is not limited to cases where the defendant has been carrying or accumulating the dangerous thing on his own land: it applies equally in a case like the present where the appellants were carrying the gas in mains laid in the property of the City (that is in the sub-soil) in exercise of a franchise to do so”.

Other cases in which the party liable under the rule had no interest in the land, other than permission to use it, include decisions of the Court of Appeal of England and Wales in West v Bristol Tramways Co [1908] 2 KB 14 and Charing Cross Electricity Supply Co v Hydraulic Power Co [1914] 3 KB 772. See also Rigby v Chief Constable of Northamptonshire [1985] 1 WLR 1242, 1255, where Taylor J observed:

“I can see no difference in principle between allowing a man-eating tiger to escape from your land on to that of another and allowing it to escape from the back of your wagon parked on the highway.”

For this reason, the Board has not found helpful submissions made, and cases cited, by the parties on the question whether Rubis was an “occupier” of the service station for the purposes of occupier’s liability. What rights Rubis possessed as owner and lessor of the service station are relevant only in so far as they bear on the question whether Rubis was responsible for bringing onto the site and keeping there the fuel which escaped.

The attribution of such responsibility is not always straightforward. On principle, it does not rest with a party who acts at the direction of another person, such as the independent contractors employed to construct the reservoir in Rylands v Fletcher. When Blackburn J, in formulating the “true rule of law”, used the phrase “for his own purposes” (see para 52 above), those words should be understood to mean that the defendant must be acting on his own account and not that the defendant must be seeking to obtain a personal benefit from keeping the dangerous thing. A public authority fulfilling statutory obligations and acting for the benefit of the public can be liable under the rule. On the other hand, the Board agrees with counsel for Rubis that it cannot be sufficient to give rise to liability that a landowner not in occupation of the land has authorised the accumulation—if by that is meant merely that the landowner has given permission for the land to be used in that way. The Board does not suppose that Fletcher could have sued Lord Wilton on the basis that Lord Wilton allowed a reservoir to be constructed on his land. A stronger causal connection is required between the actions of the defendant and the creation of the risk. The directors held liable in Rainham Chemical Works, for example, had not merely permitted the land of which they were tenants to be used by the company for the manufacture of explosives. They had entered into a personal contract with the Minister of Munitions to manufacture explosives at the site and sell them to the Minister, and the contract included obligations to take delivery at the site and use there in the process of manufacture the dangerous chemical which exploded.

Whether Rubis should be regarded as responsible for the accumulation on the service station site of the fuel which escaped must be answered, then, not by asking whether Rubis fell within the category of “occupier” (or any other legal category), but by considering the particular arrangements by which the fuel was brought onto and kept on the land from which it escaped. Those are to be found in the lease between Rubis and Fiorente.

The lease contained the following material terms:

Counsel for Rubis have emphasised that, under these arrangements, the fuel stored and dispensed at the service station, which was the dangerous thing, was owned by Fiorente. They also point out that the Court of Appeal erred in asserting, at para 52, that the storage tanks were not leased to Fiorente. In fact, all the equipment, including the tanks and riser pipes, was leased to Fiorente, which was also responsible for routine maintenance, as well as inventory recording and making weekly inspections for fuel leakage. Rubis contends that the effect of these arrangements was to make Fiorente both the owner and controller of the dangerous thing, and therefore the party potentially liable under the rule in Rylands v Fletcher.

Establishing that Fiorente was potentially liable would not show that the courts below were wrong to regard Rubis as liable. The two propositions are not mutually exclusive. There can be two (or more) persons who satisfy the test for liability, as for example in Rainham Chemical Works where both the company and its directors were held liable under Rylands v Fletcher. It is unnecessary to decide whether Rubis is correct in submitting that Fiorente was potentially liable in this case, as Fiorente has not been sued. What matters is whether or not Rubis was responsible for bringing the fuel which escaped onto the premises and for the conditions under which it was kept there.

The Board has no doubt that it was. Two features of the arrangements between Rubis and Fiorente, taken together, lead to that conclusion.

First, although the fuel stored on the site had been purchased by Fiorente, Rubis (as the lessor) caused the fuel to be brought onto and kept on the land by requiring the premises to be operated as a service station (subject to strict operating requirements) and by requiring fuel supplied (exclusively) by Rubis to be stored, handled and dispensed at the premises. Second, although the “Lessor Equipment” was leased to Fiorente, Rubis provided all the equipment and, under clauses 6 and 17 of the lease, retained the rights to inspect, test, install, repair, replace or remove any of it. Further, pursuant to clause 9(B) of the lease, Rubis was solely responsible for maintaining (other than by carrying out certain routine procedures) and repairing the tanks and other equipment. These arrangements gave Rubis a high degree of control over the integrity of the underground storage systems.

On these facts, the courts below were justified in treating Rubis as responsible for introducing and keeping on the land the fuel which escaped, and therefore as potentially liable under the rule in Rylands v Fletcher (assuming the rule to be engaged). Whether the rule is engaged depends on whether the courts below were also right to find that the use of the land as a service station for the storage, sale and supply of petroleum products was a “non-natural” use of the land.

On the appeal to the House of Lords in Rylands v Fletcher (1868) LR 3 HL 330, 338, after making the point that the defendants “might lawfully have used [the land on which the reservoir was constructed] for any purpose for which it might in the ordinary course of the enjoyment of land be used”, Lord Cairns distinguished between “natural” and “non-natural” use of the land. It appears from a careful reading of his speech that his intention was to distinguish between hazards arising on land as a result of natural features or events, such as rainfall or the natural percolation of water, and hazards created artificially by bringing onto and keeping on the land something dangerous which was not naturally there. He applied this distinction to explain how the facts of Rylands v Fletcher and of an earlier case, Baird v Williamson (1863) 15 CB (NS) 376, differed from those of Smith v Kenrick (1849) 7 CB 515, a case relied on by the defendants. In Smith v Kenrick the defendant was held not to be liable for damage caused when rainwater which accumulated in an underground lake on his land flowed into the plaintiff’s mine as a result of the defendant’s working of his own mine. Lord Cairns was distinguishing the facts of Smith v Kenrick when he said, at p 339, that:

“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 …” (Emphasis added)

Lord Cranworth distinguished the facts of Smith v Kenrick in a similar way when he said, at p 342:

“If water naturally rising in the defendants’ land (we may treat the land as the land of the defendants for the purpose of this case) had by percolation found its way down to the plaintiff's mine through the old workings, and so had impeded his operations, that would not have afforded him any ground of complaint.”

As convincingly shown in another article by FH Newark, “Non-Natural User and Rylands v Fletcher” (1961) 24 MLR 557, in referring to “non-natural use” Lord Cairns therefore appears merely to have been confining the rule—as Blackburn J had already done—to a person “who has brought something on his own property which was not naturally there” (see the second passage quoted at para 52 above).

In later cases, however, the term acquired a further connotation. As Lord Hoffmann observed in Transco, para 44, whatever Blackburn J and Lord Cairns may have meant by “natural”, the law was set on a different course by the opinion of Lord Moulton in Rickards v Lothian [1913] AC 263. This was an appeal to the Privy Council from the High Court of Australia. The facts were that water overflowed from a basin on the top floor of a building leased to the defendant and damaged books in a dealer’s storeroom two floors below. The flooding was found to have occurred after an unidentified stranger had maliciously blocked the waste pipe from the basin, turned the tap full on and left the tap running.

The Board decided that the defendant was not liable under the rule in Rylands v Fletcher because the damage was caused by the wrongful act of a third party. But the Board also held as an additional reason for its decision that the rule was not engaged because, in the words of Lord Moulton, at p 280:

“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.”

This statement of the law was treated as authoritative by the House of Lords in Cambridge Water and in Transco, save for the reference to what is “proper for the general benefit of the community”. In Cambridge Water, at p 308, Lord Goff observed that, if the interest of the local community or the general benefit of the community at large were to be taken into account, it was “difficult to see how the exception can be kept within reasonable bounds”. In Transco Lord Bingham said that “little help is gained (and unnecessary confusion perhaps caused) by considering whether the use is proper for the general benefit of the community” (para 11); Lord Hoffmann said that this consideration had resulted in distinctions being drawn that are “sometimes very hard to explain” (para 37); and Lord Walker said that “the court cannot sensibly determine what is an ordinary or special (that is, specially dangerous) use of land by undertaking some utilitarian balancing of general good against individual risk” (para 105).

The Board would add that the kind of balancing referred to by Lord Walker is the role of planning and other administrative authorities, which—unlike a court—are in a position to assess all the competing interests at stake and to take account of a broad range of environmental, social and economic considerations in judging whether a use of land is “for the general benefit of the community”. Such an assessment is not relevant to the court’s task of deciding whether one party has a private law right against another. More specifically, the principle that, as between claimant and defendant, the defendant should compensate the claimant for harm arising from the defendant’s dangerous use of land is not affected by showing that the use in question benefitted others or the community at large. The existence of such a community benefit does not provide a reason for leaving costs incurred if a dangerous substance escapes to be borne by whoever in the vicinity is unfortunate enough, through no fault of their own, to suffer damage which is a reasonably foreseeable consequence of the escape.

Even when the last words of Lord Moulton’s statement of the law are put aside, the exception that it describes is ambiguous. The expression “ordinary use of the land” is not just “lacking in precision”, as Lord Goff said of it in Cambridge Water at p 308E. It can be read either as simply the converse of a use of land “bringing with it increased danger to others” or as a further requirement. Put another way, it is unclear whether use of land is to be considered special rather than ordinary justbecause it is abnormally dangerous; or whether the scope of the rule is confined to use of land which is both abnormally dangerous and (judged by some other criterion) not ordinary use.

In Transco Lord Bingham took the latter view. He considered that, for the rule to apply, two conditions must be satisfied—although he did not think that they can be viewed “in complete isolation” from each other (para 10). First:

“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.”

The second condition was that “the defendant’s use is shown to be extraordinary and unusual” (para 11). Lord Bingham said that this could also vary according to time and place, so that “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”. Combining the two conditions, Lord Bingham summarised the overall question as being whether the defendant “has brought or kept on his land an exceptionally dangerous or mischievous thing in extraordinary or unusual circumstances” (para 11).

Lord Hoffmann, on the other hand, viewed a natural (or “ordinary” in Lord Moulton’s formulation) use of land as the converse of a use creating increased risk: see para 44. So did Lord Hobhouse, at para 64, and Lord Walker, who said that “the twin requirements [of dangerousness and ‘non-natural use’] are best understood if they are taken together” and that “[i]t is the extraordinary risk to neighbouring property, if an escape occurs, which makes the land use ‘special’ for the purposes of the principle in Rylands v Fletcher” (para 103).

The difference between these approaches did not affect the determination of the appeal in Transco because all the members of the House of Lords were agreed that the supply of water through a 3-inch pipe connecting the water main to a block of flats did not give rise to an exceptional risk or special hazard: see para 13 (Lord Bingham); para 49 (Lord Hoffmann); para 67 (Lord Hobhouse); para 91 (Lord Scott); and para 112 (Lord Walker). While Lord Bingham also expressed the view that the use by the council of its land to carry such a pipe was “entirely normal and routine” and could not be seen as “in any way extraordinary or unusual” (see also Lord Walker at para 111), it was unnecessary for the House of Lords to decide whether this factor alone would have been sufficient to preclude liability under the rule in Rylands v Fletcher.

Lord Bingham’s formulation of the law was treated as authoritative by two members of the Court of Appeal of England and Wales in Gore v Stannard [2012] EWCA Civ 1248; [2014] QB 1. In that case tyres stored on the defendant’s land caught fire and caused damage to neighbouring premises. The Court of Appeal gave several reasons for holding that the defendant was not liable under Rylands v Fletcher. One, which some might think too subtle, was that what escaped from the land was fire produced by combustion of the tyres and not the tyres themselves. Another, straightforward reason was that the storage of the tyres would not reasonably have been perceived as creating an exceptional risk. But a further reason given by Ward LJ was that keeping a stock of some 3,000 tyres on the premises of a tyre-fitting business “was not for the time and place an extraordinary or unusual use of the land”: para 50. Etherton LJ also gave this as an alternative reason for the decision, if he was wrong on the two preceding points. He said, at para 68:

“The commercial activity carried on by the defendant 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 …”

In this case it is necessary to decide the question which did not have to be decided in Transco or in Gore v Stannard whether, as suggested by Lord Bingham in Transco, unusual risk and unusual use of land are separate requirements and, if so, how—other than by asking whether it creates an exceptionally high risk—what constitutes an extraordinary use of land is to be judged. Rubis accepts that petroleum products present an exceptionally high risk of harm if they escape. But Rubis argues that the storage of petroleum products in underground tanks at a service station is nonetheless an ordinary use of land in New Providence in the present day and that this has the consequence that the rule in Rylands v Fletcher is not engaged.

Rubis invites the Board to take notice of the fact that, in The Bahamas in the present day, the primary method of transportation is by vehicles powered by petroleum products, which must be obtained from a service station like the one in this case. Such service stations are commonplace, including in urban and residential areas. Accordingly, Rubis submits, this use of land cannot be regarded as one which was, or which Rubis ought to have recognised as being, “quite out of the ordinary in the place and at the time” in question (to quote Lord Bingham in Transco: see para 102 above). Nor is there any evidence to suggest that the quantity of petroleum products stored or the particular storage arrangements at the service station were out of the ordinary for similar premises carrying on such a business. Hence, it is argued, the rule in Rylands v Fletcher does not apply.

In further support of this argument, Rubis relies on the approach taken by the courts of many states in the United States. As summarised in the reporters’ note on section 20 of the Restatement of the Law Third, Torts: Physical and Emotional Harm (October 2024 Update):

“One interesting issue that has engaged a number of modern courts involves underground gasoline storage tanks that regularly accompany service stations. Such service stations, and their storage tanks, are commonplace within American communities; and most Americans purchase their gasoline at service stations. Accordingly, there is a substantial argument, accepted by many—although not all—courts, that the maintenance of underground storage tanks is not an abnormally dangerous activity.”

The factors relied on by those courts which have accepted this argument are said by counsel for Rubis to be: (1) that the storage of petroleum products can be made safe through the exercise of reasonable care; (2) that service stations are appropriate in and near residential areas since they provide residents with necessary, desired, and convenient sources of fuel for their vehicles; and (3) that the presence and use of petrol service stations in and near residential areas is commonplace. Rubis submits that these considerations apply equally to service stations in Nassau in The Bahamas such as the one which is the subject of this case.

Of the three factors derived from US case law, the first is not a relevant factor under English common law, which is the primary source of common law in The Bahamas. That is because, as discussed earlier, the basis of liability under Rylands v Fletcher is not that the defendant’s activity creates a risk of harm which cannot be prevented by the exercise of reasonable care. It is that a person who keeps something peculiarly dangerous on land necessarily imposes on others in the vicinity a risk of harm which it is just that the person who creates the risk should bear irrespective of whether the harm that results could have been prevented by the exercise of reasonable care. What makes something dangerous for this purpose is not the risk that it will escape from the land, or that it will escape even if reasonable care is taken; it is that it creates an exceptional risk of harm if it does escape, “however unlikely an escape may have been thought to be”: Transco, para 10 (Lord Bingham).

The second factor derived from US cases—the desirability and convenience of sources of fuel in and near residential areas—is, as counsel for Rubis noted in their written case, similar to Lord Moulton’s reference in Rickards v Lothian to whether the relevant use of land is “for the general benefit of the community”. The reasons for not taking this consideration into account have been given above.

The third factor—whether the use of the land is commonplace—is the critical one for present purposes. It is by no means uncontroversial even in the United States: see eg Steven Shavell, “The mistaken restriction of strict liability to uncommon activities” (2018) 10 Journal of Legal Analysis 1. It prompts the question: why should a person who engages in an exceptionally dangerous activity be shielded from strict liability for harm arising from it by the fact that the activity is commonplace?

The justification put forward by counsel for Rubis is that, where the activity is ordinary and everyday, so too are the risks it creates, such that mutual sufferance can reasonably be expected. They adopt the reasoning of Professor Donal Nolan in The Law of Tort, 3rd ed (2014), para 23.20, in a passage also now to be found in Winfield & Jolowicz on Tort, 21st ed (2025), para 16–012. This identifies the logic of restricting liability under Rylands v Fletcher to the extraordinary or abnormal as being that:

“if the activity is one carried on by a large proportion of persons, then the incidence of harm and responsibility are so nearly co-extensive that nothing is gained by the imposition of strict liability. Unless there is a special danger created by a small minority at the expense of the general public, strict liability merely substitutes a risk of liability for a risk of loss. … Where the activity is ordinary and commonplace, so too are the risks it creates, and so (in the absence of negligence) ‘mutual sufferance’ can reasonably be expected.”

This reasoning invokes the idea of reciprocity which is fundamental to the law of private nuisance. The classic statement of the principle is that of Bramwell B in Bamford v Turnley (1862) 3 B & S 66, 83–84. After referring to various activities—such as burning weeds, emptying cess-pools and making noises during repairs—which he thought (unless done wantonly or maliciously) could not be treated as nuisances, Bramwell B said:

“There must be, then, some principle on which such cases must be excepted. It seems to me that that principle may be deduced from the character of these cases, and is this, viz, that those acts necessary for the common and ordinary use and occupation of land and houses may be done, if conveniently done, without subjecting those who do them to an action.”

Bramwell B justified this principle in the following way:

“There is an obvious necessity for such a principle as I have mentioned. It is as much for the advantage of one owner as of another; for the very nuisance the one complains of, as the result of the ordinary use of his neighbour’s land, he himself will create in the ordinary use of his own … The convenience of such a rule may be indicated by calling it a rule of give and take, live and let live.”

This rationale has force where the activity is ordinary and commonplace in the sense discussed by Professor Nolan—that is, “the activity is one carried on by a large proportion of persons” and is not one which would reasonably be regarded as exceptionally dangerous. It would apply to facts such as those of Rickards v Lothian, where the relevant use of land was supplying water to a wash basin by routine plumbing within a building. As Lord Moulton observed in that case, at pp 281–282:

“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 modern sanitary views, an almost necessary feature of town life … Such a supply cannot be installed without causing some concurrent danger of leakage or overflow. It would be unreasonable for the law to regard those who instal or maintain such a system of supply as doing so at their own peril, with an absolute liability for any damage resulting from its presence even when there has been no negligence.”

The principle of reciprocity or mutual sufferance engaged in such a case, however, does not apply to damage which results from a use of land which is specialised (rather than carried on by a large proportion of persons) and exceptionally dangerous—judged by reference both to the likelihood that damage will occur in the event of an escape and the likelihood that such damage will be severe. Requiring those injured if such a risk materialises to bear the costs themselves cannot be justified by a rule of “give and take, live and let live”.

The differences discussed earlier between the nature of liability under Rylands v Fletcher and in the law of private nuisance are relevant here. What amounts to a nuisance may depend upon the character of the locality. As Thesiger LJ famously observed in Sturges v Bridgman (1879) 11 Ch D 852, 865, what would be a nuisance in Belgrave Square (a purely residential neighbourhood) would not necessarily be so in Bermondsey (which at that time was an industrial district where many tanneries were located). Thus, in a locality where, for example, a particular type of industrial activity is reasonably to be expected, an occupier of land cannot complain that noise or fumes or other interferences which are ordinary incidents of that activity amount to a nuisance. Such inconveniences are matters which those in the locality are expected to put up with.

The same cannot be said if injury results from a use of land which, even if not extraordinary or unusual for the time and place, is nonetheless a specialised activity, not carried on by a large proportion of persons, which involves keeping on the land a dangerous thing—that is, something which gives rise to an exceptionally high risk of causing damage if there is an escape. The fact that the use in question is not out of place in the locality may be a good reason to permit it and to require those who occupy nearby land to tolerate ordinary inconveniences to which it gives rise. But it is not a good reason to allow a person who imposes this risk on others to leave those others to bear the costs of harm that results if the risk materialises. Nor is the justice of requiring the person who creates such a risk to bear those costs diminished by establishing that the locality is one in which other enterprises are also exposing innocent third parties to similar risks.

The Board therefore rejects the contention that it is sufficient to displace the rule in Rylands v Fletcher that the activity carried on by the defendant is commonplace or not for the time and place an extraordinary or unusual use of land. What brings the principle into play is that the use is special and not merely an ordinary use in the sense that the activity carried on by the defendant is a specialised one, not carried on by a large proportion of persons, which is exceptionally dangerous if it leads to an escape.

This analysis is supported by the unanimous opinion of the House of Lords in Cambridge Water, where Lord Goff (with the agreement of the other law lords) specifically disapproved the judge’s finding that the defendant’s use of land to store chemicals used in its tanning business could be regarded as a “natural or ordinary” use of land. Lord Goff made it clear that the facts that this use of land was common in the tanning industry and that the defendant’s premises were located in an “industrial village” (where tanneries had been located for over 350 years: see [1994] 2 AC 264, 270) did not justify that finding. He said, at p 309:

“Indeed I feel bound to say that the storage of substantial quantities of chemicals on industrial premises should be regarded as an almost classic case of non-natural use; and I find it very difficult to think that it should be thought objectionable to impose strict liability for damage caused in the event of their escape.”

What made this “an almost classic case of non-natural use” and justified the imposition of strict liability was that the storage of substantial quantities of chemicals on industrial premises is an activity which, as is obvious, gives rise to an exceptionally high risk of causing damage if there is an escape. As discussed earlier, the rationale for imposing strict liability is that a person who for their own purposes keeps a peculiarly dangerous thing in proximity to others necessarily imposes upon those others a risk of injury of a kind which, even if reasonable care is exercised, it is unfair to expect them to bear as part of the ordinary vicissitudes of social life. The justice of this principle is reinforced by the practical reality that the persons who undertake such activities are typically large commercial enterprises whereas the persons who suffer the consequences are likely often to be members of the public with modest resources and who have no means of knowing, when considering whether to bring a claim, what precautions the prospective defendant had taken and how these compared with what is expected in the industry.

Applying the principle to the present case, the Board is content to assume that the use of the land situated at the junction of Robinson Road and Old Trail Road in Nassau as a petrol service station is not out of keeping with the character of the locality and in that sense is an ordinary use of the land. From this it would follow that Ms Russell and others living nearby could not complain that inconvenience reasonably to be expected from the operation of the service station, such as noise from vehicles coming and going or light pollution if the premises are floodlit at night, constitutes a nuisance. But although storage of large quantities of petroleum products is an ordinary incident of using land as a service station, it is certainly not an activity carried on by a large proportion of persons. It is a specialised and dangerous activity which does not constitute an “ordinary” use of land in the sense relevant to the rule in Rylands v Fletcher. The risks of harm in the event of leakage are of an entirely different order from the type of inconveniences which those living in the vicinity can reasonably be expected to bear themselves. Leaving nearby residents such as Ms Russell to bear the costs of damage sustained if fuel escapes unless they can prove negligence cannot be justified on the basis of reciprocity or the rule of “give and take” between neighbours.

The Board concludes that, as with the storage of chemicals discussed in Cambridge Water, the storage of petroleum products in substantial quantities is a classic example of “non-natural” use of land giving rise to strict liability for damage caused in the event of an escape.

Quantum of damages

The last issue raised on this appeal is whether the Court of Appeal erred in its assessment of damages. This issue remains relevant in light of the conclusions reached above because if, on remission, the court were to find that Ms Russell’s property was contaminated by fuel from the 2012 leak, the parties and the court would need to know whether the damages recoverable are the sums assessed by the Court of Appeal or whether a fresh assessment of damages is required.

The starting point is that Rubis is liable only for loss attributable to any additional contamination of Ms Russell’s property caused by the 2012 leak and not for loss that would have been suffered in any event as a result of contamination caused by the earlier 1994 leak: see Performance Cars Ltd v Abraham [1962] 1 QB 33; Davies v Bridgend County Borough Council [2024] UKSC 15; [2025] AC 434. Because the judge had decided that Rubis was liable for damage caused by both the 1994 leak and the 2012 leak, he did not need to distinguish between the effects of the two incidents and made no attempt to do so. But once the Court of Appeal had held that Rubis is liable to compensate Ms Russell only for damage caused by the 2012 leak, and not for damage caused by the 1994 leak (see para 9 above), it became necessary to do so.

The Court of Appeal did not review all the relevant evidence and make findings about the extent of the contamination, if any, of the soil and water table of the property attributable to the 2012 leak. That is why it is necessary to remit the case for such findings to be made. Yet this meant that the Court of Appeal lacked the necessary factual basis for assessing damages for loss caused by contamination resulting from the 2012 leak.

The Court of Appeal was alert to this difficulty when considering Ms Russell’s claim for damages for diminution in the value of her property but thought that it did not preclude making such an award. At the trial evidence was given by a property valuation expert, Mr Wilshire Bethell, who had prepared an “appraisal report” dated 11 September 2014. Mr Bethel estimated that the fair market value of Ms Russell’s property at that date “under normal circumstances” would be $225,000 but that its fair market value at that date “as is, as a result of the environmental impact caused by oil seepage contamination, in the soil below” was $94,000—a difference of $131,000. As Barnett P observed, Mr Bethell made no attempt to apportion this diminution in value between the 1994 leak and the 2012 leak. Barnett P also recognised that “the state of the evidence does not permit a court to apportion diminution to each leak”. But he nevertheless decided that the contamination was “more likely than not attributable to the [2012 leak]” on the grounds that the 1994 leak had occurred some 20 years before and that “much of the effects of that leak may well have been ameliorated by remediation efforts immediately after that leak”: see para 64 of the Court of Appeal judgment. On that basis the Court of Appeal felt able to find that damage caused by the 2012 leak resulted in a loss of value of $131,000.

It was not legitimate, however, to speculate in this way and to base an assessment of damages on such speculation without any evidence to support it. To do so was inconsistent with the court’s recognition that the state of the evidence did not permit it to determine what diminution in value was attributable to the 2012 leak. Whether diminution in value is an appropriate measure of compensation at all unless there is evidence that any physical damage caused by the 2012 leak was irremediable, is a question that would need to be considered on any fresh assessment.

The judge did not award damages for diminution in value of Ms Russell’s property but awarded damages for “costs of remediation”. At the time of the trial in 2020, over seven years after the 2012 leak, no such costs had actually been incurred. But Ms Russell claimed the astonishing sum of $782,905 based on a report dated 6 November 2014 provided by Mr Bowleg of Adarie Engineering and Environmental Services (“AEES”)— the same expert who had reported on groundwater contamination (see para 20 above). This report consisted of no more than a series of large numbers—such as $300,000 for “Remediation System Installation” and $150,000 for “Remediation System O&M”—said to represent “preliminary lump sum” costs “based on a minimum three-year contract with the client”.

These figures were said to relate to two properties: the Russell property and the Edgecombe property. (The Edgecombe property was the residential property to the north of Robinson Road for which Mr Bowleg had also included analytical results in his earlier, August 2014 report: see para 21 above.) The “anticipated annual remediation cost” was said to be $146,458 per property per year. No details were given of the proposed remediation work and no explanation provided of why such work was allegedly needed when the August 2014 report had shown only minute traces of contaminants far below the levels identified as “conservative” clean-up numbers to protect groundwater. Yet the judge awarded a sum of $439,375 for which he gave no reasons at all beyond the fact that this was the amount stated in the November 2014 AEES report for a period of three years.

The Court of Appeal rightly held that this award could not be justified. As Barnett P observed, at para 65, it would be unreasonable to spend on remediation a sum far in excess of the value of the property. The figures on which the judge relied were in any case wholly unsubstantiated. The suggestion that it was necessary or reasonable to incur costs of this magnitude—or indeed any costs at all—to remove the traces of BTEX reported in the first AEES report is, to put it bluntly, absurd.

The judge also awarded $250,000 for “loss of amenity value”. The Court of Appeal noted that there was limited evidence provided of loss of amenity related to the 2012 leak—the main evidence being Ms Russell’s testimony that after this leak she had noticed a stronger odour of fuel in her home (see para 14 above). Barnet P inferred that the loss of amenity was “minimal” (para 71) and reduced the award of $250,000 to a sum of $25,000. Again, it appears that the Court of Appeal did not attempt to distinguish between the contamination from the 1994 leak and the contamination from the 2012 leak when making such a substantial reduction in the damages. This award still assumed both that Ms Russell’s evidence given long after the event was reliable and that the odour of fuel that she recalled noticing after the 2012 leak was attributable to fuel which had migrated underground from the service station to the soil on her property. Whether such a distinction should be made is a matter that can only properly be decided by considering and analysing all the evidence bearing on the question (including, for example, the Black and Veatch report). As already mentioned, the Court of Appeal did not attempt this task.

Accordingly, the Court of Appeal’s assessment of damages cannot stand and must be set aside because it was not based on findings about what, if any, physical damage to Ms Russell’s property resulted from the 2012 leak. The question of damages must be considered afresh, if it arises, on the basis of relevant factual findings made by the judge at first instance when the case is remitted.

For the reasons given, the courts below were right to hold that Rubis is liable under the rule in Rylands v Fletcher for any damage caused to Ms Russell’s property by the 2012 leak (over and above any damage resulting from the 1994 leak). But neither court assessed the evidence and made findings of fact based on the evidence about whether any, and if so what, damage was actually caused by the 2012 leak. The Board has concluded that the case should be remitted to the Supreme Court to decide that question. That said, the costs of this litigation must already far exceed the maximum amount of damages which could realistically be recovered if the claim succeeds. If the parties are well advised, they will cut their losses at this stage and bring an end to these proceedings by consent rather than incur yet more disproportionate expense.

If no such voluntary resolution is reached, directions for the new trial of the outstanding issue will need to be given at a case management conference. It will be for the judge dealing with the matter to decide whether the issue should be determined solely on the evidence adduced at the original trial or whether to admit any further evidence (bearing in mind that it is now over 13 years since the leak occurred).

The Board will humbly advise His Majesty to allow the appeal and remit the case to the Supreme Court on the basis set out above.