Supreme Court of Canada

Liverpool & London & Globe Insurance v. Canadian General Electric Ltd., [1981] 1 S.C.R. 600

Date: 1981-05-11

Liverpool & London & Globe Ins. Co. Ltd., Aetna Casualty & Surety Co., Aetna Insurance Co. of Hartford, Canadian Indemnity Co., Canadian Surety Co., Commonwealth Insurance Company, Continental Insurance Company, Dominion of Canada General Ins. Co., Eagle Star Insurance Company (of Canada), Employers’ Liability Assce. Corp. Ltd., Federal Insurance Co., of New Jersey, Fireman’s Fund Insurance Co., The General Accident Assur. Co. of Canada, General Security Insurance Co. of Canada, Glens Falls Insurance Co., Great American Insurance Co., New York, Hartford Fire Insurance Company, Home Insurance Company, Insurance Company of North America, Maryland Casualty Company, National Employers’ Mutual General Insurance Assoc. Ltd., Norwich Union Fire Insurance Society Ltd., Orion Insurance Co., Ltd., Pearl Assurance Co. Ltd., Phoenix Assurance Co. Ltd. of London, Prudential Assurance Co. Ltd., Reliance Insurance Co. of Philadelphia, Royal Exchange Assurance, St. Paul Fire & Marine Ins. Co., Scottish & York Insurance Co. Ltd., Simcoe & Erie General Insurance Co., Sun Insurance Office Ltd., The Travelers, United Provinces Insurance Co., United States Fidelity & Guaranty Co., United States Fire Insurance Co., Zurich Insurance Company (Defendants-Respondents) Appellants;

and

Canadian General Electric Company Limited (Plaintiff-Appellant) Respondent.

1981: February 17, 18; 1981: May 11.

Present: Laskin C.J. and Ritchie, Dickson, Beetz and Estey JJ.

ON APPEAL FROM THE COURT OF APPEAL FOR ONTARIO.

Insurance—Policy interpretation—Fire and explosion clause—Epoxy resin overheating—Whether or not explosion or imminent peril of explosion occurred giving rise to liability for damages.

[Page 601]

Plaintiff-respondent unsuccessfully brought an action against its insurer under an insurance policy covering damage caused by “fire, lightning, explosion”. The trial judge found that no fire or explosion had occurred and that no damage was recoverable because of action taken to meet an imminent peril arising from an insured risk. The Court of Appeal reversed that decision and ordered a new trial with respect to liability for damages resulting from actions taken to meet imminent peril and, if the doctrine of imminent peril were inapplicable, with respect to liability and damages on the question of whether or not an explosion had occurred.

An exothermic chemical process occurred in a storage tank of epoxy resin during an accidental shutdown of the tank’s refrigeration mechanism. A manhole cover was removed to inspect the contents and left open as no fumes were being emitted. Firemen, who were called later, observed hot, black vapours, and left the hatch open to prevent a build-up of pressure. Advised by a plant official, firemen directed water into the tank, but discontinued this action when a disturbance of debated violence first threw solid pieces through the hatch and then spewed out a lava-like emission. Phenol vapours spread throughout the plant causing corrosive damage to plant and materials; there were also incidental costs relating to cleanup and general overhead. Two questions arose. Firstly, did an explosion occur, and if so, did respondent suffer recoverable damages? Secondly, if no explosion occurred, was there an imminent risk of fire or explosion justifying respondent’s preventive action, and if so, did that action result in recoverable damages?

Held: The appeal should be allowed.

An appellate tribunal could not properly conclude that the trial judge committed an error in determining that an explosion, in the sense in which that term was employed in the insurance policy, had not occurred. He properly considered the relevant evidence and related it to the ordinary meaning which is given to the term “explosion”. It is not the role of an appellate tribunal to retry the factual elements of the case.

As an imminent peril of explosion neither existed nor was feared when the avertive actions were taken, those actions and the consequential damage became irrelevant. In any event the damage was caused by an unin-

[Page 602]

sured risk and not by measures taken because of an “operating” imminent insured risk. Further, insured could not have converted the uninsured risk of damage from chemical reaction into the insured risk of explosion by closing the hatch. Finally, whether or not the firemen were the insured’s agents, the insured still could be in no better position as to the closing of the hatch than if the insured itself had made the decision to leave the hatch open or closed.

White v. Republic Fire Insurance Co. (1869), 57 Maine Reports 91; The Knight of St. Michael, [1898] P. 30; Kacianoff v. China Traders Insurance Company Limited, [1914] 3 K.B. 1121; Becker, Gray and Company v. London Assurance Corporation, [1918] A.C. 101; Joseph Watson and Son, Limited v. Firemen’s Fund Insurance Company of San Francisco, [1922] 2 K.B. 355; Beresford v. Royal Insurance Company Limited, [1938] A.C. 586, referred to.

APPEAL from a decision of the Court of Appeal for Ontario[^1], reversing a judgment of Keith J. and directing a new trial. Appeal allowed.

W.L.N. Somerville, Q.C., for the (defendants-respondents) appellants.

Warren H.O. Mueller, Q.C., for the (plaintiff-appellant) respondent.

The judgment of the Court was delivered by

ESTEY J.—The plaintiff-respondent brought an action unsuccessfully against its insurer, the defendants-appellants, under a policy of insurance which covered damage caused by “fire, lightning, explosion”. The trial judge found that no fire or explosion occurred and that no damage was recoverable as being the result of action taken by the insured to meet an imminent peril arising from an insured risk. The Court of Appeal reversed the trial judge and directed a new trial with respect to liability only on the claim related to damages incurred by reason of action taken in connection with imminent peril and with respect to both liability and damages on the question as to whether an explosion, within the meaning of that term as

[Page 603]

employed in the policy of insurance, had occurred should the doctrine of imminent peril be found to be inapplicable.

The relevant portion of the policy of insurance provided for:

…direct loss or damage caused by FIRE, LIGHTNING, EXPLOSION of natural, coal or manufactured gas and the perils set out in the EXTENDED COVERAGE ENDORSEMENT(S) contained herein and subject to the conditions thereof…

While the trial lasted four days, and extensive evidence, oral and documentary, was taken, the facts and circumstances lend themselves to summary because the issues have now dwindled to but two.

The respondent manufactures electrical equipment in a plant in Peterborough, Ontario. The claim relates to a storage tank in that plant, in which, at the time in question, there were about 2,500 gallons of epoxy resin. By reason of circumstances no longer relevant to the issues, an exothermic chemical process commenced in the resin which ultimately led to phenol vapours spreading throughout the plant and occasioning considerable corrosion damage. Relatively incidental damage arose as a result of the need to clean up some epoxy resin which spilled out in the vicinity of the tank, and also in connection with the use of water in and around the tank, as well as a claim for general overhead attributable to this event.

From this point on, some detail is necessary in order to develop the ultimate issue before the Court. There had been a power failure in the City of Peterborough which caused the refrigeration and stirring equipment in the respondent’s plant to cease operating. Unfortunately no one in the standby crew was sufficiently familiar with plant operations to realize that the stirring and refrigeration equipment in connection with certain epoxy storage facilities required restarting. At 1:30 p.m. on August 1, 1971, when the plant was closed for annual vacations, a guard noticed steam escaping

[Page 604]

from a cooling system associated with the storage tank, and at 3 p.m., the acting general foreman of the respondent arrived at the plant and discovered that the mechanism for stirring the epoxy resin in the storage tank was not functioning. Unfortunately, none of the employees of the respondent knowledgeable about this storage tank and its contents could be contacted. Sometime shortly before 7:30 p.m., the general foreman of the plant was contacted and he directed that a large opening on the top of the storage tank, referred to in the evidence as a manhole cover, be opened to determine the condition of the epoxy resin. This created a circular opening about 20 inches in diameter. When the tank was so opened, a wooden prod was inserted and it was found that at least the top of the resin had solidified. As no smoke, vapour or steam was being emitted from the tank, the personnel called to the plant left at about 8:10 p.m. without closing the manhole cover. The only apparent damage to that point, according to these persons, was that the resin in the tank was no longer useable.

It was not until 9:25 p.m. that a neighbour called the fire department after observing a dense black smoke or vapour coming out of the windows of the plant in which the storage tank was located. The municipal firemen, on entering the building, found it filled with dense vapour, and their platoon chief observed steam coming from a small pipe connected with a cooling unit adjoining the storage tank. The firemen also observed hot, black vapours coming from the open manhole. The firemen immediately poured water on the exterior of the tank which was located in a concrete pit from which only the top foot or so of the tank emerged. No effort was made to close the manhole hatch as the firemen considered that such action would create a build-up of pressure in the tank. The platoon chief closed the valve on the pipe from which the aforementioned steam was emerging and after 30 seconds changed his mind and reopened it so as not to create a build-up of steam pressure in the pipe or in the facility from which the pipe led.

[Page 605]

No doubt by reason of the fact that the plant was closed for a vacation period, personnel familiar with the chemical characteristics of the epoxy resin in the storage tank could not be located. The supervisor of advanced engineering did arrive, however, at 10:15 p.m., and, while he was familiar with the storage tank, its associated equipment and its function, he had no knowledge of the chemical characteristics of the resin or of the chemical reaction taking place at that time in the storage tank. He did, however, instruct the firemen to direct water into the storage tank through the manhole, and this was done for about a minute. The operation was discontinued when a disturbance of debated violence occurred in the storage tank which threw solid pieces of epoxy resin through the manhole. This was followed by an emission of a lava-like substance which, for a brief period, continued to roll out of the manhole, across the top and down the sides of the tank into the pit in which the tank was situated, and onto the surrounding floor. All the while, the dense black fumes continued to emerge from the manhole, and this continued in diminishing quantity until the firemen left the scene about 9:45 a.m. the following morning.

From these circumstances, two questions arise:

a) Did an explosion occur, and if so, did the respondent suffer damage therefrom which it may recover from the appellant?

b) If the answer to (a) is no, was there an imminent risk of fire or explosion which would justify action by the respondent to avert these events; and if so, did the respondent take such “avertive action” which caused damage recoverable under the insurance policy?

The litigation may have become unnecessarily complicated and, to an appellate tribunal, confusing by reason of an opening statement at trial by counsel for the respondent which included the following:

I can tell Your Lordship that there are two essential underlying principles of law which are in dispute. They relate to the question of whether or not an imminent risk of fire and explosion at common law provides a claim

[Page 606]

under a fire policy for fire or explosion, the fire or explosion never having occurred or even started.

Now, I can tell Your Lordship it is not an easy point, and what I have done, My Lord, with respect to that issue is to assemble the cases which in plaintiff counsel’s submission are pertinent to that narrow issue, and I have a copy for Your Lordship and a copy for my friend, and I thought I would start not by arguing my case but simply by delivering these to Your Lordship because Your Lordship will, I think, be benefited by an understanding of that narrow legal issue which is what this whole case is all about. [Emphasis added.]

In any case, by the time the proceedings reached this Court, it had become clear that no fire actually occurred and that the respondent placed very little emphasis on argument concerning imminent risk of fire. It was contended, however, that an “explosion” occurred. The issues before this Court are for the most part confined to the conclusion reached by the Court of Appeal that the trial judge had misdirected himself as to the proper meaning of “explosion” as it appears in the above excerpt from the policy of insurance and as it relates to the actions taken by the respondent purportedly in order to avert the happening of the insured risk of explosion (and, to a lesser extent, of fire).

The appellant’s position is simply that the damage suffered by the respondent was due almost entirely to phenol corrosion of the surface of the plant structure itself and the machinery and material in the plant, and that this damage was not covered by the clause of the policy relating to fire or explosion, nor was this damage the result of action taken by or on behalf of the respondent to avert an imminent fire or explosion. The position of the respondent, on the other hand, is that an explosion did occur within the meaning of that term as employed in the policy, and that the damage suffered by the respondent related to that explosion in the same sense as smoke and water damage are regularly found to relate to and flow from fire so as to make such damage recoverable under such policies of insurance. Counsel for the respondent also advanced, but with much less

[Page 607]

emphasis, the proposition that the damages suffered by the respondent are nonetheless recoverable even if a finding were made that no explosion had occurred because they resulted either from action taken to avert an explosion or fire, or from the respondent not taking action (the closing of the manhole cover) which would have caused, in the respondent’s submission, a catastrophic explosion, the damages resulting from which would have been clearly included in the insured risks.

It should be observed at once that the trial judge found that no fire occurred, and indeed, that there was no apprehension of fire by the persons present at the time of this event. Counsel are agreed that no fire occurred, but the respondent, in this Court, pointed to some evidence which indicated the presence of flammable gases which were emitted from the storage tank. There was conflicting evidence from the experts as to whether the gases known to have been present would support a flame. One witness conducted experiments to show that a flame, once started, would be sustained by the methane and other gases present, while two other witnesses came to the opposite conclusion. There was considerable evidence that efforts were immediately made by the respondent’s employees to ventilate the building by fans and the opening of windows. There was also evidence which indicated that there were in the plant, at the time, exposed pilot lights, electric motors and lights; and that despite all these fire hazards, no fire occurred. Taking all these circumstances into account, it can hardly be said that there was no evidence to support the finding by the learned trial judge that the peril of fire was not imminent, either in the minds of those present or in fact. The trial court when on to find that there was no imminent danger of explosion.

The only knowledgeable person at the scene during the evening of the 1st August was Mr. Mulhall [senior chemist of the respondent]. In his evidence he expressed no fear of explosion, but thought that some of the fumes and vapours being emitted might be flammable, and recommended ventilation.

[Page 608]

This action, of course, was not a cause of the damage but rather the reverse.

As for the others present, their actions were based on total ignorance which cannot possibly be a basis for claiming that the steps actually taken were as a result of a real and properly founded apprehension of the insured risks—fire or explosion.

This then narrows the issue down to whether or not an explosion occurred resulting in damage to the property insured, and, failing that, whether there was an imminent peril of explosion which could justify preventive action by the insured which itself entailed recoverable damage.

The learned trial judge found that no explosion occurred.

In the result, I have no doubt that what in fact occurred and which caused the damage cannot properly be characterized as an explosion, but is correctly described, if one is speaking English other than loosely, as an eruption,—an entirely different thing and not within the insuring agreement.

Before reaching that conclusion, the Court examined the meaning of the word “explosion” and the word “eruption” in a number of dictionaries and then considered the evidence of at least one witness introduced by the plaintiff-respondent who proposed a more technical definition of “explosion”. The trial judge said:

I have thought it necessary to set out the standard dictionary definitions of the words “explosion” and related words because of the attempt of counsel for the plaintiff to lead extrinsic evidence through Mr. W.J. Cruice as to whether or not what took place immediately following the pouring of water into the storage tank, would fall within his own definition of the word “explosion”.

There is no doubt that Mr. Cruice is a highly qualified chemist specializing in the study of fire and explosion hazards. He is a senior executive of Hazards Research Corporation of Denville, New Jersey, and an officer of a national committee on Hazard Potential of Chemicals. Sub-Committee E27 of this committee is responsible for considering technical definitions or words including the word “explosion” and a list of seven alternative definitions was submitted to the members of that sub-committee in July 1977.

[Page 609]

I raised the question of the admissibility of any such evidence at once, but because the case was being tried without a jury, I was not called upon to make an immediate ruling and permitted Mr. Cruice’s evidence to be given, reserving my right to consider subsequently as to whether or not it was admissible.

I have no doubt that it was inadmissible and must be wholly disregarded.

The Court then considered the evidence of the appellants’ expert witness, Dr. Harry L. Williams, a professor in the Department of Chemical Engineering and Applied Chemistry at the University of Toronto, and concluded that introduction of water into the tank did not cause any “sudden chemical reaction” which could be an “explosion”. This witness also related the events here to the word “explosion” as will be seen later.

From this, the Court of Appeal came to the conclusion that the trial court had erroneously adopted a definition of “an explosion” from the field of chemistry. Secondly, the Court of Appeal found error in the apparent distinction made in the court below between an “eruption” and an “explosion”, the Court of Appeal being of the opinion that such terms were not “mutually exclusive categories”.

Counsel for the respondent urged upon this Court the argument that the event here amounted to an explosion in the sense the word is employed in the policy of insurance. For this he found support in various dictionaries. Counsel for the appellants likewise placed reliance upon dictionary definitions of the term and applied them to the circumstances here through the analogy of a volcanic disturbance. If the island on which the volcanic action occurred was destroyed in the process, one would say, in ordinary terminology, that an explosion had occurred; whereas if only lava were emitted, however violently, one would say that the event was an eruption. The parties were in essential agreement in this Court on the appropriate dictionary definition of the word “explosion” and the meaning of that word as employed in the insurance policy here in issue. It is the application of that meaning to the events disclosed in the evidence by the trial judge that raises differences

[Page 610]

between them. An appellate tribunal is in an invidious position when reviewing conflicting conclusions concerning an event which was required to be classified according to the terms of a policy of insurance. In the ordinary concept of judicial processes, it is no function of an appellate tribunal, such as this Court, to rework the evidence so as to revise, modify or reverse a finding of fact reached at trial. Where, however, inferences are drawn from those facts, the appellate tribunal has a role to play. Furthermore, where the analysis of the facts depends upon the correct interpretation of the document in question, an error of law amounting to misdirection of the tribunal of fact is, of course, subject to review by the appellate tribunal.

The facts regarding the alleged explosion can be easily stated. After the firemen had introduced water into the storage tank through the manhole for about one minute, a violent expulsion of steam occurred. This was accompanied by chunks of the resinous material being flung upwards through the manhole, some even reaching the ceiling some 40 to 50 feet above the storage tank. This stage was immediately followed, if not accompanied in part, by a lava-like flow of epoxy resin out through the manhole. This flow apparently lasted for a very short period of two or three minutes. The black vapours and fumes continued to escape through the open hatch in the tank before and after this event. It is clear and undenied that damage from phenolic corrosion likewise occurred before and after this event. The first question is, does this episode amount to an explosion in the sense of that word as it appears in the above quoted portion of the policy of insurance? Then, if the answer is yes, what damage, if any, resulted from such explosion?

Some facts concerning this event are in dispute; some are not. The parties agree that the water entered into no chemical reaction with the epoxy resin; there was no reported sound associated with the event; the building was not damaged; but, a fracture does appear in one photograph of the housing surrounding the rotating stem which drove

[Page 611]

the circulating paddles in the storage tank; there was a spattering of material on the ceiling immediately above the tank; the floor surrounding the tank was covered in some areas by a dark, lavalike substance; and no windows were broken. The fire platoon chief and his helper were, according to the platoon chief, thrown some 20 feet back from the storage tank when the steam was emitted. The trial judge stated with reference to this fact:

There is no evidence of even minor injuries being suffered by Ray or his unfortunate driver and I am inclined to think that their flight only seemed like 20 feet, and that what is more likely is that being understandably startled, they stepped back and fell.

The trial judge found:

…that a comparatively small amount [of resin] had burst out of the open manhole following the pouring of water into the tank.

and later concluded:

The inescapable conclusion is that the force of the eruption was directed vertically through the manhole. Neither the tank nor the building suffered structural damage, no noise was reported to have been heard by anyone, and no windows were broken. There were no further eruptions after the initial expulsion of vapours and pieces of solid epoxy and the lava like flow described by...[the platoon chief] which must have been of brief duration, having regard to the fact that most of the contents of the tank were still there when everything cooled down.

There is evidence, therefore, in support of the conclusion by the trial judge that the episode did not amount to an explosion as that word was found in the contract as construed by the Court. This Court does not ordinarily alter a finding of fact below unless in reaching it or a conclusion with reference thereto, the court below when making the finding did not follow the law in instructing itself as to the proper tests to be applied. While the learned trial judge does refer to the fact that no witness suggested that the “introduction of the water caused any chemical change in the contents of the tank”; and that “the water had no effect on the exothermic reaction”; and that “there was no

[Page 612]

sudden chemical reaction from the water”; nevertheless the Court did direct its attention to the available dictionaries in current use and did apply the definitions there examined to the facts as described by both lay and expert witnesses. The trial judge appears to have preferred the description of the episode by Dr. Williams. This expert, on more than one occasion in the evidence, described the situation in the storage tank at the critical moment as being analogous to bread dough rising because of pockets of gas developing in it. After describing that the gas evolved from the epoxy resin as a part of a degradation process in the course of the exothermic reaction, Dr. Williams then stated that the material would bubble in the course of this process, and would resemble “hot toffee”. The gas, however, did not evolve as bubbles coming up through the toffee, but “It is more like bread dough and the whole thing bubbles up like bread dough”.

He then described the actual event which followed the introduction of the water into the reactor:

My opinion, on this matter is that introducing water into the reactor cooled, at most, a little surface layer of the overheated plastic.

At about the same time, the evolution of gases reached such a stage that they were available in the space.

The boiling water and gases threw out of the reactor various amounts of black decomposition powder which we have called lava.

I don’t think the amount of water put in there had any significant effect on the total reaction.

It might have cooled a very thin layer at the top, and it would certainly, of course, cool a little bit of the material flying out, but it could not have released the heat from the interior of the mass; the deep, several feet thick layer of polymer which we had in the reactor.

I would guess that the effect was insignificant overall.

[Page 613]

It was my impression, really, that the eruption was pretty well coincidental with the application of water more than anything else.

Elsewhere in cross-examination, however, Dr. Williams expressed views on the general nature of “explosions”:

Q. Can you tell me, please, sir, in your opinion is it probable that an explosion would have occurred if the manhole cover had not been opened?

A. According to my definition of an explosion, no, sir, it would not have taken place. It would have been a rupture.

Q. What is your definition of an explosion, Doctor?

A. I use the chemical definition of a rapid chemical reaction producing gases in the sense of a gasoline explosion or an explosion of nitro-glycerine, something like that.

I do not include in my definition of an explosion the rupture of a balloon or the rupture of a tire.

I think that is, perhaps, going a bit too far.

Q. You wouldn’t include the explosion of a boiler-a steam boiler in your definition?

A. No, I would call it a rupture.

Q. That is not an explosion?

So that you would not agree with Mr. Cruice that in the modern parlance of 1970 an explosion is a rapid equilibration of a high pressure gas with the environment?

That would not be within your definition of an explosion?

A. You have two definitions of a—or I guess you have three definitions of an explosion.

His Lordship mentioned a common knowledge term explosion which is really what we are concerned with.

There is the chemical explosion where you have gases mixed with oxygen and it explodes.

And then you have to have some word, as far as ASTM is concerned, what we would call sudden releases of pressure, and they call that an explosion.

The Court of Appeal, speaking through Madame Justice Wilson, observed in connection with the evidence of Dr. Williams:

It seems to me that in adopting Professor Williams’ concept of what constituted an “explosion” the learned trial judge was led into error. Professor Williams had

[Page 614]

made it very clear that his definition of explosion was not the definition of common use but was the chemical definition. He acknowledged that he would not have included the explosion of a steam-boiler in his definition because of the absence of any rapid chemical reaction. In my view, Professor Williams’ misunderstanding of what constituted an explosion within the meaning of the policy rendered his opinion as to whether or not there was a risk of explosion if none of the avertive steps had been taken completely valueless.

The trial judge drew certain conclusions and made certain findings from this rather extensive record. First of all, as regards the consequence of putting the water through the manhole, His Lordship observed:

Nothing happened for a little over a minute and then there was a violent eruption through the “manhole” including solid pieces of material, followed by a “lava-like” mass that continued to roll out.

Later in his judgment reference was made to the amount of destroyed material recovered, and then he stated:

With a substantial quantity of water being poured suddenly on a very hot surface in a largely enclosed tank, what would one expect other than that this would be followed by a violent expulsion of steam, as Watt discovered in the 18th Century.

The inescapable conclusion is that the force of the eruption was directed vertically through the manhole.

The learned trial judge then reviewed certain of the expert testimony including that of Dr. Williams and concluded:

In the result, I have no doubt that what in fact occurred and which caused the damage cannot properly be characterized as an explosion, but is correctly described, if one is speaking English other than loosely, as an eruption,—an entirely different thing and not within the insuring agreement.

In analysing the evidence surrounding the question of explosion, the trial judge examined the dictionary definitions of the words “explosion” and “eruption” and words derived therefrom. He had earlier rejected evidence tendered by the plaintiff-

[Page 615]

respondent on the experts’ definition of an explosion. As seen from excerpts already quoted, one of the experts, Dr. Williams, indeed advanced three definitions or categories of explosions, and there is nothing in the reasons given at trial to indicate that the learned trial judge adopted in any way the restricted chemical definition of an explosion.

The interpretation of words as employed by the parties in an insurance contract is a judicial task no different in principle than the interpretation of any other contract unless the process involves the relationship between the contract and its terms, and applicable statute law, which is not here the case. Lord Sumner in Becker, Gray and Company v. London Assurance Corporation[^2] spoke about the process in this way at pp. 113-4:

One need only ask, has the event, on which I put my premium, actually occurred? This is a matter of the meaning of the contract, and not, as seems sometimes to be supposed, of doing the liberal and reasonable thing by a reasonable assured. This is why, as it seems to me, the causa proxima rule is not merely a rule of statute law, but is the meaning of the contract writ large.

Taking into account the considerations borne in mind by the trial judge as evidenced by the terms of the judgment itself, and having examined the evidence of Dr. Williams, and bearing in mind the fact that the trial judge rejected the recourse to scientific terminology to define the word “explosion” as it is found in the policy of insurance, I conclude with all deference to those who have arrived at other conclusions that there is no basis on which an appellate tribunal can properly conclude that the trial judge here committed an error in determining that the events described in the evidence in the record did not amount to an explosion in the sense in which that term is employed in the insurance policy.

This then takes one to the remaining issue as to whether there was present an imminent peril of explosion, and if so, whether action described by

[Page 616]

the Court of Appeal as “avertive steps” taken by the respondent resulted in recoverable damage. The parties have agreed upon the damage suffered in the event that in law the plaintiff-respondent is entitled to recover under the doctrine of imminent peril.

The respondent, in this Court, urged the presence of an imminent peril on the basis of three elements:

1. the actual occurrence of an explosion;

2. the threshhold event that happened, namely the commencement of the exothermic cycle which was irreversible;

3. the fire department refrained from closing the manhole in order to avoid an explosion which would have produced damage recoverable under the policy.

The respondent submits that the damages suffered by reason of its action taken to minimize damage and avert imminent perils resulted from:

a) fumes coming from the storage tank;

b) solid material thrown out of the storage tank; and

c) water discharged into the building to avert or reduce existing perils.

The respondent further asserts that five steps taken or not taken by it were to avert the insured peril or minimize the damage which would ensue from that peril materializing. These are:

1. leaving the manhole open;

2. pouring water into the tank;

3. pouring water around the tank;

4. reopening the steam valve on the refrigeration unit; and

5. opening windows in the plant.

These considerations are irrelevant if there was no “imminent peril” of an explosion. This raises questions of fact and of law. The evidence of Dr. Williams accepted by the trial judge was to the

[Page 617]

effect that there was in fact no imminent peril of explosion. The learned trial judge found that there was no reasonable apprehension of an explosion. This finding is no doubt drawn from the evidence of the senior scientist of the repondent who stated that there was the hazard of fire had the building not been ventilated, but in the words of the trial judge “gave no indication that he had any anxiety that the mass of epoxy resin itself would explode”. There is, in short, nothing in the record to indicate that those present at the time the alleged avertive actions were taken, that is from 9:30 p.m. onwards on August 1, held any concern that they were under a peril of explosion from the storage tank and its contents. The conclusion by the trial judge that there had not been a risk of explosion in existence is supported by the evidence, and therefore it is unnecessary to deal with the actions taken to avert the alleged danger. However, the respondent has argued that the affirmative decision taken not to close the hatch cover on the manhole was properly an ‘avertive’ measure, because to have done so would have caused an insured event, an explosion, which one witness for the respondent said “would have been of catastrophic proportions”. As already mentioned, the manhole was opened at 7:30 p.m. in order to examine the contents of the storage tank. It was left open at that time through inadvertence. No fumes or vapours were at that time escaping from the tank. From 9:30 p.m. onwards when the vapours and fumes were being emitted from the manhole cover, no consideration was given to closing the manhole because this would cause an accumulation of pressure. Neither the firemen nor those present testified that any consideration was given to such a measure. The respondent now seeks to argue that the firemen had the right to close the manhole cover in order to seal off the fumes, and had they done so, the corrosive damage (which was the major cause of the respondent’s losses) would have been arrested and perhaps much of it avoided. The respondent concedes that in law the respondent itself had no right to do so because, simply put, it would be an attempt by the respondent to terminate an uninsured loss by creating a risk within the coverage of the policy, namely the risk of explosion. In any event, the manhole was not opened in

[Page 618]

order to forestall an explosion, nor was it left open (after the inspection at 7:45 p.m.) for that purpose. No doubt it was left open after the fumes and vapours were discovered escaping at 9:25 p.m., in order to avoid the risk of causing an explosion, but the explosion would only become imminent in the event the manhole cover was replaced.

Because of the findings of fact we are not called upon to determine the law of imminent peril as it exists today. In view of the disposition made of this issue by the Court of Appeal, however, reference should be made to the state of the law as it bears upon the position of the parties to an insurance contract when faced with the imminence of an insured peril. As long ago as 1869, the courts of the State of Maine, in White v. Republic Fire Insurance Co.[^3], stated, at p. 95:

…the imminence of the peril must be apparent, and such as would prompt a prudent uninsured person to remove the goods; it must be such as to inspire a conviction that to refrain from removing the goods would be the violation of a manifest moral duty;

That case concerned a claim based upon the moving of goods out of a warehouse (which ultimately was not destroyed) onto the street where the goods in fact were damaged.

The doctrine began to evolve into its modern form in The Knight of St. Michael[^4] where the master of a vessel, detecting the presence of spontaneous combustion in a cargo of coal, unloaded the coal prior to reaching the contract destination, at a loss to the owners. No fire actually broke out, but the trial court determined “that there was an actual existing state of peril of fire and not merely fear of fire”. This decision came under examination by Lord Reading in Kacianoff v. China Traders Insurance Company Limited[^5], at p. 1128:

[Page 619]

The danger was present, and if nothing had been done spontaneous combustion and fire would have followed in the natural course.

From these and other cases there emerged in the courts in England the principle that the peril stated to be imminent must be within the insured risks of the policy and that the insured risk had “begun to operate as a peril” at the time of the taking of the preventive action which in fact brought about the damage to the insured. The factual situation in Kacianoff, supra, brings the rule into sharp focus. There the shipowner had contracted to deliver a cargo to Russia via Japan. The policy of insurance provided for coverage “against capture”. After the policy had been entered into and before the ship set sail from San Francisco, war broke out between Russia and Japan. The shipowner decided not to set sail from the United States and disposed of the cargo. The Court found that the loss was not occasioned by capture nor by the imminent peril of capture because that risk had not begun “to operate”. Lord Sumner, in Becker, Gray and Company v. London Assurance Corporation, supra, at p. 114, stated:

This is also why the reasonableness of the conduct of the Kattenturm’s captain and the unreasonableness of suggesting that he might have done otherwise are alike off the point. So long as his action was voluntary it was his action and not that of the captain of a British man-of-war, and the policy insures against the second, but against the first only when it amounts to barratry.

Rowlatt J. carried the matter forward in Joseph Watson and Son, Limited v. Firemen’s Fund Insurance Company of San Francisco[^6], when he stated at p. 358:

It is one thing to say that where a peril in fact existed one must take the view of the captain formed at the time the peril existed as to what would be the outcome of that peril, and must not say to him, “If you had held on you would have found that all would have come right,” or something of that sort. It is another thing to say that one must take the captain’s view whether the state of facts existed which are alleged to have constituted the peril.

[Page 620]

The damage there suffered by the insured arose when a cargo was sprayed by steam to extinguish a supposed fire. In fact, the cargo of “rosin” was not burning, and the “fire”, which the captain set out to extinguish by a steam spray, was itself only steam escaping from a broken pipe in the hold. The Court concluded that the rule “does not touch losses incurred in a mistaken attempt to avoid a peril in fact non-existent”. (at p. 359)

Although the rule may seem harsh when stated in the abstract, nonetheless it appears in distilled version in the texts, such as MacGillivray & Partington on Insurance Law, (6th ed. 1975, para. 1753) and commends itself to the application of the terms of the contract undertaken by the insurer and the insured:

Damages sustained due to the voluntary act of an insured to avoid a named peril are not a consequence of that peril and are not recoverable.

Essential to an understanding of the rule and its application is the condition that before liability arises there must be an operating peril of the type or category described in the insurance contract. The danger must be present in the sense that unless something is done, damage will ensue. It may be that in the vagaries of nature, actual damage may not have yet been suffered (as in the Maine case, supra), but if the peril has actually arisen and damage can be reasonably anticipated from the peril (assuming it to be in the contract an enumerated risk), then damage suffered as a result of the preventive measures taken by the insured will be recoverable. (The Knight of St. Michael, supra.)

Returning to the circumstances in this appeal, the critical question at this stage of the proceedings is not whether or not the insured event has occurred but whether or not the damage occasioned by the insured arose by reason of preventive action taken to avoid the imminent risk covered by the contract, namely an explosion. Here the exothermic reaction had begun and had reached an irreversible stage. This, however, did not on the evidence lead inevitably to an explosion and the

[Page 621]

evidence adduced by the plaintiff falls short of a demonstration of anything approaching inevitability. In any event, all of this in these proceedings is academic because, as has already been stated, the damage suffered by the respondent was caused substantially by phenolic corrosion (an uninsured risk) by vapours escaping through the hatch and not by an explosion, even if one in fact had occurred, or by preventive measures taken by the respondent in the face of an ‘operating’ imminent insured risk.

Madame Justice Wilson, in the Court of Appeal, stated:

…in a case where the avertive steps were successful and the insured risk was in fact averted, the damage caused in the course of taking the avertive steps would not be irrecoverable because no insured risk had preceded it.

I do not take this passage to indicate that the Court is therein stating anything different from the prerequisite of an operating peril. Rather the Court is disposing of the simple situation in which damages were incurred by reason of preventive measures whether or not the actual insured peril had culminated in damages during or in addition to the damages flowing from the preventive measures. If, on the other hand, it is intended to be an expression at variance with the requirement of the presence of an operating risk as a condition precedent to the recovery of damages, I must, with respect, disagree.

The other four steps taken (nos. 2 through 5 above) by the insured to avoid the materialization of the allegedly imminent peril themselves apparently caused little or no damage to the insured. Neither the proof of loss nor letters between the parties agreeing upon the quantum of damages attribute any significant damage to any action taken by the insured to avoid the happening of an explosion. Taking the losses claimed by the insured in the proof of loss as the yardstick, it would appear that out of a total of $324,743 claimed:

a) $6,000 was incurred in pumping out water and repairing the floor of the building;

[Page 622]

b) about $100,000 was paid out to clean up smoke and corrosion damage;

c) $5,500 was paid for the repair of the storage tank;

d) $33,000 was paid to replace the contents of the tank;

e) corrosion to materials on hand and work in process cost $180,000;

f) laboratory costs incurred in connection with recovery of materials were $95,000.

Of these damages so claimed, only the $6,000 to clean up the mess related to what might be generously called explosion preventive measures, namely the pumping of water and the opening of the windows. I do not refer at the moment to the added claim of overhead expenses incurred by the insured. The actual amount of damage agreed upon by the parties (assuming liability is found) is $238,000 rather than the $324,000 in the proof of loss. In the absence of any breakdown of the agreed damages, I have assumed the allocation to the several items to be in the same proportions as in the proof of loss. The parties, prior to trial, agreed on the amount of damages to be assigned to the preventive measures if they are found to be related to an insured risk and recoverable in law. Because of the basis upon which the parties conducted the trial (as mentioned at the outset of these reasons), no allocation was made of damages claimed by the respondent to have been suffered by reason of the happening of an explosion.

The ‘avertive’ measures enumerated earlier included the release of steam from the cooling system and the opening of the plant windows. As no damage is shown to have arisen therefrom, nothing more need be said about these measures. We then are concerned only with the consequences, if any, in law of the insured deciding not to close the hatch so as to prevent the escape of the corrosive vapours. To do so would convert a risk of damage from that source into a risk of damage

[Page 623]

from explosion which was an insured risk. It is clear, and the respondent did not raise any argument on the issue, that the law does not countenance an insured bringing on the happening of an event so as to reclassify the damages suffered from those related to an uninsured event, as those related to an insured event. As was stated by Lord Atkin in Beresford v. Royal Insurance Company Limited[^7] at p. 595:

On ordinary principles of insurance law an assured cannot by his own deliberate act cause the event upon which the insurance money is payable. The insurers have not agreed to pay on that happening. The fire assured cannot recover if he intentionally burns down his house, nor the marine assured if he scuttles his ship, nor the life assured if he deliberately ends his own life. This is not the result of public policy, but of the correct construction of the contract.

Vide MacGillivray & Parkington on Insurance Law (6th ed., 1975), p. 227, para. 553. Since the insured could not claim it had opened the hatch to prevent an explosion, and since it could not now restore the hatch to its closed position to cause the explosion, the position of the hatch has no relevance to the claim under the policy by the insured for the corrosion damage.

The respondent did, however, argue that the firemen were not the agents of the insured and that therefore, their decision to leave the hatch open to avoid the hazard of explosion and indeed the imminence of explosion, should not deprive the insured respondent of the right in law to claim the damage suffered by the insured flowing directly from the third party firemen’s decision to avoid the happening of an insured risk. Whether the firemen are the agents of the owner or not, the insured can be in no better position with reference to the possible closing of the hatch than if the insured respondent had itself made the decision to leave the hatch open or to close it. Since the insured could not, in law, claim any loss suffered by closing the hatch and causing an explosion, it could be in no position to claim damage which

[Page 624]

flowed from leaving the hatch open where the damage in that situation is uninsured. In any case, we do not have the situation where an explosion was caused by an independent third party closing the hatch.

The analysis of this transaction and its legal consequences must, in this Court, end where it began. It is not the role of an appellate tribunal, and certainly not of this one, to retry the factual elements of a case. Here the issue is essentially one of fact as to whether, given its ordinary meaning in the language, an explosion occurred so as to trigger the rights of the respondent under the policy of insurance. The trial judge has found that an explosion did not occur. He has also found that if there was an imminent peril of an explosion, the preventive actions did not give rise to any damage. The trial judge did not misdirect himself as to the law applicable in reaching these conclusions, and for these reasons, therefore, I would allow the appeal and restore the order of the trial judge with costs throughout to the appellants.

Appeal allowed with costs.

Solicitors for the defendants-respondents) appellants: Borden & Elliot, Toronto.

Solicitors for the (plaintiff-appellant) respondent: Outerbridge, Thomas, Mueller & Betts, Toronto.

[^1]: (1980), 1980 60 (ON CA), 106 D.L.R. (3d) 750; (1980), 27 O.R. (2d) 401; [1980] I.L.R. 1-1201.

[^2]: [1918] A.C. 101.

[^3]: (1869), 57 Maine Reports 91.

[^4]: [1898] P. 30.

[^5]: [1914] 3 K.B. 1121.

[^6]: [1922] 2 K.B. 355.

[^7]: [1938] A.C. 586.