Supreme Court of Canada
Schreiber Brothers Ltd. v. Currie Products Ltd. et al., [1980] 2 S.C.R. 78
Date: 1980-03-27
Schreiber Brothers Limited (Plaintiff) Appellant;
and
Currie Products Limited (Defendant) and Gulf Oil Canada Limited (Third Party) Respondents.
1980: March 10; 1980: March 27.
Present: Laskin C.J. and Martland, Ritchie, Dickson and Beetz JJ.
ON APPEAL FROM THE COURT OF APPEAL FOR ONTARIO.
Contracts—Implied condition of merchantability—Burden of proof—Right of appellate court to interfere with findings of fact by trial judge—The Sale of Goods Act, R.S.O. 1960, c. 358, s. 15(2).
The appellant, an experienced roofing contractor, undertook to replace a large sized roof at a Ford Motor Co. plant in Windsor. For this purpose it purchased a certain type of asphalt from the respondent, Currie Products Limited, a supplier of that product which was manufactured by the third party, Gulf Canada Limited, the other respondent. The action arose out of a claim for damages for breach of the implied condition of merchantability under s. 15(2) of The Sale of Goods Act. It was conceded that if Currie was found liable it would have an enforceable claim against Gulf. The trial judge, after a lengthy review of the evidence, gave judgment for the appellant for $68,291.88. The Ontario Court of Appeal set aside the trial judgment and dismissed the action.
Held: The appeal should be allowed.
Both the trial judge and the Court of Appeal were satisfied that there were no causes other than a defect in the asphalt to account for the failure of the roof, and both were satisfied that the roof failed in ordinary use. The difference of the Court of Appeal with the trial judge was on a narrow issue of the credibility of certain evidence which in turn resulted in ascribing to the appellant a burden of proof which, in the trial judge’s view of the law, had been met.
It would be open to an appellate court, where credibility of a witness was not in issue, to review findings of fact by a trial judge if they were based on a failure to consider relevant evidence or on a misapprehension of the evidence. An appeal, however, is not a complete rehearing.
[Page 79]
There is no doubt that the appellant, suing for damages for breach of the implied condition of merchantable quality, has the burden of proof of excluding its faulty workmanship as a probable cause of the blistering of the asphalt, and of excluding other probable causes after the asphalt was delivered. Whether or not there was a defect in the asphalt when it left the Gulf plant, there was still the fact of a defect in a product emanating from the respondents, and the state of the evidence was such as to make it specially appropriate for the trial judge to make inferences from what was put before him.
Annable v. Coventry (1912), 46 S.C.R. 573; Cities Service Oil Co. v. Rubel (1930), 66 O.L.R. 475; Métivier v. Cadorette, [1977] 1 S.C.R. 371; Powell v. Streatham Manor Nursing Home, [1935] A.C. 243; Prudential Insurance Co. Ltd. v. Forseth, [1960] S.C.R. 210; Hood v. Hood, [1972] S.R.C. 244; Canadian William A. Rogers Ltd. v. Lucerne Metal & Plastic Products Ltd., [1949] O.R. 135, referred to.
APPEAL from a judgment of the Court of Appeal for Ontario, reversing a judgment of the Supreme Court of Ontario. Appeal allowed.
Warren H.O. Mueller, for the appellant.
D.K. Laidlaw, Q.C., and Elizabeth Stewart, for the respondents.
The judgment of the Court was delivered by
THE CHIEF JUSTICE—This appeal, which is here by leave of this Court, arises out of a claim for damages for breach of the implied condition of merchantability under s. 15(2) of The Sale of Goods Act, R.S.O. 1960, c. 358, the relevant statute at the material time. The plaintiff is an experienced roofing contractor who undertook to replace a large sized roof at a Ford Motor Co. plant in Windsor. For this purpose it purchased a certain type of asphalt from the defendant, Currie Products Limited, a supplier of that product which was manufactured by the third party, Gulf Canada Limited. In this Court, Currie and Gulf were represented by the same counsel, it being conceded that if Currie was found liable it would have an enforceable claim over Gulf.
The trial judge, Henry J. of the Ontario Supreme Court, after a lengthy review of the
[Page 80]
evidence, gave judgment for the plaintiff for $68,291.88 and costs. He said this, in the concluding pages of his reasons:
In my opinion therefore, I am left with the probability that the fault lay with the product. I do not overlook the evidence that no such defect was experienced with Gulf’s Type 3 BUR before or since. That is a formidable record and I take it into account. However, on a common sense and practical view of all the evidence I am constrained to find on the balance of probabilities that the failure of the roof was caused by a latent defect in the Type 3 asphalt and that that defect could not be detected by prior inspection. It is not necessary to identify precisely in chemical terms what that defect was and I do not attempt to do so. It is sufficient that, however it is identified, it caused the asphalt to behave in such a way that it was useless for the purpose of constructing the built-up roof in question.
There is no issue as to the quantum of damages if the plaintiff is entitled to succeed.
The Ontario Court of Appeal, speaking through Weatherston J.A., set aside the trial judgment and dismissed the action, saying this:
The finding that the asphalt contained a latent defect was an inference based on the exclusion of other known probable causes of the blistering, and the acceptance of a scientific theory which postulated the existence of a defect in the asphalt. The scientific theory simply does not stand up to analysis. Assuming that it was not necessary to identify the gas, the plaintiff still had to show that some gas could probably be generated at roof temperatures because of a defect in the asphalt. That has not been shown, and an inference cannot fairly be drawn that the failure of the roof was caused by the existence of such a defect. The plaintiff’s case fails for want of proof.
…
The trial judge found that there was a latent defect in the asphalt, but with respect, he never directed his mind to the crucial issue whether that defect existed in the asphalt when it left the Gulf Oil plant. The burden of proof was on the plaintiff and plaintiff failed to meet it. On this ground also, I would allow the appeal.
[Page 81]
There are various findings of fact, most of them concurrent findings, which lay behind the different conclusions of the two Courts and I shall come to them shortly. The roofing process and the nature of the damage which occurred are described as follows in the reasons of the Court of Appeal (after referring to the plaintiffs contract to lay a built-up roof, being a sloping and not a flat roof):
…That is a roof by which a waterproof membrane is formed by applying alternate layers of asphalt and asbestos felt on to the roof surface. The asphalt is brushed on at a temperature of about 425°F; that is well below its distillation point, but the asphalt is viscous enough that it can be easily brushed on. The felt may contain 5% by weight, or even more, of water; the temperature of the asphalt is high enough to drive this off; that causes a certain amount of bubbling for a few seconds. The felt is perforated to allow the escape of the steam, since the layer of felt must be applied before the asphalt has had time to cool and harden.
The work was commenced in April, 1967 and continued over a period of 15 months. The roof failed because large blisters formed in the membrane, caused by the expansion of some gas or water vapour. Some blistering was noticed as early as September, 1967 but was not regarded as serious until about two years after the work had commenced, when it was realized that the problem was general and serious.
The blistering was not such as has ever been encountered before by any of the witnesses, lay or expert. A small amount of blistering is usually encountered, resulting from entrapped water vapour, but this type of blistering has usual characteristics which make it readily identifiable. The blistering which occurred on this roof was quite different.
The trial judge found that the blistering that occurred could not be attributed to faulty workmanship on the part of the plaintiff and this finding was not questioned by the Court of Appeal. The finding that there was no faulty workmanship carried with it a finding, explicitly made by the trial judge, that the plaintiff did not allow air and water vapour to get in between the plys to produce the blistering. There was also a concurrent finding that the asphalt reached the plaintiff without any intermediate intermeddling and in the
[Page 82]
same condition in which it left the manufacturer, which had followed controlled and customary procedures in producing the asphalt. The trial judge was satisfied on the evidence that there were no causes other than a defect in the asphalt to account for the failure of the roof, thus eliminating any suggestion or contention that there was anything wrong with the felts or with the baseboard or with the inspection. The Court of Appeal accepted this assessment by the trial judge.
Both the trial judge and the Court of Appeal were satisfied that the roof failed in ordinary use. However, while the trial judge concluded that since other probable causes had been eliminated there was probably a latent defect in the asphalt, the Court of Appeal concluded that it was incumbent on the plaintiff to prove, even if not required to identify the defect, that the defect existed in the asphalt when it left the Gulf plant. The cause of the failure, said the Court of Appeal, was an enigma and there was a burden of proof on the plaintiff which it failed to meet. Alternatively, the Court of Appeal held that even if there was a defect in the asphalt it was not shown that the defect had not infected the asphalt after it left the Gulf plant. This basis of decision runs, of course, counter to the finding of the trial judge excluding other probable causes after the asphalt left the manufacturer.
It was not, of course, fatal to the plaintiffs case that the deterioration of the roof through the blistering formation was unlike anything seen by the expert witnesses. Their evidence indicated that the particular blistering was not that which would be produced by trapped air or water, and this was relied on by the trial judge in coming to his conclusion. He also accepted the evidence of a Dr. Horsch who said that he had heated a sample of the roof to 180°F. (summer roof temperature) and that it gave off a gas which was like formic acid. The witness did not keep a sample and was unable to produce the result again at the trial. Gulf challenged the evidence of Dr. Horsch, alleging
[Page 83]
that what he obtained was water vapour and that it could not have been any substance that gasified at 180°F. because the reaction would be about 8,000 times faster at 425°F. to 450°F., which was the temperature at which the asphalt was applied. In such a case, any gas would have bubbled off during application of the asphalt and the extensive bubbling and foaming would have been noticed during the 15-month period over which the work proceeded, but no note was made under the heading “foaming” on the inspection sheets. The trial judge, in accepting the evidence of Dr. Horsch, said that he was fortified in doing so by his finding that no air or water had been trapped in the roofing operations. Although not knowing what the gas was or how it got into the asphalt, the trial judge held that the plaintiff did not have to establish this when other causes had been eliminated.
It was in this respect that the Court of Appeal differed from the trial judge, finding the evidence of Dr. Horsch incredible and requiring proof that a defect existed in the asphalt when it left the manufacturer’s plant. In view of the various concurrent findings which I have mentioned, the difference of the Court of Appeal with the trial judge was on a narrow issue of the credibility of certain evidence which in turn resulted in ascribing to the plaintiff a burden of proof which, in the trial judge’s view of the law, had been met.
Counsel for Currie and Gulf asserted that this Court should not interfere with the Court of Appeal’s assessment of Dr. Horsch’s evidence nor with its determination that the plaintiff had a burden of proof to show, on a balance of probabilities, that the asphalt was defective when it left the Gulf plant. A number of cases were cited in support of the Court of Appeal’s right to review the trial judge’s findings of fact, among them Annable
[Page 84]
v. Coventry[1] and Cities Service Oil Co. v. Rubel[2]. In the latter case, the Ontario Court of Appeal adopted statements in the Privy Council and in the House of Lords that an appellate court is entitled to consider whether, on the evidence, it would have come to the same conclusion as the trial judge or, to put it another way, it can make up its own mind on the evidence, not disregarding the judgment at trial and giving special weight to it where the credibility of witnesses is in issue, but with liberty to draw its own inferences from proved or admitted facts. It appears that this averment in the Cities Service case related to cases where there was no conflict in the testimony but, in any event, I regard it as going too far in the light of more modern authorities. Anglin J., as he then was, said in Annable v. Coventry, at p. 587, that he agreed with the proposition that it was the duty of an appellate court, even where the appeal turns on a question of fact, to rehear the case and not shrink from overruling the trial judge if, on full consideration the court concludes that the trial judgment was wrong. “Rehear”, taken literally, is too strong and would, in my opinion, be wrong. Anglin J. could not have meant this literally.
It would, of course, be open to an appellate court, where credibility of a witness was not in issue, to review findings of fact by a trial judge if they were based on a failure to consider relevant evidence or on a misapprehension of the evidence. An appeal, however, is not a complete rehearing. A majority of this Court held in Métivier v. Cadorette[3], at p. 382, that it was wrong for an appellate court to set aside a trial judgment where the only point at issue was the interpretation of the evidence as a whole. Powell v. Streatham Manor Nursing Home[4] was relied on in the Métivier case (as it was in Prudential Insurance Co. Ltd. v.
[Page 85]
Forseth[5]) as emphasizing the advantages of a trial judge, who sees and hears the witnesses, in coming to a conclusion on the acceptability of evidence and on the findings to which the accepted evidence leads. However clear cut the principles may be, governing the right of an appellate court to interfere with the findings of fact by a trial judge, their application may involve a difference of opinion as to whether interference is warranted in a particular case: see Hood v. Hood[6].
In the present case, I do not think that the Court of Appeal’s rejection of Dr. Horsch’s evidence gave it any warrant to interfere with the trial judge’s conclusion on the issue on which the evidence was accepted. That evidence was not the only evidence upon which the trial judge acted but it was, to him, an affirmation of his finding that no other probable cause save a defect in the asphalt at the time that it left the Gulf plant could account for the blistering that occurred when the asphalt was applied. The Court of Appeal would have the plaintiff prove affirmatively that there was a defect, but the trial judge proceeded on the basis that such proof could be made by the elimination of all other probable causes, including the elimination of any probability that some contaminant had got into the asphalt after it left the Gulf plant. No fault could be ascribed to Currie, which merely stored the asphalt after it was received from Gulf and sent it on in the same containers to the plaintiff. However, this does not absolve Currie from liability under The Sale of Goods Act if it, in fact, supplied a product which was not of merchantable quality.
Counsel for the respondents submitted further that if it was not enough to hold the judgment in their favour to stand merely on the rejection of Dr. Horsch’s evidence by the Court of Appeal, he had a stronger argument that there was no proof that the asphalt was not of merchantable quality. The
[Page 86]
failure of the roof did not, he contended, establish the failure of the product; indeed, it was not shown that any defect in the asphalt caused the damage. This, however, engages the trial judge’s finding excluding other probable causes and is, in effect, a reaffirmation of the Court of Appeal’s assignment of a particular burden of proof to the plaintiff.
There is no doubt that the plaintiff purchaser, suing for damages for breach of the implied condition of merchantable quality, had the burden of proof, in the light of the pleadings, of excluding its faulty workmanship as a probable cause of the blistering of the asphalt. This it did, to the satisfaction of the trial judge and without challenge by the Court of Appeal. Again, in the light of the pleadings and of the evidence it was an obligation of the plaintiff to exclude other probable causes after the asphalt was delivered by Currie. The trial judge found that it had done so and, again, there was no remonstrance by the Court of Appeal. This leaves the question whether the existence of a defect in the asphalt, either when it left the Gulf plant or when it left Currie, the supplier, must be shown affirmatively or may be proved by inference from the evidence as a whole. Proof of a defect must be made, of course, whether the purchaser of the product is plaintiff, suing for damages, or is defendant, resisting an action for the price. The trial judge found, as I have already noted, that the asphalt reached the plaintiff in the same condition that it was in when it left the Gulf plant; in short, that the defect in the asphalt was not attributable to anything that Currie did or failed to do. The Court of Appeal accepted this finding as well.
The difficulty in this case arises from evidence that (1) the particular blistering was unlike anything ever seen before; (2) there had been no complaints either from Currie’s other numerous customers or from other middlemen who, like Currie, were suppliers; and (3) there was no proof, apart from Dr. Horsch’s evidence, of the nature of the defect and, had there been such proof, it might have been easier to attribute its onset to a particular time or event. There is, of course, the impor-
[Page 87]
tant fact that Gulf and Currie controlled the production and delivery of the asphalt, and this undoubtedly was relevant to the trial judge’s ultimate resolution of the case in favour of the appellant.
The Court of Appeal, in placing upon the plaintiff a burden of proof to show that, although the asphalt failed as a product, there was a defect in it when it left the Gulf plant, proceeded on the view that there must be a credible theory to account for the defect and that Dr. Horsch’s theory was incredible. I do not think that, in the circumstances of this case, the absence of a credible theory that would indicate that the defect was in the manufacturing process or occurred before the asphalt left the Gulf plant is conclusive against the appellant. The rejection of the gas theory of Dr. Horsch, a theory accepted by the trial judge, does not necessarily mean that the plaintiff fails. True, the gas theory was an element in the trial judge’s conclusion in favour of the appellant but, it seems to me, that if the inference drawn by the trial judge is impermissible, in the light of the evidence of regularity adduced by the respondents, a greater burden of proof than proof on a balance of probabilities is cast on the plaintiff.
This case can, of course, be looked at as exhibiting two possible causes of the defect, each of them unlikely in view of the evidence. We are still left, however, with the fact of a defect in respect of a product emanating from the respondents. The state of the evidence was such as to make it specially appropriate for the trial judge to make inferences from what was put before him. He certainly canvassed all relevant evidence and did not misapprehend it in any material way. The concurrent findings of fact are cogent indicators that he took a proper approach. I do not think that his conclusion ought to have been disturbed on appeal on the basis on which the Court of Appeal proceeded.
It appears to me also that the Court of Appeal treated the case before it as one sounding in
[Page 88]
negligence against Gulf rather than being a contract action against Currie. If it were not for the reference by Weatherston J.A. to Canadian William A. Rogers Ltd. v. Lucerne Metal & Plastic Products Ltd.[7], it would be difficult to appreciate that this was an action under The Sale of Goods Act. As in the Rogers case, it was enough here that the plaintiff show that a defect existed when the goods were delivered by Currie, without being required to prove the cause of the defect. The Court of Appeal in the Rogers case accepted the trial judge’s holding that the plaintiff had met that burden of proof. The Court of Appeal here should have come to a similar conclusion.
I would, accordingly, allow the appeal, set aside the judgment of the Ontario Court of Appeal and restore the judgment at trial. The appellant is entitled to the costs ordered at trial and to costs in the Court of Appeal and in this Court.
Appeal allowed with costs.
Solicitors for the appellant: Schreiber, Bordonaro & Yanover, Hamilton.
Solicitors for the respondent Currie Products: White, Swaye, Mackesy & Smye, Hamilton.
Solicitors for the respondent Gulf Oil: McGarry & McKeon, Toronto.