ROGOSHESKE, JUSTICE.

Appeal from an order denying a new trial after a jury verdict for defendant, Crookston Coca-Cola Bottling Company. The action is one for personal injury and consequential damages caused by an "exploding" bottle of carbonated beverage, and the appeal raises questions concerning the trial court's submission of the issue of contributory negligence and its refusal to submit the issue of liability on the theory of strict liability, as plaintiffs requested. We find reversible error and grant a new trial.

Plaintiff Helen Lee, a waitress, while working at the Norman Steak House at Ada, Minnesota, was injured when a quart (26-ounce) Coca-Cola bottle exploded in her hand and a glass fragment of the bottle punctured the median nerve at the level of her right wrist. Mrs. Lee and her husband, Claire, brought this action to recover damages upon pleaded theories of breach of implied warranty, negligence, and strict tort liability. The case was tried and submitted to a jury on the theories of breach of implied warranty and of negligence under the doctrine of res ipsa loquitur. The jury returned a general verdict for defendant, presumably upon a finding either of no breach of warranty or no negligence or of contributory negligence. On this appeal, plaintiffs principally contend that the trial court erred (1) in submitting the issue of contributory negligence, and (2) in refusing to submit their claim against defendant upon the theory of strict liability in tort.

The evidence relating to the injury-producing incident is not in substantial dispute. Defendant's driver delivered a 12-bottle case of Coca-Cola and other beverages to the steak house shortly before noon on the day of the accident. Using a handcart, he stacked the Coca-Cola case in the dining room area of the steak house behind a bar or counter about 2 feet from a waist-high, sliding-top-three-door-type, refrigerated cooler. The case and the bottles remained undisturbed until between 12:30 and 1 p. m., when Mrs. Lee (hereafter plaintiff) undertook to transfer the bottles from the Coca-Cola case into the cooler. Without moving the case, she grasped one bottle at a time with her right hand, moved it laterally to her right, and lowered it into the cooler. While she was lowering the third bottle into the cooler, it "exploded" in her hand, the neck of the bottle with the cap intact remaining in her hand. According to the testimony of plaintiff and the manager of the steak house, the only two women present at the time, the bottle exploded with a "loud bang," which the manager heard while working in the kitchen separated from the dining area. The manager came "running" to "find out what was wrong." She saw plaintiff standing by the cooler holding the cap and neck of the bottle and found glass fragments from the bottle scattered in and about the cooler and adjacent area, some being found "under the nickelodeon machine * * * approximately * * * ten or twelve feet away" and "some around the other end of the dining area." Since plaintiff and the manager considered the accident of little consequence, requiring only the application of a band-aid to the wound after the bleeding had been stopped, the bottle fragments were swept up and discarded.

Plaintiff testified that she was "positive" and there was "not one bit" of question in her mind that she had not "struck" the "bottle on anything" as she was transferring it to the cooler. She insisted that the bottle had exploded spontaneously while in her hand, and this testimony was uncontroverted. Defendant, in support of its denial of liability, introduced evidence of the absence of temperature extremes on the date of the accident and called its plant manager and its driver, who described in detail its bottling and delivery operations, including the precautions taken to guard against the use of defective bottles. In addition, Dr. Oscar Fryer, a retired physics professor, who has done research and testing for the parent Coca-Cola company since 1939 and who has testified in numerous bottle-explosion cases, was called by defendant. He testified at length regarding the physical properties of carbonated beverages, the effect of temperature and agitation upon the internal pressure of a bottle of carbonated beverage, and the results of breakage and internal-pressure tests conducted in defendant's plant with respect to seven 26-ounce Coca-Cola bottles picked at random from defendant's supply. He explained that three forces can cause a glass bottle to break, namely, thermo-shock, internal pressure, and external forces, and that if the bottle fragments are available, an expert in glass technology can determine what force or forces caused the bottle to fracture.

Because the fragments of the bottle in question were unavailable, and in view of plaintiffs' prima facie proof of liability under the doctrine of res ipsa loquitur, defendant sought Dr. Fryer's opinion testimony in answer to a hypothetical question to support its claim that the injury-producing explosion was caused by external force due to plaintiff's mishandling. The question finally posed, after plaintiffs' repeated objections on the ground of materiality and insufficient factual foundation, was directed to elicit an opinion as to what would cause a hypothetical bottle in the hands of a hypothetical person to explode. The question was based upon hypothetical facts relating to defendant's bottling and delivery procedures as testified to by its manager and driver, the expert witness' own observations at defendant's plant and the results of his testing, and in part on the testimony of plaintiff describing the circumstances of the incident resulting in her injury. Responding to the hypothetical question, Dr. Fryer stated, "It is my opinion that the bottle was struck a blow at the time that it failed." Over plaintiffs' objection, this opinion was ruled admissible and provided the basis for the court's submission, also over plaintiffs' objection, of the issue of contributory negligence. The court also refused plaintiffs' request to instruct the jury that if the jury found that the bottle was defective, the rule of strict liability should be applied.

1. Our careful examination of the record compels us to conclude that the trial court committed reversible error in submitting the issue of contributory negligence to the jury.

Except for the opinion testimony of defendant's expert witness, the record is devoid of any evidence upon which a finding of contributory negligence could be sustained. Where, as here, the expert witness has no firsthand knowledge of the exploded bottle such as he could gain from examining its fragments, or of its age, appearance, condition, and the circumstances of its handling during the course of bottling and prior to delivery into the hands of the injured party, or the injury-producing incident, the customary way of securing the benefit of the expert's scientific skill is to ask him to assume the truth of certain facts and then give his opinion based on such hypothesis. McCormick, Evidence, § 14, p. 30. However, the universal requirement for this type of hypothetical question is that the facts assumed must be supported by competent evidence in the case. Grapentin v. Harvey, 262 Minn. 222, 114 N.W.2d 578. The record makes clear that defendant's expert witness by his own admission did not assume as facts that plaintiff did not strike, bump, or agitate the bottle. These were the undisputed facts established by the unimpeached testimony of plaintiff, the sole witness to the accident. They could not be ignored as so inherently improbable as to be incapable of belief and were clearly required to be included in the hypothesis. On this record, no assumption could be made that plaintiff had intentionally or unwittingly or unknowingly struck the bottle against the cooler or otherwise subjected it to physical impact. Moreover, such an assumption as a basis for an opinion as to the cause of an explosion of a hypothetical bottle rather than the bottle which caused the injury introduces a collateral matter, clearly irrelevant, if not immaterial, to the factual dispute and, despite skillful cross-examination of the expert by counsel for plaintiffs, likely to create confusion and therefore prejudice to plaintiffs' claim of a spontaneous explosion. Thus, the expert's ultimate opinion that the cause of the explosion was a physical impact was not based on facts supported by the record but on an assumed fact which the evidence would not permit the jury to find as true. It was therefore without proper foundation, purely speculative, immaterial, and inadmissible. Such inadmissible evidence could not justify submission of the issue of contributory negligence, and because a finding that plaintiff had been guilty of contributory negligence could conceivably have been the basis for the verdict, a new trial is required.

2. A new trial is also necessary because we hold that the circumstantial evidence justifying submission of the issue of liability to the jury on the theory of negligence also justified submission of the issue on the theory of strict liability in tort, as pleaded and requested by plaintiffs.

Our intention to apply the rule of strict liability to manufacturers and sellers of defective products was announced in McCormack v. Hankscraft Co. Inc. 278 Minn. 322, 154 N.W.2d 488. This decision was based on policy considerations expressed in our prior case law and a host of other authorities alluded to in the Hankscraft case. Summarized, these policy considerations are: (1) The public interest in safety will be promoted by discouraging the marketing of defective products which constitute a menace to consumers not equipped to protect themselves from products they are induced to purchase through modern advertising methods by persuasive representations that the product is suitable and safe for its intended use; (2) the burden of loss caused by placing a defective product on the market should be borne by the manufacturer, who is best able to distribute it by insuring against inevitable hazards as a part of the cost of the product; (3) maximum legal protection should be afforded the consumer to promote product safety and to encourage the growing practice of reputable manufacturers and sellers of settling valid claims without litigation; and (4) one injured by a defective product should be entitled to bring action directly against the party responsible for putting the product on the market without becoming involved in the delay and expense of joining other sellers in the chain of distribution, as frequently occurs when liability is sought to be determined under warranty provisions of the Uniform Commercial Code.

The rule of strict liability, as revised and adopted by the American Law Institute in 1964, is embodied in Restatement, Torts 2d § 402A. It imposes liability, without proof of negligence or privity of contract, upon a manufacturer or seller for injury caused by a dangerously defective product. To recover under the rule, the injured party must present evidence, direct or circumstantial, from which the jury can justifiably find that (1) the product was in fact in a defective condition, unreasonably dangerous for its intended use; (2) such defect existed when the product left defendant's control; and (3) the defect was the proximate cause of the injury sustained.

The greatest difficulty in establishing liability under this rule is in proving that the product was defective and that the defect existed when the product left defendant's control. While in conventional tort terms no proof of negligence is necessary, in many cases proof of a defect may simply be a substitute word for negligence. Thus, strict liability does not mean that the defendant is held liable as an insurer of his product regardless of circumstances. As is true in negligence cases with respect to the mere fact of an accident, the mere fact of injury during use of the product usually is insufficient proof to show existence of a defect at the time defendant relinquished control. Kerr v. Corning Glass Works, 284 Minn. 115, 169 N.W.2d 587; Prosser, The Fall of the Citadel (Strict Liability to the Consumer), 50 Minn. L.Rev. 791, 843. As in the Kerr case, expert-opinion testimony may, and often does, take the issue of a claimed defect under strict liability to the jury. Daleiden v. The Carborundum Co. (8 Cir.) 438 F.2d 1017; Green v. American Tobacco Co. (5 Cir.) 325 F.2d 673. Also, liability is not imposed where the injured party has not eliminated the probability that improper handling by intermediate parties might have caused the defect. Kerr. v. Corning Glass Works, supra. Moreover, as under the theories of breach of warranty and negligence, recovery is barred by concepts of contributory negligence and assumption of risk if the injured party discovers the defect, knows the dangers arising from it, and nevertheless uses the product, or if he makes abnormal use of the product. Magnuson v. Rupp Manufacturing, Inc. 285 Minn. 32, 171 N.W.2d 201.

The narrow question presented here, however, is whether circumstantial evidence, the core of the res ipsa loquitor doctrine, is sufficient to take the case to the jury on the theory of strict liability as well as on the theory of negligence. This precise issue was discussed recently by way of dicta in Holkestad v. Coca-Cola Bottling Co. 288 Minn. 249, 180 N.W.2d 860. There, plaintiff was injured by an exploding 16-ounce bottle of Bubble Up. Upon instructions permitting the jury to infer negligence under the theory of res ipsa, a finding implicit in the verdict that defendant was negligent in using a defective bottle was upheld by this court. However, at the close of the evidence, the trial court held that the evidence, which consisted only of the circumstances of the injury-producing incident, was not sufficient to permit a finding that the bottle was defective and thereupon dismissed plaintiff's alternative theory of strict liability. Because of the manifest inconsistency in the court's ruling permitting a finding of negligence with respect to using a defective bottle and, under the same evidence, denying submission of the case on strict liability, we felt compelled to discuss the issue now directly presented in this case. In concluding that the court erred in dismissing plaintiff's claim of strict liability, we observed:

"* * * When a plaintiff has proved that he was injured by a product claimed to have been defective, and where the claimed defect is such that there is circumstantial evidence from which it can be inferred that it is more probable than not that the product was defective when it left defendant's hands, absent plaintiff's own want of care or misuse of the product, there is an evidentiary basis for submitting the issue of liability to the jury on both the theory of negligence and strict liability in tort." 288 Minn. 257,180 N.W.2d 865.

For reasons which follow, we hold that recent observation to be the appropriate rule governing the disposition of the question presented in this case.

It surely must be conceded that circumstantial evidence of the type present in Holkestad and in this case justifies submission of the issue of liability on the theory of res ipsa. Johnson v. Coca-Cola Bottling Co. 235 Minn. 471, 51 N.W.2d 573. Although on post-trial or appellate review, as in Holkestad, there often may be disagreement as to whether such evidence is sufficient to sustain recovery, it would be unusual if the trial judge did not permit the jury to pass on whether or not the balance of probabilities preponderates in favor of defendant's liability. Moreover, where the case is submitted on res ipsa, under this type of circumstantial evidence it would be strange indeed if the arguments to the jury, as in this and the Holkestad case, did not include conflicting contentions by counsel as to whether the evidence established that the bottler's use of a defective bottle was the cause of the bottle's failure. That such may be the only plausible cause of an exploding bottle is demonstrated by the evidence in this case. As testified to by defendant's expert, there are three fundamental causes of bottle failure: Thermo-shock, internal pressure, and external force. According to the expert's testimony, failure because of thermo-shock could only result from drastic changes in temperature applied to the outside of the bottle, such as would be produced by placing a bottle containing hot liquid in cold water. Failure caused by external force, of course, usually results from an impact, such as striking or dropping the bottle. Failure because of internal pressure due to excessive carbonation is ordinarily unlikely because the bottle is designed to withstand approximately four times the pressure created by the gas introduced. and after the carbonated liquid is added to the syrup mixture, any excessive carbonation is equalized by exposure to atmospheric pressure during the interval between carbonation and capping. The capacity of different bottles to withstand internal pressure varies, however, in part due to bottlers' customary reuse of bottles. Some bottles have been refilled for years and might have been subjected to "rough" handling numerous times, thereby increasing the probability that even though they are designed to withstand such handling, some could develop defects which would escape detection by the most careful bottler. This may be the only plausible explanation for the bottle's failure in this case, since there is uncontradicted evidence dispelling the probability that the failure was attributable to thermo-shock or external force. Absent expert opinion, as in a case of this type, circumstantial evidence may be the only available means of establishing a claim of either negligence or a defective product.

Under the theory of strict liability, the elements of proof as noted above are few and uncomplicated. The significant difference is that under strict liability the jury need not infer from the circumstantial evidence that defendant was negligent in order to impose liability. It is sufficient that the evidence establishes that the manufacturer placed a dangerously defective product on the market, knowing that it is to be used without inspection for defects. Defendant's evidence that it exercised the highest degree of care to guard against the use of a defective product may refute plaintiff's circumstantial evidence of the injury-producing incident and support a finding that the product was not defective. Thus, in many product liability cases — especially those involving exploding bottles —, under both the theory of negligence and the theory of strict liability, the injured party, in practical effect, has the burden of proving a defect at the time the product left the manufacturer's control. To argue that the same circumstantial evidence could not support a finding of a defective bottle sufficient to justify submission on the theory of strict liability seems clearly untenable and cannot be supported by reason or authority. Prosser, The Fall of the Citadel, (Strict Liability to the Consumer), supra; Keeton, Torts, 23 Southwestern L. J. 1; 2 Frumer Friedman, Products Liability, § 26.04.

Where res ipsa is relied upon as the theory of recovery, plaintiff is not required to allege or prove specific claims of negligence or, as in this case, a specific defect. Anderson v. Eastern Minn. Power Co. 197 Minn. 144, 266 N.W. 702; Segal v. Bloom Brothers Co. 249 Minn. 367, 82 N.W.2d 359. Indeed, plaintiff's inability to determine the specific defect or cause, coupled with the fact that the defendant is in a better position to present evidence as to the cause of the accident, is itself a fundamental reason for the res ipsa rule. The same reasoning applies with equal force to excuse plaintiff from proving a specific defect when pursuing recovery on the theory of strict liability. This is especially true in exploding-bottle cases and similar situations where the product is destroyed by reason of the defect, which is also obliterated, or where the defective part alone is destroyed by the incident. Grant v. Malkerson Sales, Inc. 259 Minn. 419, 108 N.W.2d 347; Franks v. National Dairy Products Corp. (W. D. Tex.) 282 F. Supp. 528. In short, under the theory of strict liability plaintiff should not be required to prove specifically what defect caused the incident, but may rely upon circumstantial evidence from which it can reasonably be inferred that it is more probable than not that the product was defective when it left defendant's control.

In this case, plaintiffs relied on the circumstantial evidence concerning the unusual accident to support an inference that the explosion which resulted in injury to plaintiff Helen Lee was more likely than not caused by a defect in the bottle of Coca-Cola. The defendant was then faced with the task of refuting plaintiffs' evidence. Being unable to submit evidence creating issues of contributory negligence or assumption of the risk on the part of the injured party, or of causation by reason of third-party intermeddling, the defendant was left with the defense of seeking to destroy or impair the credibility of the testimony of the injured party or the eyewitnesses to the event. In other words, defendant's recourse in these circumstances is to convince the jury that the incident is the kind of occurrence which could not have happened as described by the injured party or that it was a pure accident attributable to no defect in the product or no fault of the defendant. The jury could properly have found on the evidence submitted that defendant was not causally negligent. This finding, of course, defeats plaintiffs' claim on the theory of negligence. As has been pointed out above, however, it would not necessarily preclude recovery under the theory of strict liability. Under instructions solely on negligence, a jury might conclude that the bottle was defective when it left defendant's control but that defendant was not liable because the defect did not result from negligence. Under instructions on strict liability, on the other hand, a finding that the bottle was defective when defendant put it on the market would compel a verdict for plaintiffs, absent the aforementioned defenses and without considering the question of negligence.

Thus, the trial court's refusal to submit plaintiffs' claim upon the theory of strict liability in tort must also be regarded as reversible error. The court's ruling deprived plaintiffs of a legitimate choice of theories on which to submit the case. Plaintiffs are entitled to attempt to prove their case on either or both theories — that defendant was negligent or that it put a dangerously defective product on the market.

It could be argued that the case in effect was submitted to the jury on strict liability, since the jury was instructed on implied warranty. Although strict liability in tort and in warranty are very similar, we cannot view the court's instructions as sufficient to constitute submission of the question of strict liability in tort to the jury. The jury was told that defendant warranted that the bottle of Coca-Cola "was reasonably fit for the ordinary and usual handling as it might reasonably anticipate in the exercise of reasonable care." This language falls short of conveying to the jury that if a defect existed in defendant's product when it left its control, defendant should be found liable for the injuries caused by such defect.

Reversed and new trial granted.