These consolidated appeals arise out of a single products liability action. On August 16,1978, plaintiff Wayne Vaughn was driving a fully-loaded dump truck eastbound on Missouri Highway 174. As Vaughn approached and came to a point about .2 miles east of the intersection of Highway 174 and Highway PP in Greene County, his left front tire, a Michelin tire, exploded. The truck went out of control, left the road and overturned. Vaughn’s left leg was pinned beneath the hot exhaust pipe on the left side of the cab of the truck. He sustained serious, permanent and disabling physical injuries.
Vaughn filed suit in the Circuit Court of Greene County, seeking damages for personal injury. His wife Karen joined as a plaintiff, seeking damages for loss of consortium. Michelin Tire Corporation and Truxan Parts, Inc., were joined as defendants. Initial discovery revealed that the tire involved was manufactured by Michelin’s Spanish affiliate, S.A.F.E. DeNeumati-cos Michelin (Spanish Michelin). Accordingly, plaintiffs Vaughn filed an amended petition joining defendants Spanish Michelin as manufacturer, Michelin as distributor and Truxan Parts, Inc., as retailer of the tire. Plaintiff Norma Young, owner of the truck involved in the accident, intervened as plaintiff, seeking to recover damages for the destruction of the truck.
Plaintiffs’ case was submitted to the jury on the theory of strict liability in tort. The jury found the issues submitted for the plaintiffs and against defendants Michelin and Spanish Michelin. Plaintiff Wayne Vaughn’s damages were assessed at the sum of $750,000; plaintiff Karen Vaughn was awarded $75,000 for loss of consortium. Plaintiff Norma Young’s damages were assessed at $46,000. The jury found no liability on the part of defendant Truxan Parts. Appeal No. 14250 is the appeal of Michelin and Spanish Michelin from the verdicts and judgments entered against them; Appeal No. 14248 consists of plaintiffs’ cross-appeal from the verdict and judgment in favor of defendant Truxan Parts, Inc., and their appeal from the trial court’s refusal to award prejudgment interest.
Appeal No. 14250 is the more complex of the two appeals and we shall consider it first.
*550As noted, the accident which gave rise to this action occurred on August 16, 1978, as plaintiff Wayne Vaughn was driving a 1974 Ford dump truck eastbound on Highway 174 in Greene County. He was pulling a 14-foot Marx pup trailer when the tire on the left front wheel of the dump truck exploded. Vaughn attempted to bring the truck to a stop but it went out of control, ran off the road and rolled down an embankment. Vaughn was thrown from the truck; his left leg was pinned beneath the exhaust stack on the left side of the truck. Vaughn’s injuries have required several surgical procedures.
The tire on the left front wheel of the dump truck — referred to as the accident tire — was a Michelin 12.00 R20 XZY steel-belted radial truck tire. Plaintiffs sued Michelin on the theory of strict liability in tort, specifically alleging that the tire had been defectively manufactured. Plaintiffs claimed that a manufacturing defect had, over a period of time, caused a separation of the several belts, or plies, which lie under the tread of the tire. Eventually, the outer belts became completely detached from the casing of the tire. At that point, there was no support for the sidewall of the tire, “the inner tube [came] through,” and a blowout occurred.
Defendants Michelin admitted they sold the tire, but denied that it failed because of a manufacturing defect. Michelin contended that the proper air pressure had not been maintained in the tire; this “underinf-lation” or “overdeflection” caused the belts to separate because of an excessive heat buildup within the tire. Michelin suggested that the reason for the underinflation was that the tire had been partially penetrated by a nail or other foreign object.
Michelin has briefed and argued four assignments of error in this court. The record presented is diffuse and in some respects difficult to follow. Much of the expert testimony involved the use of exhibits, but in many instances the testimony was not correlated with the exhibit. And, aside from such considerations, a great deal of evidence, including some which is not altogether clear to the court, was received in the course of the 9-day trial. To the extent possible on the record presented, we have confined ourselves to a recitation of those facts and consideration of those issues essential and necessary to an orderly disposition of the appeal. See Bloomfield Reorganized School District No. R-14 v. Stites, 386 S.W.2d 95, 97 (Mo.1960); Southwest Engineering Company v. Reorganized School District R-9, 434 S.W.2d 743, 746 (Mo.App.1968).
A pretrial conference was held in this case, and two matters taken up during that conference are of primary interest on this appeal. During the discovery process, plaintiffs had served interrogatories on Spanish Michelin including one which required Spanish Michelin to disclose “[a] general description of the tire’s composition, tread design, and anticipated use.” Michelin replied that the tire was composed of: 1 radial steel carcass ply; 2 half belts (steel); 2 working belts; 1 protector belt; 2 bead bundles; 2 ■ bead stiffeners, and “[v]arious rubber compounds both natural and synthetic.”
About 10 months before trial plaintiffs had also requested production of documents showing the design specifications of the Michelin 12.00 R20 XZY tire. Spanish Michelin resisted this request on the ground that the design specifications constituted trade secrets. The court heard argument bn the motion for production and finally concluded that the design specifications of the 12.00 R20 XZY tire were indeed trade secrets. The court refused to require Michelin to disclose the design specifications of the 12.00 R20 XZY tire.
Before the pretrial conference was concluded, counsel for the plaintiffs elicited a stipulation from counsel for Michelin.1 The court inquired:
“Any other objections to the opening graphics?
*551MR. STRONG: If they’re through with mine, I need — do need one stipulation here for Paul’s.
He has a new tire, not the accident tire, but a new tire that he’s going to use and I’m going to use a section of it in opening statement.
And I just want to be sure that we all understand that that new tire is the same size, construction, design, and that it’s a duplicate of this tire that is in this lawsuit.
Is that right Paul?
MR. BRADSHAW: Yes, it is.
But when you’re going to say you’re going to use it in your opening statement, I specifically said a minute ago that I’m not going to produce it for you to use in your opening statement.
[[Image here]]
MR. STRONG: ... In all fairness, if they — if this is going to be an exhibit in the case and if it is going to be used by them in opening statement, we ought to have a right to explain, also. I’ll be happy to give credit to the jury. I’m not going to say, ‘Hey, look what we have’—
MR. BRADSHAW: We can settle this in a hurry, then. I just won’t use it in my opening statement, and I — we won’t produce it until my case.
I thought you might like to refer to it during your case. If you want to use it then, and I can in opening statement, you can.
Otherwise, we’ll make up our minds when we get to our case whether we’íl use it.
MR. STRONG: All right. Well, we’ll let the Court rule on the two positions.
We want to use it in opening statement, whether Paul uses it or not. It’s an exhibit that will be used at the trial, and we want to use it. It’s here and the statement I made about the stipulation is correct, is it not, Paul?
MR. BRADSHAW: About what stipulation?
MR. STRONG: That it is an exact duplicate of the make and model of our tire?
MR. BRADSHAW: Yes.
MR. STRONG: And so, if it is and it will be an exhibit, we have a right to use it in opening statement. We so request.
******
THE COURT: All right. I’m going to rule that the plaintiff will not be permitted to use it in opening statement....” (Emphasis added.)
The exhibit concerning which the stipulation was made was referred to as defendants’ exhibit C at trial. It was a 12.00 R20 XZY tire from which parts of the tread and carcass had been cut away to expose the arrangement of each steel belt (there are four layers) and the alignment of the cords which form each belt. Exhibit C will be discussed in more detail presently.
The case proceeded to trial the following week and, in opening statement, plaintiffs’ counsel explained to the jury his theory of defect. Though no direct use of exhibit C could be made because of the court’s pretrial ruling, counsel demonstrated that the top steel belt on exhibit C consisted of cords that “crisscrossed” those of the underlying belt. He then noted that the top belt on the accident tire had cords that did not “crisscross” and explained that such a belt lay up eventually caused separation between the plies and ultimate destruction of the tire. Essentially, counsel inferred that defendants’ exhibit C exemplified Michelin’s intended design of all Michelin 12.00 R20 XZY radial truck tires, including the accident tire. Thus, because there was a difference in the way the top belt “crisscrossed” the underlying belt, that difference must represent a defect in the way the accident tire had been manufactured.
At this point, Michelin’s counsel apparently shared the assumption that Michelin had selected exhibit G to demonstrate the intended design of the Michelin 12.00 R20 XZY radial truck tire. In his opening statement he also referred to exhibit C as “the same kind of tire that was on the truck” and later, with additional reference to exhibit D (a whole tire similar to exhibit C), stated “[b]oth of them are like the accident tire.”
*552Before the trial commenced, Michelin produced its exhibits C and D. Exhibit C was part of a tire with a cutaway section as shown in the diagram below. Exhibit D was a whole tire with an identical cutaway section.
[[Image here]]
The body of the tire, or casing, is constructed of steel cords laid in a radial ply, i.e., a ply which runs directly around the tire from bead to bead, the bead being that part of the tire which seals against the rim of the wheel. For support and stability in the tread, several steel belts are laid up beneath the tread. There are four layers of these steel belts, viz., 1) two “transitional” or “half” belts (one not visible in the diagram) which provide support for each “shoulder” of the tire; 2) an inner “principal” or “working” belt; 3) an outer “principal” or “working” belt, and 4) a “protective” belt. Each steel belt consists of numerous steel cords laid in a tight, parallel pattern, encased in hard rubber. In the manufacture of the tire, each belt is placed so its steel cords run at a particular bias, or angle.
Plaintiffs had evidence from one George R. Edwards, an “independent tire consultant,” who had examined what was left of the accident tire. Edwards gave as his opinion that the tire failure was caused by a manufacturing defect. Edwards was allowed to testify, without objection, that the tire was unreasonably dangerous as manufactured, and he further testified that the accident tire had not failed because of a puncture.
*553Edwards’ hypothesis of the manufacturing defect was summed up in a pretrial deposition given in January 1984. Though we agree with defendants that the deposition was not introduced in evidence, Edwards’ testimony on trial was merely a convoluted expansion of the testimony he gave by deposition. For the sake of brevity, we resort to that deposition. Edwards testified:
“I fould [sic] that the subject tire failed as a result of separation developing in the laminations surrounding the two outer steel belts. The physical evidence of separation is as follows. Erratic tread wear pattern, chafe and polish, rubber reversion. These three observations indicate the separations existed within the tire for a prolonged period of time prior to the time of tire failure.
As the integrity of the four steel belt system deteriorated, the tread, protector ply and outer steel principal belt separated three hundred and sixty [degrees] around the circumference of the tire and [those belts] became completely detached from the inner belt, two transitional belts and the radial steel tire cord body.
The remaining principal steel belt did not have sufficient structural strength to hold the unsupported steel radial body plies together, [and] the internal air pressure within the air chamber overcame the strength of the remaining principal steel belt. This caused the steel radial radial [sic] plies to ‘part’ resulting in the protrusion of the inner tube [into the] atmosphere and the inner tube to ‘blow out’ and a spontaneous loss of air pressure occurred....”
Edwards explained that the accident tire had been manufactured with the top belt cords running parallel to those of the “outer working belt,” which lay directly beneath the top belt instead of crisscrossing as illustrated by exhibit C. This positioning made the two belts stiff and inflexible and, bonded together against the underlying flexible inner working belt, created opposing forces which caused separation between those belts and the rest of the tire. Edwards did admit that as long as sufficient adhesion between the belts was present, the fact that the belt plies ran parallel made no difference in proper functioning of the tire. Another expert witness for plaintiffs agreed with the conclusions reached by Edwards.
Michelin’s cross-examination of plaintiffs’ expert witnesses led plaintiffs’ counsel to believe that Michelin would abandon its stipulation that the 12.00 R20 XZY tire was designed to be constructed as illustrated by exhibit C. Plaintiffs’ counsel believed that Michelin intended to offer some evidence which would indicate it also manufactured that particular tire with the two outer belts running parallel, in addition to the “crisscross” design shown by exhibit C. Plaintiffs’ counsel then reminded the court and Michelin of the parties’ earlier stipulation and objected to the introduction of any evidence tending to show that exhibit C did not represent the intended design and construction of the accident tire. Michelin’s counsel then informed the court that, “quite frankly when I made my Opening Statement and when [plaintiffs’ counsel] made his, I didn’t even know [Michelin] made them two ways.”
The following day plaintiffs argued a motion to “limit” Michelin’s proof so as to exclude any evidence that Michelin, by design, manufactured 12.00 R20 XZY tires structurally the same as the accident tire. Michelin had been prepared to show that the 12.00 R20 tire was made and sold with both belt lay ups, and that the belt lay up was unrelated to the safe operation of the tire.
Plaintiffs’ counsel argued that Michelin had judicially admitted that the 12.00 R20 tire was designed to be manufactured as exhibit C was manufactured and should be held to that admission. The trial court held Michelin had judicially admitted that exhibr it C was the same size, construction and design as the accident tire. The court refused to allow Michelin to introduce any tire represented as a “duplicate” of the accident tire if the tire offered was not constructed as was exhibit C and took the position that any attempt to do so would violate the stipulation.
*554Thereafter, Michelin filed a written motion, requesting the court to reconsider its ruling which prohibited Michelin from introducing any evidence that the design specifications of the 12.00 R20 tire allow for two different belt lay ups. Alternatively, Michelin requested leave to withdraw any alleged admission made by counsel that exhibit C and the accident tire were identical. Michelin did not concede an admission, but suggested that if any admission had been made, plaintiffs had released Michelin by making proof contrary to the stipulation, i.e., that the accident tire and exhibit C were not identical. The motion was denied by the court.
As Michelin prepared to present its casein-chief, plaintiffs’ counsel approached the bench and advised the court that he had been shown a tire which he believed was different from exhibits C and D. Michelin’s counsel explained that he proposed to use the other tire as rebuttal to Edwards’ testimony that in his opinion, Michelin did not make tires with both belt lay ups. Because Edwards had testified that if the accident tire had a nonconforming belt lay up it indicated a manufacturing defect of some kind, Michelin argued that it should be allowed to use the tire to show that “Michelin does, in fact, make a regular practice of making tires ... with the belts running parallel to the top two belts.”
The trial court again heard argument from both parties, and at this point once more explained its ruling. We quote to avoid misstatement:
“THE COURT: Defendant’s [sic] Exhibit C was represented and I have held and will hold and continue to hold that it’s a judicial admission that that [the design of Exhibit C] was the design. By ‘design,’ I mean, the design of the direction, angle, whatever, of the safety belt, I think we're talking about. That is, the direction of the bias as prepared [sic] with the other belts, and I — that’s where I’m starting from.
I’m going to hold that that was the design of the XZY tire. That was the accident tire and no other design. And a lot of it is — it all stems from the problem we’ve — we’ve protected Michelin’s rights to their trade secrets, and that’s what I want to do.
Now, — and so I don’t want any evidence or any — any hint that that XZY cut-a-way or that the accident tire was made any other way than that Exhibit C_” (Emphasis ours.)
After further argument, defendants began presenting their evidence by offering testimony from one Robert G. Dunlop. Mr. Dunlop was a member, possibly the senior member, of a consulting firm specializing in tire development problems and he had impressive credentials as an expert on the subject of tire failure. Dunlop had examined the accident tire, or pieces of the tire, in Springfield, Missouri, in 1980. Later, Dunlop had examined the carcass of the accident tire and had, on a still later occasion, been able to fit the pieces together. Mr. Dunlop’s opinion was that the shape, condition and configuration of the casing was typical of a tire of the same type which had been run with increasingly reduced air pressure until it was finally flat and had thereafter been “run flat” for some distance. Driving on a tire which has become flat produces all the “radial splits” which were observed in the accident tire. Dunlop undertook to distinguish the tire failure in question from an ordinary “blowout” thus:
“Let’s say this tire starts at a hundred twenty pounds, and I think it did, cold. And I think in the course of maybe a hundred and fifty miles after the air started to leak out, it progressively was dropping down a hundred pounds, eighty pounds, sixty pounds, fifty pounds, forty pounds. And it finally got down to the point where it wouldn’t even stay on the rim. The rim came apart it was so flat. It was probably completely flat because once it got low enough to start splitting that carcass, any remaining air is going to come out through a big hole in the tire instead of just a little hole. There’s a big hole in the tube instead of a little hole in the tube, and its going to lose all its air very quickly. And once all its air is gone, it can still run a remarkable distance but that really chews up the tire.”
*555After Dunlop’s evidence was received, it became clear that Michelin intended to introduce a tire (exhibit OA) which had been manufactured in Spain the previous week and which had the same belt lay up as the accident tire. Michelin’s apparent intention to introduce this tire again aroused plaintiffs’ belief that Michelin would not comply with the court’s earlier ruling that Michelin could not introduce any evidence showing the accident tire was designed to be made differently from exhibit C. At this point, Michelin argued:
“... If in fact there is a judicial admission of any kind, it is only that the accident tire and the specimen tire are of the same size, construction and design. And that it’s a duplicate of this tire that is in this lawsuit. That’s what’s said, Judge. There is no representation as to which is or which is not the proper design. If in fact there is a proper design rather than two....”
Michelin strenuously argued for the tire’s admission as rebuttal to plaintiffs’ expert testimony that Michelin does not make 12.-00 R20 XZY tires with different belt lay ups. The court refused admission of the tire manufactured in Spain, but agreed to receive evidence that the 12.00 R20 was made with both belt lay ups for the sole purpose of refuting plaintiffs’ testimony that no one makes tires with both belt lay ups. This on the express condition that any tire offered be “substantially different from” the accident tire and that it in no way suggest a similar or the same tire.
Thereafter Michelin proceeded with Mr. Dunlop’s testimony. Dunlop testified that he disagreed with plaintiffs’ expert witnesses that “it was a defective practice to put the protective belts parallel with the outer working belt rather than at a bias with the outer working belt.” He testified that various tire manufacturers, including Michelin, follow the practice of placing the protective belt directly over the underlying belt and running the belts parallel would not affect the “handling ability” or the safety of the tire. Dunlop also identified defendants’ exhibit X. This document, which is not before us, contained “belt wire data.” The document was a 1976 report which contained the design specifications of various steel-belted radials manufactured by Firestone, Uniroyal, Bridgestone and Dunlop. This document showed that some tires comparable to the accident or specimen tire were routinely manufactured so the outer, or “protective” belt, ran parallel to the outer working belt. Sometimes the outer two belts were laid on the bias. In Dunlop’s opinion, it was fair to say that the industry produced some steel-belted radial tires with the protective belt and the outer working belt running parallel and some tires with these two belts running on a bias. During the course of his direct examination, Dunlop was shown Michelin’s exhibit Y. This exhibit, which is not before us, was not like the one in evidence, but was “a cross section of a Michelin tube type truck tire.” On this tire the protective belt ran parallel to the outer working belt and that configuration, Dunlop explained, was common in the industry. Mr. Dunlop concluded that the accident tire failed because it had been punctured and that the belt lay up would not have affected the operation or safety of the accident tire.
Michelin made an offer of proof, out of the hearing of the jury, through Mr. Sidney Bloor, a vice-president of the Michelin Tire Corporation. Several different tires were introduced during the course of the offer. Michelin’s exhibit C, according to Bloor, was built slightly differently from the tire which was Michelin’s exhibit B (the casing of the accident tire). There were “some angular differences in the working plies and the protective belt ... [was] on a bias to the top working ply.” Bloor was also shown the tire marked as Michelin’s exhibit OA. This tire was a 12.00 R20 constructed with the parallel belt lay up found in exhibit B (the accident tire). The substance of defendants’ offer of proof was that Michelin manufactured tires with the protective belt at the same angle and “lay” as the top working belt, and also manufactured the same tire with the protective belt on a bias with the top working belt. In the United States, the two different models were used interchangeably. This offer of proof was *556refused. Michelin later presented further evidence within the hearing of the jury. We shall presently examine that evidence in some detail.
Turning abruptly to the merits of Appeal No. 14250 — Michelin’s appeal — we note that Michelin’s first two assignments of error deal with the stipulation entered into by counsel before the trial began. Michelin’s first assignment of error is that the trial court erred in holding that Michelin judicially admitted that the accident tire was intended to be manufactured in exactly the same way as exhibit C and that Michelin was erroneously precluded from introducing evidence to the contrary. It is argued that Michelin admitted only that the two exhibits (tires) were identical because the plain language of the stipulation admits of no other meaning. Further in this connection, Michelin maintains it intended only to stipulate to the foundation for the admissibility of defendants’ exhibit C as an item of demonstrative evidence. Michelin’s second assignment of error is that it should have been allowed to withdraw from its stipulation, for three reasons: a) plaintiffs disputed many of its terms; b) the stipulation was entered into by mistake, and c) plaintiffs did not prejudicially rely on the stipulation.
Michelin calls our attention to the statements of counsel made shortly before the trial began. As we have seen, plaintiffs’ counsel called upon Michelin’s attorney to stipulate that exhibit C “[was] the same size, construction, design, and that it’s a duplicate of this tire that is in this lawsuit.” Counsel for Michelin promptly answered, “Yes, it is.” Plaintiffs’ attorney further asked Michelin’s counsel if it were correct to say that exhibit C was “an exact duplicate of the make and model of [the accident] tire.” Michelin’s counsel replied “Yes.”
As Michelin argues, stipulations should be interpreted in view of the result which the parties were attempting to accomplish. Huegel v. Huegel, 329 Mo. 571, 576, 46 S.W.2d 157, 158 (banc 1932); Pierson v. Allen, 409 S.W.2d 127, 130[5] (Mo.1966); Landers v. Smith, 379 S.W.2d 884, 888 (Mo.App.1964). Otherwise put, the interpretation of a stipulation is primarily a matter of ascertaining the intent of the parties. That intent is a question of fact to be determined by the court based on the evidence before it. Stella v. DePaul Community Health Center, 487 F.Supp. 1017, 1020 (E.D.Mo.1980), aff'd, 642 F.2d 258 (8th Cir.1981).
Without resorting to technical rules of construction, the intent of plaintiffs’ counsel is obvious. Pretrial discovery in this case was begun before one experienced circuit judge who died while the litigation was in progress. The case was assigned to another judge. The record shows that the design of steel-belted radial truck tires varies, and plaintiffs’ counsel had been advised by his own expert that the design of Michelin’s 12.00 R20 truck tire might have been altered since the accident tire was purchased. Given plaintiffs’ primary theory that one or another of the radial belts had been misaligned in the process of manufacture and given the further fact that plaintiffs had twice been denied discovery of the design specifications of the 12.00 R20 XZY tire, counsel’s need for a stipulation that exhibit C was an exemplar or specimen of the tire in issue, correctly manufactured, becomes apparent. Michelin’s counsel was under no compulsion to stipulate to any fact, but he apparently saw no harm in doing so. There is no doubt that the parties intended to stipulate that exhibit C and the accident tire were similar in one or more respects, but as our tediously extensive quotation of the record shows, counsel could never agree concerning the degree of similarity to which they had stipulated.
Plaintiffs’ counsel initially called upon Michelin’s attorney to stipulate that exhibit C was the same “size, construction, design” and a duplicate of the tire in issue. Shortly thereafter, plaintiffs’ attorney asked if it was correct to say exhibit C was “an exact duplicate” of the make and model of the accident tire. On both occasions, Michelin’s attorney answered affirmatively and without qualification. Plaintiffs’ counsel, in his opening statement, carefully explained his theory, or one of his theories, of *557a manufacturing defect by referring to exhibit C. Michelin’s counsel, in his opening statement, advised the jury that “[t]his is the same kind of tire that was on the truck.” Pointing to another exhibit (a whole tire), Michelin’s counsel advised the jury that both exhibits were “like the accident tire.”
As we have seen, trial counsel for both plaintiffs and Michelin continuously argued, throughout the trial, about the meaning of the stipulation, and the trial court seems to have been of the opinion that Michelin had judicially admitted that exhibit C represented the intended design of the accident tire. On the fifth day of the trial, Michelin’s counsel suggested that after he made his opening statement he discovered that Michelin, by design, manufactured some 12.00 R20 XZY tires with both outer belts, or plies, running parallel. The trial court advised Michelin that in its opinion, Michelin had admitted that exhibit C was a “duplicate” of the accident tire.
Michelin thereafter moved the court to be allowed to withdraw from its judicial admission, without quite admitting that any judicial admission had ever been made. Michelin argued that if it had admitted that exhibit C represented the intended design of the accident tire, plaintiffs had released Michelin by making proof to the contrary. The trial court denied the motion. Counsel kept arguing about the proper construction of the stipulation. The trial court finally modified its original ruling, as we have noted.
Some of the arguments advanced by Michelin have a degree of merit. For example, plaintiffs’ counsel cannot, in our view, have intended to stipulate that the accident tire and exhibit C were literally “duplicates.” One everyday definition of the word “duplicate,” used as a noun, is “[ajnything that corresponds exactly to something else, especially an original; a double.” The American Heritage Dictionary p. 405 (1971). Comparing the synonyms “reproduction, facsimile, copy, carbon copy and transcript,” a standard dictionary states that the word “duplicate” is a double of something else. Webster’s New Dictionary of Synonyms p. 273 (1978). If exhibit C and the accident tire were “duplicates” in the sense that they were identical, then both would contain the same flaw.
The opening statement by counsel for plaintiffs shows that counsel proposed to show a difference between the accident tire and exhibit C. As Michelin argues, when a party does not rely upon the judicial admission of his adversary, but introduces evidence which has the effect of disproving his case, the party making the admission is not bound by it. Plemmons v. Pevely Dairy Co., 241 Mo.App. 659, 673, 233 S.W.2d 426, 434[3] (1950). It may be that the plaintiffs’ proof was such as to release Michelin from the stipulation. Certainly the effect of plaintiffs’ proof of a manufacturing defect ran contrary to their assertion that exhibit C and the accident tire were duplicates. It may also be true that the circumstances would have justified the trial court in relieving Michelin from its stipulation. Nevertheless, we are convinced that the merits of this appeal do not depend upon a meticulous analysis of a poorly-worded stipulation entered into in the course of a pretrial conference.
We have been over the record. It has been read. While it is far too diffuse to summarize in a few paragraphs or pages, the fact issues to be decided were relatively simple. Plaintiffs had the burden to establish: 1) a defect in the tire; 2) that the defect existed when the tire left the manufacturer and entered the stream of commerce, and 3) the defect was the proximate cause of the injuries to the plaintiffs. Moslander v. Dayton Tire and Rubber Co., 628 S.W.2d 899, 904 (Mo.App.1981); Coulter v. Michelin Tire Corp., 622 S.W.2d 421, 427[5] (Mo.App.1981), cert. denied, 456 U.S. 906, 102 S.Ct. 1752, 72 L.Ed.2d 162 (1982). Plaintiffs’ counsel undertook to show that the manufacturing defect lay in a misalignment of the outermost, or “protective,” radial belt which made the tire abnormally dangerous. The plaintiffs had evidence that one or more of the radial belts on the subject tire was misaligned and that this misalignment caused the tire to become abnormally hot in ordinary operation. Over a period of time, the radial *558belts deteriorated or “devulcanized” and eventually the tire disintegrated.
Plaintiffs’ theory of submission was in considerable part, but not entirely, dependent upon the hypothesis that the accident tire was not built as it had been designed. On cross-examination, Edwards, one of plaintiffs’ experts, was asked if the only manufacturing defect he found was two belts running on the same bias. His reply was “No.” Being asked what else could have caused the tire to fail, Edwards gave as his opinion that insufficient adhesion between the belts, coupled with a misaligned belt, could have caused the tire to fail. Edwards also testified to a number of other manufacturing errors which might have caused the accident tire to be defective. At one point, Edwards alluded to six “possibilities” as causes for failure of the accident tire.
In any event, the sufficiency of the evidence to support the verdict is not in issue on this appeal. The real question raised, with respect to the stipulation, is whether Michelin was so limited in its presentation of evidence which tended to prove there was no manufacturing defect as to be denied a fair trial. Assuming that the trial court erroneously construed the parties’ stipulation, or erroneously refused to relieve Michelin of the burden of its admission, it is nevertheless true that error without prejudice is no ground for reversal. Mo.R.Civ.P. 84.13(b); Wilcox v. St. Louis-Southwestern Railroad Company, 418 S.W.2d 15, 19-20[3] (Mo.1967); Blanford v. St. Louis Public Service Co., 266 S.W.2d 718, 722-723 (Mo.1954); Roque v. Kaw Transport Co., 697 S.W.2d 254, 257[6] (Mo.App.1985).
Michelin’s first claim of error, reduced to its essence, is that the trial court erroneously construed the stipulation as a judicial admission by Michelin that exhibit C represented the intended design of the 12.00 R20 XZY truck tire. This error, it is argued, prevented a showing that this particular tire was designed to utilize either of two belt lay ups — one in which the outer or protective belt ran parallel to the outer working belt, another in which the protective belt ran at an angle or bias to the outer working belt. Michelin concedes that the trial court did permit it to prove that it and other tire manufacturers make tires with both belt lay ups, and that this is sometimes done for the same tire. Michelin complains, all the same, that it was not permitted to show specifically that the 12.-00 R20 XZY tire was manufactured with the protective belt and the outer working belt running parallel. We have already set forth Mr. Dunlop’s testimony. In our view, his testimony tends strongly to establish the very fact Michelin says it was not permitted to establish, viz., that the 12.00 R20 tire was made with several belt lay ups.
It is true that Michelin’s offer of proof through Mr. Bloor was rejected, but the following day Bloor testified in the presence of the jury:
“Q. Now, speaking of the radial belt, the protective belt, I understood Mr. Edwards to be of the opinion that it was improperly put on because it ran parallel with the outer working belt rather than on a bias with it. Do you remember that testimony?
A. I remember that testimony, yes.
Q. You made — do you make various kinds of Michelin tires each of those two ways?
A. All types and sizes. We make the protective belts either one way or the other way according to needs.
Q. And in some models, do you make them both ways?
A. We do.
Q. Without any different designation on the sidewall?
A. Other than the individual serial number, no, there’s no designation.
Q. So if it does happen to be the type of tire that you make in bulk configurations, would the dealer or the customer either one be able to look at the sidewall and see whether the belts ran on the bias or parallel?
A. No. No, he couldn’t tell.
Q. You have had experience with both types of layups?
A. That’s correct.
*559Q. Have you experienced any different running characteristics or wearing characteristics or difficulties of any kind as a result of the belts running parallel, the protective belt and the outer working belt?
A. No. The only difference for changing in the belt angle, let’s say from the same angle the top working ply to a bias, any real difference is the ones which run at the same angle resist shocks: right? Because by running them at the same angle they are more flexible. So you know the theory if you punch a pillow, right? If you punch a pillow the pillow gives when your fist goes into the pillow because it’s flexible, it’s absorbing or enveloping, there’s no damage done. That’s a little bit like a tire. You can make it so that it will absorb major shocks your tire will resist them better. So the ones which are made parallel are done for countries where we know there’s a lot of problems with major shock damage to the tread.
That tire also runs a little cooler than the one with the bias. And we use the bias protective belt in countries where it’s wet and we have problems with cuts to the tread. I think you all know rubber cuts very easily when it’s wet. So we put the belt on the bias to resist the cuts. Then in the States we use either indiscriminately.” (Emphasis ours.)
Further, Bloor testified:
“Q. ... Approximately when did you start manufacturing the model tires involved in this accident?
A. In the early 70’s.
Q. And is it still being manufactured?
A. Yes, it is.
Q. And has been shipped to various countries throughout the world?
A. Yes, it’s world wide distribution.
Q. Including of course a number in this country?
A. Yes. It’s not a high volume tire in the States because it’s [sic] capacity is such that you can exceed the legal limits in the States very easily. So it doesn’t have a high market in the States.
Q. It’s a more popular tire in Europe because they have higher weight limits, isn’t that true?
A. Yes, they do.
Q. During all of those years you have been making this tire, is this the first lawsuit you have ever had involving the tire?
A. That’s correct.
Q. Is this the first even claim of any kind you’ve had involving this tire?
A. That’s correct.
Q. And the first time anybody has ever alleged that there was something wrong with the way this model tire was designed or manufactured?
A. It was — it is a tire without history as we say. It doesn’t have a history of claims at all.”
Our conclusion with respect to the first two assignments of error — which deal with the stipulation — is that Michelin has failed to show that it was prejudiced by the trial court’s construction of the parties’ stipulation, even though that construction may have been erroneous. The plaintiffs’ theory of liability depended in part upon a showing that the accident tire was unreasonably dangerous because it was not manufactured as it had been designed.2 The trial court held, in effect, that Michelin had stipulated that exhibit C — the cutaway section of a tire produced by Michelin at the commencement of the trial — represented the intended design of the accident tire, although, as we have illustrated, the meaning of the parties’ stipulation was the subject of daily argument during the trial. Despite the trial court’s ruling, Michelin introduced evidence that it made “all types and sizes” of tire with “the protective belt either one way or the other according to needs.” What we intend to hold is that every fact and circumstance which Michelin says it should have been permitted to show was actually shown by it at the trial. Having presented substantially all the evidence *560it desired to present despite the trial court’s ruling, Michelin is in no position to complain. The accident sued upon in this case occurred in 1978. The cause finally came to trial almost seven (7) years later. A retrial of the case would be very difficult because of the lapse of time, if for no other reason. In such circumstances, this court should be satisfied that the error, if any, committed by the trial court materially affected the merits of the action. With respect to Michelin’s assignments of error numbers 1 and 2, we find no such error.
Michelin further assigns error to the reception of plaintiffs’ exhibit 88, which was a statistical report prepared for the United States Department of Transportation and which was titled “1976-1978 Analysis of Motor Carrier Accidents Involving Vehicle Defects or Mechanical Failure.” The document was offered as proof that a defectively manufactured tire “creates an unreasonable danger.” Counsel for defendant Truxan Parts, Inc., objected to the reception of the document on the grounds that it was “irrelevant and immaterial to any issues in this case.” It was further objected to as being “absolute hearsay.” Michelin now objects that the document was a) wholly irrelevant and immaterial to any issue in the case; b) that there was “no foundation showing that the accidents upon which the document was based were similar to the accidents in question,” and c) the document constituted hearsay. Plaintiffs argue that the document was properly admitted as a basis for Edwards’ expert opinion, citing Siebern v. Missouri-Illinois Tractor & Equipment Co., 711 S.W.2d 935, 940[6] (Mo.App.1986).
On occasion, our courts have approved the admission of properly conducted statistical surveys for one purpose or another. In Liberty Financial Management Corporation v. Beneficial Data Processing Corporation, 670 S.W.2d 40, 54-55[16][17][18] (Mo.App.1984), a statistical survey prepared by an expert witness was held to be admissible even though it was hearsay. In Siebern v. Missouri-Illinois Tractor & Equipment Co., 711 S.W.2d at 940, statistical surveys were held properly admitted to form a basis for an expert’s opinion. By contrast, in Kontz v. K-Mart Corporation, 712 F.2d 1302, 1304[2, 3] (8th Cir.1983), plaintiff brought a strict liability action alleging defective design of a folding lounge chair. Defendant had a verdict. Plaintiff appealed, contending, among other things, that certain statistical studies had been improperly excluded. The United States Court of Appeals for the Eighth Circuit held there was no abuse of discretion in excluding the results of a Consumer Products Safety Commission Study purportedly revealing approximately 8,000 injuries a year from the use of folding or beach chairs, especially in the absence of proof of similarity of circumstances.
Aside from the report prepared for the Department of Transportation, plaintiffs had competent testimony from expert witnesses that the accident tire was unreasonably dangerous. We believe it may be said upon the record before us that admission of the D.O.T. study was not prejudicial because the admission of improper evidence is harmless when the fact thereby sought to be shown is otherwise fully and properly established. Boten v. Brecklein, 452 S.W.2d 86, 95—96[20] (Mo.1970).
Michelin’s final point is that the closing argument made by plaintiffs’ counsel was so grossly improper and inflammatory that a mistrial should have been declared. The argument complained of ran to this effect: Michelin’s counsel had argued that Michelin had never had a claim on a 12.00 R20 XZY tire. If, counsel had argued, Michelin ever had a valid claim, it was paid. Counsel for plaintiffs argued that such assertions were ludicrous, stating:
“You read the papers. You’ve read about Union Carbide and Ford and General Motors and the Hyatt Regency and all kinds of people, and maybe if you have read them carefully enough, you may have read something about Michelin.”
Michelin characterizes the comparison of this casualty to the Union Carbide chemical release in India, the collapse of the skywalks at the Hyatt Regency hotel in Kansas City and other disasters as being *561so inflammatory that the trial court should have declared a mistrial. The comparison may be inapt, but of course the declaration of a mistrial is the most drastic remedy for a trial error and should be granted only when the error is so grievous that the prejudicial effect thereof cannot be otherwise removed. Brownridge v. Leslie, 450 S.W.2d 214, 216 (Mo.1970); Lewis v. Hubert, 532 S.W.2d 860, 866[11] (Mo.App.1975). In any event, no objection was interposed by defense counsel when the argument was made. A motion for mistrial was made, but it was made after the arguments had been completed and the jury had retired to deliberate. It has repeatedly been held that when a party does not object to improper argument at the time it is made, he will not be heard to object on appeal. Sandy Ford Ranch, Inc. v. Dill, 449 S.W.2d 1, 7 (Mo.1970). If the argument was overdrawn, it was not such as to deprive Michelin of a fair trial, and we find no merit in the point.
Having considered defendants Michelin’s assignments of error at length in Appeal No. 14250, and having found no error materially affecting the merits of the action, we affirm the judgments in favor of plaintiffs Wayne Vaughn, Karen Vaughn and Norma Young and against defendants Michelin Tire Corporation and S.A.F.E. DeNeumati-cos Michelin.
Appeal No. 14248 is plaintiffs’ appeal. Two assignments of error are presented. Plaintiffs first contend that the trial court erred in denying their motion for directed verdict made at the close of all the evidence and renewed in their post-trial motion for judgment n.o.v. Plaintiffs sought to hold defendant Truxan Parts, Inc., liable as a seller. It is, of course, true that the theory of strict tort liability holds “one who sells” a defective product unreasonably dangerous to the user liable for resulting injury. Zafft v. Eli Lilly & Co., 676 S.W.2d 241, 244 (Mo. banc 1984). In such cases, proof of sale of the instrumentality by the defendant to the plaintiff is an essential element of a plaintiff’s case. See M.A.I. 25.04; Means v. Sears, Roebuck & Co., 550 S.W.2d 780, 786—87[1] (Mo. banc 1977); Keener v. Dayton Electric Manufacturing Company, 445 S.W.2d 362, 366[7] (Mo.1969). While the plaintiffs moved the court for a directed verdict on all issues against all the defendants, they assign error to the denial of that motion only with respect to the issue whether Truxan Parts, Inc., sold the accident tire to Leo and Norma Young.
Our courts have recognized the power and authority of a trial court to direct a verdict for the plaintiff either at the close of the evidence or in response to after-trial motions in situations where there is no real factual dispute and when the facts establish as a matter of law that plaintiff has a right to a verdict. Alaska Federal Savings & Loan Ass’n v. Hoffman, 485 S.W.2d 118, 120[1] (Mo.App.1972). Nevertheless, our Supreme Court has been at pains to limit a plaintiff or counterclaimant’s right to a directed verdict. Only in exceptional circumstances, as in the case where the defendant in his pleadings or by his counsel in open court admits or by his own evidence establishes plaintiff’s claim, or where there is no real dispute of the basic facts supported by uncontradicted testimony essential to a claim, have the courts indicated that a verdict may be directed for the party having the burden of proof. Zagarri v. Nichols, 429 S.W.2d 758, 760 (Mo.1968); Parker v. Pine, 617 S.W.2d 536, 541 (Mo.App.1981). It has been held, and we believe, that a cause may be taken from the jury on a plaintiff’s motion for directed verdict only when there are no genuine fact issues to be submitted to the jury. Parker v. Pine, 617 S.W.2d at 541; Twellman v. Lindell Trust Co., 534 S.W.2d 83, 87 (Mo.App.1976), 93 A.L.R.3d 943 (1976).
Leo Young, deceased at trial time, was a stockholder in an organization known as Springfield Truckers, Inc. Springfield Truckers had a contract to haul coal from Oklahoma to Springfield, Missouri. The coal was used by the Springfield Board of Public Utilities. Plaintiff Vaughn’s truck was, in fact, loaded with coal, as was the pup trailer which Vaughn was pulling. The truck on which the accident tire was mounted was owned by Leo and his wife, plaintiff Norma Young. It had been leased to Springfield Truckers, Inc.
*562Stated most favorably to defendant Truxan Parts, Inc., the evidence that defendant Truxan Parts sold or furnished the defective instrumentality came primarily from Norma Young and Wayne Vaughn. Norma Young testified that the corporate records of Springfield Truckers, Inc., had been destroyed in a fire. Mrs. Young was not directly involved in the management of Springfield Truckers. Mrs. Young did know that Leo usually bought tires from defendant Truxan. She knew of no other location where Leo purchased tires within five years preceding the accident. Mrs. Young did know, however, that her husband had a charge account at another Michelin outlet in Joplin, Missouri, and yet another open credit account at a truck stop in Vinita, Oklahoma, which repaired and replaced tires. Other than her testimony concerning her husband’s customary practice and certain documents, plaintiff Norma Young had no direct knowledge that the accident tire was furnished by defendant Truxan Parts, Inc. Mrs. Young identified exhibits showing Leo's purchases from defendant Truxan, but these exhibits (invoices) did not identify the tires purchased by their serial numbers, so it was impossible to say that a particular purchase shown by the invoices represented the accident tire. Various other records were introduced indicating that the accident tire might have been purchased from defendant Truxan Parts, but no useful purpose would be served by reciting it all. Although a jury might have found that defendant Truxan Parts, Inc., sold the accident tire to Leo Young, it cannot be said that plaintiffs so clearly established that fact that there was no genuine fact issue to submit to the jury. The court did not err in refusing plaintiffs’ motions for a directed verdict.
The final point to be considered in Appeal No. 14248 is plaintiffs’ contention that prejudgment interest should have been awarded on plaintiffs’ claims. Plaintiffs cite Catron v. Columbia Mut. Ins. Co., 723 S.W.2d 5 (Mo. banc 1987), and Twin River Const. Co. v. Public Water Dist., 653 S.W.2d 682 (Mo.App.1983), among many other precedents, and argue essentially, that plaintiffs Vaughn are entitled to prejudgment interest because “equity demands that prejudgment interest be given in personal injury actions.”
We decline to enter into a theoretical discussion of the propriety of awarding prejudgment interest on unliquidated claims or demands. In Catron, it was recognized that there are several exceptions to the general rule that interest is not recoverable on an unliquidated demand. These exceptions include: a) situations in which a liquidated claim is countered with an unliq-uidated counterclaim, set-off or plea in re-coupment; b) cases in which the only contested issue is liability and the defendant does not dispute the dollar amount of damages; c) those cases in which interest is held to be a part of a taking for which compensation should be provided in condemnation cases, and d) cases in which, although the claim is in a sense unliqui-dated, the amount of damages is readily ascertainable by mathematical computation or determinable according to a recognized standard. Consistent with the controlling precedents, it cannot be said that a defendant “benefits unfairly” from protracted litigation when he truly cannot be expected to know how much he owes and thus cannot be in default. Catron, 723 S.W.2d 5, 8 (Robertson, J., concurring). This case bears no resemblance to those cases in which our courts have allowed prejudgment interest, and we decline to order its allowance here.
The judgments appealed from are not erroneous in any respect briefed or argued in this court. Accordingly, the judgments appealed from should be and are, in all respects, affirmed.
MAUS and HOLSTEIN, JJ., concur and file concurring opinions. GREENE, J., concurs. CROW, C.J., concurs in principal opinion and in concurring opinion of MAUS, J. FLANIGAN, J., dissents and files dissenting opinion. PREWITT, J., not participating.. Mr. Thomas G. Strong served as counsel for the plaintiffs; Mr. Paul L. Bradshaw, now deceased, was counsel for defendants Michelin.
. For a discussion of this order of defect in a products liability case, see W. Prosser and W. Keeton, Torts § 99, pp. 695-96 (5th ed. 1984).