The ultimate dispute here is whether plaintiff's soybean storage warehouse was damaged by a bean dust explosion (in which case defendant would be liable for the damage under a contract of insurance), or by a structural failure resulting from inadequate design, overloading, or a combination of both (risks not covered by the policy). Invoking the diversity jurisdiction of the United States District Court for the Northern District of Mississippi, plaintiff sought a judicial determination of this controversy. A jury resolved the dispute in plaintiff's favor. The trial judge, concluding that the evidence presented was inadequate to support the jury's verdict, granted the defendant's motion for judgment non obstante veredicto. Plaintiff appeals, contending that its right to a jury determination of disputed fact issues, secured by the seventh amendment to the Constitution of the United States, has been abridged by the trial judge's ruling.

The first point urged by appellant is that in determining the sufficiency of the evidence to raise a question of fact for the jury in diversity cases, federal courts must apply a federal test. The Supreme Court has yet to settle the conflict which exists among the courts of appeals on this issue.1

However, the Court's decisions in Byrd v. Blue Ridge Rural Electric Cooperative, 356 U.S. 525, 78 S. Ct. 893, 2 L. Ed. 2d 953 (1958), and Simler v. Conner, 372 U.S. 221, 83 S. Ct. 609, 9 L. Ed. 2d 691 (1963), strongly suggest that the question will eventually be resolved in favor of the federal test. In Byrd, the issue was whether a federal court sitting in diversity was bound under the rule of Erie R. Co. v. Tompkins, 304 U.S. 64, 58 S. Ct. 817, 82 L. Ed. 1188, 114 A.L.R. 1487 (1938), to follow a decision by the Supreme Court of South Carolina which, in state court, would have made the trial judge the trier of fact on the question whether Byrd was covered by workmen's compensation. The Court refused to permit state law to divest a federal jury of its normal function. In the course of the opinion, it said:

"Thus the inquiry here is whether the federal policy favoring jury decisions of disputed fact questions should yield to the state rule in the interest of furthering the objective that the litigation should not come out one way in the federal court and another way in the state court.

"We think that in the circumstances of this case the federal court should not follow the state rule. It cannot be gainsaid that there is a strong federal policy against allowing state rules to disrupt the judge-jury relationship in the federal courts." Byrd, supra, 365 U.S. at 538, 78 S. Ct. at 901. (Emphasis added.)

To permit state law to dictate when a federal trial judge must take questions of fact from a jury by means of a directed verdict or a judgment n. o. v. seems hardly less disruptive of the federal judge — jury relationship than assigning the task of resolving factual disputes to the trial judge initially.

The question posed in Simler, another diversity case, was whether federal or state law would determine if a controversy was equitable in nature and triable to the court, or legal and triable to a jury. The Court said:

"[T]he right to a jury trial in the federal courts is to be determined as a matter of federal law in diversity as well as other actions. * * * Only through a holding that the jury-trial right is to be determined according to federal law can the uniformity in its exercise which is demanded by the Seventh Amendment be achieved." Id, 372 U.S. at 222, 83 S.Ct. at 610.

If the seventh amendment requires uniformity in the exercise of the jury trial right in the federal courts, surely that subsumes uniformity in the exercise of the power to direct a verdict or grant a judgment n. o. v.

In any event, this court's position on the question is clear. In Revlon, Inc. v. Buchanan, 271 F.2d 795 (5 Cir. 1959), for example, we said:

"The quantity and quality of proof necessary to make out a case for submission to a jury in a federal court are determined by the Seventh Amendment to the Constitution of the United States, the Federal Rules of Civil Procedure and the decisions of the courts of the United States." Id. at 800.2

Careful examination of the principles involved has given us no cause to doubt the correctness of that position. We adhere to it.

One further preliminary issue remains. Appellant contends that in formulating the federal standard to be applied in determining whether the evidence in a case is sufficient to raise a question for the jury, consideration must be given to the decision in Lavender v. Kurn, 327 U.S. 645, 66 S. Ct. 740 (1946), in which the Supreme Court circumscribed the power of federal judges to dispense with a jury in resolving disputed questions of fact.

A. Lavender was an action commenced under the Federal Employers Liability Act to recover damages for the death of Haney, a switchtender employed by the Illinois Central Railroad Company. Haney was found unconscious on the ground north of the railroad track, near his switch. He had been struck in the back of the head, causing a fractured skull from which he died. Lavender, the administrator of Haney's estate brought suit against the Illinois Central and the trustees of the St. Louis-San Francisco Railway Company (Frisco). The theory asserted by the plaintiff was that Haney had been struck by the curled end or top of a mail hook extended from the outside of the mail car of a Frisco train which was backing into the station at the time along the track near which Haney was found. There were no eyewitnesses to the occurrence. It was not shown that the hook was in fact extended at the time and place of Haney's injury, and it was demonstrated that, except for a topographic elevation at one place along the track, the hook, even if extended, was too high to have struck Haney standing on the ground. The position of Haney's body when discovered appeared to contradict plaintiff's theory of causation. The defense contended, first, that plaintiff's theory was practically a physical impossibility, and second, that Haney was probably murdered. Haney's wallet, containing no money, was found six days after and about a block from the place of the injury. The examining doctor who testified that the fatal blow might have come from an object attached to a moving train also admitted that the fracture might have resulted from a blow from a pipe or club wielded by an individual. It appeared that many hoboes and tramps frequented the area at night in order to get rides on freight trains, and it was shown that Haney carried a pistol to protect himself, which was found loose under his body by those who discovered him. The jury returned a verdict for the plaintiff, which was overturned by the Supreme Court of Missouri on appeal. On writ of certiorari, the Supreme Court of the United States reversed, holding that the case was properly submitted to the jury, and that it could reasonably be inferred from the plaintiff's evidence that death was caused by the mail hook.3

In its opinion, the Court elaborated thus upon the standard to be employed in determining whether the evidence in a case raises a question for the jury:

"It is true that there is evidence tending to show that it was physically and mathematically impossible for the hook to strike Haney. And there are facts from which it might reasonably be inferred that Haney was murdered. But such evidence has become irrelevant upon appeal, there being a reasonable basis in the record for inferring that the hook struck Haney. * * * It is no answer to say that the jury's verdict involved speculation and conjecture. Whenever facts are in dispute or the evidence is such that fairminded men may draw different inferences, a measure of speculation and conjecture is required on the part of those whose duty it is to settle the dispute by choosing what seems to them to be the most reasonable inference. Only when there is a complete absence of probative facts to support the conclusion reached does a re-reversible error appear. But where, as here, there is an evidentiary basis for the jury's verdict, the jury is free to discard or disbelieve whatever facts are inconsistent with its conclusion. And the appellate court's function is exhausted when that evidentiary basis becomes apparent, it being immaterial that the court might draw a contrary inference or feel that another conclusion is more reasonable." Id. at 652-653, 66 S.Ct. at 743-744. (Emphasis added.)

We find it impossible to reconcile this rule with the language employed by the Court in Pennsylvania R. R. v. Chamberlin, 288 U.S. 333, 53 S. Ct. 391, 77 L. Ed. 819 (1932), and relied upon by the district court in this case:

"We, therefore, have a case belonging to that class of cases where proven facts give equal support to each of two inconsistent inferences; in which event, neither of them being established, judgment, as a matter of law, must go against the party upon whom rests the necessity of sustaining one of these inferences as against the other, before he is entitled to recover." Id. at 339,4 53 S.Ct. at 393.

The consequence of heeding the teachings of Lavender was well expressed, we think, by Judge Holtzoff in Preston v. Safeway Stores, Inc., 163 F. Supp. 749 (D.D.C.1958):

"The conclusion seems inescapable that the decision in Lavender v. Kurn must be deemed to constitute an abandonment of the earlier doctrine that if the evidence is capable of either of two inferences, it cannot be deemed to support either. The case substitutes the principle that in such an event, it is for the jury to determine which inference to deduce and that the jury has a right to draw either one. The prior cases * * * [such as Chamberlin] must be deemed to have been overruled sub silentio." Id. at 752-753.

We are persuaded that this view is correct. If proven facts do give support to the inference necessary to sustain a plaintiff's case, then, under the rule of Lavender, it is immaterial that they give equal support to a contrary inference. It is only when there is a complete absence of probative facts to support the conclusion reached that the jury's judgment may be ignored.

B. It has been suggested that Lavender and cases like it, arising under the Federal Employers Liability Act, are sui generis, and ought therefore to be ignored in fashioning a general rule for testing the sufficiency of evidence to raise a question for the jury.5 We are unable to agree with this proposition.

It is true, as the Court pointed out in Rogers v. Missouri Pacific R. R., 352 U.S. 500, 77 S. Ct. 443, 1 L. Ed. 2d 493 (1957), that an action under FELA is a "statutory negligence action * * * significantly different from the ordinary common-law negligence action." Id. at 509-510, 77 S.Ct. at 450. But we think the difference lies in the change in the standard of liability wrought by the Act — (from the common law concepts of proximate causation to whether the employer's negligence played any part, however small, in the injury complained of) — and not in any alteration of the jury's historic function.

The argument that the tests employed by the Supreme Court to determine the existence of a jury question vel non in FELA cases can be discounted on the ground that the Court is merely complying with a special congressional mandate to broaden the jury's powers in such cases is put to rest by the following passage from a footnote to the Court's opinion in the Rogers case:

"The inclusion in * * * [the 1908 revision of the Federal Employers Liability Act] of * * * [a] provision, `All questions of fact relating to negligence shall be for the jury to determine,' was proposed but not adopted. The view prevailed that this would be surplusage in light of the Seventh Amendment embodying the common-law tradition that fact questions were for the jury." Rogers v. Missouri Pacific R. R., supra, at 508, n. 18, 77 S. Ct. at 449. (Emphasis added.)

We are unwilling to assume that the Supreme Court, absent a statutory command, would undertake to rewrite the rule to be filled by the jury in any class of actions. In our view, FELA cases are distinguishable not because a lesser quality of evidence of fault will suffice to provide a basis for a jury's verdict, but rather because evidence of a lesser degree of fault will entitle the employee to recover.

It is the seventh amendment which governs the role of the jury in the courts of the United States, and we perceive no rational basis for the proposition that it means one thing in FELA cases and quite another in other actions.6 That the Supreme Court shares this view is indicated, we think, in the following excerpt from Atlantic & Gulf Stevedores v. Ellerman Lines, 369 U.S. 355, 82 S. Ct. 780, 7 L. Ed. 2d 798 (1962):

"This suit being in the federal courts by reason of diversity of citizenship carried with it, of course, the right to trial by jury. As in cases under the Jones Act * * * and under the Federal Employers' Liability Act * * * trial by jury is part of the remedy. Thus the provisions of the Seventh Amendment, noted above, are brought into play. * * * As we recently stated in another diversity case, it is the Seventh Amendment that fashions `the federal policy favoring jury decisions of disputed fact questions.'" Id. at 360, 82 S.Ct. at 784. (Emphasis added.)

On this analysis, we are satisfied that the principles delimiting the power of judges to grant judgment notwithstanding a jury's verdict announced by the Court in Lavender v. Kurn are fully applicable to the case before us.7

Having concluded that a uniform federal standard, materially influenced by the principles announced in Lavender v. Kurn and like cases, governs the detertermination of the sufficiency of the evidence to present a question for the jury in diversity cases, we turn now to an examination of the evidence presented in this case, and the bases upon which the trial court found that evidence insufficient to support the verdict which the jury rendered for the plaintiff. Of course, we deal here only with those portions of the evidence (and the inferences which the jury might draw therefrom) which lend support to plaintiff's theory. This follows necessarily from the principle that "the jury is free to discard or disbelieve whatever facts are inconsistent with its conclusion," unless there is a complete absence of probative facts to support that conclusion. Lavender v. Kurn, supra, 327 U.S. at 653, 66 S. Ct. at 744.

The damaged structure was a large beanhouse, erected by Muscogee Iron Works in 1951 and used by Planters for the storage of soybeans from that time until the occurrence of the event in question here on November 29, 1962. The basic design of the building and the technique employed in its loading, both of significance to the plaintiff's contentions, are described in the margin.8

The theory advanced by appellant was that a cloud of explosive bean dust near the roof in the center of the building, relatively small in comparison to the total volume of air space in the structure was ignited by a spark produced when a piece of "tramp" metal or rock falling from the overhead conveyor struck a spout baffle, resulting in an explosion which caused a momentary uplift of the roof and consequent damage to the structural members of the building when the roof settled.

The evidence adduced by appellant in support of this theory consisted of (1) the testimony of several of its employees who were in the vicinity of the beanhouse at the time of the event regarding their sensory impressions of the occurrence, (2) evidence pertaining to the position and characteristics of the remains of the structure, (3) the testimony of a chemist with respect to the results of his microscopic examination of dust samples taken from the damaged and undamaged sections of the building, (4) the opinion testimony of an expert witness regarding the cause of the collapse, and (5) evidence of extra-judicial statements made by an employee of the defendant, offered as an admission that the collapse was caused by an explosion.

A. The testimony of the employees in the vicinity of the beanhouse at the time of the occurrence was summarized by the trial court as follows:

"The lay witnesses, i. e., the eight employees, were on the grounds at the time of the collapse of the beanhouse. Three were remotely placed, one of whom saw and heard nothing until the incident was over and two of whom saw nothing at the time but heard a noise. One of these described the noise as a loud report like an explosion and the other said that it sounded like a cannon firing. The remaining five witnesses were in the vicinity of the beanhouse and they variously described the noise as being similar to a sonic boom, to the coupling of two box-cars and to dynamite when being used for blasting stumps. One of them said it was an `awful noise' and another described it as a `tremendous noise.'

"Of these five witnesses who were in a position to see, one saw dust billowing in the air and saw the building shudder. Another saw a ring of dust or smoke and a piece of sheet metal and other debris rising in the air. Two men who were on the bean dump near the south end of the building saw the south wall swell outward and a piece of sheet metal come off the south end of the monitor before they were engulfed in a cloud of bean dust.

"The fifth witness was inside the beanhouse checking the temperature of the beans. He and another employee of plaintiff (who was not called to testify) were standing on the bean pile about twenty feet from the south end of the building. A repetitive noise `like a rock or something going in the conveyor (the overhead auger)' attracted his attention to the south end of the building at the monitor area. There he saw a quick flash like a match being lit and then quickly blown out, and at that point, `it got dusty.' The substance of his testimony in this last respect is that, while it was always dusty when beans were being dropped into the beanhouse, the dust content of the atmosphere drastically increased after he saw the match-like flash of fire, so that it was almost impossible to see. Because of the dust, he and his companion could see only one way to escape, and that was through a new hole which had opened up where the conveyor entered the beanhouse. They climbed up and out, and reached the ground by climbing down the elevator. The witness was very frightened and he attributed this to the `awful noise' that he heard. Neither man was injured, and the witness was unaware of having experienced any pressure or burning while inside the beanhouse, nor did he smell any unusual odors. He was unaware that the building had partially collapsed until after he was outside."

In addition to the observations summarized above, one of the witnesses, who stated that he was standing facing the building some ninety feet away when the noise generated by the event attracted his attention, testified that the "whole top [of the beanhouse] raised up." Reference to the outline of plaintiff's theory sketched above shows the critical importance of this testimony. Another witness testified that he noticed a distinct "burnt odor" in the vicinity after the occurrence.

Damage to the beanhouse was concentrated near the center of the structure. There, the sidewalls had fallen outward, allowing beans to flow out onto the ground, and the monitor had sagged down with the roof, which in places rested on the bean pile. Some windows and ventilators were broken. The northern third of the building was virtually unharmed. The south end had leaned inward and had pulled away from the bridge supporting the conveyor between the elevator and the beanhouse.

One of appellant's expert witnesses, a licensed mechanical engineer, testified on the basis of his inspection of the remains of the structure, that the monitor section and conveyor had settled down approximately on the center line of the building, and had not shifted appreciably to either side.

Photographic exhibits were introduced in support of this testimony. The witness then indicated that, in his opinion, failure of a structural member resulting from excessive pressure against the sidewalls due to overloading (the explanation of the collapse offered by defendant) would quite probably have pulled the structure to one side or the other.

Plaintiff's evidence also included the testimony of a chemist who conducted microscopic examinations of two samples of bean dust, one from the damaged portion of the beanhouse and the other from the undamaged northern section, for the purpose of ascertaining the presence of carbon in the dust samples. The sample from the damaged section was shown to have contained carbon particles amounting to two percent by weight, while no carbon particles were detected in the other sample. The witness testified that carbon is normally the product of combustion, and that a dust explosion is a form of combustion.

B. The trial court's evaluation of these first three categories of evidence was premised upon the supposed rule9 of Chamberlin, that where proven facts give equal support to each of two inconsistent inferences, the jury cannot be permitted to draw either because its selection could be based on nothing more than conjecture. In its opinion, the trial court stated:

"The testimony of the eight employees is at best equivocal insofar as it is relied upon to prove an explosion. In sum, they testified to hearing certain sounds and seeing certain events which are all consistent both with plaintiff's theory of an explosion and the defendant's theory of structural failure. * * * Their testimony and the inferences to be drawn from it are equally consistent with both theories and probative of neither. * * *

"The evidence which describes the position and characteristics of the remains of the structure after the event falls in the same category. * * * The same may be said of the testimony of the chemist who found carbon in one of two dust samples. * * * Carbon itself is a basic element and the presence of varied kinds of foreign matter in the beans makes the presence of one more kind equivocal on the present issue."10

We think that the Supreme Court has eschewed this approach. As we read Lavender, it is immaterial that evidence may equally support an inconsistent inference, if in fact fairminded men might draw from that evidence the inference sought by the party against whom a peremptory instruction is sought. In such a case, it is the quintessential function of the jury to "settle the dispute by choosing what seems to them to be the most reasonable inference." Lavender, supra, 327 U.S. at 653, 66 S.Ct. at 744. See Tennant v. Peoria & P. V. Ry., 321 U.S. 29, 35, 64 S. Ct. 409, 88 L. Ed. 520 (1944). The question, then, is whether this evidence, together with any other supporting the plaintiff's position, constitutes a basis from which the jury might with reason have inferred that an explosion did occur, and it is no answer to that question to say that a conflicting inference might with equal probability have been drawn therefrom.

Viewed in this light, we think that the evidence, dismissed as "equivocal," cannot be so lightly disregarded. As the trial court noted, "There were no carefully placed instruments to sense and record heat, pressure and light phenomena so as to more or less conclusively demonstrate whether an explosion had occurred. Instead, plaintiff had to rely on evidence of the attendant circumstances and the results from which the ultimate fact might be inferred."

One of those results from which the ultimate fact might be inferred was the sound produced by the occurrence. That sound was variously described by witnesses as "a loud report like an explosion," "the sound of a cannon or five-inch naval gun being fired," "dynamite blasting stumps," and "a jet aircraft breaking the sound barrier." Bearing in mind the necessity for plaintiff's reliance upon inferences, we cannot say that this testimony is incapable of supporting a rational inference favorable to plaintiff in the minds of fairminded jurors. Moreover, we do not share the views of the trial court in concluding that the testimony regarding the visual observations of the employees is "all consistent both with plaintiff's theory of an explosion and defendant's theory of structural failure." This conclusion fails, we think, to take into account the observations of the witnesses who testified that the roof of the beanhouse was raised momentarily by the occurrence, and that a definite odor of burning existed in the vicinity immediately thereafter. While this testimony is patently consistent with plaintiff's theory of an explosion, it is somewhat difficult to reconcile with defendant's theory of structural failure.

We are likewise unable to agree that the evidence concerning the position of the remains is "probative of nothing." There was evidence that the roof and monitor structure did not shift appreciably to either side during the collapse. Appellant's expert witnesses expressed an opinion that had the collapse resulted from structural failure precipitated by overloading, the roof and monitor would in all probability have been pulled toward the side where the failure occurred, except in the unlikely event of simultaneous and equal failure on both sides. The question, once again, is whether fairminded jurors might draw from this evidence the inference sought by the plaintiff — not whether, in the minds of judges, the evidence might as easily be harmonized with the defendant's theory. Clearly, we think, an affirmative answer is required.

Although we think the question a closer one, we reach a like conclusion with regard to the evidence of the presence of carbon in a dust sample taken from the building after the collapse. Plaintiff's chemist testified that a bean dust explosion is a form of combustion, that carbon is normally produced by combustion, and that microscopic examination revealed particles of carbon in a dust sample taken from the damaged section of the beanhouse, while a similar sample taken from the undamaged north end of the structure was found to be free of carbon. The inference suggested by the plaintiff was, of course, that the carbon particles found were produced by the claimed explosion.

Defendant's chemist testified that his analysis of companion samples revealed only a few bits of partially burned Kleenex in the sample taken from the damaged part of the building, which, in his opinion, had no bearing on a dust explosion. But the jury might have experienced some difficulty in reconciling that testimony with the following inscription placed on the container of that sample (which was introduced in evidence) by the defendant's investigating agent MacCarmack: "Dust from between rod latticed catwalk in bottom of collapsed monitor taken from end of rods between inside angle at walk. Note carbon scale deposits and smutty dust. Exhibit A. 12/4/62. E. L. MacC."

That carbon was found in the one sample where plaintiff's theory would lead one to expect its presence and was not found in the other could be, of course, purely fortuitous. But we do not think that the jury was required to regard it so, in the circumstances present here.

C. The principal basis for the trial court's grant of defendant's motion for judgment notwithstanding the jury's verdict was its conclusion that the explosion theory postulated by plaintiff is in irreconcilable conflict with an established principle of physical law. We find no such clear and unambiguous proof of any law of physics that permits this disposition of the case, however unlikely is the plaintiff's explanation of the occurrence.

The parties agree that the presence of three factors or conditions is essential to the occurrence of a bean dust explosion. There must be (1) a combustible mixture of bean dust and air and (2) a source of ignition, and (3) the forces generated by the combustion must be adequately confined or contained. Concerning the first two of these, the trial court said:

"Viewing the evidence most favorably for plaintiff, it must be taken as established that the soybean dust in the beanhouse, if suspended in the atmosphere in correct proportions, would provide an explosive mixture. And, evidence of a spark inside the building given by one witness who was there when the incident occurred would support a finding of a source of ignition. The verdict is on sound ground as to these two points."

We agree with that evaluation.

On the question of containment, however, the trial court took a different view. It concluded that no evidence of any adequate containment was introduced, and it regarded this deficiency as conclusive against the plaintiff's claim. We think that the court erred in assuming from the evidence before it that the building itself could not have provided the necessary containment for an explosion which damaged only two-thirds of the structure.

The fundamental principle of containment is not difficult to grasp. The heat generated by the burning of a potentially explosive mix causes the surrounding gases to expand rapidly. If the gases are confined or contained in some fashion, the resulting increase of pressure against the surfaces of the container may produce an explosion. If there is no efficient container, then, through the processes of convection and conduction, the explosive potential is dissipated. Plaintiff's theory, as we have noted, is that the collapse which occurred here was triggered by the ignition of a relatively small cloud of explosive bean dust suspended in the air space between the roof and the beans near the center of the building, which caused an uplift of the roof that precipitated the structural damage. The only "container" available to plaintiff under this theory is the beanhouse itself. Analyzing this theory, the trial court said:

"* * * Returning to the relative concept of containment, it is clear that if the volume of explosive dust had been large enough, the beanhouse itself could serve as the explosion chamber, but if the dust were not that extensive its ignition would generate, at most, only a flash fire, incapable of explosively causing structural damage — the pressure and heat increases would dissipate their energy at some point between the outer limits of the dust and the interior surfaces of the beanhouse. * * * Plaintiff's expert on this matter, the engineer, testified that the explosion did not involve the whole building but instead occurred in only a small area near its center. It was his opinion that if the whole building had been involved it would have been totally destroyed. If, as plaintiff's case requires, a small explosion exerting pressure on a localized part of the beanhouse is at least sufficient to cause the failure of that part of the structure, the conclusion is inescapable that this witness was correct in the last stated opinion, for an explosion involving the whole building and depending on the beanhouse itself for the element of containment would undoubtedly exert proportionately magnified pressures on all interior surfaces, with correspondingly greater effect."

These statements indicate clearly that the court simply concluded that since the whole building was not blown apart, the damage which did occur could not possibly have been caused by an explosion of the sort claimed by the plaintiff.

Forearmed with this conclusion, the court then relied upon the responses of plaintiff's expert, the engineer, to four questions on cross examination which superficially make it appear that in expressing an opinion that an explosion did occur, the witness wholly failed to give consideration to the requirement of containment:

"Q. And this cloud of dust was out in the open?

"A. Yes.

"Q. It was not confined or contained in any fashion?

"A. We assume not, no.

"Discussing the analogy of a gas stove leaking gas, but leaking enough to cause only a puff instead of an explosion when ignited, the witness explained that this resulted from the relationship between the amount of vapor and the area which it occupied. His cross-examination continued, as follows:

"Q. In other words, you had no containment in a large area, is that right?

"A. That's right.

"Q. And the explosion you are talking about is an explosion involving no containment in a large area, isn't it, Mr. ____________________?

"A. Yes.

"In one breath the witness correctly explains the reason for the lack of explosion in the gas stove analogy, and in the next he agrees that the same circumstances apply to the explosion which he claims occurred in the beanhouse. In short, his opinion is in direct conflict with established physical laws, just as are the inferences plaintiff seeks to draw from the remainder of the evidence. * * *"

On the strength of these responses, the court concluded that the witness' opinion was in direct conflict with an established physical law, and thus devoid of probative value.

The following excerpts from other portions of the witness' testimony suffice to demonstrate that he was in fact aware of and taking into account the necessity for containment in formulating his opinion as to what caused the collapse:

"A. * * * In an explosion of this type the action involved is not so much the pressure that the explosion builds up but the rate at which it builds up and the way in which it forces the air away from it. Now, it is a kinetic action. It is not just a question of building up a pressure within a whole area of the building. If these men were down a way, they might have been completely shielded from the action. * * * In my opinion that is why a comparatively, we will say a comparatively small explosion in a building of this type, as compared to the total volume of the building, could possibly generate a sufficient amount of energy against the walls and roof to have caused an uplift."

The real conflict appears to be not between the witness' opinion and any physical law established in the record, but rather between the different meanings ascribed by the witness and the court to the notion that containment is relative. The witness — an expert trained in engineering — was clearly of the opinion that it meant that an explosion could occur without accompanying total destruction of the beanhouse, while the trial court was just as clearly of the view that any explosion which utilized the building itself for confinement would doubtless result in total destruction of the structure. We think that the jury had the right to resolve that conflict in favor of the view expressed by the expert witness, for we find no adequate basis in the record to sustain the position taken by the trial court that the explosion postulated by the plaintiff's witness was literally a physical impossibility.

In the alternative to its motion for judgment n. o. v., the defendant moved for a new trial, which was granted conditionally. Our review of the latter grant stands upon a footing entirely different from our review of a judgment n. o. v. A district judge may, in the exercise of his discretion, grant a new trial for reasons which would not support the award of a judgment notwithstanding a verdict. Hampton v. Magnolia Towing Co., 338 F.2d 303, 306 (5 Cir., 1964). Here, the trial court determined that the jury's verdict was contrary to the overwhelming weight of evidence. Appellants do not challenge this ruling, nor could they successfully do so. We find no error in it.

The judgment is reversed, and the cause is remanded for a new trial.