John L. Champeau appeals from judgment on a jury verdict1 in favor of appellees Fruehauf Corporation and Kelsey-Hayes Company following a second trial of this case in the United States District Court for the District of Minnesota. On appeal Champeau contends that the district court2 erred in granting appellees' motion for new trial following the first trial, and that the court3 erred in admitting into evidence a videotaped experiment at the second trial. We affirm.

Champeau was an over-the-road truck driver for Schultz Leasing, Inc. On October 28, 1977 he was traveling from Purdum, Nebraska to Sioux Falls, South Dakota, and was pulling a Fruehauf trailer which had computer assisted anti-lock brakes manufactured by Kelsey-Hayes. A local resident in Purdum advised Champeau that the shortest route to Sioux Falls was over a county road near Elsmere, Nebraska, and Champeau decided to take that route. The Elsmere road was under construction at the time and was closed to all but local traffic. Part of the road had been paved and part remained gravel.

The weather on the morning of October 28 included mist or light drizzle. Champeau negotiated the gravel portion of the road and entered the paved portion, which he testified had both wet and dry spots. He subsequently came upon two "S" curves in the road and successfully negotiated the first. He testified that at the time he first applied the brakes in preparation for the second "S" curve, he was approximately one-quarter mile from the start of the curve and was traveling at thirty-five miles per hour. When he applied the brakes, only the tractor brakes functioned. He released the brakes and applied them for a second time. Once again the trailer brakes did not operate. He was already into the curve when he applied the brakes for the third time. The trailer brakes failed again and the trailer pushed the tractor into a jackknife position.

Champeau suffered a neck injury which resulted in his inability to continue work as a truck driver. He later underwent disc surgery, which relieved a good deal of his pain, and he obtained employment as a janitor in a school. He has suffered a substantial decrease in earnings.

Champeau brought this products liability action against appellees alleging that the computer assisted anti-lock braking system manufactured by Kelsey-Hayes and installed by Fruehauf on its trailers was defectively designed and unreasonably dangerous. At the first trial, the jury found for Champeau and awarded him one million dollars in damages. Appellees moved for judgment n.o.v. or new trial, alleging several grounds. Following oral arguments, appellees' motion for new trial was granted by the district court without comment.

During the second trial, appellees were allowed to introduce into evidence a videotaped experiment performed at the scene of the accident which tended to show that the truck would have coasted to a stop short of the curve if, as he testified, Champeau took his foot off the accelerator one-quarter mile from the curve while traveling at thirty-five miles per hour and did not reapply the accelerator thereafter. In connection with this videotape, a list of differences in conditions between the experiment and the accident was read to the jury. As indicated, at the second trial the jury found for appellees. Champeau now appeals. We address first Champeau's challenge to the district court's grant of a new trial to appellees following the first trial.

I. NEW TRIAL.

"The authority to grant a new trial ... is confided almost entirely to the exercise of discretion on the part of the trial court." Allied Chemical Corp. v. Daiflon, Inc., 449 U.S. 33, 36, 101 S. Ct. 188, 191, 66 L. Ed. 2d 193 (1980); see also Aimor Electric Works, Ltd. v. Omaha National Bank, 727 F.2d 688, 692 (8th Cir.1984); Commercial National Bank v. Missouri Pacific Railroad, 631 F.2d 563, 565 (8th Cir.1980). The trial court's decision therefore should not be reversed absent a strong showing of abuse of discretion. Aimor Electric Works, Ltd., 727 F.2d at 692; Lane v. Chowning, 610 F.2d 1385, 1388 (8th Cir.1979); see also Commercial National Bank, 631 F.2d at 565.

We are somewhat hampered in our review of the court's new trial decision because the court failed to state on the record or in its order which of the several grounds advanced by appellees formed the basis for the grant of the new trial. While Fed.R.Civ.P. 59(a) does not require such a statement, "[i]t is in the interest of good practice and the dispatch of business for trial courts to specify in their orders granting new trials the grounds on which such orders are made." 58 Am.Jur.2d New Trial Sec. 214 (1971); see also 66 C.J.S. New Trial Sec. 210(a)(3)(a) (1950).4

Since we have few, if any, clues as to which ground or grounds formed the basis for the court's grant of a new trial to appellees, the question becomes how do we review that decision? Research has revealed few federal cases addressing similar situations. In Somerville v. Capital Transit Co., 192 F.2d 413, 414-15 (D.C.Cir.1951), cert. denied, 342 U.S. 941, 72 S. Ct. 553, 96 L. Ed. 700 (1952), the motion for new trial cited numerous grounds. In addition to the fact that the trial court failed to state on the record or in its order which grounds formed the basis of its decision to grant a new trial, no record of the evidence produced at the first trial was presented to the appellate court. The court held that "the state of the record before us precludes at the outset a finding of abuse in this case[.] ... Accordingly, we cannot characterize the action of the trial court in granting a new trial in the manner done here as an abuse of discretion." Id. at 415 (footnote omitted). In the present case we do at least have a transcript of the first trial, and the approach of the D.C.Circuit in Somerville seems unjust here. If we were to adopt the Somerville approach, a trial court could effectively avoid review of new trial decisions simply by not stating the grounds. This would deny parties the appellate review of such decisions, limited though it may be, to which they are entitled.

In Powell v. Lititz Mutual Insurance Co., 419 F.2d 62, 65 (5th Cir.1969), and Berner v. British Commonwealth Pacific Airlines, Ltd., 346 F.2d 532, 541 (5th Cir.1965), cert. denied, 382 U.S. 983, 86 S. Ct. 559, 15 L. Ed. 2d 472 (1966), the Fifth Circuit seemed to adopt an approach to this problem whereby the appellate court scans the record and determines for itself what ground was most likely the basis for the court's decision. This approach may be too broad. First, it may lead to an affirmance on a ground not even considered by the trial court. Second, if the trial court's decision was based on any ground not mentioned in the motion, it may be observed that Rule 59(d) requires the court to state specifically that ground. Failure to do so may constitute error.

We can also look upon authority from states having a new trial rule substantially similar to Rule 59 as instructive. Ark.R.Civ.P. 59 is not materially different from its federal counterpart. In addressing a similar situation, the Arkansas Supreme Court stated as follows:We have said that when an order granting a new trial is expressed in general terms without a specification of grounds, it must be affirmed if it can be supported on any ground alleged in the motion.

General Motors Corp. v. Tate, 257 Ark. 347, 349, 516 S.W.2d 602, 604 (1974). We have adopted a similar approach in the obverse situation where the district court articulates several grounds for granting a new trial, most of which would constitute abuse of discretion. If the grant of a new trial is adequately supported by one of the court's stated grounds, we have said that there is no need to review every other ground for abuse of discretion. See Commercial National Bank, 631 F.2d at 565.

In reviewing the grant or denial of a new trial where the trial court has failed to state on the record or in its order what grounds formed the basis of its decision, we will adopt the approach of the Arkansas Supreme Court in Tate. If the trial court's decision can be supported by any single ground stated in the motion, ordinarily it should be affirmed. With this standard in mind, we now turn to specifics of the grant of a new trial here.

The grounds for a new trial advanced in appellees' motion were that the verdict was against the weight of the evidence, the damages awarded Champeau were excessive, the trial court erred in instructing the jury, the trial court was guilty of prejudicial misconduct, and the trial court committed evidentiary errors during trial. From our review of the record, we conclude that the strongest of these grounds is judicial misconduct.

"Intonations, gestures and comment of the trial judge, or the comment by itself, may be so improper as to warrant a new trial. And, in general, any misconduct of the court or its officers that is prejudicial will warrant a new trial." 6A J. Moore, J. Lucas & G. Grotheer, Jr., Moore's Federal Practice p 59.08 at 59-90 (2d ed. 1986) (footnotes omitted).

A trial judge should never assume the role of advocate, and must preserve an attitude of impartiality in the conduct of a trial. It is also true, however, that a trial judge may question witnesses or comment on the evidence. A federal judge is more than a "mere moderator" or umpire in the proceedings and he may take an active role in conducting the trial and developing the evidence.

Warner v. Transamerica Insurance Co., 739 F.2d 1347, 1351 (8th Cir.1984) (citations omitted).

In their motion for new trial, appellees accused the trial court of continuous and improper questioning of witnesses, solicitation of factual and legal arguments from the attorneys in the presence of the jury, numerous improper comments on the evidence, and other prejudicial misconduct. A review of the record shows that the trial court continually interrupted the testimony of both Champeau's expert and appellees' expert. The interruption and questioning of appellees' expert was much more frequent, however. Appellees state in their brief that the court interfered with their expert's testimony no fewer than one hundred forty-five times. While this court has not tallied these occasions, we find no reason from our study of the record to doubt appellees' numbers.

On many of these occasions it appears that the court was trying to clarify the technical explanations of the expert for the benefit of the jury. Even so, appellees appear to have been prejudiced because the court at times either led the expert off into an irrelevant discussion or completely confused and perhaps misstated what the expert had said. One such confused exchange dealt with the melting of certain material on the computer system and its effect on the regular brakes on the trailer.5

These numerous bewildering attempts to clarify what to some may have been perfectly clear from the beginning likely confused the jury or led them to believe that the court found appellees' expert's testimony incomprehensible. A few such exchanges could have been overlooked, but the frequency of intervention by the court in this case is overwhelming.

The court also on several occasions solicited and encouraged factual and legal arguments between counsel in the presence of the jury. Some of these exchanges were lengthy, inherently argumentative and involved matters not in evidence. After one such exchange, the following took place in the presence of the jury:

THE COURT: Now I want to ask you: Do any of you lawyers object to clarifying the issues at this point, as I have just done?

[APPELLEES' ATTORNEY]: Well--

THE COURT: I want to know the answer to that.

[APPELLEES' ATTORNEY]: Well, it's already done, Your Honor, and we can't erase it from the jury's mind, so I think we will just have to go on. That was our final argument.

THE COURT: Do you want a mistrial?

[APPELLEES' ATTORNEY]: No, no, we feel--

THE COURT: Have I done anything wrong?

[APPELLEES' ATTORNEY] No.

In reading this exchange and others, we can literally feel the frustration of appellees' attorney, and sympathize with his reluctance to take the risk of telling the court its errors in front of the jury. Later in the trial and out of the presence of the jury, appellees' attorney took exception to the arguments before the jury and was asked once again by the court if he wanted a mistrial. Counsel responded that he just wanted to make an objection on the record.

The court also undertook on several occasions to comment on the evidence and explain to the jury its relevance to the case. These comments appear for the most part to have been largely unbiased, but they further contributed to the overall problem of the court's conduct. Also contributing to the problem was the court's celebration of a birthday with the jury during trial. As related in appellees' brief, the party was complete with cake, song and toasting, creating a "circus-like" atmosphere. Additionally, between the time the jury was charged and the time they rendered their verdict, the trial judge appeared on the CBS news magazine show 60 Minutes, where he generally lambasted manufacturers and their products.

Champeau suggests that it is unlikely that the trial court would have recognized its own misconduct and granted a new trial as a result. We disagree. First, as one of the grounds alleged in appellees' motion, we must assume that the trial court did grant the new trial for misconduct and we must then examine for an abuse of discretion. Second, the trial court's activist style was well known and widely discussed. Indeed, shortly before the first trial in this case, a majority of our court in United States v. Singer, 710 F.2d 431 (8th Cir.1983) (en banc), reversed a conviction due to judicial misconduct and criticized this same trial court for its activist style of presiding. It is therefore not unlikely that the trial court realized that it went too far in this case. We are confident that the court did not intend to be biased or to prejudice either side. This was merely the trial court's style, like it or not.

Even though not every incident was objected to and appellees did not request a mistrial, the trial court still could have granted a new trial based on plain error. See United States v. Singer, 687 F.2d 1135, 1141 (8th Cir.1982), rev'd en banc, 710 F.2d 431 (1983); Newman v. A.E. Staley Manufacturing Co., 648 F.2d 330, 334-35 (5th Cir. Unit B 1981). We feel that the overall conduct of the trial court prejudiced appellees sufficiently to prevent them from receiving a fair trial, thus affecting their substantial rights. We therefore hold that the trial court would not have abused its discretion in granting a new trial on the ground of judicial misconduct.

In coming to this conclusion we are mindful also of the size of the verdict, one million dollars, and the extremely weak case on liability. While we need not, and do not, hold that damages were legally excessive, or in present context that evidence was insufficient to support a plaintiff's verdict, both amount of award and meager evidence of liability may well have been factors in persuading the court of its judicial over-activism.

Indeed we are constrained to hold that it would not have been an abuse of discretion for the trial court to have granted a new trial on the ground that the verdict was against the weight of the evidence. "In determining if a verdict is against the weight of the evidence, the trial court may conduct its own evaluation of the evidence. It may 'weigh the evidence, disbelieve witnesses, and grant a new trial even where there is substantial evidence to sustain the verdict.' " Brown v. Syntex Laboratories, Inc., 755 F.2d 668, 673 (8th Cir.1985) (quoting Ryan v. McDonough Power Equipment, Inc., 734 F.2d 385, 387 (8th Cir.1984)). Great deference is to be given the trial court's decision. Brown, 755 F.2d at 673.

Under Minnesota law, causation is an essential element of a strict liability theory. McCormick v. Custom Pools, Inc., 376 N.W.2d 471, 475 (Minn.Ct.App.1985). Champeau attempted to prove two problems with the computer assisted anti-lock brake system. First, his expert testified that the system had not been adequately tested under all conditions that would be encountered on the road. Second, his expert testified that the system was defectively designed because screw type ends were used on the cables in order to fasten them to the computer. He testified that these connectors could work loose and allow corrosion, dust and other factors to impede conduction through the connections. Even though he had never inspected Champeau's trailer and had no idea what shape the connectors were in at the time of the accident, he opined that this design defect caused the brake failure resulting in Champeau's accident.

According to Champeau's testimony, the trailer brakes were operational when he started out the morning of the accident. Champeau also stated that the trailer brakes worked when he pulled the trailer back to Minnesota using a different tractor. No one who saw or worked on the trailer following the accident noticed anything unusual about the computers or the connectors (there is one computer per axle). Champeau's expert's theory of the design defect is suspect, but assuming that it is true, Champeau has presented no direct evidence of causation. The only circumstantial evidence of causation was his expert's opinion that the design defect caused Champeau's trailer brakes to fail. This was pure speculation and was insufficient to support an inference of causation.

Champeau presented evidence that the trailer brakes failed and he presented evidence of a potential specific defect, but he failed to connect the two with direct or circumstantial evidence. In light of Champeau's failure to prove causation, it would not have been an abuse of discretion for the trial court to have granted a new trial on the ground that the verdict was against the weight of the evidence. Unlike the situation in Crane v. Consolidated Rail Corp., 731 F.2d 1042, 1051 (2d Cir.), cert. denied, 469 U.S. 854, 105 S. Ct. 179, 83 L. Ed. 2d 114 (1984), the record here would firmly support such action.

II. ADMISSIBILITY OF THE VIDEOTAPED EXPERIMENT.

"The admissibility of evidence of experimental tests rests largely in the discretion of the trial judge and his decision will not be overturned absent a clear showing of an abuse of discretion." Sprynczynatyk v. General Motors Corp., 771 F.2d 1112, 1124 (8th Cir.1985), cert. denied, --- U.S. ----, 106 S. Ct. 1263, 89 L. Ed. 2d 572 (1986). "A court may properly admit experimental evidence if the tests were conducted under conditions substantially similar to the actual conditions. Admissibility, however, does not depend on perfect identity between actual and experimental conditions. Ordinarily, dissimilarities affect the weight of the evidence, not its admissibility." Randall v. Warnaco, Inc., 677 F.2d 1226, 1233-34 (8th Cir.1982) (citations omitted); see also Patterson v. F.W. Woolworth Co., 786 F.2d 874, 880 (8th Cir.1986).

Champeau challenges the admission of a videotaped experiment into evidence at the second trial. The experiment showed that if the driver of a tractor-trailer rig traveling thirty-five miles per hour took his foot off of the accelerator one-quarter mile from the curve that was the scene of the accident and never accelerated again, the rig would coast to a stop short of the curve. The purpose of the experiment was not to recreate the accident, but to take Champeau's distance and speed estimates and show that under the applicable laws of physics the accident could not have occurred as Champeau had described.

Some similarities in conditions existed between the experiment and the accident, but the conditions were far from identical. In order to remedy this situation the district court submitted to the jury lists of differences and similarities written by the attorneys. In light of the similarities in conditions and the lists of similarities and differences submitted to the jury, we cannot find error in the admission of the videotaped experiment into evidence.

Moreover, the experiment did not need to be performed in similar circumstances in order to be admissible because it did not purport to be a recreation of the accident and it was merely used to demonstrate general principles of physics as applied to Champeau's testimony. See Nanda v. Ford Motor Co., 509 F.2d 213, 223 (7th Cir.1974); Harkins v. Ford Motor Co., 437 F.2d 276, 278 & n. 5 (3d Cir.1970). We therefore conclude that the court did not abuse its discretion in admitting this evidence.

III. CONCLUSION.

Having affirmed the grant of a new trial to appellees following the first trial and having found no error in the admission of the videotaped experiment into evidence in the second trial, we conclude that judgment on the jury verdict of the second trial in favor of appellees should be, and it is, affirmed.