Appellant and cross-appellee General Motors Corporation (GM) appeals from a final judgment entered in the United States District Court for the District of North Dakota based upon a jury verdict in an automobile product liability and negligence action brought by appellees and cross-appellants, Vivian Sprynczynatyk and her husband Paul Sprynczynatyk (Sprynczynatyks or plaintiffs). The Sprynczynatyks were awarded $5,025,000 in actual damages from GM. We reverse and remand with directions to the district court to conduct further proceedings consistent with this opinion.

On July 16, 1980, at approximately 8:00 p.m. a one car accident occurred on a gravel road near Bismarck, North Dakota. Fourteen-year-old Rodney Sprynczynatyk (Rodney) was driving the family-owned 1980 Chevrolet Citation X-car. Rodney, who had obtained his learner's permit 2 or 3 weeks earlier, was accompanying his mother, Vivian Sprynczynatyk (Vivian) on some errands that evening. During the return trip to their house, Vivian, who had been driving, stopped and let Rodney drive the car. Rodney had driven the Citation a number of times before.

After Vivian turned the operation of the car over to Rodney, he drove approximately one mile on paved Highway 1804 and then turned north onto a gravel road. As Rodney started down a hill (approximately 11% grade), he encountered difficulties controlling the car. At some point the car left the road on the right (east) side, overturned, and came to rest on its top with the front of the car facing south. Vivian sustained injuries which rendered her quadriplegic.

Vivian and her husband Paul filed this action in district court against GM, the manufacturer of the 1980 Citation. The plaintiffs brought this action based upon theories of negligence and strict liability. They alleged that the Citation had either defectively and unreasonably dangerous or negligently designed brakes or both. The Sprynczynatyks contended that the rear brakes on the Citation locked causing the car to spin around 180? and roll as it tipped over into the east ditch. GM denied the alleged brake defect and disputed plaintiffs' theory of how the accident occurred. GM contended that Rodney did not apply the brakes, and even if he did they didn't lock, but that Rodney merely panicked as a young inexperienced driver and oversteered the car off the road.

Vivian sought actual damages for her personal injuries and Paul sought actual damages for loss of consortium and lost services. Two months before trial the plaintiffs sought leave to amend their complaint to include a request for punitive damages. Plaintiffs' motion was denied as untimely.

Rodney was not a named party to the lawsuit. The Sprynczynatyks' insurance carrier had paid them $25,000, the policy limits, and the Sprynczynatyks had executed a release in favor of the named insured, Paul, his "heirs" and "all others," which arguably ran to Rodney.

Prior to trial Rodney was hypnotized by a trained psychologist, Dr. Robert Gordon, at the request of the plaintiffs' counsel. The hypnotic session was videotaped. While hypnotized Rodney recalled applying his right foot to the brakes as hard as he could before the car spun and left the gravel road. Prior to the hypnosis session, Rodney's recollection was different. On July 28, 1980, he gave a statement to an insurance agent that he didn't apply the brakes at all, and during his September 16, 1982 deposition he testified that he could have used the brakes but his best recollection was that he did not apply them.

During the hypnotic session, after Dr. Gordon put Rodney in a trance, he told Rodney his mind could reach back and recollect things that he never thought were possible. Dr. Gordon then led Rodney back to the events of the day of the accident. Rodney recounted his actions throughout the day and then what happened during the accident. The two critical passages relating to his application of the brakes are as follows:

RODNEY: I get back into the right tracks and start going down and the car went over to the left.

GORDON: It's going to the left?

RODNEY: Um-hum. And I see that I'm gonna go in the ditch and hit the fence.

GORDON: You see you're gonna hit it. What are you going to do? Where is your right foot?

RODNEY: On the brake.

GORDON: Where's your left foot?

RODNEY: Over the clutch.

GORDON: What are you doing with your right foot and how much pressure are you putting on the brake?

RODNEY: Pushing.

GORDON: How hard?

RODNEY: Hard.

GORDON: As hard as you can?

RODNEY: Um-hum.

GORDON: Are you pumping it or pushing it?

RODNEY: No. I'm holding it steady.

GORDON: Holding it steady. What's happening now?

RODNEY: Car is sliding.

GORDON: What are you going to do?

RODNEY: Screaming, the car turns around.

GORDON: Where's your foot?

RODNEY: Still on the brake.

* * *

* * *

GORDON: What happens now?

RODNEY: The car spins all the way around. I'm looking the other way.

GORDON: Where's your left foot?

RODNEY: Over the clutch now.

GORDON: Where's your right foot?

RODNEY: On the brake.

GORDON: What happens now?

RODNEY: I look over and I see the ditch coming closer to me. I keep screaming. I take my hands off the wheel and I feel the car tipping and I put them on the ceiling and shut my eyes.

GORDON: When do you shut your eyes?

RODNEY: Right when the car is tipping over. I feel it tipping. I put my hands on the ceiling and I close my eyes.

GORDON: Where are your feet?

RODNEY: On the floor.

After undergoing hypnosis, Rodney testified at his November 14, 1983 deposition that he recalled applying the brakes hard just before the car started to spin.

GM filed a motion in limine to limit Rodney's trial testimony to his pre-hypnosis statements or to require plaintiffs to establish the reliability of Rodney's post-hypnosis testimony prior to allowing him to testify at trial. The plaintiffs opposed the motion and after oral argument the district court denied GM's motion in full.

The liability issues of the case were tried to the jury first. As part of plaintiffs' case-in-chief, following Vivian Sprynczynatyk's testimony, Dr. Gordon testified regarding Rodney's hypnosis session. During the direct examination of Dr. Gordon, plaintiffs offered the videotapes of Rodney's hypnosis session as evidence and sought to play the videotapes for the jury. GM objected on the grounds that the tapes were not admissible to prove the facts that were recited on the tapes. The court overruled the objection at which time GM orally requested that the court give a cautionary instruction that the matters that were related on the tapes were not to be taken as proof of the facts recited.1 The court gave the following oral instruction:

THE COURT: Ladies and gentlemen, the tapes that you are about to see are being received on the issue of--the ultimate issue of the credibility of the recall of the witness, Rodney Sprynczynatyk, and we aren't here--the purpose of them is to permit you to see it and you will be receiving other evidence on this subject both on the part of the plaintiff and contrary evidence from the defendant. And the purpose of viewing the tapes is to permit you to evaluate the opinions that you will subsequently hear.

The videotapes of the entire 56 minute hypnosis session were then viewed by the jury.

Rodney testified after Dr. Gordon without objection from GM. On direct examination Rodney testified that when the car was going toward the left ditch he started applying pressure to the brakes, the car kept going and then he hit them all the way and the car spun. On cross-examination Rodney testified that he now remembered things differently about the application of the brakes during the course of the accident than he did before hypnosis.

As part of its defense GM presented the expert witness, Dr. Martin Orne, a leading specialist in the field of hypnosis. Dr. Orne testified in general to the unreliability of hypnosis as a memory refresher and in particular to the absence of certain procedural safeguards in Rodney's hypnosis. Dr. Orne viewed Dr. Gordon's repeated questions regarding the location of Rodney's feet in relation to the brakes as a fatal flaw that was too suggestive and opined that Rodney's "new memory" was a creation of hypnosis and unlikely to be true.

The jury returned a verdict in favor of the Sprynczynatyks on both the strict liability and negligence counts. Following a trial on the damages, the jury awarded Vivian $4,500,000 and her husband Paul $525,000. The trial court entered judgment in accordance with the verdicts and denied GM's post-trial motions.

GM then filed this appeal. For reversal GM argues that the trial court erred 1) in admitting the videotapes of Rodney's hypnosis session; 2) in admitting Rodney's hypnotically enhanced testimony; 3) in excluding certain evidence of GM's post-crash tests of the accident car; 4) in failing to instruct on comparative fault; and 5) that there was not substantial evidence to support the verdict.

The Sprynczynatyks cross-appeal, arguing that the trial court abused its discretion when it denied them leave to amend their complaint to include a claim for punitive damages two months before trial. We discuss each of these issues in turn.

The first issue we address is GM's contention that it was prejudicial and reversible error for the trial court to admit the videotapes of Rodney's hypnosis session, and to permit the jury to view them during plaintiffs' case-in-chief without a specific instruction warning against their prohibited use to prove the truth of the matters asserted on the tapes. Plaintiffs contend the district court did not abuse its discretion in admitting the videotapes because they were admitted as a demonstrative aid to help the jury understand the expert testimony about the hypnotic process. We disagree.

In general a videotape offered to prove the truth of the matter asserted constitutes inadmissible hearsay. FED.R.EVID. 801(c). United States v. Dorrell, 758 F.2d 427, 434 (9th Cir.1985). In particular, courts have held that the contents of actual hypnotic interviews are inadmissible for the purpose of proving that the facts recounted by the hypnotized witness actually occurred. State v. Brown, 337 N.W.2d 138, 153 (N.D.1983) (en banc); State v. Beachum, 643 P.2d 246, 254 (N.M.App.1981). Thus a videotape is inadmissible as evidence unless it comes within an exception to the hearsay rule, see, e.g., Grimes v. Employers Mutual Liability Insurance Co., 73 F.R.D. 607, 611 (D. Alaska 1977), or unless it is offered for some other permissible purpose which would remove it from the definition of hearsay.

When evidence can be admitted for one purpose but cannot be admitted for some other purpose, it has limited admissibility and is governed by Rule 105 of the Federal Rules of Evidence. Rule 105 provides in relevant part that:

When evidence which is admissible * * * for one purpose but not admissible * * * for another purpose is admitted, the court, upon request, shall restrict the evidence to its proper scope and instruct the jury accordingly.

Rule 105 makes it the duty of the judge to restrict the evidence to its proper scope and entitles the opponent to a binding instruction that alerts the jury to the possibility of the forbidden use and tells them not to use it for that purpose. Since the videotapes of Rodney's hypnosis session are inadmissible to prove the truth of the matters asserted, but arguably admissible for the purpose of explaining the hypnotic procedure to the jury, the question is whether the court's limiting instruction was sufficient.

In this case GM requested in the presence of the jury a specific instruction stating the prohibited use, namely, that the videotapes were not offered to prove the truth of the facts recounted. Instead of adopting GM's proffered instruction, the district court told the jury that the tapes were being received on the ultimate issue of the credibility of Rodney's recall and that the purpose was to permit the jury to evaluate the opinions they would hear. Having viewed the videotapes and having reviewed the record in this case, we are convinced that the instruction given before the videotapes were shown was erroneous and insufficient under Rule 105 to prevent the prejudicial use of this evidence for the following reasons.

First, the court's oral instruction told the jury that the videotapes were being received on "the ultimate issue of the credibility of the recall of the witness, Rodney * * *." This was clear error in the sense that the jury was directed to consider the videotapes for the purpose of assessing Rodney's credibility. Here the videotapes were admitted as probative of the truth of what Rodney said on those tapes and that was error. Second, in the latter part of the instruction the court stated the permissible use of the tapes, namely for explanatory purposes, but failed to warn of the prohibited use. The instruction did not "restrict the scope" of the evidence. In fact, that part of the court's instruction gave a negative signal to the jury when considered in light of GM's specific request that the videotapes not be considered as truth of the matters asserted. By rejecting the prohibited use language as just articulated by GM's counsel, and utilizing only permissible use language, the jury was improperly led to think GM's proffered instruction was incorrect. Although we do not hold that Rule 105 requires all limiting instructions to contain the prohibited use language, we do state that in this case it was necessary in light of GM's specifically worded proposed request that was made in the presence of the jury. The inference that GM's requested instruction was wrong, when coupled with the erroneous issue-directing statement at the beginning of the instruction, renders the district court's instruction legally inadequate.2

Further, we find that under the circumstances of this case the showing of the videotapes of Rodney's hypnosis session without a proper cautionary instruction and during plaintiffs' case-in-chief was highly prejudicial to GM and not harmless. The issue of whether Rodney ever applied the brakes was central to the lawsuit and hotly contested. It was on the videotapes that the jury saw and heard Rodney, for the first time, state that he applied the brakes during the accident. The jury could not reasonably be expected to disregard the provocative nature of this inadmissible evidence and only consider what they saw as an aid to understanding the process of hypnosis. Other relevant evidence was available to accomplish this didactic purpose. In addition the jury could not reasonably be expected to overlook the fact that the videotapes were shown during plaintiffs' case-in-chief before GM ever challenged Rodney's hypnosis session and hypnotic recall. Because of the insufficient limiting instruction, the highly prejudicial nature, and premature presentation of the evidence, we conclude the district court abused its discretion in admitting the videotapes and that GM is entitled to a new trial.

Since this case must be retried, we turn to the other issues which are likely to arise anew, and upon which we feel compelled to comment.

II. Admission of Hypnotically Enhanced Testimony

The second hypnosis-related issue presented in this case is whether the trial court erred, as GM contends, in admitting Rodney's testimony at trial relating his post-hypnosis recollection that he applied the brakes during the accident.

As a preliminary matter the court must determine whether GM has preserved this point for appellate review. The plaintiffs contend that GM's failure to object during Rodney's direct examination precludes this court's review because FED.R.EVID. 103(a)(1) requires a timely objection be made in order to preserve an issue for appeal. GM responds that the district court's denial of its motion in limine was sufficient to preserve the issues raised in the motion.

Ordinarily in this circuit a motion in limine does not preserve error for appellate review. Hale v. Firestone Tire & Rubber Co., 756 F.2d 1322, 1333 (8th Cir.1985). However, we find that requiring a formal objection when Rodney took the witness stand at trial was not necessary under the circumstances of this case for two reasons.

First, we adhere to the Third Circuit's approach that Rule 103(a)(1) should be read in tandem with Federal Rule of Civil Procedure 46 which states that formal exceptions are unnecessary and that the test is whether an objection at trial would have been more in the nature of a formal exception or in the nature of a timely objection calling the court's attention to a matter it need consider. American Home Assurance Co. v. Sunshine Supermarket, Inc., 753 F.2d 321, 324 (3d Cir.1985). In the instant case the district court made a definitive pre-trial ruling that affected the entire course of the trial. The district court's denial of the motion was not made conditionally or with the suggestion that the matter would be reconsidered. It was not a typical motion in limine situation where a hypothetical question is posed whose nature and relevance is unclear before trial. The matter was fully briefed and argued. Under these circumstances requiring an objection when Rodney testified on the grounds raised in the motion in limine would have been in the nature of a formal exception and thus unnecessary under FED.R.CIV.P. 46. Second, we note that when the district court denied the motion in limine it implicitly denied GM's alternative request that the plaintiffs be required to establish before trial the reliability of Rodney's hypnotically enhanced testimony. There can be no question that the propriety of refusing to conduct such a preliminary hearing on the reliability issue is squarely before this court and is not contingent upon the making of any objection during the trial.

We now turn to the merits of the issue of whether Rodney's hypnotically enhanced testimony is admissible. Stated in general terms, the issue is whether the testimony of a witness who has undergone hypnosis to refresh his or her recollection is admissible, and, if so, under what circumstances. This question, previously undecided in this circuit,3 has been the subject of much discussion and disagreement.4

Before discussing the various legal approaches to this question it will be helpful to explain the phenomenon of hypnosis and to note the problems it creates in the legal context.

Hypnotism has been defined as "[t]he act of inducing artificially a state of sleep or trance in a subject * * * generally characterized by extreme responsiveness to suggestions from the hypnotist."5 In a typical hypnotic session the hypnotist leads the subject to focus his or her attention, suspend critical judgment, follow the suggestions of the hypnotist, concentrate on a past event and then recount the past event. Hypnosis has been recognized by the American Medical Association as a valid therapeutic technique since 1958.6 However, more recently hypnosis has been utilized for law enforcement investigative purposes in an attempt to aid witnesses and victims to remember forgotten details regarding an event that is the subject of a criminal or civil suit. It is when these previously hypnotized victims or witnesses take the stand to testify at trial that hypnosis becomes a legal issue. For although its use is generally accepted as therapy, the reliability of hypnosis as a truth-exacting device is controversial.7

There are many problems inherent in the hypnosis process that affect the degree in which it can be, if ever, an accurate memory restorer. Hypnosis is characterized by hypersuggestibility and hypercompliance of the subject. State ex rel. Collins v. Superior Court, 132 Ariz. 180, 644 P.2d 1266, 1269 (1982); State v. Long, 32 Wash.App. 732, 649 P.2d 845, 848 (1982) (Swanson, J., concurring). The hypnotist can consciously or unconsciously lead the subject. "[A] hypnotized subject is highly susceptible to suggestion, even that which is subtle and unintended." State v. Mack, 292 N.W.2d 764, 768 (Minn.1980). A second problem associated with hypnosis is that the hypnotized witness may be influenced by the need to "fill in the gaps" in their memory, that is, to confabulate.

The hypnotic suggestion to relive a past event, particularly when accompanied by questions about specific details, puts pressure on the subject to provide information for which few, if any, actual memories are available. This situation may jog the subject's memory and produce some increased recall, but it will also cause him to fill in details that are plausible but consist of memories or fantasies from other times. It is extremely difficult to know which aspects of hypnotically aided recall are historically accurate and which aspects have been confabulated * * *.

State v. Hurd, 86 N.J. 525, 432 A.2d 86, 92-93 (1981) (quoting Orne, The Use and Misuse of Hypnosis in Court, 27 Int'l J. Clinical & Experimental Hypnosis 311, 317-318 (1979) ).

Another concern over the use of hypnosis is the impact that it has on the hypnotized person's memory. After hypnosis neither the hypnotist nor the subject can distinguish between actual memories and those psuedo-memories confabulated under hypnosis. After hypnosis the subject has one memory of the past event, the hypnotic memory, and that becomes hardened in the subject's mind. State v. Mack, supra, 292 N.W.2d at 769.

Thus the basic problem for the courts is that hypnosis does not insure the accuracy of the witness' recall. Quite often hypnotized persons produce more information following hypnosis, but it may be accurate or inaccurate and there is no scientific technique that can reliably discriminate between true or false details recounted during hypnosis.8

Largely because of the scientific uncertainty as to the reliability of hypnotic recall, courts have taken at least three different approaches to the issue of the admissibility of testimony enhanced by hypnosis. Some jurisdictions allow such testimony to go to the jury viewing hypnosis as affecting credibility, not admissibility.9 Other courts exclude such testimony as inadmissible per se.10 Another group of courts admit hypnotically induced testimony, but only if detailed procedural safeguards are followed.11

The courts that automatically permit hypnotically enhanced testimony view the testimony as the witness' present recollection of events, refreshed by hypnosis, and conclude that the witness is competent to testify. State v. Brown, supra, 337 N.W.2d at 151; Chapman v. State, 638 P.2d 1280, 1282-1284 (Wyo.1982). That the witness' memory may have been impaired by hypnosis or that suggestive material may have been used to refresh his or her recollection is considered to be a matter affecting credibility, not admissibility. Id. It is expected that cross-examination, expert testimony on the inherent risks of hypnosis and cautionary instructions to the jury will enable the jury to accurately assess the proper weight to be given to the evidence. Id.

The second approach to determining admissibility adopted by some jurisdictions is to apply the test of general acceptance by the scientific community first enunciated in Frye v. United States, 293 F. 1013 (D.C.Cir.1923). In Frye the court stated that the evidence relating to a scientific principle or discovery is admissible when the principle is established sufficiently to have gained general acceptance in a particular field. Id. at 1014. Courts that have followed this approach have imposed a per se rule of inadmissibility to post-hypnotic testimony after finding that the process of hypnosis is not generally accepted as reliable by the scientific community. See, e.g., People v. Shirley, 31 Cal. 3d 18, 181 Cal. Rptr. 243, 641 P.2d 775, cert. denied, 459 U.S. 860, 103 S. Ct. 133, 74 L. Ed. 2d 114 (1982); People v. Gonzales, 415 Mich. 615, 329 N.W.2d 743 (1982); and State v. Mack, supra, 292 N.W.2d at 768. In general these courts reviewed the scientific research on hypnosis, focused on the problems inherent in it as an accurate memory restorer and concluded that hypnosis has not received sufficient general acceptance in the scientific community to give reasonable assurance that the results produced will be sufficiently reliable to outweigh the risks of abuse or prejudice. People v. Gonzales, supra, 329 N.W.2d at 748.

As an example, in State v. Mack, supra, the court held that the results of hypnosis used to produce hypnotically induced "memory," like the results of mechanical or scientific testing, are not admissible unless the testing has developed or improved to the point where experts in the field widely share the view that the results are scientifically reliable as accurate. Id. at 768.

The third line of authority allows for the admissibility of hypnotically induced testimony if certain safeguards are followed to insure the reliability of the testimony. The leading proponent of this "procedural safeguard" approach has been the Supreme Court of New Jersey as articulated in State v. Hurd, supra. In Hurd the court held hypnotically induced testimony was admissible if the proponent of the testimony could demonstrate that the use of hypnosis in the particular case was a reasonably reliable means of restoring memory comparable to normal recall in its accuracy. The New Jersey court also held that in reviewing admissibility of hypnotically refreshed testimony, the trial court should evaluate the kind of memory loss that hypnosis was used to restore and the specific technique employed, based on the expert testimony presented by the parties. The court also laid down specific requirements12 which must be met before a party may introduce hypnotically refreshed testimony in a criminal trial. Therefore under the "procedural safeguard" approach the testimony may or may not be admissible depending upon the trial court's determination whether such testimony is reliable under the particular circumstances.

The Fifth Circuit has recently followed an approach similar in effect to the Hurd approach. In United States v. Valdez, 722 F.2d 1196 (5th Cir.1984) the court held that post-hypnosis testimony may or may not be admissible under FED.R.EVID. 403.13 Although the Valdez court held in that particular case that post-hypnosis testimony in which a hypnotized witness identifies for the first time a person he knew was already under suspicion is inadmissible in a criminal trial, the court stated that if adequate procedural safeguards have been followed, corroborated post-hypnotic testimony might be admissible if the probative value of the testimony outweighed its prejudicial effect. Id. at 1203.

In the instant case the district court permitted Rodney to testify as to his post-hypnotic memory because North Dakota is one of the jurisdictions which has held that hypnosis affects credibility but not admissibility. See State v. Brown, supra, 337 N.W.2d at 151. Federal courts are not bound by state law on this issue. Questions such as burden of proof, presumptions, competency, and privileges are generally questions of state law, but issues of admissibility of evidence are questions of federal law. Warner v. Transamerica Insurance Co., 739 F.2d 1347, 1351 n. 6 (8th Cir.1984); Sturm v. Clark Equipment Co., 547 F. Supp. 144, 145 (W.D.Mo.1982), aff'd, 732 F.2d 161 (8th Cir.1984). Quite simply, we do not view this issue as a competency question but as an evidentiary problem within the control of the district court and governed by federal law. See United States v. Valdez, supra, 722 F.2d at 1201.

GM urges this court to follow the Frye approach and in essence establish a per se inadmissibility rule or to adopt the "procedural safeguard" approach. Plaintiffs, on the other hand, advocate the credibility approach which is, for all practical purposes, a per se rule of admissibility of such evidence. We are reluctant to establish a per se rule of inadmissibility or admissibility and we decline to do so.

A per se rule would remove the question from the discretionary realm of the district court. If we were to establish a per se rule of inadmissibility, relevant, reliable testimony would in some instances be automatically disallowed and would hamper the truthfinding function of our system. A rule of per se inadmissibility is impermissibly broad and may result in the exclusion of valuable and accurate evidence in some cases. See State v. Hurd, supra, 432 A.2d at 94; State v. Beachum, supra, 643 P.2d at 252. On the other hand, if we adopt a per se rule of admissibility, in some circumstances, evidence that was unreliable because of the methods used in hypnosis and prejudicial because the jury may be overly influenced by testimony obtained from hypnotic recall would be admitted and that too would have an undesirable effect on our judicial system. State v. Iwakiri, 106 Idaho 618, 682 P.2d 571, 577 (1984). A per se rule of admissibility does not cure the risks of undue prejudice and jury confusion. Accordingly, we adopt a flexible rule on the admissibility of hypnotically enhanced testimony that enables the district court to determine the question on a case-by-case basis. We are satisfied that, if the hypnosis session is properly conducted in appropriate cases, the hypnotically enhanced testimony does not run afoul of the Frye test to the extent it is applicable.

We adopt a rule which requires the district court, in cases where hypnosis has been used, to conduct pretrial hearings on the procedures used during the hypnotic session in question and assess the effect of hypnosis upon the reliability of the testimony before making a decision on admissibility. The proponent of the hypnotically enhanced testimony bears the burden of proof during this proceeding. In addition, we adopt a version of the Hurd safeguards14 to the extent that the district court should consider whether and to what degree the safeguards were followed when making its determination that the hypnotically enhanced testimony is sufficiently reliable.15 Other factors the district court should take into account are the appropriateness of using hypnosis for the kind of memory loss involved, and whether there is any evidence to corroborate the hypnotically enhanced testimony. The district court must then determine whether in view of all the circumstances, the proposed testimony is sufficiently reliable and whether its probative value outweighs its prejudicial effect, if any, to warrant admission. Ultimately the district court must decide whether the risk that the testimony reflects a distorted memory is so great that the probative value of the testimony is destroyed.

By our ruling today we place this hypnosis evidentiary problem directly within the control of the district court. We think the better approach is for the district court and not the jury to make the preliminary determination of admissibility as is the case with other evidentiary questions. See FED.R.EVID. 104(a). It is our hope that this case-by-case method of determining the admissibility of hypnotically enhanced testimony will guard against the problems of hypnosis, especially undue suggestiveness and confabulation, but also allow for the inclusion of reliable refreshed memory which hypnosis can at times under certain circumstances produce. In sum, we hold that the district court should, before trial, scrutinize the circumstances surrounding the hypnosis session, consider whether the safeguards we have approved were followed and determine in light of all the circumstances if the proposed hypnotically enhanced testimony is sufficiently reliable and not overly prejudicial to be admitted.

Upon retrial of this case, the district court should determine before trial the admissibility of Rodney's hypnotically enhanced testimony using the approach set forth in our discussion. In the event the district court finds that as a result of the hypnotic session Rodney's testimony on the application of the brakes has been tainted, by suggestion or confabulation, and is unreliable, Rodney will not be permitted to testify as to that matter based upon his post-hypnosis recollection. However, where it is clear that other parts of his memory of the events of the accident were present before hypnosis, and remain uncontaminated by hypnosis, the court can determine that Rodney may testify as to those events. See State v. Seager, 341 N.W.2d 420 (Iowa 1983).

In the event that the trial court finds that Rodney's post-hypnosis testimony is reliable and that its probative value outweighs its prejudicial effect, Rodney will be permitted to testify to his present recollection without reference to the fact of hypnosis. Plaintiffs will not present evidence of his hypnosis in their direct case-in-chief. If GM wishes to impeach Rodney's testimony because he was hypnotized, it may cross-examine concerning the hypnosis and both parties may then bring in experts to testify to the problems and benefits of hypnosis as rebuttal to the other party's assertions. Then and only then would the videotapes of the actual hypnosis session be considered for possible admission, subject to the district court's Rule 403 balancing test and if deemed more probative than prejudicial, subject to a limiting instruction as discussed in Part I supra.

III. Exclusion of Videotapes, Photographs, and Summaries of Post-crash Tests

GM also argues that the district court abused its discretion when it excluded videotapes, pictures and summaries of the tests it performed on the accident car in 1983. The court permitted GM's witnesses Newsock and McCarthy to talk about the tests, but the court excluded videotapes, photographs and summaries because they would not have been particularly helpful, would have been cumulative, and possibly prejudicial.

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. Hale v. Firestone Tire & Rubber Co., supra, 756 F.2d at 1333; Randall v. Warnaco, Inc., 677 F.2d 1226, 1233 (8th Cir.1982). We have read the testimony of GM's experts who fully described the tests and results and find that the tapes, photographs and summaries would, indeed, have been cumulative. We find no abuse of discretion in the district court's decision to exclude this series of cumulative evidence under those circumstances. See Borough v. Duluth, Missabe & Iron Range Railway Co., 762 F.2d 66, 70 (8th Cir.1985).

GM also contends that the district court committed reversible error when it failed to instruct on comparative fault on either the strict liability or negligence counts. Since this case was tried the Supreme Court of North Dakota has decided at least four cases which have some bearing on this issue. See Mauch v. Manufacturers Sales & Service, Inc., 345 N.W.2d 338 (N.D.1984); Day v. General Motors Corp., 345 N.W.2d 349 (N.D.1984); Andersen v. Teamsters Local 116 Building Club, Inc., 347 N.W.2d 309 (N.D.1984); and Kaufman v. Meditec, Inc., 353 N.W.2d 297 (N.D.1984). Because of these new cases and our decision to remand for a new trial on other grounds, we need not decide whether it was error for the district court to refuse to instruct on comparative fault. However, upon retrial, the district court is free to reconsider its previous decision in light of the above cases.

As a final assignment of error GM asserts that there was not substantial evidence to support the verdict.16 GM maintains that plaintiffs' theory of the accident is a collection of false premises at odds with the physical facts and that the accident could not have occurred the way plaintiffs contend whether or not Rodney applied the brakes. Thus for purposes of our discussion of this issue GM's attack on the verdict is two-fold. First, GM contends that without Rodney's testimony that he applied the brakes, there was insufficient evidence to support plaintiffs' theory that a rear brake design defect caused the accident. Second, GM argues that even with Rodney's testimony that he applied the brakes, there was insufficient evidence to support plaintiffs' theory that a rear brake design defect caused the accident.

In view of our decision to reverse on the videotape issue and to leave to the district court, upon retrial, the question of whether Rodney's hypnotically enhanced testimony is admissible, we need not determine the hypothetical question of whether the evidence, without Rodney's post-hypnosis recollection, is sufficient to support the verdict. Consequently we do not address GM's argument that there is not substantial evidence to support the verdict without Rodney's hypnotically enhanced testimony. To do so would be premature and, perhaps, unnecessary speculation on our part.

As for our assessment of the evidence that was presented to the jury in this case we hold that there was substantial evidence from which the jury could infer that the accident occurred in the manner suggested by the plaintiffs. Plaintiffs' theory was that Rodney applied the brakes, that the rear brakes locked before the front brakes causing the car to spin 180? on the road and then roll off and flip into the ditch. The following evidence was before the jury: (1) Vivian's testimony that the car spun before it left the road; (2) Rodney's testimony; (3) accident reconstructionist Arndt's opinion based on the physical damage to the car that the car swapped ends before it left the road and that the spin occurred as a result of an application of the brakes which caused the rear wheels to lock up momentarily; and (4) expert witness Manos' opinion based on the position of tire marks that the rear wheels on the car locked up. Admittedly GM proffered evidence to the contrary, but in considering the sufficiency of this evidence we must assume all facts which the plaintiffs' evidence tends to prove and all reasonable inferences fairly deducible from the facts must be drawn in the plaintiffs' favor. SCNO Barge Lines, Inc. v. Anderson Clayton & Co., 745 F.2d 1188, 1192 (8th Cir.1984); see also Ozark Air Lines, Inc. v. Larimer, 352 F.2d 9, 11 (8th Cir.1965). Given this standard of review we cannot say that the evidence supporting the jury's verdict for the plaintiffs was not substantial.

Plaintiffs' only issue on cross-appeal is their contention that the trial court erred in refusing to allow an amendment to the complaint to include a claim against GM for punitive damages. The disposition of a motion to amend is within the sound discretion of the district court and to warrant reversal there must be some abuse of discretion. Foman v. Davis, 371 U.S. 178, 182, 83 S. Ct. 227, 9 L. Ed. 2d 222 (1962). In this case the district court denied the motion mainly because it was untimely in the sense, we assume, that it would have been disruptive to the trial schedule and potentially prejudicial to GM. We find no abuse of discretion in that ruling. However on retrial the district court's rationale for refusing the amendment is no longer as strong a consideration. So, while we see no abuse of discretion in the previous ruling, we direct the district court to reconsider the amendment on retrial. "An amendment can be proper after remand to the district court even if the claim * * * had not been presented to the district court in a timely fashion." City of Columbia v. Paul N. Howard Co., 707 F.2d 338, 341 (8th Cir.), cert. denied, --- U.S. ----, 104 S. Ct. 238, 78 L. Ed. 2d 229 (1983) (citations omitted).

We conclude that the admission of the videotapes of Rodney's hypnosis session during plaintiffs' case-in-chief without a proper cautionary instruction constituted prejudicial error. Accordingly we reverse and remand for a new trial at which time the district court is directed to conduct pretrial proceedings consistent with this opinion to determine the admissibility of Rodney's hypnotically enhanced testimony and to consider plaintiffs' request to amend their complaint to include a claim for punitive damages.