BOUDIN, Circuit Judge.
Carol Fusco was injured in a car accident and brought suit against General Motors, the car's manufacturer. A jury awarded Fusco $1 million in damages and General Motors has appealed, challenging rulings on evidence and discovery made by the district Judge. We affirm.
On December 15, 1986, Fusco was driving her car, a Chevrolet Chevette, near Pelham, New Hampshire. Her car suddenly left the roadway, slid across an ice-covered embankment, and hit a telephone pole somewhere along the front left side of the car. Fusco was injured.
Fusco brought suit against General Motors in state court in New Hampshire, claiming that a key component in the steering system--the front left "ball stud"--had broken from metal fatigue and caused the disaster.1 General Motors removed the case to federal district court and took the position that the ball stud had not been the cause of the accident but rather had fractured when the car hit the telephone pole. A jury trial, begun on July 7, 1992, resulted in an evenly divided hung jury, and the district court promptly ordered a second trial for November 16, 1992.
At the second trial Fusco offered eyewitness testimony that her car had abruptly veered off the highway and collided with a telephone pole. A state trooper who arrived first at the accident testified that the car was resting against the pole near the hinge pillar on the driver's side, a location between the door and the left front fender. Fusco offered two experts (Robert Walson and Carl Thelin) who, based in part on this testimony and their examination of the broken ball stud, concluded that metal fatigue had caused the stud to break, causing the steering apparatus to fail and the car to veer into the pole.
Walson, a metallurgist, testified that the surface of the broken ball stud taken from Fusco's car was characteristic of a fatigue, rather than an impact, fracture. He supported his opinion in several ways including his pretrial examination of the surface of the ball stud under a scanning electron microscope; he was fiercely cross-examined by General Motors about this examination. Thelin, an automotive engineer, testified that General Motors' design and quality control of the ball stud were inadequate. Based on partial reconstruction of the accident, he also challenged General Motors' argument that the telephone pole impact could have broken the ball stud.
General Motors' case included testimony from its expert Jerry Chiddister who reconstructed the accident based on his experience with many crash tests. In his view, the car had "sideslipped" into the telephone pole, causing the car to slide along the pole starting at the front left fender and ending with the pole lying next to the door hinge column. He opined that on its travel down the side of the car, the pole hit the front left tire and the impact broke the ball stud, a predictable occurrence given the estimated speed of the car. Had the stud broken before the car veered, Chiddister said that there would have been a heavy black tire mark on the road because the uncontrolled tire would have dragged as the car slid off course.
Kirk Ulman, another General Motors expert, testified that he had examined the ball stud itself. He explained why the location of the break (at the neck of the stud), the surface of the break (grainy with chevron marks), and other characteristics meant that impact and not fatigue was the cause. James Willis, who worked in General Motors' facility that made the steering gear, testified to quality control and the nature of simulated fatigue fractures. Ray Schultz, a metallurgist, confirmed Ulman's testimony on key points.
The jury rendered a verdict in favor of Fusco and awarded her $1 million in damages. General Motors then appealed. In its brief General Motors does not challenge the sufficiency of Fusco's evidence but confines itself to contesting several evidentiary and discovery rulings, rulings that can only be understood against the backdrop of the testimony already described. Although these claims of error are not frivolous, we do not think that any of them warrants further proceedings.
General Motors' first claim on appeal is that the district court erred in ruling, prior to the first trial, that two videotapes--the "driving tapes"--were inadmissible. The main tape made in 1992 has two parts. In the indoor part, Ulman used a car mounted on a lift to display the function of the ball stud and tie rod and showed how in this demonstration the connection between the stud and the tire wheel or axle had been altered in the test vehicle so that the stud could be released deliberately from inside the car.
In the outdoor part, filmed at a General Motors test track, Ulman drove the Chevette while Willis, sitting in the passenger seat, intentionally disconnected the tie rod from the tire wheel. The film showed that, when the left wheel finally separated from the rod, the wheel flopped out of alignment with the right wheel and dragged on the highway apparently creating a long black skid mark. The car did not veer out of control or hit the track barrier. The other tape, made in 1986, simply showed a similar test track demonstration with a different driver and passenger. Thus, there is no need for an independent Discussion of this tape.
When General Motors produced the tapes to Fusco in June 1992, shortly before the first trial, Fusco made a motion in limine to exclude them, arguing that the test track conditions did not duplicate the conditions that existed at the time of the actual accident. In an oral ruling on July 8, 1992, the trial Judge granted the motion. Although General Motors argues that the exclusion of the tapes was error, it did not seek to offer the tapes at the second trial.
Seizing on this omission, Fusco argues that General Motors has waived its right to argue that the exclusion was error. Fusco points to cases holding that an in limine ruling on evidence may not be reviewed on appeal unless the offer or objection is renewed when evidence is actually presented at trial. This court has made general statements to this effect, e.g., United States v. Reed, 977 F.2d 14, 17 (1st Cir. 1992), in cases where the in limine motion to exclude was denied and the opponent of the evidence failed to renew the objection at trial. We have found no case in this circuit, however, where an in limine motion to exclude was granted and the proponent of the evidence, by failing to renew the offer at trial, was found to have waived the issue.
Where an objection to evidence has been overruled in limine, it makes sense to require that the objection be renewed at trial. However definite the denial of the motion to exclude prior to trial, it is child's play for the opponent of the evidence to renew the objection when the evidence is actually offered; and requiring this renewal gives the trial Judge a chance to reconsider the ruling with the concrete evidence presented in the actual context of the trial. The only criticism one might offer of the requirement is that the Federal Rules of Evidence say nothing about a second objection,2 but any practiced trial lawyer knows that much of the law of evidence is not contained in these written rules.
On the other hand, where the motion in limine is granted, and the proponent of the evidence is told that the evidence will not be admitted, the situation is different. To require that the evidence be offered again at trial would certainly give the trial court a second chance, but doing so can hardly be described as easy: on the contrary, the proponent would have to engage in the wasteful and inconvenient task of summoning witnesses or organizing demonstrative evidence that the proponent has already been told not to offer. Indeed, in many cases the prior grant of the in limine motion would make it improper to call such witnesses without prior permission. All the proponent could do would be to line up the witnesses at trial and then ask permission.
Although a symmetrical rule may be preferable if all else is equal, all else is not equal here. Where a court rules in limine that certain evidence is excluded but the ruling is merely tentative or qualified, then the proponent might well have to offer the evidence at trial in order to preserve an appeal on the issue. Fed. R. Evid. 103(a). But where the pretrial proffer is adequate and evidence is excluded unconditionally by a pretrial order, then we think that the proponent has preserved the issue for appeal and (other circumstances being unchanged) need not bring the witness to court and proffer the evidence again at trial. See McQuaig v. McCoy, 806 F.2d 1298, 1301-02 (5th Cir. 1987).3 The result is the same here where the in limine order preceded the first trial because no one disputes that the same order governed the second trial.
Needless to say, most district Judges are very cautious about making a definitive ruling in limine that evidence will not be received at trial. Trial Judges know better than most that many issues are best resolved in context and only when finally necessary. But here, as happens from time to time, the trial Judge did rule definitively that the evidence would not be admitted. The proffer was adequate since the tapes were apparently available to the court. And no change in circumstances occurred after the in limine ruling that might have affected the controlling ground for exclusion or cast any doubt on the trial Judge's intention to abide by the original ruling.
Thus we turn to the merits of the ruling excluding the driving tapes. The oral ruling was terse ("You don't have to argue it. I'm not going to let it in.") but the district Judge had a written motion from Fusco, and a written response from General Motors, and made the ruling only after allowing both sides to argue their points orally. Fusco's main objection was that the taped scene on the test track did not adequately replicate the conditions of the accident. Merely to state the obvious, no one claimed that the accident had occurred when a jury-rigged cotter pin was pulled from the ball stud by a wire leading into the passenger seat. Another facially obvious difference is that the test car was driven by an experienced driver who expected the break to occur.
General Motors readily admitted that the conditions were not the same but argued that the tape was admissible to show general scientific principles and that the dissimilarities went to weight and not admissibility.4 It repeats this argument on appeal, adding that under Fed. R. Evid. 403 the burden was upon Fusco (as the opponent of the evidence) to show that prejudice substantially outweighed probative value. General Motors says that not only did Fusco fail to show that the dissimilarities were important but, in addition, General Motors' own experts would have said that the dissimilarities were not significant.
The problems raised by demonstrations of this kind are interesting and important. The test track replication shown on the driving tapes (this court was furnished with copies of the tapes) is vivid and pertinent: one sees, in a way that no words could capture, the tire wheel flip out of alignment and the tire then dragging on the track. The impression left is that such an accident would leave a tire streak, as claimed by General Motors, and that the car is more likely to flop and drag than to veer sharply off the road. A lay juror, asked whether a look at the tapes would be helpful, would likely answer yes.
The case law in this area is muddled, as one might expect, but the tendency of the court is to treat this class of demonstrative evidence more skeptically than would the lay juror. The concern lies not with use of tape or film (the issue would be largely the same if the jurors were taken to the test track for a live demonstration) but with the deliberate recreation of an event under staged conditions. Where that recreation could easily seem to resemble the actual occurrence, courts have feared that the jurors may be misled because they do not fully appreciate how variations in the surrounding conditions, as between the original occurrence and the staged event, can alter the outcome.
In such cases the solution of many courts, including this one, has been to call for substantial similarity in conditions, or to stress the great discretion of the trial Judge to exclude the evidence where similarity is not shown, or both. E.g.. Swajian v. General Motors, 916 F.2d 31 (1st Cir. 1990); see 1 J. Strong, McCormick on Evidence § 202 (1992) (collecting cases, a number of which involve General Motors). This case law largely undercuts General Motors' claim that the burden lay with Fusco to show undue prejudice; instead, courts have created a doctrine, predating and now loosely appended to Rule 403, that requires a foundational showing of substantial similarity in circumstances. Cf. Daubert v. Merrell Dow Pharmaceuticals, Inc., 125 L. Ed. 2d 469, 113 S. Ct. 2786 (1993) (reliability requirement for expert testimony held implicit in Rule 702).
Of course the concept of substantial similarity is a flexible one, and ought to be, for the benefits of the demonstration and the dangers of misleading the jury will vary greatly depending upon the facts. We think that the trial Judge enjoys great discretion in this area. But here the circumstances were not similar: as in Swajian, the test occurred in controlled conditions, on a test track with a driver expecting the occurrence, and with a doctored piece of equipment rather than one that actually broke. 916 F.2d at 36. General Motors does not seriously claim substantial similarity, despite its reference to the testimony its experts would have given.
Instead, General Motors says that the tapes here were designed not to recreate the accident but rather to illustrate general "scientific principles," presumably the behavior of a car with a disconnected ball stud. Admittedly, "when this [scientific principles] label is attached," courts often do ask not about similarity of conditions but only whether the test was properly conducted. McCormick, supra, at 866. We think it would be a great stretch to call the tapes an abstract demonstration of scientific principles, but the critical point is not one of labels. The issue for us is whether the demonstration is sufficiently close in appearance to the original accident to create the risk of misunderstanding by the jury, for it is that risk that gives rise to the special requirement to show similar conditions.5
Here the test track demonstration was rife with the risk of misunderstanding. Whatever Fusco's counsel or experts said to the jury about differing circumstances, the drama of the filmed recreation could easily overcome the logic of the distinctions. Our case is scarcely different than the recreation in Swajian involving a broken rear axle in a similarly staged demonstration by General Motors. We there affirmed the trial Judge's exclusion of such evidence, pointing to "the sound and broad discretion of the trial Judge" in policing such videotaped evidence. Id. at 36. We reaffirm here the principle and the result.
The second issue on appeal concerns the decision of the trial Judge to exclude an additional videotape prepared by General Motors after the first trial ended in a hung jury. This tape was prepared on October 2, 1993, and provided to Fusco on October 14, 1993, approximately three months after the first trial and one month before the second. The tape shows a slow motion, close-up impact fracture of a ball stud filmed at close range. On the tape one sees the stud bend and separate from the ball. There is no commentary.
After receiving the tape, Fusco immediately moved in limine to exclude it, together with a newly created "survey" of the accident site that General Motors had tendered. The district court allowed the survey to be used but ruled that the tape would not be admissible, stating:
With regard to the importance of the exhibit to the defendant, the Court cannot fathom how defendant could not present a complete case to the jury without the videotape as defendant has already done so at the first trial. Finally, the Court does not doubt that plaintiff would have a difficult time to formulate a proper response given the obvious lack of time before retrial. It is primarily for this reason the Court obviates defendant from submitting the videotape exhibit.
On appeal, General Motors does not spend much time explaining the significance of the tape, saying only that it was "to show how the part in question bends prior to and during an overload situation." Instead, General Motors argues with some force that the district Judge has no general authority to exclude exhibits merely because their last minute appearance will inconvenience or burden the other side. This exclusion, says General Motors, was not an exercise of the court's accustomed power over discovery but an interference with the right of counsel to offer evidence at trial.
Of course, a trial court could readily exclude a witness or exhibit if some previous order had set a deadline for identification and the proponent had without adequate excuse failed to list the witness or exhibit. But here the deadline for listing exhibits at the second trial did not expire until one week before trial, well after the tape was tendered. Nor is there any general rule prohibiting a party from offering new evidence at a second trial, although few cases address the issue. See Lauritzen v. Atlantic Greyhound Corp., 8 F.R.D. 237, 238 (E.D. Tenn. 1948), aff'd 182 F.2d. 540 (6th Cir. 1950). In fact, the district court here allowed General Motors to introduce at the second trial a newly created survey of the accident site.
Nevertheless, we think that the tape did implicate the district court's authority over discovery. The tape was not a stand alone exhibit: it made sense only in the context of expert testimony to be offered by General Motors. An expert or experts would have had to explain the tape's meaning and would have used it to bolster inferences drawn by the expert. Indeed, even now it is not clear to us precisely how the tape was expected to help General Motors' experts, although we will assume that it was either intended to counter plaintiffs' expert analysis or to make more plausible General Motors' account of how the ball stud retrieved from the accident was actually damaged.
The connection of the tape to discovery now comes into view. Fusco's discovery had included a request for production of "any and all . . . videotapes" taken by, for or for the benefit of Ulman, Willis, and "any and all expert consultants." If the tape had existed when General Motors first responded to this request, the tape would have had to be produced. We think that the duty of supplementation specified in Fed. R. Civ. P. 26(e) also required General Motors seasonably to produce any later developed videotape of any importance intended for use by its expert at trial.
The "duty to supplement" provisions are not as clear as they might be, and the authorities are surprisingly sparse. A party is required to supplement--that is, update--prior responses only in limited instances but they include, pertinently, responses (1) concerning "the substance" of the expert's testimony and (2) where new information exists and failure to amend would comprise "a knowing concealment." Fed. R. Civ. P. 26(e)(1), (2). We have read Rule 26(e) generously, in light of its dual purposes, the "narrowing of issues and elimination of surprise." Johnson v. H.K. Webster, Inc. 775 F.2d 1, 7 (1st Cir. 1985).6
We think that these clauses are broad enough to embrace the tape in this case. It was closely connected to the expert's testimony, it was not previously known to Fusco, and General Motors can hardly describe the tape as unimportant since it claims that the exclusion of the tape undermined the entire trial. Thus we think that the tape was covered by General Motors' obligation to produce under the discovery rules. In our view--and this is the final step in our reasoning--the discovery obligation carries with it the implicit authority of the district court to exclude such materials when not timely produced even if there was no rigid deadline for production.
There is no suggestion that General Motors delayed unduly if one counts only the brief delay between the creation of the tape and its tender to Fusco. But practical considerations suggest that the authority of the trial Judge must be broader: otherwise it would count as adequate supplementation to create a critical new expert exhibit a day before trial and tender it on the morning of trial. It is common experience that experts, like lawyers themselves, may increase their efforts as trial approaches, and we do not suggest any general bar to an exhibit created in good faith for the expert after initial discovery. But we do think that where a discovery request for the expert's materials has been made, the later attempt to add new exhibits designed for the expert's use at trial is subject to reasonable supervision by the trial Judge.
This circuit adopted this very principle in Thibeault v. Square D Co., 960 F.2d 239, 245 (1st Cir. 1992). The factual differences between that case and this one relate to the sound exercise of the authority to exclude and not its existence. We emphasize that our ruling in Thibeault and in this case rests on the existence of a specific discovery request and the explicit duty to supplement it in timely fashion. Absent some such obligation, we agree with General Motors that courts cannot ordinarily tell counsel not to offer evidence just because the other side will be surprised and disadvantaged. Once the court's authority to exclude here is conceded, its exercise in this case cannot be faulted. The impact tape was a technical experiment made known to Fusco barely a month before trial. To prepare for cross-examination might have required further discovery by Fusco of one or more of the defense witnesses, as well as additional preparation by Fusco's experts. The tape itself is pictures without words; even to determine the precise use General Motors planned to make of it would have required more information. The district court was entitled to find that the tape would have compromised Fusco's pretrial preparations and that the supplementation came too late to be "seasonable."
General Motors' remaining claim of error also relates to the district court's authority over discovery. At the first trial Fusco's experts made use of the broken ball stud recovered from Fusco's car and offered Walson's testimony based on his examination of the part under a scanning electron microscope. That part, in Fusco's possession, was available to General Motors prior to the first trial but it chose not to make such an examination of its own. In theory, study of the surface of the broken part might cast light on whether the part had broken from impact or fatigue.
Taking note of Walson's testimony in the first trial, General Motors in preparation for the second trial requested that the part, and an additional Pontiac ball stud used by Walson for comparison, be made available to it. General Motors wanted its own metallurgist to make a similar, non-destructive examination by scanning electron microscope.7 Fusco refused to produce the parts and General Motors moved to compel production. The district court wrote a ten-page order denying General Motors' motion.
In its order, the court said that General Motors knew as early as May 1991 that Fusco intended to offer evidence of examination by scanning electron microscope. The court also said that the new examination was intended as the predicate for new expert testimony by General Motors, testimony unknown to Fusco when General Motors' experts were deposed, and that this last minute addition imposed an unfair burden on Fusco on the eve of the second trial. The court also referred to General Motors' development of new evidence on the eve of the first trial, which the court had permitted over protest by Fusco. Finally, it expressed concern that General Motors sought to overwhelm plaintiff with its unlimited legal and financial resources.
There is no doubt here of the district court's authority to deny further discovery. As noted above, no automatic bar exists against new evidence at a second trial; but the discovery deadline had long since passed and the district court had no automatic obligation to reopen the discovery period. E.g., Dabney v. Montgomery Ward & Co., 761 F.2d 494, 498 (8th Cir.), cert. denied, 474 U.S. 904, 88 L. Ed. 2d 232 , 106 S. Ct. 233 (1985). The matter was one for the informed discretion of the trial Judge, and the breadth of that discretion in managing pre-trial mechanics and discovery is very great. E.g., Fashion House, Inc. v. K Mart Corp., 892 F.2d 1076, 1082 (1st Cir. 1989); Mack v. Great Atlantic & Pacific Tea Co., 871 F.2d 179, 186-87 (1st Cir. 1989). Measured against this standard of review, the district court's quite thorough explanation for denying production easily passes muster.
Our own review of the record does not suggest to us that General Motors' request was made in bad faith or reflected any attempt to abuse its financial power. Fusco after all was seeking $8 million, and defense counsel apparently perceived in the first trial that the testimony by Fusco's expert about the scanning electron microscope carried some weight. But it was primarily for the trial Judge to balance the factors, such as the timing of the request, General Motors' earlier knowledge, the possible burden on Fusco of preparing to confront a new expert, and the importance of the testimony. The balance struck here by the trial Judge was not an abuse of discretion.
1. It appears that the ball is a spherical object with a protruding stud; that the ball and stud together form part of the elaborate connection (via the tie rod and steering gear) between the tire wheel or axle and the steering wheel. If the stud breaks entirely, the tire wheel is no longer controlled by the steering wheel.
2. The Federal Rules of Evidence do not address in limine motions at all. Instead the rules require in general terms that, to preserve a ruling on evidence for appeal, the proponent of evidence make its substance known to the court and an opponent make known the objection to the evidence and the ground of the objection. Fed. R. Evid. 103(a).
3. Preserving the claim of error based on exclusion of evidence requires an adequate proffer, so that the trial and appellate courts know what evidence is at issue. Fed. R. Evid. 103(a). There may also be cases where a change in circum-stances, after the in limine ruling but before trial, might make it unreasonable for the proponent to rely on the solely in limine ruling to preserve the issue (a new basis for admissibility might arise or the court's initial reason for exclusion might be mooted).
4. General Motors does not claim that the indoor portion of the tape was admissible independently; and indeed most of the indoor portion was merely to lay the groundwork for the test. The indoor portion did show Ulman using a straw, in place of the stud, to show how the stud could be bent by an impact on the rear portion of the wheel, but General Motors does not claim that this brief sequence justified the tape, and it could easily have been replicated in court with a mock-up.
5. Scientific principles, when demonstrated in a fairly abstract way, are quite unlikely to be confused with the events on trial. The more troublesome cases, however, are ones like this one where some principles of some kind may be demonstrated but in a fashion that looks very much like a recreation of the event that gave rise to the trial.
6. The knowing-concealment clause does not require fraudulent intent; rather it is designed to protect a party who reasonably believes "that the change that has made an answer no longer accurate is known to his opponent or that it is a matter of no importance." Fortino v. Quasar Co., 950 F.2d 389, 396 (7th Cir. 1991).
7. General Motors proposed to have one of its first-trial experts conduct the examination but, when Fusco questioned his qualifications, General Motors proposed to have another expert whose qualifications were conceded participate in the examination and testify at trial. Whether one views the testimony as that of a first-trial expert or a new expert does not greatly affect the matter.