Johnson Controls, Inc. (JCI), appeals from the district court's order denying its motion for judgment as a matter of law or new trial after a jury returned a $1.2 million verdict for McKnight in this battery explosion case. JCI argues that it was entitled to judgment as a matter of law because the only evidence to support the verdict should not have been admitted by the district court. JCI asserts that the district court erred in admitting: experimental evidence not generated under conditions substantially similar to the accident conditions; the testimony and test results of McKnight's expert without requiring the proper foundation for his qualification as an expert; and speculative testimony from McKnight's expert about a key issue--the existence of a manufacturing defect. JCI also argues that the above alleged evidentiary errors coupled with the following additional errors require a new trial: allowing McKnight's expert to testify that the accident battery was unreasonably dangerous; denying JCI's request to examine McKnight's exemplar battery; admitting exhibits and testimony regarding JCI's prior testing of other companies' batteries; admitting evidence of another similar accident; denying JCI's attempt to supplement the record on the other similar accident; instructing the jury on the elements of McKnight's negligence claim; and failing to give two jury instructions JCI requested. After a thorough review which included reading the transcript of a week-long trial, we affirm.

On August 10, 1988, Randy C. McKnight, a 16-year-old Missouri resident, was permanently blinded in one eye when an automobile battery designed and manufactured by JCI, a Wisconsin company, exploded. On the day of the accident, McKnight borrowed a friend's car and drove it to his residence. He shut off the ignition for a few minutes while he went inside. When he returned to the car, it would not start. He popped the hood and twisted the battery cable on the negative terminal attempting to get a better connection. He tried the car again, and it still would not start. He again attempted to get a better connection by twisting the battery cable and repeated this procedure two or three times. When he twisted the battery cable on the last attempt, the battery exploded. The battery had been purchased at a local store six to eight months earlier.

McKnight filed this products liability action against JCI. The case eventually went to trial on McKnight's claims of strict product liability and negligent design or manufacturing defect. The parties agreed that the explosion occurred when a spark from the battery ignited hydrogen gas escaping from the battery. The issue for trial was whether the hydrogen explosion occurred as a result of a defect in the battery before it was shipped by JCI or from damage to the battery's vent caps caused by improper care for or use of the battery after it was purchased.

The case was tried the first time in March 1992. The jury could not reach a verdict, and a mistrial was declared. On retrial in September 1992, McKnight prevailed, and the jury awarded him $1.2 million in compensatory damages. The district court denied JCI's motion for judgment as a matter of law and its alternative motion for new trial. JCI appeals.

We review the district court's denial of a motion for judgment as a matter of law de novo using the same standards as the district court.1 Amerinet, Inc. v. Xerox Corp., 972 F.2d 1483, 1505 (8th Cir.1992), cert. denied --- U.S. ----, 113 S. Ct. 1048, 122 L. Ed. 2d 356 (1993). A motion for judgment as a matter of law presents a legal question to the district court and to this court on review: "whether there is sufficient evidence to support a jury verdict." White v. Pence, 961 F.2d 776, 779 (8th Cir.1992). We view the "evidence in the light most favorable to the prevailing party and must not engage in a weighing or evaluation of the evidence or consider questions of credibility." Id. Judgment as a matter of law is appropriate only when all of the evidence points one way and is "susceptible of no reasonable inference sustaining the position of the nonmoving party." Id.2

We apply a much more deferential standard in our review of a district court's denial of a motion for a new trial under Fed.R.Civ.P. 59(a). "The [district] court's decision will not be reversed by a court of appeals in the absence of a clear abuse of discretion." Lowe v. E.I. DuPont de NeMours & Co., 802 F.2d 310, 310-11 (8th Cir.1986) (citations omitted); see also 1 Steven Alan Childress & Martha S. Davis, Federal Standards of Review Sec. 5.08, at 5-69 (2d ed. 1992) ("The trial court's decision generally is committed to its discretion."). The key question is whether a new trial should have been granted to avoid a miscarriage of justice. See Beckman v. Mayo Found., 804 F.2d 435, 439 (8th Cir.1986) ("The district court can only disturb a jury verdict to prevent a miscarriage of justice.").

JCI's argument for judgment as a matter of law focuses on the testimony of Dr. Jacobson, McKnight's expert witness. JCI argues the district court should have excluded his testimony on a number of grounds and that when the testimony is properly excluded from the record, this court must enter judgment as a matter of law because there is no evidence to support the jury verdict. Alternatively, JCI argues that the errors in admitting Jacobson's testimony or any of a number of other errors the district court committed, at a minimum, entitle JCI to a new trial.

JCI first argues that the district court erroneously allowed McKnight to introduce evidence of experiments conducted by his expert, Dr. Jacobson, without making any showing that the experiments were conducted under conditions substantially similar to the conditions existing at the time of the actual accident. JCI argues that this evidence should have been excluded and that without this evidence JCI was entitled to judgment as a matter of law. Alternatively, JCI argues that the erroneous admission of this evidence requires a new trial.

Jacobson performed his experiments on an Alpha 3 400/36 Group 24C automobile battery manufactured by JCI, the same type of battery which exploded. He performed two tests on this exemplar battery: a "coordinate test" and a "leak test." The "coordinate test" purported to provide exact measurements of both the perimeter of the sealing surface of the vent caps that plug the battery wells containing the hydrogen gas and the perimeter of the walls of the corresponding battery vent wells the vent caps are supposed to seal. (Tr. Vol. III at 71-87.) Jacobson acquired these measurements from another laboratory, which utilized a "coordinate measuring device" to take the measurements on the exemplar battery.3 (Id. at 71.) The purpose of the test was to determine how well the vent caps fit the wells, which would indicate how well the seal created between the two surfaces contained the hydrogen gas.

The "leak test" purported to determine whether any gas in fact could escape the battery's sealing mechanisms. (Id. at 88-104.) Jacobson used a "sniffer" device to detect whether any gas could escape from the battery. (Id. at 91.) In conducting the leak test, Jacobson used helium gas instead of hydrogen gas, which is normally contained in the battery. (Id. at 88.) Jacobson also drained the electrolyte fluid from the battery and pumped helium into the battery, creating pressures inside the battery ranging from one pound per square inch (psi) to five psi during the testing. (Id. at 100-101.) Jacobson also plugged the holes in the battery vent caps to obtain some of his measurements. (Id. at 92-93.) The evidence is unrefuted that at the time of the accident, the battery was full of electrolyte fluid, operating at zero psi (because the car engine was off), and that the vent caps would not have been plugged.

We have previously made the following observations about the admissibility of experimental tests in products liability cases:

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. 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.

Champeau v. Fruehauf Corp., 814 F.2d 1271, 1278 (8th Cir.1987) (internal quotations and citations omitted) (emphasis added). This foundational requirement for the admission of experimental evidence is very well established in this circuit and others. See, e.g., Cowens v. Siemens-Elema AB, 837 F.2d 817, 820 (8th Cir.1988); Hale v. Firestone Tire & Rubber Co., 820 F.2d 928, 932 (8th Cir.1987) (Hale II ); see also Gilbert v. Cosco Inc., 989 F.2d 399, 402 (10th Cir.1993) (burden is on proponent to demonstrate similarity of circumstances prior to admission of evidence). However, we also have found that where the experimental tests do not purport to recreate the accident, but instead the experiments are used to demonstrate only general scientific principles, the requirement of substantially similar circumstances no longer applies. Champeau, 814 F.2d at 1278.

JCI first contends that the conditions under which Dr. Jacobson conducted both experiments were not substantially similar to those involved in the accident because the battery Dr. Jacobson used to conduct the experiments was not substantially similar to the battery involved in the explosion.4 This argument focuses largely on the integrity of the "coordinate test," as JCI contends that the exact measurements vary from battery to battery and that the exemplar and the accident battery were made from different molds. JCI next contends that Jacobson's experiments were not performed under conditions similar to those at the time of the accident because he operated the battery under different conditions than those that existed at the time of the accident. In particular, JCI contends that the "leak test" was not performed under conditions substantially similar to the accident conditions because Jacobson used helium gas instead of hydrogen gas, drained the electrolyte fluid from the battery, substantially increased the pressure inside the battery, and plugged the vent cap holes.

McKnight first argues that both of Jacobson's experiments were offered only to illustrate the scientific principles and rationale for Jacobson's conclusions and that, therefore, the experiments did not need to be conducted under substantially similar conditions. In response to JCI's pretrial motion in limine seeking to prevent the introduction of Jacobson's experiments, McKnight specifically stated that the experiments' test results "are not intended to serve as a re-enactment of the explosion, but merely to demonstrate the physical principles of the defect asserted by plaintiffs." (Jt.App.Docket No. 321 at 8.) McKnight also points out that Jacobson never testified that he recreated or even attempted to recreate the accident.5 McKnight alternatively argues that even if the experiments needed to be performed under substantially similar conditions, that test is satisfied in this case.

Whether the evidence from Jacobson's experiments was offered and admitted as a reconstruction of the accident or only to demonstrate scientific principles will determine the foundational standard McKnight needed to satisfy. In this case, like many others, the distinction between evidence offered as a reconstruction of the accident and evidence offered to demonstrate scientific principles is very difficult to draw. See Gilbert, 989 F.2d at 403. There is no bright line distinguishing these two categories of evidence.

The First Circuit recently concluded that "the critical point is not one of labels." Fusco v. General Motors Corp., 11 F.3d 259, 264 (1st Cir.1993). Instead, the key question 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." Id.; see also Gilbert, 989 F.2d at 403. ("The more blurred that distinction [between demonstration and reconstruction] becomes, the greater the risk for prejudice."). The First Circuit went on to observe that:

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.

Fusco, 11 F.3d at 264 n. 5. The First Circuit concluded that the demonstration was "rife with the risk of misunderstanding" and that the trial court acted properly in excluding it from the record. Id. at 264.

We agree with the First Circuit that experimental evidence falls on a spectrum and the foundational standard for its admissibility is determined by whether the evidence is closer to simulating the accident or to demonstrating abstract scientific principles. See id. The closer the experiment gets to simulating the accident, the more similar the conditions of the experiment must be to the accident conditions. Here, we find this to be one of the "troublesome cases ... where some principles of some kind may be demonstrated but in a fashion that looks very much like a recreation of the events that gave rise to the trial." Id. at 264 n. 5. The experiments were conducted on the same type and make of battery as the accident battery, and the experiments were used to explain what probably happened during the accident. These tests clearly were not limited to a demonstration of scientific principles in the abstract. As such, we reject McKnight's assertion that the evidence from the experiments Jacobson conducted was offered only to demonstrate general scientific principles and that the evidence did not have to satisfy the substantial similarity requirement.6

During the trial, the trial court admitted the experimental evidence over the objection of JCI without specifically requiring a showing of substantial similarity or making findings concerning the matter. However, after trial, and in response to JCI's motion for new trial or judgment as a matter of law, the trial court made the following finding:

Regarding the alleged error in admitting evidence of experimental testing that Jacobson performed on the exemplar battery, the Court is satisfied that that battery was substantially similar to the accident battery, that the experimental testing was conducted under conditions substantially similar to the actual conditions of the accident, and that plaintiff laid a proper foundation.

(Jt.App. Docket No. 379 at 3-4.) Hence, the trial court, though belatedly, indicated that the evidence satisfied the substantial similarity test. We review that determination for abuse of discretion.

Chief Judge Arnold has described review for abuse of discretion as follows:

[W]hen we say that a decision is discretionary, or that a district court has discretion ..., we do not mean that the district court may do whatever pleases it. The phrase means instead that the court has a range of choice, and that its decision will not be disturbed as long as it stays within that range and is not influenced by any mistake of law. An abuse of discretion, on the other hand, can occur in three principal ways: when a relevant factor that should have been given significant weight is not considered; when an irrelevant or improper factor is considered and given significant weight; and when all proper factors, and no improper ones, are considered, but the court, in weighing those factors, commits a clear error of judgment.

Kern v. TXO Prod. Corp., 738 F.2d 968, 970 (8th Cir.1984). The trial court here supplied little, if any, reasoning for its finding that both the experimental battery and the experimental testing conditions were substantially similar to those involved in the accident. As such, we are left without any indication as to what factors the trial court used in reaching its decision. We are however willing to assume that the experienced trial court judge considered all the proper factors and no improper factors and accordingly we will determine whether the trial court committed a "clear error of judgment" in admitting either of these two tests.

First, we review the trial court's finding that the batteries were substantially similar, which we noted above focusses on the integrity of the "coordinate test." JCI argues that McKnight failed to introduce any evidence to establish a foundation of substantial similarity. We disagree. The evidence is undisputed that Jacobson performed his tests on the same make and type of battery manufactured by JCI as the accident battery. Jacobson specifically testified that the batteries were substantially similar. (Tr.Vol. III at 68.) Moreover, in closing argument, JCI admitted that the exemplar battery was substantially similar. (Tr. Vol. V. at 160-161.) On this record, we cannot say that the district court committed a clear error of judgment in finding that the exemplar battery was substantially similar to the accident battery.

JCI argues that under our decision in Cowens, McKnight was required to make an additional showing to establish the substantially similarity of the batteries. In Cowens, the plaintiff argued that the trial court abused its discretion in refusing to admit evidence from experiments conducted on defendant's wire mesh screens that may have been manufactured several years after the wire mesh screens allegedly involved in the accident. We found that on the record presented in that case, the disparate dates of manufacturing may have been significant and held that the trial court did not abuse its discretion in excluding the evidence on that basis. 837 F.2d at 820-21.

JCI contends that under Cowens, McKnight needed to establish that the manufacturing dates of the exemplar and the accident battery were very close in time or, more particularly, that the exemplar battery was from the same manufacturing mold as the accident battery. We disagree. We did not hold in Cowens that the trial court must refuse to admit experimental evidence unless the plaintiff established similar manufacturing dates or that the exemplar came from the same manufacturing mold as the product involved in the accident. We held only the trial court did not abuse its discretion in refusing to admit the evidence on those grounds. Moreover, unlike this case, the defendant in Cowens put on evidence that the difference in manufacturing dates would affect the measurements and data gained from the experimental tests because the dimensions of the subsequently manufactured mesh screens would differ from the ones involved in the accident.7 Furthermore, unlike this case, in Cowens, we agreed with the trial court that the experimental evidence was of little probative value, which served as an additional reason to exclude it. We cannot say that the coordinate test in this case was of little probative value. Hence, we do not find that Cowens imposed an additional foundational requirement upon McKnight in this case.

The trial court's finding that the experimental conditions Jacobson utilized in conducting the "leak test" were substantially similar to the conditions at the time of the accident is much more problematic. As noted above, Jacobson used an exemplar battery that was substantially similar, if not nearly identical, to the accident battery. However, Jacobson significantly altered the operating conditions in the battery when he performed his leak test. Jacobson drained the electrolyte fluid from the battery, increased the pressure inside the battery up to 5 psi, and plugged the vent cap holes in conducting the leak test. The record is essentially unrefuted that at the time of the accident, the battery was full of electrolyte liquid, the pressure inside the battery was 0 psi, and the vent caps were not plugged.

These differences were significant. JCI introduced evidence that the electrolyte fluid in the battery acted as part of the sealing mechanism to prevent leaks of hydrogen gas. (Tr.Vol. IV at 23.) Jacobson conducted his leak tests at pressures that far exceeded the accident conditions. The pressure caused the battery to bulge and stretch like a balloon, (Tr.Vol. III at 194), and could have caused the leaks that Jacobson then detected with his "sniffer." Given these significant differences between the accident conditions and the experimental conditions, we conclude that the district court committed a "clear error of judgment" in finding that the accident conditions and the conditions under which Jacobson performed the leak test were substantially similar. We emphasize that while the batteries met the test, the conditions did not.

McKnight argues that our cases demonstrate that the trial court acted within its discretion in finding that the experimental conditions were sufficiently similar to the accident conditions. In particular, McKnight relies on our decision in Hale II, where we found that the trial court did not abuse its discretion in admitting a videotape depicting a tire explosion and its effect on the mannequin crouched next to it even though the tire was inflated to 20 psi greater than the tire involved in the accident. 820 F.2d at 931-32. The defendant in Hale II argued that the increased psi exaggerated the force of the explosion and was unduly prejudicial.

Our decision in Hale II is easily distinguished from this case. In Hale II, we noted that while the tire was inflated to a greater pressure than the tire involved in the accident, the videotaped reenactment also used a smaller tire which would have had a smaller explosion force than the accident tire. We determined there was no error in admitting the evidence only after we were satisfied that the trial court offset the psi differential against the tire size differential before determining that the conditions between the recreation and the accident were substantially similar. Hale II, 820 F.2d at 932. In this case, however, there were no mitigating factors or offsetting factors to counterbalance against the significant departures Jacobson made from the accident conditions. Our decision in Hale II simply does not support the trial court's finding that the conditions of the leak test were substantially similar to the conditions involved in the accident. We find that the trial court committed a clear error of judgment and, accordingly, abused its discretion in finding that the conditions of the leak test were substantially similar to the conditions at the time of the accident. Hence, the trial court erred in admitting the leak test in this case over JCI's foundational objection.8

Not all errors in admitting evidence, however, are grounds for reversal. Reversal is not required when the erroneous admission of evidence is harmless. Lovett v. Shuster, 633 F.2d 98, 103 (8th Cir.1980); see also Fed.R.Civ.P. 61. "For there to be harmless error, the improperly admitted evidence cannot have had a 'substantial influence' on the jury's verdict." Lewis v. Sheriff's Dep't for City of St. Louis, 817 F.2d 465, 467 (8th Cir.1987) (quoting United States v. Iron Shell, 633 F.2d 77, 87 (8th Cir.1980), cert. denied, 450 U.S. 1001, 101 S. Ct. 1709, 68 L. Ed. 2d 203 (1981)). Stated another way, the error is harmless only if "the error did not influence or had only a slight influence on the verdict." United States v. DeAngelo, 13 F.3d 1228, 1233 (8th Cir.1994) (quotation omitted).

After careful study, we cannot say that the improperly admitted leak test evidence had a "substantial influence" on the jury's verdict. The harmful effect of the improperly admitted evidence is reduced where there is substantial evidence in the record beyond the evidence improperly admitted to support the jury's decision. See Brown v. LaCreek Elec. Ass'n Inc., 939 F.2d 623, 625 (8th Cir.1991). Here, the coordinate test, which we have concluded was properly admitted, provides substantial evidence to support the jury's decision. The coordinate test indicated that there were gaps in the fit of the caps to the battery wells which were large enough to allow significant amounts of hydrogen gas to escape the battery. The coordinate test alone supports Jacobson's theory of the cause of the explosion.

Moreover, in closing argument, McKnight relied on Jacobson's theory of the explosion with very little reference to the experimental leak test he used to reach those conclusions. We find only two brief references to the leak test in McKnight's entire closing argument. (Tr.Vol. V. at 144, 146-47.) In making those references, McKnight never argued that the leak test reconstructed the accident or otherwise demonstrated how the accident happened. McKnight focussed his closing argument instead on the results from the coordinate test and on the other evidence he introduced. McKnight argued to the jury that the coordinate test established that there was a gap in the fit between the battery wells and the cap large enough to cause the explosion, (Tr.Vol. V. at 144-45, 173-174); that JCI's own earlier tests had established the explosion theory that Jacobson advocated in this case, (id. at 142-143); that JCI never did any testing for leaks at the factory in spite of their knowledge of the possibility of explosions, (id. at 143); and that JCI's experts' testimony about the "interference fit" of the vent caps on the wells could not be believed, (id. at 145-46, 174-75).

We cannot find that the improperly introduced evidence from the leak tests was a focal point or a substantial portion of McKnight's case. Because of the testimony from the plaintiff's expert which admitted the differing conditions, and the strong countervailing evidence from JCI which pointed out the probable resulting effects of the differences between the accident and the leak test experiment conditions, we cannot say with confidence that the trial court's error affected JCI's substantial rights. Accordingly, under Rule 61 and our case law, we conclude that the trial court's error in admitting the leak test evidence was harmless and, therefore, not a basis for reversal in this case.

JCI next argues that the trial court erred in allowing Jacobson to testify as an expert without establishing a proper foundation for his qualification as an expert or for the scientific basis of his conclusions under Federal Rules of Evidence 702 and 703. McKnight argues that JCI failed to raise these issues at trial and, even assuming they were raised, that there was more than sufficient foundation for Jacobson's expert testimony.

After trial in this case, the Supreme Court of the United States addressed the standard for admitting expert testimony under Rule 702 in Daubert v. Merrell Dow Pharmaceuticals, Inc., --- U.S. ----, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993). JCI and McKnight agree that Daubert now controls the question of admissibility of expert testimony.

In Daubert, the Court held that the widely applied test for admitting expert testimony only if it was "generally accepted" within the scientific community was not consistent with Federal Rule of Evidence 702, or the other rules addressing expert testimony. --- U.S. at ----, 113 S. Ct. at 2794. The Court went on to point out the methods and standards that should govern the trial court's determination of whether to admit expert testimony. The Court held that the trial court should first determine under Rule 104(a) whether the expert's testimony rests on a reliable foundation. Id. --- U.S. at ----, 113 S. Ct. at 2796. The key question in that regard should be "whether the reasoning or methodology underlying the testimony is scientifically valid and ... whether that reasoning or methodology properly can be applied to the facts in issue." Id.9

Both parties contend that Daubert supports their position. McKnight argues that Daubert makes expert testimony more readily admissible because it is no longer subject to the "general acceptance" standard. JCI argues that Daubert requires the trial judge to be a gatekeeper for expert evidence by making a Rule 104(a) inquiry to determine whether the foundation has been satisfied and that the district court failed to make these basic findings on the foundation for Jacobson's testimony.

We need not reach this issue, however, because JCI failed to object to Jacobson's testimony on the basis that he was not qualified as an expert or that he lacked a scientific basis for his opinions. "Without an objection and a proper request for relief, the matter is waived and will receive no consideration on appeal absent plain error." Owen v. Patton, 925 F.2d 1111, 1115 (8th Cir.1991). JCI argues strenuously that it did object to Jacobson's qualifications as an expert and the scientific basis for his opinions. The record does not support JCI's assertions.

JCI argues that it objected to both Jacobson's qualifications and the scientific basis for his opinions in the following statement:

If the Court please, the objection to this testing and the opinions concerning this testing goes to the foundation for the testimony. The witness has said that the two batteries are substantially similar. And I submit to the Court that when the witness is attempting to prove what he's attempting to prove with these tests, and that is the exact fit of plugs and caps in the holes, that the foundation must be more than that. What he's trying to do here is to use a test on a battery, for which a foundation hasn't been laid sufficiently, to prove the ultimate issue in this case, and that is a defect that gaps exist between the plugs and the walls. And I submit to the Court that a sufficient foundation to use this battery in that way has not been laid in this case.

(Tr.Vol. III at 73.) This objection fails to raise any question about the scientific validity of the principles and methodology underlying his testimony. See Daubert, --- U.S. at ----, 113 S. Ct. at 2797. Likewise, this objection fails to raise any question about whether his tests were the type relied upon by experts in the field to indicate JCI was objecting to his testimony under Rule 703. An objection on one ground does not allow a party to argue on appeal that the evidence should have been excluded on different grounds. Hale v. Firestone Tire & Rubber Co., 756 F.2d 1322, 1333-34 (8th Cir.1985) (Hale I ).

To the extent that JCI is arguing that the district court was required to exercise its gatekeeping authority over expert testimony without an objection, we disagree. We have previously rejected a similar argument in United States v. Bartley, 855 F.2d 547, 552 (8th Cir.1988). We find that the principles of Bartley apply equally in this case, and accordingly we reject JCI's argument as unpersuasive.

The only remaining question is whether the district court committed plain error in allowing this evidence into the record. JCI has not argued for, nor do we find, plain error.

JCI next argues that the trial court impermissibly allowed Jacobson to offer speculative testimony that there was a manufacturing defect in the battery. JCI asserts that Jacobson testified that "[t]here could very well be a manufacturing defect," (Tr. Vol. III at 107), and that it was merely possible that a manufacturing defect occurred and that such testimony should not have been allowed, (id. at 156). McKnight claims again that JCI failed to object to this testimony and has thus waived the argument. JCI argues that it did object to the testimony when it moved to strike his testimony at its conclusion and when it made a motion for a new trial or judgment as a matter of law and asserted these issues.

The testimony that JCI claims is unduly speculative came in both during Jacobson's direct examination (id. at 107), and during JCI's cross-examination of him, (id. at 156). We first point out that the only objection JCI points to as preserving the error for appeal came at the close of Jacobson's direct testimony, before the alleged speculative testimony on cross-examination came into the record. (Id. at 146-48.) Moreover, a review of JCI's objection shows that JCI never even addressed the speculative nature of the testimony. This objection clearly and simply addressed only the causal connection between his testimony and the ultimate issue McKnight needed to prove. (Id.)

Moreover, JCI's argument essentially concedes that JCI made no contemporaneous objection. Federal Rule of Evidence 103(a)(1) requires objections to evidence to be "timely." "The rule is well settled in this circuit that for an objection to be timely it must be made at the earliest possible opportunity after the ground of objection becomes apparent, or it will be considered waived." Terrell v. Pollard, 744 F.2d 637, 638-39 (8th Cir.1984). If the ground for the objection becomes apparent while the witness is testifying, a subsequent motion to strike the testimony after the witness finishes does not preserve the issue for appeal. Id. at 639. We find that the grounds asserted for objecting to Jacobson's "speculative" testimony on direct examination were apparent at the time Jacobson provided the testimony and, thus, the motion to strike was not a timely objection, and no error can be predicated on the court's determination of it.

We must add, however, that even if we were to assume, arguendo, JCI had preserved this error for review, we would find that the trial court committed no error in admitting the testimony. In Loudermill v. Dow Chem. Co., 863 F.2d 566, 570 (8th Cir.1988), we specifically addressed Dow's argument that expert testimony should not have been admitted because it was based on speculation and conjecture, by noting that the factual basis of an expert's testimony generally goes to credibility of the testimony, not admissibility, and that the party opposing the testimony can attack the factual basis on cross-examination. We noted that the expert opinion should be excluded as too speculative only where it is "so fundamentally unsupported that it can offer no assistance to the jury." Id. Likewise, an expert may give opinions if there are sufficient facts already in evidence to take the opinion out of the realm of mere speculation or guesswork. Hurst v. United States, 882 F.2d 306, 311 (8th Cir.1989) (citing Faries v. Atlas Truck Body Mfg. Co., 797 F.2d 619, 623 (8th Cir.1986)).

We find that Jacobson's testimony that the accident was caused by a manufacturing defect was drawn from the results of his experiments already introduced into evidence. Hence, there was enough evidence in the record to allow the court to admit his opinion on manufacturing defect. Moreover, JCI extensively cross-examined Jacobson on his conclusions, and that is when he admitted that they were only possible. (Tr.Vol. III at 156.) We find that the district court properly left the ultimate determination of the credibility of his conclusions for the jury to decide.

JCI next argues that the trial court erred in allowing Jacobson to testify that the battery was "unreasonably dangerous" at various points in his testimony. (Tr., Vol. III, at 107, 111). JCI contends that Jacobson essentially testified to a legal conclusion on an ultimate issue. McKnight argues that under Federal Rule of Evidence 704(a) and this court's case law, experts are specifically allowed to testify on ultimate issues when that testimony is based on sufficient facts already in evidence.

We have discussed the admissibility of "ultimate issue" testimony at length in this circuit. "The admissibility of opinion testimony is generally committed to the sound discretion of the trial court." Kostelecky v. NL Acme Tool/NL Indus., Inc., 837 F.2d 828, 830 (8th Cir.1988). Although Rule 704(a) provides that expert opinions are not inadmissible merely because they embrace an ultimate issue in the case to be decided by the jury, not all expert testimony is admissible. United States v. Whitted, 11 F.3d 782, 785 (8th Cir.1993). The touchstone for the admissibility of expert testimony is whether it will assist or be helpful to the trier of fact. See, e.g., id.; Johnson Group, Inc. v. Beecham, Inc., 952 F.2d 1005, 1007 (8th Cir.1991); Williams v. Wal-Mart Stores, Inc., 922 F.2d 1357, 1360 (8th Cir.1990); Kostelecky, 837 F.2d at 830; Hogan v. American Telephone & Telegraph Co., 812 F.2d 409, 411 (8th Cir.1987); Strong v. E.I. DuPont de Nemours Co., 667 F.2d 682, 685 (1981).

We addressed a very similar issue in Strong where the district court excluded testimony attempting to address the question of "unreasonably dangerous." We specifically noted that "the question of whether the lack of warnings rendered the ... products unreasonably dangerous is not the kind of issue on which expert assistance is essential for the trier of fact. The jury was capable of drawing its own inferences from the available evidence." Strong, 667 F.2d at 686.

While in our view Jacobson's testimony that the battery was "unreasonably dangerous" was of little, if any, assistance to the jury in this case, (Jacobson already had testified about the probability of leaks and the combustibility of leaking hydrogen gas, and the presence of an ignition source in the sparks generated by twisting the battery cables) and came close to telling the jury what result to reach, we cannot conclude that the trial court abused its discretion in admitting the opinion. Our review of this issue would have been easier had the trial court, when faced with the objection, taken the time to place on the record the reasons it felt that Jacobson's opinion would be helpful to the jury, and we would encourage trial courts to do so in the future.

JCI next contends that the trial court erred when it denied JCI's request to examine the exemplar battery Jacobson used in conducting the experiments. The district court found that because JCI made the request to examine the exemplar after the close of discovery and because JCI had made no showing of substantial need for the discovery, the request would be denied.

"The trial court has broad discretion to decide discovery motions and we will not disturb such a ruling absent a gross abuse of discretion affecting the fundamental fairness of the trial." Pavlik v. Cargill, Inc., 9 F.3d 710, 714 (8th Cir.1993). The district court pointed out that JCI already had a similar exemplar of its own (or could easily acquire one) and could thereby easily obtain essentially the same information it was seeking from an examination of McKnight's exemplar. The denial of the request for additional discovery did not substantially affect JCI's trial preparation. Moreover, discovery had closed, and the district court had no obligation to reopen discovery. Fusco, 11 F.3d at 267 ("discovery deadline had long since passed and the district court had no automatic obligation to reopen the discovery period."). We find no gross abuse of discretion affecting the fundamental fairness of the trial on this point.

JCI next argues that the district court erred in admitting plaintiff's Exhibits 92, 94, 96, and 108, which detailed the results of JCI's own experiments on batteries manufactured by its competitors some years earlier. JCI essentially argues that these exhibits contained experimental evidence and McKnight failed to show that the experiments were conducted under substantially similar conditions.

McKnight contends that the exhibits were both offered and admitted only to rebut JCI's assertion that no one else in the world had ever come up with the explosion theory that Jacobson hypothesized. (Tr. Vol. III at 119; Tr. Vol. IV at 151.) McKnight argues that these exhibits simply show that JCI itself had contemplated this same theory in conducting its examination of other companies' batteries.

The parties addressed this issue at length at a bench conference during the trial. (Tr.Vol. III at 112-124.) McKnight specifically stated a number of times that the exhibits were being offered only to demonstrate the scientific principle Jacobson relied upon in reaching his conclusions. (Id.) The district court appeared to admit the exhibits on this basis. (Id. at 123-24.) The district court specifically instructed McKnight's counsel on the limited scope for which they could be used. (Id. at 121.) McKnight's counsel stayed within those limits during his examination of Jacobson and in closing argument to the jury. (Id. at 125-26, 127, 129; Tr. Vol. V at 142-43.) There was never any indication or argument that the experiments recorded in the exhibits were reconstructions of this accident. Quite to the contrary, McKnight's counsel was very careful to point out that underlying experiments were conducted on different batteries with different structures. (Tr.Vol. III at 125-26, 127, 129.)

As we have noted earlier, where the experimental tests do not purport to recreate the accident and instead the experiments are used to demonstrate general scientific principles, the requirement of substantially similar circumstances no longer applies. Champeau, 814 F.2d at 1278. Hence, we find unpersuasive JCI's argument that the district court erred in admitting these documents without a showing of substantial similarity.

JCI next argues that the district court erred by admitting the deposition testimony of Donna Wise, who was injured in another battery explosion accident involving a JCI battery. JCI argues that the district court erred both in failing to comply with Federal Rule of Evidence 804(b)(1) in admitting this hearsay evidence and in failing to require McKnight to show substantial similarity between the accidents.

Rule 804(b)(1) requires that in order for testimony from a prior hearing or deposition to be admissible, the party against whom it is offered must have had the motive and opportunity to develop or cross-examine that testimony. United States v. Salerno, --- U.S. ----, ----, 112 S. Ct. 2503, 2507, 120 L. Ed. 2d 255 (1992). JCI contends that McKnight failed to establish that JCI had the motive to cross-examine Wise's deposition testimony. The district court overruled this objection and found that McKnight had established JCI's motive to cross-examine Wise's statements. (Tr.Vol. II at 62.) We find no abuse of discretion. JCI had every incentive to cross-examine the testimony of Wise who, like McKnight, was suing for injuries from a JCI built battery explosion.

JCI's second argument is that McKnight failed to show that Wise's accident was substantially similar to his own. Evidence of other accidents is admissible if the other accident is substantially similar to the accident at issue in the case. Burke v. Deere & Co., 6 F.3d 497, 506 (8th Cir.1993), cert. denied, --- U.S. ----, 114 S. Ct. 1063, 127 L. Ed. 2d 383 (1994); Drabik v. Stanley-Bostitch, Inc., 997 F.2d 496, 508 (8th Cir.1993). Evidence of other accidents may be relevant to demonstrate the defendant's ability to correct known defects, the magnitude of the danger involved, the product's lack of safety for its intended use, notice of the existence of a defect, standard of care, or causation. Burke, 6 F.3d at 506; Drabik, 997 F.2d at 508. The decision whether to admit such evidence is within the sound discretion of the trial court. Drabik, 997 F.2d at 508.

JCI argues that McKnight failed to present any evidence whatsoever of substantial similarity. McKnight points out that the district court heard unreported argument and unreported in-chambers testimony about the substantial similarity of the accidents prior to making its on-the-record ruling. JCI contends that this evidence cannot be argued on appeal because it is not a part of the official record.

While we normally would agree with JCI that evidence not specifically reported in the record should not be considered on appeal, we decline to do so in this case. The record is very clear that the district court held an in-chambers, but unreported, proceeding on the question of the admissibility of the Wise deposition and other deposition testimony prior to hearing arguments and making a ruling on the record. The district court and the parties referred to this in-chambers proceeding repeatedly during the on-the-record proceedings. (Tr.Vol. II at 55, 60, 61, 62.) McKnight's counsel stated, without any objection or rebuttal from JCI, that the court had previously heard in-chambers testimony from Jacobson, who stated that the battery and the cause of the explosion in the Wise accident were the same as in this case. (Id. at 62.) The district court then ruled, stating that "as I indicated in chambers, from the record surrounding the circumstances of the proposed statement, that plaintiff has met its burden ... with regard to substantial similarity." (Id.) Moreover, in the first trial Jacobson specially testified to the district court as part of an offer of proof that the Wise accident was substantially similar to McKnight's accident. (Partial Tr. of Civil Jury Trial, Document No. 384 at 1-2.) Hence, there was an existing record of the substantial similarity of the accident before the judge making the decision. Accordingly, we find sufficient indication that McKnight submitted foundational evidence of substantial similarity and that the district court did not abuse its discretion in admitting the evidence.10

JCI also asserts that the district court committed error by denying JCI the opportunity to introduce evidence of its own account of the Wise accident to counterbalance the admission of the Wise deposition testimony. The district court excluded this evidence in order to avoid a minitrial on the merits of the Wise case. In Hale II we addressed a similar situation. The district court allowed the plaintiff to introduce evidence of other similar accidents but refused to allow the defendant to introduce evidence of the disposition of the other claims. We stated:

While this was a very close call, and while this Court would have been inclined to admit this evidence, we are not prepared to say that its exclusion was an abuse of discretion.

Hale II, 820 F.2d at 935. We think this observation applies as well in this matter.

JCI's final argument is that the district court erred in instructing the jury. The district court enjoys broad discretion in instructing the jury. Brown v. Stites Concrete, Inc., 994 F.2d 553, 559 (8th Cir.1993) (en banc). "[A] party is entitled to an instruction reflecting the party's theory of the case if the instruction is legally correct and there is evidence to support it." Farmland Indus., Inc. v. Morrison-Quirk Grain Coop., 987 F.2d 1335, 1341 (8th Cir.1993). JCI argues that the district court committed three separate reversible errors in instructing the jury in this case.

First, JCI argues that Instruction No. 7 was given in error. Instruction No. 7 provided:

Your verdict must be for plaintiff if you believe:

First, defendant designed the battery;

Second, the seal between the battery cap plug and vent well was not adequate; and,

Third, defendant failed to use ordinary care to either design the battery to be reasonably safe, or failed to adequately test the battery; and,

Fourth, as a direct result of such failure, in one [or] more of the respects submitted in paragraph third [sic], plaintiff sustained damage.

The phrase ordinary care, as used in this instruction, means that degree of care, skill and learning that an ordinarily careful and prudent expert in defendant's business would use under the same or similar circumstances.

(Tr.Vol. V at 126-27; Jt.App., Docket No. 21.) JCI asserts that there was no evidence of JCI's departure from "ordinary care" to support giving this instruction. JCI also asserts that the instruction improperly merged design defect claims and manufacturing defect claims in paragraph 3, which rendered the instruction confusing. Moreover, JCI claims that there was no evidence of a manufacturing defect and that paragraph 3 of the instruction allowed McKnight to improperly introduce inadequate testing as a basis for liability.

We find no error in giving Instruction No. 7. Jacobson testified that JCI should have used a "screw in vent cap," which is used in the battery industry as a safer alternative to the "gang vent cap" design JCI used in the accident battery. (Tr. Vol. III at 130, 132.) This is sufficient evidence that JCI failed to meet standards of ordinary care to support the instruction. We also find no misstatement of the law or undue possibility of confusion in paragraph 3.

Likewise, we find sufficient evidence of a "failure to test" manufacturing defect to support the instruction. McKnight offered evidence that JCI did not test for hydrogen leaks after the vent caps were put in place during manufacturing. (Tr. Vol. II at 105, 123, 133, 135.) Failure to test is a viable theory of recovery under Missouri law. See Cohagan v. Laclede Steel Co., 317 S.W.2d 452, 454 (Mo.1958); Zesch v. Abrasive Co., 353 Mo. 558, 183 S.W.2d 140, 145 (1944). Accordingly, we find that the district court did not abuse its discretion by giving Instruction No. 7 to the jury.

JCI's second argument is that the district court erred in declining to give JCI's proposed instruction A, which states that the jury is the sole judge of witness credibility. The court, however, gave a similar instruction in its preliminary instructions to the jury at the beginning of the trial instructing the jury that they could "give the testimony of any witness such weight and value as you believe the testimony is entitled to receive." (Tr. Vol. I at 91.) The district court committed no reversible error in failing to repeat this instruction at the end of the trial.

JCI's final argument is that the district court erred in declining to give JCI's proposed instruction B, which states that corporate defendants should not be treated any differently than other defendants and should be treated just as fairly as individuals. JCI asserts that such an instruction was especially necessary to cure some of the comments by McKnight's counsel in closing argument attacking the corporate character of JCI. This issue was covered extensively in voir dire by both the district court and by JCI's counsel. (Tr. Vol. I at 37-38, 48, 78.) Hence, we find that while it may have been desirable to reiterate these ideas in a formal instruction, the district court nonetheless did not abuse its discretion in declining to give this instruction.

Having considered all of the assignments of alleged error, the judgment of the district court is affirmed.