BARKETT, Circuit Judge:
Plaintiffs Edward and Amanda Bearint, on behalf of their minor child, Kagan Be-arint, appeal the district court’s entry of final judgment pursuant to a jury verdict finding that Cosco, the manufacturer of the automobile infant safety seat in which Kagan was seated at the time the Bearint car was involved in an accident, was not liable for the injuries Kagan sustained. The Bearints specifically contest the district court’s handling of the jury instructions and verdict form, as well as several of the district court’s evidentiary decisions.
I. Background,
Amanda Bearint was driving a Saturn automobile with her sister Tracy Irwin, her ten-week-old son Kagan Bearint, and her seven-week-old nephew Alec Bacher. Tracy was in the front passenger seat, while the children were in rear-facing child seats in the back. Kagan’s child seat, manufactured by Cosco and named the “Arriva,” was strapped into the rear center seat. Alec’s child seat, produced by another company, was secured to the right rear seat.
The four were stopped at an intersection when a Dodge conversion van rear-ended them at between
35
and 50 miles per hour. It pushed the Saturn forward with tremendous force. The stationary passengers moved towards the rear of the car as a result. This pressed Kagan against the harness of his Arriva. It also caused Amanda Bearint’s front seat to collapse backwards onto Kagan’s car seat. At the same time, the Dodge van smashed into the trunk of the car and pushed its con
tents into the back of the rear seat, which buckled forward. Further, the impact pushed the Saturn forward, into the rear of a Chevrolet Tahoe, forcing Kagan against the forward-facing shell of his car seat.
Amanda, Tracy, and Alec were not seriously injured. But for reasons that the Bearints and Cosco contest, Kagan sustained very serious brain damage. The Bearints claim that Kagan was injured when the back of the collapsing front seat struck him in the head, and sued both Saturn and Cosco for his “enhanced injuries” arising from alleged defects in the Saturn and alleged negligence and defects in the Arriva, respectively. “Enhanced injuries” do not include the injuries that Kagan sustained from the impacts themselves — they only encompass those caused by a defective device in the automobile once a crash occurs. In this case, the Bearints sought only damages from the alleged collision between the front seat back and Kagan’s head, not from those caused by the crashes themselves.
Saturn settled with the Bearints, leaving Cosco as the sole defendant at trial. Thus, the purpose of trial was to determine if Kagan had suffered enhanced injuries beyond those caused by the impacts, whether the Arriva was defective or Cosco was negligent, and if so, whether the defect or negligence was a legal cause of his enhanced injuries. The Bearints’ claims against Cosco specifically alleged that the Arriva seat failed to protect Kagan from the impact of the collapsing front seat back. Normally, infant car seats are designed to rotate backwards in rear-impact collisions to protect the children within. This cocoons the child between the back seat and the strong plastic shell of the car seat, and protects the child from the impact of a collapsing front seat. However, Kagan’s Arriva did not rotate properly, exposing his head and allegedly causing his injuries. The Bearints alleged that the risk of collapsing front seats in rear-end impacts was so well-known when Cosco designed the Arriva that it had a duty to design the seat to protect children from this danger.
Cosco disputed that Kagan’s injuries were actually from the collapse of the front seat, arguing rather that they were caused by the sudden acceleration and deceleration of his head in the child seat. It also argued that even if the collapsing front seat did cause enhanced injuries, the Arri-va was not defective. Instead, it claimed that the Arriva was trapped by the almost simultaneous collapse of the front
and
rear seats. Further, Cosco alleged that Amanda Bearint strapped her son into the Arri-va incorrectly, preventing the seat from restraining him properly.
A jury found that Kagan suffered “enhanced injuries,” but that Cosco was not liable for them.
The trial court entered a
judgment in favor of the company and denied the Bearints’ post-trial motions.
The Bearints raise four issues on appeal. First, they claim that the trial court erred by allowing Cosco to admit evidence of Saturn’s fault and by placing the issue of Saturn’s liability on the jury verdict form,
contrary to Florida’s “crashworthiness” doctrine. Second, they argue that the trial court abused its discretion when it excluded their expert’s report of crash test results. Third, they claim that the trial court erroneously refused to give their jury instruction on manufacturing defects and used a confusing jury instruction and verdict interrogatory about Kagan’s enhanced injuries. Finally, they claim that the trial court abused its discretion when it allowed Cosco to impeach their expert witness with materials that Cosco had never disclosed to the Bearints, by excluding one of their witnesses, and by excluding all of their rebuttal evidence.
II. Discussion
A. Admission of Evidence of Saturn’s Fault and Placement of Saturn on Jury Verdict Form
The Bearints first challenge the district court’s decision permitting Cosco to present evidence to the jury of Saturn’s fault in causing any “enhanced injuries” and placing Saturn on the jury verdict form. They claim that
D’Amario v. Ford Motor Co.,
806 So.2d 424 (Fla.2001), which requires the exclusion of all evidence of fault with respect to the initial accident in “enhanced injury” cases, prevents the jury from considering whether Saturn was at fault for Kagan’s enhanced injuries. The district court ruled that
D’Amario
did not exclude such evidence.
4. Do you find from a preponderance of the evidence that Saturn Corporation placed the Saturn automobile on the market in a defective and unreasonably dangerous condition, and, if so, was the defect a legal cause of enhanced injuries to Kagan Bearint?
6. State the percentage of any fault of enhanced injuries Kagan Bearint [sic], that you charge to:
Defendant Cosco-
Saturn Corporation-
Amanda Bearint-
Total must be 100%.
We review the district court’s evidentiary rulings for abuse of discretion.
Morro v. City of Birmingham,
117 F.3d 508, 513 (11th Cir.1997),
cert. denied,
523 U.S. 1020, 118 S.Ct. 1299, 140 L.Ed.2d 465 (1998). However, basing an evidentiary ruling on an erroneous view of the law constitutes an abuse of discretion
per se. Conroy v. Abraham Chevrolet-Tampa, Inc.,
375 F.3d 1228, 1232 (11th Cir.2004). We review a district court’s determinations of state law
de novo. Edic v. Century Prods. Co.,
364 F.3d 1276, 1287 (11th Cir.2004). We therefore first address the district court’s application of
D’Amario.
D’Amaño
reconciles principles of comparative fault and proximate causation in enhanced injury cases. Florida follows the doctrine of comparative fault in civil negligence, strict liability, products liability, and breach of warranty cases such as this one. Fla. Stat. § 768.81 (2004). In general, parties are liable only for the percentage of the plaintiffs damages in an amount equal to the percentage of their fault.
Id.
In
Fabre v. Marin,
623 So.2d 1182 (Fla.1993), the Florida Supreme Court held that this doctrine required the finder of fact to consider the fault of all
tortfeasors — in that case those responsible for a car accident — when apportioning liability, even if they were not a party to the lawsuit. Thus, a defendant can argue to a jury that the fault of non-parties reduces the percentage of damages for which it is liable.
This doctrine created a problem in “enhanced injury” cases, also called “crash-worthiness” suits. “Enhanced injury” claims generally occur after an automobile crash, when the failure of some device in the car causes more extensive injuries than would have been suffered had the device worked properly. In such a case, the fact-finder only determines liability for injuries inflicted by the failure of the device due to defects or negligence.
However, the principle that damages are apportioned by percentage of fault collides with the idea that society should hold manufacturers of faulty automotive equipment liable for defects regardless of what caused an underlying accident. On the one hand, the person who caused the initial accident is responsible for triggering the event that precipitated the failure of the faulty device in the first place. It follows that the jury should consider his or her fault when holding the manufacturer of the faulty device liable. On the other, the claimed injuries at issue are only those stemming from the failure of the faulty device. Regardless of whether the underlying accident resulted from someone’s fault or purely by accident, the failure of the device would have caused the enhanced injuries. And because the manufacturer has a duty to design the device to protect against clearly foreseeable accidents, it should bear full responsibility for the product’s defects. Thus, allowing a jury to allocate some of the fault to the initial tortfeasor would partially and unfairly absolve the manufacturer of liability for making a faulty device. In legal terms, the issue is whether the initial accident is a legal cause of the enhanced injuries.
The Florida Supreme Court considered this dilemma in
DAmario. D’Amario
consolidated two cases in which the plaintiffs were involved in accidents caused by drunk drivers. 806 So.2d at 427-30. In one case, the plaintiffs alleged that a faulty relay switch continued to pump fuel from the gas tank after the impact, causing a fire and an explosion that seriously injured the victim.
Id.
at 427-28. In the other, the plaintiffs alleged that a defective seat belt failed to restrain the victim properly, causing her head to strike a metal post in the car.
Id.
at 429. They sued the car manufacturer, Ford, for their enhanced injuries — those above and beyond those caused by the collisions — -resulting from the allegedly defective devices.
Id.
at 427-430.
Ford averred that
Fabre
permitted it to argue the fault of the drunk drivers in causing the initial collisions as a defense, and to include their names on the jury form. The Florida Supreme Court disagreed.
Id.
at 426 (“[Pjrinciples of comparative fault concerning apportionment of fault as to the cause of the underlying crash will not ordinarily apply in crash worthiness or enhanced injury cases.”). The court based its decision on three considerations. First, it stated that under the crashworthiness doctrine, the fault causing the first collision was not the legal cause of the enhanced injuries caused by a defective automotive device.
Id.
at 433-34. This is because the crashworthiness doctrine presupposes the occurrence of an accident that triggers the device’s failure.
Id.
Therefore, what precipitated the initial collision is irrelevant to the manufacturer’s fault.
Id.
Just as it would be unfair to hold the manufacturer liable for injuries caused by the driver’s fault, it would be unfair to hold the initial driver liable for injuries caused by the fault of the
manufacturer.
Id.
at 440.
The Court distinguished
Fabre
on these grounds, stating that whereas
Fabre
involved joint and concurrent tortfeasors, these cases concerned successive tortfeasors in which a jury can separate the injuries that each tortfeasor caused.
Id.
at 435.
Second, it reasoned that evidence of initial accident-causing fault might unduly prejudice the jury, particularly in cases of drunk driving.
Id.
at 433. Such evidence also might confuse the jury on the issue of enhanced injuries by focusing attention on the conduct creating the initial accident, when the performance of the allegedly faulty device was really at issue.
Id.
at 440.
Finally, the court was concerned that admitting evidence of the initial tortfea-sor’s fault would unfairly insulate automobile manufacturers from liability.
Id.
at 434. Otherwise, the manufacturer could seek to introduce mitigating evidence in every enhanced injury case where the culpable behavior of others caused the initial accident, behavior for which it would never bear any liability.
Id.
The court also rejected the argument that separating the damages resulting from the initial collision from those caused by the failure of the device would require the jury to consider the fault of the initial tortfeasor. Id. at 439-40. It found that the fault behind a first collision is irrelevant to determining the amount of damage stemming from the initial accident — it is enough for the jury to know the physics of the initial crash to determine the damages it caused. See id. at 437, 440. However, the
D’Amario
Court suggested in a footnote that in cases where the jury cannot separate the initial from the enhanced injuries, the doctrine of indivisible injury— where the court holds initial and subsequent tortfeasors liable for all of the plaintiffs injuries- — could still permit the defendant to argue the fault of the initial wrongdoer.
Id.
at 440 n. 16 (citing
Gross v. Lyons,
763 So.2d 276, 279-80 (Fla.2000) (noting that two car crashes three months apart caused indivisible injuries where a jury may not have been able to apportion them)).
The Bearints argue that
D’Amario
requires the exclusion of all evidence of Saturn’s fault in this case. They claim that the “triggering event that causes the product to be tested” in this case was not only the van colliding into the Saturn (and the Saturn colliding into the Chevrolet Tahoe in front), but also the collapse of the front seat due to its defective design. They contend that the collapse of the front seat in a rear-impact crash is so foreseeable that Cosco had a duty to design the Arriva to withstand that impact, just as Ford had a duty to design the relay switch and seat belt at issue in
D’Amario
to protect the cars’ occupants from a reasonably foreseeable car crash. And because the fault of the individuals causing the initial accidents was not at issue in
D’Amario,
Saturn’s fault in causing the collapse of the front seat should similarly not be at issue in this case.
The Bearints’ argument thus would distinguish the collapse of the Saturn front seat back as a separate “collision” from the collision of the Arriva with the front seat back as the Arriva rotated upwards. However, this characterization of events is a departure from the one taken by the Bearints throughout this case. The Be-arints consistently characterized the impact of the front seat back with the Arriva as a single, simultaneous “enhanced inju
ry” event. Under this characterization, the crashworthiness doctrine would only preclude evidence concerning the negligence of either driver,
not the negligence of Saturn. It was not until the Bearints filed a supplemental memorandum supporting their motion to preclude evidence regarding settlement with Saturn
that the Bearints argued that “Cosco should be precluded from attempting to blame Saturn.”
The crashworthiness doctrine does not operate to exclude evidence of Saturn’s fault for the simple reason that Saturn was alleged to have contributed to Kagan’s enhanced injuries.
D’Amario
provides only that the parties could not litigate the cause of the initial collision between the Dodge van and the Saturn.
D’Amario
does not prevent the parties from litigating the cause of the alleged enhanced injuries. Here, the only way for the jury to determine the extent, if any, to which Cosco was liable for Kagan’s enhanced injuries was for the jury to also determine Saturn’s role in enhancing the injuries.
See Fabre,
623 So.2d at 1185 (“[T]he only means of determining a party’s percentage of fault is to compare that party’s percentage to all of the other entities who contributed to the accident, regardless of whether they have been or could have been joined as defendants.”). Accordingly, we find no error in the district court’s decision allowing admission of Saturn’s negligence and including Saturn on the verdict form.
B. Exclusion of the Bearints’ Crash Test Report
At trial, the district court precluded the Bearints from introducing evidence of an expert witness report detailing a crash test recreating the conditions of the accident using an identical Saturn and Dodge van. The district court denied this proffer as untimely. It rejected the Bearints’ argument that the report was a supplemental or a rebuttal report, and stated that they failed to submit it before the deadlines required by Fed.R.Civ.P. 26(a)(2).
On appeal, the Bearints concede that the report was untimely, and make no challenge to the district court’s ruling that it was not supplemental. However, they challenge the ruling on two grounds. First, they assert that the substantial prejudicial effect that the ruling caused the Bearints requires reversal for abuse of discretion. Second, they claim that Cosco “opened the door” to this excluded evidence by asking two of the Bearints’ experts questions on cross-examination that misled the jury into believing that they had not performed such a test. We review this evidentiary ruling for abuse of discretion.
Morro,
117 F.3d at 513.
The only arguments the Bearints advance to support a finding of abuse of discretion is that despite the untimely proffer, Cosco knew of the report’s existence after the deadline for submission of both initial and rebuttal expert reports but before trial, and that excluding the report greatly -prejudiced them. Fed.R.Civ.P. 26(a)(2)(C) provides clear deadlines for the submission of expert reports to the court, and Fed.R.Civ.P. 37(c)(1) gives district courts discretion to exclude untimely submissions.
Coastal Fuels Inc. v. Caribbean Petroleum Corp.,
79 F.3d 182, 202-03 (1st Cir.1996),
cert. denied,
519 U.S. 927, 117
S.Ct. 294 (excluding testimony of rebuttal witnesses because party did not comply with Fed.R.Civ.P. 26(a)(2)(C));
see also Yeti by Molly, Ltd. v. Deckers Outdoor Corp.,
259 F.3d 1101, 1105-06 (9th Cir.2001). Although the district court may have had discretion to admit an untimely report,
see Grimm v. Lane,
895 F.Supp. 907, 913 (S.D.Ohio 1995) (admitting untimely expert evidence because no risk of unfair surprise existed), it did not abuse its discretion to exclude it as untimely in the circumstances under which the Bearints offered it. The Bearints waited until trial, about four months after the report’s publication, to submit it. Given the wide latitude the district court has to exclude untimely submissions, we cannot say that it abused its discretion to exclude this report.
The record also does not reflect that Cosco “opened the door” to this otherwise inadmissible evidence. This Circuit recognizes the concept of “curative admissibility” — also called “opening the door” or “fighting fire with fire.”
United States v. Hall,
653 F.2d 1002, 1007 (5th Cir. Unit A Aug.1981);
see Morro,
117 F.3d at 517;
see also United States v. Monroe,
437 F.2d 684, 686 (D.C.Cir.1970). Under that doctrine, when a party offers inadmissible evidence before a jury, the court may in its discretion allow the opposing party to offer otherwise inadmissible evidence on the same matter to rebut any unfair prejudice created.
Beech Aircraft Corp. v. Rainey,
488 U.S. 153, 177 n. 2, 109 S.Ct. 439, 102 L.Ed.2d 445 (1988) (Rehnquist, J., dissenting) (curative admissibility an “exercise of judicial discretion”);
California Ins. Co. v. Allen,
235 F.2d 178, 180 (5th Cir.1956) (doctrine permitted “only to the extent necessary to remove any unfair prejudice”);
Crawford v. United States,
198 F.2d 976, 978-79 (D.C.Cir.1952) (doctrine rests “upon the necessity of removing prejudice in the interest of fairness”). Consequently, the extent to which otherwise inadmissible evidence is permitted must correspond to the unfair prejudice created. Further, the trial court must also weigh the need for and value of the rebuttal evidence against the potential for undue delay, confusion, and prejudice. Fed. R. Ev. 403;
Manuel v. City of Chicago,
335 F.3d 592, 597 (7th Cir.2003).
We do not find that the district court abused its discretion in finding that Cosco had not opened the door during its cross-examination of the Bearints’ experts, Mr. Harcourt and Dr. Burton. Cosco’s questioning of Mr. Harcourt cannot be interpreted as opening the door to the report at all. The report detailed the recreation of the accident conditions by crashing an identical van into an identical Saturn sedan with an Arriva child seat similarly situated within. In contrast, Cosco’s questioning of Mr. Harcourt asked him whether he recreated the accident conditions by crashing those two vehicles together with a
different
model of child seat in the back of the Saturn:
Q But using the scientific method, we could test your hypothesis that there are no child restraint seat systems that would have protected Kagan Be-arint in this accident, right?
A You could, yes, and I did.
Q In your report, sir, do you refer— well, let’s try it this way.
A way of testing would have been to duplicate the accident that actually happened by getting a Saturn of the same make and model, and getting a Dodge van of the same make and model as the one Carol Burton was driving, and putting a different car seat instead of the Arriva in, with dummies and everything else, to see what would happen, right?
MR. BURNSIDE [Bearints’ counsel]: Objection, Your Honor. May we approach?
[The court then stopped this line of questioning.]
This is a key distinction because the Be-arints wished to introduce the crash test report to support their calculations of what the speed of the van was at the time it struck the Saturn, and to bolster their version of what happened to the Arriva after that collision. In contrast, Cosco asked this question in an effort to show that no child seat — not just the Arriva— would have functioned properly in that accident. Thus, any inferences the jury may have drawn from Cosco’s question would have suggested that the Bearints never recreated the accident using a different car seat, which they in fact never did. This did not mislead the jury or implicate the crash test report.
The second contested line of questioning involved Cosco’s cross-examination of Dr. Burton, another expert witness, as follows:
Q So if Mr. Benedict [Cosco’s expert] is right and the accident did happen at 53 kilometers per hour, that would be a speed higher than that of any of the most severe accidents you’ve reported on in this study we’re talking about, right?
A Agreed.
Q You chose to rely on Mr. McCracken’s [the Bearints’ expert] figure [as to the speed of the initial impact between the van and the Saturn], I take it?
A Yes, sir.
Q But you didn’t do anything to try to verify it, did you?
A Well, you know I’m not an engineer, so like I said, I trust the experts, and I did consider — I know Charles Benedict [Cosco’s expert], I considered what he had to say as well.
MR. GOLDSMITH [Bearints’ counsel]: Your Honor, could we approach please?
These questions likewise cannot be construed as opening the door to permit the admission of the untimely report. Moreover, up until the Bearints’ objection to the questioning, they had not mentioned Dr. Burton’s involvement in the crash test reflected in the report. Instead, they had left the impression that the report only pertained to Mr. Harcourt’s testimony. We find no abuse of discretion.
C. The District Court’s Refusal to Give the Bearints’ Jury Instruction on Manufacturing Defects and Utilization of a Confusing Jury Instruction and Verdict Interrogatory about Enhanced Injury
1. Refusal to Give Manufacturing Defect Instruction
The Bearints claim that the district court erroneously failed to give the jury an instruction on a manufacturing defect in the Arriva. We will reverse a refusal to give a requested instruction only if: (1) the requested instruction correctly stated the law, (2) the instruction dealt with an issue properly before the jury, and (3) the failure to give the instruction resulted in prejudicial harm to the requesting party.
Roberts & Schaefer Co. v. Hardaway Co.,
152 F.3d 1283, 1295 (11th Cir.1998);
Goulah v. Ford Motor Co.,
118
F.3d 1478, 1485 (11th Cir.1997) (citation omitted).
In this case, the failure to give the Bearints’ requested manufacturing defect jury instruction did not prejudicially harm them. The trial court utilized the Eleventh Circuit Standard Civil Instruction 2.1 governing product liability claims, which defines a defective product in relatively broad terms:
In this case, the Plaintiffs claim damages resulting from injuries to Kagan Bearint alleged to have occurred as a result of a defective condition in the Cosco Arriva infant seat.
In order to recover on this claim, the Plaintiffs must prove each of the following facts by a preponderance of the evidence:
First:
That, at the time of the manufacture and sale of the product (the Cosco Arriva infant seat), the product was in a defective condition making it unreasonably dangerous to the user;
Second:
That the product was expected to and did in fact reach the Plaintiffs, and was thereafter operated up to the time of the accident without substantial change in its condition as of the time Defendant Cosco sold it; ...
A product is in a defective condition, unreasonably dangerous to the user, when it has a propensity or tendency for causing physical harm beyond that which would be contemplated by the ordinary user, having ordinary knowledge of the product’s characteristics commonly known to the foreseeable class of persons who would normally use the product.
The Bearints’ suggested instruction reads:
A product is defective if ... the infant safety seat is in a condition unreasonably dangerous to the user and the product is expected to and does reach the user without substantial change affecting that condition.
The standard jury instructions already incorporate the Bearints’ requested language to a great extent. Further, the standard jury instructions are broad and encompass both manufacturing and design defects. Therefore, we find that the district court’s refusal to give this instruction did not result in prejudicial harm to the Bearints and was not an abuse of discretion.
2. Utilization of Confusing Jury Instruction and Verdict Interrogatory on Enhanced Injuries
We apply a deferential standard of review to a trial court’s jury instructions.
Eskra v. Provident Life & Accident Ins. Co.,
125 F.3d 1406, 1415 (11th Cir.1997) (citations omitted). If the instructions accurately reflect the law, the trial judge is given wide discretion as to the style and wording employed in the instruction.
Id.
Under this standard, we examine “whether the jury charges, considered as a whole, sufficiently instructed the jury so that the jurors understood the issues and were not misled.”
Carter v. DecisionOne Corp.,
122 F.3d 997, 1005 (11th Cir.1997) (citation omitted). We will reverse the trial court because of an erroneous instruction only if we are “left with a substantial and ineradicable doubt as to whether the jury was properly guided in its deliberations.”
Id.
(citation omitted). We reviéw phrasing of special jury interrogatories under an abuse of discretion standard; reversal is warranted where the interrogatories have “the potential for confusing or misleading the jury.”
Central Ala. Fair Hous. Ctr., Inc. v. Lowder Realty Co.,
236 F.3d 629, 635 (11th Cir.2000) (quoting
Petes v. Hayes,
664 F.2d 523, 525 (5th Cir. Dec.1981)).
The Bearints attack the trial court’s jury instruction on enhanced injury, which reads:
No claim is being made by Plaintiffs for damages arising out of the initial automobile collision between the Dodge van, the Saturn, and the Chevrolet Tahoe, and Defendant Cosco should not be held liable for damages caused by that initial automobile collision. Defendant Cosco is responsible for any enhanced injuries to the extent that Plaintiffs demonstrate the existence of a defective condition or negligence on the part of Defendant Cosco and that defect or negligence was a legal cause of the enhanced injuries of Kagan Bearint.
They claim that this instruction confused the jury as to what constituted the “initial automobile collision.” They allege that it suggests that the first collision was the impact of the van into the rear of the Saturn and the second accident — the one causing the enhanced injuries — was the impact of the Saturn crashing into the Chevrolet Tahoe, and not the impact of the front seat collapsing backwards onto Ka-gan.
We fail to see how the language “the initial automobile collision between the Dodge van, the Saturn, and the Chevrolet Tahoe” refers only to the collision between the Dodge van and the Saturn. Further, the instruction clearly outlines that the enhanced injuries constitute those incurred over and above those caused by the collisions between the three vehicles. Therefore, it does not raise “substantial and ineradicable doubt as to whether the jury was properly guided in its deliberations.”
Carter,
122 F.3d at 1005.
The jury interrogatory asks:
Do you find from a preponderance of the evidence that Kagan Bearint suffered enhanced injuries over and above any injuries that would have occurred as a result of the initial automobile collision alone?
Although the interrogatory only refers to a single “initial automobile collision,” given the clear presentation of the facts regarding the collisions between the Saturn, the Dodge van, and the Chevrolet Tahoe at trial and the jury instruction on “enhanced injuries,” this interrogatory did not have the potential for confusing or misleading the jury. We therefore find no abuse of discretion.
D. Other Evidentiary Rulings
1. Permitting Cosco to Impeach Plaintiff’s Experts with Materials It Never Disclosed to the Bearints
The trial court permitted Cos-co to impeach one of the Bearints’ expert witnesses with materials that it had never disclosed to the Bearints. The Bearints claim that the ruling violates Florida law requiring disclosure of all evidence to be used in trial,
see Northup v. Acken,
865 So.2d 1267 (Fla.2004) (holding that a party must disclose all evidence before trial, including materials only used for impeachment purposes), and constitutes reversible
error. However, it is federal law, and not state law, that controls discovery in diversity actions when a federal Rule is on point, provided that it does not violate the Rules Enabling Act or the U.S. Constitution.
See Hanna v. Plumer,
380 U.S. 460, 471, 85 S.Ct. 1136, 14 L.Ed.2d 8 (1965). In this case, Fed.R.Civ.P. 26(a)(3) directly addresses this matter, exempting parties from disclosing evidence they may present at trial solely for impeachment purposes, and therefore governs even if inconsistent with Florida law.
Rule 26(a)(3) exempts evidence used solely for impeachment because pretrial disclosure would significantly diminish its impeachment value.
Denty v. CSX Transp.,
168 F.R.D. 549, 550 (D.Or.2000). Here, the disputed evidence, relating to the expert witness’s qualifications as an engineer, was used solely for the purposes of impeachment rather than for a substantive purpose.
See Klonoski v. Mahlah,
156 F.3d 255, 270, 274-76 (1st Cir.1998),
cert. denied,
526 U.S. 1039, 119 S.Ct. 1334, 143 L.Ed.2d 498 (1999) (describing the difference between impeachment and substantive testimony and reiterating that only evidence used solely for impeachment is exempt from the Federal Rules’ disclosure requirement). Therefore, Cosco did not need to disclose this evidence to the Bearints.
2. Exclusion of the Bearints’ Lay Witness Steven Dyer
Next, the Bearints challenge the district court’s refusal to permit one of their lay witnesses, Steven Dyer, to offer rebuttal testimony. Dyer’s testimony was excluded because the Bearints did not disclose him on their pretrial witness list. The Be-arints argue that this constituted reversible error because Dyer’s testimony was critical to their ease and would not have prejudiced Cosco.
We review a district court’s exclusion of a witness not listed on the pretrial order for abuse of discretion, and consider: (1) the importance of the testimony; (2) the reason for the appellant’s failure to disclose the witness earlier; and (3) the prejudice to the opposing party if the witness had been allowed to testify.
Rogers v. Muscogee County School District,
165 F.3d 812, 818 (11th Cir.1999).
Regardless of the importance of Dyer’s testimony, the reasons for the delay in the Bearints’ disclosure and the consequent prejudice that his testimony would have caused Cosco require us to affirm the district court’s ruling. The Bearints first mentioned Dyer’s existence on the first day of trial. They allege that they had only recently learned of his existence and thus were unable to include him in their pretrial disclosures. In and of itself, this delay may have been excusable. But their subsequent delay in disclosing the full nature of Dyer’s testimony was not. At first, the Bearints claimed only that Dyer had witnessed the accident. As he was one of many witnesses to have done so, Cosco had no reason to believe that his testimony was important or even necessary. But as time passed, the Bearints added, little by little, to their account of Dyer’s importance. Two days later, they mentioned to the court that Dyer had also helped take Kagan out of the wrecked Saturn. Five days after that, they told the court that he was the first person present at the scene of the accident. And finally, after Cosco had rested its case a full eleven days after the start of trial, they disclosed that Dyer would testify about a central issue — whether the Arriva had actually been trapped under the driver’s seat. The Bearints could easily have made this disclosure on the opening day of trial, and offer no plausible reason for their failure to do so.
Further, permitting Dyer’s testimony at this stage of the trial would have prejudiced Cosco. After having put on its case with no reason to believe that Dyer’s testimony would materially change the Be-arints’ own presentation, Cosco would have been blindsided by a néw witness offering-facts directly contradicting its own account of events. The Bearints argue that Cosco was in fact on notice of Dyer’s testimony from the first day of trial and had ample opportunity to depose him. But because the Bearints downplayed the importance of Dyer’s knowledge at that time, Cosco had no incentive or need to do so. This does not mitigate the prejudice that allowing the testimony would have created.
The Bearints also claim in their reply brief that the parts of Dyer’s testimony impeaching the testimony of Cosco’s witnesses were admissible because a party need not disclose testimony used solely to impeach. Fed.R.Civ.P. 26(a)(3). Despite the fact that they did not raise this argument at trial and cannot do so here,
Walton v. Johnson & Johnson Servs., Inc.,
347 F.3d 1272, 1292 (11th Cir.2003),
cert. denied,
— U.S. -, 124 S.Ct. 1714, 158 L.Ed.2d 399 (2004), they could not avail themselves of this exception regardless. It only covers witnesses whose testimony is offered
solely
to impeach, and does not apply to those who provide any substantive testimony. The Bearints’ representations to the district court show that Dyer was prepared to provide substantive testimony. They cannot bifurcate Dyer’s testimony just to fit it within this exception on appeal.
S. Exclusion of the Bearints’ Rebuttal Evidence
Finally, the Bearints assert that the trial court erroneously excluded their other rebuttal evidence, which consisted of the actual Saturn seat belt and rear seat and expert testimony refuting Cosco’s assertion that Amanda Bearint did not properly secure Kagan in the Arriva. We find no abuse of discretion in this ruling. The Bearints had ample opportunity to have their expert testify on these matters in their case-in-chief. Further, they actually opposed Cosco’s motion to allow the jury to see the actual Saturn earlier at trial, contradicting their assertion that the trial court prevented them from ever presenting the car’s rear seat and seat belt. The trial court did not abuse its discretion in excluding this evidence.
AFFIRMED.