These wrongful death actions arose out of the July 20, 1992 crash of a V-22 Osprey aircraft near Quantico, Virginia. Plaintiffs are representatives of the estates of two of the seven members of the crew. Defendants designed, manufactured, and tested the Osprey and the components at issue in this case. A jury trial resulted in judgment for defendants. On appeal, plaintiffs challenge several of the District Court’s evidentiary rulings.
' For the reasons stated below, we will affirm the District Court’s final judgment.
I. FACTS
The twin-engine Osprey combines the vertical takeoff and landing capability of a helicopter with the cruising speed and flying capabilities of a fixed wing aircraft. Bell Helicopter Textron, Inc., along with Boeing Vertol Company, designed and developed the Osprey under a contract with the federal government. General Motors Corporation designed and manufactured the engines under a separate contract with *411the government. Macrotech Fluid Sealing, Inc., manufactured the torquemeter shaft seal,'known as the “617 seal,” under a subcontract with the Bell-Boeing team.
On July 20, 1992, the Osprey crashed while in the transition stage from airplane to helicopter flight. The Osprey was attempting to land at the Quantico military field after a two hour and forty-four minute flight from Eglin Air Force Basé.in Florida. The plane’s three U.S. Marine pilots, together with four Boeing engineers, were killed.
The accident was investigated by a U.S. Navy Court of Inquiry. The Court of Inquiry’s findings were then forwarded to a superior Naval authority for review, referred to as the First Endorsement. The Endorsement became part of the Court of Inquiry Report.
At trial, the District Court admitted the Report, including the Endorsement, into evidence. Both the Court of Inquiry and the Endorsement agreed that the crash occurred after a flammable fluid was ingested by the aircraft’s right engine as the craft was attempting to land. The Court of Inquiry stated that the right torqueme-ter shaft seal (the 617 seal) was installed backwards and leaked, providing “the most probable primary causal factor for the mishap.” . However, the Endorsement did “not concur” with this conclusion. The Endorsement stated that improper installation of the 617 seal was only one possible source of the leaked flammable fluid.
Plaintiffs’ theory was that the crash was caused by a transmission oil leak past a 617 seal that had been installed backwards by Boeing mechanics. They contended that Bell and Macrotech were negligent in not designing a “Murphy-proof1' seal which could not be reversed. While such a “two-way” seal has been installed in subsequent versions of the Osprey, the District Court precluded evidence of this post-incident remedial measure.
For the defense, Bell contended that a 617 seal would' not leak even if reversed and presented an alternate theory of causation that the engine failure was caused by hydraulic fluid, not transmission oil.
Bell presented evidence of three separate tests which concluded .that a reversed 617 seal did not leak. The first test (1992 test) had been performed by Bell employee Ken Wilson at the request of the Court of Inquiry. The 1992 test, which was discussed in both the Court of Inquiry’s findings and the Endorsement, concluded that a reversed 617 seal subjected to the same range of RPMs, torque, power, heat, pressure and tilt angles as the 617 seal on the Osprey did not leak. Plaintiffs’ experts criticized the 1992 test in several ways, including challenging the use of “new” seals and contending that the test’s one hour and twenty minute duration was too short. Plaintiffs did not object to the admission into evidence of the 1992 test.
The two other reversed 617 seal tests, conducted in 1997 and 1998, were performed at Bell’s request by Wilson, who had retired from Bell in 1995. These two tests were videotaped, and Bell produced these videotapes to plaintiffs five months before trial.
During the defense case at trial, when Bell attempted to elicit testimony from Wilson about the 1997 and 1998 tests, plaintiffs objected based, inter alia, on “unfair surprise.” The court excused Wilson and ordered him and defense expert Dr. Thomas Eagar to produce supplemental reports for plaintiffs by 5 p.m. that day, a Friday. Shortly thereafter, the court recessed trial until Tuesday to give plaintiffs the opportunity to consult with their own experts and depose Wilson and Dr. Eagar, if appropriate.
*412When trial resumed on Tuesday morning, plaintiffs confirmed that they had received the reports and declined to take any additional depositions. They asked the court to exclude Wilson’s videotaped tests on the ground that they were not substantially similar to the conditions on the Osprey. After hearing argument about substantial similarity, the court admitted the 1997 and 1998 tests. At the close of Wilson’s testimony, plaintiffs moved to strike his testimony regarding the 1997 and 1998 tests, again on grounds of substantial similarity, and the court denied the motion.
After Wilson testified, Dr. Eagar testified as an expert on failure analysis, testing with respect to failure analysis, and materials science. He presented an alternate theory of causation, opining that the Osprey’s engine failure was caused by hydraulic fluid, not transmission oil.
After a six week trial, the jury returned a verdict for the defendants. The District Court denied post verdict motions, and plaintiffs timely appealed to this Court.
II. STANDARD OF REVIEW
The District Court had jurisdiction under 28 U.S.C. § 1332. We have jurisdiction to review the final judgment of the District Court pursuant to 28 U.S.C. § 1291.
We review the District. Court’s evidentiary rulings principally for abuse of discretion. See General Electric v. Joiner, 522 U.S. 136, 146, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997) (decision to admit or exclude expert testimony); Glick v. White Motor Co., 458 F.2d 1287, 1294-95 (3d Cir.1972) (admission or exclusion of tests); see also Inter Med. Supplies, Ltd. v. EBI Med. Sys. Inc., 181 F.3d 446, 464 (3d Cir.1999) (reviewing district court’s admission of evidence for abuse of discretion, but exercising plenary review over evidentiary rulings with legal component); Complaint of Consolidation Coal Co., 123 F.3d 126, 131 (3d Cir.1997) (same), cert. denied, 523 U.S. 1054, 118 S.Ct. 1380, 140 L.Ed.2d 526 (1998). To show an abuse of discretion, appellants must show the district court’s action was “arbitrary, fanciful or clearly unreasonable.” Stick v. United States, 730 F.2d 115, 118 (3d Cir.1984). We will not disturb a trial court’s exercise of discretion unless “no reasonable person would adopt the district court’s view.” Oddi v. Ford Motor Co., 234 F.3d 136, 146 (3d Cir.2000).
III. DISCUSSION
Plaintiffs first challenge the District Court’s admission of the videotaped 1997 and 1998 tests which concluded that a reversed 617 seal does not leak. Next, the plaintiffs argue that Dr. Eagar’s testimony regarding a leak of hydraulic fluid lacked an adequate factual foundation. Finally, plaintiffs contend the District Court erred in precluding evidence of Bell’s post-crash two-way seal designs. We address each argument in turn.
A. The Videotaped 1997 and 1998 Tests
As proponents of the videotaped evidence, Bell had to make a foundational showing that the 1997 and 1998 test conditions were substantially similar to conditions on the Osprey. See Glick, 458 F.2d at 1294; Ramseyer v. Gen. Motors Corp., 417 F.2d 859, 864 (8th Cir.1969). However, as the term suggests, substantial similarity does not require perfect identity between actual and experimental conditions. Experimental evidence may be admitted even if conditions - do not perfectly correspond to the conditions at issue in litigation; dissimilarities may affect the weight of the evidence, but not its admissibility. See id. A ruling on substantial similarity is committed to the sound discretion of the trial judge. Id.
*413The record reflects that Bell satisfied this threshold of admissibility. Wilson, who conducted the tests, explained the protocol for each test in a written report and testified at length about the tests. The 1997 test used an actual 617 -seal installed on a replica shaft and torqueme-ter housing machined from blueprints of the Osprey. The test consisted of four phases. Each phase ran for eight to ten minutes at pressures of up to 500 psi ■(pounds per square inch), which was roughly ten times the normal operating pressure on the 617 seal on the Osprey. When oil pressure was applied to the reversed 617 seal in this duplicate housing, the seal did not leak.
The 1998 test used actual Osprey components including a 617 seal, torquemeter shaft and torquemeter housing. This configuration ran under pressure for 18 hours at 114 psi, and the reversed 617 seal did not leak. Wilson testified that centrifugal pressure did not affect the result of the tests. Any minimal additional pressure that would have been generated by centrifugal forces was instead created and surpassed by the additional pressure (500 and 114 psi) exerted in the tests. Faced with this evidence, it was a proper exercise of the District Court’s discretion to admit the 1997 and 1998 tests.
By contrast, plaintiffs did not offer any rebuttal evidence that the tests were not substantially similar to conditions on the Osprey. Instead, plaintiffs’ counsel argued that the 1997 and 1998 tests were conducted in a static environment, as opposed to a dynamic environment where they would be subject to centrifugal forces and vibration. However, as plaintiffs’ counsel admitted at oral argument, they failed to produce a witness, a report, or any evidence to support their argument that the lack of centrifugal forces imposed on the seal rendered the conditions meaningfully dissimilar. Thus, the only evidence before the District Court was defense reports and witness testimony that the tests were substantially similar to conditions on the Osprey. In light of this evidence, we will not disturb the District Court’s exercise of discretion to admit the 1997 and 1998 tests. Any dissimilarities that plaintiffs identified were properly the subject of cross-examination.
In addition, a review of the record belies plaintiffs’ alternate contention that they were unfairly surprised by the tests. The record reflects that plaintiffs had the videotapes five months prior to trial and that their own experts had viewed the videotapes: At trial, when plaintiffs objected based on unfair surprise, the court ordered Bell to produce supplemental reports from Wilson and Dr. Eagar and recessed early on a Friday to give plaintiffs additional opportunities to depose the defense witnesses and consult with their own experts. When proceedings resumed the following Tuesday, plaintiffs had declined to take additional depositions. Plaintiffs argued that the tests were not substantially similar but, as discussed above, offered no evidence to rebut1 Bell’s -evidence and support their argument that a lack of centrifugal forces rendered the tests inadmissable. Nor did plaintiffs move for a continuance in order to get the requisite testimony.
In view of the record as set out above, we find no abuse of discretion in the District Court’s rulings on the admissibility of this evidence.
B.‘ Dr. Eagar’s Testimony
Defense expert Dr. Thomas Eagar opined that the most probable source of flammable fluid which caused the Osprey’s engine failure was hydraulic fluid. Plaintiffs contend that Dr. Eagar’s testimony regarding a leak of hydraulic fluid lacked *414an adequate factual basis because there was insufficient evidence of the presence of hydraulic fluid inside the engine.
Under Rule 703 of the Federal Rules of Evidence, experts may rely on facts from firsthand knowledge or observation, information learned at the hearing or trial, and facts learned out of court. Fed. R.Evid. 703.1 If the facts are of the type “reasonably relied upon” by experts in the particular field in forming opinions or inferences upon a subject, the facts or data need not be independently admissible in evidence. Id. See In Re Paoli R.R. Yard PCB Litig., 35 F.3d 717, 747 (3d Cir.1994). Rule 705 provides for the disclosure of facts underlying the expert’s opinion. Fed.R.Evid. 705;2 see also Fed.R.Civ.P. 26(a)(2)(B) and 26(e)(1) (relating to disclosure in advance of trial of the basis and reasons for an expert’s opinion). It is an abuse of discretion to admit expert testimony which is based on assumptions lacking any factual foundation in the record. See Elcock v. Kmart Corp., 233 F.3d 734, 756 n. 13 (3d Cir.2000) (discussing Rules 702, 703, 402 and 403 and stating that foundational requirement for admissibility of expert testimony is found in the “interstitial gaps” among the federal rules). Rule 705, together with Rule 703, places the burden of exploring the facts and assumptions underlying the testimony of an expert witness on opposing counsel during cross-examination. See e.g. Ratliff v. Schiber Truck Co., Inc., 150 F.3d 949, 955 (8th Cir.1998); Toucet v. Maritime Overseas Corp., 991 F.2d 5, 10 (1st Cir.1993).
Here, the record reflects a factual foundation sufficient to support Dr. Eagar’s opinion that the most probable source of flammable fluid was hydraulic fluid. The record shows that, of the possible fluids involved in the accident, only hydraulic fluid is red. A red residue was found in the torquemeter housing. This red residue was tested for the Court of Inquiry and found to be a good match for hydraulic fluid. There was some hydraulic oil found in front of the engine and it may have gotten into the engine. Finally, a red residue containing hydraulic oil was discovered on the engine air particle separator, adjacent to the engine. Thus, the record reflects sufficient evidence of hydraulic fluid solvent in places it should not have been — outside the engine, near the engine, and in the torquemeter housing— to form the factual foundation for Dr. Ea-gar’s testimony. It was within the discretion of the District Court to admit such testimony.
Once Bell’s expert met the foundational requirements for admissibility, the burden shifted to plaintiffs to explore any deficiencies in the expert’s sources. A party confronted with an adverse expert witness who has sufficient, though perhaps not overwhelming, facts and assumptions as the basis for his opinion can highlight those weaknesses through effective cross-examination. See Ratliff, 150 F.3d at 955; Toucet, 991 F.2d at 10; cf. Daubert v. *415Merrell Dow Pharm, Inc., 509 U.S. 579, 596, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993) (“Vigorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof are the traditional and appropriate means of attacking shaky but admissible- evidence.”). The District Court properly exercised its discretion in admitting the testimony and permitting appropriate cross-examination of Dr. Eagar.3
C. Post-crash Seal Designs
The Federal Rules of Evidence expressly preclude the introduction of evidence of subsequent remedial measures to prove a party’s negligence or culpable conduct. Fed.R.Evid. 407.4 Rule 407 rests on the strong public policy of encouraging manufacturers to “make improvements for greater safety.” Kelly v. Crown Equipment Co., 970 F.2d 1273, 1276 (3d Cir.1992). A manufacturer will be discouraged from making improvements for the greater safety of its products if such changes can be introduced as evidence that the previous designs were defective. Id. Moreover, Rule 407 “operates on the presumption that undue prejudice is likely in certain situations....” Id. at 1277. Thus, courts “routinely exclude evidence of [subsequent remedial measures] to encourage people to take such measures whether or not they are at fault.” Petree v. Victor Fluid Power, Inc., 831 F.2d 1191, 1198 (3d Cir.1987) (“Petree I”).
Pursuant to Rule 407, the District Court excluded evidence that a two-way seal was used on the Osprey following the crash. However, because defendants argued that a one-way seal design was reasonable and that a two-way seal was more difficult to install and “not suited for the military environment,” Plaintiffs contend that evidence of post incident use of the two-way seal was admissible for purposes of impeachment.
While the text of Rule 407 permits admission of subsequent remedial mea*416sures for impeachment, we have cautioned against permitting the exception to “swallow” the rule. See Petree v. Victor Fluid Power, Inc., 887 F.2d 34, 39 (3d Cir.1989) (“Petree II”) (impeachment exception may not be used as “subterfuge” to prove negligence). We have recognized that, in light of the strong public policy considerations behind the rule and the risk of undue prejudice, the trial judge should be afforded a healthy deference in preserving both the rule and the exception. Id. Under Rule 407, together with the Rule 4035 unfair prejudice/probative value weighing, the trial court retains broad power to insure that remedial measures evidence is not improperly admitted under the guise of the impeachment exception. Id.
In the instant case, the record contains a significant amount of pre-incident impeaching testimony regarding one-way versus two-way seals. The court admitted considerable testimony and graphic documentation of Bell’s receipt, review, and rejection of a pre-crash alternative design of a two-way seal from a vendor, Longhorn Gasket. Plaintiffs cross-examined Beil’s witness on Longhorn’s two-way seal proposal. The jury saw an exhibit which contained a diagram of Longhorn’s two-way seal and Bell’s evaluation of the proposal. Using evidence of Bell’s rejection of a two-way seal prior to the crash, plaintiffs thus had the opportunity to impeach the defense witness’s testimony regarding the reasonableness of the 617 seal’s one-way design without resort to prejudicial post-incident evidence.
As we stated earlier, a district court retains considerable discretion in determining whether otherwise excludable remedial measures evidence should be admitted under the impeachment exception. Here, where the evidence of the existence of a two-way seal design prior to the accident was sufficient for plaintiffs to effectively cross-examine the defense witness, it was a proper exercise of the District Court’s discretion to exclude highly prejudicial post-incident evidence. In light of the availability of this pre-incident impeaching evidence, it was not error for the District Court to exclude the prejudicial post-incident remedial measures.
IV. CONCLUSION
In light of our disposition of plaintiffs’ claims, Bell’s cross appeal is moot. Likewise, we need not address Macrotech’s alternative grounds for affirmance. For the foregoing reasons, we will affirm the final judgment of the District Court.
. Rule 703 provides:
The facts or data in the particular case upon which an expert bases an opinion or inference may be those perceived by or made known to the expert at or before the hearing. If of a type reasonably relied upon by experts in the particular field in forming opinions or inferences upon the subject, the facts or data need not be admissible in evidence.
. Rule 705 provides:
The expert may testify in terms of opinion or inference and give reasons therefor without first testifying to the underlying facts or data, unless the court requires otherwise. The expert may in any event be required to disclose the underlying facts or data on cross examination.
. We appreciate our dissenting colleague's criticisms of the factual foundation for Dr. Eagar’s testimony. These criticisms, however, raise precisely the type of issues that must be resolved by a fact-finder having the benefit of the adversary process. By raising these concerns, the dissent effectively conducts an independent evaluation of the weight of the evidence — an exercise that we believe exceeds the appropriate boundaries of an abuse of discretion review. While the Federal Rules of Evidence call upon the courts to serve as gatekeepers who independently evaluate the admissibility of expert opinion testimony, they rely upon the discretion of the trial courts - not the discretion of the courts of appeals. See In re TMI Litig., 193 F.3d 613, 697 (3d Cir.1999). Because the record contains some factual basis — albeit shaky — for Dr. Eagar’s testimony, the District Court did not abuse its discretion in performing this gatekeeping function.
Moreover, we disagree with the dissent to the extent it concludes that “no reasonable expert could base an opinion” on Dr. Eagar's factual foundation. In re TMI Litig., 193 F.3d at 697. Such a conclusion would reject implicitly Dr. Eagar’s qualification to testify as an expert witness under Rule 702 — a determination that the plaintiffs do not challenge and that we have no reason to reverse. See Elcock, 233 F.3d at 756 n. 13 (explaining the relationship between the foundation requirements of Rule 703 and the qualification requirements of Rule 702).
. Fed.R.Evid. 407 provides:
When, after an injury or harm allegedly caused by an event, measures are taken that, if taken previously, would have made the injury or harm less likely to occur, evidence of the subsequent measures is not admissible to prove negligence, culpable conduct, a defect in a product, a defect in a product's design, or a need for a warning or instruction. This rule does not require the exclusion of evidence of subsequent measures when offered for another purpose, such as proving ownership, control, or feasibility of precautionary measures, if controverted, or impeachment.
. Fed.R.Evid. 403 provides:
Although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence.