Plaintiffs Joe and Phyllis Hudgens and Francis and Becky Crawford appeal from grants of summary judgment in favor of defendant DynCorp in these diversity actions arising out of a helicopter crash in Shelby County, Alabama.

Francis Crawford and Joe Hudgens were piloting a United States Army helicopter on a Medivac mission on May 1, 1999 when the helicopter's tail fin separated from the aircraft, resulting in a crash in which both men were severely injured. Each of the pilots and their respective wives filed separate suits on August 6, 1999 in the Northern District of Alabama.

After the accident, investigations conducted by the Army and the aircraft's manufacturer, Bell Helicopters/Textron ("Bell"), determined that the tail fin's separation resulted from a rupture of the helicopter's forward vertical fin spar ("fin spar" or "spar"), a component of the structural assembly by which the helicopter's tail rotor blade is attached to the pilot's carriage.1 An external skin covers the structural frame of which the fin spar is part. The investigating authorities concluded that the rupture was caused by a crack that originated at a rivet hole near the base of the fin spar.

The accident helicopter was part of the "Flat Iron" fleet of UH-1, or "Huey," helicopies maintained at the Army's Fort Rucker base in Alabama. Under the terms of a contract entered into by the Army and DynCorp, DynCorp was to maintain Army aircraft located at the base. The agreement required DynCorp to "determine the airworthiness condition of ... aircraft as required by applicable regulations and publications. Such a determination shall be based on inspection, maintenance operational checks, and test flying as required by applicable Army publications and directives." The publications contemplated by the contract have been entered in the record in the form of inspection checklists and manuals providing instructions for the maintenance and repair of UH-1 aircraft. See R.1-36, Tab H, exhibits 1-3.2 One checklist identifies a series of inspections to be completed every day a helicopter is flown; another prescribes a more intensive series of procedures to be carried out at longer intervals; and a third manual, spanning well over a thousand pages, prescribes appropriate means for repairing a variety of defects. Army personnel stationed at Fort Rucker monitored DynCorp's performance to ensure compliance with these instructions for the maintenance of the UH-1 fleet. See R.1-36, Tab G, at 2; R.1-44, Tab S, at 1.

Prior to the accident, several authorities familiar with Bell's UH-1 line of helicopters had already identified the fin spar problem that authorities later identified as the cause of the subject accident. In 1997, the Federal Aviation Administration ("FAA") circulated an "Airworthiness Directive," which reported the FAA's determination that repeated engagement of the UH-1 and related models in heavy lifting operations could create fatigue fractures that would compromise the fin spar. In an effort to prevent such failures, the FAA directive mandated that civilian operators of these models modify the fin spar in a manner designed to facilitate inspection for cracks. The directive required that such modifications be completed within the first eight flight hours subsequent to the directive's issuance. Thereafter, operators were to inspect the modified fin spar for cracks at least once during each subsequent eight hours of flight time, using a bright light and a 10X magnifying glass to enhance the visibility of any fractures. In 1998, Bell Helicopter/Textron circulated a "Military Alert Bulletin" likewise advising that cracks had been found on some aircrafts' fin spars and recommending "tap hammer" and fluorescent dye penetrant inspections within 25 flight hours of the bulletin's receipt and thereafter at 180-day intervals.

The Army is not bound by the FAA's airworthiness directives or Bell's alert bulletins in its operation of Army-owned aircraft. Nonetheless, the Army officers responsible for the development of inspection procedures pertaining to the UH-1 reviewed both the FAA and Bell warnings. R.1-36, Tab P, at 1-2. Reasoning that its own UH-1 helicopters had not been engaged in heavy-lifting operations and that its history of using UH-1 helicopters had been without accident, the Army decided not to adopt the recommended inspection protocols.3 Instead, it adhered to a regimen in which the only aids used during inspections of the fin spar were a flashlight and inspection mirror. R.1-44, Tab U.

After the crash of the accident helicopter, however, the Army departed from its inspection procedures and instituted new protocols reflective of the concerns expressed in the FAA and Bell advisories. Specifically, the Army provided for frequent inspections to be conducted with the aid of a 10X magnifying glass, and for additional periodic inspections to be made via x-ray examination, fluorescent dye penetrant, and tap hammer. Post-accident x-ray inspections of the seven other UH-1 helicopters within the Flat Iron fleet led to the discovery of cracks in the vertical fin spars of four additional aircraft. None of these cracks were visible to the naked eye at the time they were discovered via X-ray.

Although their complaints originally stated numerous additional causes of action, the Hudgens and Crawfords now primarily assert that DynCorp was negligent under Alabama law for failing to properly maintain the helicopter and/or to make necessary repairs.4 After discovery, DynCorp moved for summary judgment on the basis of its asserted entitlement to the "government contractor defense" established in Boyle v. United Technologies Corp., 487 U.S. 500, 108 S. Ct. 2510, 101 L. Ed. 2d 442 (1988). Boyle articulated a three-pronged test for courts to apply when required to decide whether protection of federal policymakers' discretion demands preemption of state tort law imposing liability on contractors for design defects in products supplied to the government.

Despite the identity of all facts implicated by DynCorp's motion for summary judgment as to the Hudgens' and Crawfords' separate suits, the plaintiffs' actions were not consolidated and thus two different district court judges ruled on the motions. Both rejected the plaintiffs' argument that the government contractor defense applies only to design defects, holding instead that the defense extends to contracts of the kind entered into by DynCorp and the Army. Additionally, both judges ruled that the evidentiary materials submitted by the parties, as modified by slightly different rulings striking certain expert opinion evidence, showed no genuine issue of material fact as to DynCorp's satisfaction of the defense's three elements.

On appeal, the Hudgens and Crawfords argue first that Boyle's government contractor defense does not apply to service contracts. Alternatively, they argue that even if the government contractor defense applies, the district court nonetheless erred in striking expert opinion testimony and in holding that DynCorp had demonstrated the absence of any genuine issue of material fact as to its satisfaction of the defense's three elements.

We review the district court's grant of summary judgment de novo, applying the same legal standards as the trial court. Chapman v. AI Transport, 229 F.3d 1012, 1023 (11th Cir.2000) (en banc). We review the district court's exclusion of expert opinion evidence for abuse of discretion. General Electric Co. v. Joiner, 522 U.S. 136, 142, 118 S. Ct. 512, 139 L. Ed. 2d 508 (1997); Allison v. McGhan Med. Corp., 184 F.3d 1300, 1306 (11th Cir.1999).

Summary judgment is appropriate when "there is no genuine issue as to any material fact and ... the moving party is entitled to a judgment as a matter of law." Fed.R.Civ.P. 56(c). DynCorp's motion for summary judgment rested exclusively on its asserted entitlement to the protection of the government contractor defense. See R.1-49. Accordingly, we need only resolve whether the district court erred in ruling that this defense protects DynCorp from the particular claims asserted by the Hudgens and Crawfords.

We begin with the district court's determination that the government contractor defense applies to service contracts like the one between the Army and DynCorp.

In Boyle, the Supreme Court addressed the issue of "when a contractor providing military equipment to the Federal Government can be held liable under state tort law for injury caused by a design defect." Boyle, 487 U.S. at 502, 108 S. Ct. 2510. The Court approached this question as one of whether the Constitution or laws of the United States had "so committed" a relevant matter "to federal control that state law is pre-empted and replaced, where necessary, by federal law of a content prescribed (absent explicit statutory directive) by the courts." Id. at 504, 108 S. Ct. 2510.5 Preemption of this kind is warranted, it held, only when the imposition of liability under state law would create a "significant conflict" with federal policy in an area of "uniquely federal interest." Id. at 507, 108 S. Ct. 2510.

The Court held that one area of uniquely federal interest is the procurement of equipment by the United States. Id. In defining what counts as a "significant conflict," the Court found it necessary to identify a "limiting principle" that would prevent the government contractor defense from interfering unduly with the operation of state law. Id. at 509, 108 S. Ct. 2510. It found such a principle in the notion of a "discretionary function" as that term is used in the Federal Tort Claims Act.6 Id. at 511, 108 S. Ct. 2510. The "selection of the appropriate design for military equipment" is a discretionary function in this sense because it "often involves not merely engineering analysis but judgment as to the balancing of many technical, military, and even social considerations, including specifically the trade-off between greater safety and greater combat effectiveness." Id. Accordingly, the Court held that the enforcement of state tort law against military contractors must be preempted inasmuch as its operation would interfere with the exercise of discretion by government officials charged with making these sensitive policy judgments.

The Hudgens and Crawfords assert that Boyle recognized a government contractor defense only in the case of parties to procurement contracts. They urge that since "[i]mmunity from tort liability for a private party is the exception to the general rule," it ought not be extended to service contracts such as the one between DynCorp and the United States "until Congress or the Supreme Court explicitly act on the precise issue." Hudgens' Initial Brief at 24; Crawfords' Initial Brief at 24. We agree that the common lawmaking power on which the government contractor defense is predicated must not be hastily invoked to limit liability under state tort law. Nonetheless, we think the government contractor defense set forth in Boyle clearly applies in the present case.

Although Boyle referred specifically to procurement contracts, the analysis it requires is not designed to promote all-or-nothing rules regarding different classes of contract. Rather, the question is whether subjecting a contractor to liability under state tort law would create a significant conflict with a unique federal interest. Glassco v. Miller Equipment Co., 966 F.2d 641, 642 (11th Cir.1992). We would be exceedingly hard-pressed to conclude that the unique federal interest recognized in Boyle, as well as the potential for significant conflict with state law, are not likewise manifest in the present case. The formulation of design specifications and the articulation of maintenance protocols involve the exercise of the very same discretion to decide how a military fleet of airworthy craft will be readied. Holding a contractor liable under state law for conscientiously maintaining military aircraft according to specified procedures would threaten government officials' discretion in precisely the same manner as holding contractors liable for departing from design specifications.

We thus hold that the government contractor defense recognized in Boyle is applicable to the service contract between the Army and DynCorp.

In order for DynCorp to have the benefit of this defense, it must establish not only the defense's general applicability to its contract with the Army, but also its satisfaction of three elements set forth by the Supreme Court to ensure that the requisite "significant conflict" between federal policy and state law exists under the particular circumstances of this case. Boyle, 487 U.S. at 512, 108 S. Ct. 2510; Glassco, 966 F.2d at 642 (stating that "three-part inquiry elaborates the `significant conflict' prong of the test"). As the Supreme Court articulated these three conditions, no liability may be imposed for design defects in military equipment if "(1) the United States approved reasonably precise specifications; (2) the equipment conformed to those specifications; and (3) the supplier warned the United States about the dangers in the use of the equipment that were known to the supplier but not to the United States." Id.

The Supreme Court's references to "specifications" reflects the nature of the case before it in Boyle, which involved an alleged defect in the design of a military helicopter's escape hatch. In the context of the present case, we rearticulate the defense's three elements to foreclose liability under state tort law if (1) the United States approved reasonably precise maintenance procedures; (2) DynCorp's performance of maintenance conformed to those procedures; and (3) DynCorp warned the United States about the dangers in reliance on the procedures that were known to DynCorp but not to the United States.

The requirement that prescribed maintenance procedures be reasonably precise is necessary to ensure that a close relationship exists between the contract duty imposed by the federal government and the state law duty that application of the government contractor defense will preempt:

If, for example, the United States contracts for the purchase and installation of an air conditioning-unit, specifying the cooling capacity but not the precise manner of construction, a state law imposing upon the manufacturer of such units a duty of care to include a certain safety feature would not be a duty identical to anything promised the Government, but neither would it be contrary. The contractor could comply with both its contractual obligations and the state-prescribed duty of care. No one suggests that state law would generally be pre-empted in this context.

Boyle, 487 U.S. at 509, 108 S. Ct. 2510. The reasonable precision requirement ensures that the government contractor defense is limited to its proper scope by requiring "that the design feature in question was considered by a Government officer, and not merely the contractor itself." Id. at 512, 108 S. Ct. 2510.

In this case, application of the defense's first prong directs us not to design specifications but to maintenance procedures. Both district judges found the Army's maintenance guidelines to constitute a comprehensive regime that DynCorp was not expected to supplement through any procedures other than those specifically set forth. They also found that the precautions advised by the FAA and Bell had been affirmatively and deliberately omitted from the prescribed maintenance procedures. We agree.

On appeal, the plaintiffs concede that the Army's maintenance procedures were reasonably precise within the meaning of the government contractor defense and that DynCorp was not required to institute any procedures not expressly prescribed by the Army. In Hudgens, the district court succinctly described the high degree of precision reflected in the Army publications setting forth inspection procedures for UH-1 aircraft. The court summarized the Army's "Daily Inspection Checklist" as requiring "over 130 individual inspections, covering eight general areas of the helicopter." See R.1-47 at 6 (internal citation omitted). Mechanics were specifically instructed to inspect the "vertical fin spar and vertical driveshaft cover attachment channel for cracks in the area directly below the 90 degree gearbox attachment fitting." Id. at 7 (citation omitted). The Army's "Phased Maintenance Checklist," prescribing 150 inspections to be completed at certain intervals, required examination of the "[v]ertical fin rib ... along rivet row at fin station 10.08 for cracks (access thru topmost lighting holes)." Id. at 6 (citation omitted). Finally, the Army's Maintenance Instructions, which exceed 1000 pages in length, provided for a complete inspection of the "tailboom" assembly of which the fin spar is part, and gave "specific instructions for repairing a variety of defects in the tailboom assembly, including cracks in the forward spar of the vertical fin." Id. at 7 (citation omitted).

DynCorp produced deposition testimony from Ralph C. Vemmer, an aeronautical engineer employed by the Army division that develops these instructions and manuals. See R.1-36, Tab F. Vemmer answered in the affirmative when asked whether the FAA and Bell advisories were "suggested deviations from the inspection and maintenance which had previously been approved by the Army." Id. at 217. Vemmer also agreed that the precautions "should not have been implemented on Army aircraft until the Army authorized such procedures." Id.

The Hudgens and Crawfords present no evidence contesting Vemmer's assertions and, as mentioned above, they conceded at oral argument that DynCorp had no duty to perform any inspections or repairs not specifically prescribed by the Army's technical manuals. Accordingly, we conclude that the Army's comprehensive publications dictated reasonably precise inspection procedures that did not incorporate the precautions enumerated in the FAA and Bell advisories.

We turn next to the third (rather than the second) element of the government contractor defense simply because it is so easily met in this case. This notice element requires contractors to inform the government of dangers "known to the supplier but not to the United States." Boyle, 487 U.S. at 512, 108 S. Ct. 2510. The plaintiffs conceded at oral argument that DynCorp has satisfied this condition. Their concession is borne out by the record, which demonstrates that the Army was well aware of the danger which the FAA and Bell had associated with a failure to adopt certain precautions not incorporated within its own procedures. In addition to Vemmer's deposition testimony, DynCorp has submitted letters written by an Army spokesperson shortly after the accident acknowledging awareness of the advisories and explaining the decision not to implement the recommended precautions. See R.1-36, Tab O, Tab P. We thus hold the third condition of the government contractor defense to be satisfied with respect to dangers engendered by the non-implementation of the procedures advised by the FAA and Bell.

Our discussion thus far establishes that DynCorp cannot be held liable for failing to institute the procedures recommended by Bell and the FAA. Army instructions were reasonably precise in their omission of these precautionary measures, and the Army was aware of the danger thus created. The Hudgens and Crawfords, however, primarily seek to defeat DynCorp's assertion of the government contractor defense by showing that DynCorp has not satisfied what might be termed the "performance prong," which requires DynCorp to show it carried out inspections in conformance with Army procedures. The plaintiffs argue that the Army's inspection procedures, despite not incorporating the precautions recommended by the FAA and Bell, still required DynCorp to find the spar crack prior to the crash.7 Since DynCorp mechanics never did so, they argue that DynCorp did not conform with the Army's procedures.

As litigation has progressively narrowed the issues involved in this case, the controversy between the parties as to DynCorp's conformance with Army procedures has been reduced to the question of whether the crack in the fin spar was visible to the naked eye at the time of any inspection required before the crash.8 Accordingly, in deciding whether DynCorp can satisfy the government contractor defense's performance prong, we need determine only whether there is a genuine issue of fact as to the crack's visibility to the naked eye.

The question of a crack's visibility is best understood to consist of two component parts. First, was the crack within the ocular capacity of the human eye to detect — that is, was the crack of such a size and nature that it could be seen? Second, was the crack present at a location in the helicopter's internal structure that was within the range of visual access achieved by mechanics carrying out the Army's procedures — that is, would mechanics have looked at the crack in the course of properly conducted inspections?

The plaintiffs point to materials they submitted from three witnesses to support the contention that the crack was visible in both of these senses.9 Each witness stated that the crack could be seen by the naked eye prior to the crash. One of them also specifically described a particular inspection during which mechanics should have looked at the portion of the fin spar where the crack was located.10

In response to a motion by DynCorp, the district court treated these materials as expert opinion evidence, see Fed.R.Evid. 702, and struck portions of the materials which it deemed inadmissible under the test set forth in Daubert v. Merrell Dow Pharmaceuticals, 509 U.S. 579, 113 S. Ct. 2786, 125 L. Ed. 2d 469 (1993), and its progeny. After striking the materials in part, the district court found that the plaintiffs had made no showing that the crack was ever visible to the naked eye prior to the crash. It accordingly determined that there existed no genuine issue of material fact as to DynCorp's entitlement to the protection of the government contractor defense. We review the district court's decision to strike certain portions of the expert opinions and then consider whether its grant of summary judgment is due to be reversed in light of the one portion of an affidavit we find to have been wrongfully excluded.

Since we review a district court's exclusion of expert opinion evidence for abuse of discretion, we will reverse only if its decision was "manifestly erroneous." Joiner, 522 U.S. at 142, 118 S. Ct. 512.

A three-pronged test controls the determination of whether expert opinion evidence is admissible. Admission is proper if "(1) the expert is qualified to testify competently regarding the matters he intends to address; (2) the methodology by which the expert reaches his conclusions is sufficiently reliable as determined by the sort of inquiry mandated in Daubert; and (3) the testimony assists the trier of fact, through the application of scientific, technical, or specialized expertise, to understand the evidence or to determine a fact in issue." City of Tuscaloosa v. Harcros Chem., Inc., 158 F.3d 548, 562 (11th Cir.1998) (footnote omitted).

The first piece of expert opinion evidence at issue appears in the affidavit of B.J. Sammons, an aircraft mechanic employed at Fort Rucker for roughly forty years before retiring on the day before the subject accident. At the time of his retirement, Sammons was the lead mechanic on the fleet of helicopters to which the accident craft belonged. The plaintiffs fault the district court for striking a range of statements from Sammons's affidavit. We find no abuse of discretion except inasmuch as the court struck Sammons's statement pertaining to an October, 1998 inspection that he describes in some detail.11 Specifically, Sammons states that properly instructed DynCorp personnel would have discovered spar cracks on this occasion because Army procedures required "a detailed inspection of all of the components of the tail boom assembly, including the vertical fin spar." R.1-38, Tab 7, at 1-3. The district court's memorandum opinion in Crawford stated no basis for excluding this portion of Sammons's affidavit.

We understand Sammons's use of the word "cracks" to refer generically to cracks of a size and nature that would permit them to be detected by inspectors whose eyesight was adequate. Thus, we read his assertion to concern only the range of visual access achieved by inspectors adhering to the Army's procedures — that is, whether inspectors would have looked at the portion of the fin spar on which the crack was located. On this reading, Sammons's opinion appears simply to apply what the district court acknowleged was his "expertise" in the "maintenance and inspection of UH-1H helicopters." Crawford Memorandum Opinion at 12. In crediting Sammons's expertise, the district court necessarily determined that the first and second prongs of the test for scientific opinion evidence — the witness's qualification as an expert and the reliability of his or her methodology — were satisfied with respect to Sammons's opinion regarding the October, 1998 inspection.

In light of this aspect of the district court's ruling, we believe its exclusion of Sammons' statement was most likely an oversight caused by the affidavit's poor draftsmanship.12 To the extent that the statement's exclusion was not an oversight, the court's determination that Sammons had satisfied the qualification and reliability prongs left exclusion to rest on a determination that the relevance prong was not satisfied. Any such determination would be an abuse of discretion. As this case has developed, the visibility of the spar crack at the time of the October, 1998 inspection is the only disputed question of fact pertinent to DynCorp's satisfaction of the elements of the government contractor defense. The Hudgens and Crawfords have failed to submit any other potentially competent evidence showing that Army inspection procedures would have led to the crack's discovery on any other occasion. Sammons's affidavit specifically identified the October, 1998 inspection as an occasion on which conformance to Army procedures would have allowed mechanics to get a view of the relevant portion of the spar.13 His specificity shows that he took account of the fact that the fin spar cannot be visually examined without opening various inspection plates, panels, or doors. R.1-38, Tab 7, at 1. Hence Sammons' affidavit, while evincing his knowledge that inspectors would not necessarily examine the relevant portion of the fin spar in the course of every maintenance function, nonetheless made clear that the "detailed inspection" called for in October 1998 was one occasion on which they would do so. The relevance of this assertion is simply not contestable. See City of Tuscaloosa, 158 F.3d at 565 (in order to satisfy relevance prong of test for expert opinion evidence, circumstantial evidence "must merely constitute one piece of the puzzle that the plaintiffs endeavor to assemble before the jury"); Allison, 184 F.3d at 1320 (remarking that expert's opinion, when offered as plaintiff's only evidence of causation in products liability action, is "more than a `piece of the puzzle'"). Since the district court's exclusion of Sammons's testimony was manifestly erroneous, the court abused its discretion in striking this portion of Sammons's affidavit.

When the wrongfully excluded evidence is re-integrated into Sammons' affidavit, the relevant passage should read as follows:

With this type repair and maintenance, there should have been a detailed inspection of all of the components of the tail boom assembly, including the vertical fin spar. Cracks would have been discovered during this inspection if the line mechanics had received proper guidance from DynCorp management.

The Hudgens and Crawfords also challenge the district court's striking of the deposition testimony of Steve Powell, a Bell Helicopters field investigator who operated an electron microscope through which he took photographs of fractured parts of the helicopter after the crash. The district court in both cases struck Powell's assertion that the crack in the fin spar would have been visible to the naked eye prior to the crash. As a rationale, the court relied entirely on Powell's admission that he had no knowledge of DynCorp's maintenance procedures. Hudgens Memorandum Opinion at 21; Crawford Memorandum Opinion at 15. This was error in that it confounded what we have identified as the two component parts of the question of visibility. The courts should have considered Powell's testimony not simply as it related to the range of visual access achieved by DynCorp inspectors, but also as to the naked eye's capacity to detect cracks of a certain size in a certain type of metal.

Despite this error, we do not believe the district court abused its discretion in excluding Powell's testimony. The Hudgens and Crawfords have not shown that Powell's experience provided a reliable basis for his opinion regarding the crack's appearance prior to the helicopter crash. The two forms of specialized expertise claimed by Powell involved using an electron microscope and applying the technique of striation counting, which the district court described as "a method used to determine the rate of growth of a crack." Crawford Memorandum Opinion at 15; see also Hudgens Memorandum Opinion at 19. We are unable to conceive of how experience in the use of an electron microscope, in and of itself, could qualify a witness to assess such matters as when a crack first appeared and whether it was visible prior to a crash. And while it is conceivable that a skilled application of striation counting could constitute a reliable means of formulating such an opinion, Powell stated that he performed no striation counting on the spar crack involved in this case. R.1-38, Tab 2, at 35. The Hudgens and Crawfords have therefore failed to identity even an arguably reliable methodology underlying Powell's determination that the crack was visible prior to the crash. Accordingly, although the district court stated an insufficient rationale for excluding Powell's testimony, we hold that such exclusion was not an abuse of discretion.

Richard McSwain is an engineer whose qualification as a materials engineer was not contested before the district court. The district court struck a portion of McSwain's affidavit in which he opined that "the cracks were visible to the naked eye and also would have been detectable by the use of non-destructive testing" at the time of an inspection performed 112 hours prior to the accident. McSwain cited several documents as his basis for reaching this conclusion. He relied on an "Aircraft 150 Hour Phase Inspection Record" for the premise that an inspection was conducted 112 hours prior to the accident.14 For the conclusion that the crack would have been visible and detectable at this time, McSwain relied on the results of a 1999 study conducted by Bell titled "UH-1 Fin Spar Crack Propagation Test Results." R.1-38, Tab 129. This test subjected another fin spar to simulated flight stress in order to study the growth of a crack at virtually the same location as the one which caused the crash of the accident helicopter.

Finding that the test spar was comparable in all relevant respects to the one installed on the accident helicopter, the district court held that "Bell's report on crack propagation ... may be used to extrapolate data" regarding the crack that both parties agree caused the crash underlying this suit. Hudgens Memorandum Opinion at 17; Crawford Memorandum Opinion at 14.15 Accordingly, the district court declined to strike McSwain's opinion as it related to the crack's detectability through non-destructive testing (such as x-rays). It did, however, strike that portion of his affidavit opining that the crack was visible to the naked eye. In Hudgens, the district court stated the following rationale for this portion of its decision: "Because there is no evidence that McSwain has seen a UH-1 aircraft before, there are no facts by which the court may infer that McSwain is competent to testify about whether the cracks would have been detectable before the crash by a visual inspection." Hudgens Memorandum Opinion at 19. In Crawford, the district court stated simply that it found no "factual basis" for McSwain's conclusion regarding the cracks' visibility to the naked eye. Crawford Memorandum Opinion at 14.

In reviewing these determinations for abuse of discretion, we are mindful of the repeated emphasis the Supreme Court has placed upon the district court's "gatekeeping" role in the determination of whether expert evidence should be admitted. See Kumho Tire Co. v. Carmichael, 526 U.S. 137, 152, 119 S. Ct. 1167, 143 L. Ed. 2d 238 (1999) (concluding "that the trial judge must have considerable leeway in deciding in a particular case how to go about determining whether particular expert testimony is reliable"); Joiner, 522 U.S. at 146, 118 S. Ct. 512 (recognizing district court's authority to exclude "opinion evidence that is connected to existing data only by the ipse dixit of the expert," should it conclude that "there is simply too great an analytical gap between the data and the opinion proffered"). Even when a decision to strike expert evidence is outcome determinative, as when it supports a grant of summary judgment, the district court's decision remains subject to deferential review. Id. at 142-43, 118 S. Ct. 512.

We begin by observing that the district court in Hudgens failed to articulate a sufficient rationale for its exclusion of McSwain's opinion regarding the visibility of the crack in the accident spar. The court's stated reason for excluding the testimony appears to have confused the two components we have identified in our discussion of the question of visibility. The fact that McSwain may never have seen a UH-1 does not detract from his potential capacity, as a materials engineer, to render an opinion about whether a crack of a particular size in a particular kind of material could be seen by the naked eye. McSwain's opinion need not be read to concern the range of visual access achieved by mechanics in the course of inspecting UH-1 aircraft. While attorneys for the Hudgens and Crawfords should have had McSwain's affidavit refer to the visibility of the crack "at the time of the inspection" conducted 112 hours prior to the crash, rather than simply "at the inspection," the context of his affidavit as a whole makes clear that it is the former opinion he means to assert. In failing to appreciate this distinction, the district court erred.

The Crawford court's opinion referred, without further elaboration, to the absence of a "factual basis" for McSwain's conclusions regarding the crack's visibility to the naked eye. Hence Crawford does not manifest the error we discern in Hudgens, but nor does it explain precisely why it found McSwain's opinion to lack a factual basis. In both cases, however, the decision to strike McSwain's opinion will not require reversal if we determine that exclusion was required by a proper application of the test for scientific opinion evidence.

We thus turn, for the purpose of determining whether McSwain's opinion arguably rested on a reliable methodology, to the Bell crack propagation study he cited as its sole basis. The Bell study reported a controlled experiment designed to reproduce precisely the same spar failure that all parties agree caused the accident underlying this suit. See R.1-38, Tab 129, at 9 fig. 6 (showing location of fin spar where crack was propagated over course of experiment). The engineers who conducted the test applied a stress load that simulated the conditions under which the spar operates in flight. Id. at 2. The report described cracks on the test spar as follows: "The fin spar flange is fabricated using five metal layers bonded together. The fatigue cracks at hole number 1 stared in the outer layer and progressed through all five layers." Id. These cracks were present even before the commencement of the controlled experiment, which utilized a spar that was received from the Army and had recorded 10,214 flight hours prior to the test. Id. Only after roughly 215 hours of additional simulated flight had the cracks grown to a point that caused the fin spar to rupture. Id. Finally, the study suggests, albeit with some ambiguity, that at least one crack was visible to the naked eye throughout the course of the experiment. See id. (reporting that "[e]ddy current was also used in conjunction with visual means to measure crack lengths on the outer layers").

In sum, the report provided a basis for the opinion that visible cracks could exist on a fin spar without developing into a complete rupture for almost twice as long as the 112 hours for which McSwain opined cracks were visible on the accident helicopter's fin spar. The study does not, however, discuss the frequency with which cracks begin on a surface layer of the multi-layer material of which the relevant portion of the fin spar is fabricated.

DynCorp has submitted materials tending to show that spar cracks often do not begin on the spar's surface layer, and that the crack on the accident craft most likely did not. One document, which DynCorp describes as a 1997 internal Bell document and which the plaintiffs have not challenged on authenticity grounds, states that inspection of the spar is "very difficult" because "visual inspection cannot find cracks in sub-laminates ..." R.1-36, Tab J, at BE018871. The document also notes that the FAA had previously informed Bell "that the laminated design did not permit adequate inspection and was an inadequate design." Id. at BE018873. It then recommends creating "a single piece [left-hand] spar cap to replace the existing laminated spar." Id. at BE018876.

We also find it telling that cracks found after the accident on four other craft within the Fort Rucker fleet were in each instance concealed beneath the surface layer of compromised spars and thus discovered only via the use of X-rays. R.1-44, Tab X (affidavit of DynCorp employee who "supervised the x-ray inspections on all UH-1 aircraft at Ft. Rucker"). None were visible to the naked eye. Id. This pattern, in conjunction with the known problem discussed in the Bell memorandum, seems to us to raise a strong inference that the crack on the accident helicopter was likewise not visible prior to the crash.

In the face of these materials, McSwain's affidavit offers no explanation of how he formed the opinion that the crack on the accident spar differed from those subsequently discovered on the rest of the Fort Rucker fleet. McSwain not only fails to identify the specific facts upon which his opinion was based, but also manifests no awareness of the pattern found in the other helicopters. We thus feel compelled to liken this case to those in which an expert's failure to explain the basis for an important inference mandates exclusion of his or her opinion. See Rider v. Sandoz Pharm. Corp., 295 F.3d 1194, 1202 (11th Cir.2002) (holding expert opinion to fall short of what "Daubert requires" when finding of reliability would require "several scientifically unsupported `leaps of faith'"). To find McSwain's opinion reliable, a district court would have to determine that he had some unstated basis for believing the spar on the accident helicopter to have exhibited the same pattern observed on the one tested by Bell. At the same time, it would have to credit his unstated (and possibly uninformed) conclusion that the accident helicopter differed in this respect from the very aircraft which its use over time most closely resembled: the other helicopters within the same Fort Rucker fleet. Evaluating the soundness of these aspects of McSwain's opinion clearly requires weighing probabilities. Yet McSwain has neither explained the likelihood he assigns to each of the unknowns nor spoken with the sort of "precision and logic" that would allow us to assess the relationship between his experience as a materials engineer and his opinion in this case. United States v. Frazier, 322 F.3d 1262, 1267 (11th Cir.2003). We therefore hold that his opinion regarding the visibility of the crack at the time of an inspection conducted 112 hours before the accident was properly excluded by the district court.

Thus, it only remains for us to review whether the district court was correct in determining that the parties' evidentiary submissions did not reveal any genuine issue of material fact. Our analysis of this matter will take into account the portion of Sammons's affidavit wrongfully struck in Crawford.

In determining whether an issue of fact is "genuine" for the purpose of defeating summary judgment, we ask whether the evidence is "such that a reasonable jury could return a verdict for the nonmoving party." Anderson v. Liberty Lobby, 477 U.S. 242, 248, 106 S. Ct. 2505, 91 L. Ed. 2d 202 (1986); Chapman, 229 F.3d at 1023. In making this determination, we view the evidence in the light most favorable to the nonmovant, drawing all reasonable inferences in that party's favor. Augusta Iron and Steel Works v. Employers Ins. of Wausau, 835 F.2d 855, 856 (11th Cir. 1988). Here, these standards must be applied to evaluate DynCorp's contention that no crack was visible to mechanics maintaining the accident helicopter in conformance with the Army's inspection procedures. We believe DynCorp has carried this burden.

Although the wrongfully stricken portion of Sammons's affidavit does tend to show at least one occasion on which DynCorp mechanics should have visually examined the relevant portion of the accident helicopter's fin spar, DynCorp has submitted a range of materials suggesting that the crack at this location was never visible to the naked eye. In addition to the the Bell memorandum and the description of the post-accident x-rays conducted on the remaining Fort Rucker fleet, it has submitted an affidavit from the mechanic who conducted the very inspection discussed in Sammons's affidavit, which states that no cracks were visible in the fin spar at that time. R.1-44, Tab U, at 1. The Hudgens and Crawfords, by contrast, have not presented any admissible materials to support their contention that the crack on the accident helicopter was visible at any point prior to the crash.

In sum, all the evidentiary materials deserving of consideration suggest that no crack could be seen by the naked eye as of October 1998, the only occasion as to which there is any indication that mechanics following Army procedures should have looked at the relevant portion of the fin spar. Since the crack's visibility remains the only disputed issue of fact pertinent to DynCorp's conformance with Army procedures, we conclude that a reasonable jury would have to find a preponderance of the evidence to demonstrate that DynCorp has satisfied this "performance" prong of the government contractor defense.

In light of our earlier discussion of the other two elements of the government contractor defense, we hold that DynCorp has demonstrated the absence of any genuine issue of material fact regarding its entitlement to the defense's protection in this case.

Having held, first, that the government contractor defense applies to the Army-DynCorp maintenance contract and, second, that DynCorp has demonstrated the absence of any genuine issue of material fact as to its satisfaction of the defense's three elements, we AFFIRM the summary judgment entered on DynCorp's behalf in the district court.

AFFIRMED.

I concur in the judgment, and I join in Judge Barkett's opinion except for Part IIC.

I concur in the judgment, and I join Judge Barkett's opinion except for Part IIC.